In this report, we have established the feasibility of producing a completely biological TEBV exclusively from human cells. This construct meets the three fundamental qualities expected of a vascular graft: mechanical strength, blood compatibility, and suturability. The first and the latter depend directly on the complex and naturally organized 3-dimensional ECM of this vessel. Although the proteins from naturally occurring ECM have been isolated and characterized, these molecules cannot be reassembled in vitro into an ECM featuring physiological density and organization. This explains the inability of previously designed biological TEBV to display physiological strength (17–19). Our model displayed a supraphysiological bursting pressure, and we have shown that this property can be modulated by the culture and maturation periods. However, compliance has also been established as a critical factor for long-term patency of vascular graft (35, 36). This TEBV was clearly more compliant than synthetic ePTFE vascular grafts but, as observed during burst pressures experiments, it was much less compliant than HSV. This is not surprising since the TEBV was optimized for strength. Nonetheless, the mechanical properties of the vessel can be modified by changing the thickness of the different layers. However, since mechanical properties of biological vascular grafts have been shown to change dramatically once implanted (37), the optimal compliance of the TEBV will have to be determined in conjunction with an in vivo study.
In their landmark paper describing the first attempt at creating a biological TEBV, Weinberg and Bell (17) hypothesized that collagenase synthesis might be a critical factor in the mechanical stability of tissue-engineered organs in vitro. Indeed, they observed that their collagen gel-based model experienced an abrupt loss of strength ( 50%) between the 4th and 12th wk of maturation in vitro. In our system, we have observed the transient induction of a gelatinolytic doublet near 50 kDa after the rolling step of the fibroblast sheets. This suggests the induction of MMP-1 (interstitial collagenase), which preferentially degrades native collagen but can also exhibit limited gelatinolytic activity (38, 39). The transient nature of the induction and the general long-term down-regulation of gelatinase activity from both SMC and fibroblasts suggest that the ECM was remodeled to provide a stable physiological environment. The long-term mechanical stability of our model is consistent with the decreased level of collagenase activity observed. A possible explanation for the sustained strength of our TEBV might be that a cell-synthesized matrix induces less collagenase expression, or is more resistant to protease activity, than a biochemically extracted and reassembled matrix such as the one used by Weinberg and Bell. Another possibility would be an increased secretion of tissue inhibitors of metalloproteinase (TIMP) by the cells in this model, which would then reduce collagenase activity. However, MMP expression is only one aspect of the question; in order to verify the above hypothesis, an assessment of ECM proteins synthesis and deposition is needed.
We have shown that a confluent endothelium, expressing normal endothelial markers, was established on the luminal surface of the IM. These ECs expressed functional thrombin receptors as judged by their increase in PGI2 synthesis when challenged with thrombin. PGI2 response to thrombin has been shown to be maximal when human ECs are cultured on natural collagen (40). Activation of the thrombin receptor has also been shown to induce nitric oxide production by ECs (41). This suggests that if platelet adhesion/activation occurs after in vivo grafting, the endothelium will respond to the ensuing release of thrombin by an increase in the synthesis of two potent inhibitors of platelet aggregation. Unfortunately, the hemocompatibility of a human endothelium cannot be properly tested in an animal model because ECs can express major histocompatibility complex (MHC) class II proteins and MHC-mismatched vascular grafts have shown very poor results, even with immunosuppression therapy (42, 43; for review, see ref 44). Consequently, we evaluated the hemocompatibility of the endothelium by using an in vitro functional assay. Although the endothelium was exposed to heparinized human blood under conditions that promote platelet adhesion (very low flow rate), ECs drastically inhibited platelet adhesion and activation. However, it is important to remember that the antithrombotic activity of the EC is useful only if the cells remain attached to the luminal surface when exposed to arterial blood flow. Indeed, endothelial cell seeding experiments of vascular grafts have been plagued with this problem for years (45, 46). Recently, two groups (8, 47) have finally reported increased patency rates for seeded small-diameter synthetic vascular grafts in humans by using EC seeding. In both cases, the grafts were heavily coated with a mixture of biological proteins. These results suggest that the biological nature of the ECM of our TEBV may favor EC adaptation to flow by providing the appropriate anchoring molecules.
Since this is the first implantable, totally biological TEBV, it is difficult to hypothesize its behavior after grafting on the basis of results obtained with other models. However, it is important to underscore the critical differences between this graft and others that share some common characteristics. For example, the Sparks-mandril graft is a tubular scar tissue induced by a subcutaneously implanted Dacron mesh (48). This model was used as an autologous small-diameter vascular graft in the middle 1970s, but was associated with high thrombogenecity and aneurysm formation (49). Although autologous in nature, this graft did include a synthetic mesh and was not endothelialized. Furthermore, this graft was essentially a scar tissue. The ECM was therefore synthesized by cells as part of an inflammatory response, which may explain why it was unable to resist the mechanical strains under physiological conditions. Finally, the closing publication on the Sparks-mandril project suggested that its mechanical instability was linked to the absence of elastin fibers and SMC (50), both of which are present in our model. More recently, a completely biological autologous small vascular graft was produced using the small intestine submucosa in dogs (51). Animal experiments showed very good results, but it appears that human intestinal tissue is not suitable for the production of such a graft (52). Nevertheless, these experiments demonstrate that an autologous tissue of nonvascular origin can be transformed to a completely biological vascular graft and successfully remodeled by the body after implantation.
In vivo experiments were undertaken in order to evaluate the suturability and early mechanical stability, which are challenging problems for biological TEBV. Our results showed that this model can be sutured by using conventional surgical techniques and that it does not tear or dilate when exposed to clinical grafting conditions for a week. However, the patency rate observed in these experiments is not indicative of the potential of this TEBV as a vascular graft. Indeed, an autologous endothelialized graft would undoubtedly show superior performances. Consequently, TEBV will have to be produced from animal cells in order to assess the long-term value of this vascular graft in an autologous system. This TEBV offers exciting future possibilities since the nonsynthetic nature of this vascular graft warrants hope of increased long-term patency rates for lower limb vascular reconstructions. Furthermore, a TEBV produced from subcultured cells may be an appealing tool for gene therapy because these cells would be in close proximity to blood and because human ECs, SMC, and fibroblasts can be safely transfected in vitro without using viruses that induce immunogenic response (53–55).
The potential clinical application of such a TEBV raises several technical questions. For example, what cell source will be used? The external jugular vein has already been used successfully to isolate vascular cells for a clinical application (8, 47). In theory, this tissue could provide the three cell types needed for the production of the TEBV. Besides veins, other sources are also available, such as skin (microvascular ECs and fibroblasts) or fat tissue (microvascular endothelial/mesothelial cells and fibroblasts) (56–60). Upscaling this model to the size needed for vascular reconstructions (i.e., 15 cm for coronary bypass and 15 to more than 30 cm for peripheral reconstructions) is also an important question. This process will either involve a strategy based on overlapping sheets or rely on larger culture containers to produce tubes from a single sheet. Upscaling also requires that a relatively large number of cells must be obtained from the initial biopsy. This brings us to the most critical factor: the time needed for TEBV production. It is obvious that this type of engineered autologous tissue is not designed to be used for emergency surgeries. However, in a nonurgent setting, a significant period of time would be available for the production of a TEBV, especially if this vessel provides excellent long-term patency. Nonetheless, the culture period needed for graft production should be as short as possible for maximal clinical applicability. Many factors can be optimized to reduce production time. For example, biopsy size, culture medium components, fetal bovine serum selection, and bioreactor designs are all elements that we are currently testing to decrease culture and maturation periods.
In addition to being a promising graft from a mechanical point of view, our model also demonstrates fundamental morphological, histological, and functional characteristics of blood vessels. For example, we have observed the expression of desmin, a marker for highly differentiated SMC considered to be irreversibly lost in cultured human SMC (61). Furthermore, a similar SMC construct showed contractile responses when challenged with vasoactive agonists (N. L'Heureux, L. Germain, F. A. Auger, and J. C. Stoclet, unpublished results), a property also considered lost in cultured human SMC (61). Another remarkable aspect of this model is the presence of elastin fibers, an important ECM component never reported in other TEBV (17–19). A possible explanation for these observations is that the cell-synthesized matrix of our TEBV may provide appropriate signals to the vascular cells through stimulation of integrin receptors, thus directing them toward a more physiological phenotype (62–64). The coculture of the various vascular cell types in this model may also favorably affect cellular behavior.
This model exhibits unique features that make it a particularly well-suited system for the study of vascular cells in vitro, especially for experiments involving physical forces. For example, it could be used to study the influence of culture conditions on the development of mechanical properties of the resulting tissue (strength, elasticity, permeability, etc.). Another possibility would be to apply mechanical forces to these engineered tissues and observe cell responses to these forces. The 3-dimensional structure and the natural characteristics of the ECM should ensure a more physiological environment for the cells, whereas the unique mechanical strength of these tissues could allow stimulation at physiological levels.
On a microscale, cultured mesenchymal cells such as SMC and fibroblasts recreate a physiological environment by secreting and organizing, in the presence of vitamin C, a 3-dimensional ECM with characteristics similar to those observed in vivo (65–67). We assembled these microstructures into macrostructures to produce engineered tissues featuring significant mechanical strength and complex histological organization. In turn, these tissues can be assembled together and with other cells to produce completely biological tissue-engineered organs. However, the shape and cell content of theses tissues can be modified to create, with the same basic technique, a wide range of completely biological tissue-engineered organs to meet the growing need for human replacement organs.