Biological and engineering design considerations for vascular tissue engineered blood vessels (TEBVs)

Highlights

• Initial clinical results with tissue engineered blood vessels are promising.

• Acellular grafts can be rapidly fabricated for applications in high flow.

• Cell-based engineered vessels are needed to reproduce full function of arteries.

• A confluent endothelium still appears to be needed to replace small diameter vessels.

• New approaches to produce vessels should address clinical challenges.

Considerable advances have occurred in the development of tissue-engineered blood vessels (TEBVs) to repair or replace injured blood vessels, or as in vitro systems for drug toxicity testing. Here we summarize approaches to produce TEBVs and review current efforts to (1) identify suitable cell sources for the endothelium and vascular smooth muscle cells, (2) design the scaffold to mimic the arterial mechanical properties and (3) regulate the functional state of the cells of the vessel wall. Initial clinical studies have established the feasibility of this approach and challenges that make TEBVs a viable alternative for vessel replacement are identified.

Introduction

Due to limitations of existing approaches to treat obstructed coronary, carotid and peripheral arteries and arteriovenous shunts for hemodialysis patients, tissue-engineered blood vessels (TEBVs) hold the potential of providing a readily available source to treat a number of complications that arise from cardiovascular disease. Although considerable progress has been made toward developing TEBVs and clinical studies have begun, the following key challenges to produce functional engineered vessels remain: (1) produce nonthrombogenic and nonimmunogenic surfaces in contact with blood; (2) develop vessels with appropriate material and mechanical properties to withstand pulsatile blood pressures without failure, permanent deformation or stenosis; and (3) enable physiological vasoconstriction and vasodilation. For clinical application, vessels should be readily available with limited processing time and at a cost competitive to existing procedures. Requirements for cell harvesting and tissue fabrication are specified by Food and Drug Administration (FDA) guidance documents (http://www.fda.gov/BiologicsBloodVaccines/GuidanceComplianceRegulatoryInformation/Guidances/Tissue/default.htm). Although these requirements are demanding, they also provide opportunities for innovative ways to design TEBVs.

Three general approaches are used to develop TEBVs for clinical applications [1•, 2] (Figure 1 and Box 1): (1) in vitro assembly of vessels with cells and degradable synthetic or biological scaffolds; (2) in vitro self-assembly from cell sheets; and (3) in vivo vessel formation of implanted acellular grafts derived from decellularized blood vessels, subintestinal submucosa or cultured allogeneic smooth muscle cells (SMCs) [3^••].

In vitro methods often require extended culture periods for cells to produce and remodel the extracellular matrix (ECM) so that TEBVs have suitable mechanical strength [2], whereas acellular approaches rely upon the in vivo growth of cells from adjacent vessels into decellularized grafts to promote remodeling. Maturation of acellular grafts may be compromised in individuals with cardiovascular disease, leading to incomplete graft remodeling and reduced vasoactivity and endothelialization. Animal studies suggest that addition of cells to acellular grafts before implantation may improve their in vivo performance [4]. Given that endothelialization of grafts by ingrowth from adjacent vessels is limited, TEBVs with inner diameters less than 6 mm may need to be seeded with endothelial cells (ECs) to prevent thrombosis.

Addressing these challenges involves identifying suitable autologous or derived cell sources for the endothelium and vascular smooth muscle cells, designing the scaffold to mimic the arterial mechanical properties and regulating the functional state of the cells of the vessel wall. After discussing recent clinical studies, we review progress in each of these areas.

The first clinical trial of TEBVs to treat single ventricle congenital defects involved 25 patients ranging in age from 1 to 24 years [5]. Biodegradable scaffolds of woven polyglycolic acid, poly-l-lactide and ɛ-caprolactone (50:50) were seeded with autologous bone marrow mononuclear cells and implanted as grafts to reconfigure portions of the pulmonary circulation. Over a mean follow-up time of 5.8 years, no graft-related deaths occurred, and all vessels remained patent. The major complication of the implanted grafts was stenosis in 24% of the patients, which could be treated with balloon angioplasty. These studies demonstrated the feasibility of using TEBVs to replace low-pressure blood vessels.

L’Heureux and colleagues at Cytograft Tissue Engineering produced an entirely autologous blood vessel using cell sheet tissue engineering [6]. Human fibroblasts were extracted from skin biopsies and ECs were harvested from a superficial vein. These small-diameter TEBVs had sufficient mechanical strength and were successfully used for hemodialysis in a clinical trial involving 10 patients with a total of 68 patient-months of patency [7^••]. Seeding of the lumen with autologous ECs also provided grafts with the necessary antithrombotic lining. In an effort to reduce the time required to produce the completely autologous vascular graft, the L’Heureux group is examining the use of allogeneic human fibroblasts, non-endothelialized vessels, and the assembly of three-dimensional vessels from threads of cell-synthesized extracellular matrix (ECM) [8].

Investigators at Humacyte [9] generated decellularized scaffolds by first growing human SMCs on a tubular polyglycolic acid (PGA) scaffold, and then removing all cellular material with detergents to leave behind a TEBV comprised completely of ECM. These TEBVs have favorable mechanical properties, in part through the incorporation of PGA, and are nonimmunogenic since all cellular material is removed before implantation. These TEBVs (inner diameter ≥6 mm) were tested in an arteriovenous (AV) fistula model between the axilliary artery and the brachial vein of baboons, and exhibited >80% patency for up to 6 months. After implantation all TEBVs showed extensive medial layer remodeling and exhibited partially endothelialized regions near the anastomotic sites with native vessels. Based on promising preclinical studies (Dahl et al. Circulation. 2013; 127: 2071–2072, doi: 10.1161/CIR.0b013e318295baf5) Humacyte began a multi-center European clinical trial in December 2012, and additional patient enrollment was approved after a safety review in April 2013. In June 2013, the FDA approved U.S. clinical trials to assess safety and function of these acellular grafts in dialysis patients who are unable to undergo AV fistula formation due to prior vessel damage.

The ideal TEBV cell sources for vascular endothelial and smooth muscle cells should be autologous, capable of many cell divisions, and able to differentiate into the mature phenotype. Adult stem populations represent a promising source of autologous cells with the capacity to differentiate to vessel wall cells [10, 11]. Autologous mesenchymal stem cells (MSCs) from bone marrow, adipose tissue or muscle can differentiate to form a contractile cell type similar to vascular smooth muscle [11].

The rapid generation of an endothelial layer is especially important for functional TEBVs with diameters less than 6 mm [2]. Autologous human microvascular ECs can be obtained from jugular or saphenous vein or liposuctioned fat, but these procedures are unattractive because of the limited number and lifespan of cells obtained, and the invasiveness of the procedures. Adipose-derived microvascular cell cultures are often contaminated with other cell types (e.g. macrophages and fibroblasts) resulting in an increased rather than decreased development of intimal hyperplasia in a dog model [12] and a decreased patency in transluminally seeded vessels in a rabbit model [13].

Promising approaches to derive autologous ECs that do not suffer from these limitations include (1) differentiating MSCs directly to ECs [14], (2) differentiating blood-derived late outgrowth endothelial progenitor cells (EPCs) into ECs [15], (3) dedifferentiating host stromal cells to intermediate induced pluripotent stem (iPS) cells that then differentiate to ECs [16, 17], or (4) transdifferentiating harvested host stromal cells directly to ECs without inducing pluripotency [18].

Bone marrow derived cells initially generated considerable enthusiasm because large numbers of mononuclear cells (MNCs) are readily available via bone marrow aspiration from the iliac crest and could be utilized on the day of TEBV implantation. The expectation was that bone marrow derived MNCs would differentiate into ECs in vivo [19]. Instead, bone marrow MNCs evoke an inflammatory-mediated process of remodeling [5]. Addition of EC-specific growth factors such as vascular endothelial growth factor and fibroblast growth factor were not successful in producing ECs [11]. Exposure of bone marrow MSCs to fluid shear stress does induce expression of EC-specific molecules such as von Willebrand Factor (vWF), platelet-endothelial cell adhesion molecule and VE-cadherin, suggesting that ECs could be derived from MSCs [20].

Our work has focused on EPCs as a readily available source of host ECs that can be obtained from adult peripheral blood or umbilical cord blood [21]. Depending on the isolation method, two functionally distinct populations known as early-outgrowth and late-outgrowth EPCs can be obtained. Only late-outgrowth EPCs, or endothelial colony forming cells (ECFCs), exhibit conventional EC behavior and EC surface markers [22, 23], as well as a high proliferative potential. Late-outgrowth EPCs isolated from healthy individuals and from patients with cardiovascular disease exhibit senescence. ECs derived from late-outgrowth EPCs elicit substantially lower alloimmune reaction than aortic ECs [24] and may be an allogeneic source. EPCs from umbilical cord blood can undergo 50–60 cell divisions before senescence [25]. Together with new protocols that permit higher yields of EPCs from small blood volumes [26], over 10¹⁰ cells could be banked from each isolation, making this an attractive cell source.

Late-outgrowth EPCs resemble mature ECs and are not derived from bone marrow [27]. ECs derived from human umbilical cord blood [28] and adult blood [25] function the same as human aortic ECs (HAECs) with regard to: expression of VE-cadherin, CD31, vWF; uptake of acetylated low-density lipoprotein; elongation and increased nitric oxide (NO) levels at physiologic shear stresses; elongation and alignment with flow direction; and up-regulation of key EC genes Krüppel-like factor 2, endothelial NO synthase (eNOS), cyclo-oxygenase 2, and thrombomodulin after exposure to flow.

Porcine late-outgrowth EPCs can spread and proliferate on titanium surfaces in vivo, protecting against thrombosis even in the low shear environment of the inferior vena cava [29^•]. Likewise, EPCs from blood of humans with cardiovascular disease seeded onto 1 mm diameter expanded poly-tetrafluoroethylene vascular grafts had 28-day patency rates of 75–88%, although intimal hyperplasia was observed near the proximal and distal anastomoses [15].

Human vascular SMCs often have limited proliferative and synthetic capability, compromising their ability to produce mechanically strong TEBVs. Cells with SMC properties could be derived from bone marrow mononuclear cells, mesenchymal stem cells from bone marrow, adipose tissue and skeletal muscle [11]. A promising source of vascular smooth muscle is iPS cells derived from late outgrowth EPCs [30^•]. These EPCs can be easily reprogrammed and do not exhibit copy number variations [30^•], raising the possibility of deriving pure populations of autologous cells to generate TEBVs.

The mechanical behavior of arteries enables their expansion after ejection of blood from the left ventricle with only a modest rise in pressure, reducing the work on the heart. The media of small-diameter muscular arteries, such as coronary arteries, contains SMCs arranged in concentric layers within an ECM comprised primarily of collagen and elastin [31]. Elastin fibers are located in concentric lamellae between SMC layers to support the compliance of the vessel during pulsatile flow [32] (Figure 2 and Box 2). Collagen fibrils in native arteries are organized in circumferential, helical, and axial directions [33], providing tensional strength to the native artery at high strains. Replicating the ECM composition of native arteries in a tissue engineered vascular construct has proven to be difficult. Transmission electron microscopy indicated that collagen in tissue engineered arteries was surrounded by glycosaminoglycans (GAGs) and SMCs, while that of native arteries was surrounded by elastin or other collagen fibrils [33]. Thus, the natural ECM environment has not yet been properly recreated within TEBVs.

Compared to native vessels, small-diameter synthetic grafts are stiffer, stronger and less compliant, and fail due to thrombosis and neointimal hyperplasia from compliance and diameter mismatch between the graft and the native artery [34]. TEBVs are generally weaker, although the compliance more closely matches values for native vessels.

A balance between mechanical strength and chemical functionality is needed when choosing a support matrix for TEBV manufacture. PGA fiber matrices have been widely used since they degrade slowly, allowing sufficient time for the maturing vessel media to develop sufficient mechanical strength [9, 35]. Fibrin matrices stimulate ECM protein production leading to stronger TEBVs [36, 37]. The burst strength of fibrin vessels with human dermal fibroblasts approaches values of native vessels after pulsatile stretch at physiological pressures for 7–9 weeks [38], emphasizing the importance of biomechanical stimuli.

To enhance the strength of collagen gels as a support matrix for TEBVs, newly formed collagen gels may be plastically compressed to produce dense collagen scaffolds with collagen fibrillar densities comparable to those of the native ECM [39]. Dense collagen gels made from plastic compression have demonstrated good incorporation with cells, allowing for good adhesion and proliferation [40, 41•]. Further conditioning under pulsatile pressure increases burst strengths to 1000 mm Hg [40].

Elastin fiber production is currently lacking from most TEBV approaches and has not been a primary focus of TEBV production until recently. However, elastin is crucial for the mechanical and signaling properties of vascular tissues. Elastic fibers comprise 30–50% of the dry weight of native vascular tissues and play important roles in the induction of actin stress fiber organization as well as the inhibition of SMC proliferation and migration [42, 43]. The proliferation of arterial SMCs is modulated by the transduction of signaling pathways activated by the interaction of soluble elastin degradation products with the elastin receptor [44]. SMC proliferation produces stenoses in arteries when extracellular elastin is not present [45].

Functional extracellular elastin formation is a complex process involving elastin protein production, secretion and fibril formation. Elastin fiber production in engineered vascular tissues is largely prevented by reduced translation of tropoelastin mRNA in cells older than neonatal cells and by inefficient tropoelastin recruitment and cross-linking into the elastic matrix [46]. Elastin synthesis in the native environment is promoted by cyclic GMP, insulin-like growth factor 1, transforming growth factor β1 (TGFβ1), and fibrin degradation products [46].

Substrate composition and topography influence the elastin production by medial cells such as vascular SMCs or fibroblasts. Elastin-based substrates are subject to enzymatic degradation without providing biochemical and biomechanical signals promoting elastin synthesis by the medial cells [43]. Although collagen promotes the quiescent, contractile phenotype of SMCs, this can limit the synthesis of elastin precursors and assembly of elastin structures in the ECM [46, 47]. 3D scaffold topography and the presence of TGF-β1 increased elastin gene expression and synthesis and expression of contractile markers by human coronary artery SMCs [48^•]. Interestingly, TGF-β1 did not affect elastin synthesis in 2D cultures. Furthermore, a correlation was found between the pore size of a substrate seeded with baboon and porcine SMCs and elastin and collagen production [49]. Vascular SMCs embedded in 3D collagen gels exposed to long-term cyclic distention increased production of elastin but not collagen [50]. Dense collagen gels have been combined with elastin protein polymer layers to create robust, mechanically strong TEBVs without the incorporation of medial cells [51].

TEBVs created from cell sheets of SMCs transduced with splice variant 3 of the proteoglycan versican and cultured with reduced exposure to ascorbate exhibited greater tropoelastin production, elastin crosslinks, and thicker collagen fiber bundles [52]. The presence of hyaluronan oligomers and TGF-β1 increases the elastin matrix deposition of adult rat aortic SMCs seeded within 3D collagen gels [53]. Rapamycin promote the contractile phenotype of SMCs and elastin synthesis in normal SMCs and iPS cells derived from patients with Williams–Beuren Syndrome, which involves a micro-deletion of one copy of the tropoelastin gene on chromosome 7 [54]. Since rapamycin is used clinically in drug eluting stents to inhibit SMC proliferation, this drug could be added to TEBVs after the vessel wall cells have proliferated sufficiently, thereby inducing a contractile phenotype and initiating elastin synthesis. Rapamycin addition would need to be coordinated so as not to interfere with EC adhesion and growth.

Endothelialized TEBVs would allow for the creation of in vitro drug testing models [55]. To date TEBVs have been studied under ideal conditions for healthy individuals. However, in the clinic, TEBVs would be implanted in patients with atherosclerosis and systemic inflammation. Therefore, evaluation of the vasoreactive response under these conditions would provide more realistic models for how the TEBV would behave after implantation. The endothelium serves as a primary target for medications for blood pressure and inflammation. An endothelialized, vasoreactive TEBV would enhance the in vitro study of new drugs to regulate cholesterol and treat hypertension, diabetes and autoimmune diseases such as lupus.

Acute inflammatory responses may be elicited through exposure to tumor necrosis factor-α (TNF-α), causing endothelium to express inflammatory markers such as vascular cell adhesion molecule-1 (VCAM-1), intracellular adhesion molecule-1 (ICAM-1), and E-selectin. Furthermore, inflammation has been shown to impair endothelial vasomotor function [2]. Statins, which are 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors, lower blood cholesterol levels and improves endothelium-dependent vasodilation [56]. Atorvastatin has reduces endothelial cell expression of ICAM-1 and VCAM-1 induced by TNF-α exposure in endothelial cells cultured alone [57]. Creating a vasoresponsive TEBV would open the door for more accurate in vitro models for testing endothelial response to medications.