Abstract

Clear cell renal cell carcinoma (ccRCC), the most common and lethal subtype of renal cell carcinoma, exhibits marked intratumoral heterogeneity and complicates clinical management. Although long noncoding RNAs (lncRNAs) regulate diverse cellular processes, their landscape and biomarker potential in ccRCC remain poorly defined. Here we performed single-nucleus and bulk transcriptomic, proteomic, and metabolomic analyses on a cohort of 100 ccRCC patients. The expression pattern of lncRNAs were described based on metacells. Malignant cells displayed broader but lower lncRNA expression, likely reflecting copy number alterations, whereas low-abundance lncRNAs in normal epithelial cells showed individual variability. Multi-omics integration was used to establish a preliminary lncRNA functional inference pipeline, identifying lncRNAs involved in metabolic and immune processes and validating their roles through functional and in vivo experiments. Candidate biomarkers lncRNAs were identified to build diagnostic (DMRlnc) and prognostic models (PMRlnc), which were validated in TCGA, CheckMate, and IMmotion151 cohorts. DMRlnc achieved high diagnostic accuracy in both discovery and TCGA-KIRC cohorts (AUC 0.98 and 0.93). PMRlnc stratified patients into distinct risk groups with significant differences (p < 0.0001) across TCGA-KIRC and IMmotion151 cohorts. PMRlnc further indicated that low-risk patients may benefit more from nivolumab, while high-risk patients might respond better to atezolizumab plus bevacizumab.