Pan-Cancer Non-Coding RNA Signatures of Tumor Microenvironment Remodeling Identify Delivery-Aware Therapeutic Targets in Pancreatic, Lung, Breast and gastric Adenocarcinomas

Abstract

The tumor microenvironment (TME) is a major determinant of progression, immune evasion, and therapy resistance in solid tumors, yet the non-coding RNAs (ncRNAs) that coordinate stromal, immune, and hypoxic programs across cancers remain poorly catalogued and have rarely been prioritized with delivery considerations in mind; we therefore performed an integrative pan-cancer analysis of long non-coding RNAs (lncRNAs) and microRNAs (miRNAs) in 2,447 primary tumors from four stroma-rich TCGA cohorts, pancreatic adenocarcinoma (PAAD, n=183), lung adenocarcinoma (LUAD, n=590), breast invasive carcinoma (BRCA, n=1,226), and stomach adenocarcinoma (STAD, n=448). For each sample we computed ESTIMATE stromal and immune scores and ssGSEA hypoxia scores, identified differentially expressed ncRNAs between expression tertiles (Wilcoxon FDR<0.05), correlated candidates with TME axes (Spearman |ρ|>0.25, FDR<0.05), reconstructed lncRNA–miRNA–mRNA ceRNA networks supported by miRTarBase/ENCORI, applied Kaplan–Meier and multivariate Cox survival models, and integrated single-cell evidence from TISCH2 to score therapeutic and delivery feasibility. Among 18,770 tested ncRNAs, we nominated 50 high-confidence candidates; LINC01614, LINC01094, LINC02544, MSCAS1, PCED1B-AS1, and AP000695.1 were the top recurrent lncRNAs, and hsa-mir-29c the top recurrent miRNA. LINC01614 was upregulated in TME-high tumors across all four cohorts (mean |log₂FC|=1.95, mean |ρ|=0.54), engaged ECM-stromal, hypoxia, TGF-β, and immune-activation modules via 294 ceRNA edges, and predicted shorter overall survival (best p=4.9×10⁻³, HR=1.17), while hsa-mir-29c was consistently downregulated and recovered canonical collagen targets. These findings establish a delivery- aware, pan-cancer ncRNA shortlist that links TME biology to actionable therapeutic strategies, providing a prioritized foundation for preclinical validation of TME-responsive ncRNA medicines.