P1-361: Improved interpretation of amyloid tracer kinetic data through model-based analysis

Tracer kinetic approaches such as stable-isotope labeled kinetics have been proposed to measure production and clearance rates of amyloid 1. In this experimental medicine study, timecourse data on CSF tracer kinetics and ELISA-based total CSF Aβ were obtained under unaltered, mild and potent production inhibition in healthy human. Mechanistic pathway model-based analysis approaches were compared with statistical approaches to enhance understanding of the utility and interpretation of tracer kinetic data. Subjects (n=5/arm) received single doses of placebo, low or high dose BACE inhibitor (MK-8511) with 13 C-labelled leucine infused from 5 to 15 hours post-dose. Serial plasma and lumbar CSF samples were obtained for determination of drug concentration, total Aβ, and fraction of leucine and Aβ labeled. Fraction-label incorporation up-slope and newly generated Aβ were determined. Additionally data were fit to a mechanistic amyloid pathway model which included terms for the major steps of the amyloid pathway (precursor production, amyloid brain production and distribution to the lumbar CSF sampling site). The ELISA-based total Aβ data exhibited a larger drug effect than the fraction-label up-slope in both BACE arms. Mean (90% CI) total Aβ AUC 12–24hr ratios relative to baseline were 0.76 (0.65, 0.86) and 0.40 (0.33, 0.48) for low and high doses, respectively, while the corresponding changes in up-slope vs placebo were 0.88 (0.76, 1.03) and 0.67 (0.53, 0.84). The model was able to simultaneously account for the timecourses of total Aβ, fraction labeled Aβ, and newly generated Aβ with a single drug action (inhibition of BACE) suggesting no disconnect in the level of brain production inhibition implied by the differing effect sizes noted above. Simulations predicted that the up-slope was dependent on factors beyond level of production inhibition such as timing of the 13 C-leucine infusion. Tracer kinetic approaches have the potential to enhance understanding of the dynamics of the amyloid pathway. The up-slope measure appears to under-estimate the level of brain production inhibition during BACE treatment. Model-based analysis accounting for tracer dynamics throughout the amyloid pathway enhances interpretation of data and allows for more accurate estimation of drug effects. Ref 1: Bateman et al, 2006 Nat. Med.