(M1030-03-15) Predicting the Pharmacokinetics of Amorphous Dispersions Using Canine PBPK Models

Purpose: One of the most common reasons for reduced therapeutic effects of oral drug candidates is poor pharmacokinetics due to solubility limitations. Enabling formulations such as amorphous dispersions can mitigate these issues by enhancing solubility, leading to increased absorption and bioavailability. This study utilizes in-silico predictions for selecting amorphous dispersion formulations to guide solubility enhancement. The dissolution profiles of these formulations were incorporated within physiologically-based pharmacokinetic (PBPK) models to predict the increase in systemic concentrations for a poorly soluble compound.

Methods: This study first employs a predictive modeling tool that leverages validated algorithms to determine appropriate polymer excipients for solubility enhancement based on the compound’s physicochemical properties. These amorphous dispersion formulations were manufactured and subsequently tested and validated in bio-relevant in-vitro dissolution experiments. The in-vitro dissolution profiles for the amorphous dispersions were used in developing canine PBPK models with GastroPlus® that considered a combination of predicted, measured, and fitted parameters. The predicted enhancement in systemic exposure was validated with in-vivo data in dogs.

Results: Compound X has an aqueous solubility of 1.2 µg/mL at pH 7, with low measured bioavailability in-vivo in dogs. Excipient selection models predicted Hypromellose Acetate Succinate – M (HPMCAS-M), and Polyvinyl acetate phthalate (PVAP) as lead polymers for improving solubility of the compound. In-vitro non-sink dissolution experiments confirmed these findings, as the solubility increased by at least five-hundred-fold with the amorphous dispersions (Figure 1). The solubility enhancements translated to the in-vivo setting with systemic exposure increasing nearly five-fold corresponding to similar increases in the fraction absorbed and bioavailability. Simulations of dog PK of the amorphous dispersions were confirmed with in-vivo studies as the predictions for C_max, AUC_0-t, and T_max were in good alignment with observed values (Table 1, Figure 2).

Conclusion: This study implemented in-silico tools for solubility enhancement and PBPK modeling to accurately predict the impact of formulation changes on the pharmacokinetics of a poorly soluble compound. In addition to providing insight into the appropriate excipient selection, PBPK modeling enabled mechanistic evaluations of changes in-vivo. The accuracy of these PBPK models demonstrated the ability to evaluate how formulation changes can influence pharmacokinetics of poorly soluble compounds. This approach can minimize the need for extensive in-vivo studies with pre-clinical species.

Acknowledgements: Jim Mullin, Simulations Plus.


Figure 1. Bio-relevant In-vitro Dissolution. In-vitro dissolution experiments of the recommended excipients for enhancement of solubility of Compound X. Only free concentrations of the compound are available to cross membranes for absorption and are shown in this figure. Carboxymethylcellulose (CMEC) had the highest maximum concentration of the active compound released, while HPMCAS-M and PVAP had similar dissolution profiles. All amorphous dispersions resulted in higher solubility compared to crystalline active pharmaceutical ingredient (API). Gastric transfer from simulated gastric fluid (SGF, pH 1.0) to fasted state simulated intestinal fluid (FaSSIF, pH 6.5) occurred at 13 minutes (orange dotted line) for every sample.


Table 1. Predicted enhancement in pharmacokinetics of Compound X with amorphous dispersion formulations. Available observed values are in parentheses.


Figure 2. Dog Pharmacokinetics of Compound X. A) Predictions of dog PK for amorphous dispersions are in good alignment with observed in-vivo data. Solid lines are simulations from each model and the symbols are observed data (error bars show the measured variability in-vivo). B) Simulations from PBPK models for fasted beagle dogs with amorphous dispersions compared to the crystalline API. The insets in both figures show plasma concentrations of the drug in log scale.