An avidity-driven, dual-targeting ADC with superior tumor selectivity for the treatment of solid tumors

Background Antibody drug conjugates (ADCs) represent one of the most promising therapeutic modalities in recent years. However, the full therapeutic potential of ADCs is hampered because many tumor antigens are also expressed in healthy tissues, causing on-target, off-tumor toxicities. Other toxicities are caused by off-target uptake of ADCs in healthy cells or by systemic payload exposure as consequence of unstable conjugation. These issues have restricted a successful implementation of highly potent payloads such as pyrrolobenzodiazepine dimers (PBD). Method We developed a bispecific ADC (bsADC) which requires the simultaneous binding to two tumor-associated antigens (TAAs) to mediate cytotoxic activity (AND-gate approach). Co-expression of the TAAs is only found in tumor cells. Cell killing is therefore restricted to tumor cells. Healthy cells, which express only one of the TAAs, are spared. Our lead candidate comprises a bispecific antibody linked to a PBD prodrug at a DAR of 2 via a tumor-selective glucuronide linker using our proprietary site-specific, click-based glycan conjugation platform, which minimizes off-target toxicity by two mechanisms: first, using a highly stable linker, the payload is not released in circulation but only inside cells expressing both targets; second, hydrophilic chemistry prevents target-independent uptake. Results We identified a TAA pair that is highly co-expressed in various solid tumors including TNBC and NSCLC, while co-expression in healthy tissues is limited. It therefore represents a promising target combination for the development of an AND-gated bsADC with enhanced tumor-selectivity. Our lead candidate makes use of a pronounced avidity binding due to low monovalent affinities for both targets. This translates into superior internalization as well as potent killing of tumor cell lines expressing both targets in comparison to healthy cells expressing one target, resulting in an increased therapeutic window. In co-culture assays, the bsADC induced bystander killing of target-negative cells, indicating the bsADC should be efficacious in tumors where co-expression of both targets is not 100% penetrant. In vivo, our candidate dose-dependently reduced the tumor burden in mice bearing cell line-derived xenograft (CDX) models of breast cancer. In addition, the bsADC was well-tolerated up to high dose levels in mice and showed high stability in vitro and in vivo. Conclusion By combining an AND-gated, dual-targeting approach with our hydrophilic, highly tumor-selective, and stable ADC conjugation platform, we have generated a bsADC with superior tumor specificity. This enables the application of a potent PBD prodrug for highly efficient anti-tumor activity. Our lead bsADC for the treatment of lung and breast cancers has entered formal preclinical development and IND-enabling studies.