Rapid endothelialization of printed vascular grafts by perivascular niche-circulating endothelial progenitors crosstalk Authors: Zhe-qian Zhang, Ping-Ping Yuan, Cong Yao, Yu-xin Lu, Yi-dong Hu, Zhen-nan Qiu, Wen-yu Zhan, Xin-chi Zhang, Yu-jiao Wang, Xiao-peng Wu, Ya-jing Zhao, Peng-yu Wang, Ning Leng, Cun Zhang, Jian-kang He, Wei Wu Published: 2025 ABSTRACT AbstractSuccessful regeneration of small-diameter vascular grafts requires rapid endothelialization to ensure patency. However, effectively recruiting circulating endothelial progenitor cells to achieve this remains a significant challenge. Here we show that accelerating the growth of adventitia into vascular grafts promotes fallout endothelialization on the blood-contacting surface. We engineered Micro-Nano structured Vascular Grafts (MnVGs) using high-resolution electrohydrodynamic (EHD) printing and electrostatic spinning. When implanted in rabbits, MnVGs rapidly regenerated neo-adventitia within the highly porous graft walls and achieved endothelialization as early as 2 weeks postoperatively, resulting in 90% patency rate (18/20). The regenerated arteries closely matched native vessels in structure and function. Mechanistically, we discovered that lymphatic endothelial cells enriched in the neoadventitia, were identified as the main contributors to the early recruitment of CD93+ ECs, which ultimately constituted the majority of the endothelium. Thus, MnVGs represent a potential strategy to preventing thrombosis in replacement of small caliber arteries.Subject terms: Implants, Translational research, Regeneration, Biomedical engineering Introduction For arterial substitutes, the critical challenge remains how to rapidly and durably achieve luminal anticoagulation, which is especially vital for small-diameter arteries1. The classic approaches to prevent thrombosis of arterial substitutes use anticoagulant agents through systematic administration, as well as locally releasing from grafts. This is reflected in autografts, synthetic grafts, and many tissue-engineered grafts. However, this strategy remains far away from clinical success in small caliber arterial reconstruction2. Naturally, vasculogenesis is pioneered by differentiation, migration, and assembly of endothelial progenitor cells (EPCs), which is then turned to be endothelial cells establishing cell-to-cell connections (e.g., tight junctions and gap junctions) and organize into tube-like structures3. Directing the movements of EPCs are orchestrated waves of signaling molecules primarily released from surrounding mesenchymal cells or tissues4. Interestingly, implanting polymeric or biological grafts that enable communication between blood and perivascular tissues may partially replicate vasculogenic events and regenerate endothelium on the graft lumen5,6. This inspired us to fully harness the body’s regenerative capacity to remodel cell-free synthetic grafts with an open porous structure, which accelerates EPCs recruitment and endothelialization. However, designing a vascular graft requiring both safety in arterial circulation and openness for perivascular tissue in-growth remains a key challenge. Ongoing research continuously emphasized the importance of perivascular cell infiltration within the graft wall regeneration as key factors in vascular design7,8. Specifically, various methods to recreate critical components for Fallout endothelialization have been developed to systematically study the cellular events and interactions involved in EPCs9,10. Laura Niklason and colleagues observed the remodeling of tubular decellularized matrix in the human arterial circulation and found that the transmural growth and infiltration of adventitia cells are crucial components for the regeneration of vascular tissues, including the intima and smooth muscle layer11. Kong et al., also discovered that increasing the intrinsic pore size of PCL electrospun scaffolds promoted endothelialization through enabling growth of peripheral tissues into graft wall. These studies collectively confirm the critical role of peripheral tissues in intimal regeneration12. However, whether it is decellularized matrix or electrospining grafts, their poor controllability of porosity and pore size is not conducive to the early interaction between perivascular tissues and circulating progenitor cells13. We previously used rapidly degradable elastomer, PGS, to construct porous scaffolds and found that an endothelial layer formed within 4 weeks. However, achieving the required strength necessitated a scaffold thickness of nearly 1 mm, PGS degradation induced inflammation and thick graft wall prolonged the time for centripetal growth of perivascular tissues and was also unfavorable for the early formation of the endothelial layer14,15. High-resolution EHD printing represents an additive manufacturing technology that leverages advanced polymer fluid dynamics to achieve high-resolution fabrication. This technique enables the precise construction of scaffolds with hierarchical architectures, such as tympanic membranes, heart valves, and vascular constructs, through the deposition of polycaprolactone (PCL) into microfibrous architectures with cellular-scale precision16,17. The resulting PCL microfibers exhibit biocompatibility and support the rapid and oriented proliferation of various cell types, such as smooth muscle cells, fibroblasts, and Schwann cells18. Artificial blood vessels require a delicate balance between mechanical integrity and porosity to ensure both biosafety and optimal cellular integration19. In this study, we engineered an artificial blood vessel through the synergistic integration of electrostatic spinning and EHD printing technologies. The construct features a bilayered architecture: an inner layer composed of PCL electrospinning filaments provides a continuous interface that facilitates blood cell deposition while preventing plasma extravasation, and an outer layer comprising ultrathin EHD printing scaffolds that deliver essential mechanical support. The graft’s unique combination of pore architecture with micro-nano scales, fiber orientation, and ultrathin configuration promotes the inward growth of neo-adventitial tissue while facilitating effective communication with the developing neo-endothelium, which acquired rapid endothelialization within 2 weeks and achieved the 90% patency rate (18/20) in a rabbit carotid artery model with 15 mm defects. Furthermore, this study revealed extensive intercellular communication networks between CD93+ ECs in neo-endothelium and adventitia derived LYVE+ endothelial cells, underscoring the critical role of regenerated adventitia in facilitating rapid endothelialization processes, and offering insights into the mechanisms underlying vascular regeneration. Result Micro-nano hierarchical vascular grafts enabled rationalized mechanical strength and cellular permeability The micro-nano vascular grafts (MnVGs) were fabricated with a bilayer structure comprising an electrospinning luminal layer and an electrohydrodynamic-printed abluminal layer, engineered to prevent blood extravasation while maintaining sufficient permeability. The luminal layer was constructed as a 30 μm thick polycaprolactone (PCL) electrospun membrane, while the outer EHD printing structure was characterized by fibers measuring 35 μm in diameter with an inter-fiber spacing of 200 μm, exhibiting lattice-like crisscross arrangement. Through precise control of EHD printing scaffold deposition, we successfully fabricated MnVGs with three distinct thickness profiles: single-layer (52.21 ± 6.75 μm), triple-layer (189 ± 21.72 μm), and deca-layer (462 ± 59.83 μm) configurations (Fig. 1a, b). Surface topography analysis using laser confocal microscopy revealed significant differences in roughness parameters between the layers, with the electrospinning layer (ES) exhibiting a roughness of 8.01 ± 2.18 μm, compared to 153.6 ± 9.46 μm for the EHD printing layer (Fig. 1c and Supplementary Fig. 1b). Fabrication and characterization of MnVGs. a Gross morphology of MnVG. b SEM images showing graft microstructure and thickness. Images are representative of results obtained from 3 independent experiments. c 3D laser confocal microscopy images depicting nanolayer and EHD-printed layer roughness. d−j Mechanical properties of grafts: stress-strain curve (d), compressive load curve of 3 layers (e), suture retention strength (f), ultimate tensile strength (g), elastic modulus (h), burst pressure (i), and compressive load (j) (n = 3 independent samples). k, l Quantification of AAF infiltration distance (k) (n = 3 independent samples) and 3D live/dead staining of AAFs in grafts 3 days post-seeding (l). m Images of AAFs and LECs infiltrating the ES layer. For (f−j) significance was determined by one-way ANOVA followed by Tukey’s post hoc analysis. The adjustments were made for multiple comparisons to get P-value. For k, significance was determined by an unpaired two-tailed Student’s t-test without made for multiple comparisons. Plotted data were presented as means ± SD. Source data are provided in Source Data file. Mechanical characterization through comprehensive testing demonstrated significant enhancements in graft properties with increasing layer number. Tensile testing revealed progressive improvements in mechanical performance, with ultimate tensile strengths measuring 0.53 ± 0.07 MPa (single-layer), 1.16 ± 0.17 MPa (triple-layer), and 3.08 ± 0.15 MPa (deca-layer) (Fig. 1d, g). Corresponding suture retention strengths were quantified at 0.13 ± 0.03 N (single-layer), 0.74 ± 0.08 N (triple-layer), and 1.25 ± 0.11 N (deca-layer), respectively (Fig. 1f). Elastic modulus measurements were quantified at 12.37 ± 1.96 KPa (single-layer), 31.57 ± 3.18Kpa (triple-layer),73.37 ± 5.61 Kpa (deca-layer) (Fig. 1h). Burst pressure assessments demonstrated substantial improvements in structural integrity, with measurements were quantified at 396.3 ± 130.4 mbar (single-layer), 752 ± 146.2 mbar(triple-layer), 2435.2 ± 409.6 mbar (deca-layer) (Fig. 1i). Cyclic compression testing yielded values of 0.02 ± 0.01 N (single-layer), 0.53 ± 0.10 N (triple-layer), and 1.19 ± 0.13 N (deca-layer) for single-, triple-, and deca-layer constructs, respectively (Fig. 1e, j). To examine the openness of MnVG with different EHD layers, cell invasion assays using rat adventitial fibroblasts were performed, which showed infiltration depths of 118.6 ± 17.36 μm in triple-layer constructs whereas only 46.2 ± 9.34 μm in electrospinning membranes with the same thickness (Fig. 1k, l). Interestingly, human umbilical vein endothelial cells (HUVECs) seeded onto all MnVGs penetrated through the ES layer and proliferated along the EHD layers, while fibroblasts (AAFs) seeded with the same approach were confined to the ES surface even after 5 days of culture (Fig. 1m). DTβ4 failed to promote penetration of both types of cells in the ES layer (Supplementary Fig. 2c). In contrast, both types of cells were unable to penetrate the 150 μm-thick electrospinning membrane over 5 days’ culture (Supplementary Fig. 2d). These findings suggested that increased EHD printing layers enhances mechanical properties, while concurrently reduces openness, potentially limiting peripheral cell ingrowth. Incorporating DTβ4 into electrospinning PCL enabled MnVGs to be antithrombogenic. Surgical implantation in rabbit models followed by 2 h explantation revealed 100% patency (5/5) in E-PCL + DTβ4 grafts, as confirmed by distal arterial refilling tests (Supplementary Fig. 3b). In vitro evaluations using recalcified whole blood incubation confirmed the antithrombogenic potential of the E-PCL + DTβ4 membranes (Supplementary Fig. 3a, c–f). The comprehensive characterization of MnVGs claimed that MnVGs with 3 EHD layers may acquire appropriate mechanical strength, surgical operability, cellular interaction, and antithrombogenic performance in vascular graft design, suggesting suitability to be grafts for in vivo experiments. MnVGs maintained high patency rates and rapid neo-adventitia Ingrowth To evaluate the impact of graft wall openness on arterial regeneration, we compared MnVGs and electrospinning-only (ES) grafts with identical thickness in a rabbit carotid artery defect model. The characterization of the ES was shown in Supplementary Fig. 4, dense vessel walls was confirmed by SEM examination. 15 mm segmental defect was surgically created in the left carotid artery of each rabbit, with MnVGs implanted in the experimental group (n = 20) and electrospinning grafts as controls (n = 6) (Fig. 2a, b). At 2 weeks after implantation, graft performance diverged markedly between groups. MnVGs acquired robust neovascularization, with complete integration with perivascular tissue. In contrast, electrospinning only grafts showed minimal neovascularization and poor formation of neointima (Fig. 2c). Doppler ultrasound revealed a 90% patency rate (18/20) in the MnVGs group whereas controls exhibited 50% patency (3/6) and significant regurgitation despite of administration of anticoagulating agents (p < 0.01, Fisher’s exact test). Hemodynamic parameters in MnVGs, including blood flow velocity (MnVG: 2 W, 72.5 ± 1.52 cm/s; 4 W, 73.0 ± 1.55 cm/s; 12 W, 74.7 ± 1.96 cm/s; native artery: 74.5 ± 1.05 cm/s) and luminal diameter (MnVG: 2 W, 1.94 ± 0.02 mm; 4 W, 1.93 ± 0.02 mm; 12 W, 1.95 ± 0.02 mm; native artery: 1.95 ± 0.2 mm) appeared similar to those of native vessels (Fig. 2d–g). Collectively, these results demonstrated that MnVGs significantly outperformed conventional electrospinning grafts, with structural openness critically enhancing patency and regenerative efficiency in 2 weeks. Gross morphology and ultrasound monitoring of MnVGs at each time point. a Schematic illustration of the MnVG implantation process. b Gross morphology of the MnVG after implantation is completed. c The gross morphology of the MnVG and control group (Con-2W) at different time points post-implantation. Yellow dotted lines indicate anastomosis sites. d Representative Doppler ultrasound images of the grafts and native carotid arteries. e Number of patent grafts per time point. f Peak blood flow velocity in MnVGs at each time point (n = 6 independent samples). g Average inner diameter of MnVGs and native carotid arteries (n = 6 independent samples) Plotted data represent means ± SD. For f, g significance was determined by one-way ANOVA followed by Tukey’s post hoc analysis. The adjustments were made for multiple comparisons to get P-value. Source data are provided in Source Data file. Transmural growth of Neo-adventitia facilitated rapid endothelialization and intima regeneration The hierarchically porous architecture of MnVGs significantly promotes host remodeling from peripheral tissues. At 2 weeks, the porous graft wall of MnVGs have been compactly filled with host cells, as confirmed by slides from both longitudinal sections and cross sections (Fig. 3a, b). Most cells appeared as spindle shape, which were evenly distributed along the whole grafts and only slight inflammation observed. As compared with remodeling tissues abluminally wrapped electrospinning membrane in MnVGs, the luminally aligned cells were more regularly arranged, and a layer of compactly aligned cells, similar to endothelial cells, completely covered the luminal surface. Quantitative analysis showed that the average thickness of luminally aligned tissue in MnVGs (111.50 ± 24.63 μm) significantly exceeded that in control group (12.66 ± 2.93 μm), confirming the significantly enhanced intima regeneration in MnVGs (Fig. 3c). Notably, the PCL residual area in MnVGs after 2 weeks of implantation (32.41 ± 2.03 × 103 μm²) was significantly lower than that observed in control group (91.59 ± 4.59 × 103 μm²) (Fig. 3d). Neointima regeneration and rapid endothelialization of MnVGs within 2 weeks. a,b H&E staining of MnVGs and control grafts (Con-2W) at the midportion (a) and longitudinal section (b) at 2 weeks. Images are representative of results obtained from 6 independent experiments. c Quantitative analysis of neointima thickness (n = 6 independent samples). d Quantitative assessment of residual PCL polymer (n = 6 independent samples). e Immunofluorescence staining of en-face preparations from different regions of MnVGs and control grafts at 2 weeks. f Quantification of luminal endothelialization at different neoarterial sites by en-face staining (n = 6 independent samples). Plotted data represent means ± SD. For c, d and f significance was determined by an unpaired two-tailed Student’s t-test without made for multiple comparisons. L: Lumen. Source data are provided in Source Data file. Immunofluorescence staining provided further insights into the cell identities. In MnVGs harvested at 2 weeks, endothelial cells immunostained with CD31 were scatteringly distributed on the luminal surface of the neointima (anastomosis, 76.35 ± 3.64%; quarter, 50.31 ± 3.93%; midportion, 38.55 ± 2.66%) (Supplementary Fig. 5a–c). Interestingly, same decreasing trend was also observed in CD31+ ECs in neoadventitia, the cell counting from the anastomosis towards the midportion of graft wall was as following, (anastomosis, 40.35 ± 3.90%; quarter, 25.64 ± 3.93%; midportion, 17.64 ± 3.47%) (Supplementary Fig. 5d). Comparatively, control group exhibited significantly fewer endothelial cells on the luminal side (anastomosis, 32.28 ± 3.39%; quarter, 16.39 ± 2.53%; midportion, 9.44 ± 2.09%), with markedly reduced endothelial cell density throughout graft wall (anastomosis, 24.70 ± 5.05%; quarter, 13.79 ± 2.95%; midportion, 6.96 ± 1.79%) (Supplementary Fig. 5b, d). These observations suggested that the optimized wall structure of MnVGs not only facilitates adventitia regeneration but also supports neo-intima tissue development, as well as endothelialization. The superior performance of MnVGs, as evidenced by their high patency rate, could be attributed to their rapid endothelialization capability. Comprehensive characterization using en-face immunofluorescence staining and SEM revealed complete cellular coverage on the luminal side of MnVGs at 2 weeks, while control group showed incomplete coverage (Supplementary Fig. 5e). En-face staining demonstrated partial coverage of CD31+ cells, indicating the emergence of mature ECs (anastomosis, 66.33 ± 2.29%; quarter, 55.98 ± 2.72%; midportion, 40.60 ± 4.23%). The spatial organization of CD31+ cells within MnVGs progressed from initial clusters to an undirected reticulation pattern, eventually forming a continuous, flow-aligned sheet at the anastomosis (Fig. 3e, f). This scattering distribution of CD31+ cells suggested two potential endothelial cell sources: ECs transmurally migrated from perivascular tissues and circulating endothelial progenitor cells from blood stream, rather than transanastomotic ingrowth. In contrast, control group showed significantly fewer CD31+ cells, as confirmed by reduced ECs density throughout the graft length (anastomosis, 36.40 ± 3.36%; quarter, 16.62 ± 2.80%; midportion, 9.92 ± 3.11%) (Fig. 3e, f). These findings collectively demonstrated that the optimized wall structure of MnVGs promotes infiltration of adventitial tissue and luminal endothelialization, which contributed to high patency rates. Adventitia enabled progressive endothelialization on luminal surface and stable vascular remodeling To evaluate the remodeling process of MnVGs, the combination of histological and immunohistochemical techniques, including hematoxylin-eosin (HE) staining, elastic Van Gieson (EVG) staining, and Masson’s trichrome staining (MTS), was used to characterize the ECM formation in neoarteries. Neoartery wall in such graft presented compactly aligned, trilayer structure and integrates with the perivascular tissue (Fig. 4a). MTS analysis revealed organization of collagen fibers, which was orderly and circumferentially aligned at 4 weeks post-implantation. By 12 weeks, the collagen fibers become densely arranged, and spindle-like cells immunostained with a-SMA were organized into well-defined bundles (Supplementary Fig. 7c). EVG staining demonstrated the initial deposition of elastic fibers at 4 weeks, which subsequently formed concentric, ring-like layers of elastic fiber bundles by 12 weeks (Supplementary Fig. 7c). Under polarized light, the polycaprolactone (PCL) structure in MnVGs appeared as bright dots or strips in neoadventitia (Supplementary Fig. 7c). The MnVGs without significant degradation persisted at 4 weeks (32.85 ± 2.16 × 103 μm²) and 12 weeks (31.86 ± 1.83 × 103 μm²), with no significant changes from the 2 week samples (Supplementary Fig. 7d). The collagen and elastin quantification for 12 week samples were quantified as 16.67 ± 1.51 μg/mg and 3.66 ± 0.24 μg/mg respectively, which was significantly >4 week samples (elastin, 1.50 ± 0.19 μg/mg; collagen, 11.12 ± 1.77 μg/mg) (Supplementary Fig. 7e, f). These results suggested that MnVGs underwent progressive remodeling and extracellular matrix (ECM) deposition. Notably, neointima formation and specific ECM deposition were only observed in the luminal side of MnVGs rather than in the grafts, with a clear boundary separating these regions from the adventitia formed in MnVGs. Host remodeling and progressive endothelialization of MnVGs within 12 weeks. a H&E staining of longitudinal sections of neoarteries. b Longitudinal views of transmural angiogenesis and luminal endothelialization in neoarteries at 4 and 12 weeks. c Immunofluorescence staining of en-face preparations from different regions of MnVGs at 4 and 12 weeks. d Quantification of the luminal endothelialization of different neoarterial sites (n = 6 independent samples). e Quantification of endothelial coverage by percentage of CD31 positive cells in the luminal of neoarteries at the anastomosis, quarter, and midportion (n = 6 independent samples). Plotted data represent means ± SD. f Quantification of luminal endothelialization at different neoarterial sites by en-face staining (n = 6 independent samples). Plotted data represent means ± SD. For d–f significance was determined by an unpaired two-tailed Student’s t-test without made for multiple comparisons. L: Lumen, the white dotted line indicated the suture sites, CD31: red. Source data are provided in Source Data file. MnVGs demonstrated a obvious capacity for guiding full endothelialization and stable vascular remodeling following implantation. None of the samples exhibited intimal hyperplasia. At 4 weeks after implantation, CD31+ cells on the luminal surface formed discontinuous monolayer strips which extended from the anastomosis to the midportion, and the ECs density on luminal surface from the anastomosis towards the midportion was gradually reduced (anastomosis, 77.74 ± 3.72%; quarter, 58.14 ± 4.18%; midportion, 42.68 ± 3.51%). By 12 weeks after implantation, ECs on luminal surface merged into continuous monolayer almost completely covering the neo-arteries (anastomosis, 93.33 ± 2.39%; quarter, 91.48 ± 2.42%; midportion, 87.61 ± 2.14%). No statistically significant differences in luminal CD31+ cell density between the anastomosis and midportion was detected (Fig. 4b, d and Supplementary Fig. 7b). At 4 weeks, the neo-capillaries in the graft walls which characaterized by CD31+ immunostaining area were significantly increased than 2 weeks (anastomosis, 71.10 ± 3.77%; quarter, 62.40 ± 3.01%; midportion, 57.21 ± 2.65%). Interestingly, after 12 weeks, the CD31+ cell population within the neo-adventitia decreased (anastomosis, 44.75 ± 2.22%; quarter, 41.42 ± 2.16%; midportion, 37.68 ± 2.14%) whereas reorganized into distinct striated, tubular structures that obviously separated from the neointimal tissue (Fig. 4e). Notably, nourishing vessels within the adventitia preferentially localized in the abluminal region of eletrospinning sheath in graft wall rather than on the neointimal side. Above spatiotemporal discrepancy between luminal and abluminal ECs was in consistence with existing literature claiming distinct cellular origins for endothelial cells in the graft lumen versus the wall14. Multiple imaging techniques provided further insights into luminal endothelialization. En-face immunofluorescence staining demonstrated the CD31+ cells, transitioning from initial clusters scatteringly distributed into a flow-oriented mesh structure (Fig. 4c). Spatial analysis showed significant increases in CD31+ cell proportions at 4 weeks (anastomosis, 85.75 ± 2.35%; quarter, 76.49 ± 2.71%; midportion, 53.29 ± 3.13%), particularly in the central graft regions where endothelial cells formed continuous, flow-aligned networks. By 12 weeks, the luminal surface was completely covered by a continuous CD31+ cells, with no regional disparities in cell distribution, as SEM presented, densely packed, flow-aligned cells appeared with similar shape of arterial-like ECs (Supplementary Fig. 7a). Further immunostaining counting revealed high ECs coverage at all examining regions (anastomosis, 90.53 ± 1.93%; quarter, 89.67 ± 1.72%; midportion, 88.75 ± 2,77%) (Fig. 4f). These observations suggested that porous graft walls enabled regeneration and infiltration of neoadventitia whereas almost all neointima formed in the luminal side of electrospinning layer in MnVGs, increasing coverage of mature ECs ensured high patency rate of MnVGs through 3 months. Neo-artery acquired similar structure and myography with native artery following long-term implantation Owing to the slow degradation rate of PCL constructs in vivo, host remodeling of MnVGs was tracked for up to 12 months. All 6 neo-arterial samples harvested at 12 months were patent. Surgical examination revealed that the neo-arteries exhibited a soft and elastic texture, with no signs of distension or aneurysms. The MnVGs completely transformed into native vessels, featuring faint yellow neo-adventitia infused with red capillaries; the lumen of the neoarteries presented a smooth surface without hyperplasia and was fully integrated with native vessels (Supplementary Fig. 8a). Histological analysis via H&E staining confirmed that nourishing capillaries regenerated through the adventitia, and the neoarteries displayed a uniform three-layer vascular structure along their longitudinal axis (Fig. 5a and Supplementary Fig. 8b). The neointima thickness was 261.4 ± 19.11 μm (Supplementary Fig. 8m). Notably, MnVGs were restricted to the neoartery adventitia, as evidenced by a significantly reduced residual PCL area under polarized light microscopy (12.05 ± 1.90 × 10³ μm² vs. 31.86 ± 1.83 × 10³ μm² in 12 week samples) (Fig. 5b). TEM images revealed the ultrastructure of neoarteries, including a waving internal elastic lamina beneath endothelial cells. In the medial layer, bundles of elastic and collagen fibers were organized into an interwoven pattern, with smooth muscle cells interspersed among them. At higher magnification, synaptic structures between smooth muscle cells and nerve axons were observed, and synaptic vesicles were visible within presynaptic membranes (Fig. 5c). Quantitative analysis of elastin and collagen content in the neoarteries (elastin, 7.01 ± 0.18 μg/mg; collagen, 20.41 ± 0.95 μg/mg) were close to those in native arteries (elastin, 7.64 ± 0.23 μg/mg; collagen, 20.18 ± 1.05 μg/mg) (Fig. 5d). Immunofluorescence staining for α-smooth muscle actin (α-SMA) demonstrated a uniform distribution of α-SMA⁺ cells surrounding the neoarteries, with no significant difference in coverage area compared to native arteries (44.13 ± 2.49% vs. 46.73 ± 2.97%) (Supplementary Fig. 8e, g and Supplementary Fig. 9a). Extracellular matrix (ECM) analysis of neoarteries via collagen type I (COL-I), collagen type III (COL-III), and elastin staining revealed well-organization of these components in vessel walls (Fig. 5e and Supplementary Fig. 8c, d, f), closely resembling native arterial ECM (Supplementary Fig. 9b–d). The fluorescence area of COL-I (52.08 ± 2.87%) showed no significant difference from that of natural arteries (52.97 ± 2.77%), while the fluorescence area of COL-III (61.93 ± 4.38%) remained lower than the native arteries (80.75 ± 3.72%), and elastic fibers (52.99 ± 2.41%) was also lower than that in natural arteries (62.19 ± 3.21%) (Supplementary Fig. 8h–j). MTS revealed organized collagen fibers and red-stained smooth muscle structures in the neointima, EVG staining showed well-aligned black-stained elastic fibers (Fig. 5f). The incomplete degradation of the PCL stent may hinder the mechanical compliance required for proper remodeling of neonatal arteries in arterial flow environment, thereby restricting organization of elastic fibers. Immunofluorescence analysis showed a continuous monolayer of CD31⁺ cells lining the entire luminal surface of the neoarteries, with endothelial cell density at the midpoint cross-section comparable to that of autologous arteries (Supplementary Fig. 8k, l). Structural and functional assessment of neoaorta. a H&E and polarized H&E staining of transverse sections at the midportion of neoarteries at 1 year post-implantation. b Quantitative assessment of residual PCL polymer by polarized microscopy (n = 6 independent samples). Plotted data represent means ± SD. c TEM morphology of neoartery structure at 1 year. Images are representative of results obtained from 3 independent experiments. In the neo-intima, red arrowheads indicate endothelial cells (ECs) and white stars mark the basement membrane. In synapses, red arrowheads indicate synapses, black stars mark smooth muscle cells, and white stars mark presynaptic membranes. d Quantitative analysis of collagen and elastin content in neoarteries at 1 year (n = 6 independent samples). Plotted data represent means ± SD. e Immunofluorescence staining of collagen III, and elastin in neoarteries. Images are representative of results obtained from 6 independent experiments. f MTS and EVG staining of neoartery cross-sections at 1 year. g–i Mechanical properties of neo-arteries: stress-strain curve (g), elastic modulus (h) and burst pressure (i) (n = 3 independent samples). Plotted data represent means ± SD. For b−i significance was determined by an unpaired two-tailed Student’s t-test without made for multiple comparisons. Source data are provided in Source Data file. The long-term mechanical integrity and in vivo remodeling were evaluated through bending, and torsion test at 12 months post-implantation. The neoartery exhibited robust pulsation under physiological blood circulation. Furthermore, the vessel maintained a patent lumen with uninterrupted blood flow following 90-degree bending and 180-degree torsion, as presented in Supplementary Movie 1-3. Following detachment, the neo-artery demonstrated robust shape-morphing into its pre-designed configuration under in vitro conditions (Supplementary Fig. 8q, r and Supplementary Movie 4). To further evaluate the functional maturation of regenerated neoarteries at 12 months, isometric force measurements and mechanical characterization tests were performed to assess their responsiveness to vasomotor signals. The stress-strain curves of neoartery and native artery groups were obtained through tensile testing (Fig. 5g). The elastic modulus measurement of the neoartery was 684.0 ± 88.37 kPa, significantly higher than that of the native artery (378.07 ± 34.31 kPa) (Fig. 5h). The burst pressure assessment revealed no significant difference between two groups, with the measured results quantified as follows: 4157 ± 281.0 mbar (neoartery), 3670 ± 196.7 mbar (native artery) (Fig. 5i). The higher elastic modulus presented in neo-arteries corresponded to lower compliance compared to the native artery. Neoarteries exhibited contractile and relaxation capabilities, suggesting vascular cells had acquired functional phenotypes similar to native vessels. However, the contractile response of neoarteries to KCl (7.50 ± 0.83 g) and phenylephrine (PE, 6.52 ± 1.05 g) remained weaker than that of native arteries (KCl: 11.94 ± 1.19 g; PE: 12.13 ± 1.26 g) (Supplementary Fig. 8n, p). Similarly, neoarteries showed a 28.42 ± 3.04% dilation in response to the endothelium-dependent vasodilator acetylcholine (ACh), which was markedly lower than the 68.44 ± 4.67% dilation observed in native arteries (Supplementary Fig. 8o). These findings demonstrated that neoarteries acquired functional vasomotor properties, albeit with reduced efficacy compared to native tissue. These findings collectively suggested that MnVGs exhibit favorable biosafety profiles and sustained regenerative effects in vivo. They facilitated the regeneration of a stable intima-media layer in damaged carotid arteries, leading to functional remodeling of the arteries. LECs enhanced the recruitment of circulating CD93+/CD34+ ECs for rapid endothelialization in grafts To precisely delineate the differences between ECs on the luminal surface and those within the neo-adventitia, vascular cells from 1, 2, and 4 week were subjected to single-cell RNA sequencing (scRNA-seq) analysis. The data obtained were meticulously analyzed, revealing host cells classified into 10 distinct subpopulations based on genetic profiling: macrophages, monocytes, dendritic cells (DCs), endothelial cells, fibroblasts, T cells, B-plasma cells, Schwann cells, mast cells, and neutrophils. These subpopulations were visualized using Uniform Manifold Approximation and Projection (UMAP) (Fig. 6a, b). Single-cell RNA-seq analysis of remodeling MnVGs at 1,2 and 4 weeks after implantation. a, b UMAP analysis atlas of carotid artery remodeling at 1, 2 and 4 weeks after implantation. The red cycle marked the EC clusters. c UMAP plot of ECs showing the number of different EC types. d The proportional composition of different cluster ECs. e The clusters of LYVE ECs and CD93 ECs in UMAP reanalysis map displays. f Strength of ligand-receptor interactions between ECs and other cell population pairs based on CellPhone analysis. Edge width is proportional to the number of ligand-receptor pairs. g Immunofluorescence staining of cell marker CD93 and LYVE in neoarteries at 2 weeks. h Immunofluorescence staining of LYVE+ cells in neo-adventitia at 2 weeks. For (g, h) images are representative of results obtained from 6 independent experiments. Source data are provided in Source Data file. ECs in the scRNA-seq mapping were identified by the expression of PECAM, CAV1, CCL21, and VWF genetic markers (Supplementary Fig. 11a, d). Differential gene markers for each endothelial cell subtype are illustrated in the accompanying heat map (Supplementary Fig. 11b, c). Enrichment pathway analysis indicated that these genes in ECs were significantly enriched in the Rap1 signaling pathway (Supplementary Fig. 11g, h). Furthermore, gene ontology (GO) analysis suggested a potential role of ECs in angiogenesis (Supplementary Fig. 11e, f). Subsequent selection of ECs allowed further classification into arterial endothelium (SELP, ICAM1, CLU), venous endothelium (DACH1, CDH13, EFNB2), activated endothelium (CCL21, ACTA2, RGS5), and lymphatic endothelium (LYVE, PROX1, MMRN1) (Fig. 6c and Supplementary Fig. 12a, b). Notably, the proportion of lymphatic endothelium increased over time following implantation (Fig. 6d and Supplementary Fig. 12c). Analysis of UMAP plots based on distinct clusters of EC markers revealed two primary populations within the neoendothelial population: CD93+ ECs and LYVE+ ECs (Fig. 6e). Pseudotime analysis indicated three distinct fates for ECs: CD93+ and LYVE+ ECs advanced to a more differentiated stage. Specifically, LYVE+ ECs were predominantly found in Pre and FATE1, while CD93+ ECs were present in FATE2 stages (Supplementary Fig. 12d, e). To delineate the heterogeneity of LYVE+ endothelial cells (ECs), we performed unsupervised clustering on our single-cell RNA sequencing data, which identified two distinct subpopulations: a canonical LYVE+ cluster (cluster 4) and a unique subpopulation (cluster 5) that co-expressed LYVE1 alongside venous endothelial markers, including NR2F2 and EPHB4 (Supplementary Fig. 15a, b). Using our spatial transcriptomics datasets, we assessed the spatial distribution of LYVE+ ECs and venous ECs within the nascent arterial tissue. We investigated the spatial relationship between venous ECs and LYVE+ LECs in the neo-vessels. The results demonstrated a significant and non-random co-localization, with venous ECs markedly enriched in the immediate proximity of LYVE+ LECs (Supplementary Fig. 15c). At 4 weeks, we observed a significant increase in the expression intensity of venous ECs markers such as NR2F2 and EPHB4 in LYVE+ LECs (P < 0.0001) (Supplementary Fig. 15d). Furthermore, the quantity of NR2F2+/LYVE+ cells at 4 weeks (33.2 ± 9.11 cells per field) was significantly more than that observed at 2 weeks (8.7 ± 3.01 cells per field) (Supplementary Fig. 15e, f). To explore context-dependent cell-cell communications that activate the remodeling process, we employed CellPhone DB to infer interactions based on the co-expression of multisubunit receptor-ligand complexes. Analysis of endothelial cells, fibroblasts, macrophages-monocytes, and pericytes revealed extensive communication between CD93+ and LYVE+ ECs with all cell types (Fig. 6f and Supplementary Fig. 12f). Specifically, LYVE+ ECs communicated with CD93+ ECs via molecules such as SEMA6A, WNT9A, and VWF (Supplementary Fig. 12g). This communication gradually diminished over time following graft implantation. Immunofluorescence at 2 weeks post-implantation showed CD93+ cell deposits on the luminal side of tubes and numerous LYVE+ cells within the walls (Fig. 6g). Some LYVE+ cells migrated through the ES layer at 2 weeks post-implantation, but they did not directly contribute to the cell source of the endothelialization (Supplementary Fig. 13c). Circulating blood-derived CD93+ ECs were identified as crucial for vascular graft endothelialization. These findings suggest that MnVGs recruits circulating CD93+ endothelial cells for early endothelialization, with close intercellular communication observed between LYVE+ and CD93+ endothelial cells in the graft. Using immunofluorescence staining to observe LYVE+ ECs in the neoadventitia of the grafts, it was found that the number of LYVE+ ECs at the midpoint site (103.8 ± 18.71 cells per field) was lower than that in the anastomosis site (222.7 ± 29.30cells per field) (Fig. 6h and Supplementary Fig. 13b). To further validate the spatial and temporal relationships of LECs in recruiting CD93+/CD34+ endothelial cells, we conducted spatial transcriptomic analysis on neoarteries at 2 and 4 weeks (Supplementary Fig. 14a, b). Gene enrichment analysis based on bubbleplot spatial transcriptomics scored the top marker genes of the single-cell dataset and projected them onto the spatial transcriptomics slices (Supplementary Fig. 14d). The main cellular components in the sample were divided into 11 clusters (Supplementary Fig. 14c, e). It is noteworthy that ECs aggregation is observed in the neointima on the luminal side, and a significant aggregation of ECs is evident in the neoadventitia of the nascent artery (Supplementary Fig. 14f, g). Based on our spatial transcriptomics analysis, we investigated the spatial relationship between CD93+/CD34+ ECs and LYVE+ LECs in the neo-vessels. The results demonstrated a highly significant and non-random co-localization, with CD93+ ECs markedly enriched in the immediate proximity of LYVE+ LECs (Fig. 7d). This precise spatial patterning, visualized by mapping endothelial subpopulations back onto the original tissue sections, provides direct anatomical evidence supporting a potential active recruitment or localization mechanism between these two cell populations during graft endothelialization. Single-cell RNA-seq analysis of MnVGs endothelialization after implantation. a UMAP analysis atlas of CD93+/CD34+ ECs at 1, 2 and 4 weeks after implantation. The red cycle marked the EC clusters. b Heatmap of cell-cell communication analysis based on CellPhone DB. c Violin plots for expression marker genes (CD93, CD34, CD14, and LYVE) in each EC cluster. d Spatial visualization of the CD93+/CD34+ ECs and LYVE+ ECs subpopulations mapped back onto the original tissue section. L: Lumen, the white dotted line indicated the suture sites. e Immunofluorescence staining of CD93+/CD34+ cell in lumine of the neoarteries at 2 and 4 weeks. The white arrowhead marked luminal CD93+/CD34+ cell. f Count of luminal CD93+/CD34+ cells at 2 and 4 weeks (n = 6 independent samples). Plotted data represent means ± SD. For (e) significance was determined by an unpaired two-tailed Student’s t-test without made for multiple comparisons. Source data are provided in Source Data file. Prior investigations have established the critical involvement of CD34+ cells in cardiovascular repair and the maintenance of endothelial homeostasis. Our UMAP analysis of neo-arteries identified a significant population of CD93+/CD34+ endothelial cells within the neo-endothelial compartment (Fig. 7a). To decipher the context-dependent cellular crosstalk that triggers vascular remodeling, we performed a systematic cell-cell interaction analysis using CellPhone DB, which evaluates the co-expression patterns of multi-subunit ligand-receptor complexes. The interaction profiling revealed that CD93+/CD34+ ECs exhibited an extroverted phenotype, engaging in frequent communication with both LYVE+ endothelial cells and fibroblasts (Fig. 7b). Gene expression analysis via violin plots demonstrated distinct expression patterns across endothelial subpopulations: CD34 and CD93 were predominantly expressed in EC subpopulations 0-3, while LYVE showed elevated expression restricted to subpopulations 4-5. Interestingly, we observed widespread CD14 expression across multiple endothelial subpopulations. Given that CD14 serves as a canonical marker for bone marrow mononuclear cells (BMMNCs), its presence in endothelial cells suggests a possible transdifferentiation of circulating hematopoietic cells into the endothelial lineage (Fig. 7c). The immunofluorescence results indicated that CD93+/CD34+ cells were only existed on graft lumen, contributing to the formation of neo-endothelium. Furthermore, the quantity of CD93+/CD34+ cells at 4 weeks (34.5 ± 8.64 cells per field) was significantly less than that observed at 2 weeks (66.7 ± 10.60cells per field) (Fig. 7e, f). Due to the limited availability of rabbit immunofluorescent antibodies, we validated the correlation between LYVE+ endothelial cells and graft endothelialization using a rat abdominal aortic defect model. Following graft implantation, SAR131675, a VEGFR3 tyrosine kinase-specific inhibitor, was administered orally at 100 mg/kg/day in the inhibitor group, while the control group received an equivalent volume of saline (Fig. 8a). Arterial exploratory surgery results revealed a 100% (6/6) patency rate in the control group compared to only 50% (3/6) in the inhibitor group (Fig. 8b). Immunofluorescence analysis demonstrated visible endothelial cell deposits on the luminal surface of grafts at 2 weeks after implantation, with a significantly higher cell count in the control group than in the inhibitor group (Fig. 8c, d). The experimental group exhibited a substantial presence of LYVE+ cells within the graft wall, surpassing the count in autologous arteries and the inhibitor group (Fig. 8e). To further investigate the impact of LYVE+ endothelial cells in the early endothelialization of neovascular endothelium within grafts, en-face staining was performed on the neoendothelium and neovascular adventitia of the grafts. CD31+/CD93+ endothelial cells were observed at the anastomosis (178.3 ± 17.70cells per field) and midpoint (82.0 ± 15.90 cells per field) in the control group with a lower count. After SAR131675 administration, the inhibitor group exhibited reduced CD31+/CD93+ cell numbers across all neointimal sites (anastomosis, 75.0 ± 12.93 cells per field; midpoint, 22.8 ± 7.14 cells per field) compared to the control group (Fig. 8f, h). The CD93+/CD34+ cell numbers of all neoadventitia sites in inhibitor group (anastomosis, 30.5 ± 6.60 cells per field; midpoint, 10.3 ± 4.18 cells per field) were lower than the control group (anastomosis, 55.0 ± 11.88cells per field; midpoint, 40.0 ± 8.85 cells per field) (Fig. 8g, i).The LYVE+ cell numbers of all neoadventitia sites in inhibitor group (anastomosis, 149.3 ± 14.50cells per field; midpoint, 54.17 ± 13.38 cells per field) were lower than the control group (anastomosis, 222.7 ± 29.30 cells per field; midpoint, 103.8 ± 18.71 cells per field) (Supplementary Fig. 13a, b). These findings indicated a promotional role of LYVE+ cells in early graft endothelialization. Inhibiting LEC proliferation significantly impeded early endothelialization, leading to decreased graft patency. Inhibiting LECs impedes rapid endothelialization within 2 weeks. a Schematic illustration of MnVG implantation and SAR131675 administration in rats. b Number of patent grafts at 2 weeks. c Cross-sectional views of LYVE+ and CD31+ cells in neoarteries at 2 weeks. d, e Quantification of CD31+ cells in the lumen (d) and LYVE+ cells in the adventitia (e) at 2 weeks (n = 6 independent samples). f, g Immunofluorescence staining of CD93+/CD31+ cells (f) and CD93+/CD34+ cells (g) in the lumen from different regions at 2 weeks (en-face). h, i Quantification of CD93+/CD31+ cells (h) and CD93+/CD34+ cells (i) at 2 weeks (n = 6 independent samples). Plotted data represent means ± SD. The significance was determined by one-way ANOVA followed by Tukey’s post hoc analysis. The adjustments were made for multiple comparisons to get P-value. LYVE: green, CD31: red. Source data are provided in Source Data file. Discussion Theoretically, synthetic vascular grafts can achieve endothelialization through cell seeding, transanastomotic EC growth, and fallout healing driven by circulating EPCs2. Given the flushing force from the bloodstream and the urgent need for antithrombogenicity, fallout endothelialization is likely more efficient for vascular grafts with clinically available lengths. Techniques such as plasma treatment, chemical grafting, and nanotexturing have been employed to create a more biocompatible luminal surface20. The incorporation of bioactive molecules, such as peptides (e.g., RGD), growth factors (e.g., VEGF), and extracellular matrix (ECM) proteins, has shown promise in enhancing endothelialization21,22. Despite the encouraging results from the aforementioned studies, providing rapid and durable signals to recruit EPCs for accelerated endothelialization in vascular grafts remains a challenge in clinical translation. Our approach diverges from these methods by emphasizing the rapid regeneration of the neoadventitia. We have found that lymphatic endothelial cells (LECs) in the adventitia play a quick and durable role in recruiting circulatory EPCs for endothelialization. To the best of our knowledge, fabricating grafts capable of regenerating LECs-enriched adventitia represents a design perspective for vascular substitutes. Fabricating artificial blood vessels that achieve an optimal balance between openness and mechanical stability remains a significant challenge. Although conventional synthetic grafts composed of polyester and expanded polytetrafluoroethylene demonstrate adequate mechanical strength, their dense architecture severely compromises wall permeability, thereby impeding cellular recruitment and endothelialization processes23. Conversely, rapidly degradable materials such as poly (glycerol sebacate) (PGS) initially provide mechanical stability and enhanced permeability through degradation-induced pore formation, their rapid degradation kinetics, however, not only compromised safety in human, but also provoke substantial inflammatory responses that hinder VPCs recruitment. This temporal mismatch between material degradation and vascular regeneration ultimately compromises endothelialization efficiency9,14. To address these limitations, we proposed that precise fabrication of slow-degrading biomaterials with controlled topological features could potentially reconcile the competing demands of wall permeability and mechanical stability. Building upon Yao et al.‘s work, which established the maximum ratio of polycaprolactone (PCL) fiber diameter to spacing in electrohydrodynamic (EHD) printing and optimized the high-resolution EHD printing technique that can fabricate cell-sized structures, we developed a integrated manufacturing approach combining high-resolution EHD printing with roller-based electrospinning. This advanced fabrication technique enabled the construction of bilayered vascular grafts with precisely controlled PCL fiber architecture, achieving a fiber diameter-to-spacing ratio of 1:518. The results proved that the high-resolution EHD printing facilitates the creation of cell-scale PCL fiber structures with enhanced mechanical properties while maintaining biocompatibility. The optimized fiber dimensions promoted directional cellular adhesion and growth of endothelial cells, fibroblasts, and Schwann cells along the fiber orientation24. Notably, increasing fiber diameter enhances mechanical properties, resulting in a thin outer scaffold with good mechanical stability and maintaining a short infiltration and growth distance for perivascular cells25. The integrated EHD-printed fibers formed a uniform, continuous external opening construct that provided mechanical support for the internal ES layer19. The surface roughness of blood-contacting materials is a critical determinant of hemocompatibility, as increased roughness has been shown to promote thrombus formation and impair endothelial cell function26. In this study, the surface roughness of the ES layer (8.01 ± 2.18 μm) was found to be comparable to that of native arterial endothelium (1.04 ± 0.47 μm) and substantially lower than that of the EHD layer (153.60 ± 9.46 μm)27. This suggests that the ES layer more closely mimics the topological features of the natural endothelium compared to the EHD layer. In line with this structural similarity, the DTβ4-modified ES layer demonstrated pronounced anticoagulant properties and supported the adhesion of circulating CD93⁺ endothelial cells. DTβ4 exerts its potent antithrombotic effect through platelet inactivation and suppression of fibrinogenesis9. Based on these findings, we strategically employed the ES layer to fabricate the luminal surface of MnVGs, aiming to minimize platelet adhesion, reduce thrombotic risk, and enhance the recruitment of endothelial progenitor cells—thereby fostering rapid and sustained endothelialization. The EHD layer, with a pore size of ~200 µm, which was substantially larger than typical cells (20 µm), may allow complete infiltration by both HUVECs and AAFs28. This opening structure facilitates the rapid formation of neoadventitial tissue, as was confirmed in grafts harvested at 2 weeks after graft implantation. Moreover, the reduced thickness of the EHD layer shortens the spatial gap between the perivascular tissue and the ES layer, thereby accelerates integration and promotes the regeneration of a structured neo-adventitia. Critically, this EHD-supported architecture enabled the implementation of the 30 µm-thick ES layer, which effectively prevented blood extravasation while maintaining selective cellular permeability. The role of the adventitia, however, is complex and cell-type-dependent. Although infiltrating fibroblasts that can differentiate into pro-fibrotic myofibroblasts risk promoting intimal hyperplasia, adventitial endothelial cells engage in constructive crosstalk with luminal endothelial cells, facilitating functional graft remodeling29,30. It is therefore essential that the ES layer exhibits selective permeability, physically blocking fibroblast infiltration while permitting beneficial interlayer communication. Our in vitro experiments confirmed that a 30 µm-thick ES layer achieves this, preventing fibroblast migration while allowing endothelial cell passage. We further established that this selectivity is governed primarily by the physical microstructure of the ES layer (e.g., pore size, fiber density). The key finding of this study is that MnVGs achieved rapid endothelialization within 2 weeks postoperatively, resulting in a 90% patency rate (18/20) in 1.5 cm carotid artery reconstruction in rabbits. We believe three key features of MnVGs are crucial: (1) A porous graft wall with precisely controlled thickness, pore size, and porosity; (2) The permeability of the electrospinning lumen, which facilitates intraluminal passage of adventitia-derived signals31 and (3) A luminal interface with mechanical and chemical stability32. We suspected that the porous graft wall is the most critical factor, as it enables rapid transmural ingrowth of the adventitia. Unlike traditional methods such as salt leaching and electrospinning, melt electrostatic writing creates highly interconnected channels for cell infiltration without requiring scaffold degradation, thereby addressing the delayed stem cell homing caused by inflammation. Here, the total wall thickness is only 150 µm, which significantly reduces the distance between the luminal surface and perivascular tissues. Most vascular grafts, such as PGS elastomeric grafts and collagenous extracellular matrix (ECM) tubes engineered by cultured smooth muscle cells (SMCs) or allogenous vessels, undergo host remodeling through scaffold degradation15,33,34. In these grafts, increased wall thickness is necessary to ensure adequate strength, which, in turn, leads to significant inflammation that compromises the early recruitment of EPCs and vascular progenitor cells (VPCs)35,36. In addition, the luminal lining of ultrathin PCL nanofibers avoids the aforementioned issues by shielding the collagenous ECM of perivascular cells, thereby preventing acute thrombosis. Incorporating DTβ4 into PCL nanofibers may provide short-term anticoagulation potential to prevent embolism. In this study, along with in vitro experiments, we further confirmed the selective permeability of the ultrathin PCL lining. Lymphatic endothelial cells (LECs) were prioritized for transmural migration, whereas fibroblasts, the primary cells responsible for producing collagenous ECM, were blocked. By leveraging the precise control of PCL fiber diameters and spacing through melt electrostatic writing, matching mechanical properties such as burst pressure and suture retention can be achieved without increasing the thickness of the electrospinning lumen. This is essential for maintaining the permeability of the graft walls and ensuring signaling communication between the neoadventitia and circulatory cells. Circulating blood, abundant in progenitor cells, including various subsets of endothelial progenitor cells (EPCs), supports arterial hemostasis and vascular repair37. In 1998, Kouchi identified islands of endothelial cells (ECs) and smooth muscle cells (SMCs) in Dacron vascular grafts used for bypasses, confirming their origin from bone marrow-derived circulating cells38. Anstine traced CD45+ cells, revealing that bone marrow-derived monocytes contribute to cardiac valve cell renewal and repair, with age-dependent variation39. We previously demonstrated that CD93+ ECs in circulating blood are crucial for early endothelialization in rabbit carotid artery grafts, primarily due to their high migratory and proliferative potentials9. In this study, we identified their cross-talk with lymphatic endothelial cells (LECs) derived from the neoadventitia, thereby identifying an approach to providing long-lasting chemotactic signals for recruiting CD93+ cells. These LECs were widely detected in microvascular grafts (MnVGs) as early as 2 weeks and play important roles in consistent intima formation on the luminal side of the grafts. The neoarteries presented circumferentially aligned elastic fibers and collagens, exhibited drug-responsive actions similar to those of native arteries, and demonstrated superiority in arterial regeneration induced by LECs enriched adventitia. In consistence with findings from our previous studies, the endothelial cells on the luminal side of MnVGs are distributed in a multi-point, scattered pattern, which aligns with the principles of fallout endothelialization9,14. However, it should be noted that the distribution of endothelial cells is not uniform, with more ECs recruited near the anastomotic sites and slower endothelialization in the midportion of the grafts. Although we did not observe uneven growth of neoadventitia in the graft walls, examination of lymphatic ECs revealed the same cellular distribution pattern in the neoadventitia, extending from the anastomosis to the midportion. This finding confirmed the interplay between LECs and luminal endothelialization, suggesting that enhancing the migration and proliferation of LECs from surrounding tissues could be a therapeutic target to improve endothelialization in vascular grafts, especially for long-distance vessel reconstruction in large animals. The adventitia contains lymphatic vessels that contribute to vascular homeostasis40. In long-distance grafts, the observed declining gradient of LEC density from the anastomosis toward the mid-graft is a common regenerative pattern41. This distribution arises not by chance but from defined biological constraints: the anastomoses act as primary entry points for LECs migrating from the native adventitia, yet their migration capacity is intrinsically limited over long distances within the early post-implantation period. Consequently, the LECs of the mid-graft region is only from the surrounding peripheral tissues. Therefore, the gradient encapsulates the spatiotemporal dynamics of a multi-source lymphatic regeneration process, where location dictates the sources of LECs41. Substantial evidence has established that lymphatic vessels play a pivotal role in maintaining tissue homeostasis and facilitating regenerative repair across various human tissues and organs, including bone, aorta, liver, and the hematopoietic system42. Zhang et al., demonstrated that the transcription factor Tbx1 is specifically activated in cardiac LECs post-myocardial infarction (MI), driving LECs migration, neolymphangiogenesis to promote tissue repair43. Similarly, in bone, intestine, and other organs, the inhibition of lymphatic ECs severely impairs post-injury repair (e.g., after radiation or chemotherapy)44,45. The findings in this research corroborated these observations, revealing that LEC inhibition impairs early graft endothelialization and diminishes patency, thereby highlighting the pivotal role of LECs in vascular graft integration. CellPhone DB analysis revealed significant cell-to-cell communication between LECs and CD93+ ECs, with stronger interactions observed during the early implantation stage compared to the late stage. We found that LECs have cross-talk with CD93+ ECs by cell adhesion-related ligands such as CDH5 and IGF. Our findings suggested that LECs may facilitate early endothelialization by promoting the adhesion and migration of CD93+ ECs. Besides, we observed a heightened abundance of LECs in the neo-adventitia compared to native adventitia, with the percentage of LECs reduced over time. This temporal dynamic suggests a sophisticated remodeling process of LECs in neo-adventitia. Further supported by pseudotime analysis, certain LYVE+ cells retained pluripotent potential and persist in an undifferentiated state. We postulated that, as adventitial remodeling advances, a subset of neo-LECs may transdifferentiate into vascular endothelial cells, thereby contributing to the genesis of adventitial nourishing vessels. This plasticity is consonant with the findings of Karina Yaniv et al., who elucidated the capacity of LECs to transdifferentiate into vascular endothelial cells, facilitating the formation of specialized vascular architectures46. Collectively, these insights underscore the critical role of LECs in vascular graft integration and regeneration, mediated through mechanisms encompassing cellular plasticity, and paracrine signaling. Targeting LEC-mediated pathways may thus herald a promising therapeutic paradigm to enhance vascular graft efficacy and advance regenerative medicine. CD93+ ECs have emerged as crucial players in angiogenesis, tumor metastasis, and artificial vascular endothelialization, representing a population of naive endothelial cells with significant regenerative potential47. Previous investigations have demonstrated that CD93 expression is markedly upregulated during endothelial cell proliferation, undergoing dimerization to enhance its binding affinity for ligands such as insulin-like growth factor-binding protein 7 (IGFBP7)47,48. This molecular interaction facilitates critical endothelial processes, including cellular adhesion, migration, and vascular regeneration. In this study, the result uncovered a communication pathway between LECs and CD93+ ECs mediated by insulin-like growth factor (IGF). This finding revealed that circulating CD93+ ECs exhibited selective deposition on the luminal surface of vascular grafts in response to LEC-derived signals, rather than through random adhesion as previously hypothesized. This targeted recruitment mechanism bears obvious similarity to the precision of aerial deployment systems guided by ground-based communication. Through systematic LEC inhibition experiments, we demonstrated that disruption of this signaling cascade significantly impairs endothelialization processes. Based on these observations, we may refine the Fallout endothelialization to “Airdrop Endothelialization”: circulating endothelial cells selectively home to sites of endothelial damage through specific molecular signals such as IGF of neoadventitia derived LECs. This process is characterized by dynamic CD93 expression patterns, with elevated levels during regeneration followed by significant downregulation upon endothelial maturation and vascular repair completion. This study has several limitations that warrant further consideration. First, MnVGs did not maintain optimal lumen morphology under bending conditions, which may restrict their flexibility with neck movement. Future refinements in the layer topology are essential to meet the mechanical and functional demands in clinical cases. Second, although the interaction between LECs and CD93+ ECs was identified here, the precise spatial localization and detailed mechanisms of intercellular communication remain elusive. Further investigations employing spatial transcriptomics are necessary to perform. Additionally, although the long-term outcomes in rabbit carotid arteries are encouraging, it should be acknowledged that the endothelialization and vascular remodeling processes in small to medium-sized animals, such as rabbits, remain different from those in humans. Therefore, finding suitable chemical agents to improve LECs migration in neoadventitia may validate vascular regeneration of MnVGs in larger animal models, such as pigs or non-human primates (e.g., cynomolgus monkeys)49. The translation of artificial vascular grafts (MnVGs) for large animal carotid artery repair is contingent upon overcoming two interrelated challenges: matching the complex hemodynamic and anatomical environment, and orchestrating timely graft remodeling within a stringent metabolic and immune milieu. While our study demonstrates that EHD printing enables precise control over graft mechanics through layered PCL fiber deposition, the extended repair lengths (>4 cm) and larger graft volumes in these models demand more than structural fidelity. They impose dual requirements for exceptional hemocompatibility to maintain patency, and for high-efficiency cellular recruitment to drive endothelialization and adventitial regeneration. To this end, we may further design “perivascular drug delivery systems”, rather than relying on systemic anticoagulation. Such systems may sustain release of anticoagulating agents at the implantation site to maintain graft patency while avoiding systemic side effects. Moreover, they can be loaded with multiple regenerative drugs to modulate key processes in graft integration and remodeling, such as recruiting adventitial lymphatic endothelial cells. These efforts may rescue the insufficient healing potential of adventitia and endithelialization in large animals and human. Methods Fabrication of Micro-Nano Hierarchical Vascular Grafts (MnVGs) The MnVGs were constructed from polycaprolactone (PCL) (Jinan Daigang Biomaterial Co., Ltd China). Stainless steel rollers were coated with a 12% weight/volume (w/v) polyvinyl alcohol (PVA) solution and subsequently dried in a vacuum oven at 25 °C for future processing. Electrospinning was performed on the rollers for 3 min using a PCL + DTβ4 spinning solution. This solution was prepared by dissolving PCL in 2,2,2-trifluoroethanol at a concentration of 12.5% w/v and mixing it with DTβ4 solution at a 4:1 ratio by volume. After electrospinning, the rollers were mounted on a custom EHD printing platform (Shaanxi Baipusheng Medical Technology Company, China) for external sheath printing. Specifically, PCL was melted at 90 °C in a heated glass syringe and extruded through a charged stainless steel nozzle with an inner diameter of 150 μm, driven by a pneumatic control system. Conductive glass was used as the printing substrate. The EHD printing process parameters were set at 3.0 kV voltage, 1100 mbar pressure, 35 μm fiber diameter, and 200 μm fiber spacing. Vascular grafts with varying wall thicknesses were produced by successively applying 1, 3, and 10 outer sheathing layers on the rollers. The grafts underwent a de-PVA process, which involved immersion in 3% (w/v) heparinized saline solution followed by three cycles of rinsing with the same solution. The resulting ultra-thin 3D micro-nano hierarchical vascular grafts were freeze-dried under vacuum, sterilized using ethylene oxide, and stored at -80°C for subsequent experiments. Characterization of MnVGs Morphology: the morphology of MnVGs was documented after fabrication using a high-definition CDD camera. The micro-morphology of the grafts was examined via scanning electron microscopy (SEM) using the following standard protocols: fixation in glutaraldehyde, dehydration in a graded ethanol series, soaking in hexamethyldisilane, and drying. Subsequently, specimens from different regions of the grafts were collected, sputter-coated with platinum, and observed under SEM to assess their microscopic morphology. SEM images were analyzed to quantify the internal diameter, wall thickness, nanolayer core thickness, and cross-sectional images using ImageJ (National Institutes of Health, version 1.53e). Mechanical characterization: the tensile and compressive properties of the vascular grafts were evaluated using a BOSE force measurement system (ElectroForce 3200 Series II, USA). Segments of MnVGs (n = 3) were prepared at 20 mm in length for mechanical testing. Tensile strength was determined by applying uniaxial force to the segments at a 5 mm/min rate, starting at 0.1 N until fracture occurred. Compressive strength was evaluated by subjecting each segment of MnVGs (n = 3) to 50% strain. The rupture pressure and compliance of the neo-artery were measured using a perfusion system (MFCS-EZ microfluidic flow control system, Fluigent, France). EHD printing exoscaffold biosafety and cell infiltration experiments: 10-layer EHD printing exoscaffolds were 3D printed on conductive glass for in vitro cell inoculation and infiltration experiments. The exoscaffolds were submerged in matrix gel to align the upper surface with the liquid level and incubated at 37°C overnight to facilitate hardening. Rat aortic adventitia fibroblasts (AAFs) and lymphatic endothelial cells (LECs) were inoculated on the gel surface and cultured for 3 days. After aspiration the culture medium, the samples were fixed with paraformaldehyde for 20 min, washed three times with PBS, and prepared for pathological sectioning and immunofluorescence staining to observe cell infiltration. The cytoskeleton staining kit of Alexa Fluor 594 phalloidin (A12381) was purchased from Invitrogen (USA). Evaluation of anticoagulant In vitro: A specially designed circular blade (R = 8 mm) was used to cut the scaffolds, which were subsequently electrospinning with solutions of PCL combined with PBS, heparin, DTβ4, and VEGF, respectively. Four scaffold types were prepared for anticoagulation assessment (G-PCL + PBS, G-PCL + HEP, G-PCL + DTβ4, and G-PCL + VEGF). Fresh anticoagulated whole blood (1 ml) was recalcifiedand incubated with each MnVG sample in an EP tube at 37 °C for 10 min. The grafts, embedded in a blood clot, were then washed five times with PBS to eliminate unbound erythrocytes. Subsequently, the samples were fixed in a 2.5% glutaraldehyde solution in PBS at 4 °C for 6 h. Following fixation, the samples were processed and imaged using a SEM (PhenomXL, Phenom-World, Netherlands) following the aforementioned SEM sample preparation protocol. In vivo: For the acute anticoagulation evaluation test, carotid artery implantation was performed using 15 mm long MnVGs (5 in the PBS-MnVGs group, 5 in the Hep-MnVGs group, 5 in the DTβ4-MnVGs group, 5 in the VEGF-3DVGs group, the implanted grafts were harvested after 2 h). All procedures were performed by an operator blinded to the graft groups. Cell cultured Rat aortic adventitia fibroblasts (AAFs) were purchased from Pricella (CP-R219, Wuhan, China) and rat lymphatic endothelial cells (LECs) were purchased from ZQXZbio (PRI-RAT-00133, Shanghai, China). Human umbilical vein endothelial cells (HUVECs) were purchased from Lonza (CC2517A, Lonza, Switzerland). Cells were expanded in their respective matching culture media (CM-R219, Pricella; PCM-R-133, ZQXZbio; CC-3162, Lonza) at 37 °C in a humidified incubator with 5% CO₂, Cultures were maintained under standard conditions, with medium changes performed every 2 days to ensure optimal growth and viability. Animal experiments All animal procedures were conducted in accordance with the guidelines approved by the Fourth Military Medical University Animal Experiment Ethics Committee and adhered to the relevant animal welfare regulations (Ethical review report: No.2024kq011). Rabbits and rats were obtained from the Animal Center of the Fourth Military Medical University in Xi’an, China. Male New Zealand white rabbits (6 months old, weighing 3.5-4.5 kg; n = 32) underwent implantation of 15 mm long MnVGs interposition grafts in the left carotid artery. The experimental cohorts included a control group (Con-2W; n = 6) and a test group (MnVGs; n = 26). Additionally, male Sprague-Dawley (SD) rats (8 weeks old, weighing 300−350 g) underwent abdominal aortic interposition with 8 mm-long MnVGs grafts. The rat cohorts were divided into an inhibitor group (n = 6) and a test group (MnVGs; n = 6). Rats in the inhibitor group were administered the VEGFR3 tyrosine kinase inhibitor SAR131675 (100 mg/kg) via gavage for 2 weeks after surgery. All animals were male. Sex was not considered in the design, as the study focused on graft mechanics and cellular responses independent of sex differences. New Zealand white rabbits received 1000 UI of heparin subcutaneously 3 days before the operation. Subsequently, heparin (800 UI/kg) was subcutaneously injected into the animals 1 week after the operation, with a subsequent reduction to 400 UI/kg over the 3 months. Animals were housed under standard laboratory conditions with a 12 h light/dark cycle and free access to food and water. Biochemical analysis Tissue samples were collected from neonatal and adult carotid arteries (n = 5) for quantitative analysis of collagen and elastin content. The Sircol Collagen Assay kit (Biocolor, Carrick fergus, UK) was used to measure the total collagen in each tissue explant. Following the kit protocol, the collagen concentration in the resulting supernatant was assessed, and the insoluble collagen per unit wet weight was calculated based on a standard collagen curve. For elastin quantification, the total elastin from each tissue explant was isolated and analyzed using the Fastin elastin assay (Biocolor, Carrick Fergus, UK) according to the manufacturer’s instructions. The measured values for collagen and elastin were normalized to the wet weight of the samples to determine the insoluble elastin content (μg/mg tissue wet weight). Ultrasound patency monitoring The patency of grafts and the integrity of blood flow were monitored using high-frequency ultrasound imaging. The diameters and blood flow velocities in the contralateral autologous carotid artery and implanted grafts were assessed at 14, 30, and 90 days post-implantation using a Vevo 2100 ultrasound imaging system (FUJIFILM VisualSonics, Toronto, Canada). Animals were anesthetized with isoflurane (1-2%) administered via a precision vaporizer and gas flow. Long-axis images of the autologous carotid arteries and grafts were acquired in B-mode, color Doppler mode, and PW mode throughout the cardiac cycle. Cinematic playback images was analyzed using Vevo 2100 software to evaluate graft patency, diameter variations, and flow velocities within and surrounding the grafts. Histological evaluation The neo-arteries were embedded in OCT (Sakura Finetek, USA) for histological analysis. Serial frozen sections (10 µm thick) were prepared using a cryostat microtome. The sections were stained with hematoxylin for 10 min, followed by immersion in 70% ethanol and eosin for 5 min, and dehydration in ethanol to complete the H&E staining protocol. Masson’s trichrome staining (G1340, Solarbio, Beijing, China) and elastic van Gieson (EVG) (ab150667, Abcam, MA, USA) were performed according to the manufacturers’ instructions. Histological images were captured using an upright fluorescence microscope (DM6000 B, Leica, Germany) under bright-field and polarized light conditions to evaluate tissue morphology and structural integrity. Quantification of PCL residues in grafts The residual PCL within vascular grafts was quantified using established protocols to assess the degradation profile and integration of the polymer over time. Polarized light microscopy was employed to visualize PCL residues, which exhibit distinct white birefringence under polarized illumination, in contrast to non-birefringent tissues. H&E staining was performed on six samples per group, and the stained sections were imaged using bright-field microscopy. The polymer residual area was quantitatively assessed using ImageJ (National Institutes of Health, version 1.53e) by analyzing the birefringent regions within the images. This approach enabled precise quantification of PCL residues, providing insights into the degradation and tissue integration of the graft material. Immunofluorescence staining Immunofluorescence staining was performed on tissue samples to elucidate the cellular and ECM components within the grafts. Samples were fixed with ice-cold acetone for 20 min to preserve cellular structures and antigens. Cellular membranes were permeabilized with 0.5% Triton X-100 in PBS for 10 min, followed by blocking of non-specific binding sites with 5% bovine serum albumin (BSA; Novus, NB 100-2076) for 60 min at room temperature. The following primary antibodies were used: Mouse Anti-CD31 (1:300 dilution; Novus, NB600-562), Mouse Anti-Elastin (1:300 dilution; Novus, NB100-2076), Mouse Anti-collagen I (1:300 dilution; GeneTex, GTX26308), Mouse Anti-collagen III (1:300 dilution; GeneTex, GTX26310), Mouse Anti-α-SMA (1:300 dilution; proteintech, 67735-1-Ig), Rabbit Anti-LYVE (1:300 dilution; Novus, NB100-725), Rabbit Anti-CD34 (1:300 dilution; GeneTex, GTX28158), Mouse Anti-CD93 (1:300 dilution; Abcam, sc365172), Goat Anti-CD31 (1:300 dilution; R&D Systems, AF3628), NR2F2 (1:300 dilution; Abcam, ab41859). Samples were incubated with primary antibodies for 12 h at 4 °C. After washing with PBS, samples were incubated with fluorescent secondary antibodies for 1 h at 37 °C. The following fluorescent secondary antibodies were used: Goat Anti-Rabbit IgG H&L(Alexa Fluor 488) (1:500 dilution; Abbkine, A23220), Goat Anti-Rabbit IgG H&L(Alexa Fluor 594) (1:500 dilution; Abbkine, A23420), Goat Anti-Mouse lgG H&L(Alexa Fluor 488) (1:500 dilution; Abcam, ab150117), Goat Anti-Mouse IgG H&L(Alexa Fluor 594) (1:500 dilution; Abcam, ab150120), Donkey Anti-Goat IgG H&L(Alexa Fluor 594) (1:500 dilution; PCM, PC-80119). Nuclei were stained with 4′,6-diamidino-2-phenylindole to facilitate cell identification. Imaging was performed using a laser scanning confocal microscope (FV 1000, Olympus, Japan) and a stereo fluorescence microscope (Leica, MF 205, Germany). Different cell populations were identified based on cell counts from each immunostained image on five different sections (center and 12, 3, 6, and 9 o’clock positions, respectively). Data were collected from at least six different MnVGs in each group. Neoarterial adventitia and intima tissue evaluation To evaluate the development and organization of the luminal endothelium, neo-arteries were collected at 14, 28 and 90 days. Samples were prepared for SEM and coated with platinum to assess endothelial cell morphology at the ultrastructural level. For immunofluorescence staining, the adventitia and intima tissues of neovascular tissue were meticulously separated under a microscope, permeabilized with 0.5% Triton X-100, and fixed with 4% paraformaldehyde. Subsequently, the tissues were blocked with 5% normal goat serum at 37 °C for 2 h. The inner membrane tissue was incubated with a 1:100 dilution of CD31 primary antibody (Novus, NB 600-562), while the outer membrane tissue was incubated with a 1:100 dilution of CD31 primary antibody and LYVE primary antibody. After incubation, the samples were treated with secondary antibodies Alexa Fluor 594 IgG (1:500) and Alexa Fluor 488 IgG (1:500). Immunofluorescence images were captured using laser confocal microscopy (FV 1000, Olympus, Japan), and analyzed using ImageJ (National Institutes of Health, version 1.53e) to quantify cellular distribution and tissue organization50. Transmission Electron Microscopy (TEM) analysis To investigate the ultrastructural features of neo-tissues in arterial grafts, TEM imaging was performed using an H7500 transmission electron microscope (Hitachi, Japan). Arterial grafts were prepared as 2 mm × 2 mm × 2 mm cubes and subjected to the following protocol: Samples were washed three times in PBS for 5 min each, fixed in 2.5% glutaraldehyde, and dehydrated through an ascending series of ethanol solutions (30%, 50%, 70%, 90%, and 100%). Subsequently, samples were infiltrated with epoxy resin. Ultrathin sections were cut using an ultramicrotome and stained with 2% uranyl acetate, followed by lead citrate staining to enhance contrast51. Images were captured at random locations across the sections to ensure representative sampling of the graft ultrastructure. Isometric force measurement To assess the vasoreactivity of neo artery, isometric force measurements were conducted using an in vitro organ bath system, which was conducted according to the methods described by Liu et al. The specific methods were as follows: neo-artery rings (3 mm in length) were mounted onto two stainless steel hooks and suspended in a four-chamber organ bath (Radnoti, USA). One hook was attached to a plastic holder, while the other was connected to a tension transducer interfaced with PowerLab (ADInstruments, Australia) for real-time recording. The chamber was filled with Krebs buffer solution, maintained at 37 °C, and continuously gassed with 95% O₂ and 5% CO₂52. After a 30 min equilibration period, rings were stimulated with 60 mM KCl and washed until the initial resting force was re-established. The artery rings were then pre-contracted with 10⁻⁶ M phenylephrine (PE), followed by concentration-response curve assessments using 10–100 mM KCl, 10⁻¹⁰–10⁻⁴ M PE, and acetylcholine (ACh)53. Vasodilatory responses were normalized to PE-induced contractions and expressed as a percentage. Vasoconstrictive and vasodilatory capacities were quantified as the area under the concentration-response curves (AUC). Single-cell RNA library preparation and sequencing Neo-arteries collected at weeks 1, 2, and 4 were rinsed thrice with PBS before undergoing single-cell RNA sequencing (scRNA-seq) analysis conducted by the NovelBio Bio-Pharm Technology Co. Ltd. Laboratory. The harvested tissue was immediately placed in MACS tissue storage solution (Miltenyi Biotec) for preservation. Subsequently, the tissue was processed by washing with PBS, cutting into ~1 mm3 fragments on ice, and enzymatically digested with collagenase I (2 mg/ml; Worthington) and dispase II (1 mg/ml; Worthington) for 50 min with agitation. Following digestion, the sample was filtered through a 70 μm cell strainer and centrifuged at 300 g for 5 min. The resulting cell pellet was then treated with red cell lysis buffer (Miltenyi Biotec) to remove red blood cells. After washing with PBS containing 0.04% BSA, the cell pellet was resuspended in PBS with 0.04% BSA, filtered through a 35 μm cell strainer, and stained with calcein AM (Thermo Fisher Scientific) and Draq7 (BD Biosciences) for viability assessment. The BD Rhapsody system is utilized for capturing transcriptome information at the single-cell level. A single-cell suspension is distributed randomly into >200,000 micropores through a limited dilution method to ensure individual cell capture. Oligonucleotide barcode-labeled beads are introduced to saturation, enabling pairing with cells within the micropores. Subsequently, cells lyse within the micropores, facilitating mRNA hybridization with barcode-tagged oligonucleotides on the beads. The beads are then pooled into a single tube for reverse transcription and Exo digestion. During cDNA synthesis, each cDNA molecule is uniquely tagged with a molecular identifier (UMI) and a cell barcode at the 5’ end (corresponding to the 3’ end of the mRNA transcript), denoting its originating cell. Comprehensive transcriptome libraries are generated using the BD Rhapsody single-cell full transcriptome amplification (WTA) protocol, encompassing random priming and extension (RPE), RPE amplification via polymerase chain reaction (PCR), and WTA indexed PCR. Library quantification is conducted using a Bioanalyzer 2200 with high sensitivity DNA chips (Agilent) and Qubit High Sensitivity DNA assays (Thermo Fisher Scientific). Subsequent sequencing is carried out on Illumina sequencers (Illumina, San Diego, CA) utilizing 150 base-pair paired-end sequencing runs. Bioinformatics analysis Single-cell RNA sequencing (scRNA-seq) data analysis was conducted by NovelBio Biopharmaceutical Technologies via the NovelBrain Cloud Analysis Platform. We utilized fastp with default settings to filter adapter sequences and eliminate low-quality reads, yielding clean data. The UMI tool facilitated single-cell transcriptome analysis by identifying cell barcode whitelists. Clean UMI-based data were mapped to the rabbit genome (OryCun2.0 Ensembl) using STAR with custom parameters from the UMI-Tools standard pipeline to obtain UMI counts per sample. Cells with over 200 expressed genes and a mitochondrial UMI rate below 20% passed quality screening, and mitochondrial genes were excluded from the expression table. The Seurat package (version 3.1.4; https://satijalab.org/seurat/) was employed for cell normalization and regression based on UMI counts and mitochondrial rates, producing scaled data. Principal component analysis (PCA) was performed on 2000 highly variable genes, and the first 10 principal components were used to construct t-distributed stochastic neighbor embedding (tSNE) and UMAP. A graph-based clustering method (resolution: 0.8) was utilized to perform unsupervised cell clustering based on the top 10 principal components from PCA. Marker genes were identified using the FindAllMarkers function and the Wilcoxon rank sum test, with the criteria of (i) lnFC > 0.25, (ii) P < 0.05, and (iii) min.pct > 0.1. To further characterize cell types, clusters representing the same cell type underwent re-tSNE analysis, additional graph-based clustering, and marker analysis. KEGG pathway analysis, based on the KEGG database, was employed to elucidate key pathways associated with differential gene expression. Fisher precision tests were applied to determine significant pathways, with significance thresholds established by P-values and false discovery rates (FDR). GO analysis: GO analysis aims to clarify the biological relevance of distinct genes within a notable subset of differentially expressed genes in an experiment. GO annotations are acquired from NCBI, UniProt, and GO databases. Fisher precision tests ascertain significant GO categories, with P-values corrected using FDR. Pseudotime Analysis: Monocle 2 was utilized for single-cell trajectory analysis employing DDR trees and default parameters. Marker genes were selected from Seurat clustering results and raw expression counts of filtered cells prior to Monocle analysis. Branch fate-determining gene analysis was conducted through branch expression analysis modeling based on pseudotime analysis. Cellphone analysis: CellPhone DB and CellChat are utilized to analyze cell-to-cell communication systematically by inferring interactions based on ligands and receptors. The pheatmap R package is employed to create a heatmap illustrating the number of interactions, while ggplot2 R is used for visualizing ligand-receptor interactions. Differential edges are processed through CircosDiff (a wrapper for the R package “circlize”) and netVisual_chord_gene in CellChat to filter receptor-ligand edges and produce a Circos map. Significance of mean values and cell communication is determined based on interactions and standardized cell matrices obtained through Seurat standardization (P < 0.05). Spatial transcriptome analysis Tissue preparation and sectioning: Resected tissues were embedded in OCT compound (Leica), flash-frozen, and stored at -80 °C. Serial 10-μm-thick sections were cut on a Leica CM1950 cryostat and mounted onto Visium spatial gene-expression slides (10X Genomics). Fixation, staining, and imaging: Sections were fixed in ice-cold methanol at -20 °C for 30 min, stained with hematoxylin and eosin (H&E), and imaged using a Pannoramic MIDI slide scanner (3DHISTECH). Determination of permeabilization time: The optimal tissue permeabilization time was established prior to the full assay using Visium Spatial Tissue Optimization Slides and the corresponding kit, following the manufacturer’s instructions. Tissue permeabilization and library preparation were performed using the Visium Spatial Gene Expression kit (10X Genomics). Following determination of the optimal permeabilization time using Tissue Optimization Slides, tissue sections on gene expression slides were permeabilized and incubated with Reverse Transcription Master Mix to capture and barcode mRNA transcripts. cDNA synthesis was performed, incorporating unique molecular identifiers (UMIs) and spatial barcodes. Whole-transcriptome amplification libraries were constructed following the BD Rhapsody WTA workflow, involving random priming and extension (RPE) and PCR amplification. Final libraries were quantified using the Qubit High-Sensitivity DNA assay (Thermo Fisher Scientific) and sized using a Bioanalyzer 2200 with a High-Sensitivity DNA chip (Agilent). Sequencing was performed on an Illumina platform with a 150 bp paired-end configuration (Illumina, San Diego, CA). Spatial Transcriptomics Data Analysis: Raw sequencing reads were quality-controlled and adapter-trimmed using fastp (default parameters). Processed reads were aligned to the Rnor6.0 genome (Ensembl 2020) and feature-barcode matrices were generated using SpaceRanger (v1.1.0)54. To mitigate batch effects, we performed a down-sampling analysis based on mapped reads per spot across samples, generating an aggregated matrix for downstream analysis. Data normalization, scaling, and regression were performed with the Seurat package (version 3.2, https://satijalab.org/ seurat/). Dimensionality reduction was carried out using PCA on scaled highly variable genes, followed by t-SNE visualization based on the top 10 principal components. Unsupervised clustering was conducted using a graph-based algorithm within Seurat. Marker genes for each cluster were identified using the FindAllMarkers function (Wilcoxon rank-sum test) with thresholds set at: adjusted p-value < 0.05, log₂ fold-change > 0.25, and min.pct > 0.1. Spatial gene expression was visualized using the SpatialFeaturePlot function (Seurat v3.1.3) and the STUtility package (v1.0.0). Gene set enrichment analysis was performed using QuSAGE (v2.16.1) to quantify pathway activity55. Statistical analysis All experiments included a minimum of three biological replicates. Statistical analyses were conducted using IBM SPSS Statistics (Version 23.0). Data are expressed as means ± standard deviation (SD). Group comparisons were assessed using the Student’s t-test for two groups and one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test for multiple groups. Statistical significance was set at P < 0.05. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary information Description of Additional Supplementary Files Supplementary Movie 1 Supplementary Movie 2 Supplementary Movie 3 Supplementary Movie 4 Reporting Summary Transparent Peer Review file Source data Source data Acknowledgements This project was financially supported by the National Natural Science Foundation of China (82071132 to W, 92268119 to W, 82525018 to W and 52125501 to H), and the Program for Innovation Team of Shanxi Province (2023-CX-TD-17 to H). We extend our gratitude to Dr. Huan Liu for her guidance on the isometric force measurement. We extend our sincere thanks to our colleagues at Liebing Company—Zhang Junquan, Li Bo, Shi Jiale, and Bi Xiaoyu—for their invaluable support and critical contributions to our work in single-cell sequencing and spatial transcriptomics. Author contributions Z.Z. and W.W. conceived the original concept and initiated this project. Z.Z., P.Y., C.Y., J.H. and W.W. designed the experiment and supervised the entire project. Z.Z. and Y.L. analyzed the scRNA-seq and spatial transcriptome data. C.Y. and Z.Q conducted the construction and characterization of MnVGs. Z.Z., P.Y., Y.L. and Y.H. conducted animal experiments and data analysis. W.Z. and X.Z. performed histological analysis. Y.Z. and X.W. carried out in vitro cell experiments. C.Z. synthesizes and supplies DTβ4. N.L., P.W. and W.Y. helped with data collection. Z.Z. and W.W. wrote the paper with edits from other authors. Peer review Peer review information Nature Communications thanks Shadi Houshyarand the other anonymous reviewer for their contribution to the peer review of this work. A peer review file is available. Data availability The single-cell RNA sequencing and spatial transcriptomics data generated in this study have been deposited in the Genome Sequence Archive (GSA) database under accession code CRA032443. These data are publicly accessible without restriction. Source data for the graphs and charts in all Figs. 1–9 and Supplementary Figs. 1–15 are provided as Source Data files. All other data supporting the findings of this study are available within the article and its Supplementary Information. Source data are provided with this paper. Schematic illustration of graft design and proposed mechanism for vascular regeneration. Fabrication: Diagram depicting the sequential fabrication of MnVGs using electrospinning (ES) to form the fibrous scaffold, followed by electrohydrodynamic (EHD) printing to create an integrated construct on a 2 mm diameter roller. Surgical implantation: Schematic of the graft being used to repair a 1.5 cm defect in a rabbit left carotid artery. At 2 weeks post-implantation: Illustration of key cellular events: (i) Recruitment of CD93+ endothelial progenitor cells to the luminal surface. (ii) Infiltration of lymphatic endothelial cells into the neo-adventitia. At 12 weeks post-implantation: Illustration depicting complete endothelialization and significant neovascular tissue maturation. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Zhe-qian Zhang, Ping-Ping Yuan, Cong Yao, Yu-xin Lu. Contributor Information Jian-kang He, Email: jiankanghe@mail.xjtu.edu.cn. Wei Wu, Email: wuweidds@126.com. Supplementary information The online version contains supplementary material available at 10.1038/s41467-025-68075-8. References Associated Data Supplementary Materials Supplementary information Description of Additional Supplementary Files Supplementary Movie 1 Supplementary Movie 2 Supplementary Movie 3 Supplementary Movie 4 Reporting Summary Transparent Peer Review file Source data Data Availability Statement The single-cell RNA sequencing and spatial transcriptomics data generated in this study have been deposited in the Genome Sequence Archive (GSA) database under accession code CRA032443. These data are publicly accessible without restriction. Source data for the graphs and charts in all Figs. 1–9 and Supplementary Figs. 1–15 are provided as Source Data files. All other data supporting the findings of this study are available within the article and its Supplementary Information. Source data are provided with this paper. Schematic illustration of graft design and proposed mechanism for vascular regeneration. Fabrication: Diagram depicting the sequential fabrication of MnVGs using electrospinning (ES) to form the fibrous scaffold, followed by electrohydrodynamic (EHD) printing to create an integrated construct on a 2 mm diameter roller. Surgical implantation: Schematic of the graft being used to repair a 1.5 cm defect in a rabbit left carotid artery. At 2 weeks post-implantation: Illustration of key cellular events: (i) Recruitment of CD93+ endothelial progenitor cells to the luminal surface. (ii) Infiltration of lymphatic endothelial cells into the neo-adventitia. At 12 weeks post-implantation: Illustration depicting complete endothelialization and significant neovascular tissue maturation.