(585e) Exploiting Reactive Chemical Functionality in Antibodies to Introduce Metalloproteinase-Targeting Functional Groups

Extracellular proteases and peptidases are known to play key roles in remodeling the tumor microenvironment, but the lack of potent, specific inhibitors of the individual active enzymes in the microenvironment limits our understanding of the underlying cancer biology. We are exploring the hypothesis that integrating additional chemical functionality into antibodies and other binding proteins will lead to potent inhibitors that simultaneously exploit the best features of proteins and small molecules. To pursue this hypothesis, we have established a platform for combining yeast display and noncanonical amino acids (ncAAs) to enable rapid generation and characterization of “hybrid” structures. As we continue to advance this platform, it is critical to understand how best to introduce and exploit chemical functionality in displayed proteins to identify strategies for efficiently identifying potent, specific enzyme inhibitors. Here, we describe our latest findings regarding the introduction of chemical functionality into antibodies displayed on the yeast surface and examinations of the functional consequences of the addition of chemical groups in previously reported antibodies targeting matrix metalloproteinase-9.

The introduction of azide functionality into yeast-displayed proteins provides a means of chemical modification via either copper-catalyzed or strain-promoted azide-alkyne cycloadditions. We previously demonstrated the introduction of an aromatic azide, p-azido-L-phenylalanine, into yeast-displayed proteins using stop codon suppression-based ncAA incorporation with a stop codon readthrough efficiency of greater than 10 percent. In characterization of an expanded range of orthogonal translation machinery, we identified a system that supports the introduction of an aliphatic azide, N⁶-((2-Azidoethoxy)carbonyl)-L-lysine, into displayed proteins. Stop codon readthrough efficiencies with this amino acid are also calculated to be upwards of 10 percent. We have confirmed that both of these amino acids support copper-catalyzed azide-alkyne cycloadditions on the yeast surface. The ability to introduce azides with distinct side chain structures in proteins will enable the exploration of the role of the linker connecting chemical functionality to proteins in the construction of “hybrids.”

To understand how best to leverage azide functionality in displayed proteins, we are investigating strategies for incorporating ncAAs into antibody fragments at positions located close to and within antibody complementarity determining regions. Using two previously reported antibodies targeting matrix metalloproteinase-9 (MMP-9), we evaluated in detail the effects of azide introduction on antibody chemical reactivity and antibody function. After confirming that the two antibodies, an inhibitory antibody named DX-2802 and a noninhibitory antibody referred to here as M0076, retain their binding function when displayed on the yeast surface, we introduced azide functionality into these antibodies at positions L1 and L93 of the antibodies. We observe partial retention of MMP-9 binding upon the introduction of aromatic azide functionality at either one of these positions in DX-2802 and M0076, while introduction of the aliphatic azide functionality appears to cause greater reductions in MMP-9 binding. These findings suggest that the existing genetic code manipulation systems are sufficient to support the generation of functional, azide-substituted antibodies in some cases. However, the results also imply that greater stop codon readthrough efficiency is needed to better exploit these functionalities in proteins, or that azide substitutions at these positions diminish antibody binding function. Enhancements to the stop codon suppression systems and engineering of antibodies to better tolerate azide substitutions could further enhance the performance of ncAA-compatible yeast display.

Simultaneously, we are examining the introduction of MMP-targeting functional groups into these azide-substituted proteins. We performed reactions with several alkyne-containing small molecules and azide-containing, yeast-displayed antibodies and showed that the extent of reaction that can be achieved with the small molecules depends on the identity of the small molecule, but changes to the reaction-promoting ligand and the azide side chain do not appear to strongly affect the extent of reaction. In preliminary studies, it appears that conjugating small molecules to some azide-substituted antibodies results in antibodies that are still able to bind to MMP-9. This retention of binding after chemical conjugation indicates that utilizing copper-catalyzed azide-alkyne conjugations will be suitable for identifying hybrids in which antibody and small molecule functionality synergistically contribute to the binding and inhibition of MMP-9 function. The studies reported here provide a strong foundation for the generation and evaluation of hybrid inhibitors using both rational and high-throughput strategies; these efforts are ongoing.