CAR-T cell signaling dynamics, rational design principles and artificial intelligence for next-generation chimeric antigen receptors.

Chimeric antigen receptor (CAR) T cell therapy has transformed the treatment of hematologic malignancies, yet its efficacy in solid tumors and durability across broader application remain limited. A central challenge lies in how CAR signaling is initiated, amplified, and regulated over time. Unlike the native T cell receptor (TCR), CARs are synthetic, modular receptors whose signaling output is dictated by the composition and spatial organization of their extracellular, transmembrane, and intracellular domains. Emerging evidence suggests that CAR signaling requirements are not static: insufficient signaling at early time points can impair activation and tumor clearance, whereas excessive or prolonged signaling promotes exhaustion, toxicity, and loss of persistence. More recent CAR designs therefore emphasize fine-tuned signaling, embracing a "less-is-more" paradigm to balance potency with durability. In this review, we summarized recent advances in CAR signaling biology, focusing on temporal signaling thresholds, modular design principles, and emerging strategies to precisely control signal strength and quality. Finally, we discuss how high-throughput screening, computational modeling, and machine learning approaches may enable disease-specific, personalized CAR designs in the future.