Evaluating In Vitro Cardiovascular Toxicity using Tissue-Engineered Blood Vessels (TEBVs)

Introduction: : Cardiovascular toxicity remains a critical yet underrepresented endpoint in environmental and pharmaceutical safety testing. In fact, cardiotoxicity is one of the leading causes of drug failure during mid-to-late-stage clinical trials. However, current preclinical models fall short in predicting human-specific vascular responses. Animal studies are costly, often poorly translatable, and lack resolution for detecting subtle vascular injury; while two-dimensional (2D) cultures fail to reproduce the physiological structure, multicellular architecture, and hemodynamic forces of human vessels. There is a pressing need for human-relevant, three-dimensional (3D) in vitro models to improve cardiovascular risk assessment.

To address these limitations, we developed a 3D, human-relevant tissue-engineered blood vessel (TEBV) model using iPSC-derived endothelial cells and human neonatal dermal fibroblasts cultured under flow conditions. This model enables simultaneous assessment of vascular function, structure, and inflammatory responses under compound exposure.

As part of a contract with the National Institute of Environmental Health Sciences (NIEHS), we evaluated five toxicants—triclosan, bisphenol A, ephedra, pazopanib, and chlorpyrifos—at concentrations of 0.3, 3, and 30 µM. The central objective was to determine whether these agents impair vascular contractility, disrupt endothelial integrity, or induce inflammation and cell death in a dose-dependent manner. By integrating functional, biochemical, and structural readouts, this model advances our ability to screen compounds for early signs of vascular injury, ultimately improving human relevance in toxicology research. This study aims to establish TEBVs as a predictive tool for cardiovascular toxicity screening, providing a mechanistic framework to evaluate chemical safety and better inform regulatory decisions.

Materials and

Methods: : TEBVs were fabricated using collagen scaffolds seeded with primary human neonatal dermal fibroblasts (hNDFs, passage 11) and iPSC-derived endothelial cells (viECs, passage 3) to establish functional endothelium and medial layers. Constructs were cultured under continuous laminar flow for 7 days in a custom perfusion bioreactor to promote physiological vessel maturation and alignment. Baseline functional assessments of vasoactivity were conducted on day 7 using live imaging to quantify vessel diameter changes in response to vasoactive agents. Phenylephrine (PE, 10 µM), a vasoconstrictor, was perfused first to assess smooth muscle-mediated contraction, followed by acetylcholine (ACh, 10 µM), a vasodilator, to evaluate endothelial-dependent relaxation.

TEBVs were then exposed to each compound at three concentrations (0.3 µM, 3 µM, 30 µM) for 4 days. Functional testing was repeated on days 9 and 11 to assess longitudinal changes. Vehicle controls (0.1% DMSO) and untreated media controls were included for normalization.

Post-treatment analyses included: Live/Dead staining using acridine orange (AO) and propidium iodide (PI) to assess cytotoxic effects on ECs and fibroblasts; immunofluorescence (IF) staining of cryosectioned vessels for markers of endothelial integrity (CD31), fibroblast structure (calponin), inflammation (ICAM-1), and nuclei density (Hoechst); and quantification of fluorescence intensity and nuclei count using ImageJ to assess cellular viability and structural integrity. Statistical comparisons were performed using non-parametric t-tests and one-way ANOVA for paired comparisons. P-values < 0.05 were considered significant. Data are presented as mean ± SEM unless otherwise noted.

Results, Conclusions, and Discussions:: Exposure to triclosan (TCS) resulted in reduced vasoactivity in TEBVs, though the trend was not strictly dose-dependent. Reduced vasoconstriction and vasodilation responses were observed at all TCS doses compared to media controls. The most pronounced PE reduction was observed at 3 µM. While partial PE responses remained at 0.3 and 30 µM, their magnitudes were diminished and neither reached statistical significance when tested against 0% change. The ACh effect seemed to follow a dose-dependent trend with some retained vasodilation at 0.3 µM, yet none of the TCS-treated groups retained statistically significant responses, suggesting impaired vessel function across doses despite biological variability.

Although DMSO vehicle exhibited slightly reduced vasoactivity compared to media, responses remained significant, indicating preserved endothelial function. DMSO-treated vessels behaved similarly to media controls in latter viability and structural assays, suggesting the solvent effects were minimal and did not confound interpretation.

Live/Dead staining demonstrated dose-dependent vascular dysfunction, with significant cytotoxicity at 30 µM triclosan (p < 0.0001), while lower concentrations primarily impaired function without immediate widespread cell death.
Immunofluorescence revealed a progressive loss of endothelial CD31 expression and increased ICAM-1 expression with increasing triclosan dose, suggesting inflammation-driven endothelial dysfunction even at sub-lethal doses. Calponin staining and fibroblast nuclei density remained largely stable, indicating selective endothelial sensitivity. Image quantification confirmed that structural damage and inflammation co-occurred with functional impairment.

Overall, triclosan induced a combination of inflammatory and cytotoxic effects, disrupting vascular function and structure in TEBVs. Similar trends were observed with other compounds, with each exhibiting distinct toxicity profiles based on mechanism: BPA, pazopanib, and chlorpyrifos caused both cytotoxicity and loss of endothelium; ephedra primarily triggered endothelial detachment with limited direct cell death, suggesting a mechanical mechanism.

This study demonstrates that TEBVs can effectively model compound-specific vascular toxicities, capturing early functional impairments and mechanistic responses. The model is scalable, reproducible, adaptable to a broad range of compounds; it is also sensitive enough to detect early functional changes in the absence of overt cytotoxicity, highlighting its potential for early toxicity prediction. Thus, this engineered vessel platform holds promise for reducing animal use, enhancing mechanistic insight, and improving the accuracy of cardiovascular safety evaluations.

Acknowledgements and/or References (Optional):: This work was supported [in part] by the Intramural Research Program of the NIH (ES103382) and by contract 75N96023P00240 with NIEHS. Special thanks to Dr. George Truskey and all members of the Truskey Lab for their mentorship and support.