Quantitative understandings of human in vitro drug response relationships to in vivo model systems

Abstract

Drug development is inefficient and expensive, with less than 10% of drug candidates entering clinical trials proceeding to market, and each costing an estimated $2.5 billion. This inefficiency is due to several reasons, including weak correlations between in vitro model systems and in vivo action. Traditionally in vitro and murine models for toxicity and efficacy of candidate drugs have not provided sufficient accuracy in prediction of human response. This is clearly demonstrated by the recurrence of safety failures in clinical trials, including post-Phase I safety trials. This weakness in predictivity may be partly addressed through the advent of three-dimensional (3D) cell culture, which can be performed at the microscale and in high throughput, allowing for efficient use of more biorelevant, potentially rare (e.g., personalized) cells.

This thesis work is focused on improving the inefficiencies of the drug development process by developing in vitro cell culture systems that better mimic the in vivo environment and may better predict human drug/drug candidate toxicity and exploiting multi-cell culture for antimicrobial action. A 532-micropillar chip system was used to culture primary human hepatocytes and screen a library of drugs for hepatotoxicity. This platform was found to be highly predictive of in vivo (mouse and rat) toxicity, with sensitivity and specificity values approaching 80%. In a similar vein and focused on antimicrobial screening, the effectiveness of a 384-pillar plate system was used to culture mesenchymal stem cells (MSCs), which have the ability to kill bacteria. MSCs killed approx. two-thirds of Escherichia coli cells using a unique and simultaneous mammalian-bacterial cell culture on a microscale platform. This approach may represent a promising avenue for gaining a fundamental understanding of the action of the innate immune system against bacterial pathogens, as well as identifying new routes to antimicrobial therapy. Overall, these results represent progress towards in vitro cell culture systems which are able to consistently predict in vivo action.