Long-term remodeling of compliance matched tissue engineered vascular grafts is sex specific

Abstract

Compliance matching of tissue engineered vascular grafts (TEVGs) is a promising technique to combat common failure modes seen clinically in small diameter vascular procedures like coronary artery bypass grafting (CABG). However, despite the influence of sex on vascular response to injury, the role of biological sex on the success of TEVG innovations such as compliance matching, namely in long-term remodeling, has been understudied. In this study we fabricated compliance matched TEVGs (CM TEVGs) and hypocompliant TEVGs (Hypo TEVGs) by adjusting the thicknesses of a high polyester urethane urea (PEUU) containing layer (80:20 PEUU:Gelatin) and a low PEUU containing layer (20:80 PEUU:Gelatin). All grafts were implanted into the abdominal aorta of both male and female Sprague Dawley rats (n = 4) as interposition grafts for 6 months and monitored using in vivo ultrasound. Upon explant, key outcomes of long-term remodeling were assessed via immunohistochemistry, multiphoton imaging, and qRT-PCR. CM TEVGs were confirmed to be compliance matched to native aorta both in vitro and in vivo upon implant. After remodeling, CM TEVGs had increased degradation and decreased calcification compared to Hypo TEVGs. CM TEVGs had increased mature VSMC genes (Acta2, Myh11, Cnn1) in males and decreased macrophage presence (CD68) in females. Independent of graft type, females had increased intimal thickening, decreased luminal presence of mature VSMCs, and increased fibrillar collagen deposition compared to males in both CM and Hypo TEVGs. Our results suggest that long-term remodeling outcomes in response to compliance matching are dependent on biological sex at both the tissue and cellular level.

1. Introduction

Coronary artery disease (CAD) is responsible for over 360,000 deaths in the US per year and affects approximately 18 million adults over the age of 20 worldwide [1,2]. Severe CAD requiring surgical intervention is often treated with coronary artery bypass grafting (CABG) as it offers improved post-surgical outcomes for patients with multivessel disease when compared to percutaneous coronary intervention [3]. The gold standard of CABG conduits is an autologous graft such as the saphenous vein or internal mammary artery. However, small diameter autologous grafts can have up to a 42% failure rate depending on the vessel used. This is largely due to intimal hyperplasia (IH) and restenosis [4,5]. There is also a shortage of usable autologous vessels and there are currently synthetic small diameter graft options that outperform autologous grafts [6]. Therefore, there has been a push to develop a tissue engineered vascular graft (TEVG) that avoids common CABG failure modes.

A key factor in successful host remodeling of small diameter vascular grafts is vascular smooth muscle cells (VSMCs) behavior. VSMCs exist on a phenotypic spectrum that allows them to shift from contractile in healthy vasculature to proliferative and ECM producing in states of injury [7]. Dysregulation of these pathways directly contributes to graft failure modes such as IH and vascular calcification [8].

A common driver of VSMC dysregulation in vascular grafts is compliance mismatch between the conduit and the host artery. Hemodynamic shifts caused by the mismatch lead to abnormal VSMC migration and restenosis [9,10]. Recent work has sought to incorporate compliance matching into TEVG design using a variety of biopolymers [11,12]. Compliance matching provides idealized hemodynamics, decreased IH markers, and decreased calcification [13,14]. Our lab has shown that computationally optimized compliance matched TEVGs (CM TEVGs) made from polycaprolactone (PCL) and gelatin have increased anti-inflammatory macrophage presence and increased contractile VSMCs in the graft matrix during acute remodeling (28 days) [15–17]. However, these grafts have increased luminal VSMC presence, indicating possible risk of IH through continued remodeling. Further investigation over long term studies is required to examine potential failure modes of these TEVGs.

Biological sex is also a key factor in VSMC behavior in vascular grafts. CABG outcomes are highly influenced by sex with post-menopausal women have an increased risk of major adverse cardiac events in the first 5 years after a CABG procedure [18,19]. This phenomenon may in part be explained by estrogen’s modulation of VSMC function and the reduction of estrogen levels after menopause [20,21]. Sex has been shown to influence the ability of vascular cells to respond to changes in matrix stiffness, which suggests that the effects of compliance matching may be sex specific [22–24]. Despite this, there are few TEVG studies that focus on this interplay between compliance, VSMC behavior, and sex during long-term remodeling.

To address this gap, this study aims to characterize how biological sex affects the long-term remodeling response towards compliance matching in TEVGs. In this study we developed electrospun CM TEVGs that were mechanically tuned using gelatin incorporation to modulate the stiffness of a polyester urethane urea (PEUU) matrix - a biodegradable polymer being used for several potential clinical cardiovascular applications [25]. The constructs also include an inner layer of polyester sulfobetaine urethane urea - 50 (PESBUU-50), a nonthrombogenic polymer with zwitterionic sulfobetaine parts in a PEUU backbone [26]. These grafts were successfully implanted into a rat small animal model and remained patent through six months of remodeling. This work has shown that compliance matching influences macrophage count, VSMC maturity, as well as graft degradation, but the magnitude of these effects is sex specific. Biological sex has a profound influence on several key aspects of remodeling independent if graft type, such as collagen deposition and calcification. Taken together, this work provides new insights specific to the role of compliance matching and its relationship with sex in a small animal model.

2. Materials and methods

2.1. CM TEVG fabrication and characterization

2.1.1. Electrospinning and crosslinking

All TEVGs were fabricated using a commercial electrospinning device (IME Technologies, Waalre, Netherlands), crosslinked, and characterized. PESBUU-50, PEUU, and porcine skin gelatin (Sigma-Aldrich) were dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) (Oakwood Chemical) to create the following three 10% w/v solutions: PESBUU-50, 80:20 PEUU:Gelatin, and 20:80 PEUU:Gelatin that were used for the inner, middle, and outer graft layers, respectively. Solutions were loaded into 5 mL syringes (BD), placed into a syringe pump (Chemyx Inc.), and connected to the system via PTFE tubing (MODDIY). To create the scaffold, the spinner head was set to a translational distance of 80 mm and translational speed of 300 mm/sec. The head was charged to 20 kV and the target was charged to −4 kV. Solutions were dispensed onto a stainless steel, Teflon coated mandrel (Applied Plastics) with a 1 mm outer diameter (OD) rotating at 1000 rpm. The working distance between the tip of the dispenser and the mandrel was 10 cm. All solutions were dispensed at a rate of 10 μL/min. Compliance was modulated by the volumes dispensed for layers 2 and 3 (Fig. 1A). After spinning, grafts were crosslinked in 0.5% w/v genipin (Wako Chemicals) in 100% ethanol on an orbital shaker at 50 rpm (Cole Parmer, Vernon Hills, USA) at 37 °C for 24 h. Grafts were then rinsed 3 times in 200 proof ethanol to remove unused crosslinker, taken off the mandrel, and stored in ethanol for further testing. Cross sectional images of the constructs were taken on a dissecting microscope (Olympus) and the average wall thickness was measured using CellSens camera software.

Fig. 1.

Overview of methods. (A) Composition of compliance matched and hypocompliant electrospun TEVGs. Each TEVG design has an identical inner layer of PESBUU-50. The thickness of layers 2 and 3 (low gelatin and high gelatin, respectively) were modulated via amount of polymeric solution dispensed in μL. (B) CM TEVG loaded into a custom CellScale Tubular Testing System (CellScale Biomaterials Testing, Waterloo, Ontario) used to test in vitro compliance. Scale bar = 1 mm. (C) TEVG small animal implant and ultrasound monitoring. Surgical images were taken post hemostasis and before closing. Ultrasound images were captured within five days of implant to confirm graft location and patency. Scale bar = 2 mm. (D) Sample preparation scheme for TEVGs upon explant (graphic made with Biorender.com).

2.1.2. Compliance testing

A Custom CellScale Tubular Testing System from CellScale Biomaterials Testing (Waterloo, Ontario) was used to measure the compliance of both native rat aortas and TEVGs (Fig. 1B). To measure native aortic compliance, the abdominal aortas of 6 Sprague Dawley rats (3M/3F) were explanted and stored in sterile saline until testing. Samples were cannulated onto 1 mm OD cannulas and secured to the device (Fig. 1B). All samples were pulled taut until the load cell registered a force of 0.01 N to control for axial tension before testing began. TEVGs and aortas were inflated from 0 mmHg to 140 mmHg with a ramp speed of 4 mmHg/sec. Testing was performed for 10 cycles, with the first 9 cycles being used as preconditioning and the 10th cycle retained for further analysis (n = 8 for each graft type). All mechanical tests were run on OpenX 12.11 software designed by CellScale. Compliance was calculated using Eq. (1).

$$\frac{\left({OD}_{120\text{mmHg}}-{OD}_{70\text{mmHg}}\right)÷{OD}_{70\text{mmHg}}}{120\text{mmHg}-70\text{mmHg}}$$ (1)

2.2. In vivo assessment of TEVGs via small animal model

2.2.1. Rat abdominal aortic interposition graft surgery

All tissue harvesting and animal procedures were in accordance with the University of Pittsburgh institutional guidelines under a protocol approved by the University’s Institutional Animal Care and Use Committee (IACUC). TEVGs were implanted as interposition grafts in the abdominal aortas of Sprague Dawley rats (Fig. 1C). In total, 12 male and 12 female rats (8 CM, 8 Hypo, 8 sham) underwent surgical procedures with an average weight of 204 ± 19 g (female: 194 ± 12 g, male: 211 ± 20 g) and average age of 69 ± 14 days (female: 71 ± 16 days, male: 66 ± 12 days). For each surgery, rats were anesthetized with isoflurane (3%–5% for induction and 1% for maintenance), ketamine (0.91 mg/kg), and xylazine (0.091 mg/kg). Abdominal fur was trimmed with an electric clipper and removed with depilatory cream (Nair^™). The exposed surgical area was prepped with a betadine scrub (Medline) before making an incision (4–5 cm) along the midline of the abdomen. The abdominal viscera were gently exteriorized and covered with saline-soaked gauze to prevent dehydration during the procedure. A 1–2 cm section of the aorta was isolated between the aortic bifurcation and the renal arteries and 1–2 aortic branches were cauterized to prevent blood loss in subsequent surgical steps.

Microvascular clamps (40 g clamping force, AROS Surgical^™, Newport Beach, USA) were placed proximally and distally on the isolated aorta. The aorta was transected, and the lumen was flushed with heparinized saline (60 UI/mL) using a 22 g flexible catheter tip (MedVet International). Anastomotic suturing began proximally using approximately 8–12 nylon sutures (10–0, AROS Surgical^™) placed in an interrupted fashion. This process was then repeated on the distal segment and the clamps were removed to restore blood flow. The average length of implant for CM TEVGs and Hypo TEVGs was 4.9 ± 0.9 mm and 5.2 ± 0.9 mm, respectively. If leakage occurred, gentle pressure was applied at the anastomotic site with sterile gauze. In some cases, an extra suture was required for complete hemostasis. Blood flow was assessed via pulsatility of the aorta distal to the graft.

A group of 8 rats (4M/4F) underwent a sham surgical procedure. The same procedure was followed until clamping, where a double artery clamp was used to secure the aorta (AROS Surgical, 40 g clamping force). The aorta was cut in the center, the clamps were slid together to account for elastic recoil, and the two aortic ends were reanastomosed. Only one anastomotic site was created in sham rats.

Once complete, the exteriorized viscera were returned to the abdominal cavity. The muscle layer was closed with a continuous over and over suture technique (3 – 0 Coated PGA Sutures, Covetrus). Before the layer was entirely closed, 5 mL of sterile saline was introduced into the abdominal cavity to prevent post-operative dehydration. Lidocaine Hydrochloride (4% topical solution) was applied onto the muscle then VetBond (Covetrus) was applied on top of the sutures. The dermis was closed using continuous subcuticular sutures (3 – 0 PGA sutures) and the skin was stapled (Covetrus) to prevent wound dehiscence. Post operatively rats were treated with subcutaneous Ethiqa XR (0.65 mg/kg) and intramuscular cefazolin (0.1 mg/kg) immediately after closing for pain management and infection prevention, respectively. For ongoing pain management, rats received subcutaneous carprofen (5 mg/kg) 24 and 48 h post-op. Aspirin (analgesic & antiplatelet) and dipyridamole (antiplatelet) were ground up with a pill crusher and stirred into an ice cream cone filled with jelly. For the first 7 days post-op, aspirin and dipyridamole were given at a high dose (200 mg and 250 mg, respectively) once a day. For the next 21 days, a low dose of aspirin and dipyridamole (100 mg each) was provided daily [15,27]. After 6 months, rats were euthanized via carbon dioxide inhalation followed by cervical dislocation as recommended by the American Veterinary Medical Association Guidelines for the Euthanasia of Animals.

2.2.2. Noninvasive monitoring of compliance and hemodynamics

A high frequency ultrasound scanner was used to assess patency, graft/vascular (inner) diameters, and in vivo velocity measurements (Vevo 3100, FUJIFILM VisualSonics Inc. Canada) using a MX400 (30 Hz) ultrasound probe. For each animal, scans were taken within 5 days post operatively and then monthly for 6 months.

To perform the scans, rats were anesthetized with isoflurane (3%–5% for induction and 1% for maintenance), the abdominal fur was removed as described above, ultrasound gel (Parker Labs Aquasonic) was applied, and images were acquired. All quantified images were captured longitudinally, and data were collected using Vevo Lab software version 5.7.1. Color Doppler was used to confirm patency in the graft. Patency was defined by observable blood flow through the conduit (Fig. 1C).

MATLAB was used to automate the measurement of the maximum and minimum (systolic and diastolic) inner diameters of the graft and surrounding native aortic sections. Briefly, an M-Mode image of the vessel (which shows diameter change over time) was exported as a traditional image file and analyzed using the Fazzy Logic function in MATLAB to find the edges of the wall. The image is then cropped to show only the upper wall during a single cardiac cycle. A Fourier series 3 is fitted over the curve of the wall, and the maximum is found. Systolic diameter is defined as the vertical distance from this max point to the lower wall calculated by the code. The same process is repeated for the diastolic diameter, but the curve minimum is used to calculate the distance. In vivo compliance was calculated using Eq. (2). Pulsed wave doppler was used to measure the average peak systolic and end diastolic velocities (PSV and EDV, respectively) proximal, within, and distal to the graft.

$$\frac{\left({ID}_{max}-{ID}_{min}\right)÷{ID}_{min}}{\mathit{\text{Pulse Pressure}}}$$ (2)

A native baseline for all measured variables was acquired by averaging scans from n = 12 (7M/5F) rats that were the same weight as the rats used for TEVG implant. We used a previously published population average pulse pressure (40 mmHg) for our calculations. No significant differences between male and female pulse pressures were noted [28,29].

2.3. Explant graft characterization

Six months post implant, animals were euthanized and TEVGs were explanted. The aortic tissue closest to the anastomotic site was also explanted. TEVGs were rinsed in saline and sectioned into three portions (Fig. 1D). The midsection of every graft was fixed in 4% paraformaldehyde (PFA) (Thermo Fisher Scientific) at 4°C for 24 h before being rinsed three times for five minutes in room temperature 1x Phosphate Buffered Saline (PBS). Midsections were then stored in PBS at 4°C until they were characterized via multiphoton imaging. To reduce the confounding effects of proximal versus distal anastomotic sites, in each group of n = 4, two proximal and two distal portions were used for each characterization method. Sections used for immunohistochemistry (IHC) were fixed in 4% PFA for 24 h before being rinsed in PBS. Samples were stored in PBS at 4 °C for an additional 24 h before being frozen in O.C.T. media (Fisher Scientific) and placed in a −80 °C freezer until sectioning. O.C.T. blocks were sectioned into 10 μm cryosections by the Histology Core at John G. Rangos Sr. Research Center at the UPMC Children’s Hospital of Pittsburgh and stored at −80 °C until staining. Superfrost Plus Gold (Fisher Scientific) slides were used to prevent TEVG containing samples from detaching during staining. Portions of the explant used for qRT-PCR were stored in RNALater (Invitrogen) at 4 °C. Shams were only assessed with immunohistochemistry or qRT-PCR with no midsection for 2P characterization.

2.3.1. Immunohistochemistry and immunofluorescence of frozen sections

All explants were stained with H&E (Sigma Aldrich), Masson’s Trichrome (Sigma Aldrich), and Von Kossa (Polysciences Inc.) stains per the manufacturer’s instructions. Briefly, frozen slides were thawed and excess O.C.T. was rinsed off in DI water before staining. After staining, all samples were mounted and coverslipped. All stained slides underwent brightfield color imaging with a EVOS M7000 (Thermo Fisher) using a 20x objective (Invitrogen, AMEP4982). Shams and anastomotic site samples were also stained with H&E, with sham males, sham females, and CM males having an n = 3 due to insufficient native tissue being explanted.

For immunofluorescent (IF) staining, slides were thawed and rinsed in saline to remove excess O.C.T. For heat induced antigen retrieval, Sodium citrate or EDTA (per recommendation of the manufacturer of each primary antibody) were brought to boil in a staining jar using a microwave. Slides were placed in the jar for 5 min. Samples were rinsed in PBS to remove the antigen retrieval solution before blocking with 10% horse serum (Gibco) in 0.1% Triton in PBS (PBST) for 1 h. Slides were stained for Alpha Smooth Muscle Actin (α-SMA, Abcam, ab7817, 1:200), CD31 (Abcam, ab182981, 1:200), Myosin Heavy Chain - 11 (MYH11, Abcam, ab224804, 1:200), Ki67 (Abcam, ab16667, 1:200), CD68 (Abcam, ab125212, 1:200), or Intracellular Adhesion Molecular - 1 (ICAM-1, Thermo Fisher Scientific, PA596365, 1:200) in blocking solution overnight at 4 °C. They were then rinsed in PBST on a shaker for 10 min before secondary staining with either Alexa Fluor^® 488 AffiniPure^® Donkey Anti-Rabbit IgG (Jackson ImmunoResearch Laboratories, 711-545-152, 1:300) or Alexa Fluor^® 488 AffiniPure Donkey Anti-Mouse IgG (Jackson ImmunoResearch Laboratories, 715-545-151, 1:300) in PBST for 1 h at room temperature. Finally slides were rinsed once more in PBST and mounted with Prolong Gold Antifade with DAPI (Thermo P36935). All IF images were taken on an EVOS M7000. Cell nuclei, graft autofluorescence, and the marker of interest were captured in the DAPI (357/44 nm excitation, 447/60 nm emission), Cy5 (628/40 nm excitation, 692/40 nm emission), and GFP (482/25 nm excitation, 524/24 nm emission) channels, respectively. Images were captured using a 10x objective (Invitrogen, AMEP4981). For all IF markers, the Hypo female group has n = 3 samples due to a cryoslicing issue that prevented samples from sticking onto the slide throughout the staining process. Additionally, there were insufficient slices of one CM Male sample to stain for all IF markers. MYH11, ICAM-1, and CD68 are specifically missing.

2.3.2. Multiphoton imaging of TEVG midsections

All three-dimensional images were taken with a TriMScope II multiphoton scanning system (LaVision BioTec, Bielefeld, Germany) mounted on an Olympus BX51 upright microscope base (Olympus, Tokyo, Japan) and coupled to a 120-femtosecond tunable pulsed laser (INSIGHT DS+DUAL, Newport, Irvine, CA). After fixing and rinsing, the midsection of each explant was placed into a PDMS well filled with PBS for imaging. All images were taken in the transverse view. An Olympus BX51 upright scanning microscope (Olympus, Tokyo, Japan) was coupled to a 120-femtosecond tunable pulsed laser (INSIGHT DS+DUAL, Newport, Irving CA) and a Nikon 16x water immersion objective (0.8 NA).

Samples were excited at 792 nm and explants were imaged for collagen deposition and graft degradation. Second harmonic generation (SHG) signal for fibrillar collagen was collected at 395/25 nm using 560, 495, and 425 long pass dichroic mirrors. TEVG autofluorescence was collected at 620/40 nm through 560 and 595 lowpass filters. Additional images were collected at 460/40 nm and 525/50 nm to capture and correct for potential signal bleeding across channels. Four sets of z-stack images were taken (located at 0°, 90°, 180°, and 270°) through the entire depth of the section in 5 μm steps. The average depth of scan was 612 ± 93 μm.

2.3.3. Image processing

All images were processed using MATLAB. A visual overview of these techniques is provided in the supplement.

H&E:

A tiled mosaic of 20x images of the entire sample was used to measure luminal growth. The inner and outer boundaries of the intima were selected using the function ROISpline [30]. These boundaries were then converted to polar coordinates (as explants are circular) and fit to a smoothed contour. Thickness was measured by subtracting the inner fitted polar line from the outer fitted line at each theta value. All these values were averaged to generate a value for mean intimal thickening for each explant (Supplemental Figure 1).

Multiphoton:

Three-dimensional z-stacks of four graft quadrants were processed slice by slice in MATLAB. Using a fused image of all four channels, an overall ROI was generated by detecting the outer boundaries of the sample and fitting a convex hull around the edges, creating a ROI mask. Next, looking at only the 395 channel (SHG), collagen fibers were extracted using the MATLAB function fibermetric [31]. Collagen fibers were then quantified in one of three ways. First, the overall SHG signal was calculated by dividing the number of positive pixels in the SHG channel by the area of the ROI mask. In the second we quantified the number of collagen fibers per ROI by finding the center of each unique fiber. Finally, tortuosity was calculated by dividing the path length between two fiber endpoints by the straight line distance between the same two points (Supplemental Figure 2).

IF Markers:

Images were taken at 10x. A tiled RGB image with all three channels blended was used to create regional masks of the luminal, graft, and abluminal areas using ROI spline. The marker of interest was defined as signal above a threshold that filters out background noise and autofluorescence in the GFP channel (482/25 nm excitation, 524/24 nm emission). This signal was binarized and fluorescent density was calculated by dividing the number of marker positive pixels by the total area of the region (Supplemental Figure 3). To measure luminal endothelial continuity (CD31 staining), positive luminal signal was skeletonized. The length of the signal skeleton was divided by the total length of the inner luminal mask.

Graft degradation:

Degradation was calculated using 10x IF images of graft autofluorescence (Cy5 channel, 628/40 nm excitation, 692/40 nm emission). Each image was binarized and skeletonized to create a centerline. Thickness was found by counting the number of pixels along a set of equally spaced lines orthogonal to the centerline. Mean sample thickness was calculated by averaging the thickness of each 10x tile for a sample. Percent degradation was calculated using Eq. (3) for each explant (Supplemental Figure 4).

$$\frac{{\mathit{\text{Thickness}}}_{pre-\mathit{\text{implant}}}-{\mathit{\text{Thickness}}}_{\mathit{\text{explant}}}}{{\mathit{\text{Thickness}}}_{pre-\mathit{\text{implant}}}}\times 100$$ (3)

2.3.4. RNA expression

Aortic tissue processing and RNA isolation:

As previously mentioned, harvested aortic tissues were immediately rinsed with sterile saline, stored in RNAlater Stabilization Solution (Thermo Fisher Scientific) at 4 °C for at least 24 h, then transferred into TRIzol reagent (Thermo Fisher Scientific), and stored in a −80 °C freezer until further processing. For mRNA isolation, samples were homogenized using Biomasher II pestles and tubes (DWK, USA) using a handheld pestle motor while on ice. RNA was then immediately isolated from tissues using TRIzol reagent and chloroform extraction. Downstream mRNA was purified using Qiagen’s RNeasy Microkit PLUS with gDNA eliminator columns. The resulting mRNA concentration was quantified using a NanoDrop One (Thermo Fisher Scientific).

Reverse transcription and preamplification:

All RNA used in this study was normalized to 30 ng as input to synthesize cDNA using the Superscript IV VILO Master Mix kit (Invitrogen). The resulting cDNA was then preamplified using the TaqMan Preamplification Kit (Applied Biosystems) according to manufacturer recommendation using a calculated input of 18.1 ng per sample and 14 cycles of preamplification.

qRT-PCR assays and analysis:

All qRT-PCR assays were performed with a QuantStudio 3 (Applied Biosystems) using TaqMan single-plex FAM-MGB assays and TaqMan Gene Expression Master Mix with a 20 μL reaction volume using manufacturer recommended setting for quantitative relative expression. All qRT-PCR reactions were performed in the MicroAmp Optical 96-Well Reaction Plates (Applied Biosystems). For analysis, Rer1, Gapdh, and Hprt, were used as potential endogenous control genes and samples were normalized to the most stable - Rer1 (Supplemental Figure 5). All qRT-PCR data was analyzed using the Livak Method, wherein ΔΔCt values are calculated and reported as relative quantification values (RQ), established by the result of 2^ΔΔCt calculation [32]. Relative quantification values (i.e. relative fold change) are represented by Log₂(RQ) where the data are displaced linearly as means with the error bars representing standard deviation. All qRT-PCR assays were completed within the laboratory, by the authors, at the University of Pittsburgh.

2.4. Statistical analysis

All data are reported as mean ± standard deviation. For ultrasound data, a repeated measures, three-way ANOVA was used where the independent arms were biological sex, graft type, and time. Two-way ANOVA was used to analyze the relationship between graft type and sex for intimal thickening, graft degradation, gene expression, and collagen characteristics. If appropriate, Tukey HSD Post Hoc tests were performed.

Due to sectioning issues, there were insufficient slices of some samples (one CM Male, one Hypo Female) to complete all IF stains, making a standard two-way ANOVA inappropriate. Due to this, all IF data was fit using a generalized linear regression model. A two-sample t-test was used Post Hoc if significant interactions were found.

3. Results

3.1. Electrospun TEVG characterization

The mean compliance of native aorta in vitro was 0.0013 ± 0.0004 mmHg⁻¹ (Fig. 2A). The mean in vitro compliance of CM TEVGs and Hypo TEVGs were (0.0011 ± 0.0002 mmHg⁻¹) and (0.0007 ± 0.00009 mmHg⁻¹), respectively. Hypo TEVGs were found to be significantly less compliant than both native aorta (p = 0.027) and CM TEVGs (p = 0.0076). CM TEVGs were not significantly different from native aorta. The average wall thickness was 151 ± 8 μm and 180 ± 16 μm for CM and Hypo TEVGs, respectively (Fig. 2B). Individual layer thicknesses measured on samples processed for IHC can be seen in Supplemental Table 1.

Fig. 2.

In vitro characterization of TEVGs. (A) Measured in vitro compliance of native rat aorta (n = 6), CM TEVGs (n = 8) and Hypo TEVGs (n = 8). *p = 0.027 **p = 0.0076. (B) Mean thickness of electrospun TEVGs measured on a surgical scope.

3.2. In vivo monitoring of TEVGs

All CM and Hypo TEVGs were successfully implanted. Using a two sample t-test, no significant differences were found in age between males and females. Males were significantly larger than females, which is expected for rats of the same age (p = 0.019). This was not accompanied by a difference in aortic diameter. CM TEVGs, Hypo TEVGs, and shams remained patent through six months. Results shown are from post implant, month 1, and month 6 in vivo (Fig. 3). Full data sets with all timepoints can be found in Supplemental Figure 6.

Fig. 3.

Non-invasive in vivo monitoring of TEVGs captures shifts in TEVG mechanics. In vivo (A) compliance, (B) systolic inner diameter, (C) peak systolic velocity, and (D) end diastolic velocity measured via ultrasound over six months of remodeling. For in vivo compliance, a significant interaction (p = 0.0072) effect between graft type and time was found. Post-Hoc testing revealed significant differences between CM TEVGs and Hypo TEVGs after implant (p = 0.018) as well as differences between CM TEVGs after implant and all subsequent time points (only months 1 and 6 shown). A total of n = 12 (7M/5F) nonsurgical rats were used to generate baseline native aortic values. Red dashed lines represent the min and max of measured native aorta values.

Post implant compliance in CM TEVGs was significantly higher than Hypo TEVGs (p = 0.018) (Fig. 3A). The mean compliance of CM TEVGs was within native rat aortic range. Hypo TEVGs experienced no significant shifts in compliance over six months of remodeling. At month 1, CM TEVGs had significantly decreased compliance compared to their post-implant timepoint (p = 0.001) and were indistinguishable from Hypo TEVGs at all subsequent time points in their respective sexes. Compliance of native aorta proximal to the graft remained within native aortic range through all six months. The native aorta distal to the graft had decreased compliance post implant, but trended upwards (non-significant) throughout remodeling back towards baseline (Supplemental Figure 7).

There were no significant shifts in systolic inner diameter across all three areas scanned (proximal, graft, and distal). Interestingly, there was a non-significant upward trend in graft inner diameter in both female graft types (Fig. 3B). Graft implantation did affect aortic hemodynamics. Proximal peak systolic velocities (PSVs) were within the native range throughout remodeling and experienced no significant shifts. PSVs within the TEVG and distal to the graft were higher than their proximally measured counterparts across all four implant groups (Supplemental Figure 7). No significant shifts in graft PSV nor EDV were noted, however CM TEVGs displayed a trend towards lower PSVs at month 6 (Fig. 3C & D).

3.3. Tissue level analysis of TEVGs

H&E images were used to quantify the degree of intimal thickening within the graft and at the site of anastomosis. The average intimal thickness in CM males and Hypo males was 17 ± 8 μm and 16 ± 7 μm, respectively. Female rats had significantly increased thickness over both graft types (p = 0.015) with a mean intimal thickness of 66 ± 36 μm and 50 ± 45 μm in CM females and Hypo females, respectively (Fig. 4A & B). Females also appeared to have increased intimal thickening at the anastomotic site (CM: 29 ± 24 μm, Hypo: 39 ± 28 μm) compared to males (CM: 23 ± 4 μm, Hypo: 6 ± 4 μm) but the relationship is not significant (Fig. 4C & D). No sex specific differences were seen in the sham groups with average thicknesses of 14 ± 9 μm and 15 ± 11 μm for males and females, respectively (Fig. 4B).

Fig. 4.

Effect of graft type and sex on intimal thickening. (A) Representative H&E image of explanted grafts. Dashed lines denote inner and outer graft boundaries. Asterisks denote inner lumen. (B) Quantification of intimal thickening within the TEVG. *p = 0.015. (C) Representative H&E image of anastomotic site adjacent to TEVG implantation. Dashed line marks where intimal thickness was measured from. Asterisks denote inner lumen. (D) Quantification of intimal thickening at the anastomotic site. Sham female, sham male, and CM Male are n = 3 and remaining groups are n = 4. Asterisks denote inner lumen. Green dashed lines denote start of intimal layer. All scale bars = 100 μm.

Masson’s Trichrome staining demonstrated no appreciable muscle tissue in the implants. Muscle (representative of developed vascular tissue) is denoted by a deep red coloring with a striated appearance which can be seen in the shams of both sexes (Fig. 5A). Collagen was present abluminally in all implant and sham groups, represented by light blue staining. Females appear to have increased collagen density abluminally compared to males.

Fig. 5.

Analysis of structural makeup of explanted TEVGs. (A) Representative Masson’s Trichrome images of explanted TEVGs. Red, blue, and gray staining represent smooth muscle, collagen, and remaining graft material, respectively. Arrows point to nondegraded TEVG. Scale bar = 100 μm. (B) Representative Von Kossa images of explanted TEVGs. Pink and brown staining represent cytoplasm and Ca2+ crystals, respectively. Scale bar = 100 μm; (C) Representative fluorescent images of graft material remaining captured in the 692 nm (Cy5) channel where PEUU is autofluorescent. Scale bar = 200 μm; (D) Percentage of graft material degraded after 6 months of implantation calculated via MATLAB. *p = 0.042, +p = 0.022. Asterisks denote inner lumen.

Von Kossa staining was used to capture calcification within TEVGs. Dark brown staining of remaining graft material indicated that some of the remaining TEVG became calcified as TEVGs before implant are free from this coloration (Fig. 5B). Female explants had reduced brown coloration compared to males in both CM TEVGs and Hypo TEVGs suggesting decreased calcification density. Hypo TEVGs also had increased calcification compared to CM TEVGs in female rats. At six months, calcification was localized to the inner and middle layers of the remaining graft. Surrounding luminal and abluminal tissues did not stain positive for calcium. No calcific deposits were noted in either sham group.

Preimplantation thickness of CM TEVGs and Hypo TEVGs measured with IF imaging was 192 μm and 220 μm, respectively. The mean thickness of remaining CM TEVGs was 63 ± 29 μm and 38 ± 11 μm for males and females, respectively. Hypo TEVGS had a remaining mean thickness of 114 ± 22 μm and 59 ± 14 μm for males and females, respectively (Supplemental Figure 8). The outermost layer of all grafts (20:80 PEUU:Gelatin) was entirely degraded after 6 months of remodeling. In 6 out of 8 CM TEVGs and 5 out of 8 Hypo TEVGs, the second layer (80:20 PEUU:Gelatin) was partially or entirely degraded. The innermost layer (PESBUU-50) remained in all explants. Overall TEVG degradation was influenced by both graft type and sex (Fig. 5C). There was increased degradation in all CM TEVGs when compared to Hypo TEVGs (p = 0.022). The percent degradation of CM TEVGs and Hypo TEVGs averaged across both sexes was 73 ± 13% and 56 ± 16%, respectively. Additionally, females experienced significantly more degradation (72 ± 15%) than males (57 ± 16%) averaged over both graft types (p = 0.042) (Fig. 5D).

3.4. Assessing cell type in TEVG explants

Visually, there were more α-SMA positive cells in female TEVG groups which corresponds to the overall increased number of cells (Fig. 6A). Luminal cells positive for α-SMA clustered at the innermost edge of the intima. There was significantly decreased fluorescent density of α-SMA, a generalized VSMC marker, in the lumen of female TEVGs compared to male TEVGs in both graft types (p = 0.047) (Fig. 6B). There was also an increase in α-SMA in the graft matrix of Hypo females when compared to both CM females (p = 0.049) and Hypo males (p = 0.015). There was no change in total marker density across the four groups (Fig. 6C).

Fig. 6.

Differential regional expression of VSMC markers in explanted TEVGs. Top: Representative immunofluorescent images (A), regional quantification of marker density (B), and total marker density (C) for generic smooth muscle cell marker αSMA. *p = 0.047 +p = 0.015 #p = 0.049. Middle: Representative immunofluorescent images (D), regional quantification of marker density (E), and total marker density (F) for contractile smooth muscle cell marker MYH11 (CM Male n = 3). ^{^}p = 0.049. Bottom: Representative immunofluorescent images (G), regional quantification of marker density (H), and total marker density (I) for synthetic smooth muscle cell marker ICAM1 (CM Male n = 3). All hypo females have n = 3 due to slicing issues. Remaining samples n = 4 unless otherwise noted. TEVG, marker, and DAPI signal are represented by orange, green, and purple coloring, respectively. Asterisks denote inner lumen. All scale bars = 200 μm.

A similar trend emerged when looking at MYH11, a motor protein associated with contractile VSMCs. There were very few MYH11 positive cells in all four groups compared to the number of α-SMA positive cells seen (Fig. 6D). Sex had a significant main effect (p = 0.049) on MYH11 luminal density as female explants of both graft types had decreased density compared to male explants of both graft types (Fig. 6E). However, neither sex nor graft type altered MYH11 density within the graft, the abluminal region, or total marker density in any group (Fig. 6F). Finally, cells positive for ICAM-1, which is a marker for more synthetic VSMCs, were more abundant than MYH11 cells, but still decreased compared to the total number of α-SMA positive cells (Fig. 6G). ICAM-1 was also most abundant in the Hypo female groups. ICAM-1 density was not significantly affected by graft type nor sex in any regions of interest. (Fig. 6H & I).

The genetic expression of VSMC behavioral markers was also analyzed. The expression of each group was normalized to the sham surgeries for that specific sex, i.e. both CM male and Hypo male expression was normalized to the sham male group. There were differences in genetic profiles of male and female shams when compared to nonsurgical native aortas (Supplemental Figure 9). All shams had decreased VSMC markers when compared to native aortas as shams are an injury site with varied cell types and decreasing VSMC density compared to unperturbed tissue. Expression of αSMA related gene Acta2 was upregulated in both CM males (p = 0.000037) and Hypo males (p = 0.0056) when compared to their female counterparts. CM males had increased expression compared to Hypo males (p = 0.011) whereas CM females had the same level of expression as Hypo females (Fig. 7A). Expression of Myh11 (gene associated with MYH11) was upregulated in CM males (p = 0.00004) and Hypo males (p = 0.0028) compared to CM females and Hypo females, respectively (Fig. 7B). CM Males also had upregulated Myh11 expression compared to Hypo males (p = 0.015). No CM vs. Hypo specific shifts were seen in females. A third gene, Cnn1, associated with the contractile VSMC protein calponin was also measured. CM males had increased Cnn1 expression compared to Hypo males (p = 0.041) and CM females (p = 0.00055) (Fig. 7C). However, Hypo males did not have significantly greater expression compared to Hypo females. There were no graft specific trends in Cnn1 expression seen in females. When compared to nonsurgical aortas, hypo males experienced downregulation compared to CM males for all three VSMC genes (Acta2, Myh11, Cnn1) (Supplemental Figure 10). Females saw no shifts in gene expression between CM females and Hypo females. However, the only sex specific shift was seen in Myh11 expression with CM females having significant downregulation compared to CM males (p = 0.0434).

Fig. 7.

Shifts in VSMC related genes in response to graft type and sex. (A) Relative gene expression for generic VSMC gene Acta2 (*p = 0.011 **p = 0.00056 ***p = 0.000037), (B) contractile VSMC gene Myh11 (+p = 0.015 ++p = 0.0028 +++p = 0.00004), and (C) contractile VSMC gene Cnn1 (#p = 0.041 ##p = 0.00055). All calculations were done in comparison to sex specific shams (n = 4).

Male grafts appeared to be more endothelialized by the time of explant as revealed by CD31 positive cells (Fig. 8A). Though not significant, there was a trend towards decreased luminal CD31 density in both CM female and Hypo female explants compared to CM males and Hypo males (Fig. 8B). This was partially due to the amount of CD31 positive cells being similar across all groups, and the luminal area of CM females and Hypo females being greater due to increased intimal thickening. However, Female TEVGs had decreased (non-significant, p = 0.05) endothelial coverage luminally (percent of CD31 positive cells on the inner lumen) compared to males (Fig. 8C).

Fig. 8.

Assessment of endothelialization and remaining macrophages in explanted TEVGs. (A) Immunofluorescent staining, (B) regional quantification of marker density, and (C) percent luminal coverage of CD31 positive cells. (D) Immunofluorescent staining, (E) regional quantification of marker density (*p = 0.031 +p = 0.0012), and (F) total marker density (#p = 0.017) for pan macrophage marker CD68 (CM Male n = 3). Hypo females with n = 3 due to staining issues. Remaining samples n = 4 unless otherwise noted. Asterisks denote inner lumen. TEVG, marker, and DAPI signal are represented by orange, green, and purple coloring, respectively. All scale bars = 200 μm.

A pan macrophage marker (CD68) was used to get a basic immune profile. Almost all CD68 positive cells were located abluminally with a higher concentration in the tissue growth closest to the remaining graft material (Fig. 8D). There were no differences observed in the lumen and graft ROIs (Fig. 8E). Abluminally, CM males had an increased density when compared to CM females (p = 0.031). There was also an increase in CD68 positive cells in Hypo female implants compared to CM female implants abluminally (p = 0.0012) and in total fluorescent intensity (p = 0.017) (Fig. 8F).

3.5. Collagen deposition characteristics

Three dimensional images (multiphoton microscopy) successfully captured fibrillar collagen in explants (Fig. 9A). These collagen fibers were largely located abluminally, with some fibers luminally and with no deposition in the remaining graft material. Collagen fibers appeared qualitatively circumferential. There was also a “fracturing” effect of the graft material in post processing due to calcification. There was an increase in the total amount of SHG signal in all female explants compared to male explants (p = 0.0093), independent of being CM or Hypo TEVGs (Fig. 9B). Similarly, females of both graft types had an increased number of fibers per volume when compared to their male counterparts (p = 0.0065) (Fig. 9C). There were no changes in fiber tortuosity noted across the four groups (Fig. 9D). There was significant downregulation in the expression of collagen I (p = 0.0011) and collagen III genes (p = 0.001) (Col1a1 and Col3a1) in females when compared to males, (Fig. 9E & F). This relationship does not occur when normalized to native rat aorta instead of sex specific shams (Supplemental Figure 10).

Fig. 9.

Increased collagen deposition in female rats. (A) Maximum intensity projections of a z-stack of 2P images. Samples were excited at 792 nm and SHG and graft autofluorescence emissions were captured at 395 nm and 620 nm wavelengths, respectively. Asterisks denote inner lumen. Scale bar = 100 μm. Characterization of (B) fibrillar collagen via SHG signal intensity (**p = 0.0093), (C) fiber number (++p = 0.0065), and (D) fiber tortuosity (Bottom). (E) Relative expression of collagen 1 (##p = 0.0011) & (F) collagen 3 (^{^^}p = 0.001) related genes in reference to sex specific shams. n = 4.

4. Discussion

Clinical failure of small diameter vascular grafts is often driven by compliance mismatch between the graft and the host artery. Despite continuing efforts to create a compliance matched TEVGs, there has been insufficient characterization of both long-term remodeling of CM TEVG and the effect of sex on TEVG remodeling. The findings of this study suggest that biological sex affects key long-term remodeling outcomes of acellular, biopolymer based TEVGs as well as the way compliance matching is received. As has been shown previously by our group, compliance matching of TEVGs improved acute remodeling outcomes [15]. In this study, we demonstrated that compliance matching improves long-term remodeling outcomes even though the compliance of CM TEVGs becomes indistinguishable from Hypo TEVGs within the first month of remodeling (Fig. 2A). Despite this mechanical shift, CM TEVGs displayed increased genetic expression of VSMC markers (Acta2, Myh11, Cnn1) - markers for contractile VSMCs (Fig. 7). Additionally, our CM TEVGs proved to be robust over long-term remodeling. Increased degradation of CM TEVGs compared to Hypo TEVGs did not lead to loss of mechanical integrity, which is indicated by a lack of aneurysm formation (Figs. 3B & 5D). There was a decrease in total CD68 density in CM females compared to Hypo females, suggesting compliance improves chronic immune response in females (Fig. 8D–F). Biological sex also influenced key remodeling outcomes independent of graft composition, which addresses a large gap in current TEVG knowledge. Most notably, female rats had increased intimal thickening and decreased calcification compared to their male counterparts (Figs. 4B & 5B).

In this study we concluded that compliance matching influences VSMC density in female rats, with increased α-SMA staining in the graft matrix in Hypo females compared to CM females. Additionally, in male rats, compliance matching affected genetic expression of VSMC markers, with CM males having more contractile VSMC markers than Hypo males. These trends were likely driven by the relationship between VSMCs and scaffold mechanics. The presence of contractile VSMCs and scaffold stiffness are inversely related — stiff scaffolds lead to decreased contractile markers and often increases in proliferation [33]. On aligned electrospun fibers, increasing stiffness leads to decreased contractile VSMC markers (MYH11, calponin, desmin) and an increase in VSMC migration and ICAM-1 expression [34]. Despite this genetic increase in contractile markers in CM males, we had a relatively low expression of MYH11 in all our explants. This was likely due to the stiffening of our TEVGs over time. The loss of pulsatility in the early stages of remodeling (approximately 28 days) in our CM TEVGs and the lack of pulsatility from implant in our Hypo TEVGs likely prevented VSMC maturation despite cellular presence [35,36].

Female rats had decreased VSMC markers at the protein and gene level, with decreased luminal αSMA and luminal MYH11, Acta2, and Myh11 compared to males. This is consistent with the literature showing that female VSMCs have decreased plasticity, meaning their ability to phenotypically switch to a more migratory phenotype in response to vascular injury is dampened [37,38]. Additionally, the presence of estrogen inhibits VSMC proliferation in cells derived from mice, rats, and humans [39–41]. Interestingly, despite this decrease in luminal VSMC density, we also found there to be an increase in intimal hyperplasia in our female rats. This is unexpected as VSMCs are the main cell type involved in IH formation. At six months it is likely that we are capturing an advanced stage of IH, which is known to have decreased cellularity throughout the intimal tissue [42]. Regardless of IH maturity, it has been shown that female rats are more susceptible to IH formation than males in vascular grafting procedures. Rodent models studying arteriovenous fistulas (AVF) demonstrated that female vein grafts had decreased luminal diameters and increased graft failure due to intimal thickening [43,44]. It is interesting to note that despite this intimal formation, no decrease in graft IDs for females was captured via ultrasound.

Through Von Kossa staining we found that our constructs were susceptible to calcification, which has been reported in several long-term TEVG studies [45,46]. VSMCs are the largest contributor to vascular calcification, phenotypically switching towards an osteogenic phenotype in response to material implant and injury [47,48]. Greater calcification was also seen in males. Male VSMCs are more susceptible to osteogenic differentiation than female VSMCs in both atherosclerotic and wound healing environments driven by androgen activation [40, 49]. Additionally, there is a relationship between degradation and calcification. Scaffolds with faster degrading electrospun outer layers made from a hybrid poly(L-lactic acid) (PLA) and polyglycolic acid have decreased calcific deposits compared to their slower degrading, PLA only counterparts [50]. In our study, female scaffolds had increased degradation and decreased calcification for both TEVGs formulations.

The decreased degradation we saw in Hypo TEVGs can partially be explained by the increased thickness of the highly synthetic second layer (80:20 PEUU:Gelatin) of Hypo TEVGs. PEUU not supplemented with gelatin is fully biodegradable on the scale of 2–3 years in a rat model which is beyond the scope of this study [51]. This layer may also contribute to the unintended increase in thickness we see in Hypo TEVGs versus CM TEVGs. In both TEVG designs, the high gelatin 3rd layer (20:80 PEUU:Gelatin) was entirely degraded, which was expected as gelatin crosslinked with genipin is almost entirely degraded within 1 month [15,52]. Conversely, our innermost layer made from PESBUU-50 is present in all explants with little to no degradation. It is also important to note that increased sulfobetaine (SB), the zwitterionic compound in PESBUU-50, in a PEUU matrix decreases the degradation rates, almost tripling the in vitro lipase degradation time compared to PEUU alone [26].

In both of our TEVG designs, female explants had increased degradation when compared to males of the same graft design. This is likely due to sex specific differences in immune response to biomaterial implantation in the vasculature, as sex is known to influence immune response and resolution of chronic wounds [53]. Generally, it is believed that estrogen has vasoprotective effects during vascular injury by reducing monocyte chemotaxis, altering macrophage phenotype, and maintaining phagocytosis, overall driving a faster resolution of immune responses [54–56]. Males are also thought to have increased M1 versus M2 macrophage presence, increasing the likelihood of an ongoing inflammatory response [57]. This is consistent with the decreased CD68 we saw in CM females versus CM males. This reduction in CD68 positive cells in CM females may also indicate an earlier resolution of inflammation. Macrophage behavior is also influenced by substrate stiffness, with stiffer substrates increasing migration of human monocyte derived macrophages and increasing anti-inflammatory responses [58,59]. This is consistent with our female population, where CD68 levels were elevated in Hypo TEVGs. Regardless of intensity, CD68 staining showed a residual macrophage presence in all grafts showing ongoing inflammation. Long term TEVG studies often have remaining biomaterial that leads to “low level” chronic inflammation at the graft interface [60].

Immunofluorescent staining also revealed a decrease in endothelial cell (EC) coverage in female explants. In opposition to our results, female ECs have often been reported to be more migratory and proliferative both basally and during wound healing in both human and rat derived cells [61,62]. However, these in vitro studies did not culture their female ECs in media containing any form of estrogen, which does not represent in vivo conditions. Another possible contributor to this discrepancy is the failure of our TEVGs to recapitulate native hemodynamics. It has been shown that female ECs are more reactive to changes in shear stress than male ECs. Lorenz et al. reported that for female ECs to downregulate pro-atherogenic markers high shear stress is required [63]. Our TEVGs may be prohibiting female ECs from responding to injury as expected.

Finally, multiphoton imaging revealed that female implants had increased fibrillar collagen compared to males independent of graft type after six months of remodeling. Injured cardiovascular tissues in females have profibrotic tendencies, which we see in the remodeling of our TEVGs For example, in a mouse model of AVF, female mice have increased fibrosis and synthetic VSMC markers after 28 days of vein graft remodeling as well as the pro fibrotic marker TGFβ-1 [43]. This is also true in atherosclerotic plaques where female plaques are characterized with a thick, collagen rich cap and male plaques are characterized by large calcific deposits with thin caps [64]. However, this increase in collagen did not coincide with an increase in collagen production genes Col1a1 and Col1a3 in female rats. These genes were significantly downregulated in females versus males, suggesting that further collagen production has been slowed. Collagen deposition may have been a response to the instability caused by rapid degradation of the TEVG in an attempt to maintain tissue integrity.

There were several limitations in this study. Most notably, there was little to no cellular infiltration into the remaining matrix of our TEVGs. The selection of polyurethane based inner layers as well as the dense nature of electrospun matrices likely contributed to this. While improvements can be made, we believe that our choice of polymers was justified. PEUU allows for mechanical support, compliance matching, and has established clinical potential [25]. Using PESBUU as a nonthrombogenic agent prevents failure via clot formation, is novel, and contributed to our high patency rate (16/16 grafts). The amount of polymer remaining was also in line with other rat implantations of polyurethane based TEVGs [65]. There has also been extensive work in altering the rates of PEUU degradation to make it more suitable for TEVG applications [66,67]. Electrospun scaffolds are often densely packed with small pores, preventing cellular infiltration into the scaffold [68]. We did not optimize our electrospinning settings to prevent this in our TEVGs. An abundance of polymer may have also prevented any mature elastin from being produced in or around our TEVGs as no elastin signal was detected with multiphoton autofluorescence. TEVGs that report successful production of elastin often utilize natural biological materials such as small intestinal submucosa and fibrin [69,70].

This lack of VSMC infiltration likely contributed to the calcification we see in our grafts. Calcification and subsequent stiffening has been a challenge in long-term TEVG studies and polyurethanes carry known risks of calcification [71,72]. However, electrospinning parameters can be adapted to modulate porosity and these changes in pore size have been shown to influence cellular infiltration as well as modulate VSMC phenotype, macrophage response, endothelialization, and calcification [46,73–75].

In our characterization of the VSMC population, no co-staining was done confirm to colocalization of contractile/synthetic markers with αSMA. This could be confounding since we chose ICAM-1 as our marker for a synthetic phenotype. In the vasculature, ICAM-1 is also expressed at low levels in endothelial cells and some immune cells, such as macrophages in the ablumen. Co-staining the future would allow for more accurate characterization of VSMCs as well as other ICAM-1 positive cell types.

Our TEVGs were compliance matched using a trial-and-error approach. This contrasts with prior work where computational optimization led to increased accuracy and repeatability in mechanical design and biomanufacturing [15–17]. Our CM TEVGs also became hypocompliant within the first month or remodeling, which limited our ability to study the effects of compliance maintenance on long term TEVG remodeling. Compliance matching outcomes should be interpreted with this in mind. Another drawback to our current TEVG design was the use of in vitro native aortic compliance values as our target compliance for graft development. We found that there was a small decrease in the range used for native aortic compliance that we measured in vitro versus in vivo. We hypothesize this was due to a mismatch in axial tension between the benchtop and during aortic implant, which was not controlled for in vivo. This is a confounding factor, as axial tension is known to influence vascular remodeling [76].

The use of a rat model to study vascular tissue engineering has some notable drawbacks despite their widespread use. Rat hemodynamics are different from human hemodynamics and rats have unique clotting pathways making it difficult to recreate failure via thrombosis [77]. Rodents also have altered rates of vascular aging and material degradation and all rats used in this study would be considered adolescent, not aged [78,79]. Further characterization using a more physiologically similar large animal model (e.g., sheep), should be used to validate the findings of any rodent study.

Finally, all sex specific differences reported are only representative of a pre-menopausal population. Most female CABG patients are postmenopausal, and changes in VSMC behavior due to estrogen loss are often cited as a driver of female CABG graft failure [80]. Therefore, future studies should incorporate a post-menopausal model for TEVG implantation. However, there are several forms of vascular repair that take place in pre-menopausal women such as endovascular aortic repair, large artery grafting (coarctation repair), and response to traumatic injury. The results of this study still provide valuable information on how cells in pre-menopausal vasculature respond to and remodel during vascular injury.

To the authors knowledge, this is the first study to evaluate the effects of sex on compliance matching of TEVGs following long-term remodeling. Our compliance matched TEVGs were successful in modulating residual immune response (females), increasing contractile VSMC genetic markers without formation of IH (males), and having increased biopolymer degradation without loss of mechanical integrity (both sexes). These are all indications of positive in vivo TEVG remodeling and restate the importance of compliance matching in small diameter vascular grafting procedures, even if compliance is not maintained. Finally, when developing new strategies for improving TEVG outcomes, researchers often focus on combating intimal hyperplasia, vascular calcification, and chronic immune response — variables clearly influenced by sex during CM TEVG implantation. Despite this, the influence of sex is understudied or overlooked, namely in small animal models. We anticipate that our emphasis on characterizing sex specific differences will help generate valuable insight into an understudied population with the long-term goal improving clinical outcomes for at risk patients.

Appendix A. Supplementary data

Supplementary material related to this article can be found online at https://doi.org/10.1016/j.actbio.2026.06.042.