Functional remodeling of an electrospun polydimethylsiloxane-based polyether urethane external vein graft support device in an ovine model

Abstract

Saphenous vein graft (SVG) failure rates are unacceptably high, and external mechanical support may improve patency. We studied the histologic remodeling of a conformal, electrospun, polydimethylsiloxane-based polyether urethane external support device for SVGs and evaluated graft structural evolution in adult sheep to 2 years. All sheep (N=19) survived to their intended timepoints, and angiography showed device-treated SVG geometric stability over time (30, 90, 180, 365, or 730 days), with an aggregated graft patency rate of 92%. There was minimal inflammation associated with the device material at all timepoints. By 180 days, treated SVG remodeling was characterized by minimal/non-progressive intimal hyperplasia; polymer fragmentation and integration; as well as the development of a neointima, and a confluent endothelium. By 1-year, the graft developed a media-like layer by remodeling the neointima, and elastic fibers formed well-defined structures that subtended the neo-medial layer of the remodeled SVG. Immunohistochemistry showed that this neo-media was populated with smooth muscle cells, and the intima was lined with endothelial cells. This data suggests that treated SVGs were structurally remodeled by 180 days, and developed arterial-like features by 1 year, which continued to mature to 2 years. Device-treated SVGs remodeled into arterial-like conduits with stable long-term performance as arterial grafts in adult sheep.

1. Introduction

With the exception of bypassing the left anterior descending coronary artery using the left internal mammary artery, autologous saphenous vein grafts (SVGs) are used for the majority of coronary and peripheral arterial bypass grafts [1]. However, since veins typically function in low-pressure hemodynamic conditions, arterial pressure can cause SVGs to distend to their maximum diameter and subsequently undergo dilation, yielding irregular lumen geometries. Irregular lumen geometry results in disturbances to hemodynamic factors, such as wall shear stress (WSS), spatial and temporal WSS gradients, and the oscillatory shear index (OSI), that contribute to intimal hyperplasia (IH) [2–4]. Arterial pressure also exposes SVGs to abnormally high levels of wall tension, exacerbating the effects of hemodynamic disturbances [5–7]. Reported SVG failure rates in coronary artery bypass graft surgery are 10%–30% at one year and approximately 50% at 10 years [8–10]. There is also a 61% primary patency rate for peripheral arterial SVGs at 1 year, and 50–70% at 5 years [11,12].

Mechanical restriction of SVGs to prevent distension under arterial pressure was first suggested in 1963 [13]. Since then, external stenting of SVGs using several different approaches, in both preclinical and clinical studies, has shown promise in attenuating IH [1,14–17]. These studies have suggested that the beneficial effect of external stenting is due to a reduction in wall tension, improved uniform lumen geometry, structural evolution and enhanced performance [1]. However, a limitation of previous SVG external stenting approaches has been their reliance on pre-fabricated tubular structures that are not conformal to the adventitial surface irregularities of SVGs [1]. During SVG harvesting, tributary ligation results in outward radial protrusions of the suture- or clip-ligated stumps. Surface irregularities may also arise if surrounding tissue is not carefully removed. Pre-fabricated external stents cause these inevitable surface irregularities to protrude into the SVG lumen and likely result in geometric and hemodynamic abnormalities, which are counterproductive to the intended benefit of external stenting. As a component of the development of a novel approach to providing external support for an SVG, we hypothesized that a conformal, external structural support composed of a polydimethylsiloxane-based polyether urethane (PDMS-PEU) fiber matrix, deposited onto the SVG by electrospinning, would induce progressive functional remodeling to an arterialized wall structure and mitigate IH [18].

Thus, the purpose of the study was to understand the evolution of structure of an SVG supported by an electrospun PDMS-PEU biomaterial, which was applied immediately prior to implantation in an ovine peripheral arterial bypass model [19]. The study was also intended to provide data to inform future development of this approach to external SVG support.

2. Materials and Methods

2.1. Device Manufacturing

A commercially available PDMS-PEU, Elast-Eon^™ E2–852 (Aortech Biomaterials, Weybridge, UK; Molecular weight approximately 100, 000 g/mol), was used to manufacture the electrospun polymer structure and embedded strain-relieving spine that represent the two components of the device. The polymer structures were electrospun from sterile solutions (31% w/v) of polymer dissolved in hexafluoroisopropanol (HFIP; DuPont, Wilmington, DE). Devices were deployed onto SVGs, which had been cannulated by a sterile, stainless-steel, rotating cylindrical mandrel, or onto bare cylindrical mandrels. The stainless-steel mandrel provided a negatively charged target for the electrospun polymer, which became positively charged as it passed through a nozzle. During the process, the nozzle was translated back and forth along the length of the rotating target. This electrospinning process was performed using a fully automated, custom-built system capable of maintaining SVG sterility within an operating room environment. The nominal electrospinning parameters used to manufacture the polymer structures were: Polymer solution flowrate = 15 mL/hour; nozzle voltage = +17 kV; mandrel voltage = −2 kV; nozzle to mandrel distance = 125 mm; nozzle translational speed = 180 mm/second; mandrel rotational speed = 250 rpm; with a typical process time = 12 minutes. The thickness of the resulting polymer structure was approximately 300 μm. The electrospinning processing time required to deposit the desired polymer thickness, for a range of vein outer diameters, was empirically determined, and used to program the electrospinning system. Prior to initiating the electrospinning process, a strain-relieving spine intended to provide kink resistance to the device was applied onto the outer surface of the target. Spines were manufactured by winding a segment of extruded polymer (0.5 mm diameter extruded filament of the same polymer used for electrospinning) onto a custom tool and setting the geometry with heat. Figure 1 is a schematic showing device fabrication and deployment processes.

Figure 1:

Schematic showing device manufacturing and deployment.

2.2. In Vitro Characterization

Twenty-nine devices (N=29), 22 cm in length and with a 4.0 mm internal diameter, were manufactured using bare mandrels. These devices were used to evaluate and verify the uniformity of their morphological and mechanical properties in vitro. The average porosity and fiber size distribution was measured for all devices. Porosity was calculated using a gravimetric method [20]. Average porosity was defined as: $\phi =\left(\frac{{\rho }_{app}}{{\rho }_{bulk}}\right)\times 100\%$, where ${\rho }_{app}=\frac{m_{wrap}}{V_{wrap}}$ is the apparent density; V_wrap = the device volume; ρ_bulk = the polymer bulk density (defined by the supplier); and m_wrap = the specimen mass measured using a microbalance (AE240S, Mettler-Toledo, Columbus, OH). Fiber size distribution was assessed via image analysis (ImageJ, NIH, USA) of scanning electron micrographs (JCM 5000 SEM, JEOL USA, Peabody, MA) at 1000X magnification (20 measurements per image, 3 images per device, representing the proximal, medial, and distal locations).

Mechanical testing was performed to determine the uniaxial tensile elastic modulus, suture retention strength, and kink radius. Briefly, rings 5–8 mm in width were cut from three locations of each polymer tube (proximal, medial, and distal). Two small, parallel round bars were inserted into the ring lumen and connected to the posts of a uniaxial tensile tester (Chatillon TCD225 Tensile Tester, AMETEK, Largo, FL). The samples were pulled while immersed in 1X phosphate-buffered saline (PBS) at 37±1°C using a constant crosshead speed of 50 mm/min until material failure was observed. The 5% fOTC6 tension modulus was defined as the tension $\left(\frac{force}{2*width}\right)$ recorded during the test while passing through the 5% strain level. Suture retention strength was tested using a 6–0 Prolene suture (Ethicon, Inc., Somerville, NJ) placed 2 mm from the free edge of each device. The suture was knotted into a loop ~3 cm in diameter and pulled uniaxially until failure along the longitudinal axis of the device, while immersed in 1X PBS at 37±1°C, using the uniaxial apparatus previously described (crosshead speed=150 mm/min). Suture retention strength was defined as the peak load recorded during the test. Finally, kink resistance to a fixed radius of 20 mm was verified by manually wrapping the unpressurized devices around the circumference of a plastic cylinder 40 mm in diameter while immersed in 1X PBS at 37±1°C and confirming absence of any kinks along the length.

2.3. Animal Study Design

All animal procedures were performed under a protocol approved by the Institutional Animal Care and Use Committee (IACUC) at American Preclinical Services (APS, 8945 Evergreen Blvd., Minneapolis, MN), and the IACUC at Concord Biomedical Sciences and Emerging Technologies (CBSET, 500 Shire Way, Lexington, MA). NIH guidelines for the care and use of laboratory animals (NIH Publication #85–23 Rev. 1985) have been observed. Autologous reversed SVGs were implanted as femoral artery interposition grafts using end-to-side anastomoses in N = 19 adult sheep (males and females >9 months old), 50–75 kg in weight. Device-treated SVG performance was assessed for up to 2 years. The 30-day, 90-day, 180-day and 365-day cohorts (N = 4 sheep/cohort) were implanted at APS. Three sheep in each cohort had a single treated SVG implanted, and the fourth sheep was implanted bilaterally with treated SVGs. The 730-day cohort (N=3 sheep) was implanted at CBSET, and each sheep in this cohort was implanted bilaterally with treated SVGs.

Details of the animal model used in this study have been previously described [19]. Briefly, the sheep were placed in a supine position on the surgical table with continuous vital monitoring. Carotid arterial access was implemented for volume administration, blood sampling, pressure monitoring, and angiography catheter insertion. Activated clotting time was monitored throughout the procedure and maintained at >350 seconds during the vein harvesting and graft implantation procedures by administering heparin intravenously (250 IU/kg). Antiplatelet therapy via aspirin (325 mg daily) and clopidogrel (75 mg daily) was administered orally three days prior to surgery and continued for the duration of the study. At the end of the designated implant times, animals were sacrificed by barbiturate overdose under deep anesthesia.

2.4. Saphenous Vein Harvest and Implantation

Saphenous vein segments 7–10 cm in length were harvested via posterolateral incisions in the hind limbs. Anti-spasm solution (1.5 mg/mL papaverine hydrochloride solution) was injected peri-adventitially using a small gauge (blunt) needle prior to dissection through the facial plane covering the adventitial plane. After two to three minutes, all tributaries were ligated using a 4–0 polypropylene suture. Prior to excision, the vein segments were flushed with preservation solution (500 mL lactated Ringer’s supplemented with 300 mg papaverine and 3000 IU heparin) using a pressure relief valve-equipped syringe, with a maximum pressure of 1.5 psi. The vein segments were stored in preservation solution for approximately 5 minutes prior to processing.

SVGs were cannulated with an atraumatic stainless-steel mandrel while immersed in preservation solution. The mandrel with mounted SVG was then placed into the electrospinning device, and the strain- relieving spine was deployed over the SVG. The electrospinning process was then initiated, depositing polymer micro-fibers until the conformal coating reached a thickness of approximately 250 μm. Next, the treated SVG was removed from the mandrel and stored in preservation solution until implantation, which was typically less than 5 minutes.

Once the SVGs were ready for implantation, femoral arteries were exposed via inguinal skin incisions. Grafts were spatulated and anastomosed in an end-to-side fashion via continuous running sutures using 7–0 Prolene. The SVGs bridged two arteriotomies on the femoral artery approximately 5 cm apart. Following the completion of both anastomoses, the femoral arteries were ligated near each anastomosis and severed between the ligations. Baseline angiograms were performed before closing the surgical incisions with 2–0 Vicryl braided sutures (Ethicon, Inc., Somerville, NJ).

2.5. Angiography

Cine-angiography was performed using a C-arm CT scanner (Arcadis Avantic, Siemens Medical Solutions USA, Malvern, PA) with contrast (IsoVue, Bracco Diagnostics, Monroe Township, NJ) for all grafts at baseline, follow-ups, and pre-termination in order to evaluate patency, and geometry. Standard J-type angiocatheters were inserted and advanced from the carotid access site until the tip of the catheter was positioned approximately 5 cm from the proximal anastomosis of each SVG. Graft patency was evaluated as a binary measurement (i.e., patent or occluded) with a qualitative description of filling defects within the graft. SVG internal diameters were also measured over time using a reference scale (the known size of the catheter used for infusion of contrast media). All angiographic measurements were made with validated software (Syngo QVA, Siemens Medical Solutions USA, Malvern, PA). Blood pressure was recorded during these measurements to ensure consistent mean arterial pressure across different grafts and time points.

2.6. Specimen Preparation for Light Microscopy

Following terminal angiography and immediately after euthanasia, SVGs were pressure-fixed in situ at 80–100 mmHg with 10% neutral-buffered formalin for 10–15 minutes. Each graft was harvested and immersion-fixed in 10% neutral buffered formalin for 24 hours prior to histological processing, performed by Alizée Pathology, LLC (20 Frederick Road, Thurmont, MD 21788). Each SVG was segmented and sectioned (4 μm thickness) at five locations (proximal anastomosis, proximal graft, mid-graft, distal graft, and distal anastomosis) and prepared for staining and light microscopic examination by standard histologic and immunohistologic methods.

2.7. Histomorphometry and Immunohistochemistry

Sections were stained using hematoxylin and eosin (H&E) and Movat’s pentachrome (MP) using standard protocols. Sections stained with MP from the mid-graft locations of each SVG were used for histomorphometry to quantify the areas of the neointima and media using Image-Pro® Plus software (Media Cybernetics, Rockville, MD). The neointima was defined as the tissue between the native internal elastic lamina (IEL) and the vessel lumen. The media was defined as the tissue subtended by the external elastic lamina (EEL) and the IEL. Neointima and media areas area were manually traced, and subsequently measured via automated software functions.

Immunohistochemistry evaluation was performed on sections from mid-graft of one representative treated SVG from the 30-day, 90-day, 180-day and 365-day cohorts. These sections were assessed for expression of CD163 (MCA1853, 1:500, AbD Serotec, Raleigh, NC, USA), a specific marker for macrophages [21]; alpha-smooth muscle actin (ab7817, 1:150, Abcam, Cambridge, MA, USA), a marker for smooth muscle cells and myofibroblasts [22]; von Willebrand factor (A0082, 1:1200, Dako, Carpinteria, CA, USA) and CD31 (ab28364, 1:150, Abcam, Cambridge, MA, USA), which are markers for endothelial cells [23]. A standard streptavidin biotin detection system (Vectastain Elite ABC Kit (Goat IgG), PK-6105, Vector Labs, Burlingame, CA) was used to detect the expression of all these markers.

2.8. Inflammation and Polymer Phagocytosis

Sections stained with H&E and MP, from each of the five locations along the graft, were used to perform a semi-quantitative evaluation of inflammation and polymer phagocytosis. Inflammation scores of 0 to 4 are based on severity as follows: 0 = Absent; 1 = Minimal; 2 = Mild; 3 = Moderate; 4 = Severe. Polymer phagocytosis scores are based on polymer fragments observed within inflammatory cells, and were scored as follows: 0 = Absent; 1 = Minimal; 2 = Mild; 3 = Moderate; 4 = Severe. Inflammation and polymer phagocytosis scores were separately reported for both the main body of the graft and for the anastomoses.

2.9. Statistical Analysis

This observational study was intended to demonstrate the technical feasibility, and the reproducibility of the healing response of the technology, and therefore statistical analysis of the data was not performed. Angiographic, histomorphometric, inflammation, and phagocytosis data are presented as mean ± standard deviation.

3. Results

3.1. In Vitro Device Characterization

All devices (N=29) were grossly free of defects (as assessed via unmagnified, corrected vision under illumination per Section 8.1 of ANSI/AAMI VP20:1994), such as the presence of holes and other discontinuities or imperfections of fabrication, and for the presence of dirt, soiled areas, spots, stains, or loose particles. A representative photograph of a treated SVG is presented in Figure 2-A. A representative scanning electron photomicrograph of the outer surface of a device is shown in Figure 2-B. The device exhibits a highly porous morphology characterized by bonded overlapping fibers approximately 5 μm in diameter. Mechanically, the material exhibits a relatively stiff, linear elastic behavior under arterial pulsatile load in the circumferential direction (Figure 2-C), with cyclic diameter excursions of less than 2% strain (by the Law of Laplace, a 4-mm diameter vessel under normotensive arterial pressure is subjected to an average wall tension of approximately 0.35 N/cm). The mechanical and morphological properties of the device are summarized in Table 1.

Figure 2:

A – Macroscopic appearance of a device-treated SVG. B – Microscopic morphology of the device obtained via SEM at 1000X magnification. C – Average circumferential uniaxial tensile properties of the device, represented by the tension (load divided by the sample width) generated by pulling along the circumferential direction.

Table 1:

Summary of mechanical and morphological properties.

Property	Units	Mean	Standard Deviation
Porosity1	%	49.3	2.0
Fiber Size	μm	7.8	1.7
5% Tension Modulus2	N/cm	1.29	0.14
Suture Retention	N	2.2	0.3
Strength
Kink Radius3	mm	< 20	n/a

N = 29 for all tests;

¹Porosity is intended as the fraction of volumetric void in the material;

²5% Tension Modulus was the tension (i.e., Load/2*Width) measured at 5% strain;

³Kink radius was only tested for a value of 20 mm.

3.2. Animal Studies

All animals (N=19) survived to their designated timepoints. Treated SVGs from all cohorts were patent except for one treated SVG in the 180-day cohort that was occluded at the 90-day follow-up angiogram, and one graft in the 730-day cohort that was occluded at the 30-day follow-up angiogram. Both of these animals, with one occluded graft each, survived to their intended timepoints. The aggregated patency rate was 92% (24/26). While all SVGs in the 90-day cohort were patent, premature (i.e., prior to 3 months) fragmentation of the polymer structure was observed in 3/5 treated SVGs. This led to localized dilations that resulted in hemodynamic disturbances, and subsequently focal stenoses of 50%−75%. Overall, 81% (2½6) of the implanted treated SVGs were free of complications (i.e., occlusion, or premature fragmentation) when considering all cohorts in aggregate.

3.3. Angiography

Qualitative angiographic assessment of all treated SVGs in all cohorts, except the three anomalous grafts from the 90-day cohort, revealed smooth and uniform luminal geometries throughout the study. Representative snapshots from angiograms of treated SVGs in each cohort at baseline and termination are shown in Figure 3.

Figure 3:

The rows show a representative angiography snapshot of treated SVGs for each cohort at baseline and termination. Note that the scale in the images is not homogeneous.

Quantitative angiography demonstrated that treated SVGs had geometric stability at baseline (4.0 mm ± 0 mm; n=26), 30 days (4.02 ± 0.38 mm; n=25), 90 days (3.63 ± 0.86 mm; n=19), 180 days (3.3 ± 0.25 mm; n=14), 365 days (3.59 ± 0.37 mm; n=10), and 730 days (3.74 ± 0.46 mm; n=5) as shown in Figure 4.

Figure 4:

Summary plot of quantitative angiography showing distributions of average graft diameters at each timepoint. Data is presented as mean ± standard deviation. Note that at implant there are (n=26) grafts; at 30 days there are (n=25) grafts; at 90 days there are (n=19) grafts; at 180 days there are (n=14) grafts; at 365 days there are (n=10) grafts; and at 730 days there are (n=5) grafts included in the distributions. Data is presented as mean ± standard deviation.

3.4. Microscopic Observations

Qualitative evaluation of the histology data showed that treated SVGs had advanced remodeling by 180 days as characterized by polymer structure fragmentation and integration; as well as the development of a neointima, and a confluent endothelium (Figure 5).

Figure 5:

Representative histology images from each timepoint using Movat’s pentachrome stain (yellow – collagen; black – elastin; blue – proteoglycans/glycosaminoglycans; pink – cell cytoplasm; and purple – cell nuclei), and hematoxylin and eosin stain (pink – extracellular matrix; and blue – cell nuclei).

The timepoints investigated in this study provided histologic snapshots of the remodeling response of treated SVGs and the external support device. In the earliest explants examined (1 month), the vein tissue was hypocellular (see reduced number of nuclei in the 30-day image in Figure 5). Between 3 months and 6 months, remodeling was characterized by progressive fragmentation of the polymer structure as it became embedded within a continuum of highly vascularized fibrous tissue. The remodeling host tissue appeared to become more vascularized with time, and more macrophages and foreign body giant cells (FBGCs) surrounded the polymer fragments. Between 6 months and 1 year, elastic fibers became more prominent and formed well-defined layers that resembled arterial EEL and IEL; these defined the neo-medial layer of the remodeled SVG (Figure 6). Also, during this period, the neo-media contained smooth muscle cells and myofibroblasts, evidenced by the expression of alpha- smooth muscle actin (a-SMA), as shown in Figure 7. Immunohistochemical staining of the longest functioning grafts with CD31 and von Willebrand factor (vWF) showed that the cells lining the lumen at the mid-graft of treated SVGs were endothelial cells (Figure 8).

Figure 6:

Movat pentachrome stained images from a graft implanted for 365 days. The letter “W” indicates the device wall. Note the significant biofragmentation of the device; the macrophages and foreign body giant cells surrounding the device fragments; the highly vascularized, fibrous capsule; the mature neo-media; and the arterial-like external and internal elastic laminae (EEL and IEL, respectively). The scale bars in the bottom two panels are 50 μm in length.

Figure 7:

Summary of immunohistochemistry staining with CD163 and alpha-smooth muscle actin (α-SMA). Representative images taken at high magnification (scale bars = 100 μm) for each cohort. The letter “W” indicates the device wall, and the letter “L” indicates the graft lumen. The 365 Day images have additional higher magnification (scale bars = 50 μm) panels shown at the bottom of the figure.

Figure 8:

Summary of immunohistochemistry staining with CD31 and von Willebrand factor (vWF). Representative images taken at high magnification (scale bars = 25 μm) for each cohort. Arrow heads indicate the expression of the marker.

The later remodeling response of the external device (up to 2 years), was predominantly characterized by the polymer structure continuing to be fragmented and integrated within a continuum of a progressively organized fibromuscular tissue capsule (Figures 5 and 6). The remodeling host tissue appeared to have developed sufficient strength to chronically prevent SVG dilation under arterial pressure, which was confirmed via angiography (see Figure 4). The EEL, IEL, and neo-media also continued to develop giving the remodeled SVG the appearance of a diametrically uniform conduit with arterial-like features (Figure 6).

There was limited inflammation associated with the device material at all timepoints. The semiquantitative inflammation and polymer phagocytosis scores over time are shown in Figure 9 and Figure 10, respectively. Inflammation was characterized as a local minimal to mild foreign body reaction mediated by macrophages and FBGCs along the inner and outer surface of the device, with no extension of the inflammatory process into the vein wall itself (see Figure 6). This was confirmed via immunohistochemical staining with CD163, which showed the presence and localization of macrophages and FBGCs over time in the treated SVGs (Figure 7). Additionally, at all timepoints, there was no evidence of local irritation, with no or very few neutrophils and eosinophils observed. Lymphocytes were likewise not prominent. Polymer degradation and phagocytosis was absent at the 30-day and minimal at the 90-day timepoints. By 180 days, there was mild fragmentation and phagocytosis of wrap fibers at all levels in all grafts. At both one year and two years, there was moderate to marked fragmentation and phagocytosis of wrap fibers at all levels in all grafts.

Figure 9:

_Summary of inflammation scores over time in the main body of the graft (▲) and the anastomosis regions (∆). Inflammation scores of 0 to 4 are based on severity as follows: 0 = Absent; 1 = Minimal; 2 = Mild; 3 = Moderate; 4 = to Severe. Data is presented as mean ± standard deviation.

Figure 10:

Summary of phagocytosis scores over time in the main body of the graft (■) and the anastomosis regions (□). Polymer phagocytosis scores are based on polymer fragments observed within inflammatory cells, and were scored as follows: 0 = Absent; 1 = Minimal; 2 = Mild; 3 = Moderate; 4 = Severe. Data is presented as mean ± standard deviation.

Histomorphometric quantification (Figure 11) suggested that the remodeled SVG was not progressive after 180 days (5.62 ± 1.86 mm²), and by 365 days (4.95 ± 1.33 mm²) there doesn’t appear to be any difference in medial and neointimal areas compared to the 180-day cohort. However, when comparing the neointimal area at 365 days to the neointimal area at 730 days (0.66 ± 0.59 mm²), there seemed to be a dramatic reduction over time. Conversely, the medial area appeared to increase between 1 and 2 years of implantation. Between 1 and 2 years there appears to be an inversion in the amount of neointimal area and medial area. That is, the medial area at 1 year is less than the neointimal area, but at 2 years the medial area is greater than the neointimal area.

Figure 11:

Summary of histomorphometry quantification of neointimal (▲) and medial (■) areas. Data is presented as mean ± standard deviation.

4. Discussion

The present study was an investigation of an electrospun PDMS-PEU external support device for SVGs. We observed that the device induced structural and functional remodeling with patency and diametrical uniformity under arterial pressure in an ovine model to 2 years. The device provides temporary mechanical support for a period of about 3 to 6 months, after which it becomes fragmented and embedded within a fibrous connective tissue capsule. As the polymer structure fragmented, the biomechanical stress was likely transferred to the remodeling host tissue. The cells within the host tissue responded to the mechanical stimulus by becoming specialized phenotypes that produced arterial-like structures. The aggregated patency rate was 92% (24/26); however, 3 of the patent SVGs had devices showing signs of early structural deterioration prior to 3 months, which will be discussed in detail below.

In vitro device characterization showed that unimplanted devices had suitable circumferential strength, suture retention strength, and kink resistance for use as an SVG external stent. The device tension modulus at 5% strain, a measurement of material stiffness, was used to help ensure that strain levels under normotensive loads were within the material’s elastic limit. Using the thin-walled formulation of the Law of Laplace [24], the circumferential wall tension in 3 mm to 5 mm diameter cylindrical conduits subjected to a normotensive systolic pressure of 120 mmHg is 0.28–0.45 N/cm. So, we estimate that treated SVGs are subjected to strain magnitudes of 1–3% under arterial pressure, well below the elastic limit of the device (see Figure 2-C). This level of cyclic strain is not anticipated to reduce device durability, as evidenced in similar materials [25]. The average suture retention strength of the device was comparable to human arteries and helps reduce the risk of tearing during suturing and surgical manipulation [26,27]. Additionally, treated SVGs have enhanced suture retention strength compared to untreated SVGs. The 20 mm minimum radius of curvature used for kink resistance measurements was selected by analyzing three-dimensional dynamic geometries of native coronary arteries. Ding et al. [28] reported curvatures of 0.39 ± 0.1 cm⁻¹ (~25 mm) and 0.48 ± 0.17 cm⁻¹ (~21 mm) for the right coronaries and left anterior descending coronaries, respectively. Therefore, 20 mm represented a conservative limit for kink radius, as SVGs used for coronary artery bypass grafting are expected to have slightly larger radii of curvature because they follow the epicardium topography.

Quantitative angiography revealed that treated SVGs maintained a uniform, and non-dilating lumen diameter throughout the duration of the study, while our previous work in the same preclinical model showed that untreated SVGs exhibited progressively increasing dilation over time, doubling the initial diameter by 1 year [19]. Therefore, the device was able to provide a sustained diametrically restrictive function against SVG dilation, which has been proposed to be a key factor in mitigating IH development by maintaining a higher average shear stress on the graft lumen [16,29,30], and by reducing wall tension [6,16]. Figure 12 shows a comparison between a treated (A & C) and an untreated (B & D) SVG implanted for one year as an ovine femoral arterial bypass graft. Note the near doubling in diameter of the untreated graft relative to the treated graft. Also, note the increased neointima formation in the dilated untreated graft in comparison to the stable neomedia in the treated graft. Finally, restricting an SVG to a controlled, uniform diameter could allow for improved size matching to the target artery and reduced flow disturbances, which could further attenuate IH [31–33]. Diametrical matching has also been reported to accelerate in vivo re-endothelialization in arteries and vascular grafts [34,35]. Thus, by preventing dilation, maintaining diametrical uniformity, and improving diametrical matching to the target femoral arteries, the external support offered by the device maintained improved hemodynamics in treated SVGs.

Figure 12:

Representative images showing a low magnification H&E stained section of a treated (A) and untreated (B) SVG after one year. Representative snapshot images from angiograms performed after one year in a treated (C) and untreated (D) SVG.

It has been suggested that electrospun materials are effective substrates for host tissue ingrowth and remodeling because they mimic native extracellular matrix (ECM) microstructure [36], and PDMS-PEU has a long history of safe use for cardiac pacemaker and defibrillator lead insulation [37,38]. Also, polyurethanes that contain both PDMS and polyether soft segments have been reported to exhibit excellent mechanical properties and enhanced biostability relative to conventional thermoplastic PEU elastomers composed only of polyether soft segments [19,21,22]. Despite the reported in vivo durability of PDMS-PEUs, they do undergo oxidation from reactive oxygen species production, and hydrolysis from hydrolytic enzyme production, by macrophages and FBGCs; and they experience environmental stress cracking (ESC) due to inherent residual stresses and imposed mechanical stresses [41–45]. Indeed, the electrospun PDMS-PEU device architecture was designed (i.e., control of: surface area to volume ratio; fiber diameter; and inter-fiber bonding) to capitalize on the ESC phenomenon in order to tune a novel process that we term “biofragmentation”. This is a process by which the structural integrity of a biomaterial diminishes due to mechanical fracture of the structural components; and these biomaterial fragments remain integrated within the remodeling host tissue. Importantly, biofragmentation differs from bioabsorption and/or biodegradation because, in the latter, the cellular processes that are responsible for foreign material breakdown, and subsequent elimination, take much longer to reduce the size of the fragments of biofragmentable materials sufficiently to allow transport via the lymphatic system and bloodstream. It is possible that some polymer fragments remain permanently embedded within the host tissue, resulting in a composite bioartificial tissue structure.

Elast-Eon films under mechanical stress in an in vitro oxidative environment have been shown to exhibit cracks approximately 5 μm deep [45], which led to specifying that the electrospun fibers of our device were between 5 μm to 10 μm in diameter in an effort to attenuate the biofragmentation rate, so that the device maintained structural integrity for at least 3 months. Residual stresses built into the polymer fibers [46], as well as imposed stresses from arterial pressure, cellular infiltration, proliferation, and ECM synthesis within the porous structure, also played an important role in biofragmentation. By controlling fiber diameter uniformity and enhancing inter-fiber bonding, through tuning of the electrospinning process, the imposed stresses were more uniformly distributed, and thereby attenuated ESC to help control the rate of biofragmentation. The devices used here exhibited relatively large fiber and pore diameters (5–10 μm and 20–30 μm, respectively), and the matrix microstructure was characterized by highly interconnected fibers and pores, as evidenced qualitatively from SEM (Figure 2-B). Importantly, large pore sizes (macro-porosity) have been shown to be associated with less IH than micro-porous external stents [47].

The results for the 30-day, 180-day, 365-day and 730-day cohorts were generally similar, however, complications in three of five devices used to treat SVGs in the 90-day cohort were unexpected. Histological observations suggested that the devices were not structurally intact, and did not support these SVGs uniformly along their length for a period of at least three months. Prior testing demonstrated that 3 months was the minimum time necessary for sufficient structural remodeling to occur, prior to the device being biofragmented, in order to enable chronic diametrical support of the graft independent from the structural integrity of the device. The 3 of 5 treated SVGs in the 90-day cohort had irregular dilations along their length that led to areas of flow disturbance, which likely contributed to the development of stenoses. The treated SVGs in the 90-day cohort were characterized by a disproportionately high rate of premature device fragmentation. The 180-day cohort had 1/5 treated SVGs occluded by the 90-day timepoint; the 730-day cohort had 1/6 treated SVGs occluded by the 30-day timepoint; and none of the treated SVGs from the 30-day or 365-day cohorts showed evidence of complications. This suggests that the 90-day cohort results are inconsistent with the other cohorts. It should be noted that the 90-day cohort SVGs were implanted within the first few days of the study, followed by the 180-day, 30-day, 365-day and 730-day cohorts. A limitation of the study may have been that a learning curve remained during the initial procedures. We believe that the strain-relieving spine caused complications in suturing the anastomoses in the first several implants, perhaps leading to undetected damage to the device. Also, the use of non-atraumatic clamps on the treated SVGs to achieve hemostasis may have caused structural damage in the initially implanted devices. Surgical handling was corrected during the course of this study, and new device features involving the spine and processing conditions were developed after this study to prevent future recurrence of similar complications.

The structural support provided by the device reduced wall tension and prevented SVG diametrical irregularity during the first three months, after which device and tissue remodeling resulted in conduits with uniform diameters and arterial-like features capable of chronically withstanding arterial pressure without dilation. These results are in stark contrast with data previously published by us and others that showed untreated SVGs with heterogeneous dilations, including an aneurysmal appearance of some grafts, using similar preclinical models [19,48]. This study proposes that after the device lost structural integrity, and the integrated tissue was exposed to pulsatile stress from arterial pressure, the cells and tissue continued to differentiate and remodel, eventually developing arterial-like features. Specifically, the neointima was remodeled into a neo-media that is subtended by an artery-like EEL and IEL (Figure 6), which have a characteristically different appearance than that normally seen in pathologically arterialized SVGs.

The intermediate hypocellular state [49,50] seen in treated SVGs by 30 days may have played a role in the overall remodeling response. Decellularized ECM used as a tissue engineering scaffold has been shown to foster rapid and functional host tissue remodeling with properties tailored to the site of implantation [51]. Functional remodeling of decellularized tissues is induced by constituent growth factors in the ECM, which might exert chemoattractant effects in the host tissue and foster cell migration, proliferation, and differentiation [51]. The remodeling response of treated SVGs appeared to involve of the development of a vascularized fibrous connective tissue capsule, reminiscent of an arterial adventitia. The device likely facilitated this by buttressing the vein against dilation under arterial pressure, thereby preventing uncontrolled IH. Over time the neointima was remodeled to a structure which had histologically well-defined EEL and IEL structures and was populated with smooth muscle cells. This appears to have supported the development of a new intimal layer that was lined with endothelial cells. This proposed remodeling paradigm is presented schematically in Figure 13.

Figure 13:

Schematic diagram summarizing the stages of device-treated SVG functional remodeling. Pre-Implant – device is deployed onto the SVG; minimal effect on vascular cell viability. One Month – cellular (macrophage, foreign body giant cell (FBGC), fibroblast, and myofibroblast) infiltration into the porous device; extracellular matrix (ECM) synthesis by the fibroblasts and myofibroblasts; native vein becomes hypocellular; fibrin coagulum develops into pseudointima; infiltration of monocytes into pseudointima and subsequent differentiation into macrophages; few fibroblasts recruited to pseudointima; and few endothelial cells or endothelial progenitor cells attach to lumen. Six Months – macrophage fusion into FBGCs; FBGCs produce reactive oxygen species and hydrolytic enzymes, and fibroblasts/myofibroblasts produce ECM, both of which contribute to environmental stress cracking and subsequent biofragmentation of the polymer fibers; microvessel development within ECM; remodeling tissue is exposed to cyclic stress from blood pressure; more myofibroblasts present in ECM; native vein internal elastic lamina (IEL) becomes diffuse; native vein external elastic lamina (EEL) condenses; pseudointima evolves into a neointima that is populated with fibroblasts, myofibroblasts and smooth muscle cells; and lumen is fully endothelialized. 1 Year – continued biofragmentation of the polymer fibers and differentiation of the remodeling tissue; continued microvessel development within the ECM; neointima develops into neo-media that is subtended by the condensed native vein EEL and the newly formed IEL; and the remodeling tissue continue to differentiate in response to cyclic stress from the blood pressure.

5. Conclusion

Functional remodeling of the electrospun PDMS-PEU device appeared to be characterized by integration within regenerating host tissue, while eliciting only a mild foreign body reaction. This mild foreign body reaction to the material may have allowed constructive tissue remodeling and differentiation to occur. In the context of this study, functional remodeling included host tissue ingrowth into the porous device, biofragmentation of the device material, and subsequent differentiation and maturation of the remodeling tissue when exposed to cyclic biomechanical stress from the blood pressure. This study suggested that an electrospun PDMS-PEU external stent device provided the appropriate combination of architectural and biomaterial properties necessary to achieve stable chronic remodeling in treated SVGs using an ovine peripheral arterial bypass model.