Abnormal blood flow in the graft has long been known as a key impairment factor that alters graft healing and overall patency (Schwarz et al., 2021). Computational flow dynamics tools, with the inputs of graft location, diameter, length and/or compliance together with graft design, have been applied to predict not only flow conditions but also consequent remodeling of a graft (Lu et al., 2014; Ramachandra et al., 2017; Donadoni et al., 2020). Variations in these physical characteristics of a graft delicately influence local fluid stresses in a blood vessel or a graft. Abnormal hemodynamic stresses may further result in impaired graft endothelialization, impaired graft healing and endothelial disruption in neighboring vessels (Chistiakov et al., 2017). Irregularities in the wall shear stress are known to induce endothelial dysfunctions, leading to negative clinical outcomes such as intimal hyperplasia, wherein smooth muscle cells (SMCs) grow to restrict the lumen of a vessel or a graft (Chistiakov et al., 2017; Ng et al., 2017; Totorean et al., 2019). Interestingly, in the case of dialysis access graft, hyperplasia develops predominantly in the vein or anastomotic regions, particularly toe and heel of distal anastomosis, where flow disturbances and altered hemodynamic factors are most prevalent (Sottiurai et al., 1989; Bassiouny et al., 1992). Quantitatively, low shear stress values (<.5 Pa) or oscillatory values (±.1–.2 Pa), as opposed to physiological values (1–2 Pa), were shown to induce endothelial dysfunction and subsequent hyperplasia development (Dhawan et al., 2010). The graft location determines flow environments around a graft (Chiu and Chien, 2011); abnormal flow patterns might occur in arteries with branches or curvatures or grafts such as looped vascular access or brachial arterial grafts as compared to aortic grafts.

The graft diameter is key to reach a hemodynamic balance. Bypass grafts with unfavorable diameter ratios to the original vessel can result in competitive flow, ultimately leading to graft failure. Competitive flows occur when a partially stenotic blood vessel is bypassed with a graft, but still retains some of its throughput, thereby competing with the newly grafted path and causing irregular flow through it (Gaudino et al., 2016). A high competitive flow decreases the amount of blood that reaches the graft, creating low shear stresses that increase the levels of hyperplasia at the graft’s anastomosis (Jin and Liu, 2019). Clinical findings suggest that more stenotic vessels with lower competitive flow, compared to less stenotic vessels, were better tolerated by coronary bypasses, and arterial grafts were more susceptible to competitive flow-induced impairment than venous grafts (Nordgaard et al., 2010; Gaudino et al., 2016; Sabik Joseph, 2016). The presence of chronic competitive flow is inversely related to internal mammary graft patency (Binns et al., 1989; Sabik et al., 2003). Intriguingly, saphenous vein graft patency did not decrease with increasing competitive flow (Ding et al., 2012; Swillens et al., 2012; Sugiura et al., 2018). Such difference could be attributed to cell and tissue differences between venous and arterial grafts (Harskamp et al., 2016; Sabik Joseph, 2016). Thus, the graft diameter in comparison with the original vessel is an important consideration for reducing flow-induced impairment, in particular for some graft types.

The roles of graft compliance and length in determining the graft patency are also well known (Skovrind et al., 2019). Compliance mismatch was recognized as a major contributor to graft failure (Tiwari et al., 2003). Though the influence of graft length on patency rates was insignificant for a similar grafting application (Tinica et al., 2018; Bosiers et al., 2020), applications that require long grafts, such as femoral artery lesion and hemodialysis access, often yielded low patency rates (Connors et al., 2011; Boutrous et al., 2019). This may be attributed to a wide range of biomechanical forces across a long graft with a large surface area. These forces consist of a complex loading combination from varying amounts of radial deformation, bending, extension/contraction, and torsion/flexion, which significantly impact the cell-graft interactions.

The graft failure in the cases of coronary bypass, dialysis access, or stent graft have been variably associated to age, diabetes mellitus, hypertension or other health conditions (Gaudino et al., 2017; AbuRahma, 2018; Misskey et al., 2018). Sex may serve as a possible compounding factor (Blum et al., 2021). Diabetes mellitus in particular causes a high risk of platelet aggregation in the blood and increases the release of von Willebrand factor, which promotes platelet aggregation and results in significant endothelial damage and arterial stiffening (Creager et al., 2003). This vascular intima damage coupled with the pathophysiological mechanism of diabetes leads to the increased formation of thrombosis (Reddy et al., 2013), and thus sooner obstruction of a graft such as arteriovenous fistula (Afsar and Elsurer, 2012). The primary failure of fistulas is associated with an enormous overall cost. Interestingly, arteriovenous fistula showed even higher primary failure rates than synthetic grafts in patients with diabetes mellites (Qin et al., 2016). To make it worse, failure of a fistula reduces the number of possible sites for an alternate access, exposing the patients to risks from salvaging attempts (Schinstock et al., 2011). Therefore, using hemodialysis access for diabetic patients as an example, it is clear that patients’ health conditions complicate cellular and molecular environments around grafts, worsening the ultimate fate of grafts, in particular biological grafts. For above-knee femoropopliteal bypass, diseases including diabetes and ischemia significantly reduced graft patency in both vein and PTFE grafts, but showing similar reduction rates for both types of grafts (Klinkert et al., 2004). Furthermore, during atherosclerosis such as coronary artery diseases, systemic biological factors (e.g., diabetes mellitus, dyslipidemia, sex, age, hypertension, and smoking) and local biological mechanisms (e.g., increased oxidative stress, vascular inflammation, and endothelial dysfunction) all contribute to the initial plaque formation and subsequent progression (Gaudino et al., 2017). These classic cardiovascular risk factors, particularly diabetes mellitus, seem to play a role in determining graft failure. For instance, the impact of diabetes mellitus on arterial and venous bypass grafts has been investigated by several groups, with most but not all, showing a detrimental effect on graft patency (Gaudino et al., 2017).

Besides comorbid diseases, the patency of a biological graft, compared to an inert synthetic graft, might be more influenced by cellular and molecular environments associated with older age. Older patient’s arteries tend to be more easily calcified making them more prone to abnormal shear stress, inflammation and other endothelial dysfunction (Tesauro et al., 2017). Due to significant comorbidities in elderly patients, a vein fistula is more likely to fail due to the deteriorating condition of blood vessels (Miller et al., 1999; Ponce, 2001; Lazarides et al., 2007). The use of synthetic PTFE grafts instead of vein fistula in elderly patients and those with compromised vessels might improve primary patency rates (Staramos et al., 2000; Rooijens et al., 2005). Clinical data point to advanced age as a significant factor in adversely affecting outcome of dialysis graft (Sugimoto et al., 2003; Moist et al., 2012). Meanwhile, almost half of all patients initiating dialysis are 65 years of age or older in the United States and the percentage of elderly patients is even higher in Canada and Europe (Ahmed and Catic, 2018), which makes the age as an important consideration for the graft design. It is increasingly recognized that an advanced age is associated with adverse graft remodeling and differentially altered remodeling of vein fistula and synthetic graft, likely through aged-related changes in vascular cells and molecular environments initiated by increases in pro-inflammatory cytokines, pro-calcification signaling, advanced glycation end products, and/or extracellular matrix stiffness. Similarly, when a bioengineered graft or TEVG include a cellular and/or molecular component, its design should consider adverse environments related to clinical conditions such as aging. In the case of vein fistula for dialysis access, its risk of failure is greatest among older patients and/or obese patients (Lok et al., 2006; Peterson et al., 2008; Shingarev et al., 2011).

Graft bioinertness, or lack of biointegration, is another major impairment factor for long-term graft patency. It is mainly related to commercial grafts such as those made of PTFE or Dacron, but could be variedly applicable to different TEVGs, particularly prior to the completion of vascular tissue regeneration. Previous review articles have elaborated the main issues with the graft inertness, which include thrombosis, inflammation and intimal hyperplasia of a graft. Infection can also be a serious issue (Kirkton et al., 2018), in particular for dialysis graft requiring continuous needle access (Halbert et al., 2020; Kostokis and Loukopoulos, 2020; Stegmayr et al., 2021). A generally accepted root cause of these issues is the poor interaction of cells and the non-specific adsorption of proteins onto the graft, which lead to platelet aggregation but fail to form organized endothelium and/or matured smooth muscle. To circumvent these issues, bioengineering strategies including chemical coatings, materials modifications with peptides, proteins or other biomolecules, and seedings of endothelial cells or stem cells on inert or degradable grafts, have been utilized, with the goal of better integrating grafts with the native vasculature (Hielscher et al., 2018; Wang et al., 2020; Ham et al., 2021; Liu et al., 2021; Yu et al., 2021). Results with these strategies demonstrated improved endothelialization and/or inhibited thrombosis, inflammation, intimal hyperplasia, or infection. More recent approaches involve degradable, bioactive materials, as described below.

The contribution of cell technology to counteract adverse remodeling is to enhance graft acceptance and integration with native vessels, including attenuating chronic inflammation, reducing thrombogenicity, and promoting fast healing and vascular regeneration. The cell technology includes the technology to isolate, culture and use different types of cells for vascular graft applications, the technology to recruit cells into grafts, and the technology to produce cell products crucial for regeneration. Herein, a brief overview is taken on each cell type used for vascular grafting, highlighting recent innovations in tissue bank, immune-tolerant cells, and bioreactors. Recent cell technology has been expanded to include the use of cell secretomes, while on the other end adding a range of in situ or in vivo cellularization technique into the realm of cell technology.

Autologous matured vascular cells, such as SMCs, endothelial cells and fibroblasts (McAllister et al., 2009; Syedain et al., 2011; Amensag and McFetridge, 2012), were employed for the TEVG creation in early preclinical and clinical studies. They are not immunogenic, but their disadvantages include invasive harvest of vascular cells, limited availability, varied quality, and low potential of replication and regeneration.

Progenitor cells, compared to matured cells, are isolated using less invasive procedures, from blood, bone marrow, umbilical cord, adipose or other tissues, and they demonstrate greater proliferative capacities despite their scarcity in tissues. For example, bone marrow-derived smooth muscle progenitor cells, compared to adult vascular SMCs, produced stronger and tougher TEVGs in vitro, and consequently progenitor seeded-grafts produced more organized elastin when implanted in vivo as jugular replacements in lambs (Swartz et al., 2005; Liu et al., 2007). Similarly, endothelial progenitor cells have been used to seed TEVG (Melchiorri et al., 2016; Muniswami et al., 2020; Tamma et al., 2020), but their high heterogeneity and rare presence could complicate the translation process (Munisso and Yamaoka, 2020). Mesenchymal stem cell (MSC), a type of progenitor cell with the potential of generating various vascular cell types (Gong and Niklason, 2008), lacks major histocompatibility complexes or other immune-stimulatory molecules while secreting anti-thrombotic and anti-inflammatory molecules, such as IL-10 (Fukunishi et al., 2018; Mirhaidari et al., 2020), and TGF-β receptor 1 (Lee et al., 2016). These characteristics make MSC an excellent allogeneic cell candidate for the TEVG production (Wang et al., 2016; Afra and Matin, 2020). In particular, adipose-derived MSCs are advantageous for their easy extraction in high quantities and high regenerative potency in a way unaffected by age (Harris et al., 2011; Zhang et al., 2011; Krawiec et al., 2016; Krawiec et al., 2017; Haskett et al., 2018). Finally, other new types of progenitor cells such as pericardial effusion-derived progenitor cells (Wang et al., 2021) can be potential candidates for cellularizing vascular grafts.

Induced pluripotent stem cells (iPSCs) have the potential of generating patient-specific SMCs, endothelial cells, and fibroblasts for TEVGs (Generali et al., 2019; Luo et al., 2020; Shi et al., 2021). The availability of iPSC-derived cell sources may eliminate the problems with variation in the cell quality and address the timing issue for “off-the-shelf” availability, both of which are concerns relevant to progenitor cells. Nevertheless, a bottleneck of using human iPSC derivatives is their tumorigenicity and immunogenicity (de Almeida et al., 2013). Thus, human leukocyte antigen-matched iPSC tissue banks would be a valuable source for personalized grafts (Jin et al., 2016; Petrus-Reurer et al., 2021). Additionally, genetically engineered master human iPSC lines, which give rise to immune-tolerant (immune-evasive or immune-suppressive) iPSC derivatives (Riolobos et al., 2013), might yield universal “off-the-shelf” cell sources for hypo-immunogenic TEVGs.

Cell seeding in vitro often requires an incubation time ranging from days to months in a dish or in a sophisticated bioreactor. Bioreactor designs vary from a rotary bioreactor for stem cell differentiation (Xing et al., 2017; Li et al., 2019a), a closed apparatus for standardizing cell technology (Kurobe et al., 2015), a biaxial stretching system for tissue matrix maturation (Huang et al., 2015; Huang et al., 2016), a semi-automated cell seeding device (Cunnane et al., 2020a), to a perfusion bioreactor for applying flow stresses (Melchiorri et al., 2016) and individualized graft conditioning (Håkansson et al., 2021). Another recent innovation in the in vitro cell seeding technology is to exploit 3D bioprinting techniques for printing patterned cell-laden hydrogels, utilizing a mixture of cells and hydrogel as the bioink to create vascular grafts with precisely-defined structures.

Recently, in situ or in vivo cellularization, utilizing human (or animal) body as a bioreactor, added a novel option into the cell technology for vascular grafts. This approach starts with placing a tube composed of Teflon (Wang et al., 2019a; Qiu et al., 2021), silicone (Fujita et al., 2020) or collagen (Nakayama et al., 2020), as a subcutaneous or peritoneal cavity implant or in an ex vivo location, allowing fibrous tissues to grow and form an tubular structure, which is then explanted and subsequently implanted as an autologous graft (Chen et al., 2019; Nakayama et al., 2020) or as an allograft after decellularization (Wang et al., 2019a; Fujita et al., 2020; Qiu et al., 2021). Preclinical or clinical studies showed some encouraging outcomes.

Last but not least is that the future cell technology for vascular grafts will surely see an increased use of stem cell secretomes (Chen et al., 2018; Cunnane et al., 2018; Cunnane et al., 2020b), as their therapeutic and regenerative values are being unveiled (Ahangar et al., 2020; Baruah and Wary, 2020; Nikfarjam et al., 2020; Tang et al., 2021) and recently exploited in other cardiovascular implants (Li et al., 2019b; Baruah and Wary, 2020; Hu et al., 2021).

The scaffold technology encompasses the techniques to improve the graft structure and mechanics and those to improve the biological integration. The former includes design innovations involving novel computational tools, biopolymers, or scaffold fabrication techniques ranging from decellularization to electrospinning and 3D printing. The scaffold technology for improved biological integration often involves the incorporation of a range of bioactive functional biomolecules on the surface or within a biopolymer, in order to modulate immune responses and/or to promote regeneration.

Related to tubular vascular implants, FDA guidance documents have specified testing criteria for a number of physical and mechanical properties of grafts, including porosity, burst strength, compliance, elastic modulus, water permeability, compliance, and kink diameter (ANSI, 2010). Vascular grafts for a specific application such as arteriovenous graft access may require additional guidelines for function testing (Vascular Access Society, 2021). For example, when the synthetic grafts are placed a loop configuration, e.g., for vascular access, preventing graft kinking is crucial (Wu et al., 2020). Nevertheless, all the test guidelines refer to traditional, permanent graft materials. When biodegradable materials, synthetic or natural, are used to construct grafts, it may be necessary to test the temporal changes of physical and mechanical properties, in order to avoid devastating events such as aneurysm. Besides physical and mechanical properties, in vitro device tests also include biological function assessment. As TEVG improvements focus on the graft development by attending to the main causes of graft failure—thrombus formation and intimal hyperplasia, in vitro thrombogenicity tests as well as proliferation and migration tests of vascular cells have been widely employed.

The goal of TEVG regeneration studies is to translate new findings to human clinical therapy, which relies on animal models to generate enormous data useful to the translation. Ideally, the selection of animal model considers the animal’s analog close to human patients in terms of similarities in anatomical, structural, pathological and/or functional aspects. Immunodeficient animal models further allow long-term accommodation of humanized vascular grafts (Itoh et al., 2019). To evaluate the biochemical, structural and functional similarities between a TEVG implant and a native artery in animals, measures taken on the implants include ultrasonic scan, flow Doppler and X-ray/angiography, as well as on explant analyses, such as mechanics (i.e., strength, compliance), histological analyses such as geometry (i.e., lumen size, thickness), layer structure, cells (e.g., inflammatory cells, circulating cells, vascular endothelial cells and SMC, cells related to fibrotic or calcified remodeling), ECM (e.g., elastin, collagen types I/III/IV), and biomolecules through gene and/or protein assays.

A closer analog to clinical conditions allows one to better predict graft healing and regeneration in clinic. A huge discrepancy exists between endothelialization outcomes in animal models and those occurring in clinic (Zilla et al., 2007; Sánchez et al., 2018; Cai et al., 2021). Despite increased graft assessments in different animal models, most studies still use young, healthy small animals. It could be attributed to drastic differences in the vascular anatomy and physiology between human and rodents, which are used for evaluations in the majority of animal studies. The re-endothelization of a vein graft was reported to complete after several weeks of post-surgery in rodents, but re-endothelization duration in human took much longer. More important than the late presence of an endothelium on TEVG in human is a mismatch in the endothelialization mechanism—mainly trans-anastomotic mechanism in short grafts for rodents versus mainly transmural or blood-borne endothelialization in much longer grafts for humans (Zilla et al., 2007). In the clinical context of PTFE grafts, trans-anastomotic endothelialization covers only the immediate peri-anastomotic graft region in patients. Based on the analysis in a senescent non-human primate model, Zilla et al. (2007) and Zilla et al. (2020), concluded that a major issue with the development of prosthetic vascular grafts lied in wrong animal models, which led to wrong questions and no healing. Large animal models such as sheep, pigs, and baboons, though less available and more costly, mimic human physiology, in terms of vascular anatomy, thrombogenicity, inflammatory responses and regenerative potentials (Fukunishi et al., 2016; Rothuizen et al., 2016; Ju et al., 2017; Ma et al., 2017; Syedain et al., 2017; Liu et al., 2018; Schleimer et al., 2018; Fang et al., 2021b).

Overall, the design of future preclinical studies would better address the clinical relevance, for example, dialysis access (Gage and Lawson, 2017; Ong et al., 2017) and meso-Rex bypass (Maxfield et al., 2017), by providing further insight into the mechanisms through experimental designs that are aligned with the purpose and scope of specific clinical applications. In selecting animal models that prematurely terminate adverse remodeling events, the factors that impair vascular regeneration around grafts in clinic may remain largely overlooked. Besides the type of animal models, the significant “impairment factors” as listed in Section 2 include age and disease conditions, e.g., diabetes and ischemia. Additionally, temporal and spatial characterizations of the host responses in animals would be crucial to improve our mechanistic understandings of TEVG remodeling.

Clinical trials are a true success measure and a determinant to the commercial value of any graft innovation. Prior trials have been performed on three types of grafts: Biological grafts (i.e., xenografts or allografts), bioreactor-manufactured grafts, and readily implantable grafts. Though trial results are encouraging and enlightening for future TEVG strategies, the overall success is still limited.

Biological grafts, referring to acellular blood vessels from human or animal (e.g., bovine) origin, have undergone clinical trials for decades. Bovine carotid artery grafts, such as those commercialized as Artegraft™, were used as arteriovenous fistula in patients undergoing chronic hemodialysis. These grafts outperformed PTFE grafts, though incrementally (Hutchin et al., 1975; Kennealey et al., 2011; Arhuidese et al., 2017). Lindsey et al. (2018) did a 15-year follow-up study on Artegaft used in lower extremity bypass surgery of 120 patients, yielding positive results for patency and salvage limb rates. Recently, the concern regarding the immunologic sensitization of artery xenografts in patients was also addressed (Dyer-Kindy et al., 2020). Similar to an artery graft, a vein xenograft such as bovine vein-derived ProCol™ demonstrated their candidacy as an alternative to synthetic grafts through clinical trials in both hemodialysis patients and patients with critical limb ischemia for 14–20 months (Hatzibaloglou et al., 2004; Schmidli et al., 2004). Vein allografts derived from human cadaveric veins were also a possible alternative. Cryopreserved allografts were used in 90 patients, showing similar patency, more resistant to infection but significantly more susceptible to aneurysms, when compared to PTFE grafts (Madden et al., 2004). More recently, 15 vein allografts seeded with autologous endothelial cells were used in coronary artery bypass surgery with 12 patients, showing patency up to 32 months (Herrmann et al., 2019). Autologous stem cells were also used to recellularize vein allografts and successfully tried on one pediatric patient for vascular vein shunts without the need for immunosuppression (Olausson et al., 2012). Overall, decades of clinical trials have shown that an artery- or vein-based xenografts or allografts can be considered as an alternative to synthetic PTFE or Dacron grafts. However, significant improvements over these graft products, in particular decellularization and recellularization as well as deproteinization and surface engineering to address complications- and/or rejections-derived from immune responses (Merola et al., 2017), are needed before they may gain a bit market share.

Bioreactor-manufactured grafts produce TEVG by culturing human autologous cells and/or allogeneic cells with or without a scaffold in a bioreactor with controlled physical, mechanical and/or biological environments. Clinical trials have illuminated promises and challenges with current methods. Lawson et al. (2016) seeded allogeneic human VSMCs onto polyglycolic acid scaffold, subjected the cell-loaded scaffold to pulsatile cyclic distension for 8 weeks before decellularization for final acellular implant. The patency of these grafts (Humacyte™) on 60 hemodialysis patients was 63% at 6 months and 28% at 18 months, with no evidence of immune response or aneurysm formation. From this trial, 16 tissue samples were acquired from 16 to 200 weeks for the characterization of vascular remodeling, showing recellularization of grafts by non-inflammatory host progenitor and vascular cells (Kirkton et al., 2019). A recent follow-up report provided 5-year (phase II) results on 11 patients who completed the study (Jakimowicz et al., 2022). Results showed that one patient maintained primary patency, and 10 maintained secondary patency. Like PTFE grafts, such acellular grafts showed better clinical outcomes when used as above-knee, femoral-to-popliteal arterial bypass conduits in 20 patients (Gutowski et al., 2020). As an example for scaffold-free cell-engineered graft, Cytograft™ employed patients’ fibroblasts to create an autologous TEVG through a prolonged fabrication procedure, either finishing with autologous endothelial seeding or implanting as it is. Its clinical trial with hemodialysis grafts on ten patients showed 1-month patency of 78% and 6-month patency of 60% (McAllister et al., 2009). Later, L’Heureux’s group modified the graft production into an “off-the-shelf” design—allogeneic Lifeline™ graft. Prior to implantation in patients for hemodialysis access, frozen cell-sheet based scaffolds were thawed, rehydrated and seeded with autologous endothelial cells the luminal side. All three implanted grafts worked well till 11 months (Wystrychowski et al., 2014).

More readily implantable grafts are desirable in clinic, as they require minimal or no need of cell culture time, relying on in vivo degradation, remodeling, regeneration and maturation. One recent approach is embedding a biotube in subcutaneous spaces of a patient for months, using patient’s body to fabricate a graft. This has been used successfully as a pulmonary artery substitute in pediatric patients (Kato et al., 2016; Fujita et al., 2020; Nakatsuji et al., 2021; Higashita et al., 2022). Another approach was employing biodegradable scaffolds in combination with autologous bone marrow mononuclear cells. Their clinical trials investigated these TEVGs in congenital heart surgery (Sugiura et al., 2018). Researchers observed postoperative growth of a TEVG as well as frequent graft narrowing in pediatric patients, but there was no graft-related mortality during the 11-year follow-up period.