Butler, P. M.; Francis, A.; Meunier, A. L.; Anderson, A. K.; Hennessey, E. L.; Miller, M. B.; Lemere, C. A.; Selkoe, D. J.; Stern, A. M.

Importance: Anti-amyloid immunotherapy is used to treat Alzheimer disease (AD) with moderate benefits and potentially serious side effects due to amyloid related imaging abnormality with effusions/edema (ARIA-E). Different anti-amyloid antibodies have different in vitro binding characteristics to different synthetic A{beta} aggregates, leading to the assumption that they bind different species in the human brain. Lecanemab is hypothesized to bind \"protofibrils,\" but these are not well-characterized in human brain. It is also unknown how binding differences correlate with ARIA-E rates. The APOE {epsilon}4 allele increases ARIA-E risk, but how it affects antibody binding characteristics is unknown. Objectives: To determine whether anti-amyloid antibodies bind different species of human brain A{beta} and whether these binding properties to human brain A{beta} explains ARIA-E rates. Design: Cross-sectional study of 18 postmortem human brains. Setting: Single tertiary care hospital. Participants: Deceased patients with AD and cerebral amyloid angiopathy (CAA). Main Outcomes and Measures: Equilibrium binding constants (KD) and total A{beta} binding (Bmax) of recombinant aducanumab, lecanemab, and donanemab equivalents to human brain soluble and insoluble amyloid plaque-enriched and CAA-enriched A{beta} aggregates. Results: Lecanemab did not bind with greater affinity to the soluble fraction of A{beta} compared to aducanumab. All three antibodies were bound essentially identical quantities of A{beta} across the 18 cases and fractions (Pearson\'s r 0.84 -- 0.97). Antibody preference for plaque vs CAA A{beta} did not differ in soluble fractions but differed slightly in insoluble extracts. The APOE {epsilon}4 allele led to a more soluble antibody-accessible A{beta} pool in a dose-dependent manner for all three antibodies. Conclusions and Relevance: The lecanemab binding target in human brain is unlikely to be distinctly \"protofibrillar\" compared to other antibodies. Differences in antibody preference for plaque vs CAA A{beta} are unlikely to fully explain differences in ARIA-E rates. The APOE {epsilon}4 allele may plausibly increase ARIA-E risk by making antibody-accessible A{beta} more soluble. These results have implications for improving the safety and efficacy of current and future anti-amyloid antibody therapies.