Electrospun Polyurethane Vascular Grafts for Cerebral Revascularization: A Pilot Study on Rats

: The current standard technique for vascular grafting in cerebral revascularization surgery employs the interposition of an autologous blood vessel. Technical complications have necessitated the development of a synthetic alternative, but classical biomaterials are not suited for small caliber vascular grafting due to the resulting neointimal hyperplasia and thrombosis. The electro-spinning of polymers is a promising technique for the development of small vascular grafts. The in vivo performance and efficacy of electrospun polyurethane (ePU) grafts with an internal diameter of <1.5 mm have thus far not been evaluated. We developed a novel ePU graft, with a diameter of 1.25 mm, for implantation into the infrarenal aorta of rats. The patency rates of grafts after a 4-month period were equal to those reported in other studies using larger ePU graft diameters and equal or higher than in studies employing other biomaterials. We observed some loss in flow velocity throughout the grafts, which suggests a decreased elasticity of the graft compared to that of the native rat aorta. However, the grafts demonstrated good neo-endothelialization and minimal neointimal hyperplasia. Their porosity promoted cellular infiltration, as observed under tissue slide examination. Our results show that ePU vascular grafts with an internal diameter of <1.5 mm are promising candidates for vascular grafting in cerebral revascularization surgery.


Introduction
Cerebral revascularization is a microsurgical procedure that is used for the augmentation or preservation of blood flow to certain brain regions.Revascularization of brain tissue can be performed either through a bypass from the extracranial circulation to the intracranial circulation (EC-IC bypass) or through the interconnection of two intracranial vessels (IC-IC bypass).One example of the augmentation of blood flow is a bypass between the superficial temporal artery and the medial cerebral artery (STA-MCA) in patients with internal carotid artery (ICA) stenosis.On the other hand, the preservation of blood flow consists of the maintenance of an existing vascular conduit, for instance, after removal of a complex brain aneurysm or a skull base tumor with vascular invasion [1].
Often, a tension free anastomosis cannot be obtained without the interposition of a vascular graft, usually the great saphenous vein (GSV) (inner diameter (ID): 4.9 ± 0.9 mm) BioChem 2024, 4 2 or the radial artery (RA) (ID: 3.55 ± 0.45 mm) [2,3].The choice of a vascular graft depends on the required flow delivery, the size of the recipient vessel, and the availability of the donor vessel [4].The advantages of the GSV include its length and the absence of atherosclerosis or vasospasm.The disadvantages of GSV include its large ID, its lower long-term patency rates (overall 80%) compared to those of RA grafts (overall 90%), as well as thrombogenic valves and varicose veins [5,6].The advantages of the RA include its anatomical similarity to cerebral arteries, the facilitation of surgical anastomosis, and its allowance for greater fluctuations in arterial blood pressure.The disadvantages include its short length and the potential vascular compromising of the hand [7].Additionally, the harvesting of both autologous grafts can be time-consuming and invasive for the patient.Thus, there is a medical need for an effective synthetic alternative for cerebral revascularization.The advantages of a synthetic alternative graft could include a reduction in vasospasm, vasculopathy, flow-mismatch, wound infection of the donor site, and the prevention of vascular compromise or availability issues.
As the IDs of cerebral vessels typically range from 1.07 ± 0.29 mm (cortical segment of anterior inferior cerebellar artery, AICA) to 8.57 ± 1.34 mm (cervical portion of the ICA), synthetic grafts with IDs < 6 mm could offer the greatest therapeutic advantages [8].To date, there is little or no data indicating the effectiveness of small caliber (ID < 6 mm) synthetic cerebral vascular grafts.Nevertheless, great progress has been made in the search for synthetic grafts in cardiovascular surgery.However, the use of classical biomaterials, such as polytetrafluoroethylene (PTFE) and polyethylene terephthalate (PET), has been limited to large caliber (ID ≥ 6 mm) cardiovascular surgery [9,10].Some animal studies using PTFE and PET grafts with ID < 6 mm have reported discouraging patency rates (overall 40% at six months) due to dynamic non-compliance, surface thrombosis, and neointimal hyperplasia (NIH), which become more critical with the gradual reduction of the graft ID [11][12][13][14].
An attempt to increase patency rates in small caliber cardiovascular grafting was recently made by the use of electrospinning, a technique that produces nanoscale fibers through the acceleration of a charged polymer fluid towards a rotating collector [15].From a morphological point of view, the porosity, high surface-to-volume ratio, and surface topography of electrospun grafts mimic the extracellular matrix of native blood vessels.This resemblance promotes tissue reorganization through nutrient transfer, gas exchange, and intercellular communication.In various in vitro and in vivo settings, these processes have been shown to counteract thrombosis and NIH and improve neo-endothelialization (NE) [10,16].
Many in vitro and in vivo animal electrospinning studies, using various electrospun polymeric biomaterials, have demonstrated the superior efficacy of electrospun polyurethane (ePU).ePU demonstrates anti-thrombogenic properties, biocompatibility, and a mechanical strength and elasticity that closely matches the native vessel capacities [12,17,18].Therefore, ePU grafts are promising substitutes for small caliber vascular replacement, and indeed, they are already used to achieve vascular access in hemodialysis patients (ID 6 mm) [10].However, the efficacy of ePU grafts with an ID < 1.5 mm has not been determined to date.To address the important question of whether ePU grafts with an ID < 1.5 mm are clinically effective, we developed a novel ePU graft with an ID of 1.25 mm, and we determined its efficacy after implantation into the infrarenal aorta of rats in a small pilot study.For this study, the infrarenal aorta was used, as it has comparable dimensions and hemodynamics to those of human cerebral vessels, hence rendering our findings clinically relevant [19].

Methods
In a preclinical pilot animal study, in-house generated ePU grafts were implanted into the infrarenal aorta of rats.After a period of four months, our novel ePU graft was assessed for patency, thrombosis, NIH, and NE.The patency and thrombosis results were compared to those of the control group comprised of rats that underwent sham surgery.As NIH is not a natural feature of native rat aorta, and endothelial coverage is typically 100%, the results of NIH and NE in regards to histology were only quantitatively described in the intervention group.The research hypothesis was that the intervention group (ePU graft implantations) would exhibit non-inferiority towards a control group (those undergoing sham surgeries) regarding the incidence of aortic patency and thrombosis.Second, we expected minimal differences in hemodynamic performance between the groups.Third, we expected minimal NIH, as well as confluent coverage of the luminal surface with NE throughout the proximal, mid-portion, and distal graft.

Graft Production: Polyurethane Solution and Electrospinning
Electrospinning of the graft was performed using a novel device designed by the technical workshop (CWFW) of Ghent University-Faculty of Sciences, consisting of three main components: (1) a power supply (Glassman High Voltage, EL50P00, High Bridge, NJ, USA) capable of generating a potential between 0 and 50 kV; (2) a 6 syringe motor driven pumping system (New Era Pump System, Multi-phaser NE1600, Farmingdale, NY, USA); and (3) a grounded mandrel in a rotating unit for the collection of the fiber jet (Figure 1).A liquid 5% (m/m) solution was prepared by dissolving non-biodegradable polyurethane pellets (Carbothane ® , Aliphatic TPU Clear Series, PC3575A, Velox Lubrizol, Hamburg, Germany) in chloroform (Chem-lab, Zedelgem, Belgium) at room temperature (15-25 • C).A 20 mL syringe was filled with the solution and positioned in the automatic dispenser unit.The flow (flow rate: 1 mL/h) was passed through a distally mounted tube with a blunt end 22-gauge needle.The needle tip was positioned vertically at a distance of 8 cm from a 1.0 mm diameter zinc mandrel.The mandrel had been coated with molten (80 • C) polyethylene glycol (PEG, Fluka-Sigma Aldrich, Buchs, Switzerland) and dried for 24 h at room temperature (15-25 • C).The PEG-coated zinc mandrel was mounted on a rotating unit (speed: 275 rotations per minute).After completion of this setup, the voltage source was set at 20 V (direct current), and consequently, the charged polyurethane solution was accelerated from the anode (needle tip) to the cathode (rotating mandrel/background electrode).Electrospinning was performed for one hour.The mandrel, circumferentially covered with a fine ePU mesh (the graft), was subsequently placed in distilled water for 1 h to dissolve the PEG coating, facilitating the detachment of the graft.The collected grafts were dried for two hours at 40 • C in atmospheric pressure to evaporate the residual chloroform.The ePU grafts were sterilized in a 100% ethylene oxide cold cycle process of 38 • C (Sint-Jan Hospital, Bruges, Belgium).Electrospinning.The polyurethane solution was pumped through a tube, distally ending in a needle.At the needle tip, the surface tension of the droplets was overcome by the power of the voltage source.The droplets were accelerated as fibers and collected onto the rotating mandrel, which was attracted by the background electrode.

SEM: ePU Graft Characterization
Before sterilization, a 1 mm piece was cut from one of the ePU grafts and sputtercoated with gold.The specimen was visualized with scanning electron microscopy (SEM) (Fei, Quanta Field Emission Gun 200, Hillsboro, OR, USA), and the images were analyzed using ZEN 2 core software for size calibration (Carl Zeiss AG, Oberkochen, Germany).The gross morphology, surface topography, fiber diameter, pore size, and wall thickness Electrospinning.The polyurethane solution was pumped through a tube, distally ending in a needle.At the needle tip, the surface tension of the droplets was overcome by the power of the voltage source.The droplets were accelerated as fibers and collected onto the rotating mandrel, which was attracted by the background electrode.

SEM: ePU Graft Characterization
Before sterilization, a 1 mm piece was cut from one of the ePU grafts and sputtercoated with gold.The specimen was visualized with scanning electron microscopy (SEM) (Fei, Quanta Field Emission Gun 200, Hillsboro, OR, USA), and the images were analyzed using ZEN 2 core software for size calibration (Carl Zeiss AG, Oberkochen, Germany).The gross morphology, surface topography, fiber diameter, pore size, and wall thickness were assessed on 24 randomly selected regions and expressed as the mean ± SD (in µm).

Animal Characterization and Grouping
For these experiments, 18 male Wistar Han International Genetic Standard (IGS) rats (Rattus norvegicus, Charles River, France) were used, and this study was approved by the Ethical Committee for Animal Experiments at the University of Antwerp (approval number 2015-55).The rats were 9 weeks of age (weighing 276-300 g) at the start of the study and were all preoperatively healthy.All experiments were conducted according to the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health [20], and the experiments were reported according to the ARRIVE guidelines [21].All the rats were housed pairwise in the animal unit of the University of Antwerp.At the start of the study, the rats were randomly divided (using computer assistance) into two groups: the intervention group (n = 9, ePU graft surgery) and the control group (N = 9, sham surgery).

Implantation of ePU Grafts
In the intervention group, an ePU graft (ID: 1.25 mm, length: 7.50 mm) was implanted into the infrarenal aorta using a surgical microscope (Carton SPZ50, Tokyo, Japan).A midline laparotomy was performed, extending from the xiphoid to just above the bladder, and the intestines were laid aside, exposing the retroperitoneal space.After dissection, two temporary aneurysm clamps (Peter Lazic, Tuttlingen, Germany) were placed on the infrarenal aorta, above the iliac bifurcation.The aorta was completely transected halfway between the clamps.The last millimeters of each stump were stripped from the adventitial layer and flushed with sterile saline.The graft was implanted using two end-to-end anastomoses with circumferentially interrupted non-absorbable 10-0 sutures (Ethilon, Somerville, NJ, USA).Before the last suture, the graft lumen was flushed with sterile saline.In the control group the infrarenal aorta was clamped, divided over the superior portion of its circumference, and reanastomosed using the same interrupted sutures.All surgeries were performed under general inhalation anesthesia using isoflurane (5% induction and 2-3% maintenance concentrations, Forene ® Abbott, Chicago, IL, USA) in an O 2 enriched gas mixture.Body temperature was rectally monitored and corrected with an underlying heating pad.Vital signs were monitored using MouseOx software (Starr Life Sciences, Oakmont, PA, USA).No anticoagulant or antiplatelet drugs were used.Subcutaneous buprenorphine (0.1 mg/mL) (Vetergesic ® Ecuphar, Oostkamp, Belgium) was administered as an analgesic premedication.Postoperatively, the following products were administered: epicutaneous chlorhexidine chloride (Astrexine ® Pierre Fabre, Anderlecht, Belgium), epicutaneous lidocaine hydrochloride 2% (Xylocaine ® AstraZeneca, Ukkel, Belgium), subcutaneous buprenorphine 0.1 mg/mL (Vetergesic ® Ecuphar, Oostkamp, Belgium), and subcutaneous amoxicillin 150 mg/mL (Duphamox LA ® Zoetis, Zaventem, Belgium).

In Vivo Assessments over a 4-Month Period: Thrombosis
Aortic thrombosis in the intervention and control group was clinically assessed over a 4-month period.The following parameters were used for the evaluation of thrombosis of the lower limb: skin/paw pad temperature (cold) and color (cyanosis/pallor), ulceration, paresis/paralysis, and limb pain (biting or shaking limb).Intestinal thrombosis was evaluated by the assessment of abdominal tenderness during handling, a loss of appetite, and alterations in fecal matter and production (blood, constipation) [22].

In Vivo Assessments after a 4-Month Period: Patency
In vivo evaluation of patency in the intervention and control group was performed after a 4-month period using an ultrasound system (Visualsonics, Vevo ® 2100, Amsterdam, The Netherlands).This system operated at a high frequency (13-75 MHz), with a spatial resolution of ~30 microns and a temporal resolution of 1000 frames/sec.Ultrasound imaging on the rats was performed under general anesthesia using sevoflurane (5% induction and 2% maintenance, Sevorane ® Abbvie, Wavre, Belgium) in an O 2 enriched gas mixture.The abdomen was shaved, and 37 • C ultrasound gel was applied.The probe was levelled perpendicular to the spine.Using a pulsed-wave 40 MHz Doppler in the coronal plane, patency was categorically evaluated by the presence (patent) or absence (non-patent) of pulsatile arterial flow in the aorta segment distal to the graft/control aorta.Absolute maximal flow velocity (Vmax) and Vmax decay from the proximal to distal graft/control aorta were measured in the sagittal plane and expressed as the mean ± SD (in mm/s).The inner/outer and proximal/distal diameters and the systolic/diastolic diameters of the graft/control aortas (IDps, IDpd, IDds, IDdd, ODps, ODpd, Odds, and ODpd) were measured using the motion-mode (in µm) and compared between the groups.The images were analyzed using ImageJ software (ImageJ, National Institutes of Health).The systolic expansion was calculated using the formula [IDps − IDpd] for the proximal segment and [IDds − IDdd] for the distal segment.The proximal to distal decay in the graft/control aorta expansion [(IDps-IDpd) − (IDds-IDdd)] was calculated and compared between groups.

Retrieval Surgery after 4-Month Period
After ultrasound imaging, retrieval surgery of the ePU grafts/control aortas was performed.The same procedure, medication, and anesthetic products as those described for the implantation were used (see above).In between two temporary clamps, either the graft (intervention group) or a 7.5 mm piece of reanastomosed control aorta (control group) was retrieved.All rats were subsequently sacrificed with an intracardiac injection of 1.5 mL sodium pentobarbital 60 mg/mL (Nembutal ® Ceva, Libourne, France).

Histological Assessments of Retrieved ePU Grafts: NIH and NE
The retrieved ePU grafts were embedded in 4% formaldehyde for 24 h and were subsequently placed in 60% isopropyl alcohol.After dehydration and tissue preparation using an automated device (Excelsior, Thermo Fisher Scientific, Breda, The Netherlands), the samples were fixed using an automated paraffin embedding device (TES 99, Medite, Burgdorf, Germany).Two longitudinal sections of 5 µm thickness were cut from the samples.A standard light microscope (Olympus BX41) with imaging software (Leica Application Suite X, Leica, Wetzlar, Germany) was used for visualization.One slide from each rat was stained with hematoxylin and eosin (H&E), for the evaluation of NIH in the ePU grafts.NIH was quantified as the thickness (in µm) of the area between the endothelial layer and luminal graft surface.It was measured in three different regions: (1) the first anastomosis, (2) mid-graft, and (3) the second anastomosis.For each region, two measurements were circumferentially performed, 180 degrees apart from each other.NIH was expressed as the mean ± SD for each region.The length of NIH (in mm) was the sum of all longitudinal NIH regions, measured as close as possible to the endothelial layer, in a longitudinal manner between both ends of the graft, superiorly and inferiorly (Figure 2).The length was compared to the total length (in mm) of the ePU graft and expressed as a percentage (length/length %).The second slide was stained with anti-Factor VIII (anti-von Willebrand Factor, Dako, Glostrup, Denmark) antibody for the identification of endothelial cells.NE was quantified as anti-F VIII positive length (in mm), compared to the total length (in mm) of the ePU graft and expressed as a percentage (length/length %).
gitudinal manner between both ends of the graft, superiorly and inferiorly (Figure 2).The length was compared to the total length (in mm) of the ePU graft and expressed as a percentage (length/length %).The second slide was stained with anti-Factor VIII (anti-von Willebrand Factor, Dako, Glostrup, Denmark) antibody for the identification of endothelial cells.NE was quantified as anti-F VIII positive length (in mm), compared to the total length (in mm) of the ePU graft and expressed as a percentage (length/length %).

Statistical Analyses
Statistical analyses were conducted using SPSS software (Statistical Package for the Social Science version 24, IBM analytics).No statistical analyses were performed to compare the incidence of aortic patency and thrombosis in the intervention group to that of the control group, and these parameters were only quantitatively described.Ultrasound data on hemodynamic performance were analyzed using a one-way ANOVA and Student's t-test to compare the mean values among single groups and between the two groups, respectively.the histological data of NIH were analyzed using a one-way ANOVA test to compare the mean values in the intervention group.NIH and NE were quantitatively described.The assumptions of the tests were confirmed by the data.The significance level was set at p < 0.05.

Pre-Implant Characterization: SEM
The features of one ePU graft were analyzed using SEM imaging.In Figure 3, the tubular gross morphology and porous surface topography of the ePU graft are presented.The fibers were randomly oriented, and small irregularly distributed beads were observed on the course of some fibers.The mean fiber diameter of the ePU graft was 2.02 ± 1.43 µm (minimum 0.30 µm, maximum 5.17 µm).The graft exhibited a mean pore size of 5.65 ± 2.30 µm (minimum 2.33 µm, maximum 11.66 µm) and a wall thickness of 223 ± 18 µm (minimum 189 µm, maximum 260 µm).The presence of PEG-residue was not evaluated because the sample size was too small.

Statistical Analyses
Statistical analyses were conducted using SPSS software (Statistical Package for the Social Science version 24, IBM analytics).No statistical analyses were performed to compare the incidence of aortic patency and thrombosis in the intervention group to that of the control group, and these parameters were only quantitatively described.Ultrasound data on hemodynamic performance were analyzed using a one-way ANOVA and Student's t-test to compare the mean values among single groups and between the two groups, respectively.the histological data of NIH were analyzed using a one-way ANOVA test to compare the mean values in the intervention group.NIH and NE were quantitatively described.The assumptions of the tests were confirmed by the data.The significance level was set at p < 0.05.

Pre-Implant Characterization: SEM
The features of one ePU graft were analyzed using SEM imaging.In Figure 3, the tubular gross morphology and porous surface topography of the ePU graft are presented.The fibers were randomly oriented, and small irregularly distributed beads were observed on the course of some fibers.The mean fiber diameter of the ePU graft was 2.02 ± 1.43 µm (minimum 0.30 µm, maximum 5.17 µm).The graft exhibited a mean pore size of 5.65 ± 2.30 µm (minimum 2.33 µm, maximum 11.66 µm) and a wall thickness of 223 ± 18 µm (minimum 189 µm, maximum 260 µm).The presence of PEG-residue was not evaluated because the sample size was too small.

Implantation of ePU Grafts
All rats were preoperatively healthy, as evaluated using standardized welfare control sheets.Throughout the experiments, a total of four rats from the intervention group were excluded from the study.In two of them, surgery was prematurely terminated without interposition of a graft, due to (1) technical problems with the microscope and (2) the inability to reach sufficient levels of anesthesia.The other two rats died because of (3) major intraoperative bleeding due to accidental injury of the vena cava and (4) anesthetic intolerance.These four subjects were replaced by four unused Wistar Han IGS rats from the same lot.In the intervention and control group, all graft/control aorta anastomoses were tension-free, and no major complications occurred.When releasing the vessel clamps after the procedure, an optical check of all graft/control aortas showed good expansile pulsations.The grafts became red and saturated with blood, indicating permeability to the erythrocytes, but without any leakage at the anastomoses (Figure 4).Small anastomototic leakage was terminated using a hemostatic patch (Surgicel ® , Ethicon, Johnson&Johnson,

Implantation of ePU Grafts
All rats were preoperatively healthy, as evaluated using standardized welfare control sheets.Throughout the experiments, a total of four rats from the intervention group were excluded from the study.In two of them, surgery was prematurely terminated without interposition of a graft, due to (1) technical problems with the microscope and (2) the inability to reach sufficient levels of anesthesia.The other two rats died because of (3) major intraoperative bleeding due to accidental injury of the vena cava and (4) anesthetic intolerance.These four subjects were replaced by four unused Wistar Han IGS rats from the same lot.In the intervention and control group, all graft/control aorta anastomoses were BioChem 2024, 4 7 tension-free, and no major complications occurred.When releasing the vessel clamps after the procedure, an optical check of all graft/control aortas showed good expansile pulsations.The grafts became red and saturated with blood, indicating permeability to the erythrocytes, but without any leakage at the anastomoses (Figure 4).Small anastomototic leakage was terminated using a hemostatic patch (Surgicel ® , Ethicon, Johnson&Johnson, Sommerville, NJ, USA).The immediate postoperative outcome for all rats was clinically favorable.
excluded from the study.In two of them, surgery was prematurely terminated without interposition of a graft, due to (1) technical problems with the microscope and (2) the inability to reach sufficient levels of anesthesia.The other two rats died because of (3) major intraoperative bleeding due to accidental injury of the vena cava and (4) anesthetic intolerance.These four subjects were replaced by four unused Wistar Han IGS rats from the same lot.In the intervention and control group, all graft/control aorta anastomoses were tension-free, and no major complications occurred.When releasing the vessel clamps after the procedure, an optical check of all graft/control aortas showed good expansile pulsations.The grafts became red and saturated with blood, indicating permeability to the erythrocytes, but without any leakage at the anastomoses (Figure 4).Small anastomototic leakage was terminated using a hemostatic patch (Surgicel ® , Ethicon, Johnson&Johnson, Sommerville, NJ, USA).The immediate postoperative outcome for all rats was clinically favorable.

In Vivo Assessments: Clinical Parameters of Thrombosis
On the first postoperative day, one rat out of nine in the intervention group developed a paraparesis of the lower limbs, suggesting graft thrombosis.Humane endpoints were met, and the rat was euthanized.Postmortem ex vivo examination of the graft could not determine the cause of thrombosis.However, subsequent histological evaluations with H&E staining demonstrated a large thrombus, occluding the graft lumen.Thrombosis was likely caused by a surgical complication (see: Section 4).During the 4-month study-period, no other clinical parameters of aortic thrombosis were observed, either in the intervention or in the control group.The paraplegic rat was considered to exhibit graft thrombosis, bringing the overall thrombosis incidence of the intervention group to 1/9 (11%) compared to 0/9 (0%) in the control group.From a clinical surgical point of view, as further described in Section 4, we found this difference in thrombosis incidence between the intervention and the control group to be insignificant.

In Vivo Assessments: Patency on Ultrasound
After a 4-month period, a pulsed-wave Doppler ultrasound was used to evaluate the patency of the ePU graft/control aortas.A pulsatile arterial flow was noted in all aorta segments located distally from the graft/control aortas.No stenotic processes were detected over the course of the graft/control aortas, and there was no evidence of ePU graft-related complications, such as graft rupture or aneurysm formation.The rat with paraparesis was clinically determined to show graft failure/non-patent, bringing the overall patency of the intervention group to 8/9 (89%) compared to 9/9 (100%) in the control group.From a clinical surgical point of view, we considered this as a nonsignificant difference in patency between the intervention and the control groups.The ultrasound imaging of the ePU grafts is shown in Figure 5.
related complications, such as graft rupture or aneurysm formation.The rat with paraparesis was clinically determined to show graft failure/non-patent, bringing the overall patency of the intervention group to 8/9 (89%) compared to 9/9 (100%) in the control group.From a clinical surgical point of view, we considered this as a nonsignificant difference in patency between the intervention and the control groups.The ultrasound imaging of the ePU grafts is shown in Figure 5.There was no statistically significant difference between the intervention and control group for the absolute Vmax in the proximal aorta (p = 0.479) and the absolute Vmax in the distal aorta (p = 0.711).However, there was a statistically significant difference (p = 0.010) in Vmax decay (from proximal to distal) between the intervention group (mean = 49 ± 40 mm/s) and the control group (mean = −14 ± 39 mm/s).These data are shown in Table 1.
Table 1.Aortic diameter and diastolic-systolic expansion with motion-mode, aortic Vmax with pulsed-wave Doppler, and the intervention and control group.The aortic systolic expansion, the aortic diastolic expansion, and the aortic expansion decay (proximal minus distal) were not significantly different between groups.There was a statistically significant difference in Vmax decay between the intervention group (49 ± 40 mm/s) and the control group (−14 ± 39 mm/s).Vmax in the mid graft was lower compared to that of the proximal and distal aorta.There was no statistically significant difference (p = 0.280) in proximal aortic systolic expansion between the grafts (mean = 123 ± 49 µm) and the shams (mean = 157 ± 75 µm).Neither was there a significant difference (p = 0.827) in distal aortic systolic expansion between the intervention group (mean = 104 ± 64 µm) and the control group (mean = 114 ± 102 µm).There was also no statistically significant (p = 0.353) difference between the intervention and control group in regards to the aortic expansion decay between the proximal and distal aorta [(IDps-IDpd) − (IDds-IDdd)].However, the means showed a relevant difference between the intervention group (mean = 52 ± 110 µm) and the control group (mean = 9 ± 75 µm).These data are summarized in Table 1.

Retrieval Surgery after a 4-Month Period
Immediately after ultrasound imaging, surgery to retrieve the ePU grafts/control aortas was performed.In vivo evaluation demonstrated good healing of the anastomoses and good expansile pulsations in the grafts/control aortas.There were no stenotic adhesions to the surrounding retroperitoneal tissue.

Histological Assessment: NIH and NE
After a 4-month period, histological analyses were performed for the evaluation of NIH and NE in the intervention group.As expected, on the first postoperative day, it was too soon to identify NIH or NE on the slides from the euthanized rat with paraparesis.Histological evaluation in this rat using H&E staining demonstrated a large thrombus, occluding the graft lumen.Therefore, only 8 out of 9 subjects were included in the histological analyses.The porous fiber mesh of the ePU graft was clearly visualized using H&E staining, and it was apparent that there was an ingrowth of different cells in the graft.NIH in the intervention group was measured on the H&E stained slides of the ePU grafts.The mean neointimal thickness of the mid graft (35 ± 40 µm) was similar (p = 0.978) to that near the anastomoses (35 ± 43 µm).In 1 of 8 ePU grafts, no NIH could be observed (0 µm).The mean percentage of the longitudinal length of the graft covered with NIH was 54 ± 29% (minimum 0%, maximum 94%).The absolute mean length of the graft covered with NIH was 6.02 ± 4.16 mm for a mean total graft length of 11.19 ± 2.48 mm.These data are summarized in Table 2.An example of NIH is shown in Figure 6.lumen were nearly confluently positive for anti-FVIII staining.These data are summarized in Table 3.An example of NE is shown in Figure 6.

Discussion
The current standard for small caliber (<6 mm ID) vascular grafting in cerebral revascularization surgery is the interposition of an autologous graft, such as the GSV or the RA [2,4].Despite the histological tissue supremacy of these autologous substitutes, technical NE in the intervention group was measured on the anti-FVIII stained sections of the ePU grafts.NE over the course of the luminal graft length was 80 ± 40%.The mean NE length was 8.95 ± 4.64 mm, covering a mean graft length of 11.19 ± 2.48 mm.One out of eight grafts did not exhibit NE.The proximal, mid-portion, and distal regions of the graft BioChem 2024, 4 11 lumen were nearly confluently positive for anti-FVIII staining.These data are summarized in Table 3.An example of NE is shown in Figure 6.Notes: NE: neo-endothelialization; SD: standard deviation.

Discussion
The current standard for small caliber (<6 mm ID) vascular grafting in cerebral revascularization surgery is the interposition of an autologous graft, such as the GSV or the RA [2,4].Despite the histological tissue supremacy of these autologous substitutes, technical issues and surgical complications necessitate the development of a synthetic alternative [5][6][7].
Yet, due to NIH and thrombosis, the use of synthetic vascular grafts using classical biomaterials such as PTFE and has mainly been restricted to large caliber vascular surgery [9,10,[12][13][14].In cardiovascular research, electrospinning is a promising technique for the development of small caliber (<6 mm ID) vascular grafts.Several reports have suggested superior in vitro and in vivo animal performance of ePU when compared to that of other organic polymers in terms of patency, absence of thrombosis, NIH, and NE [23][24][25][26][27].To date, no studies have been performed evaluating the in vivo use of ePU grafts with an ID < 1.5 mm.Considering the introduction of synthetic grafts in cerebral revascularization surgery, we developed a novel ePU graft with an ID of 1.25 mm.Detailed feature characterization was performed using SEM.The ePU graft was implanted in the infrarenal aorta of rats and evaluated for patency, thrombosis, NIH, and NE after a 4-month period.The results were compared to those of the control group who underwent sham surgery.We first hypothesized that the intervention group would exhibit noninferiority towards the control group regarding aortic patency and thrombosis.Secondly, we expected minimal differences in hemodynamic performance via ultrasound.Thirdly, upon histological examination, we expected minimal NIH and confluent NE throughout the regions of the ePU graft.
By means of SEM imaging after electrospinning, an ePU graft with small fibers of 2.0 ± 1.43 µm was observed.The pore diameter was 5.65 ± 2.30 µm, large enough for cellular infiltration, as demonstrated by the ingrowth of different cells upon tissue slide examination.The wall thickness was 223 ± 18 µm, which was similar to the wall thickness of the GSV (250-370 µm), the current standard for small caliber vascular replacement surgery [28,29].Some grafts showed an irregular topography, with beads in the fibers.These distortions suggest the presence of electrospraying during the electrospinning process, which is the breakup of low-viscosity solutions into droplets due to surface tension effects.This can be corrected by lowering the concentration of the polyurethane solution or by altering the solvent [30].Our study was an initial pilot study to assess the efficacy of our novel ePU graft.In the future, larger animal studies will be required to fine-tune more of the appropriate settings of the electrospinning process (e.g., flow rate, rotation speed, needle diameter, voltage) and to investigate their potential clinical repercussions.Furthermore, the fiber diameter was variable (2.02 ± 1.43 µm).This could be caused by temporary decreases in voltage, the interruption of the electrospinning process, or by the previously mentioned electrospraying [30,31].A regular fiber diameter of <1 µm may have been preferable, as an in vitro study by Milleret et al. demonstrated less coagulation and platelet adherence with fibers <1 µm compared to that noted using a 2-3 µm size; however, they used in vitro ePU disks instead of in vivo tubular grafts [32].The irregular fiber orientation can also be ameliorated by adjusting the rotation speed of the mandrel [24].
The implantation of the ePU was successful, as demonstrated by a favorable postoperative outcome in 8 out of 9 rats after a 4-month period.One rat was euthanized on the first postoperative day due to suspicion of graft thrombosis, which was later confirmed by histological examination.To determine the cause of the thrombotic event, the postoperative reports were examined, and these showed that temporary ligation of an undefined aortic branch (either iliolumbar or genital) close to the anastomosis region, as well as the consequent excess manipulation of the endothelium, may have led to thrombosis.Secondly, a hemostatic patch (Surgicel ® , Ethicon, Johnson&Johnson, Sommerville, NJ, USA) was used to reduce anastomosis bleeding after clamp release.The thrombogenic hemostatic fibers may have entered the graft lumen.A third possible explanation is the early graft failure on the first postoperative day, elicited by foreign body thrombosis.However, this was unlikely, as no inflammatory infiltrates were observed during histological analyses.Furthermore, immunogenic reactions are not frequently found with ePU (see introduction).Fourth, thrombosis due to NIH or insufficient NE is also unlikely because NIH in synthetic grafts only develops after two weeks and typically peaks after one month [33].Consequently, it was too soon to observe NIH or NE during histological analyses.
Apart from this single observed case of thrombosis, neither group demonstrated clinical signs of graft/control aortic thrombosis, bringing the overall thrombosis rate of the intervention group to 1/9 (11%) compared to 0/9 (0%) in the control group, which is not significantly different.Thus, we conclude (taking the small sample size of our pilot study into account, along with a single case of thrombosis due to technical reasons) that the incidence of thrombosis in the 1.25 mm ePU graft was comparable to the incidence of thrombosis in the native aorta after sham surgery.
During ultrasound imaging, using pulsatile arterial flow in the distal aortic segment as a criterion, there was a good patency of the ePU grafts after a period of 4 months.Taking into account the aforementioned single case of thrombosis and its presumed technical cause, the intervention group held a patency of 8/9 (89%) compared to 9/9 (100%) in the control group, which was not clinically significantly different.Therefore, the patency of the ePU grafts was comparable to the patency of the native aorta after sham surgery.A recent study by Bergmeister et al. reported patency rates of 95% 6 months after implantation of 1.50 mm ID ePU grafts into the infrarenal aorta of rats [34].A shortcoming in this study was the lack of a control group; however, the patency rates were good and were comparable to those obtained in our study (89% vs. 95%).There are no other published in vivo studies utilizing chemically unblended ePU grafts with smaller diameters (<1.25 mm ID).Nieponice et al. investigated the in vivo performance of 1.20 mm ID ePU grafts seeded with muscle-derived stem cells in the infrarenal aortas of rats, with patency rates of 65% after 2 months, which was lower than the 89% observed in our study [35].
Via ultrasound, there was a statistically significant (p = 0.010) difference in Vmax decay throughout the graft between the intervention group and the control group.The difference in distal aortic systolic expansion was present but was not statistically significant (p = 0.353) between groups; however, the small sample size of our pilot study may have been a source of underestimation of the effect.The loss of velocity and the limited loss of aortic systolic expansion over the course of the ePU graft suggest a loss of energy of the blood because of a higher resistance of the graft compared to that of the native aorta.A first potential cause of energy loss is graft stenosis, although this was unlikely because a stenotic segment would exhibit an increased flow velocity.The mean Vmax in the mid-graft was lower than at the proximal and distal end of the graft.Leakage of the graft may have been a second potential cause; however, in vivo evaluation during retrieval revealed a watertight anastomosis.Third, and more likely, is a lower elasticity of the graft compared to that of the native aorta, inducing an increase in resistance and loss of kinetic energy.The differences in absolute Vmax throughout the graft were not significant between the intervention and control group and neither were the proximal and distal systolic expansion rates.
Regardless of the fact that there was only one statistically significant hemodynamic difference between the groups (Vmax decay), the abovementioned findings suggest a slight mechanical incompatibility of the ePU graft in vivo.This may be a possible factor to take into account for future optimizations of the hemodynamic capacities of the graft.H&E histological analyses after the 4-month study period did not demonstrate any significant or stenotic NIH in the ePU grafts (35 ± 42 µm, 54 ± 29% of the graft length); compared to those noted in other studies in the literature, and in 1/8 grafts, no NIH was apparent.NIH is a consequence of shear stress caused by turbulent blood flow and, as mentioned previously, it normally only reaches its maximum after 1 month [36].Thus, because there was only minimal NIH after a 4-month period, it was unlikely that there would be any further increases over a longer period of time.Interestingly, in a recent report from Bergmeister et al., it was demonstrated that there were only 2/20 ePU grafts with visible NIH after 3 months (mid graft: mean 198 ± 73 µm, anastomoses: mean 159 ± 71 µm).There are two potential reasons for the smaller NIH rates found in the Bergmeister et al. study compared to those observed in our study: (1) size-mismatch; the ID of the infrarenal aorta of the rat ranged from 1 to 2 mm [19].Size-mismatch in our study may have led to shear stress conditions and subsequent NIH.However, as is evident from Figure 4, there was no size mismatch in our study [37].(2) NIH may have been caused by incomplete endothelialization, which is more likely, as described below [34].
Immunohistochemistry staining with anti-FVIII for NE demonstrated 80 ± 40% luminal coverage of the ePU graft length (mean NE length 8.95 ± 4.64 mm, covering a mean graft length of 11.19 ± 2.48 mm).This was probably an underestimation due to cutting artefacts with the microtome.The endothelial coverage is an indicator that the porous structure of the ePU graft facilitates the migration and engraftment of endothelial cells, protecting the graft from shear stress and thrombosis.Given the very low incidence of thrombotic events in our study, NE was likely sufficiently protective.The promotion of NE in ePU grafts has been demonstrated in previous studies by Uttayarat et [18,23,38,39].NE was demonstrated in vivo after a 1-month study period by Bergmeister et al. (ID 1.50 mm, infrarenal aorta of rats) [34].The added value of our study is the clear evidence of similar NE in ePU grafts with ID 1.25 mm.
There are some potential limitations to our study.Firstly, there was a small sample size (n = 18), with possible over-or underestimation of the results.Secondly, the graft had a short length (7.5 mm).In future studies, it would be interesting to investigate the ePU graft on longer trajectories because lengths starting from 2 cm provide more challenges for endothelialization.The infrarenal aorta of the Wistar rat is not long enough, so a femoral artery crossover graft could be used instead, or alternatively, larger animals, such as rabbits, sheep, or pigs, which exhibit less spontaneous endothelialization and have a greater tendency for hypercoagulability, could also be used [19].
Despite the good hemodynamic resemblance between rat aortas and cerebral blood vessels, the extrapolation of our results to humans should be performed with caution.Rats have a stronger fibrinolytic system and increased spontaneous endothelialization; subsequently, there is a potential underestimation of thrombosis incidence [19].To date, no clinical studies have been performed using ePU grafts, and studies with larger sample sizes in different (larger) animal models are needed to further confirm the good in vivo performance of ePU grafts with ID < 1.50 mm.Animal studies thus far have failed to translate into clinical trials due to insufficient in vivo mechanical testing and regulatory issues [19,50].
The ePU graft possesses several good characteristics of an arterial prosthesis (as described by Abbott et al.) [53].We, along with others, would like to reiterate the necessity for more and larger long-term studies on various animal models to determine the efficacy of small caliber ePU grafts [18].This could lead to the development of more clinically efficacious graft substitutes for use in small caliber vascular surgery, such as cerebral revascularization.

Conclusions
Our results clearly demonstrate that ePU grafts are potentially efficacious synthetic alternatives for autologous vascular grafts.We have shown that, after fine-tuning the electrospinning process, ePU vascular grafts are also potentially good substitutes for use in small caliber vascular surgery with IDs < 1.50 mm, e.g., cerebral revascularization.
The graft can be sterilized, is not expensive, is easy to suture, can be easily produce in various IDs, and shows durability.

BioChem 2023, 3 , 4 Figure 1 .
Figure1.Electrospinning.The polyurethane solution was pumped through a tube, distally ending in a needle.At the needle tip, the surface tension of the droplets was overcome by the power of the voltage source.The droplets were accelerated as fibers and collected onto the rotating mandrel, which was attracted by the background electrode.

Figure 1 .
Figure1.Electrospinning.The polyurethane solution was pumped through a tube, distally ending in a needle.At the needle tip, the surface tension of the droplets was overcome by the power of the voltage source.The droplets were accelerated as fibers and collected onto the rotating mandrel, which was attracted by the background electrode.

Figure 2 .
Figure 2. Measurement of NIH on histological slides.Graphical representation of a graft.NIH = sum of the grey parts in the superior and inferior lumen.

Figure 2 .
Figure 2. Measurement of NIH on histological slides.Graphical representation of a graft.NIH = sum of the grey parts in the superior and inferior lumen.

BioChem 2023, 3 ,Figure 3 .
Figure 3. SEM images.Left (24×): the gross morphology demonstrates a regularly shaped tubular graft with an ID of 1.25 mm.Middle (490×): smooth inner luminal curvature, with occasional small irregularly distributed beads on the course of some fibers.Right (465×): surface topography showing a nanofiber porous mesh of electrospun polyurethane.

Figure 3 .
Figure 3. SEM images.Left (24×): the gross morphology demonstrates a regularly shaped tubular graft with an ID of 1.25 mm.Middle (490×): smooth inner luminal curvature, with occasional small irregularly distributed beads on the course of some fibers.Right (465×): surface topography showing a nanofiber porous mesh of electrospun polyurethane.

Figure 4 .
Figure 4. Implantation surgery.(Left): intervention group, end-to-end anastomosis of a 1.25 mm ID ePU graft to the infrarenal aorta.(Right): control group, end-to-end anastomosis of partially transected infrarenal aorta.Images captured by a camera through the microscope lens.

Figure 5 .
Figure 5. Ultrasound imaging of an ePU graft.Velocity of aortic blood flow was measured using pulsed-wave Doppler (left), and aortic pulsatile expansion was measured using motion-mode (right).

Figure 6 .
Figure 6.Upper row: neointimal hyperplasia of the graft via H&E staining.(A) (4×): the overview shows various neointimal thicknesses along the six regions of the graft (asterisk).No stenotic processes were observed; L = luminal side.(B) (20×): the mid-portion of the graft shows various thin cell layers of NIH (double-sided arrow).(C) (20×): anastomosis site; confluent transition between aorta (white asterisk) and the ePU graft (black asterisk).Lower row: neo-endothelialization of the graft via Anti-FVIII immunohistochemistry. (D) (4×): the overview shows complete endothelialization of the graft.(E) (20×): the mid-portion of the graft shows confluent endothelialization (black arrows) and ingrowth of cells in the porous cavities (blue-stained cells).The staining on the outer side of the graft is an artifact due to fibrous staining.(F) (20×): anastomosis site; transition between aorta and graft.

Figure 6 .
Figure 6.Upper row: neointimal hyperplasia of the graft via H&E staining.(A) (4×): the overview shows various neointimal thicknesses along the six regions of the graft (asterisk).No stenotic processes were observed; L = luminal side.(B) (20×): the mid-portion of the graft shows various thin cell layers of NIH (double-sided arrow).(C) (20×): anastomosis site; confluent transition between aorta (white asterisk) and the ePU graft (black asterisk).Lower row: neo-endothelialization of the graft via Anti-FVIII immunohistochemistry. (D) (4×): the overview shows complete endothelialization of the graft.(E) (20×): the mid-portion of the graft shows confluent endothelialization (black arrows) and ingrowth of cells in the porous cavities (blue-stained cells).The staining on the outer side of the graft is an artifact due to fibrous staining.(F) (20×): anastomosis site; transition between aorta and graft.

Table 2 . NIH on H&E staining in the intervention group.
The mean NIH was approximately equal in all regions (± 35 µm) and covered ± 54% of the graft length.A total of 1/8 of the grafts showed no NIH.
Notes: a graft length is the sum of the superior and inferior luminal length; b NIH 1/2/3/4 anastomosis: NIH measured in four different regions near the proximal and distal anastomosis; c NIH 5/6 mid graft: NIH measured in two different regions of the mid graft; d significance of difference: p = 0.978; NIH: neointima hyperplasia; SD: standard deviation.BioChem 2023, 3, FOR PEER REVIEW 11

Table 3 . NE on anti-FVIII immunohistochemistry of the intervention group
. NE was 80 ± 40% over the course of the graft length.A total of 1/8 of the grafts showed no NE.

Table 3 . NE on anti-FVIII immunohistochemistry of the intervention group
. NE was 80 ± 40% over the course of the graft length.A total of 1/8 of the grafts showed no NE.