Leveraging MicroRNA Biomarkers to Dissect PD-L1 Regulation and Improve Immunotherapy Response in Lung Cancer

Non-small-cell lung cancer (NSCLC) remains a leading cause of cancer-related deaths, due to late diagnosis and therapy resistance. Immune checkpoint inhibitors (ICIs), either alone or in combinations with chemotherapy (CT), achieve durable responses in approximately 10–15% of patients with advanced disease. However, many patients do not respond and remain at risk for adverse effects. PD-L1 expression, the only FDA-approved biomarker for ICI selection, has limitations due to technical discrepancies and incomplete knowledge about its regulation. Emerging evidence suggests that microRNAs (miRNAs) influence PD-L1 expression, either by direct binding or through the modulation of regulatory genes. This study aims to elucidate miRNA-mediated PD-L1 regulation in NSCLC. To identify miRNAs involved in PD-L1 regulation in lung cancer cells, we utilized a pooled lentiviral library to overexpress the entire human miRNome (N=2,580). Transduced cells were sorted based on PD-L1 expression (PD-L1High & PD-L1Low), followed by next-generation sequencing (NGS) of genomic DNA to identify miRNAs enriched or depleted in each population. To explore the association between miRNA expression and PD-L1 abundance in patient tumor cells, we analyzed miRNA and PD-L1 expression (mRNA & protein) from the TCGA-LUAD dataset (N=507). Additionally, we performed whole miRNome profiling (miRNA-seq) on a cohort of stage IV NSCLC patients (N=35) treated with ICIs or ICI+CT as first-line therapy, categorized by PD-L1 expression levels (TPS: <1%, 1–49%, and >50%). The in vitro screening identified 550 candidate miRNAs significantly associated with PD-L1 expression (p ≤ 0.05), with 458 miRNAs enriched/depleted in PD-L1High cells and 92 miRNAs enriched/depleted in PD-L1Low cells. Of these, 21 miRNAs exhibited differential expression in the TCGA-LUAD dataset when comparing PD-L1High and PD-L1Low patient tumors (based on either PD-L1 mRNA or protein levels). These candidates were further validated in an independent NSCLC cohort receiving ICI-based therapy. More importantly, computational target prediction using the TargetScan algorithm identified 7 miRNAs with putative binding sites in the PD-L1 3′UTR, suggesting direct regulatory interactions. These findings highlight a subset of miRNAs as key regulators of PD-L1, with potential implications for refining patient selection for ICIs. Ongoing functional studies aim to validate their impact on PD-L1 expression and assess their predictive value for immunotherapy response. This study aims to identify miRNA-based biomarkers which are superior in terms of their stability in FFPE samples, capable of predicting ICI response in NSCLC patients; while unveiling novel mechanisms involved in immunotherapy response in order to enhance patient stratification and guide the development of more effective immunotherapy strategies.