In silico fragment-based peptide design targeting undruggable proteins for enhanced detection of circulating tumor cells

초록

Peptides are promising candidates for developing effective binders targeting undruggable proteins. However, conventional peptide development methods face significant challenges in precisely targeting specific protein domains. Herein, we strategically integrated in silico library screening with fragment-based drug design to identify peptides capable of effectively covering the wide and shallow PD-L1 surface from pools with randomized amino acid compositions. The identified peptide exhibited significantly enhanced PD-L1 affinity compared to its constituent fragment sequences and a validated control peptide of similar length. Furthermore, the peptide demonstrated antibody-level binding to PD-L1-expressing cancer cells when immobilized onto a surface in a multivalent manner. The remarkable PD-L1-binding capability was translated into superior detection efficacy for circulating tumor cells (CTCs) in blood samples obtained from cancer patients, surpassing that of an antibodyimmobilized surface. This high-sensitivity capture of PD-L1-expressing CTCs enabled pathological assessment, as CTC counts from the peptide-immobilized surface strongly correlated with tumor PD-L1 expression. Altogether, this peptide development strategy, which has the potential to be applied to most protein domains, provides a new perspective on the development of protein binders for diagnostic and therapeutic applications.

키워드