Cellular Responses to Flow Diverters in a Tissue-engineered Aneurysm Model

Cell Culture

The human aortic endothelial cells (ECs) and human aortic smooth muscle cells (SMCs) were purchased from the American Type Culture Collection (ATCC, Manassas, USA). The ECs were cultured in Endothelial Cell Growth medium (Vascular Cell Basal Medium supplemented with Microvascular Endothelial Cell Growth Kit – VEGF, ATCC^® PCS-100-041, Manassas, USA). The SMCs were cultured in Vascular Cell Basal Medium supplemented with Vascular Smooth Muscle Cell Growth Kit (ATCC^® PCS-100-042, Manassas, USA). All the cells were maintained at 37°C, 95% humidity, and 5% CO₂ with medium changed every 48-72 hours.

Aneurysm-like scaffold creation

Polycaprolactone (PCL) (Mn80,000, Sigma-Aldrich) solution (15% wt./v) was prepared in a mixture of glacial acetic acid and acetone (7:3) overnight in a water bath at 70°C. The PCL solution was then placed into a 12 ml syringe and dispensed at a flow rate of 0.02 ml/min through the 20-gauge blunt-tip needle. Aneurysm geometries were created using molded aluminum foil wrapped around a straight stainless-steel mandrel. The mandrel was positioned 10 cm from the needle tip and was rotated at 250 RPM. The voltage between the needle tip and the rotating mandrel was set to 13 kV. Electrospinning was performed for 2.9-3.6 hours until a saccular aneurysm-like scaffold was formed. The resulting tubular aneurysm-like scaffold (the aBVM) was then removed from the mandrel.

Scaffold conditioning and cell seeding

The aBVMs were prepared as described earlier.¹¹ The scaffolds were sterilized with 70% ethanol and pre-conditioned with cell culture media prior to seeding the cells. Bioreactor chambers were assembled as described previously, using a flow rate-controlled peristaltic pump (ISM932, ISMATEC), with the aBVM scaffolds secured with sterile suture loops. The SMCs were seeded inside the scaffold at a density of 1 × 10⁶ cells/cm², distributing the cells evenly by rotating the chambers 180° along the long axis every 30 min while incubating under static conditions. The ECs were seeded the following day, following the same procedure. Mixed cell culture media (1:3 ratio of SMC media to EC media) was placed in the reservoir tube to maintain the cells. The aBVMs were cultured for 24 hours at 1.00 ml/min flow rate of cell culture media.

FD deployment

Prior to FD implantation, the bioreactor chamber containing aBVMs were disconnected from tubing connected to the peristaltic pump and transferred from the incubator to the fluoroscopy room under aseptic conditions. The FDs (Pipeline Flex Embolization Device, Medtronic Inc., Irvine, California, USA) were aseptically deployed into each aBVM following standard methods. Briefly, a microcatheter (Marksman, Medtronic Inc., Irvine, California, USA) was placed into the “distal” end of the aBVMs parent vessel with guidance of a micro guidewire (Transcend; Boston Scientific, Marlborough, Massachusetts, USA). Iodinated contrast mixed with saline (~1 cc; 1:1 ratio) was injected to create a roadmap. The FD was advanced beyond the distal end of the microcatheter and deployed slowly across the aneurysm neck under fluoroscopy. Digital Subtraciton Angiography (DSA) was performed through the microcatheter to confirm the device location before the microcatheter was removed. The residual contrast was flushed out by the injection of saline through aBVMs containing implanted FDs. Then the bioreactor chamber with aBVMs+FDs were returned to the incubator, attached to the tubing connected to the peristaltic pump, and cultivated for a pre-specified time. The pump rate was increased form 1.00 ml/min to 5.00 ml/min after 24 hours.

Whole-Mount en Face Immunostaining

The FD-bearing aBVM samples were processed for whole-mount staining as described previously^{5, 12}. Briefly, 7, 14 and 28 days after device implantation, the FD–implanted aBVMs were harvested. Each was washed with 1x Tris-Buffered Saline (TBS) and fixed with 10% neutral buffer formalin (NBF) for at least 24 hours. The flow diverter bearing aBVMs were longitudinally cut using a scissor with extra fine blade (Extra Fine Bonn, No.14084-08, F.S.T) at the opposite to the aneurysm neck to expose the lumen. Tissue coverage on the device struts that crossed over the neck orifice and sat within the parent vessel was grossly evaluated. Immunostaining was performed using antibodies against CD31 (Dako, Carpentaria, California) and smooth muscle actin (SMA) (Dako, Carpentaria, California). Specific binding was visualized using fluorescently-conjugated secondary antibodies (Cy3- or Alexa Flour 488 conjugated donkey anti-mouse IgG; Jackson Immuno Research, West Groove, PA), with DAPI (Invitrogen) to stain the nucleus. Stained specimens were imaged on a confocal microscope (Olympus BX61, Japan). Cell/tissue coverage was evaluated and scored in the inner half as well as outer half of the neck by a trained pathologist using a 0-4 semi-quantitative scale system: 0) No cells/tissue visible; 1) cells present on device struts, but not yet confluent; no cells/tissue within device pores; 2) confluent cell growth on device struts with no gap between cells; no cells/tissue within device pores; 3) confluent cell growth on device struts with no gap between cells; some cell/tissue growth within pores; 4) struts and pores completely covered with cells/tissue.

Geometries of aBVMs

An untreated aBVM for each batch was longitudinally cut open to expose the luminal side and stained with CD31, SMA and DAPI to confirm the cellular confluency of the cell coverage in the parent vessel before proceeding to FD implantation (Figure 1). The FDs were implanted in all aBVMs (n=9) without incident. The mean diameter of aBVM parent vessels was 4.6 mm. The mean aneurysm neck diameter, and height were 5.9 mm, and 3.2 mm, respectively. Immediate post-device placement DSA visualized device coverage over the neck (Figure 2).

Morphologic findings

Cell/tissue coverage scoring results are presented in Table 1. At 7 days, 2 of 3 samples had the device completely covering the neck; the remaining one sample showed the majority of the device crossed over the neck orifice. The device within the parent vessel was partially endothelialized. Most of the device struts covering the center of the aneurysm neck were devoid of cells. Patchy interrupted tissue islands were seen on device struts and within a few of the device pores over the peripheral neck near the parent vessel. Confocal microscopy revealed these tissue islands were composed of endothelial and smooth muscle cells (CD31+/SMA+). One sample showed scattered cells on device struts at both, over the neck or within the parent vessel; these cells were also positive for CD31 and SMA. The individual and mean score for this time point is listed in Table 1.

At 14 days, similar to 7 days, 2 of 3 samples had the device completely covering the neck, the FD in one sample did not fully bridge the entire neck. The device within the parent vessel was partially endothelialized. The device struts in the outer half region of the neck were partially covered with translucent tissue (Figure 2), while the device struts in the inner half of the neck were grossly bare but were actually wrapped with one single layer of non-confluent cells (Figure 2). The cells that either covered the struts or filled in the pores between struts were positive for both CD31 and SMA stains (Figure 2), they also continued up to the cells in the parent vessels (Figure 2) suggesting that the cells from the parent vessel migrated to the neck area of non-confluent CD31, or SMA positive cells (Figure 3).

At 28 days, 1 of 3 samples had the device completely cross over the neck. The remaining two samples had device covering the majority of the neck orifice. Macroscopy demonstrated the entire parent vessel was covered with a translucent neointimal layer in two aBVMs; these two samples also had confluent tissue patches and/or interrupted tissue islands covering the device struts, as well as filling the pores between the struts over the neck (Figure 3). Confocal microscopy showed the luminal surface of the device struts located within the parent vessel was lined with a single layer of CD31+ cells. These confluent CD31+ cells lined the tissue patches/islands over the neck as well (Figure 4). SMA+ cells were identified underneath CD31+ cells at these areas. Overall, a higher degree of cellular coverage was observed on the struts and pores at the neck at 28 days compared to both 7 and 14 days (Table 1). The tissue coverage was not visible in one of the samples. It appeared that the cells did not grow well in that aBVM.

Limitations

This study had several limitations: 1) the sample size was small; 2) the “aneurysm size” utilized in this study was relatively small compared to human intracranial aneurysms; 3) this study was limited to steady flow, low-pressure conditions, instead of physiologic flow rates and pressure, which will be applied in future studies to mimic the blood vessel hemodynamics; 4) a few cells covering the device at the neck of aBVMs were positive for both CD31 and SMA indicating the potential transition of endothelial cells to mesenchymal phenotypes in response to the local shear stress, which could be alleviated by closed-loop flow models¹⁸; 5) a few of the aBVMs did not have complete device coverage across the neck due to device migration, which is commonly observed in clinical settings. The use of oversized devices could minimize the device migration; 6) as with most in vitro models, this tissue-engineered model does not fully recapitulate the in vivo environment, especially the tissue responses to endovascular implants. Future work will include co-treatment with factors involved in endovascularly treated aneurysm healing.