Comment on Binder et al, page 1975
The exact binding site of rituximab on CD20 has not yet been revealed. In this issue of Blood, Binder and colleagues demonstrate that the binding site is a discontinuous epitope, which they identified by applying mimotope technology.
Since its approval for the treatment of B-cell lymphoma in 1997, the monoclonal antibody rituximab (Rituxan; Genentech, South San Francisco, CA/MabThera; Roche, Basel, Switzerland) is widely applied. Its effector mechanisms include in vivo ablation of malignant B cells via classical antibody- and complement-dependent cytotoxicity and sensitization to chemotherapy. Moreover, induction of apoptosis through CD20 cross-linking contributes to rituximab function.1 FIG1
Although CD20 was first described in 1980, little is known about the rituximab-binding site on its target molecule. Biophysical and structural studies of CD20 were performed only recently, rendering a topologic model. CD20 is a transmembrane protein, with an extracellular region that comprises 47 amino acid residues. The tertiary structure of this extracellular loop is determined by a disulfide bond between C(167) and C(183). Disruption of this bond leads to loss of rituximab binding, indicating the recognition of a conformational epitope.2 Earlier mutation experiments indicated that the sequence AxP at positions 170 to 172 in human CD20 is critical to determining the structure of the extracellular loop and the fine specificity of anti-CD2 0antibodies.3
Mimotopes are small peptides that mimic natural B-cell epitopes structurally. A mimotope is identified by screening a random peptide phage display library with an antibody of interest. The mimotope peptide sequence sometimes differs from the original antigen sequence, because mimotopes mimic not only linear epitopes, but also conformational ones.4 Mimotopes are thus potent tools for gathering information about most B-cell epitopes.5 In this issue of Blood, Binder and colleagues generated such peptide mimics for the epitope recognized by rituximab. Only one mimotope displayed sequence homology to the 170 to 172 string. Other rituximab mimotopes shared a consensus motif between them that was not homologous to CD20. Therefore, the authors translated the aromatic and polar amino acids in the consensus motif into homologous amino acids with higher impact in antigen-antibody interactions, namely tyrosine and serine. The resulting sequence showed homology to a second section of CD20 (ie, between amino acids 182 and 186). Importantly, in the tertiary structure of CD20, both sections are brought in contact by the disulfide bond. To prove that these 2 regions together form one epitope, the authors designed a peptide that combines these 2 motifs. The recognition of this construct by rituximab was indeed substantially higher than that of either motif alone. The authors further strengthened their finding by strategically mutating the 2 identified regions of CD20, thereby reducing rituximab-binding affinity. On the structural model of CD20,2 the rituximab epitope was localized in proximity to the disulfide bond, near the plasma membrane (see figure).
What can we learn from these experiments? Even conformational B-cell epitopes that escape identification via linear peptide screens can be identified and characterized from studies with mimotopes. These epitope equivalents can further be exploited for active, epitope-specific vaccination against discontinuous epitopes.6 Knowledge of an epitope in the structural context can help to fully understand antibody effects. For example, the growth inhibitory function of the anti–epidermal growth factor receptor (EGFR) antibody cetuximab has been explained by the antibody's ability to sterically prevent the receptor from adopting the extended, active conformation.7 The identification of the rituximab epitope should trigger further studies, ultimately explaining why rituximab has such a profound effect on B-cell life. ▪