Chiara Roversi (1), Annalisa Mercuri (1), Simone Zannoni (1), Marco Borgogno (2), Annalisa Savardi (2), Marco De Vivo (2), Laura Cancedda (2), Dino Montanari (1), Marco Pergher (1), Valentina Cenacchi (1), Alessia Tagliavini (1)
(1) Evotec SE (2) IAMA Therapeutics
Introduction/Objectives:
IAMA-6 is a small molecule, selective inhibitor of the sodium potassium chloride cotransporter (NKCC1), showing encouraging in vitro and in vivo proof‑of-concept results in idiopathic autism and Down syndrome. In this work, a stepwise methodology was applied to extrapolate IAMA-6 human expected exposure levels to support the selection of the starting dose for First In Human (FIH) trial, integrating a population (popPK) and physiologically-based (PBPK) pharmacokinetic modelling strategy.
Methods:
Data coming from in vivo studies of three preclinical species (i.e. rat, mouse, and dog) treated with single or repeat dose of IAMA-6 given either orally, intravenously or intraperitoneally at different doses were pooled to develop three popPK models, one for each species, to characterize subjects PK profile. PopPK models were developed by using nonlinear mixed effects (in MonolixSuite™), thus also including the interindividual variability term on some parameters to explain inter-subject differences.
Additionally, a mechanistic PBPK model was built with GastroPlus® combining in silico predicted data (ADMET Predictor) with in vitro and in vivo measured data. Renal clearance (CL) was predicted multiplying plasma fraction unbound for glomerular filtration. Hepatic CL was estimated subtracting renal CL to the experimental total CL. All other parameters were set as default. Tissue distribution (Kp) was defined using the Lukacova (RodgersSingle) model.
Finally, a multispecies allometric scaling was applied to the PK parameters derived from popPK models to predict human drug distribution and elimination, by correcting for the protein binding difference across species and assuming a human body weight of 70 kg [1]. Two different human PK models were implemented via either GastroPlus® (PBPK model) and Simulx (popPK model) to properly characterize the absorption process and to simulate different scenarios in humans at different doses to support FIH dose selection. The predicted exposures, i.e. Area Under the Curve between 0 and 24 hours after dosing (AUC24), of the simulated population were compared to a safety threshold derived applying a 20-fold safety margin to rats’ No Observed Adverse Effect Level exposure [2]; the maximum dose for which at least 75% of the population has the exposure below the threshold was selected as the human recommended starting dose.
Results:
Developed popPK models considered various absorption models to account for the observed heterogeneous absorption process across species. Specifically, a simple first‑order process well captures mouse data, while for rat and dog models a double first‑order absorption model with a lag‑time in the onset of the second process was adopted to describe the individual PK curves, where multiple peaks appear. With regards to the distribution and disposition processes, a two compartmental model with linear elimination was adequate in each species. Overall, popPK models shown good capability in describing the PK profiles and precise estimated parameters (Residual Standard Error <50%).
The PBPK model predicted well all preclinical species investigated, especially when enterohepatic circulation was considered. In addition, the model predicted rapid precipitation of the compound in the stomach, following redissolution and absorption along the intestinal tract. Indeed, multiple peaks were observed in the in vivo profiles and the model was able to correctly predict IAMA-6 oral in vivo behaviour. Human predictions were made by using the same assumptions used for the preclinical species and the allometrically scaled CL.
Human predicted parameters derived via multi-species allometric scaling were: CL=32.95 L/h, central compartment volume V1=31.89 L, peripheral compartment volume V2=18.24 L and inter-compartmental clearance Q=5.53 L/h. The predicted human elimination half-life based on allometrically scaled parameters was T½ = 2.79 hours. The integrated PopPK-PBPK approach provided a human recommended starting dose of 30 mg/day, corresponding to median (25th-75th percentiles) predicted AUC24 following five Quaque Die (QD) administrations of 465 (223-893) ng·h/mL.
Conclusions:
An integrated approach based on popPK and PBPK modelling has been proposed to predict IAMA-6 human exposure at different doses to support FIH dose selection to be tested for upcoming clinical trials. Accordingly, a human recommended starting dose of 30 mg/day was identified.
References:
[1] Patel D, Dierks E. Single‑species Allometric Scaling: A Strategic Approach to Support Drug Discovery. Journal of Pharmaceutical Research International. 2018 May;22(3):1‑7.
[2] EMA. Guideline on the Strategies to Identify and Mitigate Risks for First‑in‑human Clinical Trials with Investigational Medicinal Products. Committee for Medicinal Products for Human Use (CHMP). 2017 Jul.
Reference: PAGE 32 (2024) Abstr 10903 [www.page-meeting.org/?abstract=10903]
Poster: Drug/Disease Modelling - CNS