I. Singh,1 W. Liu,2 D. A. Katz,1 W. M. Awni,2 S. Dutta2; 1Former AbbVie employee, North Chicago, IL, 2AbbVie, North Chicago, IL

BACKGROUND: ABT-436 is a potent V1B receptor antagonist for potential treatment of major depressive disorder, anxiety disorders and substance use disorders. In Phase I studies, ABT-436 reduced serum cortisol and demonstrated pharmacologic effect on hypothalamus-pituitary-adrenal axis. The objective of the PK/PD analyses was to quantify and predict effect of ABT-436 on cortisol concentrations to inform Phase II dose selection.
METHODS: Intensive PK and cortisol data from 96 healthy subjects in 3 multiple-dose studies which explored QD, BID and morning/evening dosing were used for conducting PK/PD analyses in NONMEM 7. To account for the circadian-episodic profiles of cortisol, an indirect response model using various biorhythmic functions (single and dual cosines) and Fourier analysis with harmonics were explored. ABT-436 effect on cortisol was characterized by an inhibitory Emax model with different IC50s estimated for morning and evening dosing. Model development was guided by diagnostic plots and likelihood ratio tests and assessed using visual predictive checks. Clinical trial simulations were conducted to predict cortisol response across different dosing regimen.
RESULTS: Baseline cortisol concentration was optimally described by Fourier series with two harmonics (24- and 12-hr period). Estimated ABT-436 IC50 was 700 ng/mL with maximum inhibition of 80% of cortisol level. Estimated IC50s for morning vs. evening were not different. Simulations showed 800 mg QD (morning dosing) would provide a maximum 46% cortisol level change from placebo.
CONCLUSION: The developed PK/PD model well-characterized the baseline and drug-induced suppression of cortisol concentrations. Combined with safety analyses in Phase I studies, it provided useful tool to inform a Phase II dose selection.