TREND: A generalizable synthetic enhancer discovery platform for targeted immunotherapy

Synthetic enhancers with high specificity are crucial for therapeutic gene control. However, experimental screens and machine learning-guided design typically require context-specific datasets, limiting generalizability. Because transcription factor (TF) activity reflects cellular state, TF-responsive enhancer libraries offer a universal starting point. Here, we developed TREND (transcription factor-responsive enhancer discovery), a massively parallel reporter assay of ~2.7 million enhancer-barcode constructs representing 57,715 designs. TREND covers 1,068 motif-annotated proteins, including 729 confirmed TFs across 49 DNA-binding domain families. Applied to ovarian cancer, TREND identified enhancers that discriminate cancer from normal epithelial cells. These enhancers enabled protein-interaction-based AND-gate circuits with reduced OFF-state leakage and amplified ON-state output, driving tumor-restricted expression of combinatorial immune effectors and robust antitumor responses in murine ovarian cancer models. TREND also identified T-cell activation-responsive enhancers with greater inducibility and lower basal activity than conventional NFAT-motif-based elements. Together, TREND provides a generalizable framework for context-specific enhancer discovery and therapeutic gene regulation.