AI Insight
Researchers analyzed genetic data from 9,288 tumors across 31 cancer types to identify nine microRNAs that consistently regulate how immune cells attack cancer. These microRNAs control 31 genes involved in cytolytic activity—the process by which immune cells kill tumor cells—and can distinguish between "immune-hot" tumors with strong immune responses and "immune-cold" tumors with weak responses. The study identified 124 potential drug compounds targeting this regulatory network, including existing immunotherapy targets like CTLA4 and PD-L1.
Why it matters
This work could improve cancer treatment by helping doctors predict which patients will respond to immunotherapy and identifying new drug targets to boost immune responses in tumors that currently evade immune attack. The findings apply across multiple cancer types, suggesting potential for broad clinical applications in precision oncology.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
Abstract Background: Cytolytic activity (CYT), a widely used transcriptomic surrogate of anti-tumor immune cytotoxicity derived from GZMA (granzyme A) and PRF1 (Perforin 1) expression, is associated with clinical outcomes across cancers. MicroRNAs (miRNAs) are key post-transcriptional regulators of tumor immunity, yet their pan-cancer roles in modulating cytolytic activity remain incompletely understood. Objective: This study aimed to identify conserved miRNA regulators of tumor cytolytic activity and their downstream gene-mediated networks across diverse cancer types, while evaluating their clinical and therapeutic relevance. Methods: Matched miRNA and mRNA expression profiles from 9,288 primary tumors across 31 TCGA cancer types were analyzed. A multi-stage framework was applied: per-cancer Spearman correlations (|{rho}| >= 0.30, FDR < 0.05) identified recurrent CYT-associated miRNAs (at least 3 cancer types); these were integrated with TargetScan-predicted targets and subjected to pan-cancer and cross-cancer triple filtering (miRNA-gene and gene-CYT associations). All associations underwent tumor purity adjustment using Consensus Purity Estimate (CPE), with LUMP (Leukocytes Unmethylation for Purity) as sensitivity analysis. Candidates were further prioritized by random forest modeling with bootstrap stability, cancer-type-adjusted Cox regression, mediation analysis, immune cell deconvolution, k-means molecular subtyping, pathway enrichment, and DGIdb-based drug-target prioritization. Results: The analysis converged on 38 high-confidence miRNA-gene-CYT regulatory triplets involving 9 conserved miRNAs and 31 target genes after stringent purity adjustment and multi-layer validation. All nine miRNAs exhibited complete bootstrap stability. Mediation analysis confirmed significant gene-level mediation in 37 of 38 triplets (FDR < 0.01), with mediated proportions up to 94%. The final miRNA signature defined two distinct pan-cancer immune subtypes (immune-hot vs. immune-cold) with significantly different cytolytic activity and overall survival (OS) (HR = 0.754, FDR = 1.12 x 10^-4). The network was enriched for T-cell activation and lymphocyte differentiation pathways and highlighted multiple druggable targets, including CTLA4 and CD274 (PD-L1), nominating 124 candidate compounds. Conclusions: In conclusion, this tumor purity-adjusted pan-cancer study defines a compact, reproducible, and clinically relevant miRNA network that regulates cytolytic activity across diverse malignancies. By linking miRNA biology to immune subtyping and actionable therapeutic targets, the present work provides a valuable foundation for advancing precision immuno-oncology.