Tissue-engineered blood vessels as promising tools for testing drug toxicity

Abstract

Drug-induced vascular injury (DIVI) is a serious problem in preclinical studies of vasoactive molecules and for survivors of pediatric cancers. DIVI is often observed in rodents and some larger animals, primarily with drugs affecting vascular tone, but not in humans; however, DIVI observed in animal studies often precludes a drug candidate from continuing along the development pipeline. Thus, there is great interest by the pharmaceutical industry to identify quantifiable human biomarkers of DIVI. Small-scale endothelialized tissue-engineered blood vessels using human cells represent a promising approach to screen drug candidates and develop alternatives to cancer therapeutics in vitro. We identify several technical challenges that remain to be addressed, including high-throughput systems to screen large numbers of candidates, identification of suitable cell sources and establishing and maintaining a differentiated state of the vessel wall cells. Adequately addressing these challenges should yield novel platforms to screen drugs and develop new therapeutics to treat cardiovascular disease.

Keywords:

• drug induced vascular injury
• drug toxicity
• endothelial dysfunction
• tissue engineered blood vessels
• vasoactivity

High attrition rates in the drug development pipeline, primarily due to safety and efficacy, are a prevalent concern for the pharmaceutical industry. Drug-induced vascular injury (DIVI) observed in animal studies often precludes a drug candidate from continuing along the development pipeline; therefore, establishing whether DIVI occurs in human vessels is critical. Several classes of drugs affect vascular function in animals and humans (Table 1). Although DIVI affects both smooth muscle cells (SMCs) and endothelium, alteration of endothelial function often triggers DIVI.

Table 1. Drugs adversely affecting vascular function.

Drug class	Mechanism of action	Clinical application	Vascular effects	Observed in humans?
Phosphodiesterase inhibitors (e.g., Roflumilast)	Suppression of proinflammatory cytokines; vasodilation of respiratory smooth muscle	Treatment of pulmonary diseases such as COPD or asthma	Periarteritis [10], endothelial loss [11] and vessel injury [5] in rats, pigs and dogs	No
Chemotherapeutic drugs (e.g., platinum compounds [Cisplatin], anthracyclines [doxorubicin], taxanes [paclitaxel])	Induce DNA damage, promoting cell death and preventing tumor cell proliferation and growth, and prevent angiogenesis	Cancer treatment	Detection of endothelial markers in human plasma, including vWF, PAI-1, t-PA and CRP; decreased nitric oxide levels; reduced brachial artery flow-mediated dilation; endothelial cell apoptosis and promotion of inflammation [12]	Yes
Adenosine agonists (e.g., adenosine, CI-947)	Cardiac vasodilation	Hypertension	Transmural necrosis and inflammatory cell infiltration in dogs [13]	No
Dopamine (DA1) agonist (e.g., fenoldopam)	Activates DA1 receptors in nephron, promoting sodium excretion	Hypertension	Vasodilation, necrosis of medial smooth muscle cells, hemorrhage; decreased caveolin-1 expression and apoptosis in rats [14]	No
Potassium channel opener (e.g., minoxidil, hydralazine, apresoline)	Opening sarcolemmal K^+ channels relaxes vascular smooth muscle	Hair growth, hypertension	Vasodilation; medial hemorrhage and necrosis and perivascular inflammation in dogs [15]	No
Endothelin receptor antagonist (e.g., SB209670, CI-1020)	Relaxes arteries by blocking smooth muscle and endothelial endothelin-1 receptors	Hypertension	Vasodilation; medial smooth muscle necrosis and arterial inflammation in dogs [16]	No

COPD: Chronic obstructive pulmonary disease; CRP: C-reactive protein; PAI-1: Plasminogen activator inhibitor-1; t-PA: Tissue plasminogen activator; vWF: von Willebrand Factor.

The vascular endothelium lines the inner surface of all blood vessels, making contact with blood and regulating thrombosis, permeability and vascular tone. Vascular SMCs are present in the cell media. Normally they are quiescent and maintain the vessel tone, although in tissue-engineered blood vessels (TEBVs) these cells need to synthesize extracellular matrix prior to become quiescent and contractile.

The endothelium modulates contraction and relaxation of the smooth muscle cells through release of nitric oxide and prostacyclin in response to changes in flow or stimuli with vasoactive compounds such as acetylcholine [1]. Conversely, the vasoconstrictor endothelin-1 is produced during inflammation or in low shear stress environments and its receptors are upregulated during disease [2]. Hypertension, diabetes, hypercholesterolemia and smoking induce endothelial activation and dysfunction, leading to reduced vasodilation in response to changes in flow. Endothelial dysfunction often results in endothelial cell death, and the dead cells are replaced largely by replication of adjoining endothelial cells. This chronic process of endothelial cell dysfunction, death and repair causes proliferation of smooth muscle cells, ultimately leading to atherosclerosis, the major pathology contributing to cardiovascular disease, the leading cause of death in the world today.

Clinically, vasodilation of the coronary or brachial arteries after an infusion of acetylcholine is used to assess endothelial dysfunction and predict future cardiac events [3]. Damage to the endothelium is caused by reactive oxygen species (ROS). Low and moderate levels of ROS promote inflammation by activating the transcription factor NF-кB, which regulates a number of proinflammatory genes. Further, ROS decrease the activity of nitric oxide synthase and react with nitric oxide to form peroxynitrite, resulting in reduced NO concentration and impaired vasodilation. Reduced NO bioavailability triggers endothelial and smooth muscle cell activation and promotes the production of proinflammatory cytokines such as TNF-α and IL-6. Activated endothelium express cell surface adhesion molecules such as vascular cell adhesion molecule 1, intercellular adhesion molecule 1, and E-selectin on their surface, promoting adhesion of leukocytes to the vascular wall [4]. A common pathway in DIVI appears to be site-specific alteration of endothelial nitric oxide synthase activity [5] that may be affected by ROS production induced by drugs.

In addition to DIVI observed in animals, various drugs to treat cancer can affect the function of vascular wall cells. Survivors of pediatric cancers often have worse outcomes due to cardiovascular complications as a result of chemotherapy and/or radiation therapy. These arterial injuries appear to result from endothelial dysfunction induced by such chemotherapeutic agents as anthracyclines and cyclophosphamides. Anthracyclines, principally doxorubicin, are a major component of chemotherapy for pediatric cancers. The cardiotoxicity of doxorubicin is well established and treatment doses have been reduced to prevent congestive heart failure and arrhythmias. Clinical studies in long-term survivors of pediatric cancer (i.e., ≥ 10 years) show altered vascular function including reduced flow-mediated dilation associated with chemotherapy [6] and carotid and femoral intimal thickening in patients who have received radiation therapy [7]. Use of a physiologically accurate in vitro model of the vasculature during preclinical toxicity testing can facilitate the production of more targeted therapies to avoid the systemic side effects induced by chemotherapeutic drugs.

Although the response of endothelium to different drugs can be screened in culture, the culture conditions need to properly mimic the physiological environment. At a minimum, drug response studies with endothelial cell monolayers should be performed following exposure to physiological shear stresses that promote alignment of the endothelium and enhanced expression of various vasoactive, anti-inflammatory and anti-thrombotic molecules. Otherwise, the reference condition of key antithrombotic and anti-inflammatory genes is incorrect and the measured response to the drug may not reflect the overall change observed in vivo, where the endothelium has adapted to the flow conditions.

Although studies with endothelium exposed to flow in two-dimensional cultures can provide some insight into the action of various drugs, and endothelial–smooth muscle co-cultures might better simulate the physiological environment, these models cannot provide a complete functional assessment of vasoactivity in response to drugs, oxidants or inflammation. As an alternative, endothelialized TEBVs made with human cells represent a promising approach to replicate the physiological environment in vitro. TEBVs consist of a medial layer of contractile cells (smooth muscle cells, mesenchymal stem cells or fibroblasts) in a three-dimensional construct composed of synthetic biodegradable polymers or biological hydrogels (e.g., fibrin or collagen). Endothelial cells and endothelial progenitor cells have been successfully seeded onto such constructs, forming confluent monolayers that remain adherent in the presence of physiological shear stresses. Although medial cell contractility and maturity has been assessed in the presence of vasoconstrictors, to date, endothelial-mediated vasodilation by medial cells has not been reported in TEBVs. Since the endothelium is a primary regulator of vascular tone, evaluation of endothelial-mediated vasodilation is a critical step toward creating an in vitro vascular injury model that may prove useful toward the discovery human biomarkers of DIVI.

Using endothelialized human TEBVs, the potential cause of inflammation and injury of the vessel wall can be assessed with human cells under physiological conditions. Furthermore, in vitro models provide the capacity to isolate and manipulate variables within the tissue microenvironment to evaluate specific tissue responses. For example, leukocytes can be added in the presence of proinflammatory cytokines (e.g., TNF-α or IL-1) to examine the effect of a drug candidate toward mediating the inflammatory response. By directly comparing TEBVs made with either animal cells or human cells, it may be possible to determine the differential response to drugs that produce DIVI in animals but not humans. Finally, induced pluripotent stem (iPS) cells or progenitor cells could be differentiated into SMCs and endothelium and used to produce TEBVs with cells from the same individual to examine the effect of genetic variations on drug responses. Such TEBVs could be used for examining therapeutic solutions for specific patient populations.

While these opportunities are exciting, several challenges remain to establish functional testing of drug toxicity with endothelialized TEBV. Foremost is the need to develop high-throughput systems to screen large numbers of candidates. This necessitates: i) reducing the maturation time before the vessels can undergo perfusion from the current period of 6 – 8 weeks [8] to ∼ 2 weeks; and ii) decreasing the size of vessels to reduce perfusion volumes and overall system size. Tissue-engineered arterioles with diameters between 50 and 200 µm would be ideal, since many drug-induced problems affect vasoactivity of the resistance vessels. For a physiological shear stress of 1.2 Pa, the flow rates would range from 17.7 to 1080 µl/h. Blood flow is quasi-steady under such conditions, so gravity-driven chambers could be used for TEBV perfusion, further simplifying the system operation. New fabrication methodologies would be needed for such small vessels. Although arteriolar-scale TEBVs can assess vasoactivity and leukocyte adhesion, pathologies observed in large arteries (accumulation of lipids, altered vessel elasticity) remain to be demonstrated. Last, detection of soluble proteins, cytokines and other small molecules produced in response to drug interactions may require reducing system volumes. Thus, validation of these systems in multiple laboratories against a set of well-established reference compounds is needed.

Primary human vascular cells may be readily differentiated toward a mature phenotype; however, their accessibility, proliferative capacity, potential for aging, and senescence limit their capacity for larger-scale production applications. Although iPS cells hold great promise, a chief limitation is that they often do not express the phenotype of differentiated cells. Methods need to be developed to ensure that endothelium and smooth muscle derived from iPS cells maintain fidelity with those in the native adult vessel wall. Some recent results with endothelial cells derived from iPS cells provide one strategy to promote differentiation of endothelial cells with highly specialized barrier function [9]. Alternatively, blood-derived endothelial progenitor cells and mesenchymal stem cells can more easily be differentiated into functioning endothelial and smooth muscle cells.

Overall, the promise of TEBVs to assess drug toxicity is significant. The challenges are real but can be addressed. Developing easy-to-use and cost-effective systems could create a viable approach to assess toxicity in blood vessels and screen drug candidates, leading to more accurate preclinical testing.

Declaration of interest

GA Truskey was supported by a NIH grant UH3TR000505 and the NIH Common Fund for the Microphysiological Systems Initiative. CE Fernandez was supported by an American Heart Association Pre-doctoral Fellowship 14PRE20500062. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.