Revisiting the Tension-Free Paradigm: Axially Prestretched Elastomeric Nanofibrillar Grafts Restore Artery-Like Axial Biomechanics in a Preclinical Model

Abstract

Healthy arteries function under substantial axial prestretch (AP), the ratio of in vivo to unloaded ex vivo length. Loss of AP promotes tortuosity and disturbed hemodynamics associated with vascular pathology, yet artery-like AP has not been incorporated in vascular reconstructions, because conventional substitutes are axially stiff and cannot safely sustain physiological AP. Here, we present Axially Prestretched Elastomeric Nanofibrillar Grafts (APENGs), fabricated by ultrahigh-speed (43,000 rpm) rotational collector electrospinning of biomedical-grade polyurethane. APENGs matched native arterial axial stiffness and tolerated elastomeric elongation beyond AP = 1.5 without failure, excessive force generation, or plastic deformation, while maintaining adequate suture retention, permeability, and burst strength. In vitro, APENGs supported endothelial cytocompatibility and low hemolysis. In a bilateral swine carotid interposition model (8 grafts, two-week follow-up), experimental APENGs (nominal AP ≈ 1.5) achieved and retained elevated prestretch and reduced tortuosity relative to contralateral controls (nominal AP ≈ 1.0). Patency was 75% (prestretched) versus 100% (controls). Histology showed rapid mid-graft endothelialization, early smooth muscle ingrowth, and limited neointimal thickening near anastomoses. This study demonstrates the first synthetic graft implanted with controlled, artery-like AP, challenging the tension-free anastomosis paradigm and positioning AP as a vascular graft design lever to restore native artery biomechanics and mechanobiology.

1. Introduction

Synthetic vascular grafts remain essential for reconstructions when endovascular or autologous options are unsuitable [1,2]. Despite decades of clinical use and technological advancements, their performance lags behind that of native vessels, with higher complication, reintervention, and failure rates [3]. Mechanical mismatch is widely recognized as a major contributor to graft dysfunction because it can perturb hemodynamics and alter cell-mediated remodeling responses [4]. This recognition has driven extensive efforts to mimic native vascular mechanics, predominantly by targeting radial compliance [5–8]. While radial compliance plays crucial roles in large elastic arteries [9–13], small- and medium-diameter peripheral vessels also experience substantial axial deformation and motion. In these settings, axial mechanics may be especially important for preventing kinking, buckling, hemodynamic disturbance, and adverse mechanobiological responses in anatomically complex and mechanically dynamic environments [14–19].

A key yet often overlooked arterial mechanical feature is axial prestretch (AP), defined as the ratio of axially loaded in vivo length to unloaded ex vivo length. In healthy young arteries, AP can exceed 1.5, which corresponds to more than 50% elongation from the relaxed length [20–22]. This prestretch is thought to be established during development: elastin-rich fibers are deposited and crosslinked early in life, and as the artery elongates during growth, the stabilized elastin network becomes effectively preloaded, thereby helping define the mature in vivo prestretched state [14,23]. Physiological AP is not merely a geometric descriptor – it is an optimized mechanical state that shapes how the arterial wall bears load and responds to pulsatile flow. AP pre-engages elastin and collagen, normalizes wall stress [24], maintains tautness [22], moderates axial force swings [14], counteracts buckling [25], and supports laminar flow [25]. Age-related AP loss correlates with tortuosity, altered hemodynamics, pathological cellular responses, and atherosclerosis progression [20–22]. Animal studies further indicate that AP reduction induces irreversible remodeling and cannot be restored naturally, and that this reduction in AP is associated with pathological responses even in native arteries [22,26]. However, the causal role of AP remains difficult to study in native arteries because AP cannot be directly prescribed, measured, or longitudinally tracked without invasive surgical exposure and axial release of the vessel. This limitation motivates engineered platforms that can restore AP in vascular reconstructions while also enabling systematic study of its biomechanical and mechanobiological effects.

Despite its physiological importance, AP is absent from graft design and testing standards, which prioritize metrics such as burst strength, suture retention, and radial compliance while ignoring axial stiffness and prestretch application. Clinically, no protocols exist to prescribe or verify target AP, and surgical practice strongly avoids anastomotic tension because conventional synthetic grafts are axially stiff and can transmit high forces to the suture line even under small elongations [27]. As a result, intentional axial prestretch is considered unsafe for existing grafts, and length redundancy is often used to prevent pulling on the reconstruction during patient movement. However, redundancy can create the opposite mechanical state, axial compression, which may contribute to graft kinking, buckling, disturbed flow, thrombosis, and loss of patency, especially in mobile peripheral locations. Engineering challenges further exacerbate this gap: achieving low axial stiffness while preserving burst strength, kink resistance, suture retention, and dilation control is difficult, as ultra-compliant constructs risk instability and overdilation [4,28]. Most synthetic and biological grafts are axially stiffer and less elastically extensible than native arteries, precluding physiological pre-extension [6,29,30]. Consequently, AP’s potential to restore native artery-like biomechanics, hemodynamic stability, and balanced remodeling remains untapped in vascular grafts [20–22].

Here, we introduce Axially Prestretched Elastomeric Nanofibrillar Grafts (APENGs), fabricated using ultrahigh-speed rotational-collector electrospinning to achieve strong circumferential fiber alignment in small-diameter tubular scaffolds. This highly aligned architecture enables anisotropic axial and circumferential mechanics, including native-like low axial stiffness, substantial recoverable prestretch, and preserved circumferential robustness. Benchtop and in vitro assays assess axial mechanics, surgical handling, permeability, burst strength, hemocompatibility, and cytocompatibility. In vivo testing evaluates feasibility in a swine bilateral carotid interposition model by comparing prestretched APENGs implanted under native-like AP with contralateral non-prestretched controls over a two-week follow-up. Primary outcomes include surgical feasibility, patency, and AP retention, while secondary outcomes include in vivo biomechanics, tortuosity, radial pulsatility, endothelialization, and cellular incorporation. This exploratory study positions AP as a practical, bioinspired graft design variable and establishes APENGs as a platform for restoring, maintaining, and studying native-like axial biomechanics in regenerative vascular reconstruction.

2. Methods

2.1. Ultrahigh-Speed Rotational Electrospinning Induces Low Axial Stiffness in APENGs

APENG scaffolds were fabricated by electrospinning using a custom setup (section S.1.) and 15% (w/w) Pellethane^® 5863–82A polyurethane (Lubrizol) in a 40/60 (w/w) DMF/THF solvent system, stirred at 40°C for ≥24 h (full component details in S.1., Supplementary Material). The Doxa Microfluidics device was used under controlled conditions (40 ± 5% RH, 30 ± 2°C), with a 25-gauge blunt-tip needle, 10 kV potential, 2 mL/h feed rate, and 10 cm needle-to-collector distance. Fibers were deposited on a foil-covered 7 mm-diameter mandrel rotating at 43,000 rpm (anisotropic APENGs) or 9,000 rpm (isotropic controls), with translational stage parameters: 0.14 mL/cm deposition, 3 mm/s velocity, 150 mm distance. Post-electrospinning, grafts were vacuum-dried for ≥12 h, released, and ethylene oxide-sterilized (Anderson AN5100/74i).

Samples were Au-sputter-coated (1 min) and imaged via SEM (NanoImages SNE-4500M PLUS) under high vacuum. Wall thickness was measured in FIJI from 200× images (n = 8 grafts, 3 locations each) [31]. Luminal and abluminal surfaces were imaged at 1000×; fiber diameter was quantified manually (30 fibers/sample on abluminal surface). Fiber alignment used FIJI Directionality plugin (local gradient; 181 bins, 0–180° range) on grayscale ROIs; histograms fitted to Lorentzian models in Microsoft Excel (Figure 1A).

Figure 1:

Ultrahigh-speed rotational electrospinning tunes APENG microstructure and enables native-like axial mechanics. A) SEM images show that increasing collector speed from 9,000 to 43,000 rpm drives pronounced circumferential fiber alignment (vertical direction). Insets show the corresponding fiber orientation distributions (polar histograms). B) Axial tension-stretch responses of native swine carotid arteries (red, n = 8) compared with conventional ePTFE grafts (dark gray, n = 3), low-rpm APENG prototypes (light gray, n = 3), and optimized high-rpm APENG prototypes (purple, n = 8). The shaded areas indicate the full range of measured responses. C) Geometrical characteristics of the optimized high rpm APENG device.

2.2. Benchtop Validation of APENG Mechanical Safety and Surgical Handling

All tests followed ISO 7198:2016. Suture retention tests were performed using BioTester 5000 (CellScale; 23 N load cells) at 100 mm/min (45°/90° cuts; n = 3 grafts, 3 replicates). Water permeability and subsequent burst pressure tests were performed using Edwards transducers (10 mmHg/s rate, n = 3).

2.3. Blood-Contacting Surface Properties and Biological Responses In Vitro

Water contact angle was measured on APENGs, ePTFE (Gore-Tex Patch), and glass (for all n = 5, 3 replicates) at 24 ± 2°C using Ossila Contact Angle software. Hemolysis assays (APENGs, Buna rubber, ePTFE; for all n = 5, 3 replicates) used 1 cm² samples stabilized in PBS (24 h), incubated with 2% porcine RBC suspension (500 μL, 37°C, 2 h), and absorbance at 545 nm (BioTek reader). Positive/negative controls: Triton X-100/PBS (Equation S1K). Platelet adhesion (APENGs, latex, Buna rubber, ePTFE; for all n = 5, 3 replicates) used PBS-stabilized samples incubated with leukocyte-rich platelet-rich plasma (500 μL, 37°C, 1 h), rinsed, fixed (2.5% glutaraldehyde), dehydrated (ethanol series), desiccated, Au-coated, and SEM-imaged (1000×). Adherent platelets quantified manually (9 images/sample, platelets/mm²).

Cytocompatibility used neonatal HUVECs (Sigma-Aldrich) in PromoCell medium + 1% penicillin-streptomycin. Cells (passage 5) were seeded (7,000/well, 96-well plates; APENG, ePTFE, and controls; for all n = 5, 3 replicates) and incubated ~16–24 h. Graft squares (4 mm²) were added for ~24 h. Assays: CyQUANT LDH/proliferation, XTT (Invitrogen protocols; BioTek reader). Cytotoxicity threshold: >30% viability reduction (ISO 10993–5). Positive control: lysis buffer; negative: media only (full materials available in S.3.1.). For qualitative analysis, HUVECs (200,000 cells/cm²) were seeded on 20 mm APENG/ePTFE inserts (pre-conditioned in media, 24 h), incubated 24 h, fixed (10% formalin), permeabilized/blocked, and stained (p120-Alexa 488, VE-cadherin-Alexa 647, DAPI). Imaged via Leica Mica.

2.4. Surgical Feasibility and Early In Vivo Mechanics in a Swine Carotid Interposition Model

All procedures were approved by the Institutional Animal Care and Use Committee at the University of Nebraska Medical Center (Protocol Number: 24-076-01-FC). Acute biocompatibility and performance were evaluated in six-month-old domestic swine (63.0 ± 2.8 kg at implantation). One pilot animal underwent a non-standard implantation with higher AP (S.4.1., Supplementary Material) and was excluded from quantitative analyses, and one animal died during postoperative recovery; thus, quantitative results reflect n = 4 animals (8 grafts; 3 females, 1 male). APENGs were ethylene oxide-sterilized and soaked in heparinized PBS immediately prior to implantation. Under general anesthesia and systemic heparinization (100–200 U/kg), bilateral common carotid arteries were exposed and clamped. An interposition segment was excised (retracted gap ~7.6 ± 0.9 cm). Nitroglycerin (0.5 mcg/kg, approximately 1 mL per dose, applied directly to the local field) was initially utilized but deemed ineffective in reversing the effect of arterial spasm. Its use was discontinued in favor of standard mechanical dilation. 7 mm-diameter APENGs (5.8–8.4 cm) were implanted with interrupted 5–0 or 6–0 polypropylene sutures. Each animal received a prestretched experimental graft (AP_N 1.50 ± 0.01) and a contralateral non-prestretched control (AP_N 1.00 ± 0.00). An experimental AP_N of 1.5 was selected because it closely approximates the physiological AP reported for healthy native arteries and therefore represents a clinically relevant design target [19]. For each animal, the implantation side (left or right carotid artery) was randomly assigned. Treatment allocation was then independently randomized such that either the prestretched or non-prestretched (control) graft was implanted first. This randomization scheme balanced implantation order and anatomical side between treatment groups, minimizing potential procedural and site-specific bias.

After flushing and suture tie-down, clamps were removed to restore flow. Animals received clopidogrel (75 mg/day) for two weeks. Patency and thrombus were assessed by duplex ultrasound immediately post-implantation and at two weeks; angiography was performed at two weeks when feasible and used for representative qualitative assessment of flow. Statistical comparison via the Wilcoxon rank-sum test was performed. Percent pulsatility (Equation S2) was quantified from implantation ultrasound videos by tracking systolic and diastolic perimeters over five cycles (this quantification was not possible for two experimental APENGs and one carotid artery as a result of over-dissection (Pig 5 Left) and failure to correctly save experimental APENG ultrasound). AP was quantified as nominal, effective (post–flow restoration), and residual (two-week) values based on graft length measurements and inter-anastomosis distances (Equations S3–S5).

2.5. Two-Week Evaluation of AP, Axial Stiffness, and Cellular Integration

Following necropsy, APENGs with adjacent carotid segments were excised, flushed with heparinized saline, and measured. Grafts (with anastomotic arterial segments) were opened longitudinally to image the luminal surface. Samples from multiple axial locations were processed for histology (three longitudinal and five transverse sections per artery-graft pair; Figure S8). Tissues were fixed in neutral buffered formalin (24–48 h), transferred to 70% ethanol, paraffin-embedded, sectioned (5 μm), and stained with H&E, α-smooth muscle actin (αSMA), and CD31. Slides were scanned at 40×; full-slide images are provided (Figures S9–S15).

Planar biaxial tensile testing was performed on square, stress-free specimens (≈11 × 11 mm) in 37 °C PBS to characterize pre-implant and explant mechanics of APENGs and adjacent carotid tissue. Specimen-specific strain limits were identified by an initial equibiaxial ramp, followed by equibiaxial stretch-controlled cycling to the defined maximum stretch (capped at λ = 1.20) at 1% stretch/s for 11 cycles (10 preconditioning cycles and one recorded cycle). Deformation was tracked using surface markers within the central region to generate membrane tension–stretch responses. The occluded graft was excluded from ex vivo mechanical analyses due to uncertainties in occlusion-associated remodeling.

For histomorphometry, only full-length longitudinal sections spanning the proximal native artery through the distal native artery were analyzed to reduce local sampling bias. Quantification was performed over the entire graft length; however, regions of interest were defined according to the biological marker being evaluated. For αSMA staining, the region of interest comprised the graft wall only, excluding the lumen and surrounding perivascular tissue. For CD31 staining, the region of interest was restricted to the luminal/neointimal interface along the entire graft length. Endothelial coverage was quantified from CD31-positive signal along the luminal/neointimal interface and reported as percent coverage normalized to a reference luminal area (lumen length × 3 μm). Smooth muscle incorporation was quantified as αSMA-positive area following DAB/hematoxylin color deconvolution and normalized to total graft area. Thresholding parameters are provided in S.5.1., Supplementary Material. For immune response characterization, H&E, CD163, and CC3 staining were quantified manually on three high-power fields (HPFs) per section. HPFs were selected from peri-graft regions immediately adjacent to the APENG and located at least 1 cm from both the proximal and distal anastomoses to minimize the influence of the anastomotic healing response.

2.6. Statistical Analysis

All statistical analyses were performed in RStudio (Version 2024.12.0). Data are presented as mean ± standard deviation unless otherwise specified. Normality was assessed using the Shapiro–Wilk test prior to statistical comparisons. For comparisons between two independent groups, Student’s t-tests were used when assumptions of normality were satisfied; otherwise, Wilcoxon rank-sum tests were applied. Paired comparisons were analyzed using paired t-tests or Wilcoxon signed-rank tests as appropriate. For analyses involving repeated measurements or multiple factors, mixed-effects ANOVA models were used to evaluate the effects of treatment group, time point, and their interaction. When significant effects were detected, post hoc pairwise comparisons were performed with Bonferroni correction for multiple testing. For non-normally distributed datasets involving more than two groups, Kruskal–Wallis tests followed by post hoc Wilcoxon rank-sum tests with Bonferroni adjustment were performed. Effect sizes were reported where appropriate, including Cohen’s d for pairwise comparisons and η² for ANOVA models. Statistical significance was defined as p < 0.05.

3. Results

3.1. Ultrahigh-Speed Rotational Electrospinning Induces Low Axial Stiffness in APENGs

Using the custom ultrahigh-speed collector, APENG fabrication at 43,000 rpm produced pronounced circumferential fiber alignment compared with 9,000 rpm prototypes (Figure 1A). The resulting alignment-driven anisotropy decoupled axial and circumferential behavior. Axial stiffness decreased by approximately 4.2-fold compared with thickness-matched isotropic controls (328.5 ± 20.0 vs. 78.3 ± 30.7 N/m). This value was comparable to that of native swine carotid arteries (74.7 ± 11.7 N/m; n = 8 each; paired t-test p = 0.26; Cohen’s d = 0.234). Optimized APENGs tolerated physiological AP at low force (<1 N at AP = 1.50; Figure S2) and maintained tubular shape during handling and axial loading (Figure 1C). Optimized APENGs had fiber diameters of 1.2 ± 0.5 μm and wall thickness of 282.3 ± 14.2 μm (Figure 1B, C). The aligned nanofibrillar wall preserved an interconnected porous architecture while maintaining a thin profile suitable for surgical manipulation.

3.2. Benchtop Validation of APENG Mechanical Safety and Surgical Handling

Given the unprecedentedly low axial stiffness of APENGs, we evaluated whether the optimized prototype could pair low-force, artery-like axial mechanics with the baseline durability and surgical handling required for implantation according to ISO 7198:2016. Suture retention exceeded the widely accepted 2 N minimum for vascular grafts [32] for both 45° and 90° cuts (3.5 ± 0.4 N and 5.4 ± 0.4 N, respectively; n = 3). Water permeability averaged 0.5 ± 0.4 mL∙min⁻¹∙cm⁻² at 120 mmHg (n = 3), indicating low permeability consistent with porous grafts designed for tissue integration and well below the 800 mL∙min⁻¹∙cm⁻² threshold that would warrant pre-clotting [33,34]. Burst pressure testing was performed up to the system maximum of 1,000 mmHg, and no APENGs failed before this limit (n = 3). This substantially exceeds upper physiological arterial pressures (typically < 200 mmHg), indicating a wide safety margin for vascular applications [35].

3.3. Blood-Contacting Surface Properties and Biological Responses In Vitro

Water contact angle measurements show that dry APENGs are comparable to ePTFE (122.9 ± 5.9° vs 134.6 ± 2.8°; p < 0.01; Table S2), and both exhibit markedly higher contact angles than glass (56.3 ± 3.6°; Figure 2A). Because grafts are routinely pre-soaked in heparinized saline before implantation, we also tested wetted samples to better reflect the intraoperative interface encountered by blood and cells. After wetting, APENGs exhibit a substantially lower apparent contact angle (50.6 ± 23.5°).

Figure 2.

Blood-contacting surface properties and in vitro biocompatibility of APENGs compared with commercial ePTFE. A) Static water contact angle measurements with representative images. The hydrophilic positive control (PC, glass) shows a significantly lower contact angle than APENGs and ePTFE. B) Hemolysis (left) shows all grafts are non-hemolytic with no differences between groups (negative control, NC: untreated RBC suspension; PC: Triton X-100–lysed RBCs). Platelet adhesion (right) on APENGs compared to ePTFE and Latex rubber (PC). C) Cell proliferation (top), LDH release (middle), and XTT metabolic activity (bottom) indicate non-cytotoxic performance across all groups with no significant differences. The LDH inset magnifies the absorbance for the graft groups and the NC. For all assays, n = 5 independent samples per group with three technical replicates each; data are average ± standard deviation, and *p < 0.05.

Higher platelet adhesion on APENGs (1,785 ± 674 platelets/mm²; Figure 2B, Figure S3, Table S1) compared to ePTFE (67 ± 30 platelets/mm²) was observed. Hemolysis remained well below the 2% ASTM F756 threshold for both materials (APENGs: 0.18 ± 0.13%; ePTFE: 0.03 ± 0.03%). The pronounced variability in the latex platelet adhesion density should also be noted, and is attributed to the aggregation and clumping of platelets typically observed on highly adherent surfaces [36,37].

Endothelial cytocompatibility is robust across all conditions (Figure 2C and Table S3). Proliferation remains near 100% (CyQUANT), LDH release stays near baseline, and XTT indicates strong metabolic activity. Immunofluorescence further shows that HUVECs form uniform, junction-positive monolayers on APENGs (VE-cadherin and p120; Figure S4), whereas cells on ePTFE are sparser and discontinuous, often appearing as clusters.

3.4. Surgical Feasibility and Early In Vivo Mechanics in a Swine Carotid Interposition Model

During conventional vascular reconstructions, the stiff synthetic graft is anastomosed on one end and cut to a length that matches or slightly surpasses the distance between anastomoses to reduce the possibility of high anastomotic tension [27]. In our interposition model, the native arterial segment between the anastomoses was excised, allowing the cut ends to retract, thereby widening the gap. To remain consistent with standard surgical practice, we defined the reference length (L₁) as the distance between these retracted ends for both prestretched experimental and non-stretched control APENGs. This practical choice results in reduced effective AP delivered to the graft after clamp removal: the tension in the graft draws the dissected arterial ends toward each other, reintroducing AP in the native artery and, correspondingly, reducing tension and reference length (L₂) in the graft (Figure 3A). Therefore, the effective prestretch experienced by grafts (AP_E experimental = 1.32 ± 0.03; AP_E control = 0.98 ± 0.02) in this model is substantially smaller than the nominal prestretch for experimental grafts (AP_N experimental = 1.50 ± 0.01; AP_N control = 1.00 ± 0.00) and the effective prestretch in the native artery (AP_E = 1.49 ± 0.09). Although reduced, the AP_E in the experimental grafts remains substantial and consistent with the average physiological levels reported for peripheral arteries of ≤ 30-year-old individuals (with ≤ 35 BMI), which is still at the upper level of normal physiological values [38]. Despite substantial extension in experimental APENGs, no suture line failure, suture hole bleeding, or macroscopic tearing was observed for the APENG or native artery.

Figure 3:

Surgical feasibility and early in vivo mechanics of APENGs in a swine carotid interposition model. A) Schematic and intraoperative images illustrate AP induction and redistribution due to geometric changes in experimental (left) and control grafts (right). Box-and-whisker plots (center) compare the nominal prestretch (AP_N) and reduced effective prestretch (AP_E) values for experimental APENGs (n = 4) and control APENGs (n = 4) with native carotid arteries (n = 8). B) Left: selected ultrasound-derived diameter waveforms over multiple cardiac cycles immediately after implantation for experimental and control APENGs, alongside native carotid artery. Right: corresponding relative pulsatility, with Wilcoxon rank-sum test indicating control APENGs have significantly reduced pulsatility (native carotid artery n = 7, experimental APENG n = 2, control APENG n = 4), summarized as box-and-whisker plots. For all box-and-whisker plots: boxes indicate the interquartile range, center lines denote the median, whiskers represent the range, and × mark the mean.

Duplex ultrasound (Table S4) unexpectedly demonstrates appreciable radial pulsatility in the prestretched experimental grafts, comparable to that of the native carotid artery (Figure 3B), whereas control APENGs exhibit significantly lower pulsatility (Figure 3, Tables S6–7). This observation is consistent with the in vitro planar biaxial mechanical tests in Figure S5, which show lower circumferential stiffness in the prestretched grafts.

3.5. Two-Week In Vivo Evaluation of Biomechanics and Patency

At two weeks, longitudinal duplex ultrasound and en face ex vivo images of the luminal surface indicate tissue integration and minimal neointimal formation, limited mainly to modest anastomotic thickening. Patency and flow are preserved in three of four experimental grafts and all control grafts (75% and 100% patency rates, respectively). It should be noted that for the single occluded graft, failure analysis shows complete luminal thrombosis with thrombotic extension approximately 3 cm into the adjacent native artery at both the proximal and distal ends, but no evidence of extensive intimal hyperplasia, graft kinking, collapse, or structural failure (Table S8). Immediate post-implantation duplex ultrasound (Table S7; Pig 2 Right/Experimental) shows small non-occlusive thrombi at both anastomoses, with more apparent flow restriction at the proximal anastomosis. Gross examination also shows a larger and more adherent-appearing thrombus near the proximal anastomosis (Table S8). Angiography suggests greater apparent tortuosity in some control APENGs than in the prestretched grafts (Figure 4), consistent with previous reports regarding AP’s stabilizing effects on arterial morphology [16].

Figure 4.

Two-week in vivo performance of control and prestretched APENGs in a swine carotid interposition model. Representative angiography at two weeks (center) shows uninterrupted contrast transit through both implants, confirming patency in control (AP_N = 1.0) and experimental (AP_N = 1.5) APENGs. Dashed lines trace the graft centerlines and highlight greater tortuosity in the control graft, which persists after explantation. Duplex ultrasound from the corresponding anastomotic regions (B-mode and color Doppler, adjacent panels) confirms a patent lumen and maintained flow within each graft. En face images of the whole graft luminal surfaces are shown at the lateral edges. In all images, cranial (distal) is up, and caudal (proximal) is down.

Angiography demonstrated visually apparent tortuosity in all control APENGs. Angiograms were additionally used to quantify tortuosity for each graft. Percent tortuosity (Equation S6) increased from experimental APENGs (n = 3; 1.99% ± 1.71%) to control APENGs (n = 4; 2.19% ± 2.47%). However, no statistically significant difference between groups was observed (p = 0.857, r = −0.17). This lack of significance is likely at least partially attributable to the lack of relevant biomechanical factors (limb flexion), small sample size, and the inherent limitations of two-dimensional angiographic imaging, which may underestimate tortuosity relative to true three-dimensional vessel geometry. Despite the lack of statistical significance, the trend toward reduced tortuosity and buckling in experimental APENGs aligns with previous reports regarding AP’s stabilizing effects on arterial morphology [16]. Reduced tortuosity is thought to lessen secondary flow and reduce pockets of low wall shear stress and flow reversal. By optimizing these hemodynamics, APENGs may reduce the risk of endothelial dysfunction, plaque accumulation, and thrombus formation—factors critically linked to the failure of small-diameter grafts [39–41]. All angiograms are available in Figure S6, while duplex ultrasound videos and ex vivo photographs of the luminal surface are available in Tables S7–9.

3.6. Two-Week Evaluation of AP, Axial Stiffness, and Cellular Integration

To define how AP shapes early in vivo evolution of graft geometry and mechanics, we tracked bypass length, effective prestretch (AP_E), and axial stiffness from implantation through two weeks. By two weeks, both control and experimental bypass segments appear to lengthen in situ (Figure 5, Table S11). Control APENGs increase from 8.1 ± 0.4 cm immediately after implantation to 9.1 ± 1.5 cm at two weeks, and experimental APENGs increase from 6.6 ± 0.5 cm to 7.7 ± 0.6 cm over the same period. Despite the increase in length, experimental grafts show a reduction rather than increase in AP_E from 1.32 ± 0.03 to 1.26 ± 0.16 at two weeks (Figure 5A, Table S11). In contrast, control grafts remain near an unloaded state, with AP_E changing from 0.98 ± 0.02 to 1.02 ± 0.03 at two weeks.

Figure 5.

AP and axial stiffness evolution over two weeks. Top box-and-whisker plots show AP_E initial and two-week values for control and experimental APENGs. AP_N for the native swine carotids removed at the time of implantation are also included for reference (n = 8). Bottom box-and-whisker plots show axial stiffness values obtained from planar equibiaxial tests pre- (left) and post-implantation (right), experimental APENGs n = 3 due to occlusion in one pig. For all box-and-whisker plots, boxes indicate the interquartile range, center lines denote the median, whiskers represent the range, × symbols indicate the mean, and unless specified, n = 4.

Consistent with the concurrent changes in graft length and AP_E, we observed changes in axial stiffness over two weeks across all implants (Figure 5B), with greater within-graft stiffening in the prestretched group (mixed-effects ANOVA, p = 0.024; Table S13). Absolute stiffness does not differ significantly between groups at baseline (p = 0.545) or two weeks (118.9 ± 30.0 N/m, 110.1 ± 29.8 N/m, 115.8 ± 28.6 N/m, for the native carotid artery, experimental, and control APENGs, respectively; p = 0.234; Table S14). Complete descriptive statistics are provided in Table S12. Although between-group differences were not statistically significant at two weeks, the combination of a large effect size (η² = 0.268), limited sample size, and substantial biological variability suggests that a Type II error cannot be excluded, as the prestretched group shows the strongest coupled shifts in lengthening, AP_E change, and stiffening. Importantly, APENGs retain pronounced anisotropy and remain far more axially compliant than conventional ePTFE (2087.4 ± 534.8 N/m, n = 3).

Experimental and control groups showed no difference in inflammatory or remodeling parameters. FBGC density (determined by H&E staining) was nearly identical between groups (experimental: 48 ± 19 cells/mm²; control: 48 ± 11 cells/mm²; p = 1.000). Similarly, lymphocyte density, also determined by H&E staining, did not differ between groups (experimental: 1778 ± 443 cells/mm²; control: 1887 ± 887 cells/mm²; p = 1.000). CD163-positive macrophages (Figure S15) were present in both graft types, with a nonsignificant trend toward lower density in experimental APENGs (experimental: 102 ± 34 cells/mm²; control: 172 ± 47 cells/mm²; p = 0.250). CC3-positive (apoptotic) cells were also observed (experimental: 92 ± 20 cells/mm²; control: 66 ± 43 cells/mm²; p = 0.500) but did not differ significantly (Figure S15 and Table S15).

Despite the relatively long interposition length (7.6 ± 0.9 cm) and short two-week timeframe, both experimental and control groups exhibited substantial endothelial coverage (21.4 ± 17.5% vs. 10.8 ± 6.5%, respectively; p = 0.28) and nuclear density (5.6 ± 2.6% vs. 4.1 ± 2.4%, respectively; p = 0.28).

4. Discussion

Native arteries are mechanically dynamic in both the circumferential and axial directions. Although vascular biomechanics is often focused around radial compliance and pulsatile diameter changes, sustained axial prestretch (AP) represents an equally important homeostatic biomechanical state that helps maintain vessel straightness and geometric stability [22], preserve efficient pulsatile mechanics [19], regulate wall-stresses, and provide mechanobiological cues that guide remodeling [34]. Reduction of AP due to aging or other factors is associated with increased vascular tortuosity, disturbed flow, altered wall stress, and maladaptive vascular responses. Despite these important associations, the causal roles and underlying mechanisms of AP remain understudied in native vessels because AP cannot be directly measured or prescribed without invasive surgical exposure and axial release of the artery, making longitudinal in vivo tracking impractical. This limits the ability to determine whether altered axial mechanics merely accompany vascular pathology or actively contribute to disease progression, remodeling, and treatment failure. Experimental platforms that allow AP to be prescribed, maintained, and tracked may therefore improve both graft design and mechanistic understanding of vascular mechanobiology.

This knowledge gap is especially important in vascular reconstruction, where axial mechanics are clinically consequential but rarely treated as a biomimetic design target. In contrast to healthy native arteries, commercial synthetic grafts completely lack physiological AP and cannot reproduce the low-force axial behavior of the arterial wall. ePTFE, in particular, has two major axial limitations: it cannot elastically accommodate more than minimal axial extension, and even small, imposed elongations generate high axial forces because of its stiffness (Figure 1B). As a result, any anastomotic elongation is currently strictly avoided in vascular bypass surgery; even slight pulling on a stiff conduit can transmit high forces to the anastomoses, increasing the risk of mechanical damage or failure. To avoid this, surgeons often introduce graft-length redundancy to prevent potential elongation, particularly in mobile anatomic regions. However, redundancy creates the opposite mechanical state to AP: axial compression. This compression can promote graft kinking, buckling, disturbed flow, thrombosis, and loss of patency. These issues are especially relevant when reconstructions cross major joints, such as the knee, where limb flexion imposes substantial foreshortening, bending, curvature, and tortuosity on the femoropopliteal segment [39–41].

Several graft technologies have attempted to mitigate these axial and geometric complications, including stretchable ePTFE [42], corrugated grafts [43], or textile architectures [44]. These designs acknowledge that grafts must tolerate compression, bending, and repetitive motion, and some can improve conformability, length adjustability, or kink resistance. However, geometric conformability is not equivalent to physiological AP. For example, stretchable ePTFE can be extended during implantation and may better accommodate subsequent compression or joint motion without buckling, but this behavior does not recreate the sustained, low-force elastomeric axial state of native arteries. Thus, these technologies primarily mitigate the consequences of absent physiological AP rather than restoring native axial biomechanics. This distinction may help explain why prosthetic conduits continue to perform poorly in biomechanically demanding peripheral locations, especially below the knee or across mobile segments, despite decades of geometric and reinforcement-based design improvements [45,46]. Restoring AP-related function requires a graft that is not merely extendable, reinforced, or kink-resistant, but capable of being prescribed, implanted, and maintained under artery-like AP without generating excessive axial force.

Electrospun elastomeric scaffolds provide an attractive foundation for this challenge because their fibrous microstructure can support tunable stiffness and anisotropy, and their interconnected porous architecture can support vascular cell infiltration and tissue integration without excessive permeability [47–54]. Furthermore, mechanical mimicry and cell-supportive microstructure in electrospun grafts can be combined with additional surface chemical optimization methods, providing a versatile platform in which bulk mechanics, microstructure, and blood-contacting surface properties can be tuned in a decoupled manner [55]. For axial biomimicry, however, fiber alignment is essential. Circumferential alignment lowers axial stiffness by allowing the graft to extend through fiber reorientation and network deformation, while simultaneously reinforcing the circumferential direction needed for radial stability and resistance to overdilation. Thus, alignment provides a route to reduce axial stiffness toward native arterial levels without simply using an ultrasoft material that could compromise burst strength, radial support, or dimensional durability.

Achieving the degree of circumferential fiber alignment required to substantially reduce axial stiffness in small-diameter tubular grafts has remained a major limitation. Conventional electrospinning collectors typically operate at relatively low rotational speeds, often only a few hundred rpm [56,57], producing modest fiber alignment and limited anisotropy. This is particularly limiting in small-diameter graft manufacturing as sufficient linear velocities for fiber alignment require rotational speeds in the tens of thousands of rpm, which can introduce vibrational, dynamic, and dimensional stability challenges, especially at clinically relevant graft lengths. The custom ultrahigh-speed rotational electrospinning system developed here addresses this manufacturing barrier by enabling stable fabrication of highly aligned, anisotropic small-diameter tubular constructs. To our knowledge, this is the first demonstration of a >40,000 rpm rotational electrospinning approach specifically developed for manufacturing small-diameter vascular grafts with native artery-like axial biomechanics.

A key mechanical design problem in biomimetic vascular graft engineering is achieving artery-like mechanics without sacrificing surgical robustness [58,59]. Across prior efforts, compliant constructs have often involved tradeoffs between mechanical compliance and mechanical strength, with some platforms, including tissue-engineered graft approaches, exhibiting dilation or mechanical failure under arterial loading [60–62]. In contrast, APENG combines native-like low axial stiffness with implantation-relevant safety benchmarks, including suture retention, low permeability, burst resistance, and no overdilation. APENG’s ability to elastically accommodate 50% elongation at low force (Figure S2A) while maintaining structural integrity distinguishes it from stiff commercial grafts such as ePTFE, in which even insignificant elongation can generate high axial forces incompatible with vascular reconstruction. APENGs also exhibit native-like nonlinear axial stiffening (Figure 1B), which may provide an intrinsic overstretch-protection mechanism by permitting low-force physiological deformation while increasing resistance at larger extensions. APENG’s native-like mechanical mimicry may also help explain the higher pulsatility observed in prestretched grafts, as Figure S5 shows that axial prestretch reduces effective circumferential stiffness by nearly 3-fold, consistent with the axial-circumferential coupling reported in native arteries [63].

Beyond mechanical performance, the graft-biological interface remains an essential complementary consideration, because a mechanically optimized graft can still fail if early blood contact promotes thrombosis or if the luminal surface does not support endothelial coverage. APENGs exhibited hydrophobicity in static water contact-angle measurements and increased platelet adhesion relative to ePTFE (Figure 2A, B, Figure S3), highlighting a familiar tension in biomimetic graft design: micro- and nano-scale features that support cell engagement and tissue integration may also increase early platelet interactions [36,37,64]. Rather than undermining the APENG concept, this finding indicates that the electrospun surface is highly biointeractive, which may support endothelial attachment and tissue integration while also motivating adjunct biointerface optimization. Importantly, the significantly reduced water contact angle after wetting suggests that clinically relevant graft pre-soaking may substantially improve APENG surface wettability, providing a more favorable blood-biomaterial interface than dry-state measurements alone would indicate. Together with hemolysis and cytotoxicity values that remained comfortably within ISO- and ASTM-defined thresholds (Figure 2B, C), these findings support an appropriate baseline biological interface for APENG without additional modification. Nevertheless, the elevated platelet adhesion observed in vitro may indicate a potential risk of early thrombus formation prior to complete endothelialization and therefore warrants further investigation in long-term in vivo studies.

Several established surface-engineering strategies could be applied to further improve APENG hemocompatibility while preserving its biomimetic mechanical properties. We have previously demonstrated that covalent heparinization of this material significantly reduces platelet adhesion without altering graft microstructure or compliance [55]. Additional approaches include hydrophilic surface modification, endothelialization-promoting coatings, and nitric oxide-generating or nitric oxide-releasing surface chemistries that suppress platelet activation while supporting endothelial cell function. Such anticoagulant and pro-endothelial surface modifications could further reduce thrombotic risk during the critical early post-implantation period before formation of a mature endothelial lining. Therefore, although the present results establish APENG as a mechanically enabling and biologically supportive platform for axially biomimetic vascular reconstruction, further optimization of the blood-contacting surface represents a promising direction for enhancing long-term hemocompatibility and graft patency.

The in vivo portion of this study provides, to our knowledge, the first preclinical feasibility demonstration of a synthetic vascular graft implanted under controlled, artery-like AP. This required revisiting the conventional “tension-free” bypass paradigm, which treats axial tension as a surgical risk to be avoided rather than a physiological mechanical state to be deliberately restored [65,66]. This paradigm is appropriate for stiff grafts such as ePTFE, but APENGs fundamentally change the relationship between elongation and anastomotic force by allowing large prescribed deformation at low axial load. Despite the 50% prescribed elongation at implantation, no suture-line failure, suture-hole bleeding, macroscopic tearing, or overt native artery injury was observed (Figure 3A, Table S9). Patency at two weeks was within the ranges reported for electrospun grafts in interposition models [59,67]. The single occlusion is most consistent with an early anastomosis-associated thrombotic event, as immediate post-implantation duplex ultrasound showed small non-occlusive thrombi at both anastomoses (Table S7; Pig 2 Right/Experimental). The proximal anastomotic thrombus was more clearly visualized and was associated with apparent flow restriction, and gross examination of the explanted graft shows a larger and more adherent-appearing thrombus near the proximal anastomosis (Table S9; Pig 2 Right/Experimental). Together, these findings suggest that the proximal anastomosis may have served as the initial or dominant site of thrombus formation; however, because thrombus was present at both anastomoses and extended into both adjacent native arteries, a single definitive initiation site cannot be assigned. This interpretation is consistent with prior reports that early graft occlusion is commonly thrombotic and can be influenced by anastomotic geometry, inflow restriction, local flow disturbance, and technical/anastomotic factors [68,69]. Nevertheless, because the occlusion occurred in the prestretched group, a contribution of axial prestretch to local anastomotic geometry, hemodynamics, or suture-line stress cannot be excluded. However, pilot observations suggest that prestretch alone is not sufficient to cause graft failure. In a smaller pilot animal that unintentionally received higher prestretch (S.4.1., Supplementary Material), APENG maintained short-term patency at AP_E = 1.57. At more extreme applied prestretch, where occlusion occurred, the graft nevertheless remained structurally intact at AP_E values up to 2.5, suggesting that loss of patency was not driven by gross mechanical rupture or anastomotic failure. Together, these findings demonstrate that substantial native artery-like AP can be surgically applied to a synthetic graft when the graft is engineered to mimic the axial mechanics of a healthy young native artery.

Beyond establishing surgical feasibility, the in vivo study provided early evidence that incorporating AP can improve graft biomechanics. Video capture, angiogram, and ultrasound imaging revealed favorable early biomechanical effects due to AP. Prestretched APENGs exhibited reduced tortuosity relative to non-stretched controls immediately after flow restoration (Figure 3A) and maintained this straighter configuration through explant (Figure 4, Figure S5), consistent with the stabilizing effects of AP on arterial morphology [16]. Although modest in this short-term carotid model, reduced buckling may be functionally meaningful because tortuous geometries can promote secondary flows, low wall shear stress, and flow reversal [25]. The stabilizing role of AP may be even more important in peripheral reconstructions exposed to substantial axial compression, bending, and surrounding tissue constraints [39–41]. As such, tortuosity will be further explored via 3D imaging in future studies. Prestretched APENGs also unexpectedly demonstrated appreciable radial pulsatility during duplex ultrasound imaging (Figure 3B, Table S4), whereas control APENGs exhibited low pulsatility. Given the high circumferential fiber alignment and corresponding circumferential stiffness, APENGs were expected to show no detectable diameter change during the cardiac cycle. The artery-like pulsatility in prestretched APENGs suggests that applied AP may reduce the effective circumferential stiffness of the wall, potentially through changes in fiber orientation, wall geometry, or load sharing between axial and circumferential directions. This axial-circumferential coupling appears to mirror an important feature of native arteries, in which AP and circumferential deformation are mechanically linked rather than isolated. Thus, incorporating AP into APENGs may not only stabilize graft geometry but also help replicate more native-like multiaxial wall mechanics under pulsatile loading. More broadly, APENGs may provide a controllable platform for investigating axial-circumferential coupling in vascular function, a phenomenon that is difficult to isolate and parametrically study in native arteries.

Another notable biomechanical observation was the redistribution of graft length and AP after clamp removal, followed by continued evolution of these parameters over two weeks. After reperfusion, axial load was redistributed between the prestretched APENG and the proximal and distal native artery segments, which had lost AP following dissection and interposition gap creation. As a result, effective AP values (AP_E) were significantly lower than nominal AP values (AP_N – Figure 3A), reflecting an important model-specific behavior of the interposition graft configuration and an opportunity to better understand in vivo axial mechanics. Importantly, this redistribution may also be leveraged as a useful feature to control AP and the axial state of native arteries in interposition-based models, whereas this behavior would be less likely in a bypass model in which the native artery segment remains intact. By two weeks, prestretched APENGs showed increased length (Table S10) but reduced AP_E at explant (Figure 5, Table S10), likely reflecting a combination of scaffold viscoelastic relaxation (Figure S2B), animal growth, positional and pressure differences during measurement, evolving peri-graft constraints, and remodeling at the graft-artery interface. In contrast, control grafts remained near their initial unloaded axial state. The coupled mechanical changes in prestretched APENGs suggest that AP may function not only as an imposed implantation state, but also as a cue influencing early graft and arterial remodeling. More broadly, this dynamic response highlights an opportunity to study the role of AP in vascular remodeling through longer-term experiments with serial in vivo assessments. Specifically, for the initial AP prescription in interposition grafts, the length of the unretracted artery removed to create the graft bed should be used as the reference rather than the intraoperative gap length alone, because the latter does not account for immediate post-implantation arterial retraction and AP redistribution. Radiopaque markers and serial imaging under standardized positioning could further help quantify immediate and longer-term post-implantation axial change.

Notably, APENGs elicited an encouraging early in vivo biological response despite relying on nanofibrillar architecture and axial mechanical design rather than biochemical surface functionalization. At two weeks, longitudinal duplex ultrasound and en face ex vivo images showed no evidence of vessel injury or anastomotic hyperplasia (Figure 4, Figure 6A, Tables S7 and S8), which are commonly associated with graft-artery mechanical mismatch [70]. APENGs supported early mid-graft endothelialization despite an average graft length exceeding 8 cm and only two weeks of implantation, an important finding because endothelialization and tissue ingrowth remain major challenges in commercial synthetic grafts and typically require optimized pore microstructure in electrospun experimental grafts [51,52]. Importantly, endothelialization was not observed as a fully confluent luminal layer at this early time point. Rather, CD31-positive regions appeared as discontinuous patches distributed throughout the graft length, including mid-graft regions distant from the anastomoses, suggesting endothelial cell presence beyond simple trans-anastomotic ingrowth. Although group differences did not reach statistical significance (Figure 6B, Table S15), prestretched grafts showed numerically higher endothelial coverage, approximately twofold greater than unstretched controls, along with modestly higher nuclear density (Figure 6B). These trends should be interpreted cautiously, but they indicate that physiological AP did not compromise early healing and warrant further investigation of AP as a potential mechanical catalyst for cellular infiltration and endothelialization, complementing biochemical surface-functionalization strategies previously used to accelerate vascular graft regeneration [71]. SEM further supports these histology findings, showing persistent adherent tissue consistent with scaffold ingrowth and intact, continuous nanofibers without visible fracture or deformation in exposed regions (Figure S7). Overall, these results support the biological potential of the highly oriented nanofibrillar architecture and suggest that axial mechanical design can be incorporated without impairing early functional remodeling.

Figure 6.

Gross and quantitative histological evaluation of APENGs. A) Representative full-length longitudinal histological images of experimental APENG stained with H&E, CD31, and α-SMA stains, indicating minimal intimal thickening, smooth muscle cell ingrowth, and endothelialization. B) Quantitative analysis of H&E (foreign body giant cell (FBGC) and lymphocyte density), as well as CD31 (endothelial cell), and α-SMA (smooth muscle cell) stains are summarized as box-and-whisker plots, with boxes indicating the interquartile range, center lines denoting the median, whiskers representing the range, and × symbols indicating the mean. Wilcoxon signed-rank paired tests indicate there is no significant difference between experimental and control APENGs for any stain. In all tests, *p < 0.05 and n = 4 independent samples, except experimental APENGs, where n = 3.

Inflammatory analysis indicates that experimental APENGs elicited a tissue response comparable to that of control APENGs during the early two-week implantation period. Neither inflammatory cell infiltration nor indicators of tissue remodeling differed significantly between groups, suggesting that the experimental modification did not adversely alter the host response. The similar FBGC densities observed in both groups indicate that the modification neither exacerbated nor attenuated the early foreign body reaction. Likewise, comparable lymphocyte infiltration suggests that adaptive immune cell recruitment remained unchanged. Together with the absence of neutrophils in the quantified regions, these findings support the conclusion that the experimental grafts do not provoke excessive acute or chronic inflammatory responses during early implantation [72]. Although not statistically significant, experimental APENGs demonstrated lower CD163-positive macrophage density than controls. Because CD163 expression is commonly associated with alternatively activated, pro-remodeling macrophages [73], this observation may reflect a difference in the progression of tissue remodeling rather than an impaired healing response, and should be explored in longer-term studies. Similarly, experimental APENGs exhibited a nonsignificant increase in CC3-positive cells. Apoptosis is a normal component of early tissue remodeling, contributing to the resolution of inflammation and turnover of infiltrating cells [74]. Therefore, the modest increase in apoptotic cells should not necessarily be interpreted as adverse, particularly given the comparable overall cellular infiltration, FBGC density, and lymphocyte density between groups. Furthermore, although none of these inflammatory metrics differed significantly between APENG groups, the observed immune response was comparable to that reported for clinically established ePTFE vascular grafts [75,76]. Together, these findings suggest that AP does not significantly exacerbate the early host inflammatory response at this sample size and that APENGs maintain a level of biocompatibility consistent with the well-established immune profile of ePTFE.

Despite these encouraging early findings, several limitations should be acknowledged. First, the in vivo evaluation was limited to a two-week implantation period, which is sufficient for assessing acute patency and early host responses but not for evaluating longer-term remodeling, including neointimal hyperplasia, endothelial maturation, extracellular matrix remodeling, calcification, graft mechanical evolution, or sustained patency. In addition, the relatively small sample size limits the statistical power to detect subtle biological differences between groups; therefore, the trends observed in cellular infiltration and tissue integration should be interpreted cautiously. The present study also employed a simplified experimental model, using a carotid interposition graft rather than a more clinically representative end-to-side peripheral bypass configuration. This design was selected intentionally to establish the feasibility of implanting axially prestretched grafts while minimizing confounding factors, including anastomotic geometry, limb flexion, and complex peripheral loading conditions. By reducing these sources of variability, this model enabled us to isolate the effects of AP, identify important considerations for controlling graft AP, particularly the redistribution of axial tension caused by native arterial retraction, and evaluate the initial in vivo feasibility of the approach. Although the ultimate goal is to investigate the effects of AP in more biomechanically demanding peripheral bypass models, where restoration of physiological axial mechanics may provide even greater benefit, the insights gained from this feasibility study establish an important foundation for future investigations in more challenging biomechanical models and smaller-diameter vascular beds. Furthermore, while histological analyses provided evidence of early endothelial coverage, functional endothelial performance was not evaluated. Measurements of nitric oxide production, vasoreactivity, and other markers of endothelial function were beyond the scope of the present study but will be important in future work to determine whether structural endothelialization translates into functional vascular regeneration. Future studies with larger cohorts and extended follow-up durations are needed to establish the durability, safety, and long-term regenerative performance of APENGs and to determine whether the early advantages associated with AP are maintained over time.

5. Conclusions

This study introduces APENGs as a mechanics-driven, bioactive vascular substitute designed to restore physiological axial prestretch, a hallmark of native arteries that has been largely overlooked in current synthetic grafts. Using a custom ultrahigh-speed rotational electrospinning method, APENGs achieve strong circumferential fiber alignment, pronounced mechanical anisotropy, native-like axial stiffness, and nonlinear stiffening while maintaining an interconnected porous microstructure conducive to endothelialization and tissue integration. Despite their low-force axial compliance, APENGs retained robust mechanical performance, tolerating substantial prestretch (AP > 1.5) and withstanding suturing, pressurization, and implantation without tearing, anastomotic failure, or leakage. Together, these properties enabled safe, axially biomimetic vascular reconstruction.

Building on this mechanically and surgically robust platform, we demonstrate, to our knowledge, the first preclinical feasibility study of a synthetic vascular graft implanted under controlled physiological AP. In a bilateral swine carotid interposition model, prestretched APENGs retained elevated AP in vivo, exhibited reduced tortuosity, preserved radial pulsatility, and achieved rapid mid-graft endothelialization within two weeks. These early functional and biological responses indicate that physiological AP can be incorporated into synthetic graft implantation without compromising circumferential strength, burst resistance, or early healing.

More broadly, this work challenges the entrenched “tension-free” anastomosis paradigm by reframing AP from a surgical risk to be avoided into a deliberate biomimetic design parameter. Beyond its potential as a regenerative vascular graft, APENGs may also provide a controllable platform for investigating AP as an underexplored regulator of vascular biomechanics, mechanobiology, and remodeling. Although this study is exploratory and limited by a small sample size, short follow-up, and the learning curve associated with a new implantation strategy, these proof-of-concept results establish a foundation for axially biomimetic vascular reconstruction. Future studies should optimize the magnitude and method of prestretch application, refine antithrombotic and biointerface strategies, and evaluate longer-term remodeling, patency, endothelial function, smooth muscle incorporation, neointimal regulation, and durability in demanding peripheral models.

Supplementary Material

Supplementary material associated with this article includes detailed methods, additional mechanical characterization, in vitro biocompatibility analyses, surgical and ultrasound imaging, angiography, histological datasets, quantitative statistical analyses, and full-slide histology images supporting the development and preclinical evaluation of axially prestretched elastomeric nanofibrillar grafts (APENGs) for restoring artery-like axial biomechanics.

Data availability statement

The datasets used and analyzed in this study are available from the corresponding authors upon request. Full histology images are openly available in Zenodo at the DOIs referenced throughout the Supplementary Material.