Artificial Gut Simulator for Simultaneous Evaluation of Drug Dissolution and Absorption

Abstract

Over the past three decades, high-throughput screening has resulted in a discovery pipeline consisting mostly of highly potent but lipophilic compounds exhibiting poor aqueous solubility and classified as Biopharmaceutics Classification System (BCS) Class II. Since poor solubility limits absorption and bioavailability, efforts have been made to develop supersaturating delivery systems such as amorphous solid dispersions (ASD) that enhance the apparent solubility of the drug without sacrificing its thermodynamic activity. The performance of these dispersions is often tested in ‘closed’, non-sink compendial dissolution testing apparatus that lacks an absorptive sink. The supersaturated solution generated upon ASD dissolution is metastable with respect to the stable crystalline phase and can undergo amorphous and/or crystalline precipitation. The rate of precipitation depends upon the degree of supersaturation. In the absence of absorption, during non-sink dissolution testing, high supersaturation can drive more precipitation than that which occurs in vivo where continuous drug absorption from the intestinal lumen decreases drug concentration, which in turn decreases the driving force for precipitation. Unsurprisingly, many in vitro studies with non-sink dissolution testing have failed to predict the in vivo performance of formulations of BCS-II drugs, which by definition, have high intestinal permeability. A simultaneous dissolution and absorption testing apparatus called the side-by-side diffusion cell allows drug to diffuse from the donor to a receiver compartment across a membrane that separates the two. However, small surface area of the membrane results in very low rates of drug absorption and very long, unphysiological experimental time scales. The first goal of this study was to develop and validate an artificial gut simulator apparatus (AGS) consisting of a hollow fiber-based absorption module suspended in the drug donor. The hollow fibers provide a large surface area for absorption, significantly improving mass transfer rate of drugs from the donor into the aqueous receiver media in the hollow fiber lumen. Continuous pumping of the drug-free receiver media into the lumen helps maintain an absorptive sink. A theory for mass transfer across the hollow fiber membrane was developed and validated using caffeine. Physiological rate of drug absorption was attained by tuning the AGS operating parameters per the theoretical model. This is an important step in developing a biorelevant test for BCS-II drugs. The next goal of this project was to understand how absorption impacts dissolution of ASDs and subsequent crystallization from supersaturated solutions of a model BCS-II compound, ketoconazole. Relative to a non-sink ‘control’, continuous drug removal by absorption enhanced ASD dissolution and significantly decreased both amorphous and crystalline precipitation. This can be attributed to both a decreased driving force for precipitation due to lower drug concentration in the AGS donor as well as to redissolution of any precipitate that is formed to replenish the drug in solution lost to absorption. On the other hand, polymer excipient added to the ASD to stabilize the drug against crystallization during storage and dissolution reduced the drug’s absorption rate by possibly interacting favorably with the free drug species and reducing the drug’s thermodynamic activity. Simple analytical techniques used in conjunction with the AGS helped decouple and understand the impact of dissolution, precipitation and speciation on absorption and vice-a-versa. The final goal of this project was to implement a scheme to establish in vitro/in vivo correlation with another BCS-II drug, dipyridamole, by inputting the drug concentration absorbed by the AGS receiver media into a compartment-based disposition model to ultimately predict the in vivo plasma concentration-time profile of the drug. The human intestinal absorption rate constant of dipyridamole, determined from Caco-2 cell monolayer permeability coefficient, was used to tune the AGS. Gastric emptying was simulated at a physiological rate to ensure a physiological rate of supersaturation generation as the weakly basic dipyridamole is solubilized and emptied from acidic gastric compartment into a neutral duodenum. This methodology of simulating gastric emptying and absorption enabled accurate prediction of drug in vivo intestinal and plasma concentration-time profiles. This approach and apparatus is anticipated to be of great utility during drug product development for screening and optimization of potential oral formulations.