An animal model to predict the immunogenicity of murine V regions in humans

Clinical trials with genetically engineered chimeric mouse/human monoclonal antibodies have demonstrated that mouse variable regions differ dramatically in their degree of immunogenicity. These observations led us to search for an animal model that could predict mouse variable region (V region) immunogenicity prior to human clinical trials. We selected monoclonal antibodies 17-1A and B72.3 for study because human trials have demonstrated the very low immunogenicity of the mouse 17-1A V region and the high immunogenicity of the mouse B72.3 V region. Random-bred New Zealand white rabbits were injected intravenously with mouse 17-1A, B72.3, or both using a dose and schedule comparable to human trials. After initial injection only two of ten rabbits developed an antibody response to mouse 17-1A, while all five animals receiving a second injection developed antibody that was entirely mouse constant region-specific. On the other hand, nine of ten rabbits demonstrated an antibody response to initial infusion of mouse B72.3 that was greater than 90% specific for the complementarity-determining region components of the V regions. Competitive inhibition with isolated heavy and light chains demonstrated specificity for heavy or light chains (approximately 60%) or a requirement for both (40%). Thus, as in humans, the V region of 17-1A elicited little or no immune response in rabbits, while the V region of B72.3 was highly immunogenic. The observation that random-bred rabbits mimic the intensity and specificity of the human immune response to these two mouse V regions suggests that this animal model could be used to identify mouse monoclonals with V regions of low immunogenicity that would be good candidates for genetically engineered chimeric antibodies.