Abstract

Background: LY3039478 is an oral Notch inhibitor that prevents release of the Notch Intracellular Domain (NICD) by inhibiting proteolytic activity of the gamma (γ)-secretase complex. LY3039478 also inhibits production of plasma amyloid-beta (Aβ), which can be used as a biomarker to determine pharmacodynamic (PD) effects of LY3039478. Covariates affecting the pharmacokinetics (PK) of LY3039478, and also PK-PD relationships, are explored in the current analysis. The analysis included PK and PD data from advanced cancer patients from the first human dose trial and healthy volunteers (HV) from the single ascending dose and pilot relative bioavailability trial. Methods: A summary of the study design elements can be found in Table 1. Population PK and PK/PD models were developed using non-linear mixed effects approaches using NONMEM Version 7.3. A sequential modelling approach was used. Study Parameters First human dose (NCT01695005) Single ascending dose and pilot relative bioavailability (NCT02659865) Population Patients with advanced cancer Healthy volunteers Dose 2.5 to 100 mg thrice weekly (TIW) in dose escalation part; 50 or 75mg TIW in cohort expansion part Part 1: single doses of 25, 50 and 75mg Part 2: single 50 mg doses in 2 period crossover study Formulation Drug in capsule (DiC) Part 1: DiC; Part 2: DiC and formulated capsule (fC) Sample collection Dose escalation patients: Intensive plasma samples up to 30 hours (after 1st and 10th dose); urine samples up to 10 hours (1st dose only) Cohort expansion patients: Sparse plasma samples up to 8 hours (1st dose only) Intensive plasma and urine samples up to 48 hours in both parts Table 1. Study Design Elements of the First Human Dose and Healthy Volunteer Trials Results: The population PK/PD model was developed using data collected from 102 patients and 34 HVs with age range 18 to 82 years and weight between 41 to 140 kg. A 3-compartment PK model with first-order absorption and lag time, plus a urine compartment, best described the LY3039478 disposition. LY3039478 was rapidly absorbed with peak concentrations occurring within 1 to 2 hours. Renal clearance contributed to approximately 20% of the apparent plasma clearance. Formulation and population effects, weight, hematocrit, and glomerular filtration rate affected the disposition of LY3039478. Precursor-dependent indirect response models were used to describe the PK/PD relationships for Aβ inhibition. The PK/PD model identified a maximal plasma Aβ inhibition of close to 100%, and 50% of maximal effects occurred at low LY3039478 concentrations. Patients with advanced cancer appeared to have higher plasma Aβ at baseline and slower loss of response than HVs. Conclusion: When given similar doses of LY3039478, healthy volunteers had higher apparent plasma clearances and therefore lower exposures than patients with advanced cancer. Weight, hematocrit, and glomerular filtration rate were significant covariates affecting the disposition of LY3039478. Patients with advanced cancer appeared to have higher plasma Aβ at baseline and longer sustained response to LY3039478 compared to healthy volunteers.

By Eunice Yuen, Robert Bell, Maria Posada, Christophe Massard, Jordi Rodon, Claire Smith, Katharine Thorn, JT Diener, Demetrio Ortega, Jeffrey Suico, Karim Benhadji