Achieving the optimal profile for Alzheimer’s immunotherapy: Rational generation of antibodies specific for toxic Aβ oligomers (P6.084)

Objective: Generate antibodies specific for toxic amyloid-β oligomers (AβO).

Background: Current evidence suggests that propagating prion-like strains of AβO, as opposed to monomers and fibrils, are preferentially toxic to neurons and trigger tau pathology in Alzheimer’s disease (AD). In addition, dose-limiting adverse effects have been associated with Abβ fibril recognition in clinical trials. These observations suggest that specific neutralization of toxic AβOs is needed to achieve optimal safety and efficacy.

Design/Methods: We employed computational simulations, using molecular dynamics with standardized force-fields to perturb atomic-level structures of Aβ fibrils deposited in the Protein Data Base. We hypothesized that weakly-stable regions are likely to be exposed in nascent protofibrils or oligomers. Clustering analysis, curvature, exposure to solvent, solubility, dihedral angle distribution, and Ramachandran angle distributions were all used to characterize the conformational properties of predicted epitopes, which quantify differences in the antigenic profile when presented in the context of the oligomer vs the monomer or fibril. The candidate peptide epitopes were synthesized in a cyclic format most likely to mimic regional AβO conformation, conjugated to a carrier protein, and used to generate monoclonal antibodies in mice. Purified antibodies were screened by SPRi and immunohistochemistry.

Results: Sixty-six IgG clones against 5 predicted epitopes were selected for purification based on their ability to recognize the cognate structured peptide and synthetic AβO, with little or no binding to unstructured peptide, linker peptide, or Aβ monomers. Additional screening identified antibodies that preferentially bound to native soluble AβO in CSF and brain extracts of AD patients compared to controls. Immunohistochemical analysis of AD brain allowed for selection of antibody clones that do not react with plaque.

Conclusions: Computationally identified AβO epitopes allowed for the generation of antibodies with the desired target profile of selective binding to native AD AβOs with no significant cross-reactivity to monomers or fibrils.

Study Supported by: PromisNeurosciences and Alberta Prion Research Institute

Disclosure: Dr. Plotkin has received personal compensation for activities with PromisNeurosciences as a consultant. Dr. Plotkin holds stock options in PromisNeurosciences, and has received research support from PromisNeurosciences. Dr. Gibbs has received supplementary stipend for activities with ProMIS Neurosciences. Dr. Gibbs has received compensation as stock option grant from ProMIS Neurosciences. Dr. Peng has nothing to disclose. Dr. Silverman has received personal compensation for activities with PromisNeurosciences as a consultant. Dr. Cashman has received personal compensation for activities with PromisNeurosciences. Dr. Cashman holds stock and/or stock options with Promis Neurosciences. Dr. Cashman has received licensing fees from University of British Columbia.