Abstract
IGHV mutated (M-CLL) and unmutated (U-CLL) CLL arise from clonal expansions of antigen-experienced B lymphocytes. As the specific IGHV genes coding the clone’s B-cell antigen receptor (BCR) and the selection for particular genetic and amino acid structures shared by different patients are not random, a role for antigenic selection in biasing BCR structure is likely. However, the binding specificities of the BCR and the nature of the selecting antigens are yet to be identified. In order to characterize actual or mimetic ligands to BCRs in CLL patients, peptide phage display technologies were used. BCRs from 3 M-CLL and 2 U-CLL patients were expressed as IgG1 mAbs in human embryonic fibroblasts and used to probe a 12-mer peptide phage display library. For all the tested mAbs, ligands were isolated after 3 rounds of selection. M-CLL mAbs bound a large number of related, non-redundant phages carrying peptide inserts with conserved consensus motifs, whereas U-CLL mAbs reacted with multiple phages expressing identical 12-mer sequences for which no consensus or a consensus of shorter length and less definition than with M-CLL mAbs was found. Repeat probings with some of the same M-CLL or U-CLL mAb yielded identical and/or similar peptide-expressing phages. Ligands binding on the phage level were synthesized with a biotin-tag at the C-terminus and tested by direct ELISA for reactivity with the selecting mAbs as well as mAbs from different IGHV subgroups. In some instances, when synthesized as peptides, binding to the mAbs changed. For example, U-CLL peptides exhibited “polyreactitvity”, reacting with other U-CLL and M-CLL mAbs, although they still bound better to their selecting mAbs. For M-CLL, one “borderline” mutated mAb (2% difference from germline) had a peptide which bound other M-CLL and U-CLL mAbs. Two peptides derived from phages specific for another M-CLL mAb didn’t react with any antibody tested. For a third M-CLL mAb (CLL 169, IGHV3-33), two peptides of considerable specificity were identified that did not bind any other mAb tested. These peptides were also synthesized in MAP configuration, with a biotin tag, in order to mimic multivalent antigens; their reactivity in ELISA was comparable to the highly specific binding of the linear peptide. Tetramers were produced using streptavidin for the CLL169 peptide having the best ELISA binding as well as its MAP peptide counterpart. Binding was examined by FACS to the original CLL cells from which the mAb was derived, as well as to cells from another case with the same stereotypic IGHV, and to cells from a distinct IGHV subgroup. These studies indicate that binding to the cells depends not only on the VH, but also on the D and J segments as well as on the configuration of the peptide (linear, MAP, tetramer linear or tetramer MAP). Binding to the cells above background could be detected for the tetramer MAP peptide but not the linear peptide. The original cells bound best to the tetramer MAP peptide. The CLL cells with the same VH subgroup exhibited similar binding. Cells from a distinct VH sharing the same D and J segments as the original cells also showed some binding to the tetramer MAP peptide. Effects on CLL cell fate after in vitro incubation with the peptides were examined using 7AAD/Annexin V. Adding the tetramer MAP peptide to the CLL cells with the BCR from which they were derived led to 63% increase in cell death after 24 hours over that seen for control cultures. Thus, ligands specific for CLL BCRs can be identified by phage display technologies. For one peptide which displayed high specificity for the selecting mAb in ELISA, binding to CLL cells as well as increased cell death in vitro have been demonstrated. In future, this may lead to the identification of BCR-targeted therapeutic agents.
Conformational Epitopes on the Diabetes Autoantigen GAD65 Identified by Peptide Phage Display and Molecular Modeling
