(M1330-03-19) A New Solid Dispersion Modelling Framework for In Vitro and In Vivo Oral Absorption Anticipation Adapted to pH-Sensitive Matrices

Purpose: Solid dispersion formulations can often be a very effective way to improve dissolution and oral absorption of a low solubility API. As they rely first on stabilizing an amorphous form of the API inside the drug product, it is also important that precipitation of the crystalline form is retarded after oral administration. Many factors may impact reprecipitation of an API in a physiological medium beyond its apparent concentration, notably pH, bile, polymer concentration or surfactants, and eventually presence of unwanted small amounts of API crystalline “seeds”. This multiplicity of factors, when added to the mechanism of release from a pH sensitive polymer matrix renders translation of in vitro dissolution/precipitation screening experiments to the in vivo situation hazardous. The purpose of this research was to expand the modelling framework previously proposed for solid dispersion dissolution in vitro to the gastrointestinal tract and its dynamics, in order to (1) allow scientists to understand better the interplay of different physical mechanisms, (2) allow for reduced lab experiments to accelerate low solubility drug development and anticipate the in vivo situation.

Methods: We have used the SIEMENS FormulatedProducts software (gPROMS platform) to simulate in vitro experiments and oral administration of solid dispersions. The framework handles the following phenomena (shown in Figure 1 scheme):- Release of the API triggered by polymer carrier dissolution extent (pH dependent) and surface offered (particle size) and loading- Primary particles dissolution (polymer and API) modelled as solubility/diffusion limited including solid surface equilibria for all species (notably buffers and bile salts), with particle size treated as population balance- Acid-base equilibria in bulk including effect of buffer strength and effect of polymer acidic groups, as well as buffer nature (carbonate mass transfer being corrected for CO2 hydration)- Complex/micelle formation reactions (treated either as fast equilibrium occurring at solid surface or forward/backward reactions in the bulk only)- supersaturation-based independent nucleation and growth mechanisms, as well as dissolved polymer modulation and bile salt modulation (EC50-like empirical relationships). The framework had been previously tested on nifedipine solid dispersions produced from different grades of HPMCAS [1]. In this study we have worked on two very low solubility bases for which there is data showing a negative effect of bile salt on precipitation. first used data on solid dispersions of ketoconazole to challenge the framework for particle size and 2 steps gastric/intestinal buffer vs intestinal buffer only [2]. As a second modelling activity we have established a model for Posaconazole solid dispersion modelling in human starting from in vitro API solution precipitation experiments in presence or absence of bile and pre-dissolved HPMCAS (originating from various sources [3-5]), which were used for calibration of nucleation and growth mechanisms.

Results: Calibration and validation of the ketoconazole model predictive power for the effect of solid dispersion particle size is presented on Figure 2. Essentially, the ability of the solid dispersion to release ketoconazole in acidic pH despite the insoluble HPMCAS_LF matrix is empirically calibrated and nucleation of the transitory amorphous form and crystalline form are adjusted (increased). The effect of loading, particle size and media are generally well reproduced, although the effect of loading is inverted for the 187µm batch, and the “spikes” observed for the smallest particle size in phosphate buffer I the first 10min are not reproduced. The mean relative error in the 10min-120min range was 25%. Main results for posaconazole modelling of solution/precipitation are presented on Figure 3a (left) and a highlight of absorption modelling in fasted human is presented on figure 3b (right). By calibrating independently nucleation and growth rate constant as a function of supersaturation, the behaviour of the molecule in buffer was well captured. HPMCAS effect was described as high for LF grade and low/negligible for MF and HF. The effect of bile (from FaSSIF), also captured by the model is very significant, although it has been shown that in real human fluid, there is a clear reversion of this effect [3]. When transposed to the human situation, simulations could offer a virtual exploration of polymer grade, PSD, amount and method of addition (in the SD itself or added externally). Figure 3b shows the interplay of PSD and grade: it is predicted that best results for absorption should be obtained with a low granule size and lipophilic HF grade.

Conclusion: The modelling framework presented relies on the integration of relatively few but diverse solution experiments (potentially small scale) and should typically require calibration on very few solid dispersion manufactured to build the model. It can be effectively conveyed to the human setting, allowing extrapolation of environment pH and bile composition, polymer grades and SD size, and contributes to explain the drivers of a successful formulation in vivo.

References: [1] G. Louit, Y. Melkeri, X. Li, E. Close, W. Chen, M. Tobyn, I. Akseli “A New Modelling Framework for pH-Sensitive Solid Dispersions Release and Dissolution” AAPS 2022 poster
[2] M. Monschke, K. Kayser, K.G. Wagner, “Influence of Particle Size and Drug Load on Amorphous Solid Dispersions Containing pH-Dependent Soluble Polymers and the Weak Base Ketoconazole” AAPS PharmSciTech (2021) 22: 44
[3] Ahmed Elkhabaz, Dana E. Moseson, Sreya Sarkar, Joachim Brouwers, Garth J. Simpson, Patrick Augustijns, and Lynne S. Taylor “Crystallization Kinetics in Fasted-State Simulated and Aspirated Human Intestinal Fluids”, Cryst. Growth Des. 2021, 21, 2807−2820
[4] Tu Van Duong, Zhanglin Ni, and Lynne S. Taylor “Phase Behavior and Crystallization Kinetics of a Poorly Water- Soluble Weakly Basic Drug as a Function of Supersaturation and Media Composition”, Mol. Pharmaceutics 2022, 19, 1146−1159
[5] D.M. Mudie, A.M. Stewart, N. Biswas, T.J. Brodeur, K.B. Shepard, A. Smith, M.M. Morgen, J.M. Baumann, and D.T. Vodak, “Novel High-Drug-Loaded Amorphous Dispersion Tablets of Posaconazole; In Vivo and In Vitro Assessment” Molecular Pharmaceutics 2020 17 (12), 4463-4472

Figure 1 - Modelling framework

Figure 2 - Ketoconazole HPMCAS ASD in vitro model vs experimental data

Figure 3 - Modelling results for posaconazole