B-cell receptors (BCR) on chronic lymphocytic leukemia (CLL) cells have stereotyped hypervariable regions, categorizing approximately 20% of all patients into one of almost 50 BCR subsets with nearly identical receptors. This suggests common antigen recognition in some CLL patients. We explored the binding pattern of recombinant CLL BCRs Fab fragments by selecting epitope mimicking peptides from phage display libraries. We performed selections with multiple libraries using different peptide lengths and constraints. Selection was successful on seven of nine Fab fragments with at least one library. The selected peptides specifically bound the Fab fragment they were selected on as well as the native BCR expressed on parental CLL cells. Subsequently, we validated in one example that the selected peptide truly mimics an epitope recognized by CLL BCRs. As CLL BCRs have long been suspected of interacting with a set of different antigens, we screened a random panel of CLL cases for binding six representatives of the selected epitope mimics described above. Surprisingly, all CLL samples recognized one or several of the epitope mimics. The majority of CLL samples could be assigned to one of 14 categories with specific epitope binding patterns. These categories were not predictable from the amino acid sequence of the complementarity-determining region 3 although samples expressing homologous BCRs preferentially clustered in the same category. In addition, we showed that this ligand-binding categorization is not only functionally but also clinically relevant. Patients belonging to the largest binding category had a poorer prognosis with a shorter time to first treatment than control patients. Our data suggest that CLL BCRs have a surprisingly confined structural diversity and that, if antigenic drive exists in CLL pathogenesis, the number of potential antigens may be much smaller than previously anticipated. Moreover, the peptides identified in this study may have potential as a targeting moiety for CLL-directed drug therapy.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author

Sign in via your Institution