Engineered microtissue platforms for modeling human pathophysiology and drug metabolism

Abstract

Over 50% of all drug candidates entering clinical trials are abandoned due to insufficient efficacy or unexpected safety issues despite extensive pre-clinical testing. Liver metabolites that cause toxicity or other side effects cannot always be predicted in animals, in part because of human-specific drug metabolism. Furthermore, while the clinical need for cancer drugs is increasing, anti-tumor activity in animals often leads to a disappointing lack of efficacy in real patients. In vitro models that can better predict human responses to drugs would mitigate the overall costs of development and help bring new therapies to market. In order to improve the predictive power of in vitro tissue models, various features of the microenvironment that modulate cell behavior have been investigated, such as cell-cell interactions, cell-matrix interactions, soluble signals, 3-dimensional (3D) architecture, and mechanical stiffness. Synthetic hydrogels offer a versatile platform within which these cues can be precisely perturbed in a 3D context; however, the throughput of these methods is quite limited. In this thesis, we explore the potential of high-throughput manufacturing and monitoring of populations of miniaturized 3D tissues, termed 'microtissues,' for modeling healthy and diseased tissues in both static and perfused systems. First, we developed a flow-based platform to test tumor proliferation in defined microenvironmental settings with large numbers of replicates (n > 1000). A microfluidic droplet generator was designed to encapsulate tumor cells with stromal cells and extracellular matrix in 100 pm-diameter poly(ethylene glycol) (PEG) microtissues (6000 microtissues/min). Upon screening a small panel of soluble stimuli, TGF-p and the TGF-pR1/2 inhibitor LY2157299 were found to have opposing effects on the proliferation of lung adenocarcinoma cells in microtissues vs. in 2-dimensional culture, affirming a potential role for 3D models in the investigation of cancer therapies. Next, we extend these techniques to the analysis of drug-induced liver injury. Phenotypic maintenance of primary hepatocytes was achieved by controlled pre-aggregation (-50 tm units) with J2-3T3 fibroblasts to establish cell-cell contacts prior to encapsulation into microtissues. Retention of both constitutive and inducible Phase I drug metabolism activity allowed detection of prototypical hepatotoxins through generation of toxic metabolites and emergence of drug-drug interactions, thereby demonstrating the suitability of hepatic microtissues for 3D, high-throughput toxicity screening. Finally, we describe efforts to bridge the gap between multi-organ models and human drug metabolism. Modular human hepatocyte microtissues were entrapped by semi-circular microsieves in a microfluidic perfusion chamber for over 3 weeks. In contrast to immortalized hepatic cell lines, primary hepatocytes stabilized in microtissues exhibited human-specific induction profiles, reflected donor hetereogeneity in CYP2D6 and CYP2C19 enzyme activity levels, and performed xenobiotic detoxification on circulating drugs, establishing the ability to incorporate hepatic functions in 'human-on-a-chip' devices. Collectively, these three applications of cell-laden microtissues demonstrate their versatility and potential impact in both drug development and fundamental studies of the cellular microenvironment.