338 Rational design of pH-dependent MSLN-targeting VHH for CAR-T therapy

Background

CAR-T cell therapies have achieved tremendous progress in hematological tumors; however, limited efficacy was observed in solid tumors. One of the critical challenges in solid tumors was the risk of clinically on target off tumor toxicity (OTOT) due to the recognition of normal tissues expressing the target antigen. The extracellular acidic characteristic of tumor tissues presents a novel mechanism to achieve target specificity.

Methods

In this study, we employed a structure-based computational approach to engineer anti-MSLN (mesothelin) VHHs with selective binding under acidic tumor microenvironment condition using methods developed as part of VHHMAb^® platform. Through in silico dual-pH His/Asp/Glu-scanning mutagenesis of the complementarity-determining regions (CDRs) and paratope amino acids, we optimized the VHH for acid pH selectivity.

Results

Testing of 20 designed variants identified four variants with more than 5-fold binding selectivity toward acidic pH. Notably, one variant (MT001) exhibited significant loss of binding at physiological pH while retaining binding activity under acidic condition in protein binding assays such as SPR. Similar pH-dependent behavior was confirmed using FACS assays at the cellular level. Furthermore, when incorporated into a chimeric antigen receptor (CAR) construct, MT001 conferred pH-dependent cytotoxicity to CAR-T cells, with enhanced cell-killing efficiency at acidic pH compared to neutral pH. This pH dependence was also observed in other CAR-T activation measures, such as CAR-T cell expansion and cytokine release after co-culture with MSLN+ tumor cell lines.

Conclusions

This study demonstrates the feasibility of computational optimization of antibodies for selectively targeting the acidic tumor microenvironment, representing a potential approach for developing safer CAR-T therapeutics.