Human Vascular Microphysiological System for in vitro Drug Screening.

Currently, over 80% of proposed pharmaceutical drug candidates that enter clinical trials fail due to concerns with human efficacy and toxicity1. While pre-clinical animal studies provide great value, animal responses to drugs may exhibit differences in toxic doses and drug metabolism2. Microphysiological systems (MPS) are perfused small-scale models of one or more human tissues or organs3 comprised of human primary cells or induced pluripotent stem cells (iPSCs) with the ultimate potential of becoming models to study disease or tools for precision medicine. In order to accurately model disease and predict drug responses on an organ scale, three-dimensional (3D) human tissue models are critical. Many pre-clinical studies are conducted on two-dimensional (2D) plastic or glass substrates; however, in vivo, tissues are comprised of complex extracellular matrices embedded with mixed populations of cells and perfused with fluids. In addition to facilitating more accurate disease models through interaction with immunological stimuli, human MPS may serve as a bridge in the drug development pipeline between 2D cell culture studies and in vivo animal studies. Assessment of toxicity within the vasculature is of particular importance, since drug-induced vascular injury (DIVI), which typically manifests in preclinical animal studies through inflammation and changes in vascular tone, precludes many drug candidates from continuing along the pipeline despite uncertain characterization of human DIVI response4. In humans, cancer chemotherapeutics cause vascular damage affecting flow-mediated vasodilation5,6. Three-dimensional (3D) tissue models have the potential to allow us to evaluate human biological interactions and diseases by taking advantage of natural spatiotemporal cues, physiological fluid perfusion, a variety of cell types, and the complex extracellular matrix that are present in tissues but are absent from 2D culture plates7. A human tissue-engineered blood vessel (TEBV) capable of responding to vasoreactive stimuli would pose a promising model for the evaluation and screening of pharmaceutical drug candidates for toxicity and efficacy within the circulatory system. An ideal TEBV for MPS applications would be comprised of human cells in a biological or biodegradable synthetic matrix, have a small inner diameter to reduce fluid volumes, exhibit enough mechanical strength to withstand physiological stresses, and be produced rapidly to facilitate efficient drug screening. The medial wall cells should exhibit a smooth muscle phenotype, be quiescent and be able to contract and relax in response to agonists or inhibitors. Most importantly, the TEBV must be endothelialized to enable physiologically relevant dilation and constriction in response to stimuli. TEBVs have been constructed using three general approaches: natural or biodegradable synthetic matrices populated in vitro with cells, self-assembled cell sheets, or in vivo repopulation of decellularized natural or synthetic vessel matrices8,9. Despite their biomimetic properties, the sizes and long culture times for fabrication of TEBVs by many of the current approaches creates challenges in applying these procedures to in vitro drug testing10. While TEBVs constructed from natural matrix components such as collagen11,12 and fibrin13,14 have traditionally exhibited poor mechanical strength, plastic compression of collagen gels embedded with smooth muscle cells increases the collagen fiber density and yields rapidly-producible tubular structures with high mechanical strength15. A functional TEBV requires a confluent endothelial layer. The endothelium plays a major role in regulating leukocyte and platelet adhesion, permeability, and vascular tone, as well as modulating vasodilation through release of nitric oxide in response to changes in flow or stimuli with vasoreactive compounds such as acetylcholine16. Endothelial coverage of TEBVs prior to implantation in animal models has been assessed with characteristic endothelial markers such as von Willebrand factor (vWF)17,18. Under static in vitro conditions, endothelium seeded on TEBVs made from cultured SMC and fibroblast sheets demonstrated nearly confluent coverage, expressed vWF and inhibited platelet adhesion10. After exposure to physiological shear stresses for 24–48 hours, endothelial progenitor cells (EPCs) remain adherent to fibrin scaffolds embedded with either human neonatal dermal fibroblasts (hNDFs) or smooth muscle cells (SMCs), deposit collagen IV and laminin, and upregulate cell adhesion molecules VCAM-1 and ICAM-1 upon exposure to pro-inflammatory cytokines such as tumor necrosis factor-α (TNF-α)19,20. Non-destructive evaluation of TEBV maturation and vasoactivity is beneficial toward effective monitoring of responses to drugs or stimuli. Table 1 summarizes previous vasoactivity assessments performed on TEBVs. Although TEBVs have been assessed for endothelium-independent vasoconstriction or vasodilation using a variety of agents, TEBVs comprised of human cells have yet to be evaluated for endothelium-dependent vasoactivity under physiological fluid perfusion conditions. One important clinical assessment of cardiovascular health, which predicts future cardiac events, involves infusion of acetylcholine to induce vasodilation of the coronary or brachial arteries21,22,23. Acetylcholine is a muscarinic cholinergic receptor agonist that stimulates release of nitric oxide, prostacyclin, and endothelium-derived hyperpolarizing factor in vessels with healthy, intact endothelium. Conversely, acetylcholine activates the muscarinic receptors on the smooth muscle cells in vessels with dysfunctional or damaged endothelium, leading to vasoconstriction21,24. Table 1 TEBV Vasoactivity. In this study, dense collagen gels15,25 embedded with human neonatal dermal fibroblasts (hNDFs) or human bone marrow-derived mesenchymal stem cells (hMSCs) were used to construct TEBVs that can be perfused under physiological conditions in a few hours. We compared hNDFs to hMSCs as a medial cell source to determine the cell source that would provide the greatest mechanical properties and contractility within 1–2 weeks. TEBV lumens were endothelialized with endothelial progenitor cells derived from patients with coronary artery disease (CAD EPCs)26,27. We used CAD EPCs because they express markers of healthy endothelium and function similar to human aortic ECs27, and provide the potential for future studies of a population pre-disposed to atherosclerosis. Perfused and endothelialized TEBVs exhibited high mechanical strength and contractility after one week and maintained these properties for five weeks in vitro. Phenylephrine and acetylcholine were used to non-destructively measure endothelium-independent vasoconstriction and endothelium-dependent vasodilation, respectively. We have used the non-specific phosphodiesterase inhibitors caffeine and theophylline to evaluate drug-induced vasodilation. We further evaluated endothelial function by measuring nitric oxide production and the acute inflammatory response following exposure to TNF-α. Finally, we have evaluated the potential for drug response testing through exposure to lovastatin for 3 days prior to exposure to TNF-α.