MULTI-ORGAN MICROPHYSIOLOGICAL SYSTEMS FOR ESTIMATING ABSORPTION, DISTRIBUTION, METABOLISM, AND EXCRETION (AMDE) OF DRUGS AND CANCER METASTASIS

Abstract

Predictive in vitro systems are essential for drug development and translational medicine. Current microphysiological systems (MPS) increasingly incorporate physiologically relevant features, such as tissue interfaces, controlled flow, and multi-organ connectivity, to better recreate human physiology. Moreover, these systems must quantify critical processes including absorption, transport, metabolism, and cellular interactions under conditions that mirror in vivo environments. However, existing MPS often require complex pumping systems, large fluid volumes, and fail to adequately bridge the gap between preclinical tests and human outcomes, limiting their predictive capacity and clinical translation potential. This dissertation addresses these critical gaps by developing novel pumpless, low liquid-to-cell ratio MPS that recreate physiologically relevant flow conditions and tissue interfaces to quantify absorption, transport, metabolism, and cancer cell invasion.First, I developed a pumpless recirculating endothelial chip with a total working volume of 400 μL, which generated unidirectional flow with an average wall shear stress of 0.588 ± 0.006 Pa (~5.9 dyn/cm²) without the need for external pumps. Under unidirectional shear, HUVECs aligned with the flow, exhibited continuous VE-cadherin borders, and secreted lower levels of IL-6 and IL-8 than under bidirectional shear, demonstrating improved endothelial phenotype control. The developed device can be integrated with other pumpless tissue-on-chip systems, allowing for the incorporation of additional barrier tissues, such as endothelial linings. Second, I designed a pumpless tumor-endothelium device to investigate HT-29 colorectal cancer cell invasion through HUVEC barriers in the presence of 50 nm carboxylate nanoparticles. Using an 8 μm PC membrane for transmigration and 20 ng/mL EGF gradient for chemotaxis, moderate nanoparticle doses (3.64 × 10¹¹-3.64 × 10¹² particles/mL, 0.025-0.25 mg/mL) preserved barrier integrity, maintained low 70 kDa dextran permeability, and trapped HT-29 cells at the basal endothelium side. High doses (3.64 × 10¹² particles/mL, 2.5 mg/mL) proved cytotoxic, eliminating both HUVECs and HT-29 on the device. Third, I engineered a pumpless GI-liver MPS to model first-pass metabolism effects. The GI chamber (~74 μL) generated physiological shear (~0.0045 dyn/cm²), while the liver chamber (~128 μL) featured parylene C coating to prevent drug loss. Primary human small intestinal epithelial cells achieved TEER 46.8 ± 4.4 Ω·cm² with optimized medium scheduling. Consequently, measured permeabilities exceeded static assays and approached ex vivo values: verapamil Papp 3.77×10⁻⁵ cm/s in MPS versus 1.24×10⁻⁵ cm/s in transwell; propranolol 3.08×10⁻⁵ cm/s versus 1.14×10⁻⁵ cm/s. Although primary human hepatocyte clearance in vitro remained lower than in vivo, the flowing MPS outperformed static cultures. Together, these pumpless unidirectional flow MPS control endothelial phenotype with precise shear, reveal how nanoparticles and EGF jointly modulate transendothelial invasion under flow, and capture first-pass processes with improved permeability and analyte recovery in sub-milliliter volumes. Overall, this dissertation advances more predictive, human-relevant experimental systems for drug development and translational medicine.