**4. Discussion**

The biological compositions of each layer of the native vessels are di fferent. Likewise, the wall thicknesses and mechanical properties of di fferent layers also vary. For instance, the wall thickness of the vessels near the heart are relatively larger. Consequently, when fabricating TEVGs, the ability to independently control each layer's composition and thickness is essential in order to achieve their biological and mechanical performances such as the biocompatibility and the compressive strength.

Today, extrusion-based printing [39], inkjet printing [40], stereolithography [41], two-photon polymerization [42] and laser-assisted printing [43] have been used wisely in vasculature engineering. While the above methods successfully shape tissue-engineered vessels having a multilayered structure or achieve vascularization in bulk constructions, there remains the challenge of fabricating curved bifurcated vascular constructions with a multilayered wall at the macro-scale in a controllable way. In the current study, a novel method of fabricating multilayered biodegradable TEVGs with a curved structure and multi-branches was proposed. The mold system was designed by computer aided design (CAD) software and fabricated by a 3D printer. Taking advantage of the merits of 3D printing technology, we can create three-dimensional structures that are closer to the native blood vessel's morphologies. The proposed method also allows the variation of the composition and dimension of each layer of the constructs. By varying the design parameters of each layer, for example, or by possibly incorporating various bioactive substances in each layer, the thickness and biocompatibility of each layer can be independently controlled. Furthermore, the fabricated tissue-engineered TEVGs possessed a bifurcated structure, which was determined by the CAD models of the mold system. Thus, it is also possible to fabricate vascular constructs with more branches by simply modifying the CAD models. In addition, the three-dimensional structure of the constructs could also be easily modified by varying the design parameters of the mold system. Because of this design flexibility, 3D printed mold systems will give researchers the ability to quickly and inexpensively design TEVGs with the shape and complexity they desire.

In recent years, several methodologies have been introduced to create channels in bulk hydrogels. Bertassoni et al. embedded agarose stripes in in the hydrogel construction to form the inner networks [44]. Negrini et al. sacrificed alginate templates through a chelating agen<sup>t</sup> to obtain a porous gelatin hydrogel [45]. Li et al. added a pre-printed bifurcated polyvinyl alcohol (PVA) structure to the hydrogel to realize vascularization [46]. Compared with the fugitive materials listed above, the adoption of the Pluronic F127 fugitive ink as the sacrificial material is an easy way of forming the inner channel, as the Pluronic F127 is a thermal sensitive material. Furthermore, by taking advantage of the strong shear thinning response of F127 at room temperature, F127 can be printed into very complex channel networks and maintains its shape fidelity without harmful chemical cross-linking, heat processing or additional templates preparing to fabricate inner networks. For TEVGs below the LCST, the ink liquefies and flows readily, while at the same time the other hydrogel materials used are sti ff and solid-like, which makes it easily removed by temperature variation without being obstructed by the closed structure and without exerting extra mechanical force that may impair the printed structure. Taking advantage of this complimentary behavior, the printed ink could be removed without influencing the form of other materials. Our work successfully solved the problem of creating multiple branches for a multi-layered TEVG, which means that the fabricated TEVGs possess more features similar to the native blood vessel. This is a big step towards creating TEVGs with the same level of complexity, which has a tremendous significance for in vitro cardiovascular research and could hopefully reduce the demand for animal experiments.

HUVECs were found to adhere well to the luminal surface of the sample. The formed continuous monolayer of HUVECs was evident after 72 h, which demonstrated the good biocompatibility of the enzymatically-crosslinked TEVGs and their potential for acting as a substitute for ECM.

From the acquired results, one can figure out that the proposed triple-layered TEVG could easily be imbedded into other porous tissue-engineered sca ffolds, thus forming a multi-scale vasculature within a whole three-dimensional structure; the scale of the whole vasculature could range from

micrometers to millimeters. The existence of this multi-scale vasculature will greatly facilitate the mass transport within the sca ffold. Furthermore, when combined with techniques such as cell encapsulation, the fabricated construct would have the potential of developing into a real functional blood vessel. Specifically, if the cell-encapsulating hydrogels were used for the fabrication of the TEVG, di fferent types of vascular cells would further develop into di fferent layers of the blood vessel with the degradation of the hydrogel materials.

As for the scale of the proposed construct, whether the smallest inner diameter could be achieved by this technique depends to a certain degree on the resolution of the 3D printing technology. Theoretically speaking, by promoting the accuracy of the 3D printer, or by adopting techniques with a higher precision such as micro-fluids technology, the scale range of the inner diameter of the TEVG could be further expanded, thus widening the application field of this novel process. Hence, this fabrication technique holds grea<sup>t</sup> potential for impacting a wide range of fields.

Finally, another promising advantage of this novel technique lies in its capacity to design patient specific vessel models because the mold can be constructed by a 3D printer which could be driven directly by the computed tomography (CT) scan data of patients' original vessels, which means that personalized TEVGs that are more suited to the actual human body conditions can be constructed. Constructing the mold system from the clinical data of native vessels is an interesting point for future exploration, and this personalized diagnostic approach has tremendous potential in clinical diagnoses and treatments.
