Abstract
The study involved the electrospinning of the copolymer poly(L-lactide-co-ε-caprolactone) (PLCL) into tubular grafts. The subsequent material characterization, including micro-computed tomography analysis, revealed a level of porosity of around 70%, with pore sizes of 9.34 ± 0.19 μm and fiber diameters of 5.58 ± 0.10 μm. Unlike fibrous polycaprolactone, the electrospun PLCL copolymer promoted fibroblast and endothelial cell adhesion and proliferation in vitro. Moreover, the regeneration of the vessel wall was detected following implantation and, after six months, the endothelialization of the lumen and the infiltration of arranged smooth muscle cells producing collagen was observed. However, the degradation rate was found to be accelerated in the rabbit animal model. The study was conducted under conditions that reflected the clinical requirements—the prostheses were sutured in the end-to-side fashion and the long-term end point of prosthesis healing was assessed. The regeneration of the vessel wall in terms of endothelialization, smooth cell infiltration and the presence of collagen fibers was observed after six months in vivo. A part of the grafts failed due to the rapid degradation rate of the PLCL copolymer.
Export citation and abstract BibTeX RIS
Corrections were made to this article on 14 January 2019. An author name was amended.
Introduction
Extensive research has been conducted worldwide on the materials and methods to be employed for the fabrication of ideal vascular grafts, a process that is particularly challenging due to the large number of requirements such as the appropriate mechanical properties, surface properties with respect to the endothelialization of the graft lumen, together with non-thrombogenicity, hemocompatibility, non-immunogenicity etc. Electrospun materials resembling the natural extracellular matrix have been found to fulfill the majority of these requirements; thus, the further investigation of such materials would seem to present a promising way forward in terms of the development of vascular grafts.
The last few years have witnessed huge progress concerning the application of biodegradable materials that enhance the healing process. Polyesters such as polycaprolactone (PCL), polylactide (PLA) and their respective copolymers (PLCL) have found uses in a broad range of tissue engineering applications. While PCL and PLA differ in terms of their surface properties and degradation rate, it is possible to tailor the properties of the copolymers of these two polymers via the composition of the final copolymer. The long-term evaluation of PCL-based electrospun vascular grafts revealed calcification on the regenerated vessel wall; therefore, it is necessary to examine the potential of other polymers for use in this application [1]. The PLCL copolymer is an elastomeric material developed for the construction of vascular grafts [2–4].
Mechanical properties make up one of the most important requirements with concern to the effective replacement of blood vessels. Studies devoted to the mechanical assessment of electrospun biodegradable vascular grafts in general have previously been conducted by Enis et al, Johnson et al, Limbert et al, Xie et al [5–8] while the mechanical properties of the material used in our study have been investigated by Johnson et al including the comparison of biodegradable polyesters such as PCL, PLA, polyglycolide and the copolymers thereof in the form of electrospun tubular structures and commercially-available grafts and human vessels. The polymers were dissolved in hexafluoroisopropanol and electrospun on a rotating mandrel to a final wall thickness of 650 ± 65 μm, and it was determined that the PLCL electrospun copolymer exhibited the highest level of compliance of all the tested grafts, attaining an average value of 8.2% mm Hg, which is higher than that of the human carotid artery with a compliance value of 5.4% mm Hg. Moreover, the burst pressure of a prosthesis made from PLCL was determined at 2.5 MPa, which is substantially higher than that of the coronary artery (around 0.4 MPa) [6].
With concern to vascular grafts, the optimal fiber diameter (and thus the resulting porosity and pore size) is crucial for cell infiltration following implantation. The morphological analysis of vascular grafts is usually performed by means of scanning electron microscopy (SEM) and image analysis tools; however, these techniques do not truly reflect the 3D structure of tubular grafts. New techniques such as micro-computed tomography (micro-CT) provide for the evaluation of 2D and 3D structural parameters including the basic parameters (e.g. volume, surface, number of objects), advanced parameters (e.g. structure thickness and separation based on the 3D sphere-fitting algorithm [9, 10], porosity) and stereology parameters [11–13]. The results obtained are based on the unique non-destructive 3D analysis of a selected specimen which allows the direct measurement of selected parameters.
The assessment of vascular grafts in vivo has been addressed by a number of research groups (see table 1); however, the results published to date suffer from certain limitations such as the animal models used, the mode of implantation and the length of the period of study. Most of the research was conducted on rats following implantation in the form of a bypass of the aorta with high blood flow. The suturing technique has been found to exert a huge impact on the healing of vascular grafts and end-to-side anastomosis appears to be more suitable in terms of further clinical use despite the turbulent flow in such anastomosis potentially creating problems with respect to vascular graft closure. In short, the detailed assessment of these materials requires the conducting of long-term studies. De Valence et al [1] and Li et al [14] investigated the performance of PCL grafts after 18 months in rats making them, to the best of our knowledge, the longest studies of their kind conducted to date. Other studies, especially those conducted employing larger animal models such as rabbits, pigs, dogs or sheep, have been limited to just weeks or a few months at most.
Table 1. Overview of in vivo-tested vascular grafts fabricated from electrospun polycaprolactone, polylactide and their respective copolymers. PDS = polydioxanone, VEGF = vascular endothelia growth factor, HGFI = class I hydrophobin, RGD = arginylglycylaspartic acid.
| Animal model, implantation area | Type of implantation | Graft type | Investigated time points | References |
|---|---|---|---|---|
| Mouse aorta | End-to-end | Electrospun PLA coated with PLCL | 4, 8, 12 months | Tara 2014 [3] |
| Rat aorta | End-to-end | Electrospun PCL | 1.5, 3, 6, 12, 18 months | De Valence 2012 [1] |
| Rat aorta | End-to-end | Electrospun PCL-plasma treated | 3 weeks | De Valence 2013 [15] |
| Rat aorta | End-to-end | Wet spun PCL + electrospun PCL | 3, 18 months | Li 2018 [14] |
| Rat aorta | End-to-end | Electrospun PCL | 7, 14, 28, 100 days | Wang 2014 [16] |
| Rat aorta | End-to-end | Electrospun PCL, PCL + PDS, PDS | 4, 12 weeks | Pan 2017 [17] |
| Rat aorta | End-to-end | Electrospun PCL | 3, 6, 12, 18, 24 weeks | Pektok 2008 [18] |
| Rat aorta | End-to-end | Electrospun PCL | 3, 12 months | Yang 2016 [19] |
| Rat aorta | End-to-end | Electrospun PCL-VEGF, HGFI modified | 1 month | Wang 2017 [20] |
| Rabbit carotid artery | End-to-end | Electrospun PCL-RGD functionalized | 2, 4 weeks | Zheng 2012 [21] |
| Rabbit carotid artery | End-to-end | Electrospun PCL | 4, 12 weeks | Wang 2016 [22] |
| Rabbit aortoiliac bypass | End-to-side | Electrospun PCL-collagen | 4 weeks | Tillman 2009 [23] |
| Beagle dogs femoral artery | End-to-end | Electrospun PLC-collagen-chitosan | 12 weeks | Wu 2015 [4] |
| Pigs carotid artery | End-to-end | Electrospun PCL | 4 weeks | Mrówczyński 2014 [24] |
However, none of the studies presented above dealt with challenging end-to-side implantation over the long term. Therefore, our study focused on the development, characterization and biological assessment of a biodegradable graft implanted end-to-side in rabbits over a period of six months.
Methods
Materials and electrospinning solution preparation
The electrospinning solution was prepared from a GMP grade copolymer of L-lactide and ε-caprolactone in a 70/30 molar ratio suitable for medical device applications (PLCL, Corbion n.v., Netherlands). Polymeric granules were dissolved in a solvent system composed of chloroform/ethanol/acetic acid in the ratio 8/1/1 v/v/v (Penta s.r.o., Czech Republic) with a final polymer concentration of 10 wt%.
The cell interactions of the newly-developed PLCL electrospun material were compared with electrospun pure polycaprolactone (PCL, average Mn 45.000, Merck KGaA, Germany) which was dissolved in the same solvent system at a final concentration of 18 wt% and electrospun on a NanospiderTM 1WS500U (Elmarco s.r.o., Czech Republic). The PLCL copolymer was also electrospun by means of this technique. Planar samples of electrospun PCL and PLCL were employed for in vitro assessment purposes.
Preparation of the vascular grafts
The electrospinning solution was stirred overnight and immediately electrospun on a custom-made device as schematically depicted in figure 1. The resulting fibers were collected on a rotating mandrel with an inner diameter of 2 mm. The distance between the needle tip and the collector was set at 20 cm. The polymer dosage applied during the preparation of the sample was maintained at 1.5 ml h−1, the needle was charged at 15 kV and the collector was grounded. The moving needle enabled the production of samples with a length of up to 15 cm. The ambient conditions consisted of a temperature of 25 °C and humidity between 50% and 55%. Electrospinning was applied until the resulting tubular samples achieved a thickness of 100 μm, whereupon they were pushed manually from the mandrel.
Figure 1. Schematic presentation of the custom-made electrospinning device used for the fabrication of tubular grafts and composed of a high speed engine (1), a torque converter (2), a grounded collector in the form of a rotating mandrel (3), a positively charged electrospinning needle (4), a pneumatic control unit (5), a polymeric solution pump (6) and a high voltage source (7).
Download figure:
Standard image High-resolution imageCharacterization of the fibrous samples
The morphology of the fibrous scaffold was analyzed by means of SEM and micro-CT. The fibrous samples were cut into pieces, placed in a SEM holder and covered with a 7 nm layer of gold. Subsequently, a cross-section of the tubular graft was captured by a TESCAN Vega 3SB Easy probe (TESCAN s.r.o, Czech Republic). The planar samples employed for cell seeding purposes were analyzed in the same way and the fiber diameter was characterized using NIS Elements software (LIM s.r.o., Czech Republic). The fiber diameter was assessed from a total of 200 measurements per material taken from three to five independent SEM pictures; the data was then presented in the form of the mean ± standard deviation.
Micro-CT scans of the tubular graft were acquired using a micro-CT SkyScan 1272 (Bruker, Kontich, Belgium). For micro-CT scanning purposes, grafts with a wall thickness of 400 μm were obtained over an extended electrospinning time under the same conditions as those described previously. The analyzed sample was scanned in air applying the following scanning parameters: 0.5 μm pixel size, camera binning 1 × 1, rotation step 0.1°, source voltage 50 kV, source current 200 μA, no filter, frame averaging (10) and 180° rotation. The scanning time was approximately 6 h for each specimen (n = 10). The flat-field correction was updated prior to each acquisition.
Cross-sectional images were reconstructed from projection images by means of NRecon software (Bruker, Kontich, Belgium) employing a modified Feldkamp algorithm. Computed tomography artifacts were reduced via the requisite setting of the correction parameters (misalignment, ring artifact and beam hardening). Visualizations were acquired by means of DataViewer (2D cross-sectional images) and CTVox (3D images; Bruker, Kontich, Belgium).
The 3D analysis of the specimen structure was conducted by means of CTAn (Bruker, Kontich, Belgium). The volume of interest (VOI) was set in the form of a cube (side = 250 μm) in the middle of the specimen so as to exclude the effect of potential superficial alterations resulting from the handling of the specimen. A total of ten specimens were analyzed for morphological parameters. Image data analysis was optimized employing TeIGen software [25]. Binarization was achieved employing global thresholding which was determined to provide a suitable method in this case. Image noise reduction was ensured using despeckle operations in 3D.
In vitro assessment
Prior to the performance of the animal experiments, the interaction of the materials was assessed using 3T3 mouse fibroblast cell lines (ATCC, USA) and human umbilical vein endothelial cells (HUVEC, Lonza Biotec s.r.o., Czech Republic). The in vitro tests were performed on planar samples prepared by means of the needle-less electrospinning of PCL and the PLCL copolymer. The samples were cut into circles with a diameter of 6 mm. Prior to cell seeding, the materials were soaked in 70% ethanol for 30 min followed by double rinsing in phosphate-buffered saline (PBS, pH 7.4).
The mouse 3T3 fibroblasts were cultivated in Dulbecco's Modified Eagle Medium (Lonza Biotec s.r.o., Czech Republic) supplemented with 10% fetal bovine serum (Lonza Biotec s.r.o., Czech Republic), 1% glutamine (Biosera, Czech Republic) and 1% penicillin/streptomycin/amfotericin B (Lonza Biotec s.r.o., Czech Republic). The fibroblasts (passage 19) were seeded on scaffolds placed in 96 well plate at a density of 5 × 103 per well.
The human umbilical vein endothelial cells were cultivated in Endothelial Basal Medium (EBM-2, Lonza Biotech s.r.o., Czech Republic) supplemented with EGM-2 Single Quots (Lonza Biotec s.r.o., Czech Republic). The endothelial cells (passage 6) were seeded on scaffolds placed in 96 well plate at a density of 7.5 × 103 per well.
The interactions of the cells with the electrospun PCL and PLCL were evaluated after 1, 3, 7 and 14 days of culturing by means of the metabolic MTT test. The MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide] was reduced to purple formazan by means of mitochondrial dehydrogenase in cells that indicated normal metabolisms. MTT solution (50 μl) was added to 150 μl of the complete medium and the samples were incubated at 37 °C for 4 h. The resulting formazan violet crystals were solubilized using acidic isopropanol and the optical density of the suspension was measured (λsample 570 nm, λreference 690 nm). Four samples of each material were incubated with MTT solution on each of the testing days and the average absorbance was calculated as the difference between absorbance measured at 570 nm and the reference wavelength of 690 nm.
In order to allow for the evaluation of cell morphology and spreading 1 day following cell seeding, the samples were rinsed twice in PBS and fixed in 2.5% glutaraldehyde for 30 min at 4 °C and, subsequently, 0.1% Triton X-100 in 0.1% bovine serum albumin solution in PBS was used as a blocking buffer for 10 min. The samples were stained using phalloidin-FITC (Merck KGaA, Germany, dilution 1:1000) which binds to the actin filaments of the cells and stains them green. DAPI (Merck KGaA, Germany, dilution 1:1000) was used for the counterstaining of the cell nuclei to blue. The stained cells were then observed by means of a Nikon Eclipse-Ti-E inverted fluorescence microscope (Nikon Imaging, Czech Republic).
The viability testing and cell quantification data obtained was processed using two-way analysis of variance (ANOVA) with the Bonferonni multiple comparison test.
In vivo testing
The materials intended for in vivo implantation were sterilized using ethylene oxide and aerated for one week at room temperature. The animals were treated according to the Czech National Convention for the Protection of Vertebrate Animals used for Experimental and other Specific Purposes Act, Collection of Legislation No. 246/1992 and its amendments concerning the Protection of Animals from Cruelty, the Public Notice of the Ministry of Agriculture of the Czech Republic and Collection of Legislation No. 419/2012 (the Keeping and Exploitation of Experimental Animals). Vascular grafts were implanted into New Zealand white rabbits (n = 10) in the form of a carotid artery bypass. The rabbits were six months old and weighed between 3.6 and 5.0 kg. The animals were anesthetized with Narcotan (Halotan)+O2 prior to implantation and Diazepam (Apaurin), Ketamin (Narketan) and Xylazin (Xylapan) were administered to the rabbits prior to surgery. The skin on the left side of the necks of the rabbits was shaved and disinfected using iodopovidon. The incision of the neck and the preparation of the subcutaneous and muscle tissue were performed followed by the isolation of the arteria carotis communis. Heparin was administered intravenously (300 mU kg–1) and the proximal part of the carotid artery was clamped. The vascular prosthesis was then sutured using Prolene 7-0 (Ethicon, Johnson & Johnson, Czech Republic) in the proximal section. The same procedure was applied to the distal section of anastomosis. The length of the implanted prosthesis was 1.5 cm. The vascular graft was sutured by means of the end-to-side technique and the bypassed section of the carotid artery was ligated. Following surgery, the animals received Meloxikam (Meloxydil 5) and Marbofloxacin (Marbocyl) medication for five days. During the time of observation, acetylsalicic acid (Kardegic 0.5 g) was administered daily (10 mg/rabbit per os).
The ten rabbits subjected to investigation were divided into two groups for the assessment of different graft healing time-points. The first group was analyzed after ten weeks of implantation (n = 5 rabbits) and the second group was monitored for long-term healing after six months (n = 5 rabbits).
Histological evaluation
For histological processing purposes, the grafts were explanted together with the proximal and distal parts of the carotid artery after ten weeks and after six months of survival. The explanted grafts with the surrounding tissue were then processed employing standard histological techniques. For the purpose of comparison, a native carotid artery without a replacement was assessed in the same way at both the ten weeks and six months time-points.
Following formalin fixation, the tissue blocks were cut into 4 μm thick histological sections with a section plane perpendicular to the long axis of the artery. The sections were stained using a variety of general (table 2) and specialized histological stains with respect to their immunohistochemical reactions (table 3). The staining methods were selected in order to be able to characterize the cellular populations of the grafts employing a similar approach to that of other research papers on the topic [1, 16, 21, 23, 24]. The visualization of the immunohistochemical reaction was based on diaminobenzidine (DAB+, Liquid; DakoCytomation, Glostrup, Denmark). The immunohistochemical sections were counterstained with Gill's hematoxylin and all the sections were dehydrated in graded ethanol solutions and mounted with a xylene-soluble medium. The sections were then observed by means of bright field microscopy and polarized light microscopy. However, when processed using standard histological techniques, it was found that most of the grafts lost their integrity which prevented the reliable quantification of the results. Due to the variability of the histological findings, all the significant features are described separately.
Table 2. Histological staining methods used in the study.
| Staining | Purpose |
|---|---|
| hematoxylin-eosin [26] | Overall morphology of the graft, foreign-body giant cells |
| Verhoeff's hematoxylin and green trichrome [27] | Overall morphology, differentiating connective tissue, elastin and vascular smooth muscle |
| Orcein (Tanzer's orcein, Bowley Biochemical Inc., Danvers, MA, USA) | Elastin fibers |
| Picrosirius red (Direct Red 80, Sigma Aldrich, Munich, Germany) [28–30] | Type I and type III collagen when observed under circularly polarized light |
Table 3. Primary antibodies used for the immunohistochemistry.
| Antibody (and staining purpose) | Manufacturer | Dilution | Pretreatment |
|---|---|---|---|
| Monoclonal Mouse Anti-Human Smooth Muscle Actin, Clone 1A4 (smooth muscle and myofibroblast marker) | DakoCytomation (Glostrup, Denmark) | 1:100 | 20 min 96 °C Dako Target Retrieval Solution, pH 9 |
| Monoclonal Anti-CD31 antibody Clone J70A (endothelial marker) | DakoCytomation | 1:40 | 20 min 96 °C Dako Target Retrieval Solution, pH 9 |
| Monoclonal Mouse Anti-Human Neurofilament Protein, Clone 2F11 (nerve fiber marker) | DakoCytomation | 1:75 | 20 min 96 °C Dako Target Retrieval Solution, pH 9 |
| Monoclonal Mouse Anti-Human CD68, Clone KP1 (macrophage marker) | DakoCytomation | 1:100 | 20 min 96 °C Dako Target Retrieval Solution, pH 9 |
The explanted native rabbit carotid arteries were characterized according to their wall thicknesses and inner diameter measurements. The average thickness of the tunica intima and the media was calculated from 4 native vessel positions. A line perpendicular to the lumen and connecting the luminal surface of the intima with the most abluminal elastic lamellar unit was drawn in each position using an Ellipse software linear measurement tool. Similarly, the average thickness of the whole of the wall was calculated from the length of the lines connecting the luminal surface of the intima with the outermost layer of the dense collagenous connective tissue of the adventitia according to Witter et al [31]. The average diameter of the native carotid arteries was calculated from the measurement of the area profile of the lumen visible in the histological sections according to Kochova et al [32]. Similar measurements could not be applied to the grafts since they usually failed to retain their round shape; moreover, cross-sections through the grafts were often deformed in the histological sections.
Determination of graft degradation in vivo
Following the histological evaluation, the explanted grafts embedded in paraffin were analyzed by means of gel permeation chromatography (GPC) in order to detect changes in the molecular weight upon implantation. The paraffin was removed via the serial rinsing and vortexing of the grafts with the surrounding tissue in hexane. The samples were then allowed to dry. Finally, the grafts were dissolved in tetrahydrofuran (THF) and the solution was left to evaporate under nitrogen so as to obtain a volume of 0.2 ml for use in the GPC analysis phase. A granulate of the PLCL copolymer and an electrospun graft were also assessed for comparison purposes. The control samples were diluted in THF so as to attain a concentration of approximately 1 g l−1.
The analysis included the use of the Dionex Ultimate 3000 HPLC system with a diode array and a Varian LC-385 ELSD detector along with a polymeric Phenomenex Phenogel 1E4 GPC column with a length of 30 cm with an i.d. of 4.6 mm and a particle size of 5 μm. The temperature of the column compartment was set at 30 °C and THF of HPLC grade purity at a flow rate of 1 ml min−1 was applied. The chromatograms were recorded at wavelengths of 200, 210, 220 and 250 nm employing an ELSD detector for 23 min. The nebulizer temperature of the ELSD detector and the evaporator temperature were set at 80 °C; the nitrogen flow rate applied was 1.3 l min−1 with an injection volume of 30 μl.
The elution of the polymeric substances was observed between 5 and 9 min; consequently, the corresponding graph was compiled so as to illustrate this specific retention time interval. The grafts with the highest peak intensity were selected for the demonstration of the shifts in the molecular weight.
Results
Vascular graft morphology
The electrospinning of the PLCL copolymer was performed using a custom-designed device with a rotating mandrel collector. The production process was very sensitive to ambient conditions especially relative humidity that was required to be maintained at between 50% and 55%. A cross-section of the resulting tubular structure is shown in figure 2. The collector in the form of a rotating mandrel had a diameter of 2 mm. However, following removal, the grafts shrank to an inner diameter of approximately 1.4 mm. The average fiber diameter of the tubular samples measured from the SEM images was determined at 5.40 ± 2.09 μm.
Figure 2. SEM photograph of a cross-section of an electrospun PLCL vascular graft; scale bars 1 mm (A), 100 μm (B).
Download figure:
Standard image High-resolution imageThe tubular samples were characterized by means of micro-CT. A visualization of the VOI (the cube inside the tubular graft wall) is depicted in figure 3 in the form of 2D (A), (B) and 3D images (C), (D). The grayscale images show plane fibers at selected sections and the pores between them; the spatial resolution and partial volume effect may have resulted in blurred fiber edges. Thus, the binarization (figure 3(B)) required prior to the image analysis may possibly influence the outcomes. The structure thickness (in 3D) can be clearly presented as a color-coded image (figure 3(D)). However, whole specimen visualization in various 2D sections or in the form of 3D volume imaging presents a unique opportunity for the evaluation of the specimens.
Figure 3. Grayscale image of a micro-CT transversal section (perpendicular to the fibrous long axis) (A) and following image binarization (B), scale bar 30 μm. 3D visualization of a binarized specimen (C), 3D visualization in color-coded mode showing differences in structure thickness: gradient transition from green to violet: green = 2 μm, yellow = 4 μm, red = 8 μm, violet = 12 μm (D), scale bar 100 μm.
Download figure:
Standard image High-resolution imageThe morphological features revealed by the micro-CT analysis can be seen in table 4. The PLCL copolymer occupied approximately 30% of the graft structure with a porosity of around 70%. The thickness of the structure corresponds to a fiber diameter of 5.58 ± 0.10 μm. The pores within the structure evince a mean of 9.34 ± 0.19 μm (represented as structure separation in the table). More than 99.99% of the porosity constituted open porosity, i.e. approximately equal to the presented total porosity value.
Table 4. Results of the micro-CT 3D analysis of selected parameters (mean, median, standard deviation).
| MEAN | MED | SD | |
|---|---|---|---|
| Percent object volume (mm3) | 30.37 | 30.22 | 0.76 |
| Structure thickness (μm) | 5.58 | 5.58 | 0.10 |
| Structure separation (μm) | 9.34 | 9.31 | 0.19 |
| Total porosity (%) | 69.63 | 69.78 | 0.76 |
MEAN = mean value, MED = median, SD = standard deviation.
Morphology of the planar samples
The electrospun morphology of PCL and PLCL is depicted in figure 4. The PLCL copolymer created homogeneous fibers with a diameter of 1.35 ± 0.57 μm in the form of electrospun planar sheets. The electrospinning of the PCL produced fibers with a diameter of 0.49 ± 0.95 μm. The high standard deviation in the PCL fibrous layer indicated the presence of tiny fibers (minimum fiber diameter of 0.3 μm) as well as thick fibers within the structure (3.54 μm maximum measured fiber diameter).
Figure 4. SEM photos of the electrospun planar samples fabricated from PLCL (A) and PCL (B); scale bar 20 μm.
Download figure:
Standard image High-resolution imageInteractions with the cell lines
The electrospun samples were prepared for in vitro assessment purposes by means of the needle-less electrospinning technique which resulted in planar samples suitable for the assessment of cell material interaction. Fibrous samples fabricated from the PLCL copolymer under study as well as PCL were seeded with two types of cell lines, i.e. 3T3 mouse fibroblasts and endothelial cells. Polycaprolactone was chosen for comparison purposes since this polymer is widely used for vascular graft purposes in its electrospun form. Viability testing revealed that the PLCL copolymer supported the proliferation of both cell types over 14 days of experimentation (figures 5(A), (B)). The higher cell viability of cells cultured on the PLCL copolymer was recorded following the culturing of both cell types over 14 days; moreover, the difference in the proliferation rate was more remarkable with respect to the fibroblasts, concerning which higher cell viability was recorded as soon as on the seventh day following cell seeding on the PLCL copolymer.
Figure 5. Results of the MTT viability assay of seeded 3T3 fibroblasts (A) and endothelial cells (B) on electrospun PLCL and PCL, **** denotes p < 0.0002, **** refers to p < 0.0001 (2 way ANOVA, Bonferonni).
Download figure:
Standard image High-resolution imageThe adhesion of cells on the surface of the fibrous materials was assessed by means of fluorescence microscopy. The staining of the actin filaments indicates cellular spreading. The counterstaining of the cell nuclei with blue DAPI allowed for the comparison of cell quantity on the tested materials. Figures 6(A)–(D) depicts cells captured on electrospun PLCL and PCL one day following seeding. More cells adhered to the PLCL copolymer than to the electrospun PCL, which also exhibited a lower degree of cellular spreading.
Figure 6. The adhesion of fibroblasts to the surface of the electrospun PLCL copolymer (A) and PCL (B), the adhesion of endothelial cells to fibrous PLCL (C) and PCL (D); scale bar 100 μm.
Download figure:
Standard image High-resolution imageImplantation of vascular grafts
The prepared vascular grafts exhibited excellent surgical handling and suture retention following implantation as a bypass of the carotid artery. No significant blood leakage was observed subsequent to the restoration of blood flow and pressure. Blood flow was monitored at the end of the study by means of palpitation. All the tested animals evinced pulsation.
Histological evaluation of the explanted grafts
Two vascular graft healing time-points were assessed, i.e. the prostheses and the control carotid arteries without replacements were explanted after ten weeks and six months (n = 5 + 5). The intima and media thickness of the native rabbit carotid artery ranged from 120 to 155 μm after ten weeks and from 145 to 180 μm after six months. The total thickness of the carotid artery, including the adventitia, ranged from 205 to 230 μm after ten weeks and from 260 to 288 μm after six months. The histological sections of the native rabbit carotid arteries are shown in figure 7. The vessel wall was composed of layers of elastin and a collagen extracellular matrix infiltrated by smooth muscle cells. The inner diameter of the histological sections of the normal rabbit carotid artery ranged from 1140 to 1360 μm after ten weeks and from 1270 to 1510 after six months.
Figure 7. Morphology of the wall of the native carotid artery ten weeks (1st column) and six months (2nd column) following the procedure. The tunica media consisted principally of repeating elastic units as demonstrated by Verhoeff's hematoxylin and green trichrome (A), (B), orcein stain showing the elastic membranes (C), (D) and vascular smooth muscle cells positive for alpha SM-actin (E), (F). Most of the type I collagen was found in the adventitia as demonstrated by picrosirius red photographed under polarized light (G), (H); scale bars 20 μm.
Download figure:
Standard image High-resolution imageMacroscopic images of prostheses prior to explanting are depicted in figures 8(A), (B). After ten weeks, four of the five tested grafts were found to be patent (an 80% patency rate after ten weeks); a cross-section of such a prosthesis is depicted in figures 8(C), (E). At the later time-point of six months, two of the tested prostheses (40%) were observed to contain invaginated and partially fibrotized islands of connective tissue surrounded by the graft lumen. The initial strength of the mechanical properties of the grafts were lost during the long-term healing process; four of the grafts were broken in the middle or near an anastomosis when explanted (80% of the implanted grafts). Full vascular graft patency was observed with respect to one graft with a freely passable lumen (a 20% patency rate after six months), as depicted in figures 8(D), (F); the whole width of the wall had been penetrated by cells. After six months, the thickness of the vessel wall decreased to 68 μm suggesting that the degradation of the prosthesis was more rapid than tissue regeneration. No apparent inflammatory infiltrates or foreign-body giant cells were discovered except for small areas associated with the stitches between the wall and the grafts (data not shown).
Figure 8. Macroscopic pictures of the prosthesis sutured as the end-to-side anastomosis of the rabbit carotid artery following ten weeks (A) and six months (B), cross-sections of the prosthesis at a lower (C), (D) and higher magnification (E), (F) stained with Verhoeff's hematoxylin and green trichrome; scale bars 500, 20 μm. The red dotted line represents the border between the lumen and the adluminal border of the graft. The outer (abluminal) border of the graft was not clearly visible since it had already been partially remodeled by newly infiltrating cells, the newly produced extracellular matrix and the partial degradation of the graft components.
Download figure:
Standard image High-resolution imageIn addition to the consideration of the overall healing process, the characterization of the cells and resultant extracellular matrices was conducted by means of specific staining as depicted in figure 9. The endothelial cells were stained via the visualization of CD 31. Despite partial damage to the graft surfaces during the histological processing phase, following ten weeks of healing, remnants of endothelial-like cells were found to be partially preserved; however, they were negative with respect to CD 31 immunohistochemistry (see figure 9(A)). The longer period of six months resulted in the positivity of the endothelial cells with respect to the staining method employed (figure 9(B)). Smooth muscle cells were presented as early as after ten weeks, at which time they were observed to have integrated within the graft wall assuming a spindle shape as can be seen in figure 9(C). The smooth muscle cells were arranged in layers resembling the natural composition of arteries during the healing process (see figure 9(D)). The extracellular matrix was composed of collagen, which was present in the regenerated vessel walls at both time-points as can be observed from the green trichrome staining depicted in figures 8(D)–(F) and the picrosirius red staining shown in figures 9(G), (H). The staining of elastin was also performed and the results are visible in figures 9(E), (F). Despite the presence of smooth muscle cells, no elastic fibers were observed. The heavy false positivity of the orcein staining was discovered at the later time-point of six months (figure 9(F)). Nervi vasorum and vasa vasorum were not present at any time within the grafts.
Figure 9. Detail of a graft wall stained for CD 31 (endothelial cell marker) ten weeks (A) and six months following suturing. The endothelial-like cells (A) were found to be CD31-negative after ten weeks. After six months, the continuous endothelium expressed the CD31 marker (B). The wall of the graft penetrated by smooth muscle cells (positive for alpha smooth muscle actin) after ten weeks (C) and six months ((D), distinct layers of smooth muscle cells are visible). Staining for elastin fibers in the vessel of the prosthesis did not reveal any elastin after ten weeks (E) or after six months ((F), heavy but non-specific background staining), collagen staining in polarized light after ten weeks (G) and after six months (H); scale bars 20 μm. The red dotted line represents the border between the lumen and the adluminal border of the graft. The outer (abluminal) border of the graft was not clearly visible since it had already been partially remodeled by newly infiltrating cells, the newly produced extracellular matrix and the partial degradation of the graft components.
Download figure:
Standard image High-resolution imageFollowing ten weeks of healing, one of the five implants was destroyed by a thrombus (as shown in figures 10(A), (C)) that was organized with connective tissue. The thrombus was recanalized by CD31-positive micro-vessels. Long-term healing over six months resulted in the prolapse of two prostheses containing invaginated and partially fibrotized islands of connective tissue surrounded by the graft lumen (see figures 10(B), (D)).
Figure 10. Thrombotic vascular graft explanted after ten weeks at lower (A) and higher (C) magnifications. Prolapsed prosthesis after six months of healing in a rabbit at lower (B) and higher (D) magnifications. At both time intervals, the thrombi were partially recanalized by CD31-positive micro-vessels (B), (D); scale bars 500 μm (A), (B), 100 μm (C), (D).
Download figure:
Standard image High-resolution imageAssessment of the molecular weight shift following implantation
The molecular weight of granules of the PLCL copolymer, an electrospun vascular graft fabricated from the same polymer and the grafts following implantation in rabbits after ten weeks and six months were assessed by means of GPC. The retention time (tR) of the peaks corresponds to the molecular weight, i.e. a shift thereof towards longer retention times indicates a decrease in molecular weight. The results are expressed in the form of relative values which compare the input materials with their stages following implantation (see figure 11).
Figure 11. GPC chromatograms of the PLCL copolymer: polymeric granulate (1), electrospun vascular graft (2), vascular graft after ten weeks in vivo (3) and after six months in vivo (4).
Download figure:
Standard image High-resolution image
The molecular weight clearly shifted towards lower values as a result of the electrospinning process (from a retention time of 5.8 to 6.2 min), and a further massive weight loss was observed following implantation, i.e. the retention times of the explanted grafts shifted to average values of 7.8 min after ten weeks and 8.0 min after six months. While the peaks of highest intensity are depicted in figure 11, it is important to note that the retention times of the analyzed samples varied from 6.8 to 8.2 min with respect to both of the assessed time-points.
Discussion
The effectiveness of vascular grafts based on electrospun biodegradable polyester were investigated due to the need for such 'off-the-shelf' materials in the field of cardio surgery. Our previous studies concentrated primarily on the development and characterization of polycaprolactone-based vascular prostheses [33, 34]. Due to the excellent elastic properties of the PLCL copolymer as reported by Johnson et al [6] and Horakova et al [35], this material is proving to be of particular interest with respect to the fabrication of vascular grafts. Moreover, cell culture experiments suggest that PLCL is preferred to PCL by fibroblasts and the endothelial cell line (see figures 5, 6). It is possible to explain this outcome by the increased surface wettability of the PLCL copolymer compared to that of the PCL. Polycaprolactone is a hydrophobic polymer with a water contact angle of around 130° [15] while polylactide is more hydrophilic due to its shorter side chain. The water contact angle of electrospun polylactide has been reported at around 100° [36]. It is supposed that the PLCL copolymer containing 70% of polylactide units is more hydrophilic than pure polycaprolactone, thus rendering it more beneficial in terms of cell adhesion. It is known that slightly hydrophilic surfaces are preferred by most cell types. Even though both materials are, nevertheless, considered hydrophobic, it is proposed that the decreased hydrophobicity of the PLCL copolymer may well enhance cell material interactions. Vascular grafts were fabricated by means of the needle electrospinning of the PLCL copolymer for implantation in rabbits in the form of a carotid artery bypass. Prior to implantation, the materials were sterilized using ethylene oxide, the effects of which on electrospun PLCL has been reported in our previous study [35]. While other studies have also considered the PLCL copolymer, a comparison with these studies was deemed inappropriate since differing ratios of L-lactide and ε-caprolactone were applied. Whereas, for example, Tara et al [3] and Wu et al [4] applied a ratio of 50/50 and Laurent et al [37] employed a ratio of 85/15, our study was based on the utilization of the medical grade GMP copolymer with a ratio of 70/30. The composition and organization of monomer units exert a huge impact on the overall behavior of the polymer. A recent study by Pisani et al considered a polymer with the same composition, i.e. 70/30, the aim being to develop a material for use in the healing of esophageal defects. Different solvents were employed for electrospinning (methylene chloride and N,N-dimethylformamide) to those used in our study and higher concentrations of the polymer were added to the electrospinning solution (15%–25%) [38]; both of these parameters are capable of influencing the behavior of the final material.
It is essential that the appropriate morphological characterization be conducted. The gold standard consists of the evaluation of the fiber diameter and pore size solely by means of SEM, which does not in fact fully correspond with reality since the consideration of the 3D structure is not included when applying this method of analysis (compare figures 2 and 3). Modern techniques such as micro-CT enable 2D and 3D visualization accompanied by the structure analysis with the advantages of non-destructivity, whole specimen evaluation, direct 3D analysis, increased time efficacy, the reduction of subjectivity and the potential for combination and comparison with standard destructive (section-based) analysis methods [11, 39–41]. However, micro-CT also presents a number of significant drawbacks, i.e. the spatial resolution is lower than that of SEM, thin structures are influenced by the partial volume effect which decreases the x-ray density thereof in reconstructed images, and micro-CT images are negatively influenced by computed tomography artifacts and image noise [42]. While the binarization process in such structures most likely influences the results, the extent of such bias is not yet known. Hence, our group developed software for virtual 3D object generation [43] aimed at enhancing the calibration of micro-CT analysis. 3D analysis is based on the sphere-fitting algorithm [9, 10], which provides a powerful tool with respect to structure characterization. In the case of more complex structures (e.g. a combination of fibers below and above micro-CT spatial resolution), it is important that we consider the above drawbacks, which may lead to the determination of significant differences between micro-CT and SEM values due to some parts of the structure not being detected via micro-CT analysis. A recent study by Bartos et al analyzed the limitations of micro-CT compared to SEM with respect to the pore size analysis of collagen-based composite scaffolds. Their results revealed significant differences regarding pore size evaluation between these two approaches, with micro-CT being considered the most beneficial overall approach [44]. Nevertheless, while this drawback applies to the planar specimens used in this study due to their smaller fiber diameters and the wide distribution of the fiber diameters (1.35 ± 0.57 μm in the case of electrospun PLCL and 0.49 ± 0.95 μm in that of electrospun PCL), it does not apply to the tubular specimens, concerning which conformity between the SEM image analysis and micro-CT analysis was determined (a fiber diameter of 5.40 ± 2.09 μm as measured by SEM compared to 5.58 ± 0.10 μm as measured by micro-CT).
The data obtained by means of micro-CT revealed that the grafts exhibited a porosity of around 70% with pore sizes of around 9.3 μm and a fiber diameter of 5.6 μm. These parameters are of prime importance in terms of the further cell infiltration of the prosthesis following implantation. A number of studies have been devoted to determining the optimal pore size for the regeneration of the vessel wall. Wang et al used electrospun PCL in their study in order to create small-pore tubular grafts with a pore size of 4.66 ± 1.63 μm (porosity of 66%) and a fiber diameter of 0.69 ± 0.54 μm, and large pore grafts with a pore size of 40.88 ± 13.67 μm (porosity of 83%) and a fiber diameter of 5.59 ± 0.67 μm. They concluded that large pores supported the regeneration of the tissue towards M2 macrophages responsible for cellular infiltration and vascularization [16]. De Valence et al also described the need for appropriate pore sizes within vascular grafts. Their study involved the covering of a high porosity graft prepared via the electrospinning of PCL from either the luminal or adventitial side with a low porosity layer of the same polymer. The high porosity layer had a fiber diameter of 2.21 ± 1.40 μm and a pore size of 9.1 ± 2.2 μm (porosity of 81%) whereas the low porosity layer had a fiber diameter of 0.83 ± 0.56 μm and a pore size of 3.3 ± 1.7 μm (porosity of around 63%). The aim of the study was to create a structure that would allow for cell infiltration through large pores and the prevention of blood leakage that may occur through highly porous materials. They concluded that an inner layer composed of low porosity fibers reduced blood leakage and did not impede cell infiltration from the adventitial side of the graft [45].
The histological part of the explanted graft analysis was characterized by severe limitations which prevented us from quantifying the histological findings. The routine processing and staining of the paraffin-processed sections produced a large number of sectioning and staining artifacts, a loss in integrity due to the presence of partially dissolved graft components and unspecific and biased staining results. Thus, an alternative approach is recommended using frozen sections which avoids the use of paraffin embedding and alcohol solvents when processing samples which include a significant proportion of PLCL fibers. However, it was possible to assess the tissue reaction of the grafts according to our previous studies of the vascular wall [31, 46, 47]. A further limitation consists of the fact that the measurement of the wall thickness and the diameter of the native carotid artery were affected by tissue shrinkage which occurred during the histological processing of the paraffin-embedded sections. According to data published by Matsumoto et al, the diameter of the carotid artery in vivo should be approximately 1.35 times greater [48], i.e. the native carotid arteries investigated in this study exhibited an inner diameter of 1140–1360 μm after ten weeks compared to 1540–1840 μm in reality and 1270–1510 μm after six months compared to 1720–2040 μm. Thus, the inner diameters of the implanted electrospun grafts (1400 μm) were even smaller than that of the native carotid arteries. The properties required of the final prosthesis were discussed with cardio surgeons and, based on their suggestions, grafts were produced with an inner diameter of 2 mm (1.4 mm following the shrinkage of the material) and wall thickness of 100 μm. With respect to our study, the walls of the implanted grafts were not thick enough to comply with the regeneration rate. Ideally, the thickness should match native carotid artery values, i.e. between 200 and 300 μm. With respect to biodegradable materials, since the degradation rate must be considered, it is reasonable to assume that an even greater wall thickness is necessary so as to ensure the regeneration of the vessel wall over the long term. That said, even though the materials were significantly thinner (wall thickness of 100 μm), the material remained mechanically stable up to ten weeks in vivo. With concern to the later time-point of six months, therefore, it can be assumed that a thicker structure would ensure the maintaining of the appropriate mechanical properties, thus leading to the required functioning of the material.
The grafts considered in our study had become fully penetrated by cells ten weeks and six months following implantation, which suggests that the pore size and porosity were appropriate (as depicted in figures 8, 9). After six months, endothelial cells were found to cover the luminal side of the graft and smooth muscle cells had assembled into layers as in normal vessels (see figures 9(B), (D)). The infiltrated cells had created their own extracellular matrix composed of collagen without the presence of elastin, an outcome that could be explained by the mechanical properties of the graft. The PLCL copolymer exhibits high degrees of elongation at break (around 500% of its initial length as measured for an inner diameter of 4 mm and a wall thickness of 200 μm as reported in [35]), compliance (8%/mm Hg) and burst pressure (2.5 MPa measured in an electrospun graft with a 6 mm inner diameter and a vessel wall thickness of 650 μm [6]). However, the elastic properties are usually converse to the mechanical strength of the material; thus, the production and deposition of collagen was preferred to that of elastin.
With respect to our study, complications were observed in terms of thrombogenicity and rapid degradation accompanied by the loss of the mechanical properties. After ten weeks in vivo, four of the five implanted grafts were found to be patent and in one case an organized thrombus was observed (figures 10(A), (C)). The interaction with blood was assessed in our previous study, which revealed the high thrombogenicity of electrospun biodegradable polymers to an extent comparable to collagen materials [49]. The further anti-thrombogenic modification of the electrospun grafts will therefore be necessary, especially with concern to small diameter prostheses.
After six months in vivo, two of the five examined prostheses exhibited prolapsed vascular walls (figures 10(B), (D)), which is associated with weakened mechanical properties. Over the long-term healing period of six months, the vascular grafts lost their mechanical strength due to the rapid degradation process in vivo as measured via GPC (figure 11). A decrease in molecular weight was observed as soon as after ten weeks despite there being no visible signs thereof during either explantation or the histological examination. However, evidence of weak mechanical properties was apparent after the longer six month period at which time four of the five grafts were found to have broken. This development may have been caused either by the explantation technique or directly in vivo due to a loss of mechanical support. Such degradation has been studied previously by, for example, Horakova et al with respect to a simulated enzyme environment [35]. A study by Pisani et al addressed the degradation of electrospun PLCL in PBS (pH 7.4) and under conditions simulating inflammation with a lower pH of 6.0. They discovered that after 28 days the material had lost 0.5%–2% of its initial weight in PBS and 0.9%–4% in the pH 6.0 environment [38]. However, all these experiments consisted of the simulation of real conditions and their value lay in the comparison of different materials with a view to the prediction of the degradation rate; a real assessment, however, can only be provided by an in vivo study. Degradation is dependent on the material properties. It is known that polyesters undergo hydrolysis that may be facilitated by certain enzymes. As a statistical copolymer, PLCL makes up one of the amorphous polymer group that degrades more rapidly than semi-crystalline polymers. Water penetrates more quickly into amorphous regions accompanied by surface erosion that alters the morphology of the fibers [35]. Moreover, other structural properties such as fiber diameter and the mass of the material present influence degradation. In addition, the biological variability of animal specimens, particularly the availability of enzymes presented within the implanted scaffold, exert a huge impact on the degradation rate. Our study demonstrated that the degradation of the PLCL statistical amorphous copolymer led to a decrease in the molecular weight in the first ten weeks following implantation that remained up to the six month time-point.
Conclusion
Vascular grafts were successfully fabricated via the needle electrospinning of the PLCL copolymer resulting in tubular structures with adequate morphological properties which facilitated the regeneration of the vessel wall in vivo. From the histological point of view, the grafts appeared to be well tolerated and no adverse tissue reaction was determined. After six months in vivo, the lumen of the vascular graft had been endothelialized and the vascular wall comprised arranged smooth muscle cells that produced collagen fibers. However, complications in the form of thrombosis and a rapid degradation rate were observed and discussed in the study with a view to the further development of small diameter bypass grafts.
Acknowledgments
The study was supported by the 'Nanofiber Materials for Tissue Engineering' project, reg. no. CZ.1.05/3.1.00/14.0308 co-financed by the European Social Fund and the state budget of the Czech Republic. The authors are particularly grateful to Dr Ondrej Novak and Josef Vosahlo for the construction of the spinning device.
TK and ZT were supported by the National Sustainability Program I (NPU I) no. LO1503 of the Ministry of Education, Youth and Sports of the Czech Republic and the Progress Q39 project of the Charles University.
MB was supported by the Ministry of Education, Youth and Sports (Progress Q29/LF1), the Charles University (First Faculty GAUK 5070/2018), the Ministry of Health of the Czech Republic (NV 15-25813A) and the 'Technological development of post-doc programs' project (reg. no. CZ.1.05/41.00/16.0346) supported by the Research and Development for Innovations Operational Program (RDIOP), co-financed by European regional development funds and the state budget of the Czech Republic.