KCNK9 could be a prognostic biomarker in uterine corpus endometrial carcinoma

Introduction

The most prevalent gynecologic cancer in the world is uterine corpus endometrial carcinoma (UCEC). Epidemiological trends are alarming: 69,120 new cases and 13,860 deaths are anticipated in the US in 2025. Two enduring negative trends define the disease: a consistent increase in both incidence (about 1% yearly since the mid-2000s) and mortality (1.5% annually from 2013 to 2022). Over the previous 40 years, 5-year survival rates have decreased. A developing and pressing public health issue is highlighted by the combination of rising burden and declining results (1). The mainstay of treatment for endometrial cancer is still surgery, which is usually complemented with biological treatments, endocrine therapy, chemotherapy, and radiotherapy. International treatment guidelines state that external beam radiation therapy, brachytherapy, or surgery can cure individuals with endometrial cancer in its early stages. However, a poor prognosis is often the result of restricted treatment options and the advancement of the disease in cases of recurrent or metastatic cancer, whether local or distant (2). Therefore, improving recovery rates requires early detection of UCEC and timely treatment. Effective biomarkers are therefore desperately needed to support UCEC prognostic assessment, early diagnosis, and risk prediction.

The two-pore domain potassium (K2P) channel KCNK9 (TASK-3) is an essential leak conductance channel for sustaining resting membrane potential and cellular excitability (3). Its deregulation has been identified as an oncogenic driver in various cancer types (4,5). By activating important survival pathways like phosphatidylinositol 3-kinase/protein kinase B (PI3K/AKT), blocking apoptosis by altering BCL-2 family proteins, and promoting the epithelial-mesenchymal transition (EMT) by upregulating SNAIL/SLUG transcription factors, KCNK9-mediated potassium efflux can mechanistically accelerate the growth of tumors (6,7). In malignancies such as breast cancer, ovarian cancer, and prostate cancer, increased KCNK9 expression is clinically linked to advanced disease stage, treatment resistance, and a poor prognosis (8-10). A study in ovarian cancer has shown that increased KCNK9 (TASK-3) expression improves cell survival, confers resistance to apoptosis, and contributes to chemoresistance through pH sensing and hypoxia adaptation (10).

Recent evidence shows that KCNK9 serves as a target for anesthetics in ovarian cancer, with lidocaine inhibiting and sevoflurane enhancing tumorigenesis, underscoring its role in malignant behavior (11). Evidence from hormonally linked malignancies suggests that KCNK9 may alter UCEC progression and responsiveness to therapy, requiring more prognostic and therapeutic studies. UCEC’s clinical importance, prognostic usefulness, and TME impact are unknown. This study examined KCNK9’s prognostic value as an independent UCEC survival predictor. We examined its interaction with the tumor microenvironment (TME) because the immune landscape affects treatment results. This could improve risk stratification models and help select patients for immunotherapy or combination medicines targeting ion channels and immunological checkpoints.

To address these questions, our study had two primary objectives: first, to comprehensively evaluate the prognostic significance of KCNK9 and determine whether it is an independent survival predictor in UCEC; second, to investigate its association with the immune landscape within the TME, a key determinant of therapeutic outcome. These goals are clinically relevant: they may enhance prognostic models for better risk stratification and guide future research into patient selection for targeted combination therapies. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-2087/rc).

Methods

Data collection

The Cancer Genome Atlas (TCGA) supplied clinical data and RNA-seq gene expression profiles for UCEC. The gene expression profiles came from 553 UCEC tumor samples and 35 normal endometrial samples. This cohort provided clinicopathological data on patient survival, duration, age, and tumor grade for our study. Nevertheless, staging data were unavailable. The work adhered to TCGA publication rules, therefore negating the necessity for ethical approval. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. For the validation group involving patient-derived tissues for immunohistochemistry (IHC), this study was reviewed and approved by the Ethics Committee of Shanghai First Maternity and Infant Hospital (approval No. KS23112). Written informed consent was obtained from all participants. Figure 1 illustrates the flowchart of the investigation.

Figure 1 This study’s flowchart. CRISPR, clustered regularly interspaced short palindromic repeats; eQTL, expression quantitative trait loci; GEO, Gene Expression Omnibus; GO, Gene Ontology; GSEA, gene set enrichment analysis; GTEx, Genotype-Tissue Expression; IHC, immunohistochemistry; KEGG, Kyoto Encyclopedia of Genes and Genomes; TCGA, The Cancer Genome Atlas; UCEC, uterine corpus endometrial carcinoma.

Identification of differential expression of KCNK9 gene

We examined pan-cancer data from UCSC Xena, which included 10,753 samples from 33 different cancer types, in order to first investigate KCNK9 expression across malignancies. The Wilcoxon signed-rank test was used to analyze the differential expression of KCNK9 across malignancies; statistical significance was defined as absolute log₂ fold-change (|log₂FC|) >1 and false discovery rate (FDR) <0.05. Visual representations of KCNK9 expression patterns across various cancer types were created using the “ggpubr” R package.

The TCGA-UCEC cohort (553 cancers, 35 normals) was then our main focus. The Wilcoxon test was used to evaluate KCNK9 transcriptome differences (Benjamini-Hochberg adjusted q-value <0.05). The study used the Kruskal-Wallis and Mann-Whitney U tests, which were suitable for the data distribution, to investigate the relationships between KCNK9 expression and clinicopathological characteristics, such as tumor grade.

The Gene Expression Omnibus (GEO) GSE17025 microarray dataset, with 12 normal controls and 79 UCEC samples, was used for independent validation. Raw CEL files were preprocessed using the “affy” package for background correction, quantile normalization, and log₂ transformation. ‘limma’ uses empirical Bayes moderation to improve variance estimates and was used for differential expression analysis. Box plots showed variation in KCNK9 expression between UCEC and normal samples.

IHC

We validated our bioinformatics findings at the protein level through IHC. Samples of adjacent tissue and endometrial cancer were collected from patients diagnosed at Shanghai First Maternity and Infant Health Care Hospital. Formaldehyde was used to fix the samples, followed by citrate buffer for antigen retrieval, blocking, and an incubation period of 1.5 hours at 22 ℃ with primary antibodies against KCNK9 (Product # SAB-55557) at a dilution of 1:100. After that, let the HRP-conjugated anti-rabbit antibody sit for 45 minutes. A semiquantitative evaluation of the staining was carried out by three different researchers, who assessed the number of positive cells as well as the staining intensity.

Independent prognostic analysis

Patients in the TCGA-UCEC cohort were classified into high and low KCNK9 mRNA expression groups by median. We used univariate and multivariate Cox proportional hazards regression to evaluate prognostic variables. Age (continuous), tumor grade (categorical), and KCNK9 expression status were considered. The P value criterion for statistical significance was less than 0.05, and hazards ratio (HRs) and 95% confidence intervals (CIs) were generated. We identified KCNK9’s independent effect by controlling for age and tumor grade in the multivariate model. There was no significant multicollinearity across predictors [variance inflation factor (VIF) <2.0] and the proportional hazards assumption was fulfilled (Schoenfeld residuals test, P>0.05 for all covariates). Results were shown in forest plots.

Nomogram and calibration

A prediction nomogram was developed to assess individual patient survival probabilities. Employ the visualization curve generated using R language (version 4.2.2) to assess patients’ overall survival (OS) probabilities at 1, 3, and 5 years.

The samples were classified into two groups based on KCNK9 expression levels: high and low, to evaluate potential differences in progression-free survival (PFS). Univariate and multivariate Cox regression analyses were conducted to determine if KCNK9 expression offers predictive information independent of other variables. Nomograms were created based on the outcomes of UVCRAs and MVCRAs to predict 1-, 3-, and 5-year PFS in patients.

Analysis of receiver operating characteristic (ROC)

The diagnostic and temporal prognostic efficacy of KCNK9 was assessed using ROC curve analysis. The R “survival ROC” package was employed to calculate the area under the curve (AUC) for differentiating tumor tissue from normal tissue and for predicting survival at 1, 3, and 5 years.

Gene co-expressed and differentially expressed genes (DEGs) analysis

We performed co-expression analysis on the TCGA-UCEC transcriptome using Spearman’s rank correlation to elucidate the functional network associated with KCNK9. Gene pairs demonstrating moderate to strong relationships (absolute correlation coefficient |ρ| >0.3) with statistical significance (FDR <0.05) were retained. We classified our tumor samples according to high and low KCNK9 gene expression for the purpose of differential expression profiling. DEGs with a |log₂FC| greater than 0.5 and an FDR less than 0.05 were considered statistically significant. The top 95 upregulated and downregulated DEGs were systematically selected based on ascending FDR values to emphasize biologically relevant targets, followed by hierarchical clustering visualization using the “ComplexHeatmap” program.

Gene Ontology (GO) and KEGG pathway enrichment analyses were performed on the identified DEGs. Gene set enrichment analysis (GSEA) was used to identify pathways associated with different KCNK9 expression levels. We used R packages for GO analysis to examine enriched terms in biological processes (BPs), molecular functions (MFs), and cellular components (CCs). GSEA was run with RNA-seq data and the KEGG gene set collection, using 1,000 permutations. Pathways with an adjusted P value <0.05 were considered significantly enriched.

Immune infiltrates analysis

The CIBERSORT algorithm was utilized to examine the association between KCNK9 and the tumor immune microenvironment. CIBERSORT employs a predefined signature matrix (LM22) to decompose the gene expression profile of a tissue sample, thereby determining the relative proportions of 22 distinct immune cell types within the sample.

Statistical analysis

We used R to analyze TCGA database statistics data. KCNK9 expression was compared between tumor and surrounding non-tumor tissues using t-test. KCNK9 expression and clinical pathology were examined using Wilcoxon and Kruskal-Wallis tests. KCNK9 expression and patient survival were assessed using Kaplan-Meier and log-rank tests. KCNK9’s endometrial cancer prediction was assessed using univariate and multivariate Cox regression. P<0.05 was significant. Using CIBERSORT, KCNK9 expression and immune cell infiltration were examined. We did three IHC investigations. Research data is reported as mean ± standard deviation (SD), and statistical comparisons are made using Student’s t-test in GraphPad Prism 8.

Supplementary multi-omics analyses

Three further bioinformatics analyses examined causation and regulatory mechanisms, supporting KCNK9’s participation in UCEC beyond correlation.

Mendelian randomization (MR) analysis

The MR Analysis used the Genotype-Tissue Expression (GTEx) expression quantitative trait loci (eQTL) Catalogue API to find single nucleotide polymorphisms (SNPs) that correlated significantly (P<5×10⁻⁸) with KCNK9 gene expression levels. To ensure independence, instrumental variables were aggregated using the European population reference panel (r²<0.001, distance =10,000 kb). The IEU OpenGWAS database provided UCEC risk outcome data (ukb-b-13545). Alleles for exposure and outcome were harmonized. The inverse-variance weighted (IVW) approach was used to estimate the principal causal influence. MR-Egger, weighted median, and weighted mode estimators were used in sensitivity analyses; the MR-Egger intercept test was used to measure horizontal pleiotropy; and Cochran’s Q statistic was used to examine heterogeneity.

Clustered regularly interspaced short palindromic repeats (CRISPR) dependency analysis

We focused on UCEC models and analyzed genome-wide CRISPR knockout screens across cancer cell lines using KCNK9 gene dependency scores (CERES, which range from 0 to 1) (DepMap Public 25Q3 release). The Wilcoxon rank-sum test and the independent two-sample t-test were used to statistically compare UCEC and non-UCEC cell lines.

DNA methylation-expression correlation analysis

We examined paired RNA-seq data for 197 UCEC samples from TCGA and DNA methylation data from the Illumina HumanMethylation450 platform in order to investigate the epigenetic regulation of KCNK9. 1,500 bp upstream to 500 bp downstream of the transcription start site (TSS) was identified as the KCNK9 promoter region. KCNK9 expression levels {log₂[transcripts per million (TPM) +1]} were associated with methylation β-values for cytosine-phosphate-guanine (CpG) sites in this area using Pearson correlation. A two-tailed t-test was used to determine statistical significance, and the Benjamini-Hochberg FDR method was used to adjust P values for multiple testing.

Publicly accessible datasets served as the foundation for all additional analysis. Appendix 1 and related Figures S1-S3 offer thorough explanations of data sources, particular methodology, quality control protocols, and thorough results.

Results

Pan-carcinoma analysis of KCNK9 mRNA expression levels

A unique expression pattern for KCNK9 was identified by our pan-cancer examination of TCGA data. The different levels of KCNK9 expression in tumor tissues relative to surrounding non-tumor tissues are shown in Figure 2A. KCNK9 was significantly upregulated in the great majority of the 33 cancer types analyzed, whereas it was downregulated in glioblastoma multiforme. This extensive overexpression points to a possible broad carcinogenic function for KCNK9 in a variety of cancers.

Figure 2 Expression landscape and prognosis value of KCNK9 in UCEC. (A) Pan-carcinomar analysis of KCNK9 mRNA expression levels in multiple tumors, the expression of KCNK9 was generally upregulated compared to adjacent tissues. (B) Differential KCNK9 mRNA expression between UCEC and normal tissues. KCNK9 exhibited elevated expression in UCEC when compared to normal endometrial tissues. (C) Differential expression analysis of KCNK9 in validation cohort GSE17025. KCNK9 exhibited elevated expression in UCEC. (D) Evaluation of KCNK9 expression in UCEC tissues and normal adjacent tissues. KCNK9 protein levels were markedly increased in UCEC tissues relative to normal adjacent tissues; a T-test was employed for the statistical analysis. (E,F) Representative images of KCNK9 expression in UCEC tissues and paracancerous tissues. Immunohistochemical techniques were utilized to assess KCNK9 expression (scale bar =100 µm). (G) Association analysis of KCNK9 expression with pathological grade. (H) Relationship between KCNK9 expression levels and 20-year survival rate of patients with UCEC. (I) Univariate and (J) multivariate independent prognostic analysis of independent risk factors for OS in individuals with UCEC. Univariate and multivariate in the TCGA dataset. (K) Development and assessment of the prognostic nomogram. A personalized OS prediction model was constructed for UCEC patients, incorporating age, grade, and KCNK9 expression levels to forecast OS at 1, 3, and 5 years. (L) The nomogram was generated to predict 1-, 3-, and 5-year OS based on risk levels and relevant variables. Calibration plots for OS prediction at 1, 3, and 5 years demonstrated the nomogram’s accuracy, showing strong alignment with the observed 1- and 3-year OS values. *, P<0.05; **, P<0.01; ***, P<0.0001. ACC, adrenocortical carcinoma; BLCA, bladder urothelial carcinoma; BRCA, breast invasive carcinoma; CESC, cervical squamous cell carcinoma and endocervical adenocarcinoma; CHOL, cholangiocarcinoma; CI, confidence interval; COAD, colon adenocarcinoma; DLBC, diffuse large B-cell lymphoma; ESCA, esophageal carcinoma; HNSC, head and neck squamous cell carcinoma; HR, hazard ratio; KICH, kidney chromophobe; KIRC, kidney renal clear cell carcinoma; KIRP, kidney renal papillary cell carcinoma; LAML, acute myeloid leukemia; LGG, lower-grade glioma; LIHC, liver hepatocellular carcinoma; LUAD, lung adenocarcinoma; LUSC, lung squamous cell carcinoma; MESO, mesothelioma; OS, overall survival; OV, ovarian serous cystadenocarcinoma; PCPG, pheochromocytoma and paraganglioma; READ, rectal adenocarcinoma; SARC, sarcoma; SKCM, skin cutaneous melanoma; STAD, stomach adenocarcinoma; TGCT, testicular germ cell tumors; THCA, thyroid carcinoma; THYM, thymoma; UCEC, uterine corpus endometrial carcinoma; UCS, uterine carcinosarcoma; UVM, uveal melanoma.

KCNK9 is highly expressed in UCEC

KCNK9 expression was significantly elevated in UCEC, consistent with the pan-cancer trend. Analysis of differential expression of KCNK9 between 35 normal endometrial samples and 553 UCEC cases from TCGA demonstrated considerable overexpression in malignant tissues (Figure 2B). The independent GSE17025 dataset confirmed this observation, revealing that KCNK9 mRNA showed a trend of overexpression (log₂FC=0.48, P=0.049) (Figure 2C). Immunohistochemical analysis of 45 matched tumor and adjacent-normal tissue slices demonstrated a significant increase in KCNK9 protein expression in UCECs (P<0.05) (Figure 2D-2F). Analysis of clinicopathological correlation demonstrated a substantial positive association between KCNK9 transcript levels and histological tumor grade. KCNK9 expression was observed to increase gradually with higher tumor grade (Figure 2G), hence supporting the association between raised KCNK9 levels and more aggressive disease features. The findings demonstrate a correlation between KCNK9 expression and advanced tumor stages.

KCNK9 is an independent predictor of unfavorable outcomes

Based on KCNK9 expression levels, TCGA-UCEC patients were sorted into two groups: high-expression and low-expression, using the median expression level as the criterion. OS rates were compared between these two groups. Kaplan-Meier analysis revealed that the group with high KCNK9 expression had significantly poorer OS than the group with low KCNK9 expression (Figure 2H).

To assess its independent prognostic value, we performed Cox regression analyses. In univariate analysis, KCNK9 (HR =1.171, 95% CI: 1.110–1.236, P=7e−09), age, and grade were strongly correlated with OS (Figure 2I). Critically, after adjusting for age and grade in the multivariate model, high KCNK9 expression remained a significant independent predictor of poor OS (HR =1.141, 95% CI: 1.063–1.223, P=0.0002; Figure 2J, Table 1).

The link between KCNK9 and OS remained statistically significant after adjusting for known clinical variables, confirming KCNK9 as an independent prognostic indicator for OS. We constructed a nomogram using age, grade, and KCNK9 expression to estimate 1-, 3-, and 5-year OS for UCEC patients (Figure 2K). Calibration plots showed close agreement between predicted and actual 1- and 3-year OS rates, validating the nomogram’s accuracy (Figure 2L). These analyses indicate that KCNK9 serves as an independent prognostic marker for UCEC.

Prognostic value of the KCNK9 score was independent of clinicopathological factors for PFS

The abscissa indicates the survival time of UCEC patients, while the ordinate represents their survival rate. According to Kaplan-Meier analysis, PFS was significantly greater in the group with low KCNK9 expression than in the group with high expression (Figure 3A). We then employed univariate and multivariate Cox regression to evaluate KCNK9 as an independent predictor of PFS. Due to the incomplete availability of tumor stage data, the models included tumor grade as the key clinicopathological covariate alongside KCNK9 expression. The findings suggested that KCNK9 (HR =1.160, 95% CI: 1.096–1.228, P=3e−07) and grade (HR =2.368, 95% CI: 1.647–3.405, P=3e−06) were markedly linked to PFS in the UVCRAs (Figure 3B). Additionally, in MVCRAs, KCNK9 (HR =1.140, 95% CI: 1.064–1.222, P=0.0002) and grade (HR =1.590, 95% CI: 1.227–2.059, P=0.0004) remained markedly associated with patient PFS (Figure 3C). Specific details are depicted in Table 2. Thus, KCNK9 expression level is an independent prognostic factor for PFS in UCEC.

Figure 3 KCNK9 expression is predictive of an unfavorable prognosis in UCEC patients. (A) PFS analysis of KCNK9 in UCEC patients. High-expressed KCNK9 is associated with poor PFS. (B) Univariate and (C) multivariate Cox regression indicated that KCNK9 was associated with prognosis in UCEC. (D) Development and verification of the predictive nomogram. Predictive nomogram for estimating PFS likelihood of patients at 1, 3, and 5 years. (E) These are calibration plots of nomograms for estimating the likelihood of 1-, 3-, and 5-year PFS in the TCGA cohort. (F) Evaluation of KCNK9 expression’s prognostic and diagnostic capabilities in UCEC. ROC examination was conducted to explore the diagnostic significance of KCNK9 expression levels. (G) ROC graph and AUC values for the predictive indicator at 1-, 3-, and 5-year survival. **, P<0.01. AUC, area under the curve; CI, confidence interval; HR, hazard ratio; PFS, progression-free survival; ROC, receiver operating characteristic; UCEC, uterine corpus endometrial carcinoma.

To facilitate clinical translation, we developed a nomogram that integrates KCNK9 expression with patient age and tumor grade to estimate individualized PFS probabilities at 1, 3, and 5 years (Figure 3D). The calibration plot indicated a high level of agreement between the actual and predicted outcomes for 1-, 3-, and 5-year PFS (Figure 3E). These data indicate that the composite model incorporating KCNK9 expression, age, and grade provided better PFS prediction than any single factor alone.

KCNK9 is a diagnostic marker in UCEC

Time-dependent ROC analysis was used to further evaluate the clinical value of KCNK9 expression in UCEC. With an AUC of 0.714 at diagnosis, KCNK9 showed a moderate capacity for discrimination (Figure 3F). Figure 3G shows survival prediction AUC values of 0.630, 0.622, and 0.716 at 1, 3, and 5 years. Prognostic accuracy improved to 0.716 at 5 years, suggesting KCNK9 may predict long-term survival. Moderate AUCs at 1 and 3 years suggest model modification or new parameters to improve short-term prognostication.

Analysis of gene co-expressed and DEGs with KCNK9 in UCEC

The co-expression study of KCNK9 mRNA profiles found 95 genes significantly connected with KCNK9, including 89 positively and 6 negatively correlated targets, with unique expression patterns in KCNK9-high and -low subgroups. Hierarchical clustering heatmap analysis showed significant variability in co-expressed genes between cohorts after differential expression analysis (Figure 4A). An abundance of KCNK9-high tumor genes distinguished the upper cluster from the lower cluster, which had low expression markers.

Figure 4 Functional analysis and KEGG pathway analysis of DEGs and Profiling tumor-infiltrating immune cells with CIBERSORT in UCEC. (A) Co-expressed and DEGs of KCNK9. Heatmap of the top 95 DEGs. (B) Circle plot showing GO enrichment analysis of DEGs. Functional annotation of DEGs between the high- and low-risk cohorts using GO enrichment analysis. (C) Bar plot and (D) bubble plot depicting markedly enriched GO pathways. The color of the bubbles denotes the P value, and their size denotes the number of genes in the relevant pathway. KEGG pathway functional enrichment results for DEGs between the high- and low-risk cohorts using KEGG enrichment analysis. (E) Bar graph, (F) bubble graph. (G) GSEA pathway enrichment analysis: the top five pathways from the GSEA enrichment analysis. Tumor-related signaling pathways were enriched in both the high- and low-risk cohorts. (H) In patients with UCEC, KCNK9 is correlated with infiltrating immune cells. DEGs, differentially expressed genes; GO, Gene Ontology; GSEA, gene set enrichment analysis; KEGG, Kyoto Encyclopedia of Genes and Genomes; UCEC, uterine corpus endometrial carcinoma.

The DEGs functional enrichment analysis found one major BP, 22 CCs, 22 MFs, and 14 KEGG pathways. Figure 4B shows DEG GO enrichment landscapes with three functional groups within ontological categories. BP showed a substantial increase in amino acid neurotransmitter reuptake, indicating neuronal transmission dysregulation.

CC annotations were mostly found in neuronal cell body microdomains, suggesting UCEC may be caused by ectopic neuronal signaling.MF profiling revealed that metal ion transmembrane transporter activity was the most significantly enriched word, possibly indicating disrupted ion homeostasis in KCNK9-high tumors (Figure 4C,4D).

Two mechanistically different but complementary pathways were identified by KEGG pathway analysis: the Oxytocin signaling route and Axon guidance (Figure 4E,4F). These pathways have both been linked to cytoskeletal remodeling and hormone-responsive tumor growth in the past. Together, these multi-modal enrichment patterns imply that deregulation of the KCNK9-mediated pathway may cause UCEC by activating neuroendocrine-like signaling and potentiating ionic imbalance. Five important processes linked to KCNK9 expression were found using GSEA. High KCNK9 expression was favorably associated with tumor-promoting pathways such as cardiac muscle contraction and neuroactive ligand-receptor interaction. On the other hand, there was an inverse link between tumor-suppressive pathways, such as ribosome biogenesis and xenobiotic metabolism (Figure 4G).

The expression of KCNK9 is linked to the levels of immune cell infiltration in UCEC

According to recent research, the density of lymphocytes invading tumor tissue can predict improved overall and disease-free survival in women with endometrial cancer (12). We used CIBERSORT to measure immune cell infiltration in order to clarify the possible immunomodulatory function of KCNK9. Strong positive associations with activated dendritic cells and M1-polarized macrophages were found to be associated with elevated KCNK9 expression, while inverse relationships were found with CD8⁺ cytotoxic T lymphocytes, regulatory T cells (Tregs), neutrophils, and resting dendritic cells (Figure 4H). This dichotomous infiltration pattern implies that KCNK9 overexpression may paradoxically increase antigen-presenting cell activation while simultaneously promoting an immunosuppressive milieu through T cell exhaustion, possibly reflecting compensatory anti-tumor immune responses.

Supplementary multi-omics analyses support a causal and functional role for KCNK9 in UCEC

To go beyond the correlation and find stronger proof of KCNK9’s involvement in UCEC, we did extra analyses.

MR suggests a potential causal link

MR analysis using three distinct genetic instruments indicated that genetically predicted higher KCNK9 expression is associated with increased UCEC risk [IVW odds ratio (OR) =1.0096 per SD increase, 95% CI: 1.0079–1.0113, P=1.05×10⁻²⁹]. Sensitivity analyses showed consistent estimates, with no significant pleiotropy or heterogeneity (Figure S1, Tables S1,S2).

CRISPR dependency reveals context-specific relevance

Analysis of DepMap data across 1,186 cancer cell lines (including 20 UCEC lines) showed that while KCNK9 is not a pan-essential gene, UCEC cell lines exhibited a significantly lower mean dependency score compared to non-UCEC lines (0.029±0.0229 vs. 0.0485±0.0654; t-test P=0.002). However, heterogeneity was observed among UCEC models, with 40% (8/20) showing dependency scores above the median of non-UCEC cancers, suggesting context-specific functional relevance in a subset of UCEC (Figure S2, Tables S3,S4).

DNA methylation analysis reveals epigenetic regulation

DNA methylation and expression data from 197 UCEC samples showed strong negative associations between KCNK9 promoter methylation and expression. Sixteen CpG sites in the promoter region showed significant relationships after FDR correction (FDR <0.05). The cg25831071 locus showed a significant connection (Pearson r=−0.331, P=2.01×10⁻⁶, FDR =2.01×10⁻⁴) (Figure S3, Table S5). This inverse connection shows that DNA methylation may affect KCNK9 transcriptional control in UCEC, adding mechanistic insight beyond genetic associations.

The additional studies offer genetic, functional genomic, and epigenetic evidence indicating that KCNK9 may contribute to the pathogenesis of UCEC and exhibit context-dependent effects.

Discussion

The association of KCNK9 with survival outcomes in multivariate Cox models, following adjustments for age and tumor grade, indicates its potential as an independent prognostic marker. The time-dependent AUC values for OS at 1, 3, and 5 years (0.630, 0.622, 0.716) indicate moderate predictive accuracy, especially concerning long-term outcomes. Although the TCGA cohort shows promise, independent prospective validation is necessary to establish clinical utility. The increasing prevalence of UCEC highlights the necessity for innovative biomarkers and treatment approaches. This study examines the knowledge gap concerning the potassium channel KCNK9 in UCEC. We utilized MR, CRISPR dependency assessment, and DNA methylation-expression analysis to advance from correlation to the investigation of causality and regulation. Utilizing established bioinformatics methods, our integrative approach, which combines transcriptomic, proteomic, and clinicopathological data, offers a comprehensive foundation (13).

Compared to single-platform analysis, pan-cancer expression profiling, immune microenvironment deconvolution, and experimental protein-level validation offer more proof. By connecting molecular findings to clinical implications, KCNK9’s dual function as an immune regulator and oncogenic driver was clarified using standard bioinformatics techniques in a hypothesis-driven framework.

Machine learning and deep learning have accelerated biomarker development by providing reliable analysis of high-dimensional data. We used conventional bioinformatics tools rather than advanced machine learning algorithms, but our integrative framework—combining transcriptomic, proteomic, and clinicopathological data—provides a foundation for future machine learning predictive model research.

The study found that UCEC consistently and significantly upregulates KCNK9. IHC on patient-derived tissue samples, transcriptome analysis in the TCGA discovery set, and an independent GEO cohort (GSE17025) supported this increase. High KCNK9 expression is linked to lower survival and greater tumor grade. Researchers found that KCNK9 is a reliable independent predictive biomarker. This strategy could improve prognosis when paired with recent risk-stratification methods like the molecular categorization-based European Society of Gynaecological Oncology-European Society for Radiotherapy & Oncology-European Society of Pathology (ESGO-ESTRO-ESP) guidelines (14). KCNK9 expression may help choose adjuvant therapy in the heterogeneous “no specific molecular profile” NSMP category, where hormone receptor status and histologic grade determine risk. Elevated KCNK9 levels indicate aggressive tumor biology, especially in recurrence, patients may need more surveillance or timely systemic therapies such as platinum-based chemotherapy or poly (ADP-ribose) polymerase (PARP) inhibitors. Using KCNK9 expression, age, and tumor grade, the nomogram predicts survival probability at 1, 3, and 5 years with AUC values between 0.630 and 0.716. Since KCNK9 has an AUC of 0.714, it may be a non-invasive biomarker for liquid biopsies. This may supplement invasive procedures like endometrial sampling. KCNK9’s progressive improvement in prognosis accuracy, shown by an AUC of 0.716 after 5 years, can drive individualized follow-up measures.

After controlling for age and tumor grade, increased KCNK9 expression was associated with lower overall and PFS (HR =1.141, 95% CI: 1.063–1.223, P=0.0002). Advanced MR analysis, beyond mere correlation, uncovered preliminary genetic evidence suggesting a potential causal link between genetically predicted elevated KCNK9 expression and heightened UCEC risk (IVW OR =1.0096, P=1.05×10⁻²⁹). Potassium channels, including KCNK9, are attractive pharmaceutical targets because of their accessibility from the extracellular environment (15). Using the integrative method and looking at genetic (MR) evidence points to KCNK9 as a good gene to study more as a possible candidate for therapy. KCNK9 is found to be a possible independent prognostic biomarker and a good option for future studies in drug development. While our study provides strong associative and multi-omics evidence linking KCNK9 to poor prognosis and an immunosuppressive microenvironment, definitive establishment of KCNK9 as a bona fide therapeutic target requires direct experimental validation of its tumor-promoting functions in UCEC models. Our findings, including the potential causal link suggested by MR and the context-specific dependency revealed by CRISPR analysis, provide a strong rationale and a clear hypothesis to guide these essential future functional studies.

As a potassium channel, KCNK9 is well-positioned to influence both intrinsic tumor cell fitness and the extrinsic immune environment. The observed aggressive phenotype and immunosuppressive TME may be caused by KCNK9-induced changes in the ionic environment.

KCNK9, as a potassium channel, is likely to affect intrinsic tumor cell fitness and the extrinsic immune environment. KCNK9-mediated potassium efflux influences intracellular ion homeostasis and membrane potential, impacting cell proliferation, migration, and the secretion of factors that modify the TME (16). These ionic shifts may alter the electrochemical gradient within the TME, thereby influencing immune cell activity (17). This theory, focused on ion channel-mediated alterations of the tumor’s electrochemical and immunological environment, links our observational and genetic data to a possible biological cause in UCEC and warrants examination in forthcoming mechanistic research.

CIBERSORT analysis indicated that elevated KCNK9 expression correlates with an immunosuppressive TME, marked by reduced cytotoxic CD8+ T cell infiltration (r=−0.32, P<0.001) and heightened Tregs (r=0.28, P=0.002). The ionic alterations caused by KCNK9 may play a role in this niche, possibly through mechanisms such as the C-C chemokine receptor type 4/C-C motif chemokine ligand 22 (CCR4/CCL22) chemokine axis that facilitates Treg recruitment (18,19). Simultaneously, the association with inhibited dendritic cell maturation (resting dendritic cells: r=−0.19, P=0.01) presumably hinders antigen presentation, hence diminishing anti-tumor immunity (20). The increased infiltration of M1 macrophages (r=0.24, P=0.008) is typically associated with pro-inflammatory responses (21). KCNK9-high UCEC may have a compensatory immunological response to accelerated tumor growth, which is consistent with the complex and sometimes contradictory activities of immune cells in the TME (22,23). This complex immune modulation aligns with observations in malignancies such as triple-negative breast cancer, where ion channel dysregulation generates conflicting signals (24).

A testable hypothesis linking our genetic and observational data to a possible biological driver in UCEC is provided by this suggested mechanism, which is based on ion channel-mediated modification of the tumor electrochemical and immunological environment. We propose that the observed extrinsic immunosuppressive effects and intrinsic carcinogenic pathways (e.g., PI3K/AKT) may be supported by KCNK9-mediated ionic changes. Testing this theory in UCEC models should be the top priority of future mechanistic research.

Because KCNK9 plays a dual role in promoting intrinsic tumor cell proliferation and cultivating an immunosuppressive microenvironment, treatment approaches that just target it may produce insufficient results. Therefore, we suggest that combination strategies, including co-targeting KCNK9 with immune checkpoint inhibitors [like programmed cell death protein 1/programmed death-ligand 1 (PD-1/PD-L1) blockers], could more successfully interfere with the tumor’s immune-evasive strategies as well as its proliferative engine.

First, co-administration of PD-1/PD-L1 checkpoint blockers and KCNK9 inhibitors (e.g., a particular monoclonal antibody that targets its extracellular domain) may concurrently disrupt intrinsic tumor growth pathways and restore CD8⁺ T cell function (25,26).

Secondly, targeting KCNK9-mediated hypoxia adaptation pathways, in which KCNK9 overexpression is associated with enhancing tumor cell survival in hypoxic environments, may mitigate platinum-based chemoresistance (27). This approach is based on ovarian cancer methods that fight chemoresistance by stopping hypoxia-induced glycolysis or targeting mitochondrial metabolism. This gives hope to UCEC patients whose disease keeps coming back or has spread to other parts of the body (28).

The strategies illustrate the potential of integrating immunotherapy with ion channel modulation to improve clinical outcomes in UCEC. An illustration of a bridging technique that links conventional bioinformatics with contemporary multi-omics systems is the amalgamation of TCGA bulk RNA-seq data with IHC validation. GSEA-enabled pathway enrichment analysis indicated that KCNK9 is associated with axon guidance pathways and neuroactive ligand-receptor interactions. These paths have not been studied much in UCEC before now, but GSEA’s ability to make sense of complicated signaling networks has changed that.

The KEGG analysis found that the oxytocin signaling system is one example of this kind of uniqueness. It is known to play a role in hormone-related cancers like breast carcinoma, but this is the first time it has been linked to uterine cancers through KCNK9 dysregulation (29). Studies have demonstrated that OT increases cell migration and invasion in a PI3K/AKT-dependent manner, suggesting that oxytocin receptor activation may activate EMT through PI3K/AKT signaling (30,31). KCNK9’s pro-metastatic effects, independently verified, are explained by this mechanism. The negative connection between KCNK9 and ribosome biogenesis pathways suggests a metabolic reprogramming that prioritizes ion channel activity over anabolic pathways for proliferation. This is increasingly recognized by machine learning-based metabolic flux measurements. GSEA showed that higher KCNK9 expression negatively correlated with ribosome biogenesis and protein translation pathways (Figure 4G). This shows significant metabolic reprogramming in KCNK9-high UCEC cancers. These cells may use bioenergetic resources to sustain ion channel function and cellular stress responses at the expense of bulk protein synthesis machinery to proliferate quickly. This change fits with the part of the KCNK9 gene in controlling the potential mitochondrial membrane and apoptosis (32).

Consequently, TASK-3 activity, observed in cancer cell mitochondria, has the potential to influence critical parameters that dictate cellular metabolic fitness (33). The downregulation of ribosomal biogenesis in aggressive KCNK9-high tumors may indicate a strategic metabolic trade-off, reallocating biosynthetic capacity to support KCNK9-driven proliferation signals and survival pathways. This reordering draws attention to a crucial aspect of metabolic flexibility that needs more research. Even if these advantages are significant, it is important to consider the drawbacks. Caucasians are overrepresented while uncommon UCEC histological subtypes are underrepresented due to our reliance on the TCGA database.

Our findings may not apply to heterogeneous populations or rare disease variants due to these factors. Bulk RNA-seq data automatically averages signals across all tumor cells, hiding intra-tumoral heterogeneity and unique stromal-tumor interactions. Future studies using single-cell RNA sequencing or spatial transcriptomics are needed to determine KCNK9’s cellular expression and immune cell affiliation. The location of KCNK9 expression in tumor nests vs invasive fronts may explain its role in localized immune evasion. IHC confirmed KCNK9 overexpression, but functional investigations are needed to prove causality.

The necessity for customized functional validation was highlighted by CRISPR dependency analysis, which revealed context-specific relevance but not uniform essentiality in UCEC models. The function of KCNK9 in immune cell recruitment and Wnt/β-catenin pathway modulation may be clarified using CRISPR knockout models in UCEC cell lines or patient-derived xenografts. Fourth, it is impossible to assess KCNK9’s prognostic capacity for chemotherapy or immunotherapy results in the absence of therapeutic response data.

The constraint is significant. Clinical trials that stratify UCEC patients based on KCNK9 expression levels are necessary to confirm its use in therapy selection and to implement our biomarker discoveries in clinical practice. The intricate complexity of tumor biology is insufficiently captured by bioinformatics analysis, despite its strength; hence, the integration of many omics is necessary to enhance mechanistic understanding.

In summary, this study was designed to address the independent prognostic value of KCNK9 in UCEC and its correlation with the TME. The clinical implications are twofold: first, KCNK9 expression could augment existing prognostic models for improved risk stratification. Second, its association with an immunosuppressive microenvironment suggests it may inform future research into patient selection for immunotherapy or combination therapies targeting ion channels and immune checkpoints.

Conclusions

This multi-omics study establishes KCNK9 as a significant contributor to the aggressiveness and immune evasion of UCEC. It is consistently overexpressed, serves as an independent prognostic factor, and correlates with an immunosuppressive TME. Supplementary genetic analyses, notably MR, provide initial evidence for a potential causal link between KCNK9 and UCEC risk. KCNK9 is a strong candidate biomarker that warrants further functional validation and therapeutic exploration.

Our results offer a framework for exploring ion channel biology in gynecologic oncology. The integration of layered computational evidence elucidates the clinical and mechanistic associations of KCNK9, establishing a foundation for subsequent functional validation in experimental models and prospective clinical evaluation to benefit patients.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-2087/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-2087/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-2087/prf

Funding: This work was supported by the Shanghai First Maternity and Infant Hospital (No. 2022A07).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-2087/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Specifically, the use of patient-derived tissues for immunohistochemical validation was approved by the Ethics Committee of Shanghai First Maternity and Infant Hospital (approval No. KS23112), and written informed consent was obtained from all participants.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Siegel RL, Kratzer TB, Giaquinto AN, et al. Cancer statistics, 2025. CA Cancer J Clin 2025;75:10-45. [Crossref] [PubMed]
2. Abu-Rustum N, Yashar C, Arend R, et al. Uterine Neoplasms, Version 1.2023, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw 2023;21:181-209. [Crossref] [PubMed]
3. Lin H, Li J, Zhang Q, et al. C-type inactivation and proton modulation mechanisms of the TASK3 channel. Proc Natl Acad Sci U S A 2024;121:e2320345121. [Crossref] [PubMed]
4. Rinné S, Schick F, Vowinkel K, et al. Potassium channel TASK-5 forms functional heterodimers with TASK-1 and TASK-3 to break its silence. Nat Commun 2024;15:7548. [Crossref] [PubMed]
5. Pardo LA, Stühmer W. The roles of K(+) channels in cancer. Nat Rev Cancer 2014;14:39-48. [Crossref] [PubMed]
6. Konakov MV, Teplov IY, Levin SG, et al. Anti-hypoxic effect of interleukin-10 in hippocampal neurons is mediated by modulation of TASK-1 and TASK-3 channels activity. Biochem Biophys Res Commun 2022;615:17-23. [Crossref] [PubMed]
7. Cheng Y, Tang Y, Tan Y, et al. KCNK9 mediates the inhibitory effects of genistein on hepatic metastasis from colon cancer. Clinics (Sao Paulo) 2023;78:100141. [Crossref] [PubMed]
8. Zou Y, Xie J, Tian W, et al. Integrative Analysis of KCNK Genes and Establishment of a Specific Prognostic Signature for Breast Cancer. Front Cell Dev Biol 2022;10:839986. [Crossref] [PubMed]
9. Zhai S, Zhao J, Cai J. Integrative Bioinformatics and Experimental Validation Reveal the Mechanistic Action of Patchouli Alcohol in Prostate Cancer Treatment. Curr Pharm Biotechnol 2025; Epub ahead of print. [Crossref] [PubMed]
10. Innamaa A, Jackson L, Asher V, et al. Expression and prognostic significance of the oncogenic K2P potassium channel KCNK9 (TASK-3) in ovarian carcinoma. Anticancer Res 2013;33:1401-8. [PubMed]
11. Hu Z, Jia Q, Yao S, et al. The TWIK-related acid sensitive potassium 3 (TASK-3) channel contributes to the different effects of anesthetics on the growth and metastasis of ovarian cancer cells. Heliyon 2024;10:e34973. [Crossref] [PubMed]
12. Skok K, Maccio U, Martin SD, et al. Tumor-Infiltrating Lymphocytes in Breast and Female Genital Tract Cancers: Overlooked Potential and Unexplored Frontiers. Cancer Med 2025;14:e71023. [Crossref] [PubMed]
13. Angi B, Muccioli S, Szabò I, et al. A Meta-Analysis Study to Infer Voltage-Gated K(+) Channels Prognostic Value in Different Cancer Types. Antioxidants (Basel) 2023;12:573. [Crossref] [PubMed]
14. Concin N, Matias-Guiu X, Cibula D, et al. ESGO-ESTRO-ESP guidelines for the management of patients with endometrial carcinoma: update 2025. Lancet Oncol 2025;26:e423-35. [Crossref] [PubMed]
15. Sun H, Luo L, Lal B, et al. A monoclonal antibody against KCNK9 K(+) channel extracellular domain inhibits tumour growth and metastasis. Nat Commun 2016;7:10339. [Crossref] [PubMed]
16. Chirra M, Newton HS, Gawali VS, et al. How the Potassium Channel Response of T Lymphocytes to the Tumor Microenvironment Shapes Antitumor Immunity. Cancers (Basel) 2022;14:3564. [Crossref] [PubMed]
17. Zúñiga L, Cayo A, González W, et al. Potassium Channels as a Target for Cancer Therapy: Current Perspectives. Onco Targets Ther 2022;15:783-97. [Crossref] [PubMed]
18. Chen S, Cui W, Chi Z, et al. Tumor-associated macrophages are shaped by intratumoral high potassium via Kir2.1. Cell Metab 2022;34:1843-1859.e11. [Crossref] [PubMed]
19. Zhan S, Yung MMH, Siu MKY, et al. New Insights into Ferroptosis Initiating Therapies (FIT) by Targeting the Rewired Lipid Metabolism in Ovarian Cancer Peritoneal Metastases. Int J Mol Sci 2022;23:15263. [Crossref] [PubMed]
20. Yang X, Wang Y, Lin Y, et al. A Multi-Task Self-Supervised Strategy for Predicting Molecular Properties and FGFR1 Inhibitors. Adv Sci (Weinh) 2025;12:e2412987. [Crossref] [PubMed]
21. Lamorte S, Quevedo R, Jin R, et al. Lymph node macrophages drive immune tolerance and resistance to cancer therapy by induction of the immune-regulatory cytokine IL-33. Cancer Cell 2025;43:955-969.e10. [Crossref] [PubMed]
22. Garg P, Ramisetty SK, Raghu Subbalakshmi A, et al. Gynecological cancer tumor Microenvironment: Unveiling cellular complexity and therapeutic potential. Biochem Pharmacol 2024;229:116498. [Crossref] [PubMed]
23. Wang Y, Liu N, Guo X, et al. The immune microenvironment in endometrial carcinoma: mechanisms and therapeutic targeting. Front Immunol 2025;16:1586315. [Crossref] [PubMed]
24. Li X, Guo Q, Chen Q, et al. Reconciling the Cooperative-Competitive Patterns among Tumor and Immune Cells for Triple-Negative Breast Cancer Treatment Using Multimodule Nanocomplexes. Adv Mater 2024;36:e2312219. [Crossref] [PubMed]
25. Fu Z, Zhang X, Gao Y, et al. Enhancing the anticancer immune response with the assistance of drug repurposing and delivery systems. Clin Transl Med 2023;13:e1320. [Crossref] [PubMed]
26. Yang M, Cui M, Sun Y, et al. Mechanisms, combination therapy, and biomarkers in cancer immunotherapy resistance. Cell Commun Signal 2024;22:338. [Crossref] [PubMed]
27. Nan Y, Wu X, Luo Q, et al. OTUB2 silencing promotes ovarian cancer via mitochondrial metabolic reprogramming and can be synthetically targeted by CA9 inhibition. Proc Natl Acad Sci U S A 2024;121:e2315348121. [Crossref] [PubMed]
28. Sriramkumar S, Sood R, Huntington TD, et al. Platinum-induced mitochondrial OXPHOS contributes to cancer stem cell enrichment in ovarian cancer. J Transl Med 2022;20:246. [Crossref] [PubMed]
29. Shi Q, Yang Z, Yang H, et al. Targeting ion channels: innovative approaches to combat cancer drug resistance. Theranostics 2025;15:521-45. [Crossref] [PubMed]
30. Ali TM, Yassin NYS, El Askary A, et al. Efficacy and Mechanisms of Action of Oxytocin in NA/STZ-induced Diabetic Wistar Rats: Roles of Adipocytokines, PI3K/AKT, and AMPK Signaling Pathways. In Vivo 2025;39:3333-50. [Crossref] [PubMed]
31. Hutchings CJ, Colussi P, Clark TG. Ion channels as therapeutic antibody targets. MAbs 2019;11:265-96. [Crossref] [PubMed]
32. Skaar DA, Dietze EC, Alva-Ornelas JA, et al. Epigenetic Dysregulation of KCNK9 Imprinting and Triple-Negative Breast Cancer. Cancers (Basel) 2021;13:6031. [Crossref] [PubMed]
33. Bachmann M, Rossa A, Antoniazzi G, et al. Synthesis and cellular effects of a mitochondria-targeted inhibitor of the two-pore potassium channel TASK-3. Pharmacol Res 2021;164:105326. [Crossref] [PubMed]