Development and in vivo validation of tissue-engineered, small-diameter vascular grafts from decellularized aortae of fetal pigs and canine vascular endothelial cells

Fetal pigs of 100-day gestational age were delivered by cesarean section. Adult mongrel dogs were used as both vascular EC donors and graft recipients. All animals received humane care in compliance with the Guide for Care and Use of Laboratory Animals published by the National Institutes of Health (NIH publication No. 85-23, revised 1996).

Pregnant sows were intravenously anesthetized with 1% pentobarbital sodium (10 mg/kg) and delivered by cesarean section. After removal of fetuses, sows were euthanized by excessive anesthesia. Fetal pigs with crown-rump lengths of 25–30 cm were selected and 5-cm sections of aorta were excised under sterile conditions. These specimens werecryopreserved immediately transported to the laboratory with a warm ischemia time of <30 min [14].

Fetal pig aortae with an outer diameter of 4 mm and a length of 6 cm were selected. After removal of surrounding tissues under sterile condition, the blood vessels were washed with sterile saline several times to remove residual blood. DAFPs were prepared by continuous enzymatic digestion using trypsin, DNase, and RNase as previously described [15] with few modifications. Aortae were first digested with a solution containing 0.1% trypsin/0.02% EDTA in phosphate-buffered saline (PBS) (without Ca²⁺ and Mg²⁺) for 36 h (all reagents from Sigma, St. Louis, Mo., USA), and the solution was changed every 12 h. The aortae were then decellularized with 20 μg/mL RNase and 200 μg/mL DNase (Boehringer, Mannheim, Germany) for 4 h at 37 °C under a humidified atmosphere of 5% CO₂ and 95% air with constant gentle shaking. DAFPs were then washed with sterile PBS several times to remove the digestion residue. Following DAFP preparation, approximately 3 mm of the tissue at the end of each decellularized specimen was taken and cut into two semi-rings. Half of the tissue was examined following conventional hematoxylin–eosin (HE) staining, and the other half was used for DNA quantification [16]. Finally, DAFPs were subjected to vacuum freeze-drying and ethylene oxide sterilization.

Seven domestic dogs aged 6 months and weighing 25 kg were used. Intravenous administration of 1% pentobarbital sodium (10 mg/kg) was used for anesthesia. Approximately 10 cm of the left external jugular vein was excised under aseptic condition as previously described [17]. Primary vascular ECs were obtained by enzymatic digestion as previously described [18]. In brief, the outer membrane layer of the vessel was removed and the vessel was cut longitudinally to form a sheet. The intima layer was flattened on the culture dish. A solution of 0.2% collagenase A (Boehringer Mannheim, Germany) in PBS with Ca²⁺ and Mg²⁺ was injected beneath the inner membrane surface and subsequent digestion performed in an incubator for 15 min. The digested EC suspension was collected and centrifuged at 1000 rpm for 5 min. The EC pellet was resuspended and cultured in medium 199 (GIBCO) supplemented with 10% fetal bovine serum, 100 units/mL penicillin (Sigma), 100 mg/mL streptomycin (Sigma), 0.25 mg/mL amphotericin B (Sigma), 5 ng/mL endothelial growth factor (Boehringer Mannheim) and 1% L-glutamine (Sigma). Cells were passaged using trypsin digestion and the fourth generation was used as seed cells (the number of vascular seed cells in the fourth generation was sufficient for seeding purposes, and ECs still possessed abundant proliferative capacity at this passage). Cells were confirmed to be vascular ECs by morphology as well as by immunohistochemistry and immunofluorescence to detect the endothelial marker von Willebrand factor (vWF) as previously described [19].

Vacuum freeze-dried DAFP material was soaked in sterile PBS for 24 h. Vascular ECs were seeded onto DAFPs by rotational precipitation. Briefly, ECs were trypsinized to form a cell suspension and the EC concentration was adjusted to 3 × 10⁶ cells/mL. Both ends of DAFP tubular stents were clamped and the cell suspension was injected into the arterial lumen using a syringe. The arteries were then rotated 120° after standing for 10 min and seeding was completed after a total rotation of 360°. The grafts were transferred to a vascular bioreactor system for 7 days of dynamic culture following 3 days of static culture. A vascular bioreactor system was designed and produced by our research group as depicted in Fig. 1. This bioreactor system was driven by a peristaltic pump [20] and specimens were placed in the processing chamber (Fig. 1b). The liquid flow in the system was gradually increased by adjusting the speed of the peristaltic pump so that the perfusion rate increased from 20 to 60 mL/min (the perfusion rate was 20 mL/min on the first day of dynamic culture and increased by 10 mL/min daily to 60 mL/min on the fifth day of dynamic culture, at which point the perfusion rate was maintained at 60 mL/min through to the seventh day). The static pressure due to height differences of the culture solutions was 10 mmHg, and the dynamic pressure generated by the peristaltic pump was 60 mmHg. After culture completion, the fixed ends (lacking ECs) were removed and then the grafts were used for surgical implantation into the same dogs from which ECs had been derived.

The same seven dogs from which left external jugular veins were harvested to obtain ECs received general anesthesia. Routine preoperative preparation was followed and the right common carotid artery was exposed. A location of the carotid artery with an external diameter of approximately 4 mm at diastole was selected as the site for implantation of the tissue-engineered vascular grafts. Vascular clamps were placed on both sides of the selected position, and the artery (4 mm in diameter) was transected. The tissue-engineered vascular grafts were used to bridge the common carotid artery using microsurgical techniques. The incisions were closed, and routine intramuscular injection of heparin as well as intravenous injection of antibiotics were applied postoperatively. Doppler ultrasonography was conducted at 1, 3, and 6 months and contrast-enhanced CT at 6 months, following implantation. Vascular grafts were sampled for histological and electron microscopic examination after the dogs were sacrificed at 6 months post-implantation.

All experimental animals received general anesthesia prior to each procedure and neck skin preparation prior to ultrasonic examination. Bilateral common carotid artery ultrasound was performed using a Doppler ultrasound device. Animals were injected with contrast medium prior to three-dimensional CT angiography.

DAFPs and sampled vascular grafts were fixed in 10% neutral-buffered formalin solution, embedded in paraffin, transversely sectioned (5 μm) and stained using HE for histological assessment of general morphology.

For transmission electron microscopy (TEM), specimens were fixed in 2.5% glutaraldehyde solution, rinsed with 0.1 M phosphate buffer, and then fixed in 1% osmic acid for 2 h. Subsequent rinsing in double-distilled water was followed by dehydration in a graded series of water–acetone solution. Specimens were then impregnated with Epon 812 resin, embedded, and polymerized. Semi-thin sections were counterstained with azure II and basic fuchsin and examined with light microscopy to determine orientation. Ultra-thin sections were stained with uranyl acetate and lead citrate and then examined with a Zeiss EM10 electron microscope (Zeiss, Oberkochen, Germany).

Immunohistochemistry and immunofluorescence showed positive vWF staining in most cells (Fig. 2).

Histological staining of fetal porcine aortae revealed a typical three-layer structure of the arteries (Fig. 3A-a), and SEM demonstrated a complete layer of ECs with a cobblestone phenotype (Fig. 3B-a, b). The DNA content of DAFPs was <0.1%. Histological staining of DAFPs showed that the extracellular matrix remained intact; however, no nuclei or intact cells were present (Fig. 3A-b). In addition, SEM examination showed that after decellularization, endogenous ECs were completely removed and the structure of the inner elastic membrane was clearly visible (Fig. 3B-c, d).

After 1 week of dynamic culture in the custom-made dynamic vascular bioreactor system, vascular grafts were successfully constructed using DAFPs and canine vascular ECs (Fig. 4). The vascular grafts possessed a complete, intact EC layer. Histological findings showed that a continuous cell layer was present on the inner surface of the grafts (Fig. 3A-c), and SEM revealed that ECs were tightly connected, forming a continuous cell layer that completely covered the inner elastic membrane (Fig. 3B-e, f).

Seven dogs were numbered and all were subjected to extensive postoperative evaluation (Table 1). Vascular diameter of the grafts matched that of the common carotid artery into which they had been implanted (Fig. 4). Following implantation, vascular grafts exhibited good patency and no obvious thrombi were found attached to the vascular wall, as assessed by Doppler ultrasound and enhanced CT examination (Fig. 5). Six months after implantation, the grafts showed no obvious stenosis and expansion. Additionally, dynamic Doppler sonography revealed good graft compliance, and vascular grafts underwent spontaneous rhythmic vasodilation and contraction in synchrony with heartbeat.

Sampling results at 6 months post-transplantation showed that the tissue-engineered vascular grafts had smooth surface and no thrombi. As depicted in Fig. 6, histological staining showed that the grafts had been remodeled within the animals, and a continuous inner cell layer was identified. The tunica media was structurally dense and similar to natural arterial structure, and cells had evenly infiltrated and were identified as muscular fibroblasts by immunohistochemistry (smooth muscle actin expression was used as a classical marker for myofibroblasts). In addition, TEM examination of the inner cells showed that their surfaces were rich in finger, spherical, and villous protrusions and that they contained abundant cytoplasmic organelles, indicating that they were indeed ECs (Fig. 7). Moreover, Weibel–Palade bodies, which are membrane-enclosed rod-shaped organelles specifically found in ECs, were also identified.