The population PKPD of unacylated and acylated ghrelin following single rising doses of BI 1356225 in healthy subjects is impacted by the time since last meal
Robin Svensson (1), Andrea Henrich (1), Susanna Bianzano (2), Donald Sarubbi (3), Daniel Röshammar (1,5), Nadine Beetz (4)
(1) Pharmetheus AB, Uppsala, Sweden, (2) Boehringer Ingelheim International GmbH, Ingelheim, Germany, (3) Boehringer Ingelheim Pharmaceuticals Inc, Ridgefield, CT, USA, (4) Boehringer Ingelheim Pharma GmbH & Co.KG, Biberach, Germany, (5) contracted as an external consultant by Translational Medicine and Clinical Pharmacology, Boehringer Ingelheim Pharmaceuticals
Introduction/Objectives: BI 1356225 is an inhibitor of the ghrelin O-acyl transferase (GOAT) enzyme. GOAT activates unacylated ghrelin (UAG) via acylation of its amino acid serine (acylated ghrelin, AG). Both, GOAT and UAG are expressed in endocrine cells of the stomach mucosa, as well as in specific regions of the brain. An increase in AG in these tissues results in stimulation of appetite, food seeking behaviour, and deposition of energy in fat tissue. The objective of this analysis was to describe the PKPD relationship between BI 1356225 and ghrelin (UAG and AG). From preclinical data it was predicted that at least 80% GOAT inhibition is needed to result in a clinically meaningful body weight loss after long-term treatment.
Methods: The data for the model development originated from a single rising dose study which investigated BI 1356225 in doses of 0.1, 0.3 and 0.9 mg (oral solution) and 2.5, 5, 10 and 20 mg (tablets) in healthy male subjects. In total, the dataset included 65 subjects of which 14 subjects were randomized to placebo. Ten out of 16 subjects in the 2.5 mg arm were part of a cross-over relative bioavailability part of the study and did not contribute with any PD samples. A sequential PKPD modelling approach was applied where, initially, a PK model was developed on 1067 plasma concentration measurements of BI 1356225 from 51 subjects (excluding subjects randomized to placebo). In a second step, a joint UAG and AG PKPD model was developed on 813 AG and 813 UAG samples from 55 subjects (including subjects randomized to placebo). PK was incorporated in the PKPD model using the ‘individual PK parameters (IPP)’ [1] approach where the continuous individual predicted plasma concentration was used as input. The final PKPD model was used to predict the required daily BI 1356225 dose in order to achieve at least 80% GOAT inhibition.
The analysis was conducted in NONMEM version 7.3. Model selections were guided by changes in the objective function value (OFV) using the likelihood ratio test as well as by using various diagnostic plots. In particular, visual predictive checks (VPCs) stratified on dose group were considered a key diagnostic plot during model development.
Results: The population PK of BI 1356225 was described by a two-compartment disposition model with first-order elimination and absorption. The UAG and AG data were described by two indirect response PKPD models, respectively. The formation of AG was dependent on the amount of UAG. The drug effect was incorporated as inhibition of the elimination of UAG and the formation of AG. The drug concentration associated with half of the maximum effect (estimated at 91% inhibition) was estimated at approximately 10 nmol/L. The elimination rate of both UAG and AG peaked right after a meal and then decreased until the next meal, which described the observed multiple daily peaks seen for both biomarkers. The timing of the drug effect assessment in relation to any food intake should hence be taken into consideration.
Conclusions: After accounting for the diurnal ghrelin variation related to food intake, inhibition of the GOAT enzyme resulted in a dose-dependent increase in UAG and a decrease in AG levels. A daily BI 1356225 dose of approximately 1 mg was predicted to achieve a median GOAT inhibition of at least 80% over the steady-state dosing interval. At this dose, the between-subject variability in the GOAT inhibition ranged from 55.8% (5th percentile) to 89.5% (95th percentile).
References:
[1] Zhang L, Beal SL and B SL, 2003, Simultaneous vs. sequential analysis for population PK/PD data I: best-case performance. JPKPD vol. 30: 387–404.