The Design and Fabrication of Three-Chamber Microscale Cell Culture Analog Devices with Integrated Dissolved Oxygen Sensors Aaron Sin, † Katherine C. Chin, † Muhammad F. Jamil, † Yordan Kostov, ‡ Govind Rao, ‡ and Michael L. Shuler* ,† School of Chemical and Biomolecular Engineering, Cornell University, 120 Olin Hall, Ithaca, New York 14850, and Department of Chemical Biochemical Engineering, University of Maryland Baltimore County, Baltimore, Maryland 21250 Whole animal testing is an essential part in evaluating the toxicological and pharmacological profiles of chemicals and pharmaceuticals, but these experiments are expensive and cumbersome. A cell culture analog (CCA) system, when used in conjunction with a physiologically based pharmacokinetic (PBPK) model, provides an in vitro supplement to animal studies and the possibility of a human surrogate for predicting human response in clinical trials. A PBPK model mathematically simulates animal metabolism by modeling the absorption, distribution, metabolism, and elimina- tion kinetics of a chemical in interconnected tissue compartments. A CCA uses mammalian cells cultured in interconnected chambers to physically represent the corresponding PBPK. These compartments are connected by recirculating tissue culture medium that acts as a blood surrogate. The purpose of this article is to describe the design and basic operation of the microscale manifestation of such a system. Microscale CCAs offer the potential for inexpensive, relatively high throughput evaluation of chemicals while minimizing demand for reagents and cells. Using microfabrication technology, a three-chamber (“lung”-“liver”-“other”) microscale cell culture analog (μCCA) device was fabricated on a 1 in. (2.54 cm) square silicon chip. With a design flow rate of 1.76 μL/min, this μCCA device achieves approximate physiological liquid- to-cell ratio and hydrodynamic shear stress while replicating the liquid residence time parameters in the PBPK model. A dissolved oxygen sensor based on collision quenching of a fluorescent ruthenium complex by oxygen molecules was integrated into the system, demonstrating the potential to integrate real-time sensors into such devices. 1. Introduction Toxicological and pharmacological testing are crucial in the chemical and pharmaceutical industries. Potential pharmaceuticals have to be screened for efficacy as well as possible toxicity. In chemical industries, toxicity profiles are important in determining the safe exposure level and first aid mechanisms for household and com- mercial chemicals. Whole animals are commonly used to determine these toxicological and pharmacological pro- files, but these experiments are generally expensive and lengthy to perform. There is also considerable doubt whether results from animal tests can be extended reliably to human beings. Alternatives to animal studies include in vitro cell cultures and computer models. To help solve this problem, mammalian cell cultures (which include human cells) have been used to obtain mechanistic information for xenobiotic (foreign chemical) metabolism. Researchers have also combined this mecha- nistic information with physiological information such as blood flow and organ volumes to create physiologically based pharmacokinetic (PBPK) models (1, 2). A PBPK model mathematically simulates the absorption, distri- bution, metabolism, and elimination (ADME) processes of living systems, providing a method to link mechanistic data obtained in in vitro cell cultures to system-wide toxicological and pharmacological information. However, a realistic PBPK model often requires parameters, par- ticularly those associated with the kinetics of metabolism, that are difficult to estimate. Further, a prime limitation on a PBPK model is that all relevant mechanisms, whether direct or indirect, must be anticipated and included in the model. Often, secondary effects are not explicitly included. A cell culture analog (CCA) system is a physical replica of the PBPK model (3). Mammalian cells are cultured in different compartments to represent organs, which in turn are interconnected by circulating cell culture me- dium that acts as a blood surrogate. Design parameters such as compartment residence times and flow distibu- tion are based on the corresponding PBPK model. Using representative cell types in the compartments, one can obtain system-wide rate parameters that can then be used to refine the PBPK model. In addition, because mammalian cells from different species of origin can be applied to the CCA, this system provides a potential means of studying cross-species extrapolation of toxico- logical and pharmacological profiles. The accuracy of the PBPK parameters obtained from the CCA depends very much on the authenticity of tissue * To whom correspondence should be addressed. E-mail: mls50@ cornell.edu. † Cornell University. ‡ University of Maryland Baltimore County. 338 Biotechnol. Prog. 2004, 20, 338-345 10.1021/bp034077d CCC: $27.50 © 2004 American Chemical Society and American Institute of Chemical Engineers Published on Web 11/05/2003