Current biofabrication methods for vascular tissue engineering and an introduction to biological textiles

Cardiovascular diseases are the leading cause of mortality in the world and encompass several important pathologies, including atherosclerosis. In the cases of severe vessel occlusion, surgical intervention using bypass grafts may be required. Synthetic vascular grafts provide poor patency for small-diameter applications (< 6 mm) but are widely used for hemodialysis access and, with success, larger vessel repairs. In very small vessels, such as coronary arteries, synthetics outcomes are unacceptable, leading to the exclusive use of autologous (native) vessels despite their limited availability and, sometimes, quality. Consequently, there is a clear clinical need for a small-diameter vascular graft that can provide outcomes similar to native vessels. Many tissue-engineering approaches have been developed to offer native-like tissues with the appropriate mechanical and biological properties in order to overcome the limitations of synthetic and autologous grafts. This review overviews current scaffold-based and scaffold-free approaches developed to biofabricate tissue-engineered vascular grafts (TEVGs) with an introduction to the biological textile approaches. Indeed, these assembly methods show a reduced production time compared to processes that require long bioreactor-based maturation steps. Another advantage of the textile-inspired approaches is that they can provide better directional and regional control of the TEVG mechanical properties.


Introduction
According to the American Heart Association, cardiovascular diseases (CVDs) were the leading cause of mortality in the United States in 2021 [1]. The most common CVD is atherosclerosis. This pathology is a progressive inflammatory disease that develops in the vascular media and causes hyperplasia and calcification of blood vessels [2,3]. The progressive obstruction of the vessel lumen is often asymptomatic until the formation of a thrombus [2]. In addition, at an advanced stage of development, the atheromatous plaque can break and release clots into the blood circulation. These can lead to the occlusion of smaller arteries, such as the brain arterioles, and thus causes ischemic injuries. One option to restore blood circulation is percutaneous angioplasty, which consists of introducing a balloon within a peripheral artery, bringing it to the pathological artery, inflating it, compressing the atheromatous plaque, and restoring the artery lumen [4,5]. This procedure is usually associated with using a stent that keeps open the lumen of the vessel [5]. Nevertheless, in response to the injury of the vessel wall, the removal of the endothelium, and the presence of the stent, the vessel often develops restenosis [6]. Nonetheless, in endovascular peripheral artery procedures, stents are not always necessary, and balloon angioplasty can be adequate for most of the procedures, especially in below-the-knee arteries [7]. Although more costly and risky than percutaneous interventions, a more long-term solution is to bypass the blockage with a conduit [8,9].
Synthetic vascular grafts (SVGs) are commercially available and have been successfully used as large and medium internal diameter (ID) prosthetics (ID ⩾ 6 mm) (table 1) [10]. Generally, these SVGs are composed of expanded polytetrafluoroethylene (e-PTFE, also known as Gore-Tex®) or woven/knitted polyethylene terephthalate (PET, also known as Dac-ron®) yarn. With a medium ID (⩾ 8 mm), these grafts º Widely available º Low cost of production º Easy to produce º On the market º Pro-thrombogenic º Cause a chronic inflammatory reaction º Intimal hyperplasia can occur at the graft connection sites º Exposed to infection º Cannot be remodeled by the host's cells Tissue-engineered vascular grafts (expected advantages) º Widely available º Native-like or superior mechanical properties º Low thrombogenicity º Resist infections º Can be remodeled by the host's cells º Can integrate native tissue º Costly to produce º Complex to produce º Complex to be commercialized showed primary patency rates ranging from 66%-85% (e-PTFE and Dacron, respectively) after three years for femorofemoral bypass [11]. Unfortunately, when applied to small-diameter vessels (ID < 6 mm), these artificial grafts displayed far lower patency (e.g. 20%-25% for a 1 mm diameter PTFE vessel) (table 1) [10,12]. Indeed, these materials, recognized as foreign bodies by the host's innate immune system, cause a chronic inflammatory environment that promotes thrombosis, intimal hyperplasia, calcification, and fibrosis (table 1) [13]. Despite new surface treatments designed to reduce thrombogenicity, current commercial SVGs are made from rigid materials that cannot be remodeled by the patient's cells and are also susceptible to bacterial infections (table 1) [14]. As a result, these synthetic conduits perform very poorly in small-diameter (< 6 mm) applications (table 1) [15]. As a result, these are not used for coronary bypass and only for lower limb bypass) or as arteriovenous shunts.
Autologous vascular grafts, such as internal mammary arteries or saphenous veins, remain the 'gold standard' for vascular repairs. For example, it was demonstrated that saphenous vein grafts display a low occlusion rate (10%-15%) after one year in the context of coronary artery bypass surgery, and 50% of these conduits are still patent after ten years of implantation (table 1) [16]. However, the availability of these grafts is limited due to previous use or the poor quality of the vessel because of the cardiovascular pathology itself (table 1) [17][18][19]. In addition, harvesting these grafts causes substantial donor site morbidity that can be particularly significant in elderly and/or diabetic patients (table 1) [19]. Consequently, there is a critical clinical need for a longlasting but readily available small-diameter vascular conduit.
An alternative to synthetic prostheses or autografts is to use tissue-engineered vascular grafts (TEVGs) made of scaffolds associated with living cells (figure 1). Vascular tissue engineering is a continuously growing field. The number of research articles (PubMed) and patents (Google Patents) reached a total of 680 and 334, respectively, at the end of 2021. Historically, the first TEVG described in the literature was 'a model of blood vessel' produced in vitro by Weinberg & Bell in 1986 [20]. However, Hubbell et al used the 'tissue-engineered vascular graft' term for the first time in 1991 to describe an endothelial cell-selective material for tissue engineering [21]. The domain emerged in the 2000's at an academic level with a constantly increasing interest globally (figure 2, black line). Meanwhile, industrial developments in the domain emerged during the same period, with a constantly increasing number of patents deposited over the years (figure 2, red line), demonstrating the strong interest of academics and industries in developing TEVGs. While these grafts are costly as well as complex to produce and commercialize, they have the potential to display both adequate mechanical and biological properties (table 1). Indeed, a biological composition would provide much-improved integration with the host and allow tissue remodeling that would lead to functional blood vessel regeneration. Additionally, this type of graft could be reendothelialized to have low thrombogenicity and should resist infections (table 1). Assuming this level of performance, these grafts would be an attractive option for revascularization applications involving arteries with IDs ranging between 1 to 6 mm [22].
Textile approaches represent a promising option for the development of TEVGs. Indeed, they allow to achieve high mechanical strengths while using a starting material (the thread) that is restively fragile Figure 1. Current biofabrication methods for producing human tissue-engineered vascular grafts (TEVGs). Many approaches have been developed for TEVG production. As scaffold-free approaches, cell aggregates can be 3D printed onto a support or grid of needles (Kenzan method) to generate a tube. These aggregates can also be directly assembled and fused into a tubular mold. A cell-assembled extracellular matrix (CAM) can be used to fabricate TEVGs. Indeed, a CAM sheet can be rolled and fused or cut as threads and woven around a mandrel as a tubular structure. A tubular structure can also be obtained by 3D printing a mix of cells in a soluble support solution. As scaffold-based approaches, human cells can be seeded onto biological scaffolds (e.g., collagen or fibrin gels) that have been molded as tubes. They can also be seeded onto biodegradable synthetic or biohybrid electrospun scaffolds. Finally, living fibers containing cells obtained using a double-coaxial laminar flow system and reconstituted/cross-linked collagen threads can be used to produce a TEVG. 'Tissue-engineered vascular' and 'tissue-engineered blood vessel' names, both singular and plural, were used as keywords for the call on PubMed and Google Patents. by itself. These techniques allow control of the ID, the length, the wall thickness, and the porosity of the grafts. Biological textile approaches allow the production of TEVGs, which can be clinically relevant and transferable to the industry. Several reviews have presented the promise of textile strategies for vascular tissue engineering [23][24][25][26][27]. However, they focused on textile-inspired approaches involving biodegradable plastics (polylactic acid [PLA], polyglycolic acid [PGA], or PCL) or biohybrid scaffolds, including denatured proteins. This review presents current biofabrication methods for vascular tissue engineering (figure 1), emphasizing the advantage of using biological materials, particularly for completely biological textile strategies.

The first TEVG
Tissue engineering is a multidisciplinary research field aiming to produce biological tissues by, most commonly, combining cells and biomaterials [28]. The first TEVG was developed in 1986 by Weinberg and Bell using a multi-step process [20]. Firstly, a bovine collagen gel was mixed with bovine smooth muscle cells (SMCs) and placed within a tubular mold. The tubular tissue was obtained after one week of maturation around a mandrel. Subsequently, a Dacron® mesh was slipped over the tissue to provide mechanical support, and a second gel was cast around the first, this time containing bovine fibroblasts. After two additional weeks of maturation, the tissue was removed from the mandrel and cannulated to inject a suspension of bovine endothelial cells within the luminal surface. Finally, a three-layered TEVG was obtained after one more week of dynamic maturation (rotation around a longitudinal axis at 1 rev min −1 ). While transmission electron microscopy images showed a well-organized endothelial layer covering the luminal surface, the TEVG had a burst pressure that was ten times lower than native arteries (≈ 300 mmHg). The mechanical properties of the vessel were entirely dependent on the synthetic mesh, as the burst pressure without the mesh was < 10 mmHg. Despite the poor mechanical properties of this model, they demonstrated the possibility of reproducing in vitro the general architecture of a living muscular artery with a mostly biological scaffold.

In situ engineering of vascular grafts
Although this may not be strictly a tissue engineering approach, another technique that relies on an extracellular matrix as a scaffold was developed even sooner in the late 1960s. This approach, often referred to as the 'Sparks' mandril' , is based on using the patient's foreign body reaction for vascular graft production [29]. In this method, a knitted Dacron® tube on a metal rod was implanted onto the patient's rib cage through a stab wound to elicit an encapsulating fibrotic response. The resulting graft embedded in fibrous collagenous tissue can be removed and used as an autologous vascular substitute. Various versions of this approach have been developed, including one using peritoneal cavity implantations without a synthetic scaffold. This method produced mechanically strong vessels in rats, rabbits, and dogs [30,31]. Arterial implantations in animal models have given relatively good results up to 6.5 months.
In 2020, Dr Rotmans's group showed a subcutaneous method for engineering an autologous fibrocellular tissue capsule that can be used as a vessel for arteriovenous grafting in a goat model [32]. Compared to the 'Sparks method' , which uses a metal rod and a synthetic knitted scaffold for the tissue capsule formation, they used only polymeric rods composed of a copolymer poly(ethylene oxide terephthalate)poly (butylene terephthalate) and [33] that were subcutaneously implanted in the neck of the animals for one month. In vitro characterizations of their completely biological vessels showed suitable mechanical properties for implantation (burst pressure: ≈ 2,400 mmHg and suture retention strength: ≈ 2N). In addition, ex vivo blood flow perfusion assay demonstrated that their graft can be rapidly endothelialized and subsequently be less thrombogenic than e-PTFE grafts. After two months of implantation as an arteriovenous connection, their tissue capsule graft and e-PTFE graft showed patency rates of ≈ 66% (n = 3) and ≈ 33% (n = 3), respectively.
Meanwhile, Nakayaka et al showed the first application of an autologous 'biotube' vascular graft as hemodialysis access in humans [34]. This 'biotube' was also engineered in situ by grafting two molds, assembled with a silicone rod and a stainless steel pipe, into the abdominal subcutaneous tissue of two patients for two months. The clinical follow-ups showed stenosis of both tubes after 3-4 months for the two patients and occlusion after 30 months for one patient. While avoiding the costly in vitro steps of classical tissue engineering, this approach has significant limitations, including risk for the patient during the implantation of the mandrel or the mold, reproducibility problems associated with patient-topatient variability, and difficulty in performing quality control during production.

Biodegradable synthetic scaffolding
Because of the limitations associated with nondegradable SVGs and autologous vessels, many research groups developed vascular grafts using alternative tissue engineering approaches in the late 1990s. One popular approach was to use degradable polymers (e.g., PLA or PGA) associated with living cells [35]. This approach has the advantage Association for the Advancement of Science. From [36]. Adapted with permission from AAAS. (B) and (C) Canine cell-derived 3 mm-diameter TEVGs (g) were implanted as carotid artery (ca) bypass and as a coronary shunt. (B) From [38]. Adapted with permission from AAAS. of having a chemically and physically well-defined matrix, which can be tailored to specific mechanical and geometric requirements. However, the downside of this approach is that the degradation of this polymeric matrix creates an inherently unstable tissue with changing properties over its maturation in a bioreactor.
Currently, Dr Niklason's group leads this approach and uses long-term culture (10 weeks) in a complex bioreactor (figure 3(A)) [36]. During the culture period, the biomaterial composed of a PGA mesh is largely degraded and replaced by extracellular matrix (ECM) produced by the cells themselves. In 2011, a decellularization step was introduced in the production process to engineer vascular grafts 'off-the-shelf ' [37]. Blood vessels made using this method have shown adequate mechanical properties, good mid-term results in animal studies, and promising results in two single-arm clinical trials (figures 3(B) and (C)) [37][38][39][40]. One of the limitations of this approach is that the subproducts of this synthetic matrix degradation create a microenvironment that negatively affects cell growth and ECM production [41]. In addition, incomplete degradation of this matrix could lead to adverse in vivo effects. Furthermore, the specific approach of Dr Niklason's group is based on using pooled early-passage SMCs from human aortas collected from deceased donors, which complicates cell sourcing by adding complex risk assessment and quality control issues. The next challenge for this approach will be to achieve a graft performance-to-price ratio that will support successful commercialization.

Biological scaffolding
Based on the previous works of Weinberg and Bell, in 1993, Dr L'Heureux developed a tubular vascular model exclusively composed of human cells and collagen [42]. This equivalent was formed with the three main cell types composing native vessels: SMCs, fibroblasts, and endothelial cells. Briefly, a medialike tissue was formed by seeding the human SMCs (2 × 10 6 cells ml −1 ) onto a tubular structure composed of a human collagen gel (3 mg ml −1 ) molded within a plastic tube containing a glass mandrel in its center. After two days, the SMCs had compacted the collagen around the mandrel. In the resulting empty space, an additional layer of collagen containing human fibroblast cells at a concentration of 5 × 10 5 cells ml −1 was added onto the outer part of the media tissue to form an adventitia-like structure. To recreate the endothelium layer, the tubular structure was slipped off the mandrel and cannulated at one extremity. Then, a solution containing human endothelial cells at 1 × 10 6 cells ml −1 was injected into the lumen of the vessel. Subsequently, the vessel was closed, placed within an incubator, and rotated. The artificial blood vessel was canulated and perfused in vitro for one to two additional weeks [42]. This first completely biological and human model triggered a strong enthusiasm for developing autologous TEVG for translational applications. However, this graft displayed relatively weak mechanical properties. Dr Tranquillo's group, which also developed vascular constructs using collagen gels with cells [43], developed more promising vessels by turning to fibrin gel [44]. To overcome the low mechanical strength of this type of construct, they chose a long-period culture approach (seven to nine weeks) in a complex pulsatile bioreactor to stimulate ECM production by the cells (figure 4) [45]. In addition, they adopted the use of fibroblast cells instead of SMCs to produce mechanically stronger vascular grafts [45,46]. They reported that vascular substitutes made from sheep cells had shown good mechanical resilience and good results after implantation in sheep for up to six months [47]. More recently, Vascudyne Inc. announced the successful first-in-human use of its TRUE TM Vascular Graft, which was produced using Dr Tranquillo's technology [48]. This first procedure was performed in ten end-stage renal disease patients requiring hemodialysis access, with a completion date of their phase 1 clinical trial estimated in 2022. Although these unpublished results are promising, this approach has limitations, such as using a complex bioreactor system exposed to contamination risks over a long period of culture and the gel contraction that could require very large culture bioreactors.

Cell-assembled extracellular matrix approach
A method that does not rely on the exogenous scaffold for TEVG production was published by Dr L'Heureux's group in 1998 [49]. This approach, sometimes referred to as tissue engineering by selfassembly, aims to produce a cell-assembled extracellular matrix (CAM) by channeling the ability of mesenchymal cells to secrete and assemble endogenous ECM in standard tissue culture flasks [49,50]. The accumulation of ECM on the culture substrate leads to forming a robust yet completely biological sheet (figure 5(A)). This was the first demonstration of a so-called 'scaffold-free' approach that produced mechanically strong tissues. To produce a TEVG, SMCs or fibroblasts were cultured in the presence of 50 µg ml −1 sodium L-ascorbate for approximately 30 day [49]. Both cell types formed cell sheets that could be detached from the culture flask and rolled around an inert mandrel (3 mm diameter) to produce a cylinder composed of concentric sheet layers (figure 5(A)). After additional maturation periods, the layers fused and formed a cohesive TEVG comprising vascular media and adventitia. After that, the TEVG was removed from the mandrel, and endothelial cells were seeded within the lumen. The resulting biological tubular structure displayed burst strength comparable to native vessels (≈ 2,500 mmHg), as well as good suturability and handling characteristics evaluated as 'tissue-like' by an experienced vascular surgeon [49]. Unlike matrices made of proteins chemically extracted from living tissue, the CAM has a native-like organization and composition, giving it physiological levels of mechanical strength and a potential for better in vivo integration/remodeling. These substitutes have been implanted in ten patients with end-stage renal failure and have shown a high luminal permeability rate (up to three years) [52,[55][56][57]. Clinical results support the hypothesis that a TEVG free of synthetic material can integrate with the native tissue, be remodeled by host cells, and resist infection. The durability of autologous TEVGs shows that the CAM does not trigger an innate immune response. Notably, non-living TEVGs composed of CAM produced by allogeneic fibroblasts were implanted in three patients without eliciting a specific immune response [55]. In agreement with the literature, the data demonstrated that allogeneic fibroblast remnants alone do not create a specific immune response [58][59][60]. Taken together, these open the door to a more economically viable strategy based on the large-scale production of allogeneic substitutes that can be stored until needed (off-the-shelf). Furthermore, recent data on CAM production variability between patients suggests that an allogeneic approach would significantly improve reproducibility compared to an autologous approach [61]. While a non-living allogeneic approach is an important step towards making production more commercially viable, TEVGs produced by CAM sheet rolling/fusion can be prohibitively costly and complex. The mechanical integrity of this TEVG depends on the fusion of sheet layers during a lengthy maturation process (eight to 12 weeks). With two steps of rolling/maturation, the production time of the TEVGs used in the clinical trial was about six to seven months. In addition, layer fusion is a cell-driven process that can be unreliable due to patient-to-patient variability and high sensitivity to culture conditions. Finally, the rolling approach of the CAM has limited ability to control the geometry and mechanical properties of the TEVG.

Cell aggregate assembly approaches
In 2010, Kelm et al (from Dr Hoerstrup's group) described a novel approach of manufacturing based Figure 5. Tissue-engineered vascular graft (TEVG) production process using scaffold-free approaches. (A) A sheet of cell-assembled extracellular matrix is produced and rolled around a mandrel to form a tubular structure. Then, a TEGV is obtained after a period of maturation. Adapted with permission from [52]. (B) Microtissues composed of cell aggregates can be generated using specific plates for spheroid formation. The microtissues are seeded within a mold and cultured in a bioreactor. After maturation, the TEGV is removed from the mold and ready to use. Adapted from [53], Copyright (2010), with permission from Elsevier. (C) The cells can be directly deposited into a tubule-like shape using a layer-by-layer bioprinting process. The final construct results in a TEVG that can be matured in a bioreactor. Adapted with permission from [54]. on cell aggregation for the production of smalldiameter scaffold-free TEVG (figure 5(B)) [53]. In their study, micro-aggregates composed of myofibroblasts and endothelial cells were produced in 60well plates. Then, the spheroids were loaded into the bioreactor and cultured for seven or 14 days under dynamic conditions and 14 days in static. Their results showed that unattached individual aggregates were still visible within the medium after seven days of dynamic conditions. A completely-fused tubular tissue was observed when the aggregates were cultured for 14 days in either dynamic or static culture conditions. To assemble a tubular tissue of 5 mm-length with a three-mm ID and a wall thickness of 1 mm, 4,000-5,000 aggregates were necessary. This implies the need for many cells, which can be a limit and makes an industrial translation economically challenging. In addition, one of the main issues is the lack of ECM formation after 14 days of culture and, consequently, mechanical strength.
Dr Alexis' group developed an interesting cell aggregate approach for vascular tissue engineering using magnetic forces to control tissue organization and assembly [62]. This approach requires the incorporation of magnetic nanoparticles (iron oxide) within the cellular spheroids made of cells (rat aortic fibroblasts or rat aortic SMCs) embedded within a bovine collagen gel. However, magnetic nanoparticle incorporation methods can involve cellular uptake, which can induce adverse effects on cell activity, viability, and phenotype [63]. In their first work, they described the development of a Janus structure of magnetic cellular spheroids (JMCSs), which can limit the nanoparticle internalization compared to a current uptake strategy (35% vs. 83%), improve the cell viability for at least seven weeks, and preserve cell phenotype [62]. Multiple steps are required for the biofabrication of a tubular structure using these spheroids. First, a magnet is used to assemble JMCS into a strip (two to three-spheroids-thick). After four days of fusion, the magnet is removed, and the tissue is wrapped as a tube around a glass cylinder containing magnets (5 mm-diameter). An additional maturation time of six days is required for the complete assembly of the tubular structure. While their approach requires a short two-step maturation time (ten days total), the spheroid fusion quality depends on the cell and the collagen concentrations [64]. For example, a combination of a low collagen concentration and a high cell number results in more fused tissue with better cellular intermixing over time compared to higher counterparts. However, dense collagenic tissue is necessary to obtain sufficient mechanical properties. They finally improved their process by adding a cyclic longitudinal stretching force using rod magnets (cyclic longitudinal translation at 1 Hz with a 10% magnitude [65]. This postprocess aimed to stimulate ECM deposition by cells during a third maturation step (three to seven days) and consequently enhance mechanical properties. While Dr Alexis's group has not shown, to our knowledge, a TEVG with clinically relevant dimensions and mechanical properties, their approach can theoretically allow the production of promising TEVGs. However, as for Dr Hoerstrup's approach, producing a clinically efficient TEVG would require a substantial quantity of cell aggregate and a longer stimulation time, which can limit an industrial translation.

The 3D bioprinting technologies
Bioprinting technologies are attractive for their advantageous characteristics, such as the capability to produce geometrically complex forms at high resolution and with an elevated reproducibility of the cell deposition process (figure 5(C)) [66]. To our knowledge, the first bioprinted scaffold-free vascular tissue was developed by Norotte et al in 2009 [67]. They reported using a fully biological self-assembly approach for bioprinting small-diameter vascular constructs (outside diameter: 0.9-2.5 mm). To realize the prototype, various vascular cell types, such as SMCs and fibroblasts, were used as cell aggregates or cylinders for the bioinks. These bioinks were extruded as spherical fragments on cylindrical rods (diameter: 300-500 µm) using a bioprinter, and agar-based rods were used as supports for the structure during the printing. Subsequently, multicellular cylinders composing the structure were fused within two to four days of maturation, and the supporting agar-based rods were manually removed. This method allowed the engineering of small-diameter vessel-like structures with control of the shape and the possibility of generating hierarchical trees composed of combined tubes of distinct diameters. However, their model showed an apoptotic pattern throughout the tissue after three days of post-printing maturation, which is directly related to a lack of nutrient and oxygen supplies within the core of the construct. In addition, these cell-rich constructs can only have limited strength due to the limited amount of ECM.
Kucukgul et al establish an interesting road map to develop a model of scaffold-free bioprinted macrovascular structure (figure 6) [68]. The first step requires procuring geometric and topological information on the targeted tissue using medical imaging techniques ( figure 6(A)). Then, a numerical model of the imaged tissue is generated using computer-aided design software ( figure 6(B)). Subsequently, the design can be numerically modified and optimized to obtain the desired structure for tissue reconstruction (figure 6(C)). Finally, the terminal numerical model is loaded to the bioprinter to coordinate bioink and support structure depositions ( figure 6(D)). Similar to the study of Norotte et al [67], their bioprinting method resulted in a lack of nutrient and oxygen diffusions within the construct because of the very high cell density and the presence of thick support pillars surrounding the tissue.
In 2012, Xu et al have biofabricated a tubular zigzag structure using a different bioprinting method, the scaffold-free ink-jetting process [69]. This printing modality uses a platform-assisted 3D inkjet system composed of a motorized XY stage, including a nozzle dispenser, a motorized Z stage where the structure is printed, a Z stage-fixed container with a solution of gelatinization (calcium chloride), and an imaging system. Using a pneumatic controller, the bioink is ejected from the computer-controlled nozzle dispenser as droplets in the direction of the Z stage, which is positioned within the container at the airliquid interface. The bioink was composed of NIH 3T3 mouse fibroblasts at 3 × 10 6 cells ml −1 resuspended in a 2% sodium alginate solution. After printing the first layer, the Z stage moves down vertically by a distance of the layer thickness (70 µm), resulting in the gelatinization of this submerged layer by the calcium chloride solution. Finally, the operation is repeated layer-by-layer until the obtention of the final construct, which stands on itself in the solution. Two main failures were observed in the process of biofabrication. Indeed, the first failure can occur after an imbalance between the buoyant and gravity forces due to the droplet impact at the maximal angled pivot point, resulting in the collapse of the construct. Another latter failure can occur if the thickness of the structure is too thin to resist the longitudinal deformation generated by the impact force of the droplets. Nonetheless, this promising bioprinting system allows the fabrication of complex forms, such as zigzag tubes, without the need for sacrificial supporting structures and directly within a solution gelatinization, which may supply the cells with nutrients and oxygen.
A different bioprinting approach, named the Kenzan method, also presents the advantage of not using sacrificial supports that can impact the final structure after removing it [70][71][72][73]. This method was inspired by the traditional Ikebana art, where the stems of floral arrangements are impaled in a domeshaped metal needle array called Kenzan. In Japanese, Kenzan means sword (ken) and mountain (zan). In the context of bioprinting, spheroids composed of cell aggregates are deposited on a microneedle array, which is used as a support (figures 7(A) and (B)) [71]. For the generation of spheroids, the cells or a mix of different cell types (cardiomyocytes derived from induced pluripotent stem cells, human umbilical vein endothelial cells, and normal human dermal fibroblasts) are seeded in non-adhesive round-bottomed 96-well plates in the presence of their appropriate culture medium. After 24 h, the cells have aggregated and started to secrete their own ECM to form large spheroids (400-600 µm diameter) that can be aspirated by a robotically-controlled suction nozzle (Bio-3D Printer, Cyfuse Biomedical K.K., Japan). Then, the spheroids are impaled in the needle and positioned to generate an initial tubule-like spheroid assembly that can be matured for seven days (figures 7(A)-(E)) [71,72]. Right after the initial phase of maturation, the spheroids are fused together to form a tubular tissue that can be removed from the needle array and positioned in a bioreactor for two additional days of maturation (figure 7(F)) [72]. The tubular tissue is perfused within the bioreactor using a perforated catheter connected to a roller pump that initiates a dynamic flow at a rate of 2 ml min −1 for two days that has been doubled for the two-terminal days (figure 7(F)) [72]. Finally, scaffold-free tubular tissue is obtained with personalized dimensions (length and ID) or shape. However, this method has some limitations, such as limited nutrient diffusion bioprinter is used to depose the spheroids on the needle array. The role of the needle array is to support the position of the spheroids in space during the printing process and maturation phases. (F) After an initial phase of maturation (seven days), the construct can be placed within a bioreactor. The construct is generally cannulated by a catheter with side holes and is perfused with a culture medium. (G) A scaffold-free tubular tissue is generated after four additional days of maturation within the bioreactor. Adapted from [72]. CC BY 4.0.
at the core of the spheroids and heterogeneous distribution of the cells within the spheroid. In addition, the construct showed poor tensile mechanical properties (939 mN), which was half of the native vessel value. While this promising type of graft has been implanted in the neck of immunodeficient pigs as a very short arteriovenous shunt (for three weeks, one month, and three months (n = 2 by time points)) [74], further preclinical investigations regarding its long-term in vivo effectiveness need to be addressed before clinical application.
More recently, in 2018, Maina et al have also taken advantage of a bioprinted model of a cylindrical vessel (figure 5(C)) [54]. This tissue was biofabricated using an Organovo dual-head printer, which can deposit the appropriate cell types, such as fibroblasts and SMCs, into a soluble support solution by microextrusions. The authors stated that they could print, in approximately six minutes, 10 cmlong vessel-like structures, which can be matured in 24-36 h. Their review also announced that preliminary data showed similar tensile and mechanical strengths between the printed structures and native blood vessels [54]. While this clearly announced the promising potential of this new generation of TEVGs produced by bioprinting, no mechanical or in vivo data were published yet to our knowledge. 6. Biological textile approaches for vascular tissue engineering 6.1. Biohybrid textile approaches Biohybrid textile approaches are based on the association of synthetic biodegradable polymer-based fibers with biological materials, such as collagen, to provide a proper environment for cell adhesion, migration, and survival.
Electrospun nanofiber-based textiles are widely developed for vascular tissue engineering [75][76][77][78][79]. To our knowledge, the first biohybrid techniques consisted of electrospinning a blend of synthetic and biological polymers onto a mandrel and exposing the construct to glutaraldehyde solution to improve its stability and mechanical properties [79]. From our point of view, it is difficult to believe that these chemically treated constructs preserved their biological advantages since these proteins would be encapsulated or degraded once implanted.
Dr Ramakrishna's group showed another electrospinning technique that did not include chemical cross-linking [80]. Indeed, they demonstrated that a plasma treatment of electrospun poly(L-lactic acid)co-poly(e-caprolactone) (PLLA/PCL; 70:30 ratio) nanofibers meshes allows collagen solution absorption by making the scaffold hydrophilic. This coated structure does not require chemical treatment to improve its stability, preserving the biological properties of the added protein (improved cell viability and attachment) [80]. In 2009, they implanted a collagen-coated nanofiber tubular structure (ID: 1 mm) for seven weeks to replace the inferior superficial epigastric vein in a rabbit model [81]. While they demonstrated that their graft could allow endothelial cell adhesion in vitro [80,81], surprisingly, no endothelial-like cells were observed at the inner surface of the graft after seven weeks. Despite this, they did not observe thrombosis sign and obtained good graft patency after seven weeks. However, animals were under anti-thrombotic treatment during this implantation time. These positive results should be taken cautiously because without graft endothelialization, ensuring the lack of thrombogenic events would be challenging after stopping the medicine.
Despite intense efforts in testing different process parameters and post-processing treatments, the mechanical ability of the scaffold to maintain structural and functional integrity immediately after implantation and during the early remodeling phases remains an issue. Combining electrospinning and fused deposition modeling represents a promising solution. For example, Dr Trombetta's group reinforced their model of heparin-loaded PLLA tubular scaffolds (ID: 5 mm, length: 6 cm) by extruding helical PCL rings on its outer surface [82]. As expected, this reinforcement significantly increased the mechanical properties of the graft (ultimate tensile stress [UTS]: 0.72 vs. 1.58 MPa or burst pressure: 0.08 vs. 0.30 MPa). While the burst pressure of this reinforced electrospun graft is twice that of a saphenous vein (0.30 and 0.16 MPa, respectively), it remains below the value of an internal mammary artery (0.42 mPa [83]). However, this armored graft still displays superior mechanical properties compared to other electrospun scaffolds with the advantage of using the biological effect of heparin. They finally evaluated the effectiveness of this armored heparinreleasing vascular graft in an aortic reconstruction model in rabbits [84]. They obtained good patency with no sign of thrombosis or structural failure after only four weeks of implantation without antiplatelet therapy. Promising results included a graft wall uniformly colonized by endogenous fibroblast-like cells and remodeled with the deposition of newlyform ECM. Moreover, the inner surface of the graft was covered with endogenous elongated cells, which may be endothelial cells. While further analysis must be conducted to confirm these results, this preliminary data showed the promising potential of these biohybrid vascular grafts.
Dr Jockenhoevel is one of the leaders of the biohybrid textile approaches. In 2009, his research group published the first model of small-diameter TEVG (ID: ≈ 5 mm) made of a cellularized biodegradable composite fibrin-polylactide scaffold without using electrospinning [85]. Indeed, they developed a warp-knitted macroporous mesh of poly(L=D)lactide 96=4, which was combined with an autologous fibrin cell carrier biomaterial. The warp-knitted tubular synthetic mesh was integrated into a mold containing a fibrin-based biomaterial loaded with arterial SMCs and fibroblasts. After 45 min of polymerization, the inner casting cylinder of the mold was removed for seeding endothelial cells on the luminal surface of the graft. Once the endothelial cell adhesion was completed (30 min), the graft was placed in a bioreaction under a pulsatile flow for 7, 14, or 21 days of maturation/ conditioning. Their results showed that the burst pressure of the graft slightly increases over the time of maturation but remains under the value of a native carotid (day 7: 399 ± 64 mmHg, day 14: 450 ± 90 mmHg, day 21: 466 ± 78 mmHg, and native carotid: 1,275 ± 122 mmHg). In 2010, they implanted their cellularized biodegradable composite fibrin-polylactide TEVGs (the 21 days-matured version) as interpositional carotid grafts in a sheep model for up to six months [86]. These autologous grafts gave promising preclinical data with good graft patency for up to six months with no sign of graft failure. In addition, histological analyses revealed excellent graft remodeling with mature autologous proteins, a homogeneous cell distribution through the graft wall, and a confluent monolayer of aligned endothelial cells into their luminal surfaces. While these biohybrid grafts performed well in vivo, their biofabrication process requires complex molding and bioreactor systems, which are costly and involve significant contamination risks. In addition, an autologous strategy including living cells can be a strong limitation for industrial commercialization.

Completely biological textile approaches
In 1994, Cavallaro et al, a collaborative work with the pioneering tissue engineering company Organogenesis, reported an innovative method for fabricating collagen threads [87]. These biological threads were produced by extruding bovine collagen (solubilized in 8.8 mM acetic acid) into a buffered solution of 20% polyethylene glycol (pH 7.55), where the collagen precipitated into filaments. While these threads were composed of collagen, only a small fraction appeared to be assembled as fibrils, and the resulting matrix did not display a native-like ultrastructure. Indeed, these biological threads showed a low UTS of 1.2 MPa (wet), and they had to be chemically cross-linked for eight hours in a 50 mM ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride solution to obtain a stronger biomaterial (UTS of 23.9 MPa) that were compatible with textile approaches. These cross-linked threads were used to fabricate impressive fabrics using textile methods such as knitting, weaving, or braiding. Braided and bundled structures composed of collagen threads were used to replace an anterior cruciate ligament in a dog model. While an inflammatory response was immediately identified after implantation, the immune reaction was said to be resolved after 12 weeks, although little evidence was provided to support this statement. Nonetheless, the implants provided the necessary mechanical strength and led to the production of new collagenous tissue that was similar to a native ligament. They also demonstrated that knitted collagen threads could be used to repair herniation after 12 weeks of implantation in a rat abdominal repair model.
In 2008, Zeugolis et al evaluated the influence of cross-linking methods (chemical, physical, and biological) on the mechanical properties of extruded collagen threads [88]. Their results demonstrated that bifunctional agents, such as epoxy and hexamethylene diisocyanate (HMDC), improve the mechanical resilience of wet collagen threads (stress at the break; untreated: 3 MPa, epoxy: 30 MPa, and HMDC: 17 MPa) by the creation of linear cross-linking within the collagenous matrix. Cornwell et al also investigated the effects of various cross-linking strategies on the mechanical properties and the in vitro rate of new tissue ingrowth on bundle threads extruded from type I collagen [89]. Their results indicate that while physical cross-linking techniques (dehydrothermal or ultraviolet light) increase thread mechanical strengths, these treatments significantly reduce the migration rate of fibroblast cells. Only carbodiimide chemical cross-linking achieved sub-optimal strength generation with a similar cell migration rate compared to uncross-linked threads. While these treatments interestingly enhanced the strength of the threads, these cross-links are not natural, and the tissue would be expected to be perceived as a foreign material by the innate immune system. While this treatment has the added advantage of making the implant harder to degrade, it creates the conditions to generate a chronic inflammatory environment. While this may be acceptable for some applications where a strong fibrous/scarring response is desired (e.g., hernia repair), this is not suitable for small-diameter vascular repairs.
In 2014, Younesi et al developed interesting biological threads using electrochemically aligned collagen (ELAC) fibers that enable the fabrication of woven biotextile scaffolds [90]. These threads were fabricated by a kinematically rotating linear electrode pair and then chemically cross-linked for three days (0.625% genipin in 90% v/v ethanol solution at 37 • C). Their results demonstrated that a reduction in the thread diameter increases the UTS (150 µm ⊘: ≈ 21 MPa, 130 µm ⊘: ≈ 35 MPa, and 100 µm ⊘: ≈ 51 MPa). By twisting three threads as a yarn, they provided a significant gain of the UTS (≈ 65 MPa) to reach values comparable to that of the native tendon. In addition, they showed that these threads support cell adhesion and differentiation (tendogenesis), as well as ECM deposition in vitro. Recent in vivo biomechanical, histochemical, and immunohistochemical metrics showed interesting outcomes after three months of functional loading of this ELAC scaffold in critical-size tendon defects (New Zealand white rabbits) [91]. ELAC threads have been originally developed for tendon repairs. More recently, these collagenous filaments have been used to produce a biohybrid vascular graft using the circular knitting method [92]. The ELAC threads were assembled together with traditional PLA yarns to reinforce the structure. The collagen component promoted endothelial cell adhesion and proliferation in vitro, and the PLA provided sufficient mechanical properties and structural stability. Once knitted to form a tubular structure, the vessel exhibited excellent bursting strength (≈ 2 MPa), suture retention strength (≈ 11 N), and compliance (≈ 4%/100 mmHg). However, the knitting method generated a structure with a ≈ 250 µm pore size. While a high porosity may facilitate cell infiltration necessary for tissue remodeling, this macro-scale pore size will lead to a lack of hemostasis and ultimately result in unacceptable blood leakage. In addition, the in vivo success of this type of graft will depend on the right balance between degradation and ECM neoformation since it is composed of biodegradable synthetic PLA and chemically-denatured collagen. But again, and maybe more importantly, unlike in soft tissue repair (ligament/fascia), the vascular environment does not cope well with inflammatory reactions. The inflammation triggered by these foreign materials may compromise the vascular graft function by causing thrombosis and intimal hyperplasia.
In 2019, Zhang et al used knitting to engineer a bilayer vascular graft (figure 8) [93]. This graft was initially made of knitted filaments composed of type I collagen extracted from rat tail tendons. However, this assembly showed an extremely high porosity (pore size: 427 µm), so rat collagen nanofibers were electrospun to cover the vascular graft (pore size: 2 µm). Once hydrated, this covered graft showed supra physiological mechanical properties (bursting pressure of covered graft: 0.66 ± 0.09 MPa, human internal mammary artery: 0.42 ± 0.17 MPa, and the human saphenous vein: 0.21 ± 0.12 MPa [83]). While this graft demonstrated superior cytocompatibility compared to a knitted PLA-based synthetic graft (human umbilical vein endothelial cells expanded 4.5-fold from day one to day nine after seeding), the processed collagen will trigger inflammation and be rapidly degraded once implanted.
Using a different biofabrication process, Onoe et al showed a technique for producing core-shell hydrogel microfibres, which include encapsulated ECM components and living cells [94]. This technique involved a double-coaxial laminar flow system [95], which allows the fabrication of completely biological meter-long living microfibers. Indeed, this microfluidic system will shape the fiber by exposing the core (ECM protein gel and cells) and the shell (Na-alginate solution) to a sheath (CaCl 2 solution). Then, the shell will be polymerized, and the fiber will be continually extruded. Finally, the shell can be selectively removed by enzymatic digestion after a maturation time to obtain living fibers only composed of cells (e.g. fibroblasts, myocytes, endothelial cells, nerve cells, or epithelial cells) and ECM components (pepsin-or acid-solubilized collagens or fibrin). The viability assay showed elevated cell survival (> 90%) after fiber extrusions. They also showed that these living fibers are compatible with a woven approach. A total of 2.5 m of the cell fiber was used to weave a planar fabric-like structure that can be further processed to create a 3D woven structure (approximately 5 mm-width and 5 mm-height). They also presented a fiber assembly method using a double helical coil (approximately one mm in diameter and approximately 10 mm-length). Using this approach, they can assemble two different cell fibers by reeling them with a glass rod. The structure of these helical cellular coils would enable the control of the cell orientation within this tubular structure. While this weaving approach could be effective in creating woven tubular structures, these living fibers may not have the mechanical strength to be assembled in a graft that could be implanted in the arterial circulation.
Also inspired by the textile industry methods, Dr L'Heureux's group is developing a woven TEVG model using CAM-based threads, which is completely biological, human, and displays appropriate mechanical properties [96]. Compared to a rolled/fused CAM sheet approach (presented in section 5.1), a textile method reduces the TEVG production time 3-fold by eliminating the maturation step (fusion) of the sheets and allowing precise adjustment of its mechanical properties. Among the main textile industry methods, weaving has been identified as being the most promising for its ability to generate tubular structures with low porosity walls (unlike knitting grafts as engineered by Zhang et al [93]) and whose diameters are not drastically affected by the longitudinal tension (unlike braiding) (figures 9(A) and (B)) [96]. In this process, a set of longitudinal CAM-based threads can be woven with a single circumferential thread to generate a human TEVG that was sutured to the carotid of a sheep (figures 9(C)-(H)). The graft was leak-proof and allowed to restore normal blood flow immediately post-implantation. This process greatly improves the applicability of the CAM for TEVG production by developing a faster, cheaper, and more versatile method. Based on the clinical results of TEVGs made with rolled CAM sheets, the host immune system will accept these woven CAM-based TEVGs and completely integrate them with the native tissue. Recent data showed that human CAMbased materials are long-lasting (up to six months) and, unlike processed collagen implanted in parallel, do not trigger the innate immune response in nude rats [60]. Unpublished data shows largely untouched CAM material one year after implantation in the aorta of nude rats. Advancements in CAM production made the transition to sheep possible, and ovine CAM sheets with mechanical properties similar to human CAM sheets have been produced [97]. This will allow evaluation of the effectiveness of the woven CAMbased TEVGs in a clinically relevant allogeneic context for long-term studies in large animals to justify translation to humans. A recent approach to produce human woven TEVG, which is more economical, uses biological yarns made from human amniotic membranes [98].

Discussion
LeMaitre company is one of the leading actors in biological vascular graft commercialization with two products on the US market (Artegraft® and ProCol®). These products are Food and Drug Administration (FDA)-approved vascular grafts for hemodialysis accesses and low-extremity bypasses. Another company, Cryolife, Inc., is accredited by the American Association of Tissue Banks and approved by the FDA for the commercialization of cryopreserved and decellularized human cadaver vascular allografts (Cryovein®) and decellularized bovine ureter (Syn-erGraft®). Although these xenogeneic and allogeneic grafts have been used in patients, their adoption by surgeons has not been established since revues, and clinical studies showed that they did not demonstrate significant improvement in patency compared to their synthetic counterparts [99][100][101][102][103][104]. In addition, these grafts are strongly susceptible to thrombosis and infection. Since these grafts are obtained from native vascular tissues, their shape, size, and geometry are limited and vary between grafts making quality controls relatively challenging. While several publications call these grafts TEVGs, this is arguable since they are composed of xenogeneic or allogeneic decellularized natural matrices, and their structures have not been engineered.
Developing a small-caliber arterial substitute with a tissue engineering approach represents a critical challenge in cardiovascular surgery to provide a functional conduit for coronary and peripheral bypass when autologous vessels are unavailable [105]. Before attempting the commercialization or the clinical translation of TEVGs, they must have specific characteristics to withstand the vascular environment. The ISO 7198:2016 standard describes the methodology for validating various properties of vascular prostheses [106]. The parameters evaluated include the structural characteristics of the wall (thickness, composition, porosity, transmural permeability, and diameter), mechanical properties (tensile strength, suture retention, compliance, and bending radius), and biocompatibility. This standard has been established for synthetic vascular prostheses. However, it does not consider the immunological aspect of graft integration in the host, which is particularly important for tissue reconstruction. It also does not specify minimum or acceptable values for each parameter under evaluation. While the mechanical and physical properties of TEVGs are important, endothelialization is a crucial process for long-term TEVG effectiveness [107], and its consideration is not part of the ISO. Most of the TEVGs presented in this review were not endothelialized (table 2). Indeed, this process makes more complex the biofabrication by adding an important cell culture step in a bioreactor. In addition, endothelialized grafts are not compatible with an 'off-the-shelf ' production which would limit their industrial development. While scaffold-based and scaffold-free TEVGs showed promising mechanical and/or biological results, the time and cost of production, as well as governmental regulations, continue to challenge their development (table 2).
In the US, before getting authorization for new therapeutic product commercialization, a new drug application must be submitted to the FDA after completing a Phase III clinical trial. To our knowledge, only five biotech companies have succeeded in bringing TEVGs to humans and are closer to commercialization. Cytograft Tissue Engineering, Inc., founded in 2000 by Dr L'Heureux, was the first company to implant a TEVG (Lifeline TM Graft, made from CAM) in humans as an arteriovenous shunt. While they obtained successful results in Phase I and II clinical trials, unfortunately, the company closed before realizing a phase III study. Founded in 2004, Humacyte, Inc., used Dr Niklason's TEVG biofabrication approach. Their human acellular vessel, based on a biodegradable synthetic scaffold remodeled by cells in culture, was implanted in humans for the first time in 2013 and is now in phase III clinical trial initiated in 2017 to ensure its efficacy and safety as autologous arteriovenous fistula in approximately 240 patients. This company is the most advanced in clinical development and the closest to commercializing its TEVG model. Recently, Vascudyne, Inc., which uses Dr Tranquillo's technology for producing a TEVG (TRUE TM Graft), has completed a phase I clinical trial including ten patients to test the preliminary safety of their graft for hemodialysis access. Finally, Biotube Co., Ltd, and Xeltis, which industrialized in situ engineering and electrospinning approaches, respectively, have also implanted small-diameter vascular grafts in humans. The path from the bench to the commercialization of tissue-engineered products has become longer and more expensive than it was first envisioned by pioneers who expected that a biological solution would receive more support than plastic implants.
Biological textile approaches (weaving, knitting, braiding, or crocheting) can be automated, allowing rapid and reproducible production, and facilitating the commercialization of these products (table 2). This automation can be realized based on existing machines used in the textile industry that are less complex and better defined than bioreactors. In addition, this approach provides excellent control over the mechanical properties of the TEVGs, including directional and regional control. This allows, among other things, the production of tapered TEVGs, which are clinically relevant. Unlike the rolled CAM approach, it will enable the control of wall permeability/density, which can influence, for example, the flexibility of the graft and its ability to be colonized by cells. However, these approaches require using compatible biomaterials that are stable and strong enough to resist the biofabrication process. Finally, biological textile approaches are not limited to the creation of TEVGs. They can also lead to innovation in many other applications, such as a simple non-inflammatory  [96] suturing material, a ligament, or a hollow structure to support organ development.

Data availability statement
The data that support the finding of figure 2 are available upon reasonable request from the authors.