Highly compliant biomimetic scaffolds for small diameter tissue-engineered vascular grafts (TEVGs) produced via melt electrowriting (MEW)

Biofabrication approaches toward the development of tissue-engineered vascular grafts (TEVGs) have been widely investigated. However, successful translation has been limited to large diameter applications, with small diameter grafts frequently failing due to poor mechanical performance, in particular mismatched radial compliance. Herein, melt electrowriting (MEW) of poly(ϵ-caprolactone) has enabled the manufacture of highly porous, biocompatible microfibre scaffolds with physiological anisotropic mechanical properties, as substrates for the biofabrication of small diameter TEVGs. Highly reproducible scaffolds with internal diameter of 4.0 mm were designed with 500 and 250 µm pore sizes, demonstrating minimal deviation of less than 4% from the intended architecture, with consistent fibre diameter of 15 ± 2 µm across groups. Scaffolds were designed with straight or sinusoidal circumferential microfibre architecture respectively, to investigate the influence of biomimetic fibre straightening on radial compliance. The results demonstrate that scaffolds with wave-like circumferential microfibre laydown patterns mimicking the architectural arrangement of collagen fibres in arteries, exhibit physiological compliance (12.9 ± 0.6% per 100 mmHg), while equivalent control geometries with straight fibres exhibit significantly reduced compliance (5.5 ± 0.1% per 100 mmHg). Further mechanical characterisation revealed the sinusoidal scaffolds designed with 250 µm pores exhibited physiologically relevant burst pressures of 1078 ± 236 mmHg, compared to 631 ± 105 mmHg for corresponding 500 µm controls. Similar trends were observed for strength and failure, indicating enhanced mechanical performance of scaffolds with reduced pore spacing. Preliminary in vitro culture of human mesenchymal stem cells validated the MEW scaffolds as suitable substrates for cellular growth and proliferation, with high cell viability (>90%) and coverage (>85%), with subsequent seeding of vascular endothelial cells indicating successful attachment and preliminary endothelialisation of tissue-cultured constructs. These findings support further investigation into long-term tissue culture methodologies for enhanced production of vascular extracellular matrix components, toward the development of the next generation of small diameter TEVGs.


Introduction
Cardiovascular disease (CVD) remains the leading cause of death worldwide, despite significant advances in surgical procedures, preventative approaches and public awareness [1].Vessel bypassing procedures utilising autologous vascular grafts remain the gold-standard treatment option in the management of CVD.Commercially available synthetic alternatives are frequently used in cases where such autologous vasculature is inaccessible or unsuitable for surgical use.However, there are significant limitations associated with these approaches in small diameter (<6 mm) vessel bypassing applications.In particular, small diameter synthetic grafts have been reported to fail at up to 75% within 3 years of implantation, with clinical complications including thrombosis, infection, intimal hyperplasia, aneurysm formation, and calcification, the principal causes [2,3].Such complications are primarily associated with poor biological integration and biomechanical deficiencies associated with small diameter synthetic grafts, with secondary intervention often required.
In the manufacture of synthetic vascular grafts, mechanical and biocompatibility testing, among other characteristic assessments, are critical for comprehensive evaluation.Detailed mechanical testing standards for the characterisation of manufactured vascular grafts are described within ISO 7198:2017 [4].In accordance with the methods described, longitudinal and circumferential tensile tests for assessment of uniaxial mechanical properties facilitate derivation of graft strength and elastic modulus based on the stress-strain response.Such tests may also be utilised for the estimation of burst pressure where pressurised methods are unsuitable [5].
Furthermore, specifications of dynamic test methodologies for determination of mechanical compliance, enabling assessment of graft anisotropy across physiological pressures, are detailed.As mechanical characterisation of healthy human vasculature is dependent on physiological vessel function and morphology, as well as donor age and condition, a broad range of published mechanical properties are reported in the literature.As such, human arterial vessel burst pressures range between 1200 and 4200 mmHg [6][7][8], while arterial compliance range between 8.0% and 17.0% per 100 mmHg [9].
It is well established that while overall mechanical performance is important to graft patency, mismatched radial mechanical compliance, specifically at the anastomosis site, is one of the most critical contributing factors to the development of intimal hyperplasia and in turn mechanical graft failure [10,11].The onset and progression of intimal hyperplasia is attributed to inconsistent haemodynamics, directly associated with inadequate graft compliance and often geometric incompatibility with patient requirements [12,13].
In order to overcome the limitations associated with current clinical treatment strategies, substantial research into regenerative medicine approaches, incorporating new and existing technologies toward the biofabrication of tissue-engineered vascular grafts (TEVGs) has been conducted.Many such approaches have investigated the use of various material compositions to integrate mechanical compliance, while others have utilised dynamic culture processes toward enhanced compliance characteristics in TEVGs [14,15].While these have been somewhat successful, none have been approved for use as products in clinical practice.
Existing manufacturing methods toward the biofabrication of small diameter TEVGs have frequently employed bioprinting approaches [16][17][18][19], however such constructs often lack the required mechanical properties for clinical translation.Contrastingly, scaffold-based TEVGs fabricated from polymers which are biocompatible, have shown clinical potential [14], though optimisation of scaffold architecture is required to ensure homogeneous morphology and mechanical performance.Scaffold substrates to be utilised for the biofabrication of vascular grafts must be able to be manufactured via processes which enable precise control of scaffold morphology and mechanics [20].Additionally, material biocompatibility is essential to ensure optimal cellular adherence and growth in vitro, while ensuring a non-immunogenic in vivo response [21].
Further, for the incorporation of cellular components, it is essential that scaffolds exhibit sufficient porosity for cellular infiltration and nutrient diffusion, while supporting extracellular matrix (ECM) production in vitro [15].The regenerative potential of such TEVGs is dependent on substrate biodegradability, hence the material choice must be thoroughly considered for appropriate tissue integration and eventual replacement by host tissue over time.
Recently, melt electrowriting (MEW), an electrohydrodynamic additive manufacturing technology has been investigated for use in the manufacture of tissue-engineered heart valves [22,23].As a meltbased extrusion technology, MEW overcomes the toxicity issues associated with solution electrospun scaffolds, while also enabling exponentially greater control over the morphological architecture of scaffolds produced.While there are a number of biocompatible polymers suitable for use in MEW, poly(εcaprolactone) (PCL) is the polymer utilised most frequently in MEW for soft tissue engineering applications.PCL has been used in approved medical device manufacture by both the U.S. Food and Drug Administration (FDA) and Australian Therapeutic Goods Administration (TGA) [24][25][26], and possesses a low melting point (60 • C) above which the material remains stable for long periods [24,26,27].
Furthermore, through the investigation of bioinspired, spatially heterogeneous MEW scaffolds as templates for heart valve reconstruction, design methodologies for the incorporation of biomimetic mechanical properties have been developed [28,29].Such anisotropic mechanical properties are characteristic of soft tissue, often represented by a J-shaped stress-stain response to mechanical loading, as collagen fibre recruitment is load dependent.Hence, the design of biomimetic architectural scaffolds with precisely defined serpentine architecture have been shown to impart structural-functional relationships which mimic the fibrillar morphology and directionality of collagen and elastin components of vascular tissue; thus exhibiting viscoelastic behaviour [22,23,29].
Herein, we aim to enable translation of such biomimetic MEW scaffold architecture toward small diameter TEVGs, as a potential solution to the current lack of mechanically compliant, biofabricated grafts available.It is expected that the continual advancement of biomaterials science, micro-scale additive manufacturing and tissue culture technologies will enable production of the next generation of small diameter TEVGs, overcoming the limitations associated with current clinical treatment strategies in the management of CVD.

MEW of tubular scaffolds
Medical grade PCL granules (Mw = 81 kg mol −1 , Mn = 54 kg mol −1 ) (Purasorb PC 12, Corbion, The Netherlands) were sourced and used as received.A custom-built MEW device described previously [20,29] fitted with a 4.0 mm diameter rotating stainless steel collector was utilised for the manufacture of tubular scaffolds (figure S1, supplementary data).GCode patterns for four architectural groups were designed for fabrication of scaffolds with pore spacing of 0.5 and 0.25 mm with straight perpendicular fibres resulting in scaffolds with square pores, and scaffolds with sinusoidal and straight perpendicular fibres produced for assessment of integrated mechanical compliance.The two pore sizes were selected based on review of published literature, whereby scaffolds with pore sizes of 0.5 mm or less have been demonstrated to facilitate suitable cellular attachment and infiltration within a reproducible in vitro culture period [30,31].Scaffolds with pore spacing less than 0.25 mm were not included in this study, due to limitations associated with the MEW fabrication process, whereby at small inter-fibre distances, the accumulation of electrostatic charges resulted in fibre bridging and inconsistent scaffold architecture as described in the literature [32].Stereomicroscope images of scaffolds with reduced 0.125 mm pores (figure S2, supplementary data) depict the poor morphology of such constructs despite attempts toward optimisation.As a result, pore spacing of 0.25 mm was determined as the minimal spacing for accurate production of highly consistent sinusoidal architectures suitable for tissue culture studies alongside mechanical assessment of differing scaffold geometries in this work.
For extrusion of PCL as a molten polymer, a 3 cc syringe barrel (Nordson, USA) fitted with a 23 G general purpose dispensing tip (Nordson, USA) was filled with PCL granules and heated to 80 • C in the MEW print head, with the temperature allowed to stabilise for 15 min prior to printing.Electrohydrodynamic extrusion of the polymer jet was controlled through use of a pneumatic regulator (smooth muscle cell (SMC), Japan) set to 0.025 MPa, with an electric potential of 4.0 kV applied between the nozzle and the grounded collector by the high voltage power supply within the MEW device.A 4.0 mm working distance was maintained between the tip and rotating collector, with translational speed maintained at a constant 280 mm min −1 .Printing was conducted in ambient conditions.

Scaffold characterisation
Images of tubular scaffolds were obtained using a stereomicroscope (SMZ25, Nikon, Japan), while morphological characterisation of fibre architecture was determined via scanning electron microscopy (SEM) (TM3000, Hitachi, Japan).PCL scaffolds imaged via SEM were analysed at 8 mm working distance under vacuum conditions, with an accelerating voltage of 15.0 kV.Scaffold geometries (n = 12) from each group were analysed using ImageJ [33] to obtain fibre diameter, wall thickness and pore spacing measurements from multiple regions.Values were reported as mean ± SD (n = 12).

Mechanical testing
Mechanical characterisation of the tubular scaffolds was conducted in accordance with ISO 7198:2017 [4].Uniaxial tests were performed using an Instron 68TM-30 mechanical testing system (Instron, USA) equipped with a 100 N load cell.Longitudinal and circumferential tensile properties were determined via standardised test methods as subsequently described, using fixtures manufactured as per ISO 7198:2017 [4].Dynamic radial compliance testing facilitated characterisation of the anisotropic behaviour and mechanical compliance of the tubular scaffold designs.Data obtained for each of the methods described throughout were summarised alongside physiologically reported ranges (table S1, supplementary data).

Longitudinal tensile testing
In accordance with ISO 7198:2017 section A.5.2.3 [4], tensile testing was performed to determine the longitudinal mechanical properties of the tubular scaffold constructs.Suitable test fixtures were manufactured to interface directly with the upper and lower load cell fittings, within which the 4.0 mm inner diameter tubular scaffolds were secured (figure S4, supplementary data).Tubular scaffold samples (n = 6) from each group were cut to a suitable length, ensuring an initial separation between fittings of 50 mm.Testing was performed using the Instron 68TM-30 testing system (Instron, USA) equipped with a 100 N load cell.The upper fixture was raised at a uniform rate of traverse of 100 mm min −1 until failure.For determination of stress-strain from the force-displacement data obtained, the cross-sectional area and length of each sample was used, with the elastic moduli determined from the linear region of the stress-strain data.Values were reported as mean ± SD (n = 6).

Circumferential tensile testing
For characterisation of the circumferential mechanical properties of the tubular scaffold constructs, tensile testing was performed in accordance with ISO 7198:2017 section A.5.2.4 [4].Matched upper and lower fixtures were manufactured to interface directly with the load cell adapter, supporting 1.5 mm diameter pins, over which the 4.0 mm diameter tubular scaffolds were positioned (figure S5, supplementary data).Tubular scaffold samples (n = 6) from each group were cut to a suitable length of 5.0 mm.Testing was performed using the Instron 68TM-30 testing system (Instron, USA) equipped with a 100 N load cell.
From an initial position with nil separation between pins, the upper fixture was raised at a uniform rate of 100 mm min −1 until failure.Stressstrain data was derived from the force-displacement data obtained, with the area used in calculations defined as the cross-sectional area of the samples aligned with the applied load.In turn, the elastic moduli of samples were determined from the linear region of the stress-strain data, while the ultimate tensile strength (UTS) and corresponding percentage strain were calculated at failure.Values were reported as mean ± SD (n = 6).
Further, while pressurised testing is the gold standard for vascular graft burst strength, alternative tests are permissible in accordance with ISO 7198:2017 [4] if pressurised tests cannot be readily conducted; as required for MEW scaffolds due to the highly porous architecture.Such alternatives test methods have been validated throughout literature to provide effective approximation of graft burst strength where pressurised tests are unsuitable [34][35][36].It has been noted that indirect methods may overestimate graft burst pressure; however, if graft homogeneity can be assumed, as is the case with the MEW scaffolds produced, these methods remain viable [5].
Thus, based on Laplace's law, denoted in equation (1), where σ θ is the circumferential wall stress, in Pa; P is the internal pressure, in Pa; t is the wall thickness, in m; and R i is the inner radius, in m, Additionally, circumferential wall stress, σ θ can be calculated based on the force, F in N, and the initial perpendicular cross-sectional area which is derived from the sample thickness, t and initial length L 0 , both in m, By combining these equations, equation ( 3) can be derived, allowing the burst pressure to be calculated.Where F max is the maximum circumferential load recorded prior to failure, in N; L 0 is the initial length of the sample; and D the diameter, in m,

Probe burst testing
Probe burst tests were performed for comparison with previously described methods appropriate for the determination of tubular scaffold strength.
Various studies have investigated the suitability of the probe burst test method, however not all comply with the dimensional requirements specified by ISO 7198:2017 section A.5.2.6 [4].Consequently, direct comparisons cannot be made with existing studies unless probe test apparatus conform to these specifications, as these factors directly influence the data obtained [36].As such, the components for the probe burst test were manufactured in accordance with ISO 7198:2017 [4].The cylindrical traversing probe exhibited a hemispherical tip with diameter of 9.5 mm, designed to interface directly with the load cell adapter to ensure concentric alignment with the 11.3 mm diameter opening in the clamp and base plates (figure S6, supplementary data).Tubular scaffold samples (n = 6) from each group were cut along the longitudinal axes and flattened to form single thickness sheets, with the flattened width equal to the circumference of the 4.0 mm inner diameter tubular constructs, calculated to be approximately 12.56 mm.Samples were secured between the clamp and base plates, covering the central opening.Testing was performed using the Instron 68TM-30 testing system (Instron, USA) equipped with a 100 N load cell.The probe was lowered through each sample at a constant rate of traverse of 100 mm min −1 until failure, with the maximum load determined.Scaffold stiffness was then calculated from the linear region of the force-displacement curves obtained.Values were reported as mean ± SD (n = 6).

Dynamic compliance testing
Radial compliance of scaffolds was determined using a pulsatile test setup enabling comparative pressurediameter measurements to be obtained across physiologically relevant pressure ranges.Dynamic testing effectively simulated in vivo pressures during cyclic loading, in accordance with test methods described in ISO 7198:2017 [4] (figure S7, supplementary data).Testing was performed within a sealed chamber (TA Instruments, USA), with continuous pressure monitoring enabled through use of an in-line pressure transducer (PendoTech, USA), interfaced with Statys Pressure Monitor software (BDC Laboratories, USA).Highly reproducible cyclic flow was facilitated through the use of a programmable syringe pump (World Precision Instruments, Germany) and a 60 ml syringe filled with phosphate buffered saline (PBS).
For testing, tubular scaffold samples (n = 6) from each group were cut to a test length of 40 mm.Due to the highly porous architecture of the scaffolds, an elastic, non-permeable liner was utilised throughout dynamic testing of scaffold compliance, as permitted by ISO 7198:2017 [4], whereby at a pressure of 120 mmHg the liner exhibited a diameter notably greater than the nominal pressurised diameter of the scaffolds.Throughout testing, a digital microscope (Dino-Lite, Australia) was used to capture video recordings at 30 Hz for analysis of scaffold diameter using a video processing protocol incorporating the VasoMetrics [37] macro for ImageJ [33].Video frames were incrementally sectioned to directly measure sample diameter along the length of the test specimen at cyclic pressures between 0 and 160 mmHg for at least six cycles.Scaffold compliance, expressed as a percentage per 100 mmHg, was calculated according to equation ( 4) at three pressure ranges of 50-90 mmHg, 80-120 mmHg and 110-150 mmHg to assess non-linear behaviour, as per ISO 7198:2017 [4], with overall compliance also calculated.Values were reported as mean where R p is the internal radius, in m; p 1 is the lower pressure valve, in mmHg; and p 2 is the higherpressure value, in mmHg.As direct measurements of the external scaffold diameter were obtained, the R p values were calculated according to equation ( 5), where D is the measured diameter, and t is the known scaffold wall thickness, both in m, Hysteresis curves for each scaffold group were produced for comparative analysis of the pressurestrain variations between samples, to evaluate the reproducibility in the anisotropic properties of the sinusoidal scaffolds produced.

Scaffold surface modification
To validate the capacity of the PCL scaffolds with varying architectures to support cellular attachment, growth, and ECM deposition, a 10 day in vitro study was performed.Firstly, scaffold samples were treated with O 2 /Ar 2 (15/5) plasma for 2 min at 38 W in a vacuum plasma cleaner (PDC-002-HP, Harrick Plasma, USA) to induce hydrophilicity within PCL scaffolds prior to cell seeding.The protocol was developed in accordance with published methods for the plasma treatment of PCL scaffolds; whereby exposure to plasma for 2 min was shown to induce complete scaffold wettability suitable for cell suspension infiltration, while only reported to affect a minimal 11% increase in tensile modulus and negligible change in polymer crystallinity, when compared to untreated PCL scaffolds produced via MEW [38,39].
Samples were stored under vacuum and used for cell seeding within 24 h.For validation of PCL scaffold plasma treatment, contact angle measurements were performed using 5 and 10 µl droplets of dH 2 O. Contact angle images (figure S8, supplementary data) were acquired with a digital microscope (AD7013MT, Dino-Lite, The Netherlands) and analysed with ImageJ [33].Values were reported as mean ± SD (n = 3) (figure S9, supplementary data).

Cell seeding and tissue culture
A 10 day cell compatibility study was performed using human mesenchymal stem cells (MSCs) purchased from the Tulane Center for the Preparation and Distribution of Adult Stem Cells (Tulane University, USA), isolated from bone marrow aspirates of normal adult donors via approved protocols [40,41].MSCs were cultured in vitro with alpha minimum essential media (MEM-α) (Gibco, USA), containing 10% foetal bovine serum (FBS) (Gibco, USA) and 1% penicillin streptomycin (Invitrogen, USA).MSCs were expanded in 175 cm 2 flasks for 7 d prior to seeding, plated at a density of 5000 cells per cm 2 .Individual scaffold samples were sterilised in 80% ethanol for 30 min and allowed to dry via evaporation inside a biosafety cabinet for 30 min.MSCs between passage 4-6 were then deposited onto scaffold samples at a seeding density of 10 million cells per ml, using 5 µl per sample.Samples were placed in an incubator for 3 h at 37 • C, 5% CO 2 and 95% humidity for attachment.Some attachment of cells to culture plates was observed, though predominantly, cells attached to the plasma treated scaffold substrates, with samples washed and transferred to new plates for staining, imaging and analyses.Subsequently, 0.5 ml of MEM-α was added to each sample within 24-well plates, with exchanges made every 2-3 d thereafter.
Additionally, to validate the compatibility of the scaffold constructs to support endothelial cell growth, human umbilical vein endothelial cells (HUVECs) (PromoCell GmbH, Germany) were expanded for 7 d at a density of 5000 cells per cm 2 in Endothelial Cell Growth Medium (EGM-2) (CC-3162, Lonza, USA) containing 0.04% hydrocortisone, 0.4% hFGF-B, 0.1% vascular endothelial growth factor, 0.1% recombinant insulin-like growth factor (R3-IGF-1), 0.1% ascorbic acid, 0.1% hEGF, 0.1% GA-1000, 0.1% heparin, and 2% FBS.At the completion of the MSC culture, HUVECs between passage 5-6 were seeded onto the surface of the MSC-cultured constructs of each architectural group, at a density of 10 million cells per ml, using 5 µl per sample.The HUVECseeded constructs were placed in an incubator for 2 h at 37 • C, 5% CO 2 and 95% humidity for attachment, with 0.5 ml of EGM-2 media added to each sample within 24-well plates, and media exchanges made every 2-3 d thereafter for a 7 day period.

Cell viability
Live and dead cell imaging MSC-seeded scaffolds was performed on days 1 and 10 with FDA (Sigma-Aldrich, USA) and propidium iodide (PI) (Sigma-Aldrich, USA) to determine viability.Prior to staining and imaging, samples (n = 6) from each architectural scaffold group were washed in PBS, and transferred to new well plates containing fresh PBS, to ensure cellular components imaged were only those attached to scaffolds and not culture plates.Samples were then stained in 10 µg ml −1 FDA and 5 µg ml −1 PI in PBS for 5 min, then rinsed with PBS and stored in fresh PBS in new well plates.Images were obtained via inverted fluorescence microscopy (AxioObserver Z1, Carl Zeiss, Germany).Cell viability was determined based on the percentage of live and dead stained cell counts analysed via ImageJ [33], using a processing method by which images were binarized with particle analysis to identify individual cells, with cell density per mm 2 calculated with respect to total area within each region of interest.Values were reported as mean ± SD (n = 6).

Histology
For histological staining and imaging, fixed tissuecultured samples were embedded in 2% w/v agarose (Meridian Life Science, USA) for tissue processing (Leica 300S, Leica Microsystems, Germany).Samples were then embedded in paraffin blocks, with 5 µm sections obtained for routine hematoxylin and eosin (H&E) staining.Stained sections were imaged via brightfield microscopy (AxioObserver Z1, Carl Zeiss, Germany).

Statistical analysis
Statistical analyses were performed using GraphPad Prism software (Dotmatics, USA).Measurements performed in replicate were reported as mean ± standard deviation (SD).After confirmation of normality, differences between groups were determined using one-and two-way analysis of variance (ANOVA) tests as appropriate, with values of p < 0.05 considered significant.Assessment of differences between groups was performed based on Tukey multiple comparison tests, with a confidence level of 95% (p < 0.05) considered statistically significant.Throughout figures, symbols (p < 0.001 * * * , 0.001 < p < 0.01 * * , 0.01 < p < 0.05 * ) indicate level of significance.

Scaffold morphology
Observed under both brightfield microscopy (figures 1(A) and (B)) and SEM (figure 1(C)), the PCL scaffolds produced via MEW exhibited highly consistent morphology, with minimal deviation in pore spacing from the intended GCode designs for both the 0.5 and 0.25 mm control (denoted A) and sinusoidal (denoted B) designs (figure 1).No significant differences in fibre diameter were exhibited between any of the groups, with measured diameters of 14.9 ± 2.9 µm, 14.5 ± 1.2 µm, 14.8 ± 1.7 µm, 14.3 ± 1.8 µm for the 0.5A, 0.5B, 0.25A, and 0.25B groups, respectively (figure 1(D)).The measured pore spacing of the 0.5A and 0.5B architectural groups were analysed as 501.1 ± 7.1 µm and 503.1 ± 9.6 µm, respectively; while the corresponding 0.25A and 0.25B pore spacings were 251.2 ± 4.3 µm and 250.2 ± 6.4 µm (figure 1(E)).Statistically significant differences in these results were observed only between intended pore size design, with negligible variation between control and sinusoidal arrangements at each specified size, demonstrating the accuracy of the MEW fabrication technique to produce scaffolds in accordance with intended GCode pattern inputs.Correspondingly, the 0.5A and 0.25A samples exhibited minimal deviation of 1.08 ± 0.89% and 1.40 ± 1.02% from the control GCode design patterns; while the 0.5B and 0.25B samples exhibited slightly greater deviation from the sinusoidal designs, at 1.74 ± 0.90% and 2.31 ± 0.92% deviation (figure 1(F)).
Such minor deviations in the PCL scaffolds with respect to their GCode designs further validate the high geometric accuracy achieved through use of MEW additive manufacturing technologies.No significant differences in the wall thicknesses of the tubular scaffold groups were noted, with thicknesses of 404.7 ± 26.9 µm, 395.0 ± 23.8 µm, 398.7 ± 9.4 µm, 401.0 ± 9.9 µm, for the 0.5A, 0.5B, 0.25A, and 0.25B groups respectively (figure 1(G)).Notably, greater deviation in measured thicknesses were determined within the scaffolds designed with larger pores.The data corresponded with qualitative observations, whereby the thicknesses of the intersection points of the MEW fibres were more pronounced within the 0.5A and 0.5B constructs due to the greater separation between parallel fibres, evident within supporting SEM images obtained (figure S3, supplementary data).

Uniaxial mechanical testing
Longitudinal tensile testing enabled characterisation of the axial mechanical performance of scaffold groups.Assessment of the stress-strain response (figure 2(B)) of each group subject to longitudinal tensile loading revealed a significant increase in elastic moduli derived from the linear region of each curve with respect to reduced pore size, while only limited increase was observed with the incorporation of sinusoidal circumferential fibres.As presented in figure 2(C), the 0.5A and 0.5B scaffold groups exhibited longitudinal elastic moduli of 2.81 ± 0.11 MPa and 4.76 ± 1.37 MPa, respectively; while statistically greater elastic moduli of 10.24 ± 0.21 MPa and 13.50 ± 2.07 MPa were exhibited by the 0.25A and 0.25B scaffolds, attributed to the greater reinforcement of the smaller pore sizing.These results are similar to the longitudinal modulus reported for the internal mammary artery (16.8 ± 7.1 MPa) and much greater than that reported for the coronary artery (1.48 ± 0.24 MPa) [42].An equivalent trend in the longitudinal UTS of samples was observed (figure 2(D)), with the 0.5A, 0.5B, and 0.25B groups exhibiting relatively low UTS of 0.49 ± 0.03 MPa, 0.59 ± 0.06 MPa, and 0.89 ± 0.05 MPa, respectively; while the 0.25B group exhibited UTS of 1.21 ± 0.12 MPa, tending toward physiological longitudinal strength of the coronary artery (1.44 ± 0.87 MPa) and the internal mammary artery (4.3 ± 1.8 MPa) as reported in literature [42].
Corresponding circumferential tensile testing enabled further characterisation of the overall mechanical performance of the various architectural arrangements incorporated into the tubular scaffolds.The anisotropic mechanical properties of all scaffolds were evident, with the circumferential stress-strain responses (figure 2(F)) characterised by reduced linear elastic gradient relative to the longitudinal results obtained.Statistically similar radial elastic moduli were determined for the 0.5B (0.79 ± 0.11 MPa), 0.25A (0.84 ± 0.12 MPa) and 0.25B (0.79 ± 0.05 MPa) groups, while the 0.5A constructs exhibited significantly reduced elastic modulus (0.50 ± 0.06 MPa).Notable differences in the curvature of the stress-strain responses were observed within the toe region prior to linear elastic deformation, particularly in the scaffold groups exhibiting sinusoidal architectural arrangements, as intended.Interestingly, the magnitude of the biomimetic J-shaped stress-strain response of the sinusoidal constructs was observed to be slightly greater in the 250 µm group than the 500 µm group with the same design, while the control geometries exhibited negligible toe region elongation.
Given the high porosity of MEW scaffolds, direct pressurised burst methods were unsuitable, hence the circumferential tensile strength of the constructs was utilised for the calculation of equivalent burst pressures, using the derivations previously described.Burst pressures for the 0.5A and 0.5B scaffold groups were below physiological requirements, calculated to be 303.43 ± 50.25 mmHg and 631.16 ± 105.66 mmHg, respectively.Notably however, the calculated burst pressures of the 0.25A and 0.25B scaffold groups were statistically similar at 1027.32 ± 108.64 mmHg and 1078.70 ± 235.76 mmHg, respectively, at the lower range of values reported within literature (1200 mmHg) [6][7][8] though the methods used herein were not direct measures, with such tests essential to assess burst pressure of long-term tissue-cultured constructs in future studies.
Probe burst tests performed using flat sections of scaffold geometries, provided indicative assessment of the biaxial properties of the various architectural arrangements.Failure of the flattened scaffold samples occurred centrally in the probe burst tests performed, with the maximum force prior to failure recorded.While stress-strain responses were unable to be determined due to the testing method, the force-displacement curves (figure 2(J)) were obtained.It was observed that the sinusoidal scaffold groups at both pore sizes, exhibited significantly greater stiffness (figure 2(K)) within the linear elastic region, while demonstrating an ability to withstand greater maximum tensile forces (figure 2(L)).
The 0.5A and 0.5B groups exhibited stiffness of 0.41 ± 0.01 N mm −1 and 0.42 ± 0.01 N mm −1 respectively; while 0.25A and 0.25B stiffness was significantly greater at 1.16 ± 0.04 N mm −1 and 1.14 ± 0.02 N mm −1 .Notably similar trends were observed between scaffold groups in terms of maximum force, with the 0.5A and 0.5B groups exhibiting a maximum force of 2.39 ± 0.07 N and 2.57 ± 0.03 N, compared to 4.62 ± 0.10 N and 4.57 ± 0.07 N for the 0.25A and 0.25B scaffold groups.

Dynamic compliance testing
Comprehensive investigation of the dynamic mechanical performance of the constructs was conducted cyclically across a physiological pressure range, using a pulsatile setup as shown in figure 3(A).Radial compliance calculated at three pressure ranges of 50-90 mmHg, 80-120 mmHg and 110-150 mmHg enabled assessment of the non-linear behaviour (figure 3(B)), while overall radial compliance was also determined (figure 3(C)) based on analysis of the pressure-diameter relationships recorded.
The overall radial compliance of the 0.5A and 0.25A geometric control scaffolds was determined to be 1.25 ± 0.15% per 100 mmHg and 5.5 ± 0.1% per 100 mmHg, respectively with a general downward trend exhibited in the ranged compliance of the control groups, at the increased pressure variations.The corresponding hysteresis curves of both control groups (figures 3(D) and (F)) exhibited a relatively linear response, indicative of negligible compliance.Contrastingly, the sinusoidal scaffold constructs 0.5B and 0.25B, exhibited overall dynamic compliance of 8.62 ± 1.26% per 100 mmHg and 12.9 ± 0.6% per 100 mmHg, significantly greater than the rectangular controls.These results were reflected in the ranged compliance calculations, with the maximum ranged compliance for the 0.25B architectural group, exhibited within the physiological pressure range of 80-120 mmHg, being 14.5 ± 1.5% per 100 mmHg.This is further supported by the shape and data portrayed within the hysteresis curve of the 0.25B scaffold group, with consistent J-shaped cyclic compliance exhibited across the tested pressure range.Thus, through incorporation of the circumferential sinusoidal fibres patterning, geometric straightening was observed resulting in enhanced radial compliance prior to material deformation of the MEW scaffolds with sinusoidal designs.The compliance data obtained for the sinusoidal constructs closely match the physiological requirements of native vessels, with values reported between 8.0% and 17.0% per 100 mmHg [9].Further, the compliance of the scaffolds produced were significantly greater than reported values for synthetic Dacron (1.8 ± 1.2% per 100 mmHg) and polytetrafluoroethylene (1.2 ± 0.3% per 100 mmHg) grafts [43].

Cell culture
Live and dead cell staining of the MSC-seeded scaffold samples enabled quantification of viability at Day 1 and 10 timepoints, determined through ImageJ [33] analysis.Scaffold samples imaged via fluorescence microscopy (figure 4(A)) indicated few dead cells (stained red), with substantial attachment and growth of live cells (stained green).Viability, expressed as a percentage of live to total cells stained, was high (>90%) across timepoints, with no significant differences in viability between architectural scaffold arrangements (figure 4(B)).Furthermore, cell densities were calculated for each group based on these results (figure S10, supplementary data), indicating substantial increases between timepoints across groups.
While FDA/PI staining provided a qualitative indication of cell coverage, analysis of DAPI/Phalloidin fluorescence images (figure 4(C)) enabled quantification of F-actin coverage after 10 days in vitro culture (figure 4(D)).Across all architectural scaffold groups, high scaffold coverage (>85%) was observed at the Day 10 timepoint, with only the 0.5A scaffold geometry exhibiting mean coverage below 95%.Extensive migration and branching of MSCs along PCL fibres and across scaffold pores was exhibited by all groups.The high cell viability and coverage exhibited by all groups demonstrates the suitability of the micro-fibrous PCL scaffolds produced by MEW to support cell attachment and growth in vitro, assessed over a short time period, validating the use of such scaffolds as substrates for long-term tissue culture.Additionally, H&E-stained sections (figure S11, supplementary data) complement the fluorescence staining results obtained, with early accumulation of ECM components throughout the thickness of the scaffold substrates, after the brief culture period.
Furthermore, MSC-cultured samples subsequently seeded with HUVECs, and cultured for an additional 7 day period, were stained for expression of CD31, to assess the compatibility of the scaffold substrates with vascular endothelial cells.Representative images obtained via CLSM imaging (figure 4(E)) indicate successful attachment and survival of CD31+ endothelial cells on MSCcultured samples.While more extensive in vitro studies are required for complete endothelialisation of the luminal surface of TEVGs, these results verify the compatibility of the scaffold-tissue constructs for culture of vascular endothelial cells.

Discussion
Biofabrication techniques incorporating synthetic biomaterials as substrates for tissue growth and regeneration continue to develop in complexity and biomimetic accuracy.While many processes and technologies utilised in the manufacture of TEVGs exhibit success in large diameter applications, limitations associated with mechanical properties, particularly radial compliance, have restricted the successful translation of such approaches in the manufacture of small diameter TEVGs.
Herein, MEW of medical-grade PCL was utilised for the fabrication of tubular scaffold constructs with physiological dimensions, as substrates for small diameter TEVGs.Through the generation and production of micro-fibrous architectural arrangements with circumferential sinusoidal geometries, anisotropic mechanical properties have been incorporated into the small diameter scaffolds produced, based on processes established by Saidy et al in the biofabrication of tissue-engineered heart valves [22,23].The MEW manufacturing methodology utilised herein enabled production of highly reproducible scaffolds, with negligible variation in fibre diameter and minimal pore spacing deviation, irrespective of architectural arrangement, as depicted throughout figure 1.
Ensuring mechanical performance which mimics that of native human vasculature is critical to graft patency, with mechanical mismatch a key contributor to hemodynamic failure.Herein, MEW of PCL scaffolds which exhibit control and sinusoidal circumferential fibrous architecture were developed to investigate the influence of sinusoidal fibre straightening on substrate anisotropy.Through mechanical characterisation, performed in accordance with international standard protocols, the integrated compliance of the sinusoidal constructs, which exhibited Jshape stress-strain responses, were assessed.Further, as indicated within figure 2, bulk mechanical performance was evaluated, demonstrating physiologically relevant properties.
Most notably, the 0.25B scaffold group exhibited longitudinal elastic modulus of 13.50 ± 2.07 MPa, similar to the longitudinal elastic modulus of 16.8 ± 7.1 MPa exhibited by human mammary arteries [44].Additionally, while the longitudinal UTS of the 0.25B scaffold group (1.21 ± 0.12 MPa) remained slightly below physiological strength (1.4-6.3MPa) [42], the mechanical properties of the constructs are expected to be enhanced with the incorporation of ECM components produced through extended in vitro tissue culture of relevant cell types, such as MSCs as utilised herein and throughout literature [45][46][47], as well as SMCs which have been extensively studied for use in TEVG biofabrication [48][49][50].Incorporation of these specifically relevant cell types in the long-term formation of collagen and muscle tissue components in vitro are required, to be investigated in future work.Further, future investigations into the effects of mechanical stimuli provided by pulsatile flow are proposed to assess their influence on TEVG performance.
While the mechanical properties of scaffolds produced as substrates for tissue-engineering are critical, the biocompatibility of such constructs is also essential for long-term viability.Independent of whether biomaterial-based grafts are intended for implantation in combination with biological components or without, constructs must exhibit suitable properties for host cell repopulation and eventual complete tissue regeneration in vivo.A common limitation associated with many TEVG fabrication approaches which incorporate biological components, is the inability for constructs to support bulk cell attachment and infiltration, leading to poor tissue formation in vitro.These limitations are primarily associated with intrinsic scaffold properties, such as insufficient porosity or poor material compatibility for use with in vitro cell culture processes [51].
To overcome such limitations, several key factors were considered prior to scaffold manufacture; namely, the material selection of medical-grade PCL as a biocompatible, biodegradable polymer used in the manufacture of approved medical devices [24].Further, PCL remains the gold-standard material for the fabrication of scaffold constructs via MEW.Additionally, while MEW continues to develop, the fabrication process has been thoroughly validated throughout literature as specialised additive manufacturing technology uniquely suited to the production of microscale scaffold constructs with high geometric accuracy and porosity, ideal for use in soft tissue regeneration [19,20].
Morphologically, the scaffolds produced in this study exhibited highly reproducible fibre diameter and pore spacing, as represented in figure 1, with negligible deviation from the intended geometric architectures, input as GCode patterns within the MEW machine control software.The geometric homogeneity translated through to the highly consistent mechanical properties of the constructs with minimal variation observed between samples within groups.The high degree of consistency demonstrated further supports the continued application of MEW as a key technology in the manufacture of soft tissue substrates suitable for regenerative medicine applications.
Mechanical characterisation, performed in accordance with international standard protocols, ensured reproducibility within the results obtained while enabling direct comparisons to be made with existing and future studies performed.The MEW scaffolds with 250 µm pores showed sufficient strength to warrant further investigation toward their use as substrates for TEVG biofabrication.The architectural scaffold groups with reduced pore spacing consistently demonstrated greater mechanical performance, with significant emphasis on the highly compliant sinusoidal scaffold geometries (figures 2 and 3).The 0.25B scaffold group exhibited physiological compliance, burst pressure and strength, similar to published mechanical properties for coronary and mammary arteries [42,51].A limitation associated with the mechanical testing performed in this study, is the indirect estimation of burst pressure, due to the porous nature of the scaffold substrates limiting the direct measure of pressurised burst strength.However, all appropriate aspects were considered in the development of the test method, in accordance with ISO 7198:2017 [4], whereby the assumption that manufactured constructs exhibited homogeneous morphology, was validated through geometric characterisation.Hence, the estimation of burst pressure based on the circumferential tests can be assumed to provide results indicative of physiological relevance.An associated benefit in the mechanical characterisation of TEVGs where indirect methods may be utilised, is that the protocols described require markedly less scaffold or graft material for testing, which is important in the biofabrication of TEVG substrates which often require considerable resources.Ultimately however, direct burst pressure measurements remain the gold standard, with such tests to be performed in future studies for complete characterisation of the mechanical properties.This is a limitation associated with performing mechanical tests on highly porous scaffold substrates prior to the addition of biological components.
Overall, characterisation of the 0.25B architectural scaffold group suggests the PCL substrates produced via MEW may pose a viable option for the biofabrication of small diameter TEVGs, with biomimetic mechanical properties tending toward native vasculature.Additionally, the high degree of control enabled by MEW, which facilitated the production of highly reproducible geometries and correspondingly consistent mechanical properties, support continued use and optimisation of the manufacturing methodology.
The ability to tailor mechanical compliance based on the circumferential architectural arrangement has been demonstrated herein (figure 3), enabling suitable mechanical compliance to be directly incorporated to small diameter tubular scaffolds based on microfibre architecture.As such, there exists potential for further incorporation of anisotropic spatial regions within scaffold constructs, such that the performance may be adjusted for a particular grafting application.
Through preliminary in vitro culture of human MSCs and HUVECs, indicative results were obtained for validation of scaffolds as substrates to support vascular tissue formation, as represented in figure 4. Irrespective of architectural scaffold arrangement, both MSC viability and scaffold coverage were high across timepoints, with extensive branching covering the micro-porous MEW scaffolds.The enhanced surface hydrophilicity induced by plasma treatment ensured appropriate initial cell adherence, in accordance with published results [38].Furthermore, successful attachment of endothelial cells was observed after seeding onto MSC-cultured samples, with the HUVECs beginning to cover the surface of the tissueculture structures, indicating preliminary endothelialisation.The findings of this brief co-culture study are promising for extended in vitro culture of multicellular TEVGs utilising the scaffold substrates engineered here.Further research is required into the optimisation of TEVG endothelialisation to develop a smooth, continuous surface resembling the native lumen, with studies into the effects of physiological flow conditions on luminal surface morphology [52].
Future studies are warranted to investigate the long-term effects of architectural scaffold morphology on the differentiation of MSCs and their paracrine signalling.Further, dynamic tissue culture conditions which imitate the physiological mechanical environment, must be assessed in future to ascertain the influence of such conditions on tissue formation in vitro.Additionally, in future work it is critical to ensure the TEVGs developed via MEW and cultured in vitro exhibit clinically applicable properties in regard to leakage and burst pressure.Hence, future tissue culture experiments must be performed over an extended period, in accordance with current literature which report culture times up to 10 weeks for SMCs [53].Optimisation of culture time of the TEVGs reported herein is therefore critical to ensure sufficient time for the accumulation of dense ECM components throughout the scaffold substrates in accordance with studies in the literature [54].Both mechanical and histological assessment in accordance with appropriate methods must be performed to evaluate the performance of the TEVGs cultured longterm, prior to in vivo tests for graft patency.
It must be noted that, while the small diameter sinusoidal scaffolds with 250 µm pores exhibited mechanical properties similar to that of native vasculature, with physiologically high compliance and encouraging preliminary cell culture results, the data reported are limited to a relatively small sample size.Further testing is required, with evaluation of cellseeded scaffold construct mechanical performance and overall tissue production through in vitro culture required for more complete assessment.Ultimately, in vivo studies for critical assessment of mechanical performance, graft patency, biocompatibility, and immunogenicity, remain the benchmark to ensure comprehensive evaluation of alternative grafting options.
In addition to the results presented throughout this investigation, it is critical that future studies assess material degradation under physiological conditions for complete characterisation of TEVG performance, such that trends in long-term mechanical performance may be identified.Degradation studies were not performed herein, as the degradation kinetics of PCL and corresponding mechanical properties have been extensively investigated throughout literature [24,26,55].Most notably, Bartnikowski et al report that pre-clinical assessment represents the most applicable test for characterisation of physiological material degradation, with in vitro test results noted to differ significantly from those ascertained in vivo [26].As such, while PCL is well-established as a biocompatible and biodegradable polymer suitable for soft tissue engineering applications [24,55], pre-clinical in vivo assessment of scaffold degradation is required for complete characterisation of graft performance in future studies.Thus, it may be suggested that with the slow degradation of PCL [24], in combination with host cell infiltration and tissue remodelling, the mechanical integrity of the constructs will be maintained.Hence, future studies are necessary for further evaluation of scaffold substrate performance throughout extended in vitro tissue culture, to determine the influence of ECM production on mechanical and biological performance, as a prelude to in vivo assessment.Furthermore, it is expected that, with the incorporation of physiological pulsatile conditions in a dynamic tissue culture environment, the production of essential ECM components such as elastin and collagen will be enhanced.

Conclusion
Highly reproducible biocompatible tubular scaffolds exhibiting physiologically accurate morphology and customisable architectures, were successfully manufactured via MEW.Comprehensive mechanical characterisation has demonstrated that biomimetic performance may be achieved within small diameter tubular scaffold substrates, through design variations.Furthermore, the MEW scaffold constructs have been validated as substrates for cellular attachment and growth in vitro, substantiating further research toward extended tissue culture practices.Ultimately, the successful production of highly compliant polymeric scaffold substrates which also exhibit physiologically relevant bulk mechanical properties, represents an important step toward the biofabrication of reproducible scaffolds for the biofabrication of small diameter TEVGs utilising MEW technology.

Figure 2 .
Figure 2. Mechanical characterisation of small diameter tubular PCL scaffolds manufactured via MEW (4 mm inner diameter), tested using Instron 68TM-30 testing system in accordance with ISO 7198:2017 test methods.(A) Longitudinal tensile test setup (i) photograph and (ii) diagram (not to scale).Uniaxial mechanical data obtained via longitudinal tensile testing showing (B) stress-strain curves to failure; (C) elastic moduli calculated from the linear region of the stress-strain curves; and (D) ultimate tensile strength (UTS), of tubular scaffold samples (n = 6) from each of the scaffold groups.(E) Circumferential tensile test setup (i) photograph and (ii) diagram (not to scale).Uniaxial mechanical data of tubular scaffold samples (n = 6) from each group obtained via circumferential tensile testing, depicting the (F) stress-strain curves to failure, where the inset depicts the toe region from 0% to 60% strain; (G) the elastic moduli calculated from the linear region of the stress-strain curves; and (H) ultimate tensile strength (UTS), of tubular scaffold samples (n = 6) from each of the groups.(I) Probe test setup (i) photograph and (ii) diagram (not to scale).Uniaxial mechanical data of scaffold samples (n = 6) from each group, depicting the (J) force-displacement curves to failure; (K) the material stiffness calculated from the linear region of the force-displacement curves; and (L) maximum force prior to failure, scaffold samples (n = 6) from each of the groups.All values reported as mean ± SD (n = 6), where symbols (p < 0.001 * * * , 0.001 < p < 0.01 * * , 0.01 < p < 0.05 * ) indicate significance levels, based on ANOVA tests.

Figure 4 .
Figure 4. Viability, proliferation, and morphology of MSC-seeded PCL scaffold samples, with various architectural arrangements, after 10-days of in vitro culture.(A) Representative micrographs of FDA/PI-stained scaffold samples showing live (green) and dead (red) cells for each of the scaffold groups at D10. Scale bars: 100 µm.(B) Quantification of cell viability assessed via ImageJ [33] analysis of FDA/PI-stained samples (n = 4-6) from each of the architectural scaffold groups at D1 and D10 timepoints, reported as a percentage live to total cells.(C) Representative micrographs, stained with DAPI and Phalloidin to visualise cell nuclei (blue) and F-actin (red) deposition, respectively at D10. Scale bars: 250 µm.(D) Quantification of scaffold F-actin coverage of stained samples (n = 6) from each of the architectural scaffold groups at D10, reported as a percentage (%) F-actin-stained area to total area, analysed via ImageJ.(E) Representative high magnification (40X) images of MSC-cultured samples subsequently seeded with HUVECs for 7 d showing positive expression of CD31 (green).Scale bars: 100 µm.All values reported as mean ± SD (n = 4-6), where symbols (p < 0.001 * * * , 0.001 < p < 0.01 * * , 0.01 < p < 0.05 * ) indicate significance levels, based on ANOVA tests.