Prognostic Value and Immune Relevance of Ultrafiltration Failure–Related Genes in Clear Cell Renal Cell Carcinoma

Prognostic Value and Immune Relevance of Ultrafiltration Failure–Related Genes in Clear Cell Renal Cell Carcinoma | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Prognostic Value and Immune Relevance of Ultrafiltration Failure–Related Genes in Clear Cell Renal Cell Carcinoma Yifei Liu, Yuheng Tao, Jiajia Wang, Wei Liu, Silin Jiang, Rongjiang Jiang, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7839463/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Clear cell renal cell carcinoma (ccRCC) is a common malignancy characterized by intratumoral heterogeneity, affecting tumor progression and immune regulation. Although ultrafiltration (UF) failure has been clinically associated with kidney disease and ccRCC, the underlying molecular mechanisms remain poorly understood. Here, we developed a UF failure–related gene signature (UFFRGS) by integrating single-cell and bulk RNA sequencing data from TCGA-KIRC and E-MTAB-1980 cohorts. A total of 162 UF failure–related genes were identified, and 117 machine learning algorithm combinations were applied to construct a robust prognostic model comprising 37 key genes. UFFRGS stratified patients into high- and low-risk groups with significant survival differences. High-risk patients exhibited increased infiltration of CD8⁺ T cells, regulatory T cells, and activated CD4⁺ memory T cells, while low-risk patients showed enrichment of M2 macrophages and monocytes, reflecting immune heterogeneity. Mutation analysis revealed distinct patterns in VHL, PBRM1, and SETD2 between risk groups. Drug sensitivity analysis further indicated differential responses to multiple chemotherapeutic agents, providing potential guidance for individualized therapy. Among the core genes, CSNK1E was highlighted as a key prognostic factor, with high expression associated with poor survival and elevated in T4 tumors. GSVA suggested CSNK1E may promote tumor progression via Wnt/β-catenin and hormone-related pathways. In summary, UFFRGS is a stable, reliable prognostic tool that captures molecular and immune heterogeneity in ccRCC, offering insights for risk stratification and potential therapeutic intervention. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 1. Introduction Clear cell renal cell carcinoma (ccRCC) is the most common subtype of Renal Cell Carcinoma, accounting for approximately 70%–80% of all cases [1]. Despite significant advances in surgical resection, targeted therapies, and immunotherapies in recent years [2], the overall prognosis for ccRCC patients remains unsatisfactory, largely due to the pronounced molecular heterogeneity and clinical tendencies toward recurrence and metastasis [3]. Accordingly, the establishment of precise prognostic assessment systems and the development of individualized therapeutic strategies have become central objectives in both clinical practice and fundamental research. Patients with chronic kidney disease (CKD) have a significantly higher risk of developing malignancies than the general population, with ccRCC being the most common type [4]. Epidemiological studies have shown that dialysis patients have an approximately 3-6-fold higher incidence of renal cancer compared with non-dialysis individuals, and this risk increases progressively with longer dialysis duration [5]. During long-term dialysis, acquired cystic kidney disease (ACKD) frequently develops [6], and cystic transformation has been widely recognized as an important precursor lesion for renal carcinogenesis [7]. Ultrafiltration failure (UF failure) was originally defined as impaired peritoneal function and fluid clearance during peritoneal dialysis [8]. As a typical complication of long-term peritoneal dialysis, UF failure not only indicates prolonged dialysis vintage but is also associated with more severe chronic inflammation and fibrosis, thereby serving as a clinical signal for identifying patients at high risk of renal cancer. Notably, the pathological processes underlying UF failure overlap substantially with the molecular mechanisms of tumorigenesis, including epithelial–mesenchymal transition (EMT) [9], oxidative stress [10], and the formation of a chronic inflammatory microenvironment [11]. These alterations play critical roles in the invasion, metastasis, and immune regulation of ccRCC [12, 13]. Therefore, ultrafiltration failure–related genes (UFFRGs) may act not only as molecular markers of peritoneal dysfunction but also as important regulators in the initiation and progression of ccRCC. However, systematic studies investigating the prognostic value of UFFRGs in ccRCC are still lacking. Against this backdrop, the present study integrates single-cell RNA sequencing data with bulk RNA sequencing data to construct a stable and reliable prognostic risk prediction model,termed UF failure–related genes’ signature (UFFRGS), employing 117 machine learning algorithms. Moreover, we further explore the associations between the model and the tumor immune microenvironment, as well as drug sensitivity profiles, with the aim of providing novel theoretical insights and potential therapeutic targets for both the prognostic evaluation and personalized treatment of ccRCC. 2. Materials and Methods 2.1. Sources and preprocessing of datasets The transcriptional profiles and corresponding clinical information were obtained from the Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/) under the TCGA-KIRC project. Following the study design, the TCGA-KIRC cohort was randomly divided into a training set (70%) and an internal validation set (30%) for model development and internal validation, respectively. For external validation, the study incorporated the E-MTAB-1980 dataset from the ArrayExpress database (https://www.ebi.ac.uk/arrayexpress/), which contains gene expression profiles and follow-up clinical data from patients with ccRCC. To enhance the robustness of validation, the TCGA internal validation set was combined with the E-MTAB-1980 dataset to establish an independent validation cohort. In addition, single-cell RNA sequencing (scRNA-seq) data were retrieved from the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE248762. This dataset included samples from six long-term dialysis patients without UF failure and four long-term dialysis patients with UF failure. To ensure cross-platform comparability, gene expression levels from both TCGA-KIRC and E-MTAB-1980 (originally quantified as FPKM, Fragments Per Kilobase of transcript per Million mapped reads) were uniformly converted to TPM (Transcripts Per Million), followed by log2(TPM + 1) transformation for downstream analyses. 2.2. Model Construction and Validation To establish an accurate and robust prognostic prediction system, we employed an ensemble learning strategy that integrated ten distinct machine learning algorithms in various configurations, yielding a total of 117 modeling approaches . The algorithms included stepwise Cox regression, Lasso regression, Ridge regression, CoxBoost, random survival forest (RSF), elastic network (Enet), partial least squares regression for Cox (plsRcox), generalized boosted regression modeling (GBM), supervised principal components (SuperPC), and survival support vector machine (survival-SVM) . Notably, several methods—such as Lasso, stepwise Cox, RSF, and CoxBoost —incorporated intrinsic feature selection mechanisms, thereby enabling the automatic identification of key prognostic variables during model training. Model development were conducted in a stepwise manner. Candidate gene screening ： Genes significantly associated with prognosis and related to ultrafiltration failure (UFFRGs) were identified as initial candidates in the TCGA-KIRC-traning cohort using univariate Cox regression. Model construction:A total of 117 algorithmic combinations were applied to the candidate gene set.The modeling process was conducted under a 10-fold cross-validation framework, ensuring comprehensive variable selection and multi-model development. Performance evaluation:Model performance was assessed using both internal validation (TCGA) and external validation (E-MTAB-1980) cohorts.Harrell’s concordance index (C-index) was used to quantify predictive accuracy. (d) Optimal model selection:The model with the highest average C-index across validation cohorts was selected.This final model, termed UF failure–related genes’ signature (UFFRGS) , demonstrated both superior predictive performance and potential clinical applicability. For clinical risk stratification , patients were categorized into high-risk and low-risk groups according to the optimal cut-off score determined using the survminer R package. Kaplan–Meier survival curves was then performed to compare overall survival between the two risk groups, with statistical significance assessed by the log-rank test [14-16]. To further evaluate predictive performance, time-dependent ROC curves were generated [17, 18], providing a dynamic assessment of model accuracy across multiple follow-up intervals. 2.3. Enrichment analysis in high‑ and low‑risk groups Following application of the optimal prognostic model, patients were stratified into high-risk and low-risk groups based on the predetermined cut-off score. To investigate the biological differences between the two cohorts, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed to identify differentially activated biological processes, molecular functions, and signaling pathways. In addition, Gene Set Variation Analysis (GSVA) was applied to calculate enrichment scores for each sample across a wide range of functional pathways, allowing for a comprehensive assessment of pathway activity differences between the high-risk and low-risk cohorts. 2.4. Immune infiltration analysis The ESTIMATE package in R was applied to calculate the stromal score, immune score, and overall ESTIMATE score for each sample, thereby evaluating the cellular composition of the tumor microenvironment (TME). To further characterize the immune landscape of ccRCC patients, we utilized the CIBERSORT algorithm [19] in combination with single-sample Gene Set Enrichment Analysis (ssGSEA) [20] to quantitatively assess the levels of immune cell infiltration. Based on the ssGSEA results, we then compared the expression profiles of human leukocyte antigen (HLA) family genes between high-risk and low-risk groups, and further examined differences in immune function across the two cohorts. 2.5. Tumor Mutation Burden and Drug Sensitivity Analysis Somatic mutation data were systematically analyzed using the “maftools” R package , and differences in tumor mutation burden (TMB) between the high-risk and low-risk subgroups were compared. To further explore potential therapeutic implications, the “oncoPredict” R package was employed to predict the sensitivity of each cohort to a panel of anti-tumor drugs, thereby identifying potential treatment strategies. 2.6. Identification and Functional Analysis of Core Genes The core gene was initially selected using univariate Cox regression analysis . Based on its expression levels, patients were stratified into high-expression and low-expression groups. The prognostic significance of the core gene was subsequently evaluated using Kaplan–Meier survival curve analysis . In addition, grouped bar plots incorporating patients’ clinical characteristics were generated to examine differences in core gene expression across various clinicopathological parameters . To further explore the potential biological roles of the core gene, GO and KEGG enrichment analyses were performed. GSVA was also applied to assess associated biological processes and signaling pathways . Moreover, by integrating the core gene expression data with immune infiltration profiles , the correlation between the core gene and the tumor immune microenvironment was evaluated. 3. Results 3.1. Identification of Ultrafiltration Failure – Related Genes Using scRNA-seq Data Using the publicly available dataset GSE248762, we analyzed samples from long vintage patients without ultrafiltration failure (LV_NOT_UF) and those who developed ultrafiltration failure (LV_UF). The single-cell RNA sequencing data were first aggregated into pseudobulk expression profiles , followed by differential gene expression analysis . Applying stringent filtering criteria— log₂ fold change (logFC) > 1 , adjusted p-value (adj.p) 10 —we identified a set of 168 genes significantly upregulated in the LV_UF cohort . The results were visualized using volcano plots and MA plots (Fig. 1A, B) . Subsequent GSVA revealed that this high-expression gene set was significantly enriched in pathways related to signal transduction and hormonal regulation (Fig.1 C) . To further refine the candidate genes, an intersection analysis was performed between these 168 upregulated genes and gene sets from the TCGA-KIRC and E-MTAB-1980 datasets, resulting in 162 genes consistently upregulated in LV_UF samples . These genes were designated as UFFRGs . Fig. 1 Differential gene expression and pathway enrichment between high- and low-expression groups. (A) Volcano plot showing the distribution of differentially expressed genes between the high-expression and low-expression groups. (B) MA plot displaying the relationship between mean expression and fold change of genes in the two groups. (C) GSVA enrichment analysis highlighting the differential activation of biological pathways between the high-expression and low-expression groups. 3.2. Comprehensive Development and Evaluation of the Prognostic Model Initially, univariate Cox regression analysis was performed on the 162 UFFRGs , identifying 76 genes significantly associated with prognosis. Subsequently, an integrated machine learning strategy was applied to systematically evaluate 117 algorithmic combinations based on these genes, with the goal of identifying the most predictive and stable model (Fig.2 A). Among all models, the top five exhibited high average C-index . Notably, the “StepCox[both] + CoxBoost” model, which included only 37 genes , was selected as the final signature after comprehensive evaluation, owing to its superior predictive performance and simplicity . Finally, we developed UFFRGS based on the model, which demonstrated high predictive accuracy across multiple datasets. Using the optimal cutoff score determined by the survminer R package, patients were stratified into high-risk and low-risk groups. Kaplan–Meier survival curves revealed significantly better survival in the low-risk group compared to the high-risk group (Fig. 2 B, D, F). To further assess predictive performance, time-dependent ROC curves were performed (Fig. 2 C, E, G). In the training cohort, the area under the curve (AUC) values for 1-, 2-, and 3-year overall survival were 0.846, 0.763, and 0.791 , respectively. Consistent results were observed in the validation cohort, confirming the robust predictive value of the model. Overall, these findings indicate that the UFFRGS model exhibits strong generalizability and holds considerable potential for clinical application across multiple independent cohorts. Fig. 2 Comprehensive development and validation of the prognostic model. (A) C-index distribution of 117 prognostic models. (B,C) Kaplan–Meier survival curve and time-dependent ROC curve of the optimal model in the TCGA-KIRC training cohort. (D,E) Kaplan–Meier survival curve and time-dependent ROC curve in the TCGA-KIRC validation cohort. (F,G) Kaplan–Meier survival curve and time-dependent ROC curve in the E-MTAB-1980 cohort. 3.3. Association Between Risk Score and Clinical Characteristics In clinical practice, clinical characteristics are commonly used for prognostic assessment. We further explored the relationship between the risk score (RS) and various clinical characteristics. In the training cohort , multivariate Cox regression analysis showed that the RS remained an independent prognostic factor for overall survival (OS) after adjustment for clinical variables (Fig. 3 A). This finding was consistently validated in the validation cohort (Fig. 3 B) , supporting the robustness of the RS as a prognostic biomarker for ccRCC . Moreover, patients with advanced pathological stage (Stage IV) , larger or more invasive tumors (T3/T4) , distant metastasis (M1) , or lymph node involvement (N1) tended to have significantly higher risk scores (Fig. 3 C-F) . These results indicate a strong correlation between the RS and disease progression , suggesting that the RS may serve as a valuable indicator for assessing tumor aggressiveness and predicting clinical outcomes . Fig. 3 Clinical relevance of the prognostic risk score. (A) Multivariate Cox regression analysis of risk score and clinical factors in the TCGA-KIRC training cohort. (B) Multivariate Cox regression analysis in the TCGA-KIRC validation cohort. (C–F) Box plots showing the associations between risk score and key clinical characteristics. 3.4. Functional Enrichment Analysis To explore the molecular features distinguishing the high-risk and low-risk cohorts, we performed functional enrichment analysis on genes associated with risk stratification. Using the GO and KEGG databases, we systematically evaluated differences in gene expression profiles between the two cohorts (Fig. 4 A, B). GO enrichment analysis revealed significant functional divergence: genes in the high-risk group were predominantly involved in biological processes such as signal release , extracellular matrix remodeling , and ion homeostasis dysregulation . KEGG pathway analysis further demonstrated that high-risk group genes were significantly enriched in pathways including neuroactive ligand–receptor interaction and cytokine–cytokine receptor interaction , suggesting aberrant activation of neural signaling and immune response pathways in this subgroup. To quantitatively compare cancer-related pathway activity between the risk groups, GSVA was conducted to calculate pathway enrichment scores for each sample (Fig. 4 C). The results indicated that the high-risk group showed significant enrichment in immune-related and signal transduction pathways , such as the KRAS signaling pathway , IL-6/JAK/STAT3 signaling , interferon-gamma response , and the complement system . In contrast, the low-risk group exhibited marked enrichment in metabolic and signaling pathways , including heme metabolism , fatty acid metabolism , TGF-β signaling , and the PI3K-AKT-mTOR pathway , reflecting characteristic alterations in energy metabolism and signal regulation . Fig. 4 Pathway Enrichment and Activity Differences Between High- and Low-Risk Groups. (A) GO enrichment analysis showing significantly enriched biological processes, molecular functions, and cellular components. (B) KEGG pathway enrichment highlighting key signaling pathways associated with the risk groups. (C) GSVA analysis depicting pathway activity differences between the high- and low-risk groups. 3.5. Immune Infiltration Analysis Building on the established association between risk scores and immunopathological features, we conducted a comprehensive analysis of the tumor immune microenvironment . The results showed that the high-risk group exhibited significantly elevated stromal, immune, and ESTIMATE scores , accompanied by reduced tumor purity (Fig. 5 A) . Using the CIBERSORT algorithm , we quantitatively assessed the distribution of immune cell subsets in the TCGA training cohort and visualized differences between the risk groups using boxplots (Fig. 5 B). Specifically, the low-risk group displayed higher proportions of resting CD4⁺ memory T cells, monocytes, M1/M2 macrophages, and resting mast cells , whereas the high-risk group showed increased infiltration of CD8⁺ T cells, regulatory T cells (Tregs), and activated CD4⁺ memory T cells . These findings were consistently validated in the validation cohort (Fig. 5 C) , confirming the robustness of the observed immune landscape. Further analysis of human leukocyte antigen (HLA) family genes revealed significant downregulation of several key HLA genes—including HLA-B, HLA-C, and HLA-E —in the high-risk group (Fig. 5 D) . Examination of immune functional activity indicated that the low-risk group was characterized by enhanced mast cell-related functions , whereas the high-risk group exhibited heightened activity in CD8⁺ T cell and macrophage-associated functions (Fig. 5 E) . Collectively, these results demonstrate distinct immune microenvironment profiles and differential immune response patterns between the two risk groups. Fig. 5 Immune infiltration analysis in high- and low-risk groups. (A) Boxplots of stromal, immune, ESTIMATE scores, and tumor purity. (B) CIBERSORT results in the training cohort. (C) CIBERSORT results in the KIRC validation cohort. (D) Differential expression of HLA-related genes. (F) Immune function differences between groups. 3.6. Tumor Mutational Burden (TMB) Analysis We first constructed a comprehensive mutational landscape of the entire cohort (Fig. 6 A). Missense mutations represented the predominant variant type, with C>T substitutions being the most frequent single-nucleotide variation. Among all mutated genes, VHL exhibited the highest mutation frequency (46%), underscoring its potential role as a pivotal driver in the pathogenesis of ccRCC. We next compared TMB between the risk subgroups (Fig. 6 B, C). In the training cohort, TMB levels showed no statistically significant differences between high- and low-risk patients. However, in the validation cohort, the high-risk subgroup displayed significantly elevated TMB levels compared with the low-risk subgroup. To further characterize subgroup-specific mutation patterns, we applied waterfall plots to depict and rank the gene mutation profiles in each risk group (Fig. 6 D-F). Distinct distributions were observed among the top 15 most frequently mutated genes. For example, VHL , PBRM1 , and SETD2 exhibited heterogeneous mutation profiles between the high- and low-risk groups, suggesting that divergent molecular mechanisms may contribute to the stratified risk classifications. Fig.6 Tumor mutational burden (TMB) analysis in different risk groups. (A) Mutation landscape of the entire cohort. (B) TMB differences between high- and low-risk groups in the training cohort. (C) TMB differences in the KIRC validation cohort. (D) Mutation profile of the top 15 genes in the training cohort. (D-F) Mutation profiles of the top 15 genes in high- and low-risk groups. 3.7. Drug Sensitivity Analysis We next evaluated the differences in drug response between the risk subgroups (Fig. 7). The analysis revealed that patients in the high-risk group exhibited significantly lower half-maximal inhibitory concentration (IC50) values for several agents, including AZD7762, Camptothecin, CDK9 inhibitors, Staurosporine, Topotecan, Vinorelbine, Dactinomycin, and Epirubicin, indicating a heightened sensitivity to these compounds. In contrast, the low-risk group displayed markedly reduced IC50 values for agents such as AT13148, Cediranib, Daporinad, and Osimertinib, suggesting a preferential sensitivity to these therapies. Collectively, these findings highlight distinct drug sensitivity patterns between the high- and low-risk subgroups. Such differences may provide a foundation for developing risk-adapted therapeutic strategies and support the clinical application of precision medicine in ccRCC. Fig.7 Comparison of drug responses between high- and low-risk groups. 3.8. Core Gene Analysis Multivariable Cox regression analysis of the 37 genes incorporated into the prognostic model identified CSNK1E as a key gene with a significant impact on survival outcomes (Fig. 8 A). According to its expression levels, patients in the TCGA-KIRC cohort were stratified into high- and low-expression groups, with differential expression profiles illustrated by volcano and heatmap plots (Fig. 8 B, C). Kaplan–Meier survival analysis further confirmed that CSNK1E served as a robust predictor of overall survival (OS): patients with low CSNK1E expression exhibited a significantly more favorable prognosis compared with those in the high-expression group (Fig. 8 D). Associations between CSNK1E expression and clinical characteristics were then evaluated (Fig. 8 E-G). No significant differences were observed across pathological stage, M stage, or N stage. However, in terms of T staging, median CSNK1E expression was markedly elevated in T4 tumors compared with T1–T3 tumors, suggesting that CSNK1E expression is more closely linked to local tumor invasiveness than to lymph node or distant metastasis. To further explore the biological implications of CSNK1E, functional enrichment analyses were conducted. GO analysis revealed that differentially expressed genes (DEGs) in the high-expression group were significantly enriched in biological processes such as steroid metabolic pathways and monoatomic ion channel complexes (Fig. 9 A). KEGG pathway analysis demonstrated that these DEGs were predominantly involved in neuroactive ligand–receptor interactions and hormone-related signaling pathways (Fig. 9 B), highlighting a potential role of CSNK1E in neuroendocrine regulation. GSVA further showed distinct pathway activity patterns between the groups (Fig. 9 C). The high-expression group exhibited marked activation of oncogenic signaling pathways, including the Wnt/β-catenin signaling pathway , which was consistent with their poorer clinical outcomes. Conversely, the low-expression group displayed enhanced activity in multiple metabolic and immune-related pathways. Finally, immune infiltration analysis using the CIBERSORT algorithm revealed distinct immune landscape profiles according to CSNK1E expression levels (Fig. 9 D). Notably, resting CD4⁺ memory T cells and resting NK cells were positively correlated with CSNK1E expression, indicating that CSNK1E may exert an important role in shaping the immune microenvironment. Fig. 8 Core gene analysis in the TCGA-KIRC cohort. (A) Multivariate Cox regression analysis of the 37 model-included genes. (B) Volcano plot of differential gene expression. (C) Heatmap showing expression patterns across samples. (D) Kaplan–Meier survival curve comparing high- and low-expression groups of CSNK1E. (E-G) Boxplots showing associations between CSNK1E expression and key clinical features. Fig. 9 Functional and immune analyses of CSNK1E. (A) GO enrichment analysis of genes associated with CSNK1E expression. (B) KEGG pathway enrichment analysis. (C) GSVA showing pathway activity differences between high- and low-expression groups. (D) CIBERSORT analysis of immune cell infiltration associated with CSNK1E expression. 4. Discussion In ccRCC, most existing prognostic models rely on bulk transcriptomic data or limited immune-related gene panels [21, 22], which fail to adequately capture tumor cellular heterogeneity and the complexity of the immune microenvironment. Notably, few studies systematically incorporate ultrafiltration failure–related mechanisms. There is, therefore, a pressing need for a robust prognostic tool that integrates single-cell and bulk transcriptomic data while accounting for immune context. In this study, we developed the UFFRGS by integrating single-cell and bulk transcriptomic datasets. The model, comprising 37 key genes, demonstrated strong prognostic performance across multiple independent cohorts. Patients were stratified into high- and low-risk groups based on risk scores. High-risk patients not only exhibited significantly poorer overall survival but also displayed distinct immune microenvironment characteristics: increased infiltration of CD8⁺ T cells in high-risk patients versus enrichment of M2 macrophages in low-risk patients. While CD8⁺ T cells generally exert antitumor effects, prolonged activation in ccRCC can lead to an exhausted phenotype, adversely affecting prognosis [23]. In contrast, M2 macrophages are associated with immunosuppressive signaling and tumor aggressiveness [24]. These immune patterns provide a biological basis for UFFRGS-based stratification, indicating that the model captures heterogeneity within the tumor immune microenvironment and supporting its potential utility in clinical risk assessment and personalized therapy. Tumor mutational burden analysis further revealed heterogeneous mutation patterns in VHL, PBRM1, and SETD2 between risk groups, consistent with prior studies describing the genomic and transcriptomic landscapes of ccRCC, including mutations in key driver genes and alterations in immune evasion–related pathways [25]. Moreover, correlations between risk scores, clinical features, and drug sensitivity provide an integrated framework that links molecular characteristics to potential therapeutic responses. Among the 37 core genes, CSNK1E emerged as a notable candidate. As a Casein kinase 1 family member [26], CSNK1E participates in cell cycle regulation, signal transduction, and tumorigenesis [27, 28]. Although its roles in neural and endocrine regulation are documented [29, 30], its function in ccRCC, particularly in relation to UF failure, remains unclear. Our findings indicate that CSNK1E overexpression correlates with poor survival and is associated with locally invasive T4 tumors rather than lymph node metastasis, suggesting a key role in local tumor progression and microenvironmental regulation. Pathway analysis suggested that CSNK1E may promote tumor progression by activating Wnt/β-catenin signaling and perturbing hormone-related and neuroactive ligand–receptor interactions, consistent with the well-established role of Wnt/β-catenin in ccRCC progression [31]. Previous studies report that CK1δ/ε inhibition suppresses ccRCC cell proliferation and enhances treatment sensitivity [32], indicating potential druggability of CSNK1E. As a circadian rhythm gene, CSNK1E dysregulation may disrupt rhythm-dependent metabolic and immune homeostasis [33, 34], aligning with its identification as a prognostic factor in our integrated analysis. Notably, high CSNK1E expression correlated with increased resting CD4⁺ memory T cells and NK cells, suggesting a “hypoactive” immune state [35]. While these findings are correlative, functional experiments are needed to validate the direct effects of CSNK1E on Wnt signaling, immune regulation, and circadian rhythm in ccRCC. Despite these insights, this study has limitations. First, it is based on retrospective data and predominantly computational analyses, lacking prospective clinical validation to confirm UFFRGS's prognostic value. Second, functional and immune analyses are inferred from bioinformatics; in vitro experiments such as gene knockdown/overexpression and immunological assays are required to elucidate the roles of CSNK1E and other core genes. Additionally, drug sensitivity predictions rely on computational models without validation in patient-derived or pharmacological data, limiting immediate clinical translatability. Future studies should address these limitations by validating UFFRGS in prospective cohorts, supporting computational predictions with targeted functional experiments, and integrating public drug-response datasets or limited clinical samples for drug sensitivity validation. Further incorporation of multi-omic data, including epigenomics, metabolomics, and spatial transcriptomics, may optimize risk modeling and expand its application across diverse renal cancer subtypes. 5. Conclusion Using machine learning algorithms, this study constructed an ultrafiltration failure–related gene prognostic model for clear cell renal cell carcinoma (UFFRGS) and validated its robust predictive performance across multiple independent cohorts. The results demonstrated that high-risk patients not only exhibited significantly poorer overall survival but also displayed distinct immune microenvironment features and key driver gene mutation patterns. The core gene CSNK1E may contribute to tumor progression through multiple signaling pathways. Overall, this study provides novel insights into the potential molecular links between ultrafiltration failure and clear cell renal cell carcinoma, and offers a theoretical basis for clinical risk stratification and the development of personalized therapeutic strategies. Declarations Ethics approval and consent to participate Not applicable Consent for publication Not applicable Availability of data and materials The datasets generated and analysed during the current study can obtained from the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/), the Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/), and the ArrayExpress database (https://www.ebi.ac.uk/arrayexpress/). Competing interests The authors declare that they have no competing interests. Funding This study was supported by the Medical Research Project of the Jiangsu Provincial Health Commission (No. Z2023076). Authors' contributions YL was responsible for data collection, bioinformatics analysis, and drafting of the manuscript. YT and JW assisted with data preprocessing, statistical analysis, and figure preparation. WL and SJ contributed to literature review and interpretation of biological findings. RJ and LS participated in model validation and result discussion. XR provided technical support and made critical revisions to the manuscript. JY provided overall guidance and performed the final review of the paper. All authors read and approved the final manuscript. Acknowledgements Not applicable References Hsieh JJ, Purdue MP, Signoretti S, Swanton C, Albiges L, Schmidinger M, Heng DY, Larkin J, Ficarra V: Renal cell carcinoma . Nat Rev Dis Primers 2017, 3 :17009. Schiavoni V, Campagna R, Pozzi V, Cecati M, Milanese G, Sartini D, Salvolini E, Galosi AB, Emanuelli M: Recent Advances in the Management of Clear Cell Renal Cell Carcinoma: Novel Biomarkers and Targeted Therapies . Cancers (Basel) 2023, 15 (12). Li Y, Lih TM, Dhanasekaran SM, Mannan R, Chen L, Cieslik M, Wu Y, Lu RJ, Clark DJ, Kołodziejczak I et al : Histopathologic and proteogenomic heterogeneity reveals features of clear cell renal cell carcinoma aggressiveness . Cancer Cell 2023, 41 (1):139-163.e117. Åkerlund J, Holmberg E, Lindblad P, Stendahl M, Ljungberg B, Thorstenson A, Lundstam S: Increased risk for renal cell carcinoma in end stage renal disease - a population-based case-control study . Scand J Urol 2021, 55 (3):209-214. Levine E: Renal cell carcinoma in uremic acquired renal cystic disease: incidence, detection, and management . Urol Radiol 1992, 13 (4):203-210. Kabaria R, Klaassen Z, Terris MK: Renal cell carcinoma: links and risks . Int J Nephrol Renovasc Dis 2016, 9 :45-52. Przybycin CG, Harper HL, Reynolds JP, Magi-Galluzzi C, Nguyen JK, Wu A, Sangoi AR, Liu PS, Umar S, Mehra R et al : Acquired Cystic Disease-associated Renal Cell Carcinoma (ACD-RCC): A Multiinstitutional Study of 40 Cases With Clinical Follow-up . Am J Surg Pathol 2018, 42 (9):1156-1165. Smit W, Parikova A, Krediet RT: Ultrafiltration failure in peritoneal dialysis. Causes and clinical consequences . Minerva Urol Nefrol 2005, 57 (3):165-174. Aroeira LS, Aguilera A, Sánchez-Tomero JA, Bajo MA, del Peso G, Jiménez-Heffernan JA, Selgas R, López-Cabrera M: Epithelial to mesenchymal transition and peritoneal membrane failure in peritoneal dialysis patients: pathologic significance and potential therapeutic interventions . J Am Soc Nephrol 2007, 18 (7):2004-2013. Oh SH, Yook JM, Jung HY, Choi JY, Cho JH, Park SH, Kim CD, Kim YL, Lim JH: Autophagy caused by oxidative stress promotes TGF-β1-induced epithelial-to-mesenchymal transition in human peritoneal mesothelial cells . Cell Death Dis 2024, 15 (5):365. Baroni G, Schuinski A, de Moraes TP, Meyer F, Pecoits-Filho R: Inflammation and the peritoneal membrane: causes and impact on structure and function during peritoneal dialysis . Mediators Inflamm 2012, 2012 :912595. Fiaschi T, Chiarugi P: Oxidative stress, tumor microenvironment, and metabolic reprogramming: a diabolic liaison . Int J Cell Biol 2012, 2012 :762825. Suarez-Carmona M, Lesage J, Cataldo D, Gilles C: EMT and inflammation: inseparable actors of cancer progression . Mol Oncol 2017, 11 (7):805-823. Liu H, Tang T: MAPK signaling pathway-based glioma subtypes, machine-learning risk model, and key hub proteins identification . Sci Rep 2023, 13 (1):19055. Liu H, Weng J: A comprehensive bioinformatic analysis of cyclin-dependent kinase 2 (CDK2) in glioma . Gene 2022, 822 :146325. Liu H, Weng J, Huang CL, Jackson AP: Is the voltage-gated sodium channel β3 subunit (SCN3B) a biomarker for glioma? Funct Integr Genomics 2024, 24 (5):162. Li Y, Liu H: Clinical powers of Aminoacyl tRNA Synthetase Complex Interacting Multifunctional Protein 1 (AIMP1) for head-neck squamous cell carcinoma . Cancer Biomark 2022, 34 (3):359-374. Liu H, Weng J: A Pan-Cancer Bioinformatic Analysis of RAD51 Regarding the Values for Diagnosis, Prognosis, and Therapeutic Prediction . Front Oncol 2022, 12 :858756. Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, Hoang CD, Diehn M, Alizadeh AA: Robust enumeration of cell subsets from tissue expression profiles . Nat Methods 2015, 12 (5):453-457. Hänzelmann S, Castelo R, Guinney J: GSVA: gene set variation analysis for microarray and RNA-seq data . BMC Bioinformatics 2013, 14 :7. Hua S, Xie Z, Zhang Y, Wu L, Shi F, Wang X, Xia S, Dong S, Jiang J: Identification and validation of an immune-related gene prognostic signature for clear cell renal carcinoma . Front Immunol 2022, 13 :869297. Ren S, Wang W, Shen H, Zhang C, Hao H, Sun M, Wang Y, Zhang X, Lu B, Chen C et al : Development and Validation of a Clinical Prognostic Model Based on Immune-Related Genes Expressed in Clear Cell Renal Cell Carcinoma . Front Oncol 2020, 10 :1496. Qi Y, Xia Y, Lin Z, Qu Y, Qi Y, Chen Y, Zhou Q, Zeng H, Wang J, Chang Y et al : Tumor-infiltrating CD39(+)CD8(+) T cells determine poor prognosis and immune evasion in clear cell renal cell carcinoma patients . Cancer Immunol Immunother 2020, 69 (8):1565-1576. Liu J, Geng X, Hou J, Wu G: New insights into M1/M2 macrophages: key modulators in cancer progression . Cancer Cell Int 2021, 21 (1):389. Hakimi AA, Ostrovnaya I, Reva B, Schultz N, Chen YB, Gonen M, Liu H, Takeda S, Voss MH, Tickoo SK et al : Adverse outcomes in clear cell renal cell carcinoma with mutations of 3p21 epigenetic regulators BAP1 and SETD2: a report by MSKCC and the KIRC TCGA research network . Clin Cancer Res 2013, 19 (12):3259-3267. Knippschild U, Gocht A, Wolff S, Huber N, Löhler J, Stöter M: The casein kinase 1 family: participation in multiple cellular processes in eukaryotes . Cell Signal 2005, 17 (6):675-689. Yang WS, Stockwell BR: Inhibition of casein kinase 1-epsilon induces cancer-cell-selective, PERIOD2-dependent growth arrest . Genome Biol 2008, 9 (6):R92. Lin SH, Lin YM, Yeh CM, Chen CJ, Chen MW, Hung HF, Yeh KT, Yang SF: Casein kinase 1 epsilon expression predicts poorer prognosis in low T-stage oral cancer patients . Int J Mol Sci 2014, 15 (2):2876-2891. Bryant CD, Parker CC, Zhou L, Olker C, Chandrasekaran RY, Wager TT, Bolivar VJ, Loudon AS, Vitaterna MH, Turek FW et al : Csnk1e is a genetic regulator of sensitivity to psychostimulants and opioids . Neuropsychopharmacology 2012, 37 (4):1026-1035. Varghese RT, Young S, Pham L, Liang Y, Pridham KJ, Guo S, Murphy S, Kelly DF, Sheng Z: Casein Kinase 1 Epsilon Regulates Glioblastoma Cell Survival . Sci Rep 2018, 8 (1):13621. Xu Q, Krause M, Samoylenko A, Vainio S: Wnt Signaling in Renal Cell Carcinoma . Cancers (Basel) 2016, 8 (6). Lin YC, Sun DP, Hsieh TH, Chen CH: Targeting CK1δ and CK1ε as a New Therapeutic Approach for Clear Cell Renal Cell Carcinoma . Pharmacology 2024, 109 (6):330-340. Fulcher LJ, Sapkota GP: Functions and regulation of the serine/threonine protein kinase CK1 family: moving beyond promiscuity . Biochem J 2020, 477 (23):4603-4621. Li S, Wang X, Wang Q, Ding K, Chen X, Zhao Y, Gao Y, Wang Y: Effects and Prognostic Values of Circadian Genes CSNK1E/GNA11/KLF9/THRAP3 in Kidney Renal Clear Cell Carcinoma via a Comprehensive Analysis . Bioengineering (Basel) 2022, 9 (7). Giraldo NA, Becht E, Remark R, Damotte D, Sautès-Fridman C, Fridman WH: The immune contexture of primary and metastatic human tumours . Curr Opin Immunol 2014, 27 :8-15. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 04 Dec, 2025 Editor invited by journal 10 Nov, 2025 Editor assigned by journal 07 Nov, 2025 Submission checks completed at journal 07 Nov, 2025 First submitted to journal 12 Oct, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7839463","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":555452240,"identity":"555e8463-6991-422e-a0fe-1819a492343b","order_by":0,"name":"Yifei Liu","email":"","orcid":"","institution":"Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yifei","middleName":"","lastName":"Liu","suffix":""},{"id":555452241,"identity":"dfea0af9-4489-45d3-b613-2a9854a5ba88","order_by":1,"name":"Yuheng Tao","email":"","orcid":"","institution":"Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yuheng","middleName":"","lastName":"Tao","suffix":""},{"id":555452242,"identity":"f2d4131c-0588-41ea-9625-a3aa708c8345","order_by":2,"name":"Jiajia Wang","email":"","orcid":"","institution":"The First Affiliated Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jiajia","middleName":"","lastName":"Wang","suffix":""},{"id":555452243,"identity":"430098b8-9670-467e-b034-029f79922f52","order_by":3,"name":"Wei Liu","email":"","orcid":"","institution":"Second Affiliated Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Wei","middleName":"","lastName":"Liu","suffix":""},{"id":555452244,"identity":"859ff8f8-3a85-4517-b233-6e861524cedb","order_by":4,"name":"Silin Jiang","email":"","orcid":"","institution":"Second Affiliated Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Silin","middleName":"","lastName":"Jiang","suffix":""},{"id":555452246,"identity":"baa5e69e-bed6-4a38-8933-5108c43b5d3a","order_by":5,"name":"Rongjiang Jiang","email":"","orcid":"","institution":"Second Affiliated Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Rongjiang","middleName":"","lastName":"Jiang","suffix":""},{"id":555452250,"identity":"2d08fe23-222a-40c1-8b50-fe0c22bd887f","order_by":6,"name":"Lianlian Shen","email":"","orcid":"","institution":"Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Lianlian","middleName":"","lastName":"Shen","suffix":""},{"id":555452251,"identity":"c744d39c-243f-446b-bb24-52e326e4e334","order_by":7,"name":"Xiaohan Ren","email":"","orcid":"","institution":"The First Affiliated Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xiaohan","middleName":"","lastName":"Ren","suffix":""},{"id":555452253,"identity":"0bf349be-d2d8-4546-babd-2418d5d7188e","order_by":8,"name":"Jian Yang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6ElEQVRIiWNgGAWjYLACCQMGBn4GhgQQm7GBaC2SDSRpAQGDAxCasBaD42cPv7AosMszvt3wdDMPg43shgPMzx7g1XImL81CwiC52OzOgbTbPAxpxhsOsJkb4HdPjpmBhAFz4rYbCSAthxM3HOBhk8Cr5fwbkJb6xM0zwFr+E6HlRo7xAwkDoOESYC0HCGuRvPHGDBjIxxNnAB12c45BsvHMw2xmeLXwnc8x/izxpzqxf0ZO2o03FXayfcebn+HVonCAgU0aooInAehOIM2MTz0QyDcwMH/8AGayHyCgdhSMglEwCkYqAABxcU4O0Ebf6wAAAABJRU5ErkJggg==","orcid":"","institution":"Second Affiliated Hospital of Nanjing Medical University","correspondingAuthor":true,"prefix":"","firstName":"Jian","middleName":"","lastName":"Yang","suffix":""}],"badges":[],"createdAt":"2025-10-12 08:53:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7839463/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7839463/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97892894,"identity":"dac9c98c-e3f2-4439-93b0-5ae59b818e2e","added_by":"auto","created_at":"2025-12-10 15:24:00","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2269886,"visible":true,"origin":"","legend":"\u003cp\u003eDifferential gene expression and pathway enrichment between high- and low-expression groups. (A) Volcano plot showing the distribution of differentially expressed genes between the high-expression and low-expression groups. (B) MA plot displaying the relationship between mean expression and fold change of genes in the two groups. (C) GSVA enrichment analysis highlighting the differential activation of biological pathways between the high-expression and low-expression groups.\u003c/p\u003e","description":"","filename":"Fig.1.png","url":"https://assets-eu.researchsquare.com/files/rs-7839463/v1/42520db7f7c17b7d1f6df89d.png"},{"id":97894116,"identity":"c1131dc1-9c05-4b71-a5b3-b382f1c0cfc5","added_by":"auto","created_at":"2025-12-10 15:31:57","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":7383560,"visible":true,"origin":"","legend":"\u003cp\u003eComprehensive development and validation of the prognostic model. (A) C-index distribution of 117 prognostic models. (B,C) Kaplan–Meier survival curve and time-dependent ROC curve of the optimal model in the TCGA-KIRC training cohort. (D,E) Kaplan–Meier survival curve and time-dependent ROC curve in the TCGA-KIRC validation cohort. (F,G) Kaplan–Meier survival curve and time-dependent ROC curve in the E-MTAB-1980 cohort.\u003c/p\u003e","description":"","filename":"Fig.2.png","url":"https://assets-eu.researchsquare.com/files/rs-7839463/v1/db0a19f078b6f7de860cc634.png"},{"id":97894541,"identity":"e847e291-437b-4a6a-9acc-b635b8f9e2ed","added_by":"auto","created_at":"2025-12-10 15:32:41","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1557983,"visible":true,"origin":"","legend":"\u003cp\u003eClinical relevance of the prognostic risk score. (A) Multivariate Cox regression analysis of risk score and clinical factors in the TCGA-KIRC training cohort. (B) Multivariate Cox regression analysis in the TCGA-KIRC validation cohort. (C–F) Box plots showing the associations between risk score and key clinical characteristics.\u003c/p\u003e","description":"","filename":"Fig.3.png","url":"https://assets-eu.researchsquare.com/files/rs-7839463/v1/2bc557081cb51c383923bf7d.png"},{"id":97697480,"identity":"79e8b765-2b64-4b17-aa81-fcaa535f2e0d","added_by":"auto","created_at":"2025-12-08 11:44:12","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":4090181,"visible":true,"origin":"","legend":"\u003cp\u003ePathway Enrichment and Activity Differences Between High- and Low-Risk Groups. (A) GO enrichment analysis showing significantly enriched biological processes, molecular functions, and cellular components. (B) KEGG pathway enrichment highlighting key signaling pathways associated with the risk groups. (C) GSVA analysis depicting pathway activity differences between the high- and low-risk groups.\u003c/p\u003e","description":"","filename":"Fig.4.png","url":"https://assets-eu.researchsquare.com/files/rs-7839463/v1/95d4fdc4710b1aef137f815b.png"},{"id":97697483,"identity":"ec27ecb3-910d-4767-943f-5d0b61f04df5","added_by":"auto","created_at":"2025-12-08 11:44:12","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2871556,"visible":true,"origin":"","legend":"\u003cp\u003eImmune infiltration analysis in high- and low-risk groups. (A) Boxplots of stromal, immune, ESTIMATE scores, and tumor purity. (B) CIBERSORT results in the training cohort. (C) CIBERSORT results in the KIRC validation cohort. (D) Differential expression of HLA-related genes. (F) Immune function differences between groups.\u003c/p\u003e","description":"","filename":"Fig.5.png","url":"https://assets-eu.researchsquare.com/files/rs-7839463/v1/7dcc277e2fba2a4d96203bad.png"},{"id":97697488,"identity":"6a4efc25-f493-45b7-bed2-26397fe91088","added_by":"auto","created_at":"2025-12-08 11:44:13","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":3665814,"visible":true,"origin":"","legend":"\u003cp\u003eTumor mutational burden (TMB) analysis in different risk groups. (A) Mutation landscape of the entire cohort. (B) TMB differences between high- and low-risk groups in the training cohort. (C) TMB differences in the KIRC validation cohort. (D) Mutation profile of the top 15 genes in the training cohort. (D-F) Mutation profiles of the top 15 genes in high- and low-risk groups.\u003c/p\u003e","description":"","filename":"Fig.6.png","url":"https://assets-eu.researchsquare.com/files/rs-7839463/v1/cac1b217536808e2358cd965.png"},{"id":97697485,"identity":"53aac14b-6882-40cc-8a9d-8e821db533a8","added_by":"auto","created_at":"2025-12-08 11:44:13","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1424305,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of drug responses between high- and low-risk groups.\u003c/p\u003e","description":"","filename":"Fig.7.png","url":"https://assets-eu.researchsquare.com/files/rs-7839463/v1/47a5dd27b5787f38bf971715.png"},{"id":97892939,"identity":"fc6fa4d9-16fa-4faa-a116-0f55348ff5d4","added_by":"auto","created_at":"2025-12-10 15:24:20","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":4264575,"visible":true,"origin":"","legend":"\u003cp\u003eCore gene analysis in the TCGA-KIRC cohort. (A) Multivariate Cox regression analysis of the 37 model-included genes. (B) Volcano plot of differential gene expression. (C) Heatmap showing expression patterns across samples. (D) Kaplan–Meier survival curve comparing high- and low-expression groups of CSNK1E. (E-G) Boxplots showing associations between CSNK1E expression and key clinical features.\u003c/p\u003e","description":"","filename":"Fig.8.png","url":"https://assets-eu.researchsquare.com/files/rs-7839463/v1/1c775d812f393f02149f7df5.png"},{"id":97697484,"identity":"6b7542c8-07af-4084-a2b7-8fbbadcede5d","added_by":"auto","created_at":"2025-12-08 11:44:12","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":4700270,"visible":true,"origin":"","legend":"\u003cp\u003eFunctional and immune analyses of CSNK1E. (A) GO enrichment analysis of genes associated with CSNK1E expression. (B) KEGG pathway enrichment analysis. (C) GSVA showing pathway activity differences between high- and low-expression groups. (D) CIBERSORT analysis of immune cell infiltration associated with CSNK1E expression.\u003c/p\u003e","description":"","filename":"Fig.9.png","url":"https://assets-eu.researchsquare.com/files/rs-7839463/v1/459fc8bdc0f5095ba560b984.png"},{"id":98420889,"identity":"15a99712-123a-48de-875d-b1317cc9576a","added_by":"auto","created_at":"2025-12-17 16:18:02","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":36165509,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7839463/v1/aaf28652-eba2-4b13-86c9-ceafb589aa59.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Prognostic Value and Immune Relevance of Ultrafiltration Failure–Related Genes in Clear Cell Renal Cell Carcinoma","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eClear cell renal cell carcinoma (ccRCC) is the most common subtype of Renal Cell Carcinoma, accounting for approximately 70%–80% of all cases [1]. Despite significant advances in surgical resection, targeted therapies, and immunotherapies in recent years [2], the overall prognosis for ccRCC patients remains unsatisfactory, largely due to the pronounced molecular heterogeneity and clinical tendencies toward recurrence and metastasis [3]. Accordingly, the establishment of precise prognostic assessment systems and the development of individualized therapeutic strategies have become central objectives in both clinical practice and fundamental research.\u003c/p\u003e\n\u003cp\u003ePatients with chronic kidney disease (CKD) have a significantly higher risk of developing malignancies than the general population, with ccRCC being the most common type [4]. Epidemiological studies have shown that dialysis patients have an approximately 3-6-fold higher incidence of renal cancer compared with non-dialysis individuals, and this risk increases progressively with longer dialysis duration [5]. During long-term dialysis, acquired cystic kidney disease (ACKD) frequently develops [6], and cystic transformation has been widely recognized as an important precursor lesion for renal carcinogenesis [7].\u003c/p\u003e\n\u003cp\u003eUltrafiltration failure (UF failure) was originally defined as impaired peritoneal function and fluid clearance during peritoneal dialysis [8]. As a typical complication of long-term peritoneal dialysis, UF failure not only indicates prolonged dialysis vintage but is also associated with more severe chronic inflammation and fibrosis, thereby serving as a clinical signal for identifying patients at high risk of renal cancer. Notably, the pathological processes underlying UF failure overlap substantially with the molecular mechanisms of tumorigenesis, including epithelial–mesenchymal transition (EMT) [9], oxidative stress [10], and the formation of a chronic inflammatory microenvironment [11]. These alterations play critical roles in the invasion, metastasis, and immune regulation of ccRCC [12, 13].\u003c/p\u003e\n\u003cp\u003eTherefore, ultrafiltration failure–related genes (UFFRGs) may act not only as molecular markers of peritoneal dysfunction but also as important regulators in the initiation and progression of ccRCC. However, systematic studies investigating the prognostic value of UFFRGs in ccRCC are still lacking.\u003c/p\u003e\n\u003cp\u003eAgainst this backdrop, the present study integrates single-cell RNA sequencing data with bulk RNA sequencing data to construct a stable and reliable prognostic risk prediction model,termed UF failure–related genes’ signature (UFFRGS), employing 117 machine learning algorithms. Moreover, we further explore the associations between the model and the tumor immune microenvironment, as well as drug sensitivity profiles, with the aim of providing novel theoretical insights and potential therapeutic targets for both the prognostic evaluation and personalized treatment of ccRCC.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cp\u003e\u003cstrong\u003e2.1. Sources and preprocessing of datasets\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe transcriptional profiles and corresponding clinical information were obtained from the Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/) under the TCGA-KIRC project. Following the study design, the TCGA-KIRC cohort was randomly divided into a training set (70%) and an internal validation set (30%) for model development and internal validation, respectively.\u003c/p\u003e\n\u003cp\u003eFor external validation, the study incorporated the E-MTAB-1980 dataset from the ArrayExpress database (https://www.ebi.ac.uk/arrayexpress/), which contains gene expression profiles and follow-up clinical data from patients with ccRCC. To enhance the robustness of validation, the TCGA internal validation set was combined with the E-MTAB-1980 dataset to establish an independent validation cohort.\u003c/p\u003e\n\u003cp\u003eIn addition, single-cell RNA sequencing (scRNA-seq) data were retrieved from the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE248762. This dataset included samples from six long-term dialysis patients without UF failure and four long-term dialysis patients with UF failure.\u003c/p\u003e\n\u003cp\u003eTo ensure cross-platform comparability, gene expression levels from both TCGA-KIRC and E-MTAB-1980 (originally quantified as FPKM, Fragments Per Kilobase of transcript per Million mapped reads) were uniformly converted to TPM (Transcripts Per Million), followed by log2(TPM + 1) transformation for downstream analyses.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2. Model Construction and Validation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo establish an accurate and robust prognostic prediction system, we employed an \u003cstrong\u003eensemble learning strategy\u003c/strong\u003e that integrated \u003cstrong\u003eten distinct machine learning algorithms\u003c/strong\u003e in various configurations, yielding a total of \u003cstrong\u003e117 modeling approaches\u003c/strong\u003e. The algorithms included \u003cstrong\u003estepwise Cox regression, Lasso regression, Ridge regression, CoxBoost, random survival forest (RSF), elastic\u0026nbsp;\u003c/strong\u003enetwork\u003cstrong\u003e\u0026nbsp;(Enet), partial least squares\u003c/strong\u003e regression for Cox\u003cstrong\u003e\u0026nbsp;(plsRcox), generalized boosted regression modeling (GBM), supervised principal components (SuperPC), and survival support vector machine (survival-SVM)\u003c/strong\u003e. Notably, several methods—such as \u003cstrong\u003eLasso, stepwise Cox, RSF, and CoxBoost\u003c/strong\u003e—incorporated intrinsic feature selection mechanisms, thereby enabling the automatic identification of key prognostic variables during model training.\u003c/p\u003e\n\u003cp\u003eModel development were conducted in a stepwise manner.\u0026nbsp;\u003c/p\u003e\n\u003col\u003e\n \u003cli\u003e\u003cstrong\u003eCandidate gene screening\u003c/strong\u003e\u003cstrong\u003e：\u003c/strong\u003e\u003cstrong\u003eGenes significantly associated with prognosis and related to ultrafiltration failure (UFFRGs) were identified as initial candidates\u0026nbsp;\u003c/strong\u003ein the TCGA-KIRC-traning cohort using univariate Cox regression.\u003c/li\u003e\n \u003cli\u003eModel construction:A total of 117 algorithmic combinations were applied to the candidate gene set.The modeling process was conducted under a 10-fold cross-validation framework, ensuring comprehensive variable selection and multi-model development.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003ePerformance evaluation:Model performance was assessed using both internal validation (TCGA) and external validation (E-MTAB-1980) cohorts.Harrell’s concordance index (C-index) was used to quantify predictive accuracy.\u003c/strong\u003e\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003e\u003cstrong\u003e(d) Optimal model selection:The model with the highest average C-index across validation cohorts was selected.This final model, termed\u0026nbsp;\u003c/strong\u003eUF failure–related genes’ signature (UFFRGS)\u003cstrong\u003e, demonstrated both superior predictive performance and potential clinical applicability.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor \u003cstrong\u003eclinical risk stratification\u003c/strong\u003e, patients were categorized into \u003cstrong\u003ehigh-risk\u003c/strong\u003e and \u003cstrong\u003elow-risk\u003c/strong\u003e groups according to the \u003cstrong\u003eoptimal cut-off score\u003c/strong\u003e determined using the \u003cstrong\u003esurvminer\u003c/strong\u003e R package. \u003cstrong\u003eKaplan–Meier survival curves\u003c/strong\u003e was then performed to compare overall survival between the two risk groups, with statistical significance assessed by the \u003cstrong\u003elog-rank test\u003c/strong\u003e [14-16]. To further evaluate predictive performance, \u003cstrong\u003etime-dependent ROC curves\u003c/strong\u003e were generated [17, 18], providing a dynamic assessment of model accuracy across multiple follow-up intervals.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3. Enrichment analysis in high‑ and low‑risk groups\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFollowing application of the optimal prognostic model, patients were stratified into high-risk and low-risk groups based on the predetermined cut-off score. To investigate the biological differences between the two cohorts, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed to identify differentially activated biological processes, molecular functions, and signaling pathways. In addition, Gene Set Variation Analysis (GSVA) was applied to calculate enrichment scores for each sample across a wide range of functional pathways, allowing for a comprehensive assessment of pathway activity differences between the high-risk and low-risk cohorts.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4. Immune infiltration analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe ESTIMATE package in R was applied to calculate the stromal score, immune score, and overall ESTIMATE score for each sample, thereby evaluating the cellular composition of the tumor microenvironment (TME). To further characterize the immune landscape of ccRCC patients, we utilized the CIBERSORT algorithm [19] in combination with single-sample Gene Set Enrichment Analysis (ssGSEA) [20] to quantitatively assess the levels of immune cell infiltration. Based on the ssGSEA results, we then compared the expression profiles of human leukocyte antigen (HLA) family genes between high-risk and low-risk groups, and further examined differences in immune function across the two cohorts.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.5. Tumor Mutation Burden and Drug Sensitivity Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSomatic mutation data were systematically analyzed using the \u003cstrong\u003e“maftools” R package\u003c/strong\u003e, and differences in \u003cstrong\u003etumor mutation burden (TMB)\u003c/strong\u003e between the high-risk and low-risk subgroups were compared. To further explore potential therapeutic implications, the \u003cstrong\u003e“oncoPredict” R package\u003c/strong\u003e was employed to predict the sensitivity of each cohort to a panel of anti-tumor drugs, thereby identifying potential treatment strategies.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.6. Identification and Functional Analysis of Core Genes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe \u003cstrong\u003ecore gene\u003c/strong\u003e was initially selected using \u003cstrong\u003eunivariate Cox regression analysis\u003c/strong\u003e. Based on its expression levels, patients were stratified into \u003cstrong\u003ehigh-expression\u003c/strong\u003e and \u003cstrong\u003elow-expression\u003c/strong\u003e groups. The \u003cstrong\u003eprognostic significance\u003c/strong\u003e of the core gene was subsequently evaluated using \u003cstrong\u003eKaplan–Meier survival curve analysis\u003c/strong\u003e. In addition, \u003cstrong\u003egrouped bar plots\u003c/strong\u003e incorporating patients’ clinical characteristics were generated to examine differences in core gene expression across various \u003cstrong\u003eclinicopathological parameters\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eTo further explore the potential biological roles of the core gene, \u003cstrong\u003eGO\u003c/strong\u003e and\u003cstrong\u003e\u0026nbsp;KEGG enrichment analyses\u003c/strong\u003e were performed. \u003cstrong\u003eGSVA\u003c/strong\u003e was also applied to assess associated \u003cstrong\u003ebiological processes\u003c/strong\u003e and \u003cstrong\u003esignaling pathways\u003c/strong\u003e. Moreover, by integrating the core gene expression data with \u003cstrong\u003eimmune infiltration profiles\u003c/strong\u003e, the correlation between the core gene and the \u003cstrong\u003etumor immune microenvironment\u003c/strong\u003e was evaluated.\u003c/p\u003e"},{"header":"3. Results","content":"\u003cp\u003e\u003cstrong\u003e3.1. Identification of Ultrafiltration Failure\u003c/strong\u003e\u003cstrong\u003e–\u003c/strong\u003e\u003cstrong\u003eRelated Genes Using scRNA-seq Data\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUsing the publicly available dataset GSE248762, we analyzed samples from long vintage patients without ultrafiltration failure (LV_NOT_UF) and those who developed ultrafiltration failure (LV_UF). The single-cell RNA sequencing data were first aggregated into \u003cstrong\u003epseudobulk expression profiles\u003c/strong\u003e, followed by \u003cstrong\u003edifferential gene expression analysis\u003c/strong\u003e. Applying stringent filtering criteria—\u003cstrong\u003elog₂ fold change (logFC) \u0026gt; 1\u003c/strong\u003e, \u003cstrong\u003eadjusted p-value (adj.p) \u0026lt; 0.05\u003c/strong\u003e, and \u003cstrong\u003elog₂-transformed average expression \u0026gt; 10\u003c/strong\u003e—we identified a set of \u003cstrong\u003e168 genes significantly upregulated in the LV_UF cohort\u003c/strong\u003e. The results were visualized using \u003cstrong\u003evolcano plots\u003c/strong\u003e and \u003cstrong\u003eMA plots (Fig. 1A, B)\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eSubsequent \u003cstrong\u003eGSVA\u003c/strong\u003e revealed that this high-expression gene set was significantly enriched in pathways related to \u003cstrong\u003esignal transduction\u003c/strong\u003e and \u003cstrong\u003ehormonal regulation (Fig.1 C)\u003c/strong\u003e. To further refine the candidate genes, an \u003cstrong\u003eintersection analysis\u003c/strong\u003e was performed between these 168 upregulated genes and gene sets from the \u003cstrong\u003eTCGA-KIRC\u003c/strong\u003e and \u003cstrong\u003eE-MTAB-1980\u003c/strong\u003e datasets, resulting in \u003cstrong\u003e162 genes consistently upregulated in LV_UF samples\u003c/strong\u003e. These genes were designated as \u003cstrong\u003eUFFRGs\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eFig. 1 Differential gene expression and pathway enrichment between high- and low-expression groups. (A) Volcano plot showing the distribution of differentially expressed genes between the high-expression and low-expression groups. (B) MA plot displaying the relationship between mean expression and fold change of genes in the two groups. (C) GSVA enrichment analysis highlighting the differential activation of biological pathways between the high-expression and low-expression groups.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2. Comprehensive Development and Evaluation of the Prognostic Model\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInitially, \u003cstrong\u003eunivariate Cox regression analysis\u003c/strong\u003e was performed on the \u003cstrong\u003e162 UFFRGs\u003c/strong\u003e, identifying \u003cstrong\u003e76 genes\u003c/strong\u003e significantly associated with prognosis. Subsequently, an \u003cstrong\u003eintegrated machine learning strategy\u003c/strong\u003e was applied to systematically evaluate \u003cstrong\u003e117 algorithmic combinations\u003c/strong\u003e based on these genes, with the goal of identifying the most predictive and stable model (Fig.2 A). Among all models, the top five exhibited \u003cstrong\u003ehigh average C-index\u003c/strong\u003e. Notably, the \u003cstrong\u003e“StepCox[both] + CoxBoost”\u003c/strong\u003e model, which included only \u003cstrong\u003e37 genes\u003c/strong\u003e, was selected as the final signature after comprehensive evaluation, owing to its \u003cstrong\u003esuperior predictive performance and simplicity\u003c/strong\u003e. Finally, we developed UFFRGS based on the model, which demonstrated high predictive accuracy across multiple datasets.\u003c/p\u003e\n\u003cp\u003eUsing the \u003cstrong\u003eoptimal cutoff score\u003c/strong\u003e determined by the \u003cstrong\u003esurvminer\u003c/strong\u003e R package, patients were stratified into \u003cstrong\u003ehigh-risk\u003c/strong\u003e and \u003cstrong\u003elow-risk\u003c/strong\u003e groups. \u003cstrong\u003eKaplan–Meier survival curves\u003c/strong\u003e revealed significantly better survival in the low-risk group compared to the high-risk group (Fig. 2 B, D, F). To further assess predictive performance, \u003cstrong\u003etime-dependent ROC curves\u0026nbsp;\u003c/strong\u003ewere performed (Fig. 2 C, E, G). In the training cohort, the \u003cstrong\u003earea under the curve (AUC)\u003c/strong\u003e values for \u003cstrong\u003e1-, 2-, and 3-year overall survival\u003c/strong\u003e were \u003cstrong\u003e0.846, 0.763, and 0.791\u003c/strong\u003e, respectively. Consistent results were observed in the validation cohort, confirming the \u003cstrong\u003erobust predictive value\u003c/strong\u003e of the model.\u003c/p\u003e\n\u003cp\u003eOverall, these findings indicate that the \u003cstrong\u003eUFFRGS model exhibits strong generalizability\u003c/strong\u003e and holds \u003cstrong\u003econsiderable potential for clinical application\u003c/strong\u003e across multiple independent cohorts.\u003c/p\u003e\n\u003cp\u003eFig. 2 Comprehensive development and validation of the prognostic model. (A) C-index distribution of 117 prognostic models. (B,C) Kaplan–Meier survival curve and time-dependent ROC curve of the optimal model in the TCGA-KIRC training cohort. (D,E) Kaplan–Meier survival curve and time-dependent ROC curve in the TCGA-KIRC validation cohort. (F,G) Kaplan–Meier survival curve and time-dependent ROC curve in the E-MTAB-1980 cohort.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3. Association Between Risk Score and Clinical Characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn clinical practice, clinical characteristics are commonly used for prognostic assessment. We further explored the relationship between the \u003cstrong\u003erisk score (RS)\u003c/strong\u003e and various clinical characteristics. In the \u003cstrong\u003etraining cohort\u003c/strong\u003e, \u003cstrong\u003emultivariate Cox regression analysis\u003c/strong\u003e showed that the RS remained an \u003cstrong\u003eindependent prognostic factor\u003c/strong\u003e for \u003cstrong\u003eoverall survival (OS)\u003c/strong\u003e after adjustment for clinical variables (Fig. 3 A). This finding was consistently validated in the \u003cstrong\u003evalidation cohort (Fig. 3 B)\u003c/strong\u003e, supporting the robustness of the RS as a prognostic biomarker for \u003cstrong\u003eccRCC\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eMoreover, patients with \u003cstrong\u003eadvanced pathological stage (Stage IV)\u003c/strong\u003e, \u003cstrong\u003elarger or more invasive tumors (T3/T4)\u003c/strong\u003e, \u003cstrong\u003edistant metastasis (M1)\u003c/strong\u003e, or \u003cstrong\u003elymph node involvement (N1)\u003c/strong\u003e tended to have significantly \u003cstrong\u003ehigher risk scores (Fig. 3 C-F)\u003c/strong\u003e. These results indicate a strong correlation between the RS and \u003cstrong\u003edisease progression\u003c/strong\u003e, suggesting that the RS may serve as a \u003cstrong\u003evaluable indicator\u003c/strong\u003e for assessing \u003cstrong\u003etumor aggressiveness\u003c/strong\u003e and predicting \u003cstrong\u003eclinical outcomes\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eFig. 3 Clinical relevance of the prognostic risk score. (A) Multivariate Cox regression analysis of risk score and clinical factors in the TCGA-KIRC training cohort. (B) Multivariate Cox regression analysis in the TCGA-KIRC validation cohort. (C–F) Box plots showing the associations between risk score and key clinical characteristics.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.4. Functional Enrichment Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo explore the molecular features distinguishing the \u003cstrong\u003ehigh-risk\u003c/strong\u003e and \u003cstrong\u003elow-risk\u003c/strong\u003e cohorts, we performed \u003cstrong\u003efunctional enrichment analysis\u003c/strong\u003e on genes associated with risk stratification. Using the \u003cstrong\u003eGO\u003c/strong\u003e and \u003cstrong\u003eKEGG\u003c/strong\u003e databases, we systematically evaluated differences in gene expression profiles between the two cohorts (Fig. 4 A, B). \u003cstrong\u003eGO enrichment analysis\u003c/strong\u003e revealed significant functional divergence: genes in the high-risk group were predominantly involved in biological processes such as \u003cstrong\u003esignal release\u003c/strong\u003e, \u003cstrong\u003eextracellular matrix remodeling\u003c/strong\u003e, and \u003cstrong\u003eion homeostasis dysregulation\u003c/strong\u003e. \u003cstrong\u003eKEGG pathway analysis\u003c/strong\u003e further demonstrated that high-risk group genes were significantly enriched in pathways including \u003cstrong\u003eneuroactive ligand–receptor interaction\u003c/strong\u003e and \u003cstrong\u003ecytokine–cytokine receptor interaction\u003c/strong\u003e, suggesting aberrant activation of \u003cstrong\u003eneural signaling\u003c/strong\u003e and \u003cstrong\u003eimmune response pathways\u003c/strong\u003e in this subgroup.\u003c/p\u003e\n\u003cp\u003eTo quantitatively compare \u003cstrong\u003ecancer-related pathway activity\u003c/strong\u003e between the risk groups, \u003cstrong\u003eGSVA\u003c/strong\u003e was conducted to calculate pathway enrichment scores for each sample (Fig. 4 C). The results indicated that the \u003cstrong\u003ehigh-risk group\u003c/strong\u003e showed significant enrichment in \u003cstrong\u003eimmune-related and signal transduction pathways\u003c/strong\u003e, such as the \u003cstrong\u003eKRAS signaling pathway\u003c/strong\u003e, \u003cstrong\u003eIL-6/JAK/STAT3 signaling\u003c/strong\u003e, \u003cstrong\u003einterferon-gamma response\u003c/strong\u003e, and the \u003cstrong\u003ecomplement system\u003c/strong\u003e. In contrast, the \u003cstrong\u003elow-risk group\u003c/strong\u003e exhibited marked enrichment in \u003cstrong\u003emetabolic and signaling pathways\u003c/strong\u003e, including \u003cstrong\u003eheme metabolism\u003c/strong\u003e, \u003cstrong\u003efatty acid metabolism\u003c/strong\u003e, \u003cstrong\u003eTGF-β signaling\u003c/strong\u003e, and the \u003cstrong\u003ePI3K-AKT-mTOR pathway\u003c/strong\u003e, reflecting characteristic alterations in \u003cstrong\u003eenergy metabolism\u003c/strong\u003e and \u003cstrong\u003esignal regulation\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eFig. 4 Pathway Enrichment and Activity Differences Between High- and Low-Risk Groups. (A) GO enrichment analysis showing significantly enriched biological processes, molecular functions, and cellular components. (B) KEGG pathway enrichment highlighting key signaling pathways associated with the risk groups. (C) GSVA analysis depicting pathway activity differences between the high- and low-risk groups.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.5. Immune Infiltration Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBuilding on the established association between \u003cstrong\u003erisk scores\u003c/strong\u003e and immunopathological features, we conducted a comprehensive analysis of the \u003cstrong\u003etumor immune microenvironment\u003c/strong\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe results showed that the \u003cstrong\u003ehigh-risk group\u003c/strong\u003e exhibited significantly elevated \u003cstrong\u003estromal, immune, and ESTIMATE scores\u003c/strong\u003e, accompanied by \u003cstrong\u003ereduced tumor purity (Fig. 5 A)\u003c/strong\u003e. Using the \u003cstrong\u003eCIBERSORT algorithm\u003c/strong\u003e, we quantitatively assessed the distribution of immune cell subsets in the \u003cstrong\u003eTCGA training cohort\u003c/strong\u003e and visualized differences between the risk groups using \u003cstrong\u003eboxplots (Fig. 5 B).\u0026nbsp;\u003c/strong\u003eSpecifically, the \u003cstrong\u003elow-risk group\u003c/strong\u003e displayed higher proportions of \u003cstrong\u003eresting CD4⁺ memory T cells, monocytes, M1/M2 macrophages, and resting mast cells\u003c/strong\u003e, whereas the \u003cstrong\u003ehigh-risk group\u003c/strong\u003e showed increased infiltration of \u003cstrong\u003eCD8⁺ T cells, regulatory T cells (Tregs), and activated CD4⁺ memory T cells\u003c/strong\u003e. These findings were consistently validated in the validation\u003cstrong\u003e\u0026nbsp;cohort (Fig. 5 C)\u003c/strong\u003e, confirming the robustness of the observed immune landscape.\u003c/p\u003e\n\u003cp\u003eFurther analysis of \u003cstrong\u003ehuman leukocyte antigen (HLA) family genes\u003c/strong\u003e revealed significant downregulation of several key HLA genes—including \u003cstrong\u003eHLA-B, HLA-C, and HLA-E\u003c/strong\u003e—in the high-risk group \u003cstrong\u003e(Fig. 5 D)\u003c/strong\u003e. Examination of \u003cstrong\u003eimmune functional activity\u003c/strong\u003e indicated that the low-risk group was characterized by enhanced \u003cstrong\u003emast cell-related functions\u003c/strong\u003e, whereas the high-risk group exhibited heightened activity in \u003cstrong\u003eCD8⁺ T cell\u003c/strong\u003e and \u003cstrong\u003emacrophage-associated functions (Fig. 5 E)\u003c/strong\u003e. Collectively, these results demonstrate distinct immune microenvironment profiles and differential immune response patterns between the two risk groups.\u003c/p\u003e\n\u003cp\u003eFig. 5 Immune infiltration analysis in high- and low-risk groups. (A) Boxplots of stromal, immune, ESTIMATE scores, and tumor purity. (B) CIBERSORT results in the training cohort. (C) CIBERSORT results in the KIRC validation cohort. (D) Differential expression of HLA-related genes. (F) Immune function differences between groups.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.6. Tumor Mutational Burden (TMB) Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe first constructed a comprehensive mutational landscape of the entire cohort (Fig. 6 A). Missense mutations represented the predominant variant type, with C\u0026gt;T substitutions being the most frequent single-nucleotide variation. Among all mutated genes, \u003cem\u003eVHL\u003c/em\u003e exhibited the highest mutation frequency (46%), underscoring its potential role as a pivotal driver in the pathogenesis of ccRCC.\u003c/p\u003e\n\u003cp\u003eWe next compared TMB between the risk subgroups (Fig. 6 B, C). In the training cohort, TMB levels showed no statistically significant differences between high- and low-risk patients. However, in the validation cohort, the high-risk subgroup displayed significantly elevated TMB levels compared with the low-risk subgroup.\u003c/p\u003e\n\u003cp\u003eTo further characterize subgroup-specific mutation patterns, we applied waterfall plots to depict and rank the gene mutation profiles in each risk group (Fig. 6 D-F). Distinct distributions were observed among the top 15 most frequently mutated genes. For example, \u003cem\u003eVHL\u003c/em\u003e, \u003cem\u003ePBRM1\u003c/em\u003e, and \u003cem\u003eSETD2\u003c/em\u003e exhibited heterogeneous mutation profiles between the high- and low-risk groups, suggesting that divergent molecular mechanisms may contribute to the stratified risk classifications.\u003c/p\u003e\n\u003cp\u003eFig.6 Tumor mutational burden (TMB) analysis in different risk groups. (A) Mutation landscape of the entire cohort. (B) TMB differences between high- and low-risk groups in the training cohort. (C) TMB differences in the KIRC validation cohort. (D) Mutation profile of the top 15 genes in the training cohort. (D-F) Mutation profiles of the top 15 genes in high- and low-risk groups.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.7. Drug Sensitivity Analysis\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe next evaluated the differences in drug response between the risk subgroups (Fig. 7). The analysis revealed that patients in the high-risk group exhibited significantly lower half-maximal inhibitory concentration (IC50) values for several agents, including AZD7762, Camptothecin, CDK9 inhibitors, Staurosporine, Topotecan, Vinorelbine, Dactinomycin, and Epirubicin, indicating a heightened sensitivity to these compounds.\u003c/p\u003e\n\u003cp\u003eIn contrast, the low-risk group displayed markedly reduced IC50 values for agents such as AT13148, Cediranib, Daporinad, and Osimertinib, suggesting a preferential sensitivity to these therapies.\u003c/p\u003e\n\u003cp\u003eCollectively, these findings highlight distinct drug sensitivity patterns between the high- and low-risk subgroups. Such differences may provide a foundation for developing risk-adapted therapeutic strategies and support the clinical application of precision medicine in ccRCC.\u003c/p\u003e\n\u003cp\u003eFig.7 Comparison of drug responses between high- and low-risk groups.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.8. Core Gene Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMultivariable Cox regression analysis of the 37 genes incorporated into the prognostic model identified \u003cstrong\u003eCSNK1E\u003c/strong\u003e as a key gene with a significant impact on survival outcomes (Fig. 8 A). According to its expression levels, patients in the TCGA-KIRC cohort were stratified into high- and low-expression groups, with differential expression profiles illustrated by volcano and heatmap plots (Fig. 8 B, C). Kaplan–Meier survival analysis further confirmed that \u003cstrong\u003eCSNK1E\u003c/strong\u003e served as a robust predictor of overall survival (OS): patients with low CSNK1E expression exhibited a significantly more favorable prognosis compared with those in the high-expression group (Fig. 8 D).\u003c/p\u003e\n\u003cp\u003eAssociations between \u003cstrong\u003eCSNK1E\u003c/strong\u003e expression and clinical characteristics were then evaluated (Fig. 8 E-G). No significant differences were observed across pathological stage, M stage, or N stage. However, in terms of T staging, median CSNK1E expression was markedly elevated in T4 tumors compared with T1–T3 tumors, suggesting that CSNK1E expression is more closely linked to local tumor invasiveness than to lymph node or distant metastasis.\u003c/p\u003e\n\u003cp\u003eTo further explore the biological implications of CSNK1E, functional enrichment analyses were conducted. GO analysis revealed that differentially expressed genes (DEGs) in the high-expression group were significantly enriched in biological processes such as steroid metabolic pathways and monoatomic ion channel complexes (Fig. 9 A). KEGG pathway analysis demonstrated that these DEGs were predominantly involved in neuroactive ligand–receptor interactions and hormone-related signaling pathways (Fig. 9 B), highlighting a potential role of CSNK1E in neuroendocrine regulation.\u003c/p\u003e\n\u003cp\u003eGSVA further showed distinct pathway activity patterns between the groups (Fig. 9 C). The high-expression group exhibited marked activation of oncogenic signaling pathways, including the \u003cstrong\u003eWnt/β-catenin signaling pathway\u003c/strong\u003e, which was consistent with their poorer clinical outcomes. Conversely, the low-expression group displayed enhanced activity in multiple metabolic and immune-related pathways.\u003c/p\u003e\n\u003cp\u003eFinally, immune infiltration analysis using the CIBERSORT algorithm revealed distinct immune landscape profiles according to CSNK1E expression levels (Fig. 9 D). Notably, resting CD4⁺ memory T cells and resting NK cells were positively correlated with CSNK1E expression, indicating that CSNK1E may exert an important role in shaping the immune microenvironment.\u003c/p\u003e\n\u003cp\u003eFig. 8 Core gene analysis in the TCGA-KIRC cohort. (A) Multivariate Cox regression analysis of the 37 model-included genes. (B) Volcano plot of differential gene expression. (C) Heatmap showing expression patterns across samples. (D) Kaplan–Meier survival curve comparing high- and low-expression groups of CSNK1E. (E-G) Boxplots showing associations between CSNK1E expression and key clinical features.\u003c/p\u003e\n\u003cp\u003eFig. 9 Functional and immune analyses of CSNK1E. (A) GO enrichment analysis of genes associated with CSNK1E expression. (B) KEGG pathway enrichment analysis. (C) GSVA showing pathway activity differences between high- and low-expression groups. (D) CIBERSORT analysis of immune cell infiltration associated with CSNK1E expression.\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eIn ccRCC, most existing prognostic models rely on bulk transcriptomic data or limited immune-related gene panels [21, 22], which fail to adequately capture tumor cellular heterogeneity and the complexity of the immune microenvironment. Notably, few studies systematically incorporate ultrafiltration failure–related mechanisms. There is, therefore, a pressing need for a robust prognostic tool that integrates single-cell and bulk transcriptomic data while accounting for immune context.\u003c/p\u003e\n\u003cp\u003eIn this study, we developed the UFFRGS by integrating single-cell and bulk transcriptomic datasets. The model, comprising 37 key genes, demonstrated strong prognostic performance across multiple independent cohorts. Patients were stratified into high- and low-risk groups based on risk scores. High-risk patients not only exhibited significantly poorer overall survival but also displayed distinct immune microenvironment characteristics: increased infiltration of CD8⁺ T cells in high-risk patients versus enrichment of M2 macrophages in low-risk patients. While CD8⁺ T cells generally exert antitumor effects, prolonged activation in ccRCC can lead to an exhausted phenotype, adversely affecting prognosis [23]. In contrast, M2 macrophages are associated with immunosuppressive signaling and tumor aggressiveness [24]. These immune patterns provide a biological basis for UFFRGS-based stratification, indicating that the model captures heterogeneity within the tumor immune microenvironment and supporting its potential utility in clinical risk assessment and personalized therapy.\u003c/p\u003e\n\u003cp\u003eTumor mutational burden analysis further revealed heterogeneous mutation patterns in VHL, PBRM1, and SETD2 between risk groups, consistent with prior studies describing the genomic and transcriptomic landscapes of ccRCC, including mutations in key driver genes and alterations in immune evasion–related pathways [25]. Moreover, correlations between risk scores, clinical features, and drug sensitivity provide an integrated framework that links molecular characteristics to potential therapeutic responses.\u003c/p\u003e\n\u003cp\u003eAmong the 37 core genes, CSNK1E emerged as a notable candidate. As a Casein kinase 1 family member [26], CSNK1E participates in cell cycle regulation, signal transduction, and tumorigenesis [27, 28]. Although its roles in neural and endocrine regulation are documented [29, 30], its function in ccRCC, particularly in relation to UF failure, remains unclear. Our findings indicate that CSNK1E overexpression correlates with poor survival and is associated with locally invasive T4 tumors rather than lymph node metastasis, suggesting a key role in local tumor progression and microenvironmental regulation.\u003c/p\u003e\n\u003cp\u003ePathway analysis suggested that CSNK1E may promote tumor progression by activating Wnt/β-catenin signaling and perturbing hormone-related and neuroactive ligand–receptor interactions, consistent with the well-established role of Wnt/β-catenin in ccRCC progression [31]. Previous studies report that CK1δ/ε inhibition suppresses ccRCC cell proliferation and enhances treatment sensitivity [32], indicating potential druggability of CSNK1E. As a circadian rhythm gene, CSNK1E dysregulation may disrupt rhythm-dependent metabolic and immune homeostasis [33, 34], aligning with its identification as a prognostic factor in our integrated analysis. Notably, high CSNK1E expression correlated with increased resting CD4⁺ memory T cells and NK cells, suggesting a “hypoactive” immune state [35]. While these findings are correlative, functional experiments are needed to validate the direct effects of CSNK1E on Wnt signaling, immune regulation, and circadian rhythm in ccRCC.\u003c/p\u003e\n\u003cp\u003eDespite these insights, this study has limitations. First, it is based on retrospective data and predominantly computational analyses, lacking prospective clinical validation to confirm UFFRGS's prognostic value. Second, functional and immune analyses are inferred from bioinformatics; in vitro experiments such as gene knockdown/overexpression and immunological assays are required to elucidate the roles of CSNK1E and other core genes. Additionally, drug sensitivity predictions rely on computational models without validation in patient-derived or pharmacological data, limiting immediate clinical translatability.\u003c/p\u003e\n\u003cp\u003eFuture studies should address these limitations by validating UFFRGS in prospective cohorts, supporting computational predictions with targeted functional experiments, and integrating public drug-response datasets or limited clinical samples for drug sensitivity validation. Further incorporation of multi-omic data, including epigenomics, metabolomics, and spatial transcriptomics, may optimize risk modeling and expand its application across diverse renal cancer subtypes.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eUsing machine learning algorithms, this study constructed an ultrafiltration failure–related gene prognostic model for clear cell renal cell carcinoma (UFFRGS) and validated its robust predictive performance across multiple independent cohorts. The results demonstrated that high-risk patients not only exhibited significantly poorer overall survival but also displayed distinct immune microenvironment features and key driver gene mutation patterns. The core gene CSNK1E may contribute to tumor progression through multiple signaling pathways. Overall, this study provides novel insights into the potential molecular links between ultrafiltration failure and clear cell renal cell carcinoma, and offers a theoretical basis for clinical risk stratification and the development of personalized therapeutic strategies.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated and analysed during the current study can obtained from the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/), the Cancer Genome Atlas (TCGA, \u0026nbsp;https://portal.gdc.cancer.gov/), and the ArrayExpress database (https://www.ebi.ac.uk/arrayexpress/).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by the Medical Research Project of the Jiangsu Provincial Health Commission (No. Z2023076).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors' contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eYL was responsible for data collection, bioinformatics analysis, and drafting of the manuscript. YT and JW assisted with data preprocessing, statistical analysis, and figure preparation. WL and SJ contributed to literature review and interpretation of biological findings. RJ and LS participated in model validation and result discussion. XR provided technical support and made critical revisions to the manuscript. JY provided overall guidance and performed the final review of the paper. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eHsieh JJ, Purdue MP, Signoretti S, Swanton C, Albiges L, Schmidinger M, Heng DY, Larkin J, Ficarra V: \u003cstrong\u003eRenal cell carcinoma\u003c/strong\u003e. \u003cem\u003eNat Rev Dis Primers \u003c/em\u003e2017, \u003cstrong\u003e3\u003c/strong\u003e:17009.\u003c/li\u003e\n\u003cli\u003eSchiavoni V, Campagna R, Pozzi V, Cecati M, Milanese G, Sartini D, Salvolini E, Galosi AB, Emanuelli M: \u003cstrong\u003eRecent Advances in the Management of Clear Cell Renal Cell Carcinoma: Novel Biomarkers and Targeted Therapies\u003c/strong\u003e. \u003cem\u003eCancers (Basel) \u003c/em\u003e2023, \u003cstrong\u003e15\u003c/strong\u003e(12).\u003c/li\u003e\n\u003cli\u003eLi Y, Lih TM, Dhanasekaran SM, Mannan R, Chen L, Cieslik M, Wu Y, Lu RJ, Clark DJ, Kołodziejczak I\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eHistopathologic and proteogenomic heterogeneity reveals features of clear cell renal cell carcinoma aggressiveness\u003c/strong\u003e. \u003cem\u003eCancer Cell \u003c/em\u003e2023, \u003cstrong\u003e41\u003c/strong\u003e(1):139-163.e117.\u003c/li\u003e\n\u003cli\u003e\u0026Aring;kerlund J, Holmberg E, Lindblad P, Stendahl M, Ljungberg B, Thorstenson A, Lundstam S: \u003cstrong\u003eIncreased risk for renal cell carcinoma in end stage renal disease - a population-based case-control study\u003c/strong\u003e. \u003cem\u003eScand J Urol \u003c/em\u003e2021, \u003cstrong\u003e55\u003c/strong\u003e(3):209-214.\u003c/li\u003e\n\u003cli\u003eLevine E: \u003cstrong\u003eRenal cell carcinoma in uremic acquired renal cystic disease: incidence, detection, and management\u003c/strong\u003e. \u003cem\u003eUrol Radiol \u003c/em\u003e1992, \u003cstrong\u003e13\u003c/strong\u003e(4):203-210.\u003c/li\u003e\n\u003cli\u003eKabaria R, Klaassen Z, Terris MK: \u003cstrong\u003eRenal cell carcinoma: links and risks\u003c/strong\u003e. \u003cem\u003eInt J Nephrol Renovasc Dis \u003c/em\u003e2016, \u003cstrong\u003e9\u003c/strong\u003e:45-52.\u003c/li\u003e\n\u003cli\u003ePrzybycin CG, Harper HL, Reynolds JP, Magi-Galluzzi C, Nguyen JK, Wu A, Sangoi AR, Liu PS, Umar S, Mehra R\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eAcquired Cystic Disease-associated Renal Cell Carcinoma (ACD-RCC): A Multiinstitutional Study of 40 Cases With Clinical Follow-up\u003c/strong\u003e. \u003cem\u003eAm J Surg Pathol \u003c/em\u003e2018, \u003cstrong\u003e42\u003c/strong\u003e(9):1156-1165.\u003c/li\u003e\n\u003cli\u003eSmit W, Parikova A, Krediet RT: \u003cstrong\u003eUltrafiltration failure in peritoneal dialysis. Causes and clinical consequences\u003c/strong\u003e. \u003cem\u003eMinerva Urol Nefrol \u003c/em\u003e2005, \u003cstrong\u003e57\u003c/strong\u003e(3):165-174.\u003c/li\u003e\n\u003cli\u003eAroeira LS, Aguilera A, S\u0026aacute;nchez-Tomero JA, Bajo MA, del Peso G, Jim\u0026eacute;nez-Heffernan JA, Selgas R, L\u0026oacute;pez-Cabrera M: \u003cstrong\u003eEpithelial to mesenchymal transition and peritoneal membrane failure in peritoneal dialysis patients: pathologic significance and potential therapeutic interventions\u003c/strong\u003e. \u003cem\u003eJ Am Soc Nephrol \u003c/em\u003e2007, \u003cstrong\u003e18\u003c/strong\u003e(7):2004-2013.\u003c/li\u003e\n\u003cli\u003eOh SH, Yook JM, Jung HY, Choi JY, Cho JH, Park SH, Kim CD, Kim YL, Lim JH: \u003cstrong\u003eAutophagy caused by oxidative stress promotes TGF-\u0026beta;1-induced epithelial-to-mesenchymal transition in human peritoneal mesothelial cells\u003c/strong\u003e. \u003cem\u003eCell Death Dis \u003c/em\u003e2024, \u003cstrong\u003e15\u003c/strong\u003e(5):365.\u003c/li\u003e\n\u003cli\u003eBaroni G, Schuinski A, de Moraes TP, Meyer F, Pecoits-Filho R: \u003cstrong\u003eInflammation and the peritoneal membrane: causes and impact on structure and function during peritoneal dialysis\u003c/strong\u003e. \u003cem\u003eMediators Inflamm \u003c/em\u003e2012, \u003cstrong\u003e2012\u003c/strong\u003e:912595.\u003c/li\u003e\n\u003cli\u003eFiaschi T, Chiarugi P: \u003cstrong\u003eOxidative stress, tumor microenvironment, and metabolic reprogramming: a diabolic liaison\u003c/strong\u003e. \u003cem\u003eInt J Cell Biol \u003c/em\u003e2012, \u003cstrong\u003e2012\u003c/strong\u003e:762825.\u003c/li\u003e\n\u003cli\u003eSuarez-Carmona M, Lesage J, Cataldo D, Gilles C: \u003cstrong\u003eEMT and inflammation: inseparable actors of cancer progression\u003c/strong\u003e. \u003cem\u003eMol Oncol \u003c/em\u003e2017, \u003cstrong\u003e11\u003c/strong\u003e(7):805-823.\u003c/li\u003e\n\u003cli\u003eLiu H, Tang T: \u003cstrong\u003eMAPK signaling pathway-based glioma subtypes, machine-learning risk model, and key hub proteins identification\u003c/strong\u003e. \u003cem\u003eSci Rep \u003c/em\u003e2023, \u003cstrong\u003e13\u003c/strong\u003e(1):19055.\u003c/li\u003e\n\u003cli\u003eLiu H, Weng J: \u003cstrong\u003eA comprehensive bioinformatic analysis of cyclin-dependent kinase 2 (CDK2) in glioma\u003c/strong\u003e. \u003cem\u003eGene \u003c/em\u003e2022, \u003cstrong\u003e822\u003c/strong\u003e:146325.\u003c/li\u003e\n\u003cli\u003eLiu H, Weng J, Huang CL, Jackson AP: \u003cstrong\u003eIs the voltage-gated sodium channel \u0026beta;3 subunit (SCN3B) a biomarker for glioma?\u003c/strong\u003e \u003cem\u003eFunct Integr Genomics \u003c/em\u003e2024, \u003cstrong\u003e24\u003c/strong\u003e(5):162.\u003c/li\u003e\n\u003cli\u003eLi Y, Liu H: \u003cstrong\u003eClinical powers of Aminoacyl tRNA Synthetase Complex Interacting Multifunctional Protein 1 (AIMP1) for head-neck squamous cell carcinoma\u003c/strong\u003e. \u003cem\u003eCancer Biomark \u003c/em\u003e2022, \u003cstrong\u003e34\u003c/strong\u003e(3):359-374.\u003c/li\u003e\n\u003cli\u003eLiu H, Weng J: \u003cstrong\u003eA Pan-Cancer Bioinformatic Analysis of RAD51 Regarding the Values for Diagnosis, Prognosis, and Therapeutic Prediction\u003c/strong\u003e. \u003cem\u003eFront Oncol \u003c/em\u003e2022, \u003cstrong\u003e12\u003c/strong\u003e:858756.\u003c/li\u003e\n\u003cli\u003eNewman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, Hoang CD, Diehn M, Alizadeh AA: \u003cstrong\u003eRobust enumeration of cell subsets from tissue expression profiles\u003c/strong\u003e. \u003cem\u003eNat Methods \u003c/em\u003e2015, \u003cstrong\u003e12\u003c/strong\u003e(5):453-457.\u003c/li\u003e\n\u003cli\u003eH\u0026auml;nzelmann S, Castelo R, Guinney J: \u003cstrong\u003eGSVA: gene set variation analysis for microarray and RNA-seq data\u003c/strong\u003e. \u003cem\u003eBMC Bioinformatics \u003c/em\u003e2013, \u003cstrong\u003e14\u003c/strong\u003e:7.\u003c/li\u003e\n\u003cli\u003eHua S, Xie Z, Zhang Y, Wu L, Shi F, Wang X, Xia S, Dong S, Jiang J: \u003cstrong\u003eIdentification and validation of an immune-related gene prognostic signature for clear cell renal carcinoma\u003c/strong\u003e. \u003cem\u003eFront Immunol \u003c/em\u003e2022, \u003cstrong\u003e13\u003c/strong\u003e:869297.\u003c/li\u003e\n\u003cli\u003eRen S, Wang W, Shen H, Zhang C, Hao H, Sun M, Wang Y, Zhang X, Lu B, Chen C\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eDevelopment and Validation of a Clinical Prognostic Model Based on Immune-Related Genes Expressed in Clear Cell Renal Cell Carcinoma\u003c/strong\u003e. \u003cem\u003eFront Oncol \u003c/em\u003e2020, \u003cstrong\u003e10\u003c/strong\u003e:1496.\u003c/li\u003e\n\u003cli\u003eQi Y, Xia Y, Lin Z, Qu Y, Qi Y, Chen Y, Zhou Q, Zeng H, Wang J, Chang Y\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eTumor-infiltrating CD39(+)CD8(+) T cells determine poor prognosis and immune evasion in clear cell renal cell carcinoma patients\u003c/strong\u003e. \u003cem\u003eCancer Immunol Immunother \u003c/em\u003e2020, \u003cstrong\u003e69\u003c/strong\u003e(8):1565-1576.\u003c/li\u003e\n\u003cli\u003eLiu J, Geng X, Hou J, Wu G: \u003cstrong\u003eNew insights into M1/M2 macrophages: key modulators in cancer progression\u003c/strong\u003e. \u003cem\u003eCancer Cell Int \u003c/em\u003e2021, \u003cstrong\u003e21\u003c/strong\u003e(1):389.\u003c/li\u003e\n\u003cli\u003eHakimi AA, Ostrovnaya I, Reva B, Schultz N, Chen YB, Gonen M, Liu H, Takeda S, Voss MH, Tickoo SK\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eAdverse outcomes in clear cell renal cell carcinoma with mutations of 3p21 epigenetic regulators BAP1 and SETD2: a report by MSKCC and the KIRC TCGA research network\u003c/strong\u003e. \u003cem\u003eClin Cancer Res \u003c/em\u003e2013, \u003cstrong\u003e19\u003c/strong\u003e(12):3259-3267.\u003c/li\u003e\n\u003cli\u003eKnippschild U, Gocht A, Wolff S, Huber N, L\u0026ouml;hler J, St\u0026ouml;ter M: \u003cstrong\u003eThe casein kinase 1 family: participation in multiple cellular processes in eukaryotes\u003c/strong\u003e. \u003cem\u003eCell Signal \u003c/em\u003e2005, \u003cstrong\u003e17\u003c/strong\u003e(6):675-689.\u003c/li\u003e\n\u003cli\u003eYang WS, Stockwell BR: \u003cstrong\u003eInhibition of casein kinase 1-epsilon induces cancer-cell-selective, PERIOD2-dependent growth arrest\u003c/strong\u003e. \u003cem\u003eGenome Biol \u003c/em\u003e2008, \u003cstrong\u003e9\u003c/strong\u003e(6):R92.\u003c/li\u003e\n\u003cli\u003eLin SH, Lin YM, Yeh CM, Chen CJ, Chen MW, Hung HF, Yeh KT, Yang SF: \u003cstrong\u003eCasein kinase 1 epsilon expression predicts poorer prognosis in low T-stage oral cancer patients\u003c/strong\u003e. \u003cem\u003eInt J Mol Sci \u003c/em\u003e2014, \u003cstrong\u003e15\u003c/strong\u003e(2):2876-2891.\u003c/li\u003e\n\u003cli\u003eBryant CD, Parker CC, Zhou L, Olker C, Chandrasekaran RY, Wager TT, Bolivar VJ, Loudon AS, Vitaterna MH, Turek FW\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eCsnk1e is a genetic regulator of sensitivity to psychostimulants and opioids\u003c/strong\u003e. \u003cem\u003eNeuropsychopharmacology \u003c/em\u003e2012, \u003cstrong\u003e37\u003c/strong\u003e(4):1026-1035.\u003c/li\u003e\n\u003cli\u003eVarghese RT, Young S, Pham L, Liang Y, Pridham KJ, Guo S, Murphy S, Kelly DF, Sheng Z: \u003cstrong\u003eCasein Kinase 1 Epsilon Regulates Glioblastoma Cell Survival\u003c/strong\u003e. \u003cem\u003eSci Rep \u003c/em\u003e2018, \u003cstrong\u003e8\u003c/strong\u003e(1):13621.\u003c/li\u003e\n\u003cli\u003eXu Q, Krause M, Samoylenko A, Vainio S: \u003cstrong\u003eWnt Signaling in Renal Cell Carcinoma\u003c/strong\u003e. \u003cem\u003eCancers (Basel) \u003c/em\u003e2016, \u003cstrong\u003e8\u003c/strong\u003e(6).\u003c/li\u003e\n\u003cli\u003eLin YC, Sun DP, Hsieh TH, Chen CH: \u003cstrong\u003eTargeting CK1\u0026delta; and CK1\u0026epsilon; as a New Therapeutic Approach for Clear Cell Renal Cell Carcinoma\u003c/strong\u003e. \u003cem\u003ePharmacology \u003c/em\u003e2024, \u003cstrong\u003e109\u003c/strong\u003e(6):330-340.\u003c/li\u003e\n\u003cli\u003eFulcher LJ, Sapkota GP: \u003cstrong\u003eFunctions and regulation of the serine/threonine protein kinase CK1 family: moving beyond promiscuity\u003c/strong\u003e. \u003cem\u003eBiochem J \u003c/em\u003e2020, \u003cstrong\u003e477\u003c/strong\u003e(23):4603-4621.\u003c/li\u003e\n\u003cli\u003eLi S, Wang X, Wang Q, Ding K, Chen X, Zhao Y, Gao Y, Wang Y: \u003cstrong\u003eEffects and Prognostic Values of Circadian Genes CSNK1E/GNA11/KLF9/THRAP3 in Kidney Renal Clear Cell Carcinoma via a Comprehensive Analysis\u003c/strong\u003e. \u003cem\u003eBioengineering (Basel) \u003c/em\u003e2022, \u003cstrong\u003e9\u003c/strong\u003e(7).\u003c/li\u003e\n\u003cli\u003eGiraldo NA, Becht E, Remark R, Damotte D, Saut\u0026egrave;s-Fridman C, Fridman WH: \u003cstrong\u003eThe immune contexture of primary and metastatic human tumours\u003c/strong\u003e. \u003cem\u003eCurr Opin Immunol \u003c/em\u003e2014, \u003cstrong\u003e27\u003c/strong\u003e:8-15.\u003c/li\u003e\n\u003c/ol\u003e\n"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-urology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"buro","sideBox":"Learn more about [BMC Urology](http://bmcurol.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/buro/default.aspx","title":"BMC Urology","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7839463/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7839463/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Clear cell renal cell carcinoma (ccRCC) is a common malignancy characterized by intratumoral heterogeneity, affecting tumor progression and immune regulation. Although ultrafiltration (UF) failure has been clinically associated with kidney disease and ccRCC, the underlying molecular mechanisms remain poorly understood. Here, we developed a UF failure–related gene signature (UFFRGS) by integrating single-cell and bulk RNA sequencing data from TCGA-KIRC and E-MTAB-1980 cohorts. A total of 162 UF failure–related genes were identified, and 117 machine learning algorithm combinations were applied to construct a robust prognostic model comprising 37 key genes. UFFRGS stratified patients into high- and low-risk groups with significant survival differences. High-risk patients exhibited increased infiltration of CD8⁺ T cells, regulatory T cells, and activated CD4⁺ memory T cells, while low-risk patients showed enrichment of M2 macrophages and monocytes, reflecting immune heterogeneity. Mutation analysis revealed distinct patterns in VHL, PBRM1, and SETD2 between risk groups. Drug sensitivity analysis further indicated differential responses to multiple chemotherapeutic agents, providing potential guidance for individualized therapy. Among the core genes, CSNK1E was highlighted as a key prognostic factor, with high expression associated with poor survival and elevated in T4 tumors. GSVA suggested CSNK1E may promote tumor progression via Wnt/β-catenin and hormone-related pathways. In summary, UFFRGS is a stable, reliable prognostic tool that captures molecular and immune heterogeneity in ccRCC, offering insights for risk stratification and potential therapeutic intervention.","manuscriptTitle":"Prognostic Value and Immune Relevance of Ultrafiltration Failure–Related Genes in Clear Cell Renal Cell Carcinoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-08 11:44:08","doi":"10.21203/rs.3.rs-7839463/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2025-12-04T13:56:36+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-11-10T21:38:54+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-11-07T06:51:21+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-11-07T06:51:20+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Urology","date":"2025-10-12T08:42:46+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-urology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"buro","sideBox":"Learn more about [BMC Urology](http://bmcurol.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/buro/default.aspx","title":"BMC Urology","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0bfcfa15-be11-43f4-9bd5-62c4956282f1","owner":[],"postedDate":"December 8th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-12-08T11:44:08+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-08 11:44:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7839463","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7839463","identity":"rs-7839463","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}