Category: Preclinical Development
Purpose: The prediction of in vivo dissolution of weakly basic poorly soluble drugs represents a major task in today’s drug discovery and development, in particular when precipitation of the drug in the gastrointestinal tract is likely to occur. A large number of newly developed NCEs are weak bases. For them, reliable and predictive in vitro and in silico approaches are unquestionably required, in order to accurately capture the risk of pH-dependent precipitation, and finally enable fast and successful lead compound selection and subsequent development. Despite this clear need, up to date, reliable input parameters for PBPK simulations derived from sophisticated in vitro assays are rare. Therefore, we recently demonstrated an increased predictability of a transfer model which employs a physiologically relevant bicarbonate compared to a standard phosphate buffer for simulating intestinal conditions .
Methods: The in vitro transfer model was used to simulate drug transfer from the donor (stomach) into the acceptor (small intestine) compartment. Therefore, 80 mg pazopanib, as hydrochloride (HCl)-salt, (one tenth of the human dose of 800 mg) were suspended in SGF and subsequently transferred into FaSSIF-phosphate in a small-scale setup. A large-scale USP II-based transfer model (800 mg pazopanib), equipped with a pHysio-grad® device, was employed to perform transfer model experiments in FaSSIF-bicarbonate. The use of the pHysio-grad® allowed to integrate a dynamic and physiologically relevant bicarbonate buffer into a transfer model setup. GastroPlusTM simulations were performed using the advanced compartmental absorption and transit model. Next to compound specific properties, e.g. basic physicochemical, solubility, and permeability parameters etc., results from in vitro dissolution experiments of pazopanib-HCl tablets and from in vitro transfer model experiments in FaSSIF-phosphate and FaSSIF-bicarbonate were used for the simulations.
Results: Compared to the clinical data of 800 mg pazopanib in the fasted state , the bottom-up PBPK simulation yielded an underprediction of pazopanib plasma concentrations (Figure 1). The integration of the in vitro dissolution profile of Votrient® tablets, two tablets with each 400 mg pazopanib as HCl-salt, in FaSSIF yielded even lower plasma concentrations (Figure 1). Using the in vitro supersaturation, obtained from transfer model experiments in FaSSIF-bicarbonate , as representative FaSSIF solubility, yielded a high correlation between the predicted and clinically observed plasma concentrations (Figure 2). This positive correlation was only observed using in vitro supersaturation data from experiments with FaSSIF-bicarbonate. In case of using in vitro supersaturation from experiments employing FaSSIF-phosphate , plasma levels where again underpredicted (Figure 2). These observations can be explained by the higher solubility of pazopanib in a physiologically relevant bicarbonate buffer, which more accurately predicts pazopanib intestinal solubility and, consequently, the fraction absorbed.
Conclusion: In the present study, a bottom-up prediction or the integration of in vitro dissolution data of pazopanib-HCl tablets did not yield a positive correlation of predicted and observed plasma concentrations. In contrast, a positive prediction of in vivo plasma concentrations of pazopanib was observed, when integrating in vitro intestinal supersaturation derived from transfer model experiments applying FaSSIF-bicarbonate. However, the integration of in vitro supersaturation obtained from transfer model experiments using FaSSIF-phosphate, did not yield a comparable high correlation. The experimental data clearly highlights the importance of physiologically relevant dissolution and precipitation testing, i.e. using a physiologically relevant bicarbonate buffer, with the aim to generate reliable input parameters for PBPK simulations.
 Jede, C. et al., Improved Prediction of In Vivo Supersaturation and Precipitation of Poorly Soluble Weakly Basic Drugs Using a Biorelevant Bicarbonate Buffer in a Gastrointestinal Transfer Model, Mol Pharm, 2019, DOI:10.1021/acs.molpharmaceut.9b00534.
 GSK, Clinical Pharmacology Study Report, VEG10005, An Open-Label, Two-Period, Randomized, Crossover Study to Evaluate the effect of Food on the Pharmacokinetics of Single Doses of Pazopanib in Cancer Subjects, 2018.