(W1130-02-10) The Advantage and Disadvantage of In Vivo Predictive Dissolution Profiles of BCS Class II Drugs Generated by a Transfer Method and Compendial Method with Different Buffer Species

Purpose: The rising abundance of low-solubility APIs in the drug discovery pipeline necessitates the development of an in vivo predictive method that can be applied to them.^1,2 Traditionally, in vitro dissolution testing using the compendial dissolution apparatuses I and II has been the gold-standard for evaluating drug dissolution and assuring drug quality.³ However, these methods don’t accurately represent the complex physiology of the GI tract, making it difficult to predict in vivo behavior of low-solubility drugs. This is where animal testing comes in, however, it has the downside of being very costly in terms of time and resources. Standardizing an in vitro dissolution method for low-solubility APIs can minimize the use of animal models and could lead to faster assessment of oral formulations.

Methods: In this study, the in vivo predictive transfer method, gastrointestinal simulator alpha (GIS-α), was used to study the dissolution of commercially available BCS Class II drugs, danazol, fenofibrate, celecoxib, and ritonavir. The rate of transfer between the stomach, duodenum, and jejunum were set to follow a first order-like kinetics with a gastric half-emptying time of 8 mins. Two sets of media were used: the simple media, SGF and SIF, and the biorelevant media, FaSSGF and FaSSIF. In both set-ups, the stomach initially contains 50 mL of SGF (or FaSSGF) and 250 mL of water corresponding to the standard of taking one glass of water with each dosage. The duodenum compartment contains 50 mL of SIF (or FaSSIF) and this is kept constant throughout the experiment. The jejunum simply collects the transfer coming from the duodenum. The paddle speed was set at 100 rpm in the stomach but was kept at 50 rpm in the duodenum and jejunum. The whole set-up was kept at 37° C by using a water bath. To initiate the experiment, the dosage form was dropped into the stomach compartment. A spiral sinker was used for capsule dosage forms to prevent them from floating. A 750 μL sample was taken from all three chambers at 0, 5, 10, 20, 30, 60, and 90 min. The samples were then spun at 16400 × g for 1 min.  Five hundred μL of the supernatant was taken and mixed with 500 μL of methanol to prevent precipitation. The samples were then analyzed using HPLC and the resulting dissolution profiles were then compared with results obtain from the compendial USP II apparatus. To assess in vivo predictability, in silico modeling was employed to predict in vivo plasma concentrations, and these were compared to existing clinical data for all four drugs. Table 1 below summarizes the physicochemical properties of each drug used in the simulation. The computational software, GastroPlus^TM 9.8.3 (SimulationPlus, Inc., Lancaster, CA), was ran using a Lenovo ThinkPad computer with Intel Core i5 processors. With this software, one can input in vitro dissolution profiles alongside known physicochemical and pharmacokinetic (PK) parameters obtained in the literature (Table 1). A standard physiological condition model within the software was used: Human Physiological-Fasted alongside the Opt LogD Model SA/V 6.1.

Results: The dissolution profiles are shown in Figure 1. Comparison of results show that, in all cases, the GIS-α with the biorelevant buffers gave superior dissolution profiles, and in silico modeling yielded better prediction when utilizing the results from the GIS-α as input compared to the dissolution profiles obtained from the USP II apparatus. The in vivo predictions shown in Figure 2 are still far from what was observed clinically, suggesting that there is still room for improvement for this in vivo predictive method especially when applied to low-soluble drugs. Nonetheless, results show that the GIS-α an edge over USP II in generating drug dissolution profiles of BCS class II drugs. This is especially useful in the early stages of drug and formulation development.

Conclusion: Dissolution testing is standard in the pharmaceutical industry as it can inform decisions and aid in formulation testing. Based on the results of this study, the more physiological conditions included in the model, the better the in vivo prediction will be. The GIS-α with the biorelevant buffers produced the best results with the simple buffers coming in second followed by the USP II. However, the results are not close to clinical outcome yet for those model drugs. This indicates further improvement is needed in this in vivo predictive dissolution method for low-solubility drugs.

References: 1 Emami, S., Siahi-Shadbad, M., Adibkia, K. & Barzegar-Jalali, M. Recent advances in improving oral drug bioavailability by cocrystals. BioImpacts 8, 305-320, doi:10.15171/bi.2018.33 (2018).
2 Bhalani, D. V., Nutan, B., Kumar, A. & Singh Chandel, A. K. Bioavailability Enhancement Techniques for Poorly Aqueous Soluble Drugs and Therapeutics. Biomedicines 10, 2055, doi:10.3390/biomedicines10092055 (2022).
3 Matsui, K., Tsume, Y., Takeuchi, S., Searls, A. & Amidon, G. L. Utilization of Gastrointestinal Simulator, an in Vivo Predictive Dissolution Methodology, Coupled with Computational Approach To Forecast Oral Absorption of Dipyridamole. Mol Pharm 14, 1181-1189, doi:10.1021/acs.molpharmaceut.6b01063 (2017).
4 Sunesen, V. H., Vedelsdal, R., Kristensen, H. G., Christrup, L. & Mullertz, A. Effect of liquid volume and food intake on the absolute bioavailability of danazol, a poorly soluble drug. Eur J Pharm Sci 24, 297-303, doi:10.1016/j.ejps.2004.11.005 (2005).
5 Zhu, T., Ansquer, J. C., Kelly, M. T., Sleep, D. J. & Pradhan, R. S. Comparison of the gastrointestinal absorption and bioavailability of fenofibrate and fenofibric acid in humans. J Clin Pharmacol 50, 914-921, doi:10.1177/0091270009354995 (2010).
6 Patel, N. P., Sanyal, M., Shrivastav, P. S. & Patel, B. N. Estimation of Celecoxib in Human Plasma by Rapid and Selective LC-MS/MS Method for Bioequivalence Study. Int J Pharm Pharm Sci 10, 16-22 (2018).
7 Arora, S. et al. Biopharmaceutic In Vitro In Vivo Extrapolation (IVIV_E) Informed Physiologically-Based Pharmacokinetic Model of Ritonavir Norvir Tablet Absorption in Humans Under Fasted and Fed State Conditions. Mol Pharm 17, 2329-2344, doi:10.1021/acs.molpharmaceut.0c00043 (2020).

Acknowledgements: This work was made possible through the MRL post-doctoral program of Merck & Co., Inc., Rahway, NJ, USA. The authors declare no conflict of interest.

Table 1. Physicochemical and pharmacological properties of each drug for in silico modeling.

Figure 1. Comparison of dissolution methods used in this study.

Figure 2. Predicted plasma concentration-time profile for all drugs tested. Clinical data for danazol (100 mg), fenofibrate (145 mg), celecoxib (400 mg), and ritonavir (100 mg) are represented as open circles and included for visual comparison.