(229d) Can We Predict How Humans Will Respond to Drugs?

Shuler, M. L. - Presenter, Cornell University
Tatosian, D. A. - Presenter, Cornell University
Sung, J. H. - Presenter, Cornell University

Modern biology has provided a more mechanistic basis to better understand how cells function. While pharmacologists can often provide a potential mechanistic basis to understand how a drug functions, they cannot usually relate that understanding to a quantitative prediction of what a safe and effective dose might be. Generally it has been difficult to relate molecular mechanisms to cell response to organ/tissue response to physiological system response. Our ultimate goal is to build mathematical models of molecular mechanisms, embedded in cellular models, embedded in an organ or tissue model, embedded in a system or physiologically based pharmacokinetic model (PBPK). This quantitative, multiscale approach uses a specific mechanistic model embedded in increasingly coarse grain models. Ideally the overall system model could predict, for example, that if a certain dose of drug was administered orally that the time dependent concentration at the organ/cells of interest could be estimated and the specific response of that organ/cell target could be predicted.

Due to the complexity of the human body and our lack of knowledge of all of the interacting mechanisms, such a model by itself would not be credible. Animal studies are expensive, difficult to control, take long times, and are not very predictive of human response. Human clinical studies are terribly expensive and humans are poor model systems. Thus, a physical system to test the credibility of such a mathematical model would be useful. Our proposed approach is to build a cell culture analog in which each tissue or organ compartment in a PBPK is represented by a bioreactor with a cell type or several cell types representative of that organ or tissue. These compartments are interconnected with a surrogate blood flow that mimics the circulatory pattern in the PBPK and the body. Each compartment acts effectively as a reactor, absorber, or surge tank. Since the number of cells, the level of key enzymes per cell, the flow rate, and residence time are known, the comparison between the cell culture analog and a PBPK model should be straightforward. If the proposed mechanism and corresponding model predict the observed behavior, it suggests that the mechanism and model are plausible. If not, then alternative mechanisms can be evaluated and in an iterative manner tested with the in vitro experimental system.

We have implemented the cell culture analog concept using a microfabricated system with compartments in which living cells were placed. This has been termed a ?Body-on-a-Chip?. We have used this system to probe response of cancer to treatment with a combination of drugs. For multidrug resistant (MDR) uterine cancer the system shows a synergistic response using the chemotherapeutic agent, doxorubicin, in combination with two MDR suppressors (cyclosporine and â-nicardipine). For example, when the system is treated with 1 µM doxorubicin the MDR cancer cells grow 3.5 fold over 72 hr. When either 1 µM doxorubicin with 10 µM cyclosporine or 1 µM doxorubicin with 10 µM nicardipine are used, the MDR cancer cells increase 1.6 fold. However, with 1 µM doxorubicin and 5 µM cyclosporine and 5 µM nicardipine the number of MDR cells decrease to 0.8 of the original number. Note that in the combination treatment the total amount of MDR suppressors remains the same as when a single MDR suppressor was used demonstrating a synergistic response. These MDR suppressors have no adverse effects on cells representing liver and bone marrow.

We have also constructed PBPK-PD models and the corresponding micro CCA's to test a combination treatment with a Tegafur (prodrug for 5-flurouracil) and uracil to predict the range of responses that may be observed within the human population.

This approach is a beginning step to predicting human response, including individual human response, to a variety of drug treatments including those with combinations of multiple drugs.