Structure-based Optimized Antibody for Targeting Pancreatic Cancer

Diagnosis and staging of pancreatic cancer are commonly assisted by FDA approved in vitro assays conducted with monoclonal antibody (mAb) 1116NS19.9. This mAb recognizes an aberrant tetra-glycan carbohydrate antigen 19-9 (CA19-9) also known as Sialyl-Lewis A antigen, which decorates cell-surface proteins as well as proteins that are found in cancer patient sera. Shortcomings of this mAb include a failure to recognize the early stages as well as an insufficient positive predictive value. To address these shortcomings, we structurally characterized the recognition between the 1116NS19.9 mAb and its CA19-9 antigen. The structure revealed high chemical and geometrical complementarities between the mAb and the CA19-9 antigen. Structural analysis suggested very limited options for modifying the CDRs to enhance CA19-9 binding. As an alternative approach, we used our structure and an algorithm, called AbLIFT*, which uses phylogenetic analysis and Rosetta atomistic design to optimize the interface between the antibody variable domain light and heavy chains. Two designs exhibited tenfold improvement in K_D and are currently being further tested in ex vivo patient samples. Another well characterized mAb for targeting CA19-9 is mAb 5b1. We solved the structure in the bound form and found that it has a different solution for CA19-9 binding, suggesting a synergistic effect for combining Ab 1116NS19.9 and Ab 5b1 in treatment. Furthermore, our newly improved 1116NS19.9 designs and structural insights offer a potential reagent for earlier diagnosis of pancreatic cancer and immunotherapy.

* Web site of server: http://ablift.weizmann.ac.il/bin/steps