Silvia Grandoni (1), Nicola Cesari (2), Giandomenico Brogin (2), Paola Puccini (2), Paolo Magni (1)
(1) Università degli Studi di Pavia, Dept. Electrical, Computer and Biomedical Engineering, Pavia, Italy, (2) Chiesi Farmaceutici S.p.A, Pharmacokinetic, Biochemistry and Metabolism Department, Parma, Italy.
Objectives: the study of locally acting inhaled drugs pulmonary pharmacokinetics (PK) is challenging primarily for the lung poor accessibility and the absence of easily accessible fluids for direct or surrogate measurement of compounds concentration in the lung tissue [1]. Physiologically-Based PharmacoKinetic (PBPK) modelling has been deemed advantageous in this field, mainly for the possibility to predict the local exposure and has been recently applied [2]. The trend emerged is the building of models with an increasing complexity, to describe more accurately the spatial heterogeneity of lung anatomy, physiology and, consequently, of the related PK processes. However, this approach requires a great number of model parameters, some of which are unknown and their estimation on in vivo data is required. Furthermore, the published models have not been evaluated in the perspective of predicting lung concentration-time profiles in absence of in vivo data. Hence, the objective of this work was the development of a PBPK modelling framework that can be used from the discovery phase, when the in vivo data have not yet been generated, to prospectively predict the rat and human PK of different inhaled compounds, starting from the in vitro data, speeding up the drug development and reducing the end-stages failures.
Methods: an in-house pulmonary model, inspired by the one of Boger et. al [3] has been developed to describe the PK of inhaled drugs and integrated on a previously built and validated PBPK model for intravenous and oral experiments [4]. The main processes governing the pulmonary PK [5] i.e., deposition, dissolution, mucociliary clearance and absorption through the lung are included. The respective equations are parameterized using early in silico and in vitro compounds information.
As a first step, the model was built to predict the PK in rat and evaluated on preclinical data, therefore it has been extended to human. Its assessment was performed on rat total lung and plasma concentration data after intratracheal administration of 9 different compounds. As solubility and lung retention are two key aspects in the PK optimization at least in the early stages, test compounds have been selected to have different degree of solubility and lung retention, to obtain a more robust model. They were classified as highly and poorly soluble by using a criterion based on the dissolution number, similar to that used in Biopharmaceutical Classification System for orally administered drugs, being the definition of criteria for inhaled drugs still an open issue [6]. To discern highly retained compounds from the poorly retained ones, a classification based on the evaluation of the in vivo mean residence time (MRT) has been adopted.
After the evaluation on preclinical data, the model has been applied to predict the plasma drug concentration after inhalation in human. The core of the strategy lies in the reparameterization of the model used for rats with human physiological parameters and with in vitro compounds specific parameters related to human ADME (where possible). The assessment was performed on healthy volunteer plasma concentration data after inhalation of two different compounds.
Results: the model evaluation was based on the visual inspection of the total lung (for rats only) and plasma concentration predicted profiles against the in vivo data and on a fold error analysis. This last was performed on some PK metrics (e.g. Area Under Curve, MRT) and considering for each compound its class, to investigate possible trends in model misspecifications with respect to the compound classification. With few exceptions, the PK metrics calculated on the model predicted profiles are inside the two fold error limit and for both the species of interest.
Conclusions: for the first time, a PBPK modelling framework for the extrapolation of rat and human inhaled compounds PK from the in vitro data generated during the discovery phase has been proposed and evaluated on several compounds with different properties. The very encouraging results regarding rat PK prediction suggest its application for purposes such as compounds prioritization prior to first-time-in-animal studies and/or their optimization. The promising preliminary assessment results on healthy volunteer data encourages further evaluation to increase confidence for its potential use to predict the human PK in the early stages and for the preclinical to clinical translation.
References:
[1] B. Forbes et al., “Challenges in inhaled product development and opportunities for open innovation,” Adv. Drug Deliv. Rev., vol. 63, no. 1–2, pp. 69–87, 2011.
[2] P. Bäckman, S. Arora, W. Couet, B. Forbes, W. de Kruijf, and A. Paudel, “Advances in experimental and mechanistic computational models to understand pulmonary exposure to inhaled drugs,” Eur. J. Pharm. Sci., 2017.
[3] E. Boger et al., “Systems Pharmacology Approach for Prediction of Pulmonary and Systemic Pharmacokinetics and Receptor Occupancy of Inhaled Drugs,” CPT Pharmacometrics Syst. Pharmacol., vol. 5, no. 4, pp. 201–210, 2016.
[4] Grandoni S., N. Cesari, G. Brogin, P. Puccini and Magni P., “Building in-house PBPK modelling tools for oral drug administration from literature information,” ADMET & DMPK, vol. 7, no. 1, pp. 4–21, 2019.
[5] J. M. Borghardt, B. Weber, A. Staab, and C. Kloft, “Pharmacometric Models for Characterizing the Pharmacokinetics of Orally Inhaled Drugs,” AAPS J., vol. 17, no. 4, pp. 853–870, 2015.
[6] J. E. Hastedt et al., “Scope and relevance of a pulmonary biopharmaceutical classification system”, AAPS/FDA/USP Workshop March 16-17th, 2015 in Baltimore, MD,” AAPS Open, vol. 2, no. 1, p. 1, 2016.
Reference: PAGE () Abstr 9504 [www.page-meeting.org/?abstract=9504]
Poster: Oral: Drug/Disease Modelling