Ex vivo endothelialized cECM-enriched core–shell fibrous vascular graft promotes rapid regenerative remodeling in vivo

Abstract

Occlusive arterial disease continues to be a leading contributor to morbidity and mortality, and the ongoing lack of small-diameter vascular grafts (<6 mm) significantly restricts surgical interventions. In this work, we present a new core/shell fibrous vascular graft (C/S PE) created through coaxial electrospinning, consisting of a polycaprolactone (PCL) core and a cardiac extracellular matrix (cECM)-enriched shell designed for ex vivo endothelialization. Proteomic profiling confirmed that the cECM preserved angiogenic and cell-adhesive components capable of guiding vascular lineage commitment. To enhance luminal cellular organization prior to implantation, a dynamic ex vivo perfusion strategy was applied to induce mesenchymal stem cell differentiation under gradually increasing shear conditions designed to emulate early physiological adaptation. This approach enabled the formation of a confluent endothelial-like layer after two weeks. RNA-seq analysis revealed activation of pathways associated with endothelial proliferation, angiogenesis, extracellular matrix remodeling, etc. Following implantation in the rat abdominal aorta model, the ex vivo endothelialized (C/S PE-EC) grafts maintained 100% patency and progressive host-driven remodeling characterized by endothelial regeneration, smooth muscle layer formation, and a shift toward a pro-healing macrophage phenotype. Importantly, the C/S PE (EC) grafts reached nearly complete endothelial coverage and functional eNOS expression within one month, suggesting the synergistic role of biomimetic matrix composition and biomechanical conditioning in accelerating vascular integration. By integrating synthetic strength, ECM bioactivity, and stem cell–derived endothelialization, this approach offers a promising pathway toward clinically translatable small-diameter vascular grafts, particularly appropriate for scheduled procedures such as Fontan surgery and hemodialysis access.

Graphical abstract

Highlights

• •A porous cECM-enriched core/shell fibrous vascular graft (C/S PE) was developed.
• •A custom bioreactor enabled >95% uniform endothelial luminal coverage within 2 weeks.
• •The cellularized C/S PE-EC graft showed 100% patency in the rat abdominal aorta.
• •The C/S PE-EC exhibited rapid endothelialization and uniform smooth muscle layer formation.
• •Promoted rapid M2 polarization for pro-regenerative remodeling.

1. Introduction

Globally, occlusive arterial disease remains a leading cause of morbidity and mortality. Despite progress in cardiovascular therapies, more than 350,000 coronary or peripheral arterial grafting surgeries continue to be performed annually in the United States [1]. The internal mammary artery and the saphenous vein are among the most frequently utilized autologous vessels. Nevertheless, invasive harvesting methods pose increasing challenges, particularly in elderly patients, those with previous bypass surgeries, or individuals with congenital heart defects, making autologous vessel collection more difficult [2,3]. Although large-diameter (ID > 6 mm) synthetic polytetrafluoroethylene (ePTFE) vascular grafts are routinely used and readily available on the market, their application in small-diameter arteries (ID < 6 mm) continues to yield unsatisfactory clinical results, largely due to thrombosis, suboptimal endothelialization, and intimal hyperplasia [4]. Current research efforts focus on addressing these limitations by modifying synthetic graft surfaces with growth factors, proteins, and pharmacological agents [5]. Additional innovative strategies involve tissue-engineered vascular graft (TEVG) methods, whereby synthetic or decellularized scaffolds are pre-endothelialized or seeded with stem cells to mitigate early graft failure [[6], [7], [8]]. Cellularized TEVG emerges as a highly promising option for Fontan procedures in pediatric patients with congenital heart defects [9]. Because the Fontan operation is elective and predominantly performed in children, cellularized TEVG is appealing due to its potential to accommodate patient growth. However, the uniform culture of cells throughout the graft lumen wall and achieving complete surface coverage remain major technical challenges and require considerable time (average percent coverage of 64%) [8]. Consequently, a more reliable and refined strategy is needed to enhance clinical outcomes.

Mesenchymal stem cells (MSCs) exhibit minimal MHC-I expression and virtually no expression of MHC-II, thus they are regarded as immunoprivileged [10]. Indeed, multiple studies have demonstrated that MSCs suppress both syngeneic and allogeneic T-cell proliferation to a similar extent, indicating that their immunomodulatory properties are independent of MHC expression [11,12]. Notably, preclinical and clinical investigations of allogenic MSCs have produced inconsistent outcomes. Some studies demonstrated favorable effects in skin graft applications for wound regeneration and in the management of systemic lupus erythematosus, while others reported inconsistent results in the context of solid organ transplantation [[13], [14], [15]]. In contrast, vascular grafts seeded with undifferentiated MSCs have yielded promising outcomes [16,17]. These MSC-seeded grafts demonstrated strong antithrombotic properties and successfully contributed to the regeneration of the endothelial layer in vivo. Additionally, a recent study reported that extracellular vesicles derived from MSCs improved graft patency via immunomodulation in a rat model [18]. Nevertheless, the in vivo effects of endothelial cells (ECs) derived from MSCs under dynamic culture conditions remain poorly understood.

It is also important to recognize that off-the-shelf application of a cell-seeded vascular graft remains unattainable, as successful vascular grafting necessitates dynamic 3D cell culture to ensure uniform cell distribution across the lumen wall. It is widely acknowledged that cell-free biomaterial-based vascular grafts are frequently selected as the preferred option in urgent clinical scenarios [19]. In our previous studies, we developed cell-free vascular grafts with superior bio-functional properties suitable for emergency settings [20,21]. The aim of this research is to develop a pre-endothelialized vascular graft designed for hemodialysis access and Fontan procedures. Unlike earlier strategies that relied on common dense ePTFE, polyglycolic acid (PGA), or decellularized arterial matrices for ex vivo endothelialization, the current investigation emphasizes engineering a core-shell fibrous vascular scaffold characterized by a highly porous architecture [[22], [23], [24]]. Co-axial electrospinning presents a valuable approach for producing core-shell fibers [25]. In comparison to conventional electrospinning, coaxial electrospinning provides specific benefits for generating core–shell micro/nanofibers. This technique permits the simultaneous inclusion of multiple biomolecules within both the core and shell layers, thereby supporting the fabrication of fibers designed for specific and multifunctional applications. Our previous research demonstrated that the cardiac extracellular matrix (cECM) sourced from porcine tissue represents a highly effective material for constructing vascular grafts [21]. cECM is rich in various collagen types that support cell adhesion and proliferation, as well as angiogenesis-associated proteins such as actin gamma 1 (ACTG1), fibronectin 1 (FN1), cadherin 13 (CDH13), and plasminogen (PLG).

Herein, we developed an innovative core-shell fibrous polycaprolactone (PCL)-cECM vascular graft (C/S PE) suitable for both cellular and acellular applications, comprising a PCL core surrounded by a cECM shell (Scheme 1). A custom-designed 3D bioreactor system was utilized to culture rat bone marrow-derived MSCs (rBMSCs) within the graft lumen. After 1 and 2 weeks of dynamic culture using endothelial conditioning media, the endothelial differentiation capacity of rBMSC-derived endothelial cells (rBMSC-ECs) was assessed. Following optimization, both pre-endothelialized (C/S PE-EC) and acellular C/S PE grafts were implanted into the rat abdominal aorta for one month to assess their in vivo efficacy.

Scheme 1.

Schematic illustration depicting the fabrication of the core/shell PCL-cECM (C/S PE) vascular graft and the cell seeding process. A novel bioreactor was constructed to culture rBMSCs under dynamic conditions to promote endothelial differentiation. Subsequently, the pre-endothelialized C/S PE (EC) was implanted into the rat abdominal aorta for biological assessment.

2. Results

2.1. Biochemical and proteomic characterization of decellularized cardiac extracellular matrix (cECM)

Fig. 1A outlines the decellularization of minced cardiac tissues. H&E, Masson's trichrome, and safranin O staining demonstrated a lack of nuclear material, confirming that decellularization was both effective and maintained the integrity of collagen and GAGs (Fig. 1B–D). The ELISA assays confirmed that the collagen type I and the sGAG contents were significantly higher in cECM compared to the native, whereas the elastin content was similar between the two groups (Fig. S1). The observed DNA content was less than 50 ng/mg dry weight, further supporting successful decellularization (Fig. S2A).

Fig. 1.

Extraction & characterization of the cardiac ECM (cECM). (A) Schematic illustration outlining the entire extraction and characterization process. (B) H & E staining, (C) Masson's trichrome, and (D) safranin O histological stains are presented to confirm decellularization outcomes. (E) LC/MS-MS spectra detail the molecular composition. (F) Scatter plot indicating the number of proteins with up- or down-regulated expression. (G) Gene Ontology (GO) enrichment analysis based on the combined set of differentially expressed proteins (both upregulated and downregulated), classified according to biological process (BP), cellular component (CC), and molecular function (MF). Heatmaps highlight several critical GO sub-categories: (H) wound healing (n = 3), (I) angiogenesis (n = 3), (J) extracellular matrix (n = 3), and (K) antioxidant activity (n = 3).

Fig. 1E presents LC-MS chromatograms derived from native and decellularized cardiac tissues. Proteomic analysis revealed 86 proteins with increased expression in the decellularized cECM, 100 upregulated proteins in native tissue, and 19 proteins without significant expression changes between groups (Fig. 1F). Proteins were classified by gene ontology (GO) enrichment analysis according to biological process (BP), cellular component (CC), and molecular function (MF) (Fig. 1G). This GO enrichment analysis identified key subcategories pertinent to endothelial cell proliferation and wound healing. To provide additional detail, protein expression heatmaps are presented for important subcategories, including GO:0042060 Wound Healing (Fig. 1H), GO:0001525 Angiogenesis (Fig. 1I), GO:0031012 Extracellular Matrix (Fig. 1J), and GO:0016209 Antioxidant Activity (Fig. 1K). These heatmaps indicate preservation of most proteins after decellularization.

2.2. Preparation and Physicochemical Characterization of C/S PE

The coaxial (core/shell) electrospinning method was utilized to produce small-diameter C/S PE grafts. Fig. 2A shows a 10 cm length of C/S PE graft with an estimated inner diameter of 2 mm. In our previous studies, we have reported that 10% PCL concentration produces thinner nanofibers, whereas 20% PCL concentration produces thicker microfibers [20,21]. The primary objective was to obtain a porous, relatively thick core-shell fibrous structure, optimal for subsequent cell seeding and promoting a cellularized lumen. That's why we decided to use a 20% PCL concentration for the core and a 12% PCL and 10% cECM blend as the shell to coat the core PCL. SEM images in Fig. 2B and C demonstrate the porous (average pore size is around 60.52 ± 4.74), fibrous, and thick cross-sectional architecture of the graft. The mean fiber diameter measured was 3.14 ± 0.88 μm (Fig. 2D). The surface morphology of the fibers is illustrated in Fig. 2E. Notably, SEM analysis revealed that the fiber cross-sections in some regions consisted of a distinct central core (PCL) and an external shell (PCL & cECM), as presented in Fig. 2F. The core-shell structure was further validated through fluorescence imaging (Fig. 2G) and TEM imaging (Fig. 2H), clearly displaying a red exterior shell encasing a black inner core (Fig. 2G). The distribution of the thickness of the core and shell regions was provided in Fig. 2I & J, respectively. The average thickness of the core was 0.54 ± 0.32 μm, and the average thickness of the shell was 0.25 ± 0.09 μm. The C/S PE group demonstrated a significantly reduced water contact angle (WCA) when compared to the PCL control group (p < 0.001), as shown in Fig. 2K. The WCA for PCL and C/S PE were 119.7 ± 5.6° and 82.6 ± 6.5°, respectively, indicating that the C/S PE graft is hydrophilic (<90°).

Fig. 2.

Preparation and Physicochemical Characterization of C/S PE. (A) Optical image of the C/S PE graft. SEM images representing the (B) whole cross-sectional and (C) magnified cross-sectional views. (D) Fiber diameter distribution of the C/S PE graft, quantified from randomly selected fibers measured from SEM images (n = 100 fibers). (E) SEM image showing the surface fibers. Core/shell morphology of the fiber (F) SEM image, (G) fluorescence image (red = shell and black in the middle = core), and (H) TEM image. Thickness distribution of the (I) core and (J) shell regions of the fibers (n = 50). (K) Water contact angle for evaluating hydrophilicity (n = 4). (L) Percentages of the residual mass in the PCL & C/S PE groups (n = 3). (M) Stress vs strain curve measured in circumferential and longitudinal directions (n = 3). (N) Ultimate tensile strength (O), percentages of elongation, and (P) Young's Modulus of the fabricated grafts (n = 3). (Q) Hemolysis ratio for hemocompatibility (n = 4) and (R) blood clotting index for thrombosis assessments (n = 5). (S) SEM images of platelets on the surface of the (i) PCL and (ii) C/S PE grafts. Statistical significance was calculated by two-way ANOVA with Tukey's test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. “N.S” means not significant.

Fig. 2L demonstrates the degradation profile of the PCL and C/S PE groups. The residual masses of the PCL and C/S PE groups after 28 days were 95.8 ± 0.7 and 90.5 ± 1.2, respectively. That suggests the degradation of C/S PE is significantly (p < 0.0001) greater than that of the PCL group. During the initial 7 days, both groups exhibited nearly identical degradation kinetics, suggesting that early-stage mass loss was governed primarily by surface hydration and limited polymer chain relaxation rather than material composition. However, from day 14 onward, the degradation trajectories diverged markedly, with the C/S PE group showing progressively greater mass loss at days 14, 21, and 28. This delayed divergence reflects the onset of biologically mediated degradation of the cECM-rich shell, while the PCL-only scaffold remained largely resistant due to its hydrophobicity and slow hydrolytic cleavage. This difference is particularly important in the context of small-diameter vascular graft design, where degradation must be synchronized with neotissue formation. Excessively rapid degradation can compromise mechanical integrity and promote aneurysmal dilation, whereas overly slow degradation, as seen in conventional PCL grafts, can delay cellular infiltration, prolong foreign body reactions, and limit constructive remodeling.

The tensile properties (stress vs strain) of the grafts were analyzed in both the longitudinal and circumferential directions (Fig. 2M). The ultimate tensile strengths for the C/S PE group were measured as 18.3 ± 1.6 MPa longitudinally and 8.3 ± 0.7 MPa radially (Fig. 2N). When compared with the PCL group, the C/S PE demonstrated a significant increase in longitudinal tensile strength (p < 0.001), but a significant reduction in circumferential strength (p < 0.001). In addition, as illustrated in Fig. 2O, the percentage elongation for C/S PE was markedly lower in both directions than that of the PCL group (p < 0.001). Specifically, the percentage elongation in the longitudinal direction was 114.5 ± 3.8% for the PCL group and 81.3 ± 2.5% for the C/S PE group. In the circumferential direction, the values were 111.4 ± 3.6% and 85.1 ± 1.1% for the PCL and C/S PE groups, respectively. The Young's modulus for the C/S PE graft exhibited a similar directional trend to that observed in tensile strength (Fig. 2P). PCL grafts exhibited a slightly higher suture retention force compared to C/S PE grafts; however, the difference was not statistically significant (6.39 ± 0.36 N vs. 5.91 ± 0.51 N, p = 0.25) (Fig. S2C).

Hemocompatibility was evaluated using a hemolysis ratio test in accordance with ISO 10993-4:2017 guidelines. The hemolysis ratios of PCL and C/S PE samples were determined to be 4.16 ± 0.93% and 2.79 ± 1.8%, respectively (Fig. 2Q), demonstrating that both materials caused minimal red blood cell lysis and met the criteria for acceptable hemocompatibility (threshold <5% as set by ISO guidelines). We further assessed the degree of thrombogenicity by performing blood clotting index measurements, as presented in Fig. 2R. After 30 min of contact with blood, the C/S PE group displayed a significantly elevated blood clotting index (77.96 ± 6.7%) compared to the PCL group (33.36 ± 3.5%) (p < 0.001). A higher blood clotting index indicates a greater antithrombogenic property [26]. SEM images showed less platelet adhesion and restricted spreading morphology on both PCL and C/S PE surfaces (Fig. 2S). Even though these findings point to decreased platelet activation tendencies, morphological evaluation by itself cannot offer conclusive functional proof.

2.3. Assessment of in vitro biocompatibility of the C/S PE graft

The MTT cytotoxicity assay presented in Fig. 3A demonstrated that, on day 1, none of the experimental groups differed from the control (tissue culture plate). However, by day 3 (p = 0.0012) and day 7 (p < 0.05), the PCL group appeared to support slower endothelial cell proliferation relative to the control. In contrast, C/S PE was able to maintain cell viability at levels similar to the control by day 3. On day 7, C/S PE sustained viability comparable to the control (N.S.) and was significantly greater than that observed in PCL (p < 0.01), whereas the PCL group continued to display reduced viability compared to the control. These results suggest superior biocompatibility of C/S PE compared to PCL during the experimental timeframe. Additionally, fluorescence imaging of the cells on the grafts was conducted over these intervals (blue = nucleus, green = F-actin) (Fig. 3B). Across all groups, cells presented with elongated morphologies and prominent actin filaments, indicative of healthy cellular status. By day 3, cell clusters on C/S PE and Control surfaces showed a tendency toward forming larger and more confluent aggregates, while PCL still featured noticeable cell-free spaces. At day 7, the proportion of F-actin area was markedly higher than in the PCL group (p < 0.05) (Fig. 3C), and nuclear counts were also significantly increased (p < 0.05) (Fig. 3D). Results of the LIVE/DEAD™ Viability assay confirmed the viability of cells in all groups, but viability was substantially greater in C/S PE compared to PCL following 5 days of culture (p < 0.05) (Fig. 3E). Fluorescence micrographs in Fig. 3F reveal a higher number of dead cells (dead = red, live = green) on PCL grafts relative to both the control and C/S PE groups. Fig. 3G provides a schematic overview of the primary outcomes from the in vitro biocompatibility assessments.

Fig. 3.

in vitro biocompatibility evaluation. (A) MTT assay shows enhanced proliferation on C/S PE at day 7 (n = 5). (B) F-actin (green) and Hoechst (blue) staining reveal improved spreading and confluence compared with PCL and control. (C–D) Quantification of F-actin area (n = 3) and nuclei number (n = 3) confirm higher cytoskeletal organization and cell density. (E) Viability assay demonstrates increased survival on C/S PE at day 5. (F) Live/Dead staining shows predominantly viable cells with fewer dead cells (n = 5). (G) Schematic summary of C/S PE promoting endothelial proliferation, biocompatibility, and reduced cytotoxicity. Scale bars: 200 μm. Statistical significance was calculated by two-way ANOVA with Tukey's test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. “N.S” means not significant.

2.4. Assessment of rBMSCs culture in C/S PE lumen via perfusion bioreactor

A custom-made perfusion bioreactor was utilized to culture rBMSCs within the graft lumen. To assess endothelial cell differentiation, a specialized cell culture medium supplemented with VEGF, EGF, and FBS was used. Fig. 4A outlines the overall experimental workflow. The complete experimental procedure is detailed in the methods section. H&E staining images shown in Fig. 4B demonstrate that after 1 week of culture, a thin cellular layer began to form, although it did not cover the entire circumference. By week 2, the cellular layer became more clearly defined due to enhanced cell proliferation, and circumferential coverage of the lumen was achieved. A significant increase in the cell count within the C/S PE graft was observed by week 2 compared to week 1 (p = 0.0013) (Fig. 4C). Approximately 97% of the circumferential lumen area was covered by cells after 2 weeks, whereas coverage at 1 week was limited to 55% (p = 0.03) (Fig. 4D). Masson's trichrome staining in Fig. 4E reveals a greater extent of collagen deposition following 2 weeks of culture. Fig. 4F presents SEM images display the inner lumen wall of the C/S PE graft before and after cell seeding (2 weeks). These images clearly demonstrate the formation of a continuous cellular layer. The graft measured approximately 6 cm in length. To obtain a more comprehensive assessment, we performed sectioning and fluorescence staining along the graft, extending from the proximal region (media inlet) to the distal (media outlet) region (Fig. 4G). According to Fig. 4G, a distinct and substantive cellular layer formed and covered the entire circumference of the lumen from the proximal to distal regions after 2 weeks.

Fig. 4.

Assessment of rBMSCs Culture in C/S PE Lumen via Perfusion Bioreactor. (A) Graphical representation of the whole perfusion-based dynamic cell culturing process. (B) H&E staining revealed complete cellularization after 2 weeks of culturing. (C-D) Number of nucleus (n = 4 sections) and percentages of cell coverage in the circumferential direction (n = 3 sections). (E) Masson's trichrome staining suggests higher collagen deposition after 2 weeks. (F) SEM images of lumen surface, (i) before cell seeding and (ii) after 2 weeks of cell culturing. (G) DAPI immunofluorescence staining of different regions (proximal to distal) to ensure complete cellularization of the whole graft. Statistical significance was calculated by one-way ANOVA with Tukey's test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. “N.S” means not significant.

2.5. Evaluation of differentiation abilities of rBMSCs to endothelial cells under dynamic conditions

We assessed the endothelial differentiation potential of rBMSCs after 1 week and 2 weeks in dynamic culture conditions on the lumen wall of the C/S PE graft (Fig. 5A). Initially, immunofluorescence staining for specific endothelial markers—(i) CD31, (ii) ICAM1, (iii) flk1, and (iv) eNOS—was conducted to identify endothelial-like cells (Fig. 5B). As shown in Fig. 5B, rBMSCs cultured on the grafts following endothelial induction exhibited clear expression of endothelial cell-specific markers. The cells displayed intense positive staining for CD31 (red), ICAM1 (red), Flk1 (green), and eNOS, with nuclei counterstained using DAPI (blue). Quantitative fluorescence analysis indicated that the circumferential coverage of CD31 (Fig. 5C), ICAM1 (Fig. 5D), and flk1 (Fig. 5E) exceeded 95% after 2 weeks when compared to 1 week (p < 0.0001). While Fig. 5F suggests that almost 80% of the circumferentially covered cells expressed eNOS after 2 weeks. A continuous and uniform layer of endothelial marker-positive cells was observed along the luminal surface, which structurally confirms successful endothelialization after 2 weeks.

Fig. 5.

rBMSCs to endothelial differentiation and activation of different pathways. (A) Schematic representation of rBMSCs differentiated into ECs. (B) Immunofluorescent detection of (i) CD31, (ii) ICAM1, (iii) Flk1, and (iv) eNOS (scale bars: 200 μm). Quantitative analysis of circumferential coverage for (C) CD31, (D) ICAM1, (E) Flk1, and (F) eNOS (n = 4 sections). (G) Venn diagram displaying differentially expressed genes (DEGs) between rBMSCs and differentiated endothelial-like cells in C/S PE grafts analyzed via RNA sequencing. (H) Scatter plot visualizing the distribution of upregulated and downregulated DEGs. (I) Gene ontology (GO) analysis indicating enrichment of terms linked to endothelial proliferation, angiogenesis, and blood vessel development. (J–M) Heatmaps presenting clustered DEGs associated with cell differentiation (J), endothelial cell proliferation (K), angiogenesis (L), and blood vessel development (M). (N–O) Gene set enrichment analysis (GSEA) highlighting significant gene enrichment in angiogenesis and vascular remodeling pathways. (P) Bubble plot showing pathway enrichment and signaling activation, including VEGF, MAPK, PI3K-Akt, mTOR, HIF-1, Notch, TGF-β, and JAK-STAT pathways. (Q–R) Heatmaps illustrating the activation of Notch (Q) and VEGF (R) signaling pathway genes, supporting robust pathway engagement. (S) Circular plot showing marked upregulation of major endothelial genes (Vegfa, Nos3, Flt1, Kdr) compared to MSC-specific markers.

To further substantiate these findings, we performed RNA-seq analysis on the C/S PE graft after 2 weeks of dynamic culturing (C/S PE-EC) as well as the C/S PE graft seeded with rBMSCs (C/S PE-MSC) that were maintained for 2 weeks under static conditions without endothelial media. The Venn diagram in Fig. 5G demonstrates that 16,606 genes were shared between the two groups; meanwhile, 1235 genes were exclusive to C/S PE (MSC), and 1424 were specific to C/S PE (EC). In contrast, the scatter plot indicates that 4572 genes exhibited upregulation in C/S PE (EC), 1472 genes were upregulated in C/S PE (MSC), and 12,031 genes did not display significant changes in expression (Fig. 5H). For functional categorization, Gene Ontology (GO) enrichment analysis was conducted to group proteins by biological process (BP), cellular component (CC), and molecular function (MF) (Fig. 5I). This analysis identified important subcategories such as cell differentiation, cell migration, EC proliferation, angiogenesis, extracellular matrix, and growth factor binding. To further enhance understanding, heatmaps were generated to illustrate the expression patterns within these subcategories. The heatmap in Fig. 5J shows hierarchical clustering of differentially expressed genes related to cell differentiation (GO:0030154), with upregulation marked in red and downregulation in blue. As shown in Fig. 5J, gene expression associated with cell differentiation (GO:0030154) was significantly elevated in the C/S PE (EC) group relative to the C/S PE (MSC) group (p < 0.00001). Gene set enrichment analysis (GSEA) demonstrated a marked enrichment of the Cell Differentiation gene set in the experimental group (Fig. S3A). The gene set for EC proliferation (GO:0001935) was significantly expressed in the C/S PE (EC) group (p < 0.001) (Fig. 5K). This was supported by GSEA, which showed a significant enrichment score for the C/S PE (EC) group (Fig. S3B). A substantial proportion of pro-angiogenic genes were highly upregulated in C/S PE (EC) compared to C/S PE (MSC), indicating activation of angiogenic signaling in the C/S PE (EC) group (Fig. 5L). GSEA further validated a highly significant enrichment of the angiogenesis gene set (p < 0.00001), with the enrichment score achieving its maximum among the top-ranked genes (Fig. 5N). The placement of gene set members (black bars) within the ranked list metric illustrates that angiogenesis-related genes are predominantly found among the most upregulated transcripts in C/S PE (EC). The subset of genes related to blood vessel (BV) development (GO:0001568) was also substantially upregulated in the C/S PE (EC) group (Fig. 5M). GSEA analysis provided additional confirmation of a significant enrichment score in the C/S PE (EC) group (Fig. 5O). Collectively, these observations suggest the presence of a strong transcriptional program supporting EC differentiation, EC proliferation, and angiogenesis. Additional heatmaps and GSEA results for other subcategories are available in Fig. S3. The bubble plot in Fig. 5P presents the results of KEGG pathway enrichment analysis. Multiple prominent pathways were identified, including P13K-Akt, MAPK, VEGF, HIF-1, and Notch signaling pathways. These signaling pathways are all recognized for their established roles in cell differentiation and endothelial cell proliferation. To facilitate deeper insight, the expression levels resulting from hierarchical clustering of differentially expressed genes within the Notch and VEGF signaling pathways are presented. The hierarchical clustering and associated heatmap visualization revealed a unique gene expression pattern, with a significant upregulation of multiple Notch pathway (p < 0.00001) and VEGF pathway (p < 0.001) genes in the C/S PE (EC) group (Fig. 5Q and R). Enrichment of PI3K–Akt and VEGF signaling pathways is consistent with the upregulation of Flk1 and eNOS (Fig. 5B) and supports endothelial lineage commitment. Specifically, enrichment of VEGF signaling pathways supports the upregulation of Flk1 (VEGFR2), a key endothelial receptor mediating angiogenic responses. Activation of PI3K–Akt signaling is known to regulate eNOS expression and nitric oxide production under shear stress, aligning with the increased eNOS staining observed. Furthermore, upregulation of genes associated with cell adhesion and endothelial activation correlates with enhanced CD31 and ICAM1 expression, indicating acquisition of endothelial-associated intercellular junction and inflammatory-responsive characteristics. Chord diagram analysis of the RNA-seq dataset illustrated clear transcriptional reprogramming during the endothelial differentiation of rBMSCs (Fig. 5S). Endothelial markers such as Vegfa, Flt1, Flt4, Eng, and Nos3 (eNOS) showed substantial upregulation (logFC moving toward the red zone), whereas mesenchymal stem cell markers, including Lepr, Nes, Nt5e, and Cxcl12, were downregulated (logFC moving toward the blue zone). This observed transcriptional transition from MSC-associated genes (blue arcs) to those related to endothelial lineage (purple arcs) demonstrates a phenotypic shift consistent with definitive endothelial-like lineage commitment.

2.6. Pathophysiological condition of grafts in vivo

The cell-free C/S PE graft and the pre-endothelialized C/S PE (EC) grafts were implanted into the rat abdominal aorta model as illustrated in Fig. 6A. Following a 30-day implantation period, color Doppler analysis was performed to assess graft patency and blood flow velocity (Fig. 6B). Both groups demonstrated regular, broad peaks with uninterrupted forward flow, indicating maintained patency. In the C/S PE group, diastolic flow was minimal but detectable, whereas the C/S PE (EC) group exhibited persistent diastolic flow. As depicted in Fig. 6C, the peak systolic velocity observed in the C/S PE group was significantly higher, reaching approximately 78 cm/s, compared to approximately 66 cm/s in the C/S PE (EC) group. Overall, while both groups retained patency, the C/S PE group displayed higher systolic velocity and reduced diastolic flow, which are indicative of lower compliance and increased vascular resistance. H&E staining shown in Fig. 6D (i) revealed neointima formation in the proximal and distal segments of the C/S PE graft; however, one graft exhibited focal thrombus formation in the mid-region (full cross-sectional images are available in Fig. S4). It is noteworthy that no anticoagulants were used during the investigation period. Conversely, the C/S PE (EC) group demonstrated uniform neointima formation from the proximal to distal regions (Fig. 6D (ii)). Higher-magnification images confirmed continuous endothelial linings throughout the C/S PE (EC) graft with an absence of thrombus. Consequently, data from both color Doppler analysis and histological examination indicate that, of the 3 grafts in the C/S PE group, 2 remained completely patent, and 1 of them remained patent but contained a small thrombus (Fig. 6E). In contrast, all grafts demonstrated full patency in the C/S PE (EC) group. Fig. 6F indicates that 91.3 ± 1.1% of the lumen was open in the C/S PE group compared to 96.8 ± 1.6% in the C/S PE (EC) (p = 0.013). The thickness of the regenerated neo-tissue in the C/S PE (EC) group was significantly greater than that in the C/S PE group throughout the entire region (p < 0.00001) (Fig. 6G). Masson's trichrome staining images for the (i) C/S PE and (ii) C/S PE (EC) groups are presented in Fig. 6H. Both groups demonstrated substantial collagen (blue) deposition. Quantification of collagen coverage from the staining images showed values of 36.99 ± 4.4% in the C/S PE group and 42.4 ± 8.3% in the C/S PE (EC) group (Fig. 6I). Additionally, Von Kossa staining was performed to assess calcification (Fig. S4C). No calcium (black) deposition was detected in either group.

Fig. 6.

Graft Pathophysiological Condition in vivo. (A) Implantation of the graft in the rat abdominal aorta model. (B) Doppler ultrasound assessment to evaluate blood flow within the graft (n = 3 rats). (C) Measurement of peak systolic velocity in the graft at 1 month post-implantation (n = 3 rats). (D) H&E staining performed on the explanted graft. (E) Assessment of graft patency at 1 month after implantation. (F) Quantification of the percentage of open lumen area. (G) Measurement of regenerated tissue thickness. (n = 5 sections) (H) Masson's trichrome staining of explanted grafts, with blue indicating collagen. (I) Quantitative analysis of collagen coverage in the extracted grafts (n = 5 sections). Statistical significance was calculated by two-way ANOVA with Tukey's test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. “N.S” means not significant.

2.7. Evaluation of endothelial layer formation

Immunofluorescence staining for eNOS (endothelial nitric oxide synthase) (Fig. 7A) and CD34 (Fig. 7B) was performed to identify endothelial cells. Functional ECs were marked by eNOS expression (red), whereas CD34 expression (red) was used to identify endothelial and early-stage endothelial cells. Nuclei were stained with DAPI (blue). In the C/S PE group, eNOS⁺ and CD34⁺ cells appeared infrequently and were discontinuous along the luminal interface, with the middle region largely lacking an endothelial lining.

Fig. 7.

Endothelialization and Smooth muscle cell regeneration. Endothelialization was assessed by staining for (A) eNOS (red) and (B) CD34 (red). Quantitative analysis of positive staining area for (C) eNOS and (D) CD34. While smooth muscle cell coverage and maturation were examined using (E) αSMA (green) and (F) MYH11 (green). Quantitative analysis of the positive area fraction for (G) αSMA and (H) MYH11 (n = 5 sections). Nuclei were counterstained with DAPI (blue). Scale bars: 200 μm. Data are presented as mean ± SEM, with individual data points shown. Statistical significance was calculated by two-way ANOVA with Tukey's test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. “N.S” means not significant.

Conversely, the C/S PE (EC) group displayed strong and uniform luminal staining for both markers, indicating enhanced endothelial presence. Quantitative analyses confirmed that the C/S PE (EC) group exhibited significantly greater luminal coverage of eNOS⁺ (Fig. 7C) and CD34⁺ (Fig. 7D) cells in the middle (p < 0.0001) and distal regions (p < 0.0001) compared to the C/S PE group; no significant difference was observed in the proximal region. Overall, the C/S PE (EC) grafts facilitated more rapid and uniform formation of the endothelial layer along the lumen, while the cell-free C/S PE grafts similarly exhibited notable host-derived endothelialization after one month, underscoring the graft's inherent capability to facilitate vascular regeneration. Immunofluorescence staining of the eNOS and CD34 in the native rat abdominal aorta is provided in Fig. S5.

2.8. Vascular smooth muscle layer regeneration

To assess vascular smooth muscle regeneration, immunofluorescence staining was performed targeting αSMA, an established marker of contractile smooth muscle cells (Fig. 7E), and myosin heavy chain 11 (MYH11), which indicates mature and highly differentiated vascular smooth muscle cells (Fig. 7F). Both markers were visualized in green, with nuclei counterstained by DAPI (blue). In the C/S PE group, αSMA⁺ and MYH11⁺ cells were observed along the medial layer; however, the staining was discontinuous and only partially covered the mid-region, in contrast to the C/S PE (EC) group. By comparison, the C/S PE (EC) grafts demonstrated a more continuous and uniform smooth muscle layer, as well as stronger expression of both αSMA and MYH11, indicating superior recruitment and maturation of vascular smooth muscle cells. Quantification corroborated these results, revealing a significantly greater extent of αSMA⁺ (Fig. 7G) and MYH11⁺ (Fig. 7H) cell coverage in the C/S PE (EC) group versus the C/S PE group (p < 0.01). Nonetheless, the substantial presence of αSMA⁺ and MYH11⁺ cells in the C/S PE group after one month suggests that even acellular grafts promote host-derived smooth muscle regeneration, although at a comparatively slower and less organized rate than C/S PE (EC) grafts. Immunofluorescence staining of the αSMA and MYH11 in the native rat abdominal aorta is provided in Fig. S5.

2.9. Macrophage polarization effect of rBMSCs-derived EC seeded C/S PE on vascular regeneration

Macrophages play an essential role in both inflammation and tissue regeneration. Because vascular grafts typically act as foreign bodies, their implantation often elicits an inflammatory response at the graft site. As a result, an initial detection of proinflammatory (M1) markers is commonly observed; nevertheless, the graft's capacity to transition M1 markers to proregenerative M2 markers is fundamental for achieving long-term applicability. Immunofluorescence staining was performed to detect CCR7 (green), a well-established marker of M1 macrophages (Fig. 8A), and CD206 (red), a recognized M2 macrophage marker (Fig. 8B) [27]. As shown in Fig. 8A, there was a greater abundance of CCR7⁺ cells across the proximal, middle, and distal segments of the cell-free C/S PE graft after one month. In contrast, only a limited number of CCR7⁺ cells were present in the pre-endothelialized C/S PE (EC) group, localized mainly to the adventitial layer of the graft wall. Notably, substantial numbers of CD206⁺ cells were observed in both groups, predominantly concentrated in the medial layer of the graft (Fig. 8B), with a consistent spatial distribution from proximal to distal regions. These findings indicate that proinflammatory macrophages infiltrating the graft wall were polarized towards a proregenerative phenotype. Nevertheless, the cell-free C/S PE group exhibited a more sustained inflammatory response than the C/S PE (EC) group. Quantitative assessments were consistent with these outcomes. Fig. 8C demonstrates a significantly greater number of CCR7⁺ cells in the cell-free C/S PE group relative to the C/S PE (EC) group throughout all regions (p < 0.0001). No significant difference in CD206⁺ cell numbers was noted in the proximal region (p = 0.99), whereas notably greater numbers of CD206⁺ cells were found in the middle (p = 0.0002) and distal (p = 0.006) regions of the C/S PE (EC) group compared to the C/S PE group. Fig. 8E demonstrates that M1 markers were distributed throughout the graft wall, while M2 markers were predominantly confined to the medial region of the C/S PE graft wall. In contrast, few M1 markers appeared within the adventitial region of the C/S PE graft wall, and numerous M2 markers were localized to the central and inner areas.

Fig. 8.

Macrophage polarization in implanted vascular grafts. (A) Immunofluorescence staining for CCR7 (M1 macrophage marker, green) with nuclear counterstaining (blue) highlights the infiltration of pro-inflammatory macrophages at the graft location. The insets at higher magnification show concentrations of M1 macrophages. (B) Immunofluorescence staining for CD206 (M2 macrophage marker, red) with nuclear staining (blue) reveals the presence of anti-inflammatory, tissue-regenerative macrophages along the graft wall. Insets emphasize the distribution pattern of M2 macrophages. Quantitative analysis for (C) M1 (CCR7⁺) and (D) M2 (CD206⁺) macrophages is presented as fractions of positive cells (n = 5 sections). (E) A schematic summarizes macrophage polarization in C/S PE grafts versus C/S PE (EC) grafts, illustrating a transition from a pro-inflammatory (M1)-dominant environment to a pro-regenerative (M2)-dominant phenotype in the presence of MSCs. Scale bars: 200 μm (overview) and 50 μm (insets). Statistical significance was calculated by two-way ANOVA with Tukey's test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. “N.S” means not significant.

3. Discussion

In this study, we developed and evaluated a novel small-diameter vascular graft (C/S PE) consisting of a polycaprolactone (PCL) core and a PCL-cECM shell produced via coaxial electrospinning. The design allows for both acellular and pre-endothelialized implantation, offering adaptability for emergency applications. Nonetheless, translation of endothelialized vascular grafts into clinical practice remains difficult due to challenges with establishing uniform EC coverage, sourcing ECs, and complex harvesting methods. To address these limitations, we established a dynamic culture system that efficiently generates EC-like phenotype from rBMSCs within the lumen of C/S PE vascular grafts.

Multiple studies have demonstrated that decellularized ECM derived from cardiac sources (cECM) exhibits a strong capacity to stimulate the proliferation of vascular ECs and also enhances angiogenic processes [28,29]. To the best of our knowledge, we are the first to fabricate an acellular SDVG employing cECM through the conventional electrospinning technique, as detailed in our recently published research [21]. LC/MS analysis in our previous investigation identified trace amounts of cytoplasmic proteins; therefore, for the current study, we optimized the extraction protocol by utilizing equivalent concentrations of sodium dodecyl sulfate and Triton X-100. The revised LC/MS assessment verified the complete removal of cytoplasmic proteins. In pursuit of a highly porous scaffold with thicker core-shell fibers, we selected 20% (w/v) as the fiber core. Given that one major goal of this work is to assess the ex vivo endothelialization potential of the graft, this high porosity was specifically engineered to enhance cellular adhesion and proliferation [4,30].

The coaxial electrospinning process generated a consistent core–shell fiber architecture with hydrophilic surface properties (water contact angle less than 90°), significantly lower than that observed in bare PCL. Recent investigations into ECM-based scaffolds have also demonstrated that incorporating ECM into synthetic materials appreciably improves the surface wettability of the constructs [31,32]. The mechanical analysis of C/S PE indicated enhanced longitudinal tensile strength, which is likely attributable to the predominant alignment of fibers in the longitudinal orientation (Fig. S2D).

Implanted vascular grafts are required to maintain excellent hemocompatibility and demonstrate antithrombogenic properties. The C/S PE demonstrated a hemolysis ratio of less than 5%, indicating strong hemocompatibility [33]. This graft also achieved a markedly higher BCI value compared to PCL, signifying lower in vitro thrombogenicity. This observed antithrombogenicity may be attributable to cECM-derived proteins including heparan sulfate (HSPG2) and plasminogen (PLG), as revealed by LC/MS analysis [34]. The C/S PE graft revealed improved biocompatibility relative to the control group, attributed to the incorporation of cECM. Previous investigations have shown that structural ECM proteins facilitate cell adhesion and proliferation via specific peptide domains [35], such as the Arg-Glu-Asp-Val (REDV) and Arg-Gly-Asp (RGD) sequences present in fibronectin and the Val-Ala-Pro-Gly (VAPG) sequence in elastin.

Our grafts demonstrated greater than 95% endothelial marker (CD31, flk1, and ICAM1) coverage after 2 weeks of incubation using our custom bioreactor. This result aligns with a recent report by Park et al., which achieved similar levels of endothelial marker coverage in decellularized vascular grafts under equivalent culture conditions [7]. Saunders, Sarah K. et al. designed a bioreactor-based perfusion system and attained approximately 72% EC coverage [36]. Batty, Luke et al. demonstrated that their culturing protocol produced greater than 90% coverage of CD31⁺ cells. Our RNA sequencing data indicated robust eNOS (NOS3) production following the culture of rBMSCs (Group MSC 11 TPM, Group C/S PE (EC) = 89 TPM). Enhanced eNOS expression is characteristic of viable, functional ECs [37]. Gene ontology (GO) and KEGG pathway analyses revealed a marked distinction between rBMSCs and the ECs derived from rBMSCs.

The C/S PE (EC) grafts demonstrated a 100% patency rate at 1 month in the rat abdominal aorta model. In comparison, a meta-analysis of small-diameter vascular grafts reported a median patency rate of 83% (n = 873), with a median follow-up duration of 56 days [38]. The recorded peak systolic velocity for the C/S PE (EC) graft was approximately 66 cm s⁻¹. According to Dokuchaeva et al., normative Doppler values in Wistar rats are as follows: central/infrarenal PSV ≈55.9 ± 17.6 cm s⁻¹ at 6 months. Our measurement slightly surpasses their central mean but remains within one standard deviation; it is, however, lower than their proximal aortic mean (≈80.1 ± 23.7 cm s⁻¹) [39]. These results are consistent with velocities reported in current rat graft studies that utilize Doppler monitoring for patency assessment [40]. It is essential to acknowledge that the exact measurement location (proximal vs. infrarenal), anesthetic regimen, and angle correction can impact velocity values, and therefore, should be standardized for accurate comparisons. The neointimal thickness in the C/S PE (EC) graft exceeded 100 μm in all measured orientations. Notably, the increased neo-tissue thickness did not compromise lumen diameter (with over 97% of the lumen remaining open) because the graft wall degraded inward and was remodeled by native tissue. In a recent publication, Zhang, Chunliang et al. reported that their collagen and red blood cell membrane vascular graft generated a neointima approximately 168 μm thick after 1 month of implantation [41]. Nevertheless, histological analysis indicated a reduction in lumen diameter. In another recent investigation, Stahl, Alexander et al. observed that their Polyurethane-based vascular graft led to a reduction in lumen diameter from about 2 mm to 1.3 mm, attributed to more pronounced neointima regeneration after 28 days of implantation [40]. In mouse studies, implantation of a tropoelastin-loaded polyglycerol-sebacate vascular graft (1.5 mm diameter) resulted in a neo-intima approximately 100 μm thick [42]. In that study, graft patency remained comparable to native controls due to rapid graft wall degradation, paralleling findings in our work. Current evidence indicates that our pre-endothelialized C/S PE (EC) graft achieves strong neointimal regeneration capacity while preserving lumen diameter.

Rapid endothelialization in vivo suppresses neointimal hyperplasia and reduces thrombus formation. The endothelial monolayer regulates smooth muscle cell proliferation and releases nitric oxide (NO), supporting an antithrombotic effect [43]. Our in vivo assessment indicates more than 97% endothelial coverage as early as 1 month post-implantation. In a recent publication, Zhou et al. found that a passivated hydrogel coating on small-diameter vascular grafts facilitated uniform and stable vascular regeneration in rabbits, achieving 84.5 ± 14.0% endothelial coverage and significant neovascularization at 3 months [44]. Wang et al. observed that vascular grafts incorporating endothelial cell–derived vesicles reached nearly 90% endothelial coverage after 3 months of implantation in a rabbit model [45]. Rafique et al. reported that DMOG-loaded PCL grafts achieved 90.98 ± 4.21% endothelial coverage following 4 weeks of implantation in a rat model [30]. Collectively, these results underscore that our grafts demonstrated significantly enhanced endothelialization relative to previously reported approaches, achieving near-complete luminal coverage within just one month.

Macrophage polarization has emerged as a critical factor governing biomaterial remodeling and the successful integration of vascular grafts. In this study, immunofluorescence staining revealed a significant decrease in CCR7⁺ (M1) macrophages and a corresponding increase in CD206⁺ (M2) macrophages within the C/S PE (EC) grafts relative to the acellular C/S PE group. Subsequent quantitative analysis verified that rBMSC-derived EC incorporation promoted a shift in macrophage phenotype toward an anti-inflammatory and pro-regenerative state. While macrophages are recognized for their role in endothelialization, we did not observe a robust direct association with endothelialization [46]. The majority of these polarized M2 macrophages were localized to the perivascular region, rather than the intimal area, indicating a limited effect on endothelialization. Xiao, Weiwei et al. recently documented a similar distribution pattern in their study of a recombinant DTβ4-inspired vascular graft [47]. They further noted that M1/M2 macrophage infiltration was predominantly confined to the adventitial region of the grafts, suggesting that these cells primarily regulate perivascular progenitor cells.

Certain limitations should be considered when interpreting the findings of this study. Despite the fact that the functional enzyme eNOS and endothelial-associated markers (CD31, Flk1, ICAM1) were elevated after dynamic perfusion training, direct functional tests like tube formation, acetylated LDL uptake, or nitric oxide release were not carried out. Consequently, rather than displaying fully developed endothelial activity, the differentiated cells are better defined as displaying an endothelial-like phenotype. The small sample size (n = 3 per group) used in the in vivo trials may have restricted statistical power and increased sensitivity to inter-animal variability. This study lacks cell-tracking and lineage-tracing methods. As a result, it is impossible to distinguish between host-derived cells and seeded rBMSC-derived EC-like cells as the source of the regenerated endothelium and smooth muscle layers. The specific regulatory factors and macrophage subpopulations involved were not investigated in this study and warrant further exploration. It should be noted that the transcriptome analysis in this study was conducted using a limited number of samples and, therefore, primarily provides exploratory insights into potential molecular responses.

4. Conclusions

In this study, we developed and validated a novel small-diameter vascular graft (C/S PE) by coaxially electrospinning a polycaprolactone (PCL) core with a shell enriched in cardiac extracellular matrix (cECM). Compared to PCL alone, the graft exhibited excellent physicochemical properties, demonstrated enhanced hemocompatibility, and achieved superior biocompatibility. Rat bone marrow-derived MSCs were efficiently induced to differentiate into endothelial cells within a custom-designed bioreactor, resulting in over 95% luminal coverage after two weeks of dynamic culture. RNA-seq analysis confirmed strong activation of pathways associated with endothelial differentiation and angiogenesis. Both acellular (C/S PE) and pre-endothelialized (C/S PE (EC)) grafts remained patent for one month following in vivo implantation in a rat abdominal aorta model; notably, the pre-endothelialized grafts exhibited enhanced hemodynamic function, more uniform endothelialization, improved regeneration of smooth muscle, and promoted macrophage polarization toward a pro-regenerative M2 phenotype. Collectively, these findings support the potential of the C/S PE graft as a versatile platform for ex vivo endothelialization in planned surgeries, such as Fontan procedures and hemodialysis access. This strategy addresses several persistent challenges in small-diameter vascular graft engineering by integrating the bioactivity of cECM with the mechanical stability of synthetic polymers and leveraging stem cell-based endothelialization. Future investigations will focus on extending implantation duration, scaling the approach, and evaluating its efficacy in large-animal models to advance clinical translation.

5. Methods

5.1. Cardiac tissue decellularization

Fresh porcine hearts (locally sourced in Cheonan, South Korea) were cut into cube-shaped pieces (approximately 1 cm³). The tissue pieces were rinsed with distilled water (DW) to eliminate debris and residual blood. Decellularization was initiated by treating the samples with 1% SDS for 24 h, with the solution replaced every 12 h. Subsequently, 1% Triton X was applied for an additional 24 h. Once all pieces became translucent, they were washed with PBS for 48 h, refreshing the solution every 12 h. The resultant translucent decellularized cardiac tissues (cECM) were then freeze-dried and stored at −20 °C until needed.

5.2. Characterization of the cECM

To assess the thorough removal of nuclear material and ECM proteins, we conducted H&E, Masson's trichrome, and safranin O histological staining according to standard protocols. The Exgene™ DNA extraction kit was employed to quantify the total DNA content in both natural cardiac tissue and cardiac extracellular matrix (cECM), as per the manufacturer's instructions. LC-MS/MS analysis of cECM and native tissues (n = 3) was carried out as previously described in our published method [21]. Gene ontology analysis was performed using Shiney GO 0.82 (https://bioinformatics.sdstate.edu/go/) [48]. Gene clustering and heatmap visualization were generated using TBtools-II [49].

5.3. Core/shell PCL-cECM (C/S PE) vascular graft fabrication

The C/S PE vascular graft was fabricated by means of the coaxial electrospinning technique. For the core, a solution of 20% PCL (70000-90000 Mn, Sigma-Aldrich) in hexafluoro-2-propanol (HFIP) was prepared. The shell solution comprised a blend of PCL-cECM. Initially, 10% cECM and 12% PCL were dissolved in HFIP separately. These two solutions were subsequently combined at a ratio of 10:90 (cECM: PCL). Two 12 mL (lure-lock type) syringes were connected to a coaxial nozzle (inner: 21G, outer: 16G). The following electrospinning parameters were used: nozzle to collector distance: 14 cm; collector rotation speed: 500 rpm; core injection rate: 2 mL/h; shell injection rate: 2.5 mL/h; voltage: 24 kV.

5.4. Core/shell fiber morphology

For cross-sectional analysis, the fabricated C/S PE graft was snap-frozen in liquid nitrogen and sectioned before examination by a scanning electron microscope (SEM) (JEOL, JSM-6701F, Tokyo, Japan). Prior to SEM imaging, the samples were coated with platinum using a sputter coater (Cressington, 108 Auto Sputter Coater). From a variety of images captured, we randomly selected 100 surface fibers for diameter measurement with ImageJ software. Anisotropy of the fiber alignment was evaluated using the OrientationJ plugin. To visualize the core/shell architecture, rhodamine B (Sigma Aldrich, R6626-25G) was incorporated into the shell solution during fabrication, and the graft's surface was assessed under a fluorescence microscope (Nikon, ECLIPSE, Ti2). Pore size (n = 50 pores) was calculated from the SEM images using ImageJ software.

For the transmission electron microscopy (TEM) analysis, Samples were fixed in 2.5% glutaraldehyde for primary fixation, followed by secondary fixation with osmium tetroxide (OsO₄) conducted under a fume hood. Various ratios of acetone and epoxy resin mixture were used to incubate fixed samples. Then, samples were incubated at 60 °C and 100 °C for an entire night after being dipped in the 100% resin. An ultra-microtome was used to segment the resin samples, and section slides were then adhered to the grids. Uranyl solution was used to stain the samples, which were then securely fastened to the TEM holder for examination.

5.5. Water contact angle

The hydrophilicity of the graft surfaces was assessed using a water drop shape analyzer (EasyDrop, KRÜSS). Specifically, a 3 μL water droplet was applied onto the surface of each graft. A CCD camera captured the process, and the contact angle was analyzed concurrently using the drop shape analyzer software.

5.6. In vitro degradation assay

To evaluate in vitro degradation, triplicate samples (1 × 1 cm²) from each vascular graft category were first weighed and positioned in individual Petri dishes containing 2 mL of simulated body fluid buffer (SBF, pH 7.4). The specimens were incubated at 37 °C on a reciprocal shaker (SH30L), with the buffer solution replaced every 24 h. At specific intervals (Days 1, 3, 7, 14, 21, and 28), the samples were removed, rinsed thoroughly, and lyophilized. Finally, the remaining mass was measured to determine weight loss.

5.7. Tensile test

We evaluated our grafts in both longitudinal and circumferential orientations using a universal tensile testing machine (RB 302, MICROLOAD, South Korea). Briefly, each sample was sectioned into pieces measuring 30 × 8 mm². Both ends of the sample were secured in the machine clamps. Stress-strain curves were generated by applying up to a maximum load of 30 N at a crosshead speed of 0.5 mm/s. For suture retention strength assessment, 5-0 Prolene® filaments were sutured onto both graft edges and anchored to the clamps as open loops. The suture retention test was conducted under identical loading parameters as the tensile test.

5.8. Hemolysis test

Whole blood was harvested from Sprague–Dawley rats into tubes preloaded with acid citrate dextrose (ACD) as an anticoagulant. Samples underwent centrifugation at 3000 rpm for 10 min to separate plasma, after which residual blood cells were washed two times with PBS. Purified erythrocytes were prepared by an additional two rounds of centrifugation. For the hemolysis assay, 0.2 mL of a 2% erythrocyte suspension in PBS was mixed with 10 mL of PBS extract from each test sample, followed by centrifugation at 3000 rpm for 10 min. Triton X-100 and PBS served as the positive and negative controls, respectively. Absorbance of the supernatant was determined at 540 nm using a UV microplate reader, and hemolysis ratio (HR) was calculated as follows:

$$HemolysisRatio=\frac{Absorbanceofthesample-AbsorbanceoftheNegativecontrol}{\text{Absorbance}\text{of}\text{the}\text{Positive}\text{Control}-AbsorbanceoftheNegativecontrol}$$

5.9. Blood clotting index

Precisely weighed graft specimens (5 mg per sample) were placed in centrifuge tubes, to which 100 μL of freshly collected rat blood pre-mixed with 10% 0.1 M calcium chloride was added. After 30 min of incubation, 3 mL of distilled water was added to each tube. Next, 200 μL of the resultant supernatant was transferred into a 96-well plate and absorbance was recorded at 540 nm by a UV microplate reader. The blood clotting index (BCI, %) was determined using the equation below:

$$\%BCI=\frac{Absorbanceofthesample}{\text{Absorbance}\text{of}\text{the}\text{control}}\times 100$$

5.10. Platelet activation test

Platelet-rich plasma (PRP) was generated by centrifuging anticoagulated whole blood at 2000 rpm for 10 min before collecting the upper PRP fraction. Sterilized grafts were incubated with 200 μL PRP at 37 °C for 1 h, washed with PBS, fixed in 3% glutaraldehyde, dehydrated through graded ethanol concentrations (50–100%), and air-dried before SEM evaluation.

5.11. MTT assay

Cow pulmonary artery endothelial (CPAE, CCL-209, ATCC) endothelial cells were used for initial cytocompatibility screening to evaluate endothelial adhesion and material safety using a standardized mature endothelial model. Samples underwent pre-incubation in culture medium for 30 min, then CPAE cells (1 × 10⁴ cells/mL) were seeded onto the surfaces. After 1 h of cell attachment, cultures were refreshed with new medium and maintained at 37 °C, 5% CO₂ for periods of 1, 3, and 7 days. At each designated time, 100 μL of MTT solution (5 mg/mL) was introduced and incubated for 4 h. Subsequently, the medium was aspirated and replaced by 400 μL of DMSO to dissolve the formazan product. Absorbance was measured at 595 nm via a UV spectrophotometer.

5.12. Observation of cell morphology and proliferation

Briefly, CPAE cells (1 × 10⁴ cells/mL) were seeded onto the surfaces of the samples and maintained at 37 °C with 5% CO₂ for 1, 3, or 7 days. At each designated time point, samples were washed with PBS, then fixed with 4% paraformaldehyde, and permeabilized using 0.5% Triton X-100. Following blocking with 2.5% BSA for 1 h, actin filaments (F-actin) were labeled with FITC–phalloidin (25 μg/mL) overnight at 4 °C, and nuclei were counterstained using Hoechst 33342 (1 μg/mL) for 10 min. After further washing and mounting, the samples were visualized using a fluorescence microscope (Nikon, ECLIPSE, Ti2). The cell count and F-actin positive area percentage were assessed using ImageJ software.

5.13. LIVE/DEAD™ viability assay

Cell viability was evaluated using a Live/Dead viability kit (Invitrogen, USA). CPAE cells (1 × 10⁵) were seeded onto the samples and cultured for either 1 or 5 days. At each designated time point, 200 μL of assay solution containing calcein-AM (live, green) and ethidium homodimer-1 (dead, red) was applied, and samples were incubated at room temperature for 45 min. Samples were subsequently rinsed with PBS and imaged by fluorescence microscopy (Nikon, ECLIPSE, Ti2). Live and dead cells were quantified using ImageJ software.

5.14. Custom bioreactor design & 3D cell culture

The bioreactor chamber was designed using Autodesk Fusion software. The fabricated vessel chamber is connected to a peristaltic pump and a media reservoir bottle. The path through the chamber for the silicon tube was designed with a 2 mm inner diameter (I.D.). The schematic design is provided in the supplementary document.

Fabricated vascular grafts were cannulated with male/female fittings and sterilized in an ethylene oxide (EO) gas chamber. Sterilized grafts were preconditioned by soaking in culture media containing VEGF (0.1%), EGF (0.1%), and FBS (2%) for 30 min prior to recellularization. rBMSCs were employed in perfusion culture experiments to investigate stem-cell-mediated endothelial differentiation and regenerative potential under physiologically relevant chemical and mechanical stimulation. Expanded rBMSCs were harvested using 0.05% trypsin/0.02% EDTA and centrifuged to form pellets. rBMSCs used in this study were isolated following our previously reported protocol [50,51]. Detailed protocol is provided in the supplementary file. The cell pellets were resuspended to yield 1 × 10⁵ cells/mL. 200 μL of rBMSCs suspension was slowly injected using a syringe, followed by incubation in a CO² incubator at 37 °C for 1 h. After the initial incubation, the grafts were carefully rotated to change orientation (bottom-to-top), and a second cell suspension injection (200 μL) was performed. The grafts underwent a further 1 h of incubation at 37 °C in culture medium. Cellularized vessel grafts were then placed into the custom-designed vessel bioreactor system. The chamber and media reservoir were filled with culture medium, and the peristaltic pump was initially set to deliver media at a flow rate of 100 μL/min, which was gradually increased to 1 mL/min. The perfusion flow rate was gradually increased from 100 μL/min to 1 mL/min to mimic the stepwise mechanical conditioning strategy commonly used in vascular tissue engineering. A low initial flow rate enabled stable cell attachment and prevented shear-induced detachment, while progressive elevation of flow introduced laminar shear stress to stimulate endothelial differentiation. The estimated wall shear stress range during conditioning was approximately 2 dyn/cm². Media were replaced every 3 days, and the cultivation was performed for either 1 or 2 weeks.

5.15. Histological analysis after cellularization

To assess cellularization, grafts were cryo-sectioned using a cryostat (CM1860, Leica). The grafts were divided into three regions (proximal, middle, and distal). The proximal region is 2 cm from the inlet, the middle region is the central 2 cm, and the distal region is 2 cm from the outlet. H&E and Masson's trichrome staining were conducted on cross-sections of the grafts using established protocols. The luminal surfaces of the grafts were coated with platinum and imaged using SEM. Samples were also stained with DAPI (1 μg/mL) for 5 min and examined under a fluorescence microscope (Nikon, ECLIPSE, Ti2). The total cell number was quantified using the cell counter plugin in ImageJ software. The cell coverage percentage was determined via ImageJ software analysis.

5.16. Immunofluorescence staining of the endothelial marker positive cells

To evaluate endothelial differentiation of rBMSCs, immunofluorescence staining was performed for CD31, ICAM1, and flk1. In brief, cryo-sectioned samples were washed with PBS and permeabilized using 0.5% Triton X-100 for 30 min. After blocking with 2.5% BSA for 1 h, primary antibodies to CD31 (1:50, ab28364), ICAM1 (1:100, ab33894), and Flk1 (1:50, sc-393163) were applied overnight at 4 °C. Subsequently, corresponding secondary antibodies were applied for 50 min. Nuclei were counterstained with Hoechst 33342 (1 μg/mL) for 10 min. Following further rinsing and mounting, the samples were visualized with a fluorescence microscope (Nikon, ECLIPSE, Ti2). Marker coverage was quantified using ImageJ software. Five sections from each of the three animals were considered for imaging and quantification.

5.17. RNA sequencing and bioinformatic analysis

Total RNA was isolated from the samples with a lysis buffer. Poly(A)-tailed mRNA was selectively captured and purified using oligo(dT) magnetic beads. Subsequently, the isolated RNA was fragmented and first-strand cDNA was synthesized using random N6 primer–driven reverse transcription. The complementary strand was then produced, with dUTP incorporated in place of dTTP. Following adenylation at the 3′ ends, cDNA fragments underwent end repair and adaptor ligation. Incorporation of dUTP enabled selective degradation of this strand with uracil-DNA glycosylase (UDG), leaving the complementary strand for amplification. Amplification was performed using specific primers to construct a cDNA library of high quality. PCR products were denatured and single-stranded DNA was circularized with a splint oligonucleotide and DNA ligase, generating DNA nanoballs for sequencing on the DNBSEQ platform.

The raw sequencing reads first underwent a quality control (QC) assessment to determine their appropriateness for subsequent analyses. Following the removal of low-quality sequences, the high-quality (clean) reads were mapped to the reference genome. An additional QC step was then conducted to assess mapping efficiency and examine the distribution of aligned reads throughout the reference sequence. Filtered expression data were subsequently normalized and log-transformed [log2(read count + 1)]. Differential gene expression analysis was performed by comparing normalized values between experimental and control groups to measure fold-changes. Functional enrichment analyses, including Gene Ontology (GO) annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway mapping, were conducted using the ShinyGO 0.80 tool (http://bioinformatics.sdstate.edu/go/). Heatmap generation and clustering of gene expression profiles were carried out utilizing Tbtools-II software [49].

5.18. In vivo implantation and Doppler analysis

The constructed grafts were surgically implanted into the rat abdominal aorta. This in vivo implantation was carried out with permission from Soonchunhyang University's Animal Ethics Committee (AEC) (SCH25-0089). Prior to surgery, anesthesia was induced in all rats with 3% isoflurane. Oxygen and N2O gas were used to maintain anesthesia. Postoperative administration of anticoagulant drugs was avoided. Tramadol (50 mg/mL, IM) in the dose of 0.1 mL/kg was used two times a day for two days to control postoperative pain. Briefly, a laparotomy was performed to expose the abdominal aorta. Vascular grafts, approximately 2 mm in ID and 1 cm in length, were anastomosed end-to-end using 8–10 interrupted sutures with 9-0 monofilament nylon (ETHILON™, ETHICON, USA). After the procedure, the abdominal incision was closed with 4-0 synthetic monofilament absorbable sutures (SURGIFIT, Ailee Company Limited, South Korea). The grafts remained implanted for 1 month. Prior to removal, graft patency and blood velocity were evaluated by Doppler ultrasonography [52]. After 1 month, the grafts were harvested, and all rats were euthanized with CO_2. The specimens were then fixed in 4% paraformaldehyde for 24 h and stored at −80 °C until further analysis.

5.19. Cryo-sectioning & histological analysis

All harvested grafts were placed in 30% sucrose for 24 h. Subsequently, samples were embedded in molds containing O.C.T. compound (Sakura, USA). Cryo-sectioning was performed to obtain 5 μm slices using a cryostat (CM1860, Leica). Prepared slides were preserved at −20 °C until analysis.

H&E and Masson's trichrome staining were carried out on cryosections of the grafts according to established protocols.

5.20. Immunofluorescence staining of the extracted grafts

To assess endothelialization, smooth muscle layer development, and macrophage polarization, immunofluorescence staining was performed to identify relevant markers. The detailed methodology is presented in section 4.19. For the evaluation of endothelialization, expressions of CD34 and eNOS were analyzed. For the identification of smooth muscle cells, the presence of αSMA and MYH11 was examined. Lastly, to determine the polarization status of macrophages, the expression levels of CCR7 (M1) and CD206 (M2) were evaluated.

5.21. Statistical analysis

Statistical analyses were carried out using GraphPad Prism version 8.0 (San Diego, CA, USA). The Shapiro–Wilk test was applied to assess data normality. For comparisons involving a single factor, one-way ANOVA followed by Tukey's post hoc test was employed, whereas two-way ANOVA with Tukey's multiple comparison test was utilized for analyses involving two factors. A p-value of less than 0.05 was considered statistically significant. All results are presented as mean ± standard error of the mean (S.E.M).

CRediT authorship contribution statement

Md Abdullah Al Fahad: Writing – original draft, Visualization, Validation, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Minji Choi: Writing – review & editing, Methodology, Investigation, Formal analysis, Data curation. Hyun-Yong Lee: Methodology, Investigation, Formal analysis. Prayas Chakma Shanto: Methodology, Data curation. Md Sowaib Ibne Mahbub: Formal analysis. Nusrat Jahan: Formal analysis. Myeongki Park: Formal analysis. Namhun Kim: Formal analysis. Sang Ho Bae: Supervision, Methodology, Investigation. Byong-Taek Lee: Writing – review & editing, Visualization, Supervision, Resources, Project administration, Funding acquisition, Conceptualization.

Ethics approval and consent to participate

Adult male Sprague-Dawley rats were housed in standard cages with free access to equal amounts of food and water, according to the Animal Ethics Committee of Soonchunhyang University, South Korea (Number: SCH23-0058). All the experiments related to animals in vivo complied with the arrival guidelines and regulations.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: