Background T cell engagers (TCEs) activate T cells by bridging the CD3 complex to a tumor associated antigen (TAA). TCEs have shown clinical efficacy for hematological malignancies and hold promise for solid-tumors. Dose limiting toxicities, including Cytokine Release Syndrome (CRS) and Tumor Lysis Syndrome (TLS) remain major concerns.

Methods We engineered two ROR1 specific TCEs with properties to increase the size of their therapeutic windows using a unique ROR1 specific VHH from Camelus bactrianus (clone 5A1) that contains a non-canonical disulfide-bond between CDR1 and CDR3 and ROR1 specific valency that (1) decouples cytotoxicity from cytokine release and (2) facilitates the discrimination between ROR1 cell surface density on normal versus tumor cells.

Results When 5A1 was paired with an anti-CD3 Fab in two orientations that include a half-life extending Fc domain, we were able to establish a T cell/tumor synapse that resulted in sufficient activation to induce degranulation/tumor-killing at TCE concentrations of 10-30 pM (EC50 ~ 150-200 nM). In contrast, expression of CD69, CD25 and IFNg, the latter of which is directly upstream of CRS, was not observed until the concentration of TCEs reached ~1 nM (EC50 ~ 5-10 nM). Compared to the non-decoupling control TCE, 5A1 TCEs demonstrated 3.5-4.5 fold greater decoupling. In the context of JeKo1 CD19+/ROR1+ tumor targets, the 5A1 TCEs were 6-7.5 fold better decoupled compared to blinatumomab. We also observed an EC50-cytotoxicity for blinatumomab of below 100 fM, suggesting that the risk of both CRS and TLS are multiple folds higher than for 5A1 TCEs. Compared to the non-decoupling control TCE, the 5A1 TCEs mediated superior cytotoxicity/serial killing at an effector to target ratio (E:T) as low as 1:16, more consistent with solid-tumors. The 5A1 TCEs were also utilized ROR1 specific binding valency to discriminate between cell surface densities of ROR1 typical to tumor cells (>15k molecules/cell) versus normal cells (generally <10,000 molecules/cell). The concentration of TCEs required for maximum IFNg release and killing of tumor cells registered in the pM range, as opposed to the supraphysiologic nM/uM range that was insufficient to induce more than 25% and 35% of IFNg release and cytotoxicity, respectively, on cells expressing <10,000 molecules of ROR1.

Conclusions We hypothesize, that our ROR1 TCEs will provide a superior therapeutic index and safety profile associated with reduced risks of CRS and TLS, the capacity to dose higher, avoid continuous infusion and potentially mediate better efficacy in solid tumors.