Figure 1

Comparisons of central nervous system (CNS) monoclonal antibody (mAb) exposure in a static influx model versus a cycling influx and efflux model. Based on estimates that ~20 μM of amyloid β (Aβ) (~100 mg) are deposited in the Alzheimer’s disease brain, the potential target engagement in each model is shown. The parameters used correspond to human studies using a 400 mg dose of anti-Aβ mAb. A method to estimate exposure based on a trapezoidal method for estimating the area under the curve was used with a theoretical cycling time of 1 hour (complete exchange) and antibody half-life of 21 days. In the cycling model, the estimate of how much Aβ could be targeted in the brain is almost certainly an overestimate as the model does not take into account the efficiency of antibody binding within the brain and the extent to which binding of plasma Aβ or other peripheral sources could decrease the amount of free mAb entering the brain. The issue of how much binding of plasma/peripheral Aβ might decrease free mAb exposure in the brain is complex and will be related to the target epitope and antibody affinity. Given an estimate for the daily turnover of Aβ in plasma of ~50 nmol, if the mAb bound all plasma Aβ produced in a day and the binding was essentially irreversible (as has been observed for several anti-Aβ mAbs), then that binding would be predicted to reduce the exposure of free antibody in the brain by ~50%. BBB, blood-brain barrier.