MECHANISTIC MATHEMATICAL MODEL OF THE IMPACT OF ANTI-AMYLOID-BETA DRUGS AND BACE INHIBITORS IN ALZHEIMER’S DISEASE

Abstract

Clinical trials for Alzheimer's Disease (AD) have had a high failure rate (99.6%). The reasons may include insufficient disease understanding, lack of targeting to the right stage of disease or patient population, and insufficient drug exposure in the brain. To maximize the learning from past clinical experience, we established a single mechanistic mathematical model of mild/moderate AD describing molecular mechanisms leading to plaque formation with the help of clinical biomarkers and capture the effects of multiple anti-amyloid beta (Ab) monoclonal antibodies and amyloid targeting β-secretase (BACE) inhibitors. The AD model includes three different pools of Ab species: monomer, soluble oligomer, and plaque and describes Ab production, aggregation, the intercompartment transport as well as antibody-dependent cellular phagocytosis of plaque. The model was informed by data from stable isotope labeling kinetics experiments, molecular binding and clinical biomarker data for four anti-Ab monoclonal antibodies (Aducanumab, Crenezumab, Solanezumab, Bapineuzumab) and three small molecule BACE inhibitors (Elenbecestat, Verubecestat, Atabecestat). The model was calibrated to capture plasma and CSF drug PK, Ab levels and plaque burden from published clinical studies. Because of the complexity of Ab species, binding affinities of antibodies to these species can be difficult to measure, and often a wide range of values is reported for the same antibody. By fitting the model to all available Ab data in different compartments, we were able to estimate apparent in vivo affinities to different Ab species for the four anti-Ab antibodies. Our analyses indicate that for drugs that showed no or minimal plaque reduction in clinical studies, doses that are higher than those tested could lead to plaque reduction. Drugs that induce ADCP can lead to significant plaque reduction at lower doses and with faster kinetics than drugs that do not include this mechanism. We have developed a mechanistic mathematical model to analyze treatment effects of anti-Ab antibodies and BACE inhibitors on soluble Ab species and plaque. The results allow us to compare the effects of various drugs, mechanisms and combinations on plaque burden. In the future, the model will be expanded to capture drug effects at different stages of disease.