Feras Khalil1, Klaus Lindauer1, Kishore Pasikanti2
1Grünenthal GmbH, 2GRT Therapeutics Inc
Introduction: First-in-human (FIH) clinical trials are designed to assess the pharmacokinetics (PK), tolerability, and safety of an investigational medicinal product in humans [1]. These trials typically follow a stepwise exposure escalation approach, employing single ascending dose (SAD) and multiple ascending dose designs. Given the paramount importance of participant safety, dose selection is usually guided by quantitative modeling techniques to predict human exposure based on preclinical data, supplemented by real-time assessments of emerging clinical data to refine subsequent dose levels if needed. Quantitative modeling approaches, including allometric scaling and population pharmacokinetic (popPK) modeling, play a crucial role in predicting human PK while informing and optimizing key design elements of the FIH trial. These include determining an appropriate starting dose, defining a dose range to be explored that is expected to be safe and efficacious, optimizing the PK sampling schedule, and selecting tablet strengths for manufacturing. Accurate human exposure and dose predictions streamline decision-making and reduce operational burden. Retrospective evaluation of human PK prediction is essential for improving prediction methods and strategies for future candidate drugs [2]. Objectives: This analysis aimed to predict human PK and dose for the FIH trial of GRT6018, a potent and selective agonist of the nociceptin/orphanin FQ peptide (NOP) receptor, supporting the determination of the starting dose and the anticipated therapeutic dose (ATD) range. Additionally, we present a retrospective evaluation of the pre-trial exposure predictions. Methods: Human PK was predicted using an integrated analysis of in vivo concentration-time data from multiple animal species, applying a nonlinear mixed-effects modeling approach combined with allometric scaling. PK data were collected from four different animal species (mouse, rat, monkey, and dog) across 11 preclinical studies following intravenous or oral administration of single or multiple doses, up to 35 mg/kg. Data analysis was conducted using NONMEM software (version 7.4; ICON Development Solutions), supplemented with the PsN toolkit (version 3.6.2) [3]. Following popPK model development, clinical trial simulations (CTS) were performed using the estimated model parameters to predict human drug exposure, assuming a reference body weight of 60 kg and oral bioavailability (F) of 35%. Additional data, including plasma protein binding, nonclinical drug biodistribution, and results from native NOP binding assays, were integrated to estimate systemic and tissue target engagement (TE), supporting the determination of the starting dose and the ATD range. Results: GRT6018 PK was best described by a two-compartment model with linear absorption and elimination. Bodyweight-based allometric scaling was applied to all disposition parameters. The allometric exponents for clearance (CL) and intercompartmental clearance (Q) were fixed at 0.75, while volume parameters (Vc and Vp) were fixed at 1, with no significant difference in results if these exponents were estimated. Predicted human systemic CL and steady-state volume of distribution (Vss) in a 60-kg subject were 28.2 L/h (0.47 L/h/kg) and 184.2 L (3.07 L/kg), respectively, with a predicted terminal half-life of 4.5 hours. CTS supported the selection of a starting dose (1 mg) based on minimal anticipated biological effect level and a TE-guided dose range (1–150 mg). Predicted PK parameters, concentration-time profiles, and key exposure metrics (maximum concentration [Cmax], area under the concentration-time curve extrapolated to infinity [AUCinf]) aligned well with observed values.The absolute prediction error (APE) for mean Cmax across all SAD cohorts ranged from 16% to 59% (average 40%) and from 3% to 37% for mean AUCinf (average 25%). Based on interim PK data for the first dose escalation meeting, adjusting assumed oral bioavailability from 35% to 45% improved the a priori prediction accuracy, reducing average APE from 40% to 25% for mean Cmax and from 25% to 15% for mean AUCinf. No adjustments to the planned doses were required during the trial. Conclusions: Allometric scaling, combined with popPK modeling and CTS, enables robust translation of preclinical PK data into human PK and dose predictions for FIH trials. This case highlights the value of allometric popPK modeling for a Biopharmaceutics Classification System Class I compound. Accurate assumptions of oral bioavailability are critical for optimal exposure predictions, suggesting that physiologically based PK modeling may be required to support FIH predictions for drugs with more complex PK profiles, such as those with non-linear or solubility-limited absorption.
[1] European Medicines Agency. Guideline on strategies to identify and mitigate risks for first-in-human and early clinical trials with investigational medicinal products (europa.eu). EMA/CHMP/SWP/28367/07 Rev.1; 2017. [2] Davies, Michael, et al. Trends in pharmacological sciences 41.6 (2020): 390-408. [3] Lindbom et al. Computer methods and programs in biomedicine 79.3 (2005): 241-257.
Reference: PAGE 33 (2025) Abstr 11500 [www.page-meeting.org/?abstract=11500]
Poster: Drug/Disease Modelling - Other Topics