Clear cell renal cell carcinoma (ccRCC) is the most prevalent and lethal subtype of kidney cancer. It is notable for being one of the most immune-infiltrated tumor types, with the tumor microenvironment (TME) playing a critical role in disease progression, metastasis, and response to therapy. Despite this, the precise immunological mechanisms that drive patient outcomes remain poorly understood.
Immune checkpoint inhibitors (ICIs), including PD-1 and CTLA-4 blocking antibodies, have transformed treatment for advanced ccRCC. However, response to immunotherapy varies widely among patients. Identifying which patients will respond to these agents is critical, as non-responders face treatment toxicity without benefit. A quantitative immune signature could guide this selection more precisely than current clinical criteria.
The tumor microenvironment contains numerous immunological actors, including mast cells, regulatory T cells, macrophages, lymphangiogenesis mediators, and angiogenic pathways. Each of these contributes to a complex immunosuppressive or immunostimulatory balance that shapes tumor behavior. Quantifying 53 distinct immunological terms across six ccRCC patient cohorts enabled a systematic search for the prognostically most relevant immune features.
Six independent ccRCC patient cohorts were combined into a unified dataset totaling 1,207 samples from databases including TCGA, ICGC, GSE73731, GSE40435, E-MTAB-1980, and CPTAC. Batch effects between cohorts were corrected using the ComBat algorithm, and RNA-seq data were normalized to Transcripts Per Million (TPM) to ensure comparability across microarray and sequencing platforms.
The ssGSEA algorithm was applied to quantify 53 immunological terms for every patient, capturing immune infiltration across cell types including mast cells, T cell subsets, natural killer cells, and dendritic cells, as well as functional immune activities such as angiogenesis, CSR activation, and IL-4 signaling. This provided a comprehensive snapshot of each patient's immune landscape.
Through 1,000 iterations of L1-penalized (LASSO) estimation followed by multivariate Cox regression, five immune terms were selected to construct the Immune-Related Risk Score (IRS): angiogenesis (negative coefficient), CSR activated (positive), IL4 score (positive), lymphocyte PCA (positive), and mast cells (negative). The IRS formula weights and combines these terms to produce a continuous patient-level risk estimate.
Optimal cutoff values for classifying patients as high or low IRS were determined using ROC analysis, yielding cutoffs of 1.35 and 3.19 in the training and validation cohorts with AUC values of 0.807 and 0.776 respectively. The final model's predictive performance improved further when IRS was combined with patient age and stage, indicating complementary information content.
High IRS patients had significantly worse baseline clinical characteristics, including higher tumor grade and more advanced pathological T stage. The risk score was not associated with patient age, but was strongly correlated with grade and stage, validating that it reflects genuine tumor aggressiveness rather than confounding demographic factors.
Pathway analysis revealed that EMT, JAK/STAT signaling, and G2/M checkpoint pathways were upregulated in high IRS patients. These pathways are well-established drivers of tumor invasion, cell cycle dysregulation, and resistance to therapy in ccRCC. Conversely, KRAS signaling was downregulated in low IRS patients, reflecting the distinct biology underlying each group's clinical trajectory.
Genomic instability was higher in the high IRS group, with a greater burden of focal and arm-level copy number variations (CNV). Eleven specific genes were more frequently mutated in the high IRS group, including RELN, NF2, MYH4, and MUC16, which may contribute to increased tumor aggressiveness and immune evasion through co-occurring mutational interactions.
Tumor stemness indices (mRNAsi and mDNAsi) were negatively associated with IRS, meaning that low IRS patients exhibited higher cancer stem cell characteristics. While this might appear paradoxical, it reflects a distinct biology: low IRS tumors may depend on stemness pathways for growth, whereas high IRS tumors rely on immune evasion and genomic instability as their primary drivers of progression.
Analysis of immune checkpoint expression showed that high IRS patients had significantly elevated levels of CTLA-4 and PD-1, the two primary targets of approved immunotherapy drugs in ccRCC. The high immune infiltration in these tumors, while suggesting greater inflammation, may paradoxically represent a context of immune exhaustion rather than effective anti-tumor immunity.
CPA4, a gene with the highest differential expression between IRS groups, was selected for functional validation. CPA4 knockdown in ccRCC cell lines 786-O and Caki-1 significantly reduced cell proliferation, colony formation, migration, and invasion in vitro, confirming CPA4 as an oncogene that promotes ccRCC progression and representing a potential therapeutic target in high IRS patients.
TIDE (Tumor Immune Dysfunction and Exclusion) analysis, which models immune evasion mechanisms, found that low IRS patients had significantly lower TIDE scores, meaning they were more likely to be immunotherapy responders. This was corroborated by subclass mapping against a panel of 47 known immunotherapy responders, which confirmed that low IRS patients resembled PD-1 immunotherapy responders more closely than high IRS patients.
Drug sensitivity analysis using the Genomics of Drug Sensitivity in Cancer (GDSC) dataset estimated IC50 values for 12 targeted therapy drugs across IRS groups. The results revealed differential drug sensitivities that could inform tailored treatment selection, with high IRS patients showing a decreased IC50 for sunitinib, suggesting they may derive more benefit from this first-line targeted agent.
The combination of low TMB and low CNV, consistent with the low IRS profile, was associated with favorable responses to immune checkpoint inhibitor therapy, aligning with findings in other cancer types where genomic stability supports a productive rather than exhausted immune response. This genomic-immune interaction provides a rationale for combining TMB, CNV, and IRS in future patient stratification algorithms.
The IRS signature was built from six diverse international cohorts and validated across both training and independent validation sets, providing a more robust evidence base than single-cohort studies. Its ability to stratify patients across clinical subgroups including different grades, stages, and treatment contexts suggests it captures a fundamental immunobiological axis of ccRCC.
The identification of the five IRS components including mast cells, angiogenesis, IL-4 signaling, lymphocyte activity, and CSR activation provides biologically interpretable targets for future mechanistic investigation. Each of these pathways has independently established roles in tumor progression, immune regulation, or therapy response, reinforcing the biological plausibility of the IRS.
Future directions should include prospective validation of the IRS in patients receiving standardized immunotherapy or targeted therapy regimens, and integration of the score into clinical decision tools. If validated, the IRS could serve as a companion diagnostic for ccRCC immunotherapy trials, enabling more precise patient selection and improving the cost-effectiveness of expensive biologic treatments.