Associate Chair and Associate Professor of Chemical Engineering University of Michigan, Michigan, United States
Antibody drug conjugates (ADCs) are coming of age with the approval of six new drugs in the past three years. These therapeutics combine a potent small molecule payload, often a cytotoxic drug, with a tumor antigen targeting monoclonal antibody paired through a chemical linker. However, the gains in this field have been hard won following a significant investment and many initial failures in the clinic. The increased complexity of ADCs compared to small molecule drugs or biologics makes them challenging to develop, and previous methods of drug screening and down-selecting candidates along the pipeline don’t work well for ADCs. The tumor delivery challenges of large molecules can limit the tumor uptake and distribution of ADCs in vivo, and toxicity from the small molecule can limit tolerability. Therefore, the most potent compounds in vitro may be too toxic for efficacy in vivo. Similarly, the antibody cross-reactivity and payload tolerability in animals is different than humans, meaning preclinical results may not adequately capture the clinical therapeutic window. All this results in traditional development pipelines selecting for drugs that may lack a therapeutic window, leading to many failures in the field. Even worse, this approach can screen out compounds that may be clinically effective but don’t perform well in preclinical assays. However, by combining experimental data with computational simulations of drug distribution, the clinical efficacy at tolerable dose can be predicted to identify promising ADC designs. These include selecting the right payload potency, linker type, drug to antibody ratio (DAR), and internalization rate for a given target. The success of these designs is illustrated with currently approved ADCs for solid tumors. In particular, we highlight how the physicochemical properties of the released payload determine its ability to reach all cells within a tumor, including antigen-negative cells that can escape targeting by the ADC. The significant investment in ADC development over the past two decades has generated an enormous number of useful tools for constructing ADCs. By utilizing this knowledge in computational ADC drug design, we can assemble the next generation of ADCs tailored to a particular target to increase the rate of clinical success.
