(M1530-10-55) Using Dissolution-Permeation Systems and In Silico PBPK Modeling for Food Effect Prediction of a BCS Class II Drug

Purpose: Assessing food effect (FE) on oral drug absorption is critical to optimize the safety and efficacy of the final product. However, the complex interplay between physicochemical compound characteristics, formulation properties and physiological changes along the gastrointestinal tract induced by food intake make it challenging to predict FE on oral drug absorption. Several approaches have been used to predict FE (ref 1, 2) including in-vitro tools and physiologically-based pharmacokinetic (PBPK) modeling, but a single approach that would give fully reliable predictions of FE has not yet been determined. In this study, we used in-vitro dissolution-permeation assay combined with in-silico PBPK modeling for FE prediction.

Methods: A weak base BCS class II drug with solubility of 12 and 56 µg/mL in fasted and fed state simulated intestinal fluid (FaSSIF, FeSSIF) respectively, was used for this study. In-vitro dissolution-permeation studies were carried out using a µFLUX or BioFLUX apparatus (Pion Inc., Billerica MA, USA) which consists of donor and receiver chambers separated by a biomimetic membrane. Measurements were performed using FaSSIF or FeSSIF in the dissolution vessel and Acceptor Sink Buffer (ASB pH 7.4) in the receiver chamber. A small amount of unformulated drug powder was added in the donor chamber of the µFLUX. and film coated tablets (dose strength of 10, 20, 40, 80 mg) were added into the donor compartments of the BioFLUX. Concentration was monitored in both chambers using in-situ monitoring technique implemented with Rainbow FO UV-Vis spectrometer (Rainbow R2D8, Pion Inc). In-silico modeling was performed using GastroPlus™ software (v.9.8, Simulations Plus Inc. Lancaster, US). Drug physicochemical properties were determined in-vitro and used as input parameters for simulation. Fasted and fed state human physiologies were tested. Disposition parameters were determined based on in-vivo human plasma concentration-time profiles after peroral solution administration using PKPlus software. Drug permeability was calculated using in-vitro flux and used as input parameter for simulations. The tested compound is a substrate and inhibitor of P-gp transporter. The model was validated by comparing the predictions with clinical plasma concentration-time profiles (ref 3).

Results: Dissolution-permeation profiles of unformulated drug from µFLUX were higher in FeSSIF with a 2.5-fold increase in dissolution and 1.8-fold increase in flux compared to FaSSIF, suggesting a positive food effect. Dissolution profiles of drug from film coated tablets tested in BioFLUX showed a higher percent of drug released in fed state compared to fasted state at all dose levels tested, where lower percent of drug was released for higher doses in both prandial states. The same trend was predicted by PBPK modeling. Donor drug concentration at plateau (drug released at 2 hours) under fed condition showed a steeper increase with dose compared to drug concentration in the fasted state condition. Similar trend was observed for maximal plasma concentration (C_max) in the clinic. Flux, in the fasted and fed states, increase as a function of dose with a larger dose dependence seen in the fed state. Based on flux data, positive FE was successfully predicted for all doses (fed/fasted flux ratio was 1.6-2.0, depending on a dose). In-silico simulations also predicted positive FE, which was stronger for higher doses. This was also observed in-vivo, where AUC ratio (fed/fasted) was 1.96 (20 mg dose), 2.89 (40 mg dose) and 3.84 (80 mg dose).  The stronger FE seen for higher doses, is most likely due to the solubility-limited absorption of the tested drug. In addition, the strongest FE observed at the highest dose can be explained with inhibition of P-gp at high drug concentrations, achieved after 80 mg tablet administration in fed state.

Conclusion: Based on the above, this study shows that: Small-scale dissolution-permeation system of the unformulated drug was predictive of FE. In-vitro dissolution-permeation setup could be used as a screening tool to assess the impact of food on oral drug absorption under biorelevant conditions when passive permeability is the major pathway for oral absorption. Effective permeability value calculated based on in-vitro flux data was successfully used as input parameter for establishing the PBPK model. The developed PBPK model predicted adequately the direction and extent of FE. Overall, the flux-based approach in combination with PBPK modeling, could be a helpful tool in formulation development for prediction of the extent of food effect at various doses.

References: Ref 1: Jankovsky et al, ADMET&DMPK (2022) 10(4)
Ref 2: Zhang et al, Curr Pharmacol Rep (2020) 6:267–279
Ref 3: Luo et al, AAPS PharmSciTech (2022) 23:185