Introduction

Clear cell renal cell carcinoma (KIRC), the most prevalent histological subtype of kidney cancer, accounts for approximately 75% of all renal malignancies1,2. As a primary neoplasm originating in the renal cortex, KIRC is often detected incidentally during routine imaging examinations including ultrasonography or computed tomography3. However, the absence of systematic screening protocols has resulted in a considerable number of cases being diagnosed at advanced stages upon symptom manifestation4. The disease carries a concerning five-year survival rate ranging between 50% and 70%5. Clinical management of advanced KIRC presents substantial therapeutic challenges, particularly following disease progression when treatment selection becomes crucial for both quality of life and overall survival (OS)6,7,8. Thus, KIRC has emerged as a major health concern, necessitating the urgent exploration of effective strategies to optimize therapeutic decision-making and enhance patient outcomes.

Mitochondrial autophagy (mitophagy) represents a highly selective quality control mechanism that specifically identifies and eliminates damaged or dysfunctional mitochondria through lysosomal degradation9,10. This evolutionarily conserved process plays a pivotal role in maintaining cellular homeostasis by: (1) removing compromised mitochondria, (2) preventing excessive reactive oxygen species (ROS) generation, and (3) sustaining optimal cellular bioenergetics11,12. Emerging evidence has established mitophagy dysregulation as a pathogenic factor in diverse disease states, including neurodegenerative diseases, cardiovascular pathologies, and various malignancies13,14,15. Particularly in clear cell renal cell carcinoma (KIRC), recent studies have revealed significant associations between autophagic processes and tumor progression, positioning mitophagy as both a critical pathophysiological mediator and a potential therapeutic target16,17. While the precise mechanistic involvement of mitophagy in KIRC pathogenesis requires further elucidation, its emerging role in oncogenic modulation presents compelling opportunities for targeted cancer intervention strategies.

This study employed an integrative analytical approach to investigate KIRC through differential expression profiling combined with Cox proportional hazards and LASSO regression analyses. We developed a prognostic risk stratification model based on mitophagy-related genes (MRGs), with rigorous validation performed across both internal and independent external cohorts. Subsequently, we constructed and validated a clinically applicable nomogram incorporating the MRG signature and key clinical parameters to enable individualized survival probability estimation. Comprehensive characterization of the tumor immune microenvironment associated with these prognostic genes was conducted, including evaluation of potential immunotherapeutic responsiveness and pharmacological susceptibility. Functional validation studies using METTL24 as a prototype MRG demonstrated significant suppression of KIRC cell proliferation and migratory capacity in vitro. Our findings provide a novel framework for enhancing prognostic assessment and therapeutic strategy optimization in KIRC management. This work advances the paradigm of precision oncology by establishing mitophagy-related biomarkers that may guide personalized therapeutic interventions, ultimately aiming to improve clinical outcomes for KIRC patients.

Results

Identification of gene subtypes based on MRGs

Employing the ConsensusClusterPlus algorithm, we stratified KIRC patients into two molecularly distinct subtypes (C1 and C2) based on MRG expression profiles (Fig. 1A). Kaplan-Meier survival analysis demonstrated significantly poorer clinical outcomes in C2 compared to C1 (p < 0.001; Fig. 1B), with the survival curves exhibiting pronounced divergence. Unsupervised hierarchical clustering analysis revealed distinct gene expression patterns characterizing each subtype (Fig. 1C). Notably, comprehensive immune profiling uncovered significant inter-cluster differences in tumor-infiltrating immune cells. C2 displayed markedly elevated infiltration of natural killer (NK) cells and T lymphocyte populations compared to C1 (Fig. 1D). This enhanced immunogenic profile in C2 suggests: (1) an activated anti-tumor immune response, and (2) potential heightened immunotherapeutic susceptibility relative to C1.

Fig. 1
figure 1

(A) Consensus matrix derived from the ConsensusClusterPlus algorithm (k = 2). (B) Overall survival (OS) of patients with KIRC stratified by distinct MRGs-based classifications. (C) Heatmap visualizes the differential expression of MRGs within each cluster with different clinical features. (D) Disparities in various immune cell populations across the delineated distinct clusters.

Full size image

Development of prognostic signature for MRGs using LASSO Cox regression analysis

Differential expression analysis using the ‘limma’ package (thresholds: |log2FC| > 1, FDR < 0.05) identified 867 significantly differentially expressed genes (DEGs) between the two clusters. The TCGA-KIRC cohort was randomly partitioned into two equally sized subsets (training and testing sets) with balanced clinical characteristics (p > 0.05 for all baseline features). Univariate Cox regression analysis of the DEGs revealed 399 genes with significant prognostic value (p < 0.05). Through LASSO regression analysis of these candidate genes in the training cohort, we identified 10 key prognostic markers based on optimal λ values determined by minimum criteria (Figs. 2A, B). Subsequent multivariate Cox regression analysis yielded a refined 6-gene prognostic signature comprising METTL24, PABPC1L, FAM20A, PTX3, LRRC19, and SCX, with their respective regression coefficients used to calculate risk scores (RS). Patients in the training set were dichotomized into high- and low-risk groups based on median RS thresholds (Fig. 2C). Risk stratification analysis demonstrated: (1) clear separation of survival outcomes by RS category (Fig. 2D), (2) distinct expression patterns of signature genes between risk groups (Fig. 2E) - with PABPC1L, FAM20A, PTX3 and SCX showing elevated expression, while METTL24 and LRRC19 were downregulated in high-risk patients, and (3) significantly poorer overall survival in the high-risk group (p < 0.001; Fig. 2F). Time-dependent ROC analysis confirmed the model’s robust predictive accuracy, with AUC values of 0.772 (1-year), 0.761 (3-year), and 0.829 (5-year) (Fig. 2G). These results establish our MRG-based signature as a clinically relevant prognostic tool with potential utility for risk stratification and therapeutic decision-making in KIRC management.

Fig. 2
figure 2

(A) Coefficients of predictor variables were derived using LASSO Cox regression. (B) The optimal regularization parameter (lambda) was identified, exhibiting with the first black dashed line in the logarithmic scale plot. Risk scores (RS) (C) and patient survival status (D) were examined in relation to the MRG prediction model. (E) Gene expression patterns of MRGs were compared in different risk cohorts, visualized in a heatmap. (F) Kaplan–Meier survival analysis on prognosis and risk groups (p < 0.001). (G) ROC curves assessed the performance of the MRG signature for predicting 1-, 3-, and 5-year survival outcomes.

Full size image

Validation of the MRG prognostic signature

We rigorously assessed the prognostic performance of the MRG-based risk model through comprehensive internal and external validation. The TCGA-KIRC cohort served for internal validation, while the E-MTAB-1980 dataset provided independent external validation. Both cohorts were stratified into distinct risk categories based on established thresholds. Risk distribution analysis revealed: (1) Significant association between risk stratification and clinical outcomes, with high-risk patients demonstrating markedly elevated mortality rates (Figs. 3A, B); and (2) Consistent MRG expression patterns across validation cohorts that recapitulated the training set findings (Figs. 3C, D). Kaplan-Meier survival curves confirmed significantly poorer prognosis in high-risk groups for both validation cohorts (log-rank p < 0.001; Figs. 3E, F). The model demonstrated robust temporal predictive accuracy, with 5-year AUC values of 0.740 (TCGA-KIRC) and 0.747 (E-MTAB-1980), indicating excellent generalizability across independent datasets (Figs. 3G, H). Moreover, Kaplan-Meier survival analysis conducted on the two external validation cohorts (GSE29609 and ICGC-RECA-EU) further demonstrated that the high-risk group exhibited significantly poorer prognosis, indicating robust performance of the model across multiple independent datasets (Supplementary Fig. 1). These outcomes cumulatively substantiate the reliability of the MRG risk model as an effective instrument for forecasting outcomes in patients with KIRC, highlighting its potential utility in guiding clinical management.

Fig. 3
figure 3

Validation of the prognostic utility of the MRG signature in both internal (TCGA-KIRC cohort) and external (E-MTAB-1980 cohort) datasets. (A–B) Visualization of risk score distribution and associated survival outcomes in the TCGA-KIRC and E-MTAB-1980 cohorts, respectively. (C–D) Heatmaps illustrating differential MRG expression between risk groups in each dataset. (E–F) Kaplan–Meier survival analysis demonstrating significant differences in survival probabilities between the risk groups in the internal and external cohorts. (G–H) Predictive performance of the MRG signature for 1-, 3-, and 5-year survival is assessed through ROC curves across both validation cohorts.

Full size image

Correlation between the MRG prognostic signature and clinical features

To elucidate the clinical relevance of the MRG-based prognostic signature, we conducted a comprehensive survival analysis stratified by key clinical parameters in the TCGA-KIRC cohort. The analyzed variables included age (≥ 65 or < 65 years), sex (female or male), group status (G1–2 or G3–4), metastatic status (M0 or M1), stage (I–II or III–IV), and TNM stage (T1–2 or T3–4). Stratified survival analysis demonstrated that high-risk patients consistently exhibited significantly poorer overall survival compared to low-risk counterparts across all clinical subgroups (Fig. 4). Furthermore, quantitative analysis revealed a progressive elevation in risk scores (RS) corresponding with advanced tumor stage, higher histological grade, and more severe TNM classification (Supplementary Fig. 2). These observations underscore the consistency and dependability of this prognostic risk signature and affirms its utility in clinical settings.

Fig. 4
figure 4

Associations between the MRG-based prognostic model and clinical characteristics in patients with KIRC. (A–F) Kaplan–Meier survival analyses evaluating overall survival (OS) in subgroups stratified by distinct clinical parameters: (A) age (> 65 vs. ≤65 years), (B) sex (female vs. male), (C) tumor grade (low grade: G1–2 vs. high grade: G3–4), (D) metastatic status (non-metastatic: M0 vs. metastatic: M1), (E) clinical stage (early stage: I–II vs. advanced stage: III–IV), and (F) primary tumor classification (T1–2 vs. T3–4).

Full size image

Nomogram establishment based on the MRG signature

To rigorously evaluate the predictive performance of our mitophagy-related prognostic model, we conducted comparative analyses with four established signatures18,19,20. The results revealed that our model exhibited superior discriminative ability, achieving a significantly higher concordance index (C-index = 0.755) compared to the alternative models (C-indices: 0.701, 0.654, 0.684, and 0.701; Fig. 5A), demonstrating its enhanced prognostic accuracy for clinical application. Subsequent Cox regression analyses confirmed the MRG signature’s robust clinical utility, with risk score (RS) showing statistically significant associations with patient outcomes in both univariate and multivariate analyses (Fig. 5B, C), establishing it as an independent prognostic factor even after adjustment for confounding clinical variables. We subsequently developed a clinically applicable nomogram (Fig. 5D) that demonstrated excellent predictive accuracy, as evidenced by strong concordance between predicted and observed survival outcomes in calibration analyses (Fig. 5E) and outstanding time-dependent discrimination (AUCs: 0.910 [1-year], 0.826 [3-year], and 0.857 [5-year]; Figs. 5F-H). These comprehensive validation results substantiate the MRG-based nomogram as a reliable tool for prognostic assessment and disease progression monitoring in clinical practice.

Fig. 5
figure 5

Nomogram construction integrating the MRG prognostic signature and its clinical relevance in the TCGA-KIRC cohort. (A) Comparative analysis of C-indices to evaluate the predictive power of the MRG signature against four alternative prognostic models. (B, C) Cox regression analyses (univariate and multivariate) examining the associations between risk scores (RS) and clinical parameters in the TCGA-KIRC dataset. (D) A nomogram was constructed to predict 1-, 3-, and 5-year overall survival (OS), assigning scores to individual variables and calculating a total score to estimate survival probabilities over time. (E) Calibration plots comparing the nomogram-predicted OS probabilities at 1, 3, and 5 years with observed survival outcomes, demonstrating alignment between predictions and reality. (F-H) ROC curves comparing the 1-, 3-, and 5-year predictive performance of the nomogram, clinical features, and the risk model, highlighting the superior prognostic accuracy of the nomogram in patients with KIRC.

Full size image

Interaction between the prognostic signature and TME

To systematically characterize the immunological correlates of our risk signature, we employed seven established computational approaches to evaluate the relationship between risk scores (RS) and immune cell infiltration patterns. Comprehensive analyses demonstrated that RS exhibited significant positive correlations with NK T cells, M1 macrophages, regulatory T cells (Tregs), myeloid dendritic cells, and CD4 + Th1 T cells, while showing negative associations with neutrophils, hematopoietic stem cells, CD8 + T cells, and endothelial cells (Fig. 6A). For more refined immunological characterization, we stratified TCGA-KIRC patients into five distinct immune subtypes using a published framework21revealing substantial differential distribution between high- and low-risk groups (p < 0.001; Fig. 6B), suggesting profound interplay between risk stratification and tumor immunology. ssGSEA-based evaluation of tumor microenvironment composition demonstrated that high-risk tumors were enriched in CD8 + T cells, macrophages, follicular helper T cells (Tfh), and tumor-infiltrating lymphocytes (TIL), whereas low-risk tumors showed predominance of B cells, immature dendritic cells (iDC), mast cells, and neutrophils (Fig. 6C). Functional immune profiling further identified significant differences in critical immunological processes including T cell co-stimulation, type II interferon responses, pro-inflammatory signaling, and parainflammatory responses between risk groups (Fig. 6D). Assessment of immune checkpoint molecules revealed downregulation of CD274, KDR, IDO1, and HAVCR2 in high-risk tumors (Fig. 6E), while TIDE analysis indicated superior predicted immunotherapy response in high-risk patients (Fig. 6F). These comprehensive immunological evaluations collectively establish our risk signature as a robust biomarker for tumor immune landscape characterization and immunotherapy response prediction in KIRC.

Fig. 6
figure 6

Analysis of the association between the prognostic signature and the tumor microenvironment in KIRC. (A) Evaluation of immune cell infiltration in TCGA-KIRC patients using multiple computational approaches. (B) Relevance between immune-related subtypes and risk classification based on the prognostic model. (C) Comparative analysis of immune cell abundance between high-risk and low-risk groups. (D) Functional differences in immune activities across the two risk groups. (E) Assessment of immune checkpoint expression in relation to risk stratification. (F) Association between risk scores and tumor immune dysfunction and exclusion (TIDE) metrics. Statistical significance denoted as *p < 0.05, **p < 0.01, ***p < 0.001.

Full size image

Pathway analysis associated with MRG signature

Gene Ontology (GO) analysis demonstrated significant enrichment of the identified genes in key biological processes including immune response mediation through molecular effector production, immunoglobulin biosynthesis, and organic anion transmembrane transport (Fig. 7A, B). Cellular component analysis revealed predominant localization to immunoglobulin complexes, collagen-enriched extracellular matrices, and the plasma membrane exterior surface. Molecular function assessment identified substantial enrichment in antigen recognition, glycosaminoglycan binding, and sulfur compound interactions. Complementary KEGG pathway analysis further substantiated the immunological relevance of these genes, with prominent involvement in cytokine-cytokine receptor networks, complement and coagulation cascades, viral protein-cytokine receptor interplay, and Wnt signaling transduction (Fig. 7C, D). These molecular pathways represent fundamental mechanisms governing immune regulation and have established roles in oncogenesis and therapeutic responsiveness. Collectively, these functional annotations validate the biological plausibility of our risk signature as a potential biomarker for evaluating immunological status and predicting immunotherapy efficacy in KIRC patients, with significant implications for precision oncology approaches.

Fig. 7
figure 7

Pathway enrichment analysis of DEGs based on risk classification in this model. (A–B) Pathway enrichment results of the GO analysis are presented using bar charts and circle plots. (C–D) Pathway enrichment outcomes of the KEGG analysis are visualized through bar charts and circle plots57,58,59 (KEGG permissions: Ref: 251930 ).

Full size image

Pharmacological susceptibility assessment

Subsequently, we explored the response to various anticancer drugs in patients with KIRC based on the risk stratification derived from this model using the GDSC database. The results indicated that Bortezomib exhibited a lower IC50 value in the high-risk group, suggesting its potential for these patients’ treatment, whereas the low-risk group might benefit from treatments with Bryostatin, Genentech Cpd 10, Lapatinib, Imatinib, IPA-3, Midostaurin, Obatoclax Mesylate, and Pazopanib (Supplementary Fig. 3). These drugs target different molecular pathways and mechanisms; thus, our findings may provide effective therapeutic options for targeted treatment of patients with KIRC and point to new directions for future research.

Potential role of METTL24 in KIRC

Among the identified genes, METTL24 exhibited the highest weight and was thus selected for in vitro investigations to explore the potential role of MRG in KIRC. Successful transfection of the METTL24-GFP construct was initially confirmed by fluorescence microscopy in both ACHN and CAKI-1 cell lines (Supplementary Fig. 4, Fig. 8A), with subsequent western blot analysis of the Flag-tagged protein providing additional verification (Fig. 8B). Quantitative RT-PCR analysis demonstrated robust METTL24 overexpression in transfected cells (Fig. 8C). Notably, METTL24 overexpression induced upregulation of the canonical mitophagy marker PINK1, an effect that was significantly attenuated by 3-MA treatment (Fig. 8D). Functional characterization revealed that METTL24 overexpression exerted potent anti-proliferative effects as quantified by CCK-8 assay (Figs. 8E-F) and significantly impaired cellular migration capacity in both wound healing and Transwell assays (Fig. 8G). These comprehensive findings establish METTL24 as a key regulator of KIRC cell proliferation and motility, highlighting its potential as a therapeutic target for KIRC intervention.

Fig. 8
figure 8

Identification of the potential role of METTL24 in KIRC. (A) Flow cytometry was performed to verify the transfection efficiency of the METTL24 plasmid. (B) Western blot analysis was conducted to evaluate the expression of the Flag tag in KIRC cells after plasmid transfection. (C) qRT-PCR confirmed a significant overexpression of METTL24 in ACHN and CAKI-1 cell lines (n = 3). (D) Expression of the classic autophagy marker PINK1 was analyzed via western blot in METTL24-overexpressing KIRC cells, with additional evaluation of its modulation by the autophagy inhibitor 3-MA. (E) Proliferative capacity of METTL24-overexpressing KIRC cells was measured using the CCK-8 assay (n = 4). The scratch wound assay (F) and Transwell assay (G) were used to assess the impact of METTL24 overexpression on the migratory capacity of KIRC cells. Statistical significance denoted as *p < 0.05, **p < 0.01, ****p < 0.0001.

Full size image

Discussion

Mitophagy represents a crucial cellular quality control mechanism that maintains cellular homeostasis through selective elimination of dysfunctional mitochondria, thereby preventing subsequent cellular damage22. Dysfunctional mitophagy leads to pathological accumulation of impaired mitochondria, a process increasingly recognized as contributing to tumor initiation and malignant progression, which has consequently become a major focus of oncological research23. The evolutionarily conserved PINK1/Parkin pathway serves as the predominant mitophagy regulatory cascade, with PARKIN deficiency being particularly prevalent in colorectal carcinoma and demonstrating strong association with adenomatous polyposis coli deficiency24. Key MRGs such as PINK1, BNIP3L, and PARKIN have been implicated in tumor therapy resistance mechanisms, positioning them as promising molecular targets for anticancer therapeutic development25. Pharmacological modulation of mitophagy through the PHB2-PARL-PGAM5-PINK1 axis may yield significant therapeutic advantages in oncology26. Emerging research has systematically elucidated the pathophysiological connections between mitophagy dysregulation and diverse malignancies including multiple myeloma, colorectal carcinoma, hepatocellular carcinoma, and ovarian cancer27,28,29,30. Elucidating the mechanistic role of mitophagy in oncogenesis provides novel diagnostic and therapeutic opportunities, with several validated MRG-based prognostic signatures already demonstrating clinical utility in guiding precision oncology approaches31. MRG signatures demonstrate robust diagnostic and molecular classification capabilities across diverse malignancies, where they reliably discriminate between neoplastic and normal tissues while identifying clinically relevant subtypes characterized by distinct immune microenvironments and therapeutic sensitivities32. The prognostic models developed from these genetic markers exhibit exceptional predictive performance for clinical outcomes in multiple cancer types, offering clinically actionable insights for treatment stratification33. Notably, these signatures maintain significant associations with tumor immune infiltration landscapes and immune checkpoint molecule expression patterns, establishing their utility as predictive biomarkers for immunotherapeutic efficacy32. Moreover, they provide a molecular framework for designing tailored combination therapies, including synergistic approaches combining immune checkpoint inhibition with cytotoxic chemotherapy or molecularly targeted agents, while facilitating treatment optimization based on individual mitophagy profiles34. These versatile applications underscore the translational value of MRG signatures as multidimensional tools for enhancing precision medicine approaches and improving clinical responses to immunotherapy.

This study elucidated the correlations between MRGs and KIRC prognosis, establishing a novel six-gene MRG signature for evaluating clinical outcomes and predicting therapeutic response in KIRC patients. Kaplan-Meier analysis demonstrated consistent associations between risk stratification and prognostic outcomes across both training and validation cohorts. Furthermore, ROC curve analyses revealed that our risk model exhibited superior prognostic performance compared to existing models utilizing alternative gene signatures. Notably, we developed an innovative nomogram incorporating key clinical parameters, which demonstrated robust predictive accuracy and significantly enhanced AUC values for KIRC patients. The MRG signature effectively prognosticated survival outcomes and predicted immunotherapy responsiveness in KIRC, underscoring its clinical utility for patient management.

The six-gene prognostic signature for KIRC comprises METTL24, PABPC1L, FAM20A, PTX3, LRRC19, and SCX, all of which demonstrate established associations with tumor progression and represent potential prognostic biomarkers in KIRC. The methyltransferase-like family (METTL), particularly METTL24, plays a pivotal role in tumorigenesis, with prior evidence supporting its prognostic value in colorectal cancer35. PABPC1L has been characterized as a critical determinant of Gleason score progression in prostate cancer36. FAM20A exhibits significant downregulation in lung squamous cell carcinoma, demonstrating an inverse correlation with DNA repair pathway proteins and potential implications for radiotherapy resistance37. PTX3 functions as an extrinsic tumor suppressor, mediating complement-dependent inflammatory responses in malignancies including leiomyosarcoma and colorectal cancer38. LRRC19, predominantly expressed in intestinal epithelial cells, modulates inflammatory cascades, with its deficiency attenuating intestinal tumorigenesis39. However, the mechanistic contributions of these genes to KIRC pathogenesis warrant further investigation. Given METTL24’s dominant weighting in our prognostic model, we performed functional validation through METTL24 overexpression in Caki-1 and ACHN KIRC cell lines. This intervention significantly attenuated cellular proliferation, invasion, and migratory capacity, while concurrently upregulating the canonical autophagy marker PINK1, suggesting METTL24’s potential involvement in autophagic regulation within KIRC. These results position METTL24 as a critical tumor suppressor in KIRC and highlight its therapeutic potential as a molecular target.

The tumor immune microenvironment (TIME) plays a critical role in determining immunotherapy efficacy by modulating therapeutic responses, tumor progression, and clinical outcomes40,41. Pan-cancer single-cell analyses demonstrate that immune-infiltrated microenvironments display enhanced tumor immunogenicity and improved predictive accuracy for treatment response42. Quantitative assessment of immunosuppressive microenvironments through integrated scoring systems like iMLGAM can predict therapeutic resistance via mechanisms including CEP55-mediated T-cell suppression43. Furthermore, systemic immune-inflammation indices serve as independent prognostic indicators across multiple malignancies44. Mitophagy emerges as a key regulator of TIME dynamics by orchestrating immune cell development, activation, and differentiation. Impaired mitophagy induces mitochondrial ROS accumulation in macrophages, activating HIF-1α signaling and driving M2 polarization33. Within the TIME, elevated lactate levels and hypoxia inhibit T cell mitophagy, leading to dysfunctional mitochondrial accumulation, metabolic stress, and cytochrome c release. Conversely, BNIP3-mediated mitophagy activation restores dendritic cell function, enhances T cell priming, and improves immunocompetence45. To evaluate mitophagy-related heterogeneity in KIRC TIME, we employed seven computational algorithms to correlate risk scores with immune infiltration patterns. Immune subtype stratification revealed significant risk classification disparities, suggesting MRG-based risk scores exhibit strong associations with immune infiltration and warrant further characterization of their immunological profiles. Quantitative analyses demonstrated elevated infiltration of CD8 + T cells, macrophages, Tfh cells, and TILs in high-risk patients. Enhanced CD8 + T cell infiltration potentiates tumor-specific cytotoxicity and improves immune checkpoint inhibitor sensitivity. Increased macrophage infiltration with M1 polarization remodels the TIME through pro-inflammatory cytokine secretion46. Augmented Tfh cell activity promotes B cell-mediated antibody production and tertiary lymphoid structure formation. Higher TIL density directly correlates with improved tumor immunogenicity and treatment response rates47. These collective findings suggest high-risk patients may exhibit superior immunotherapeutic responses and clinical outcomes compared to their low-risk counterparts.

Immune checkpoint blockade (ICB) therapy, which reactivates dysfunctional immune responses, has emerged as a transformative approach in oncology with demonstrated clinical efficacy48,49,50. FDA-approved ICB agents targeting key immune checkpoints—including nivolumab (anti-PD-1) and ipilimumab (anti-CTLA-4)—have become established treatments for advanced KIRC51. However, substantial interpatient variability in treatment response due to tumor heterogeneity necessitates the development of precision medicine strategies52,53. Our analysis revealed that KIRC patients in the high-risk group exhibited upregulated expression of immune checkpoint genes (PDCD1 and CTLA4), indicating their enhanced potential for ICB responsiveness compared to low-risk patients. This observation was further corroborated by TIDE scoring, where high-risk patients demonstrated significantly lower scores, predictive of superior immunotherapy outcomes. These collective findings establish that mitophagy heterogeneity modulates the tumor immune microenvironment and serves as a predictive biomarker for ICB response in KIRC, revealing a novel mechanistic link between mitochondrial autophagy and antitumor immunity. This work provides critical insights into mitophagy regulation in KIRC pathogenesis, facilitates clinical decision-making for personalized therapy, and opens new avenues for investigating immunotherapeutic applications.

GO and KEGG enrichment analyses of DEGs revealed associations between risk grouping and related signaling pathways, including production of molecular mediators of immune responses, immunoglobulin production, and cytokine–cytokine receptor interactions. Immune mediators including cytokines and chemokines directly modulate the inflammatory profile and antitumor immune activity within the microenvironment through activation of pattern recognition receptors and regulation of immune cell functional states54. Immunoglobulins such as IgG and IgA facilitate antigen-specific humoral immune responses via antibody-dependent cellular cytotoxicity, complement activation, and immune complex formation, while concurrently influencing tertiary lymphoid structure development55. Cytokine-receptor signaling pathways including PD-1/PD-L1 and IL-6/IL-6R dynamically regulate immune cell migration, activation, and exhaustion processes, thereby determining the immunosuppressive or immunostimulatory characteristics of the tumor microenvironment56. Our study identified potential drugs for subgroups with distinct mitophagy patterns, thereby offering a basis for the combined use of ICB drugs and targeted therapies.

In summary, these findings demonstrate that the MRG-based risk signature serves as a robust prognostic biomarker for assessing immunological profiles and predicting immunotherapy responsiveness in KIRC patients, thereby potentially enhancing clinical outcomes. Future therapeutic approaches could integrate targeted mitophagy regulation with immunotherapeutic interventions or explore mitochondrion-specific nanodrug formulations to optimize treatment effectiveness. Additionally, non-invasive biomarkers such as circulating mtDNA may facilitate real-time monitoring, offering novel opportunities for personalized KIRC management. Nevertheless, this study presents certain limitations, including potential selection bias stemming from the exclusive use of publicly available, predominantly retrospective datasets. Further investigations are warranted to elucidate the precise mechanistic roles of MRGs in KIRC pathogenesis. Ultimately, more precise and individualized therapeutic strategies should be investigated through meticulously designed clinical trials to validate these findings.

Materials and methods

Data acquisition

Gene expression data and clinical profiles for 514 KIRC patients were collected from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/). The TCGA-KIRC cohort was divided into two equally random groups: training (50%) and internal testing (50%). The E-MTAB-1980 cohort (obtained from the ArrayExpress database (https//www.ebi.ac.uk/arrayexpress/)), GSE29609 cohort (obtained from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/)) and ICGC-RECA-EU cohort (obtained from the International Cancer Genome Consortium (ICGC) database (https://dcc.icgc.org/)) were designated as external validation groups. All gene expression data were normalized, and the expression values of duplicate genes were averaged. In total, 29 MRGs were retrieved from the Pathway Unification online resource and corresponding literature sources.

Construction of KIRC molecular subtype

The molecular subtypes of KIRC were classified according to MRG expression patterns through consensus clustering analysis performed with the ConsensusClusterPlus package in R. Prognostic disparities among the established subtypes were then systematically evaluated. Immune microenvironment heterogeneity across subtypes was interrogated using the MCP-counter algorithm. Cluster-specific differentially expressed genes (DEGs) were identified via the limma package, applying stringent significance thresholds (absolute log2 fold change > 1 and false discovery rate < 0.05) to ensure robust biomarker selection.

Development and validation of MRG signature

To identify key prognostic genes, we first performed univariate Cox regression analysis on differentially expressed genes (DEGs), retaining those with statistical significance (p < 0.05). To refine gene selection, we implemented a rigorous two-stage analytical approach: first applying LASSO regression for dimensionality reduction and feature selection, followed by multivariate Cox regression to establish the final prognostic gene set. KIRC patients were stratified into high- and low-risk groups using median risk scores as the cutoff. Survival disparities between these cohorts were assessed through Kaplan-Meier analysis with log-rank testing. The model’s predictive capacity was evaluated using time-dependent receiver operating characteristic (ROC) curves at 1-, 3-, and 5-year intervals. We constructed a comprehensive nomogram integrating clinical parameters with RSs to facilitate individualized survival prediction, with model discrimination validated through ROC analysis. Finally, we externally validated the prognostic genes using the E-MTAB-1980 dataset, comprising over 100 samples with complete clinicopathological and survival data, thereby ensuring the robustness of our findings. Kaplan–Meier survival analysis was additionally performed on two external validation cohorts (GSE29609 and ICGC-RECA-EU) to assess the model’s prognostic efficacy. This external validation confirms the robustness and applicability of our findings.

Comparative analysis of the tumor microenvironment across subgroups

To systematically investigate the association between immune cell infiltration patterns and risk stratification, we employed multiple computational algorithms including XCELL, QUANTISEQ, CIBERSORT, EPIC, MCPCOUNTER, and TIMER. These complementary methodologies facilitated a comprehensive evaluation of immune cell composition and their functional dynamics within the tumor microenvironment (TME). Immune cell functional states were further characterized through single-sample gene set enrichment analysis (ssGSEA), while transcriptomic profiling of immune checkpoint molecules was performed to evaluate potential implications for immunotherapy response. This integrative analytical framework yields critical insights into the immunological landscape and its potential clinical relevance for therapeutic interventions.

Analyzing the pathway of the identified signature

Differentially expressed genes (DEGs) among distinct groups were systematically identified using the limma package in R. Subsequent functional annotation of these DEGs was conducted through comprehensive pathway enrichment analyses, including Gene Ontology (GO) for biological process categorization and Kyoto Encyclopedia of Genes and Genomes (KEGG) for pathway mapping. To further characterize group-specific biological variations, we performed gene set variation analysis (GSVA) employing both GSVA and GSEABase packages in R, thereby elucidating the differential molecular pathways and biological processes underlying the observed transcriptional variations.

Pharmacological susceptibility assessment

To evaluate the clinical utility of the MRG signature as a predictive biomarker for therapeutic response, we systematically investigated its association with drug sensitivity in both conventional chemotherapy and molecularly targeted agents. Drug response profiles were quantitatively assessed through calculation of half-maximal inhibitory concentration (IC50) values, representing the drug concentration required for 50% growth inhibition. These pharmacological parameters were computationally derived using the pRRophetic prediction algorithm, which leverages large-scale gene expression data to generate drug sensitivity estimates across a comprehensive panel of therapeutic compounds.

Cell culture and transient transfection

The KIRC cell lines Caki-1 and ACHN, obtained from the Cell Bank of the Central Laboratory at Tianjin Medical University’s Second Hospital, were cultured under standard conditions: Caki-1 in RPMI 1640 complete medium and ACHN in DMEM medium, both at 37 °C with 5% CO2. For genetic manipulation experiments, the METTL24 overexpression plasmid and corresponding negative control vector were procured from Shanghai GeneChem, with transfection performed using Lipofectamine 3000 (Invitrogen). Successful METTL24 overexpression was confirmed through comprehensive molecular characterization employing both western blot analysis and reverse transcription quantitative polymerase chain reaction (RT-qPCR). These genetically modified cell lines served as essential experimental models for elucidating METTL24’s functional role in regulating cellular proliferation dynamics and mitophagy pathway activation in KIRC.

Western blot

Proteins from METTL24 overexpression and control KIRC cells were extracted using RIPA buffer containing protease inhibitors. After separation by 10% SDS-PAGE, the proteins were transferred onto PVDF membranes. Blocking was followed by overnight incubation with primary antibodies against Flag (F1804, Sigma-Aldrich) and PINK1 (23274-1-AP; Proteintech). The membranes were then treated with HRP-conjugated secondary antibodies, washed, and then visualized using chemiluminescence.

RT-qPCR

RNA from METTL24-overexpressing and control KIRC cells was extracted with TRIzol and converted to cDNA using a reverse transcription kit. Quantitative PCR was conducted using a SYBR Green-based Master Mix, with gene expression levels determined through the 2−∆∆CT calculation method. The primers used for amplification included the forward sequence AGCCTGGCTACTGACTCTAGT and the reverse sequence TGTTGCGGATTTGATGAGCTAA.

Cell proliferation viability

A total of 1500 METTL24-overexpressing KIRC cells or control cells were seeded into 96-well plates. Cell proliferation was assessed at 24, 48, and 72 h post-incubation using the CCK-8 assay, allowing for a comparative analysis of growth rates between the two groups.

Wound healing experiment

The 12-well plates were seeded with METTL24 overexpression or control KIRC cells and incubated under optimal conditions. Once the cells reached an appropriate density, a sterile pipette tip was used to create a scratch along the ruler at the center of the cell monolayer. The medium was replaced with 2% low-serum medium. The width of the scratches was observed under a microscope at 0 and 24 h to assess cell migration.

Transwell assay

Approximately 2 × 105 METTL24-overexpressing or control KIRC cells were seeded into specialized Transwell chambers. Serum-free medium was added to the upper chamber, and DMEM/1640 medium with 12% serum was added to the lower chamber to establish a nutrient gradient. Following 24 h, the inserts were detached, and the cells adhering to the upper side were gently scraped away. The cells attached to the lower side of the insert were subsequently fixed and subjected to staining with a 0.1% crystal violet solution. After thorough rinsing, the cell count was examined microscopically to evaluate alterations in the migratory capacity of KIRC cells.