The potential for chemical injury to the liver is of great concern in the development of new therapeutic drugs. Currently, approximately 20 % of adverse drug reactions (ADRs) are unpredictable and idiosyncratic. In this poster, we present a microfluidic gradient chamber for in-vitro studies on hepatocyte metabolism and response to drug-induced toxicity. A key feature of the designed reactor is that it affords selective localization of micro-scale collagen gels, which serve as three-dimensional (3D) scaffolds for cell culture. The selective localization is accomplished by fabricating regions of high (convective) transport resistance as cell culture compartments. The culture compartment is flanked by columns of asymmetrically positioned posts. These columns form micro-fabricated “membranes” that separate the culture compartment from medium flow channels. The spacing between the posts affords diffusive exchange of essential nutrients, growth factors and drug molecules with the flow channels, but prevents convective transport. Leveraging the convection-free culture chamber design, stable diffusion gradients can be generated across the culture compartment by introducing parallel media streams with different chemical, i.e. drug, compositions into each flow channel.

To optimize the dimensions of the gradient chamber, a coupled convection-diffusion-kinetic model with oxygen as the limiting substrate was developed. The model was used to describe hepatocyte proliferation and metabolism in the gradient chamber. Kinetic parameters were obtained from the literature for human hepatocarcinoma (HepG2) cells. Simulation results predicted that the oxygen tension could be maintained above the effective Km for a culture compartment volume of up to 4.8x10^-4 cm^3 by distributing the medium inlets along the length of the reactor. The maximum attainable density was 10^5 cells.

Based on the simulation results, gradient chambers were fabricated using standard soft-lithography techniques and PDMS molding. To maintain consistency with the mathematical model, the HepG2 cell line was used as the model cell system. Cultures remained viable and exhibited hepatocyte metabolic function for at least 10 days. On day 7 after seeding, cultures were treated with the test drug troglitozone (TGZ), an insulin sensitizing, anti-diabetic compound first approved for clinical use, but eventually withdrawn due to indications of idiosyncratic ADR. A linear gradient of the drug was applied. Viability and metabolic assessments using fluorescent probes estimated that TGZ causes significant toxicity above 100 uM, in good agreement with conventional well-plate assay results. Taken together, our results to date indicate that the gradient chamber is an efficient platform for in vitro drug toxicity studies.