Restriction lift date: 2030-12-31
Advancing in vitro and computational pharmaceutics tools for guiding lipid-based formulation design
Loading...
Files
Date
2025
Authors
Ejskjær, Lotte
Journal Title
Journal ISSN
Volume Title
Publisher
University College Cork
Published Version
Abstract
Purpose: Computational chemistry has become a well-established method in drug discovery, guiding the selection of lead drug candidates with optimal binding affinity and uncovering a broader chemical design space than classical synthetic chemistry approaches. However, this focus on high binding affinity often results in lead candidates with poor water solubility. As a result, increasing number of these lead candidates are emerging from drug discovery that exhibit poor biopharmaceutical properties, necessitating bio-enabling formulation strategies for the development of effective oral drug products. Lipid-based formulation (LBF) is a bio-enabling formulation strategy which can improve the bioavailability of certain poorly water-soluble drugs. However, the formulation design of LBFs is challenging as there are multiple solubilities to consider to ensure a robust and effective formulation: drug lipid solubility in the lipid vehicle, drug solubility in the aqueous colloidal dispersion that forms on dispersion of the LBF in gastrointestinal fluids, and subsequently drug solubility within the post digestive colloidal phase after the LBF has undergone digestion. In early-stage formulation screening and development, there is a growing need to create rapid high-throughput models for screening formulation excipients and assessing solubilities under these conditions. This approach helps streamline the formulation development process and identify drug candidates for a LBF approach. A significant limitation of many existing in vitro assessment tools for LBF is their low throughput, making them challenging to routinely use when lead drug candidates are in scarce supply. The overarching aim of this thesis was therefore to develop industrial scalable in vitro tools for assessment of LBF and exploring the use of computational pharmaceutics tools for guiding LBF design.
Method: A range of immobilized and liquid lipases were screened in the classical pH stat lipolysis assay to identify a lipase with comparable extent of digestion and digestion kinetic profile as the commonly used porcine pancreatic lipase. This was followed by a feasibility study using the UV in-line probes of the MicroDISS ProfilerTM to monitor drug concentration upon dispersion and digestion. Subsequently, a bigger data set was screened upon dispersion and digestion for utility of the method. This data set was used for statistical modelling, using partial least squares analysis and multiple linear regression to produce computational models for predicting solubility change upon dispersion and digestion of the LBF. Lastly, the digestion method using the MicroDISS ProfilerTM was further developed with monitoring of pH throughout the experiment and used for screening of precipitation inhibitors in LBF. The influence of the pH of the biorelevant media in lipolysis experiments was also investigated.
Results: Firstly, this thesis demonstrated that Palatase® 20000 L is a good alternative to the classical lipase, porcine pancreatin. Palatase® 20000L shows a comparable extent of digestion, digestion profile, and influence on drug concentration post-digestion of a model compound. Furthermore, Palatase® 20000L is compatible with in-line UV detection of drug concentration upon dispersion and digestion of LBFs. Successful utility of the method to screen a larger dataset of 30 drugs in two different LBF upon dispersion and digestion showed a tendency towards cationic drug increased solubility post-digestion, whereas anionic and neutral drugs decreased or had no influence of digestion. Predictive models were successfully developed to estimate the solubility change upon dispersion and digestion using a linear regression equation with good predictability (pre-digestion R2 = 0.70, and post-digestion R2 = 0.68). The use of precipitation inhibitors in LBFs were shown to have a moderate effect in a Type IV LBF, however only a minor effect or negative effect in a Type III LBF. In addition, using a pH of 6.5 or 7.5 in the in vitro lipolysis assay influences the outcome of the experiment in relation to extent of digestion, where a higher pH result in higher digestion.
Conclusion: This thesis introduces an innovative in vitro digestion method designed to enable real-time monitoring of dynamic drug concentrations during digestion of LBFs. This approach allows for the examination of the effects of digestion of LBFs in a higher throughput approach, eliminating artifacts and the labour-intensive sample preparation required by comparison to the classical pH-stat lipolysis methods. The thesis demonstrated the utility of the high-throughput in-line digestion method as a screening and characterization tool for LBF dispersion and digestion. This novel in vitro digestion method was used to establish a computational model for predicting the solubility ratio post-digestion of LBFs. To our knowledge, this study is the first to create a validated predictive model of drug concentration during LBF digestion, established using 30 poorly water-soluble model drugs. This computational pharmaceutics approach advances the application of data-driven models in lipid-based drug product development. The research conducted in this thesis, using a range of PIs suggest that in the majority of cases PIs offered limited capacity to maintain supersaturation during dispersion/digestion of LBFs.
Description
Keywords
Lipid-based formulation , Lipolysis , Computational modelling
Citation
Ejskjær, L. 2025. Advancing in vitro and computational pharmaceutics tools for guiding lipid-based formulation design. PhD Thesis, University College Cork.