Microfluidic culture of human hepatocytes and endothelial cells with applications in drug toxicity screening

Abstract

Drug-induced liver injury (DILI) continues to be a major problem for patient health and pharmaceutical expenditures, partially due to inadequacies of current model systems for predicting hepatotoxicity prior to clinical trials. In the drug development pipeline, many platforms are implemented depending on the stage of development, the number of compounds in question, and the specific hypothesis being studied. Primary human hepatocytes (PHHs) are considered the 'gold standard' for in vitro screening, as they retain a full complement of drug metabolizing enzymes and transporters. However, PHHs are in limited supply and lack the genetic diversity representative of the human population. In this dissertation, we explore alternative cell sources to PHHs such as iPSC-derived hepatocytes, mouse hepatocytes, and the hepatocarcinoma cell line HepaRG in an engineered liver platform. We found that each of these cell types showed a high level of hepatic functions when incorporated into a micropatterned co-culture (MPCC) of the hepatocyte type in question with 3T3-J2 murine embryonic fibroblasts. MPCCs of PHHs and 3T3-J2 fibroblasts were then challenged with known hepatotoxins and their non-toxic structural drug analogs before undergoing global gene expression analysis. These analyses revealed that hepatotoxins caused a differential expression of significantly more genes than the non-toxic analogs, and the corresponding pathways could reveal underlying mechanisms of drug toxicity. Next, these in vitro models were supplemented with endothelial cells to give a more complete representation of liver physiology. We found that co-cultures of hepatocytes and endothelial cells rapidly lost functionality, but tri-cultures of hepatocytes, endothelial cells, and 3T3-J2 fibroblasts were stable for multiple weeks. However, endothelia in the body experience shear stress from fluid flow, a phenomenon not mimicked with traditional in vitro cultures. Thus, we developed an in vitro platform for perfusing cultures with a physiologic level of shear stress. This system, constructed of tissue culture polystyrene with polydimethylsiloxane, was modeled using computational software and compared alongside static controls. Ultimately, we believe these platforms can be incorporated as the liver compartment into a "body-on-a-chip" platform used to understand multi-organ effects of drugs and diseases that impact the liver including diabetes, hepatitis B/C, and malaria.