Epitope-specific anti-PrP antibody toxicity: a comparative in-silico study of human and mouse prion proteins.

Abstract

Despite having therapeutic potential, anti-PrP antibodies caused a major controversy due to their neurotoxic effects. For instance, treating mice with ICSM antibodies delayed prion disease onset, but both were found to be either toxic or innocuous to neurons by researchers following cross-linking PrP^C. In order to elucidate and understand the reasons that led to these contradictory outcomes, we conducted a comprehensive in silico study to assess the antibody-specific toxicity. Since most therapeutic anti-PrP antibodies were generated against human truncated recombinant PrP^91-231 or full-length mouse PrP^23-231, we reasoned that host specificity (human vs murine) of PrP^C might influence the nature of the specific epitopes recognized by these antibodies at the structural level possibly explaining the 'toxicity' discrepancies reported previously. Initially, molecular dynamics simulation and pro-motif analysis of full-length human (hu)PrP and mouse (mo)PrP 3D structure displayed conspicuous structural differences between huPrP and moPrP. We identified 10 huPrP and 6 moPrP linear B-cell epitopes from the prion protein 3D structure where 5 out of 10 huPrP and 3 out of 6 moPrP B-cell epitopes were predicted to be potentially toxic in immunoinformatics approaches. Herein, we demonstrate that some of the predicted potentially 'toxic' epitopes identified by the in silico analysis were similar to the epitopes recognized by the toxic antibodies such as ICSM18 (146-159), POM1 (138-147), D18 (133-157), ICSM35 (91-110), D13 (95-103) and POM3 (95-100). This in silico study reveals the role of host specificity of PrP^C in epitope-specific anti-PrP antibody toxicity.