De novo engineered disulfide bond supersedes native interchain linkage to enhance TCR pairing and anti-tumor efficacy in T cell therapy.

T-cell receptor-engineered T (TCR-T) cell therapy is considered highly promising for treating solid tumors. However, it still has significant limitations; one is exogenous-endogenous TCR chain mispairing, which could substantially compromise cell surface expression and the efficacy of the engineered TCR while raising clinical safety concerns. To address this obstacle, we developed a disulfide-substituted TCR (DSS-TCR), in which the native disulfide bond is replaced with artificially designed disulfide bonds within constant domains to increase the fidelity and functionality of TCR pairing. Our study demonstrated that ablation of the native interchain disulfide bond significantly reduces mispairing but severely impairs tumor-killing activity. Using structure-guided computational prediction, we designed nine pairs of artificial interchain disulfide bond-forming sites within TCR constant domains de novo. When introduced into native disulfide-deficient TCRs, four pairs reversed the decrease in cell surface expression. Notably, compared with the wild-type human TCR, two of them significantly enhanced both TCR surface expression and cytotoxic activity. By further combining different pairs of mutations and incorporating hydrophobic substitutions in the α-chain transmembrane domain, DSS-TCRs achieved superior pairing efficiency and antitumor efficacy that were comparable to those of TCRs incorporating mouse-derived constant regions. DSS-TCR significantly decreases the TCR mismatching rate while theoretically reducing immunogenicity. This superior optimization effect was also confirmed for other TCR-based receptors. Therefore, this engineering approach offers a safer and more potent paradigm for TCR-based T-cell therapeutics.