1. Introduction
Research is continuously carried out to create the ideal tissue scaffold. This is particularly important in the investigation of vascular prostheses [
1]. Vascular grafts larger than 6 mm in diameter are widely available [
2]. The production of artificial vascular grafts of smaller diameter (diameter less than 6 mm [
3])—in contrast to those of larger sizes—is associated with many difficulties [
4]. For this reason, it is necessary to search for new techniques for the production of small vascular prostheses. Vascular prostheses with small diameters, like those with larger diameters, must have appropriate parameters. In addition to bio- and hemocompatibility, one of these criteria is a surface with proper pore sizes [
5,
6,
7,
8]. A surface pore size adapted to the cell size will facilitate the process of their adhesion to a material and further proliferation [
9,
10]. Moreover, a well-chosen pore size will accelerate and improve the prosthesis’s integration with the surrounding tissues [
11]. Appropriate surface pores additionally will allow cells to exchange nutrients, a limiting factor in small diameter prostheses [
5,
12]. This is a crucial aspect since the endothelialization of the material is the most effective method to prevent thrombosis [
13].
Vascular prostheses with large diameters are usually made of polymers, such as polyethylene terephthalate (PET) and expandable polytetrafluoroethylene (ePTFE) [
14,
15]. However, in the case of vascular prostheses with smaller diameters, these polymers’ use is problematic due to clot formation, clogging the prosthesis’s lumen. Therefore, for the preparation of prostheses of small diameter, polyurethanes (PU) are increasingly used [
16,
17,
18]. PUs are already used in cardiological devices, such as heart valves or vascular stents [
19,
20]. Apart from acceptable bio- and hemocompatibility [
7,
21], polyurethanes are characterized by good mechanical strength [
22]. The biocompatibility of numerous PUs has been studied In Vitro and in vivo for many applications [
23]. They also maintain a good cell adhesion and rate of cell proliferation on their surfaces [
24].
Many different methods can be used to produce vascular prostheses, e.g., electrospinning, solution blow spinning, phase inversion, and solvent casting/particulate leaching [
5,
7,
25,
26]. Phase inversion is a simple method that does not require expensive and complicated equipment. Phase inversion is a process during which a third component (polymer nonsolvent) is introduced into the two-component polymer-solvent system. During the process, the polymer is turned from a solution into a solid state. The basis of the process is the difference in the polymer solubility between the solvent and nonsolvent. As a result, phase separation occurs, and two phases are created—rich and poor—in a polymer. This process is caused by the diffusion of solvent from the polymer solution to the precipitation bath, and the nonsolvent from the precipitation bath to the polymer solution [
27,
28,
29]. Pores are formed due to the loss of solvent and nonsolvent [
30]. Changing the process parameters allows obtaining surfaces with different morphologies. The porosity can be increased by adding a porogen [
31]. In the case of vascular prostheses, in addition to an appropriate surface morphology, a crucial aspect is the good mechanical properties of the scaffold is needed to maintain the vessel’s structural integrity under high pressure and flow [
32].
The study aimed to assess the effect of selected process parameters on the morphology of polyurethane structures that can potentially be used as a small diameter vascular prosthesis. The appropriate morphology of such a scaffold and its mechanical properties are crucial, as outlined above. ChronoFlex polyurethane was used in the presented study. It is often used in cardiac implant investigations [
33,
34,
35,
36,
37], most often to produce vascular prostheses by electrospinning [
38,
39,
40]. Cylindrical structures were obtained using the phase inversion process. Various types of polyurethanes (differing in flexibility), different polymers concentrations, different nonsolvents, and different process times were investigated. Moreover, the influence of porogen addition on scaffold morphology and elasticity was studied. The cytotoxicity of the obtained scaffolds was also analyzed.
3. Discussion
Vascular prostheses can be obtained using various techniques. Fibrous materials are the most popular materials studied so far for use as vascular prostheses [
42,
43]. Fibrous vascular prostheses are fabricated to closely match natural blood vessel behavior. Electrospinning [
44,
45,
46,
47] and solution blow spinning [
48,
49] are most often used; these methods provide scaffolds with different fiber diameters and a high porosity, which are very good for substrates in terms of cell growth. Moreover, such materials are usually characterized by their high flexibility and breaking strength [
43]. However, special equipment is needed to manufacture fibrous scaffolds. In addition, selecting the most favorable parameters, such as flow, voltage, etc., can be very labor intensive and the process depends on many variables [
50]. The need to use highly volatile solvents is also a major limitation of the process. For some polymers, such as ChronoFlex, the selection of such a solvent is difficult. Our group also conducted research on the production of fibrous structures from ChronoFlex using the solution blow spinning (SBS) technique [
51,
52]. The only solvent that made it possible to obtain fibers using this method was 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP). The disadvantage of this solvent is its high cost; which forced us to look for alternative methods of obtaining cylindrical structures.
The phase inversion method is an interesting alternative; however, it also has disadvantages—the complexity of the process and possible uneven wall thickness—which can be counteracted using standardized methods, e.g., the extrusion process [
53] or thermal-induced phase inversion [
54,
55]. In addition, the differences between the various phase inversion techniques make it possible to achieve different microporosities [
3,
56]. As already mentioned, the phase inversion process depends on many parameters. The use of polyurethane and phase inversion methods to produce scaffolds that may be a base for vascular prostheses allows the obtaining of a material that meets the criteria for a suitable scaffold, i.e., good mechanical properties, biocompatibility, ease of manufacture, and appropriate morphology [
1,
57].
The primary purpose of the present research was to complement fundamental knowledge on the production of cylindrical polyurethane structures with surface pores using the phase inversion method. This allows investigating the effect of individual process parameters on the distribution and size of surface pores. The number of papers describing the impacts of various process parameters on pore distributions in the resulting structure is not sufficient, especially in terms of using ChronoFlex. In addition, an important aspect of the proposed study is the evaluation of material toxicity. The goal of this part of the work was to create structures that could be used as prostheses of blood vessels, so they should have a specific internal diameter, wall thickness that allows the cylindrical structure to be also maintained under dynamic conditions as well as appropriate flexibility and porosity. The inner surface of such materials deserves special attention. The importance of surface porosity in artificial blood vessels has been emphasized for many years [
58].
Our aim was to create a porous structure that would promote the restoration of the endothelial structure. For the prosthesis’ inner surface, the minimal recommended pore size was 10 µm, which will allow the formation of a monolayer of endothelial cells [
59]. Restoration of the endothelium will increase the hemocompatibility of such a surface and reduce the risk of blood clotting [
10,
60]. The outer side should also be uniform; cracks and tears are undesirable and indicate a structure’s poor mechanical properties.
ChronoFlex is often used in studies on materials contacting with blood [
19,
61,
62]. The presented research aimed to evaluate the selected parameters’ influence on polyurethane cylindrical scaffolds’ morphology. Khorasani and Shorgashti presented similar studies on the influence of parameters on the morphology of obtained flat surfaces. They investigated the effects of polyurethane concentration, process temperature and composition nonsolvent solution [
63]. Here, various types of ChronoFlexes at several concentrations, various nonsolvents and with different process times were tested. First, the influence of the nonsolvent composition on the morphology of the maintained structure was investigated. Six solutions that differed in water:ethanol ratios were tested. A comparison of all samples showed that 0:100 water/EtOH was the most preferred choice of nonsolvent. Selected variants had a microporous structure. In addition, the materials were free from cracks, and the artifacts that were found occurred in relatively small amounts. The flexibility of the samples, assessed manually, was also the best in these samples. It could be noticed that the higher the ethanol concentration in the nonsolvent solution was, the better the obtained structure was, which met the assumed criteria. The result obtained corresponded to that of Khorasani and Shorgashti [
63]; who studied water and various alcohols as nonsolvents. They observed that when an alcohol solution replaces water in nonsolvent, macrovoids are reduced.
Another analyzed process parameter was polymer concentration. At low analyzed polymer concentrations, the system was a low viscosity liquid [
56]. Lower concentrations (10% and 15%) were inadequate as the polymer dripped from the matrix, and thin walls were obtained with very uneven thicknesses. At 25% PU, the inner surface had more surface porosity; however, the outer surface material was also more porous. A 25% PU solution was too dense, so material walls were uneven, and the materials were also stiff. Therefore, we decided to use 20% PU. The literature reports that, as a polymer concentration increases, surface porosity decreases [
63,
64]. However, in the experiments presented here, a smaller surface porosity was observed on the surface of materials made using lower tested concentrations. The differences probably resulted from the nonsolvent used, which proves how strongly it affected the resulting structure.
With the longest tested process time, i.e., 24 h, the materials had the most favorable properties. During this process, structures were obtained without defects (interruption of the wall), as well as with the most favorable surfaces for both the inside and outside. The dependencies were similar for all three analyzed polyurethanes.
As mentioned above, the three analyzed polymers differ in terms of Shore hardness. According to this scale, the C75A polymer is a soft polymer while C45D and C75D are hard ones. This is in line with the results where C75A materials were the most flexible and C75D materials were the stiffest.
When analyzing the surface morphology of the obtained materials, a set of parameters was selected: 0:100 water/ethanol, 20% PU concentration, and 24 h process time. After analyzing the mechanical properties of the materials from the three tested PU materials, it was decided that PU C45 D would be used for tests with a porogen.
The addition of a porogen met the expectations—the surface porosity was increased by about 9%. The average pore diameter also increased. However, there was a deterioration in the outer surface morphology; it was more folded and had many more surface irregularities. The washed-out porogen produced larger voids, resulting in a more uneven wall thickness. The wall thickness of the obtained structures was similar to that found in literature reports [
9]. The use of pre-mixed polymer and the porogen can often result in irregular pore shapes while retaining residual porogen in the structure [
65]. Therefore, by changing the type of porogen (its size), the pore size can be controlled [
64,
66]. In the studies presented here, small porogen residues can be seen in the obtained materials, hence we observed stiffening (changes of mechanical parameters) of the structure and an increasing wall thickness. Ahmed et al. [
67] showed that porogen addition (NaHCO
3) stabilized polyurethane cylindrical structures obtained via phase inversion, with water as a nonsolvent.
Vascular prostheses must have adequate mechanical properties to withstand blood pressure. In addition, they should be resistant to deformation and compression and have sufficient tensile strength to resist tensile loads when implanted into the body [
68]. Our study showed that scaffolds with the addition of porogen were stiffer and less flexible than those without it. PVA addition lowered the mechanical properties of the materials. This result is not surprising since the greater the material’s porosity, the lower the scaffold’s mechanical properties [
69]. The greater variability between samples and the higher standard deviations were due to the materials’ pore sizes and the porosity’s random nature. One of the most important mechanical parameters to be remembered when constructing vascular prostheses is the Young modulus [
70]. The Young’s modulus for natural arteries is in the range of approx. 1.0–3.4 MPa [
70,
71], which is comparable to the PU C45D material. The reduction in tensile strength of the material with porogen corresponds with the work of Ahmed et al. [
67]. The presence of macrovoids can reduce the mechanical properties of the materials [
72].
Fibroblasts and endothelial cells are the most commonly used for cytotoxicity tests, and the ISO standard suggests L929 cells for these tests [
73]. MTT analysis confirmed the lack of cytotoxic effects on cells of the material made of three examined polymers and the material obtained with 10% PVA. This was an entirely expected result. The PU used for preparation of the scaffolds was of medical grade. In the materials’ pores, no solvent or porogen remained, which would have leaked into the extract, giving a negative result.
Cylindrical scaffolds obtained using the phase inversion technique tend to be a promising material in testing for use as an artificial blood vessel. The next stage of research related to the obtained scaffolds is the analysis of material-blood interactions and the adhesion of human endothelial cells.
4. Materials and Methods
4.1. Materials
Polyurethane, ChronoFlex C75A, C45D and C75D, was bought in the form of pellets, from (AdvanSource Biomaterial, Wilmington, MA, U.SA). N,N-dimethylacetamide (DMAC), sodium chloride (NaCl, 99%), polyvinyl alcohol (PVA), were purchased from (Sigma Aldrich, Poznań, Poland). Hexane and ethanol (EtOH) were purchased from (POCH, Gliwice, Poland).
4.2. Preparation of Polyurethane Scaffolds—Selection of Process Parameters
The research was carried out in three parts for each type of tested polyurethane. First, the variable was nonsolvent concentration, second—the PU concentration, and third—the time of the process. Three types of polyurethane were used each time (
Figure 12).
Three types of PUs that differed in hardness were analyzed: C75A, C45D and C75D. Polyurethane pellets were washed with 70% EtOH/water solution, dried to constant weight at 40 °C and dissolved in DMAC to a given concentration. Four concentrations of PU solutions were examined: 10% (
w/
v), 15% (
w/
v), 20% (
w/
v) and 25% (
w/
v). Afterward, the stainless-steel matrix (with 6 mm diameter) was dipped in a PU solution and then immersed in a nonsolvent solution for a given time. Six nonsolvent solution differ in water/EtOH ratios were analyzed: 100:0 water/EtOH, 20:80 water/EtOH, 40:60 water/EtOH, 60:40 water/EtOH, 80:20 water/EtOH and 100:0 water/EtOH. Three process times were examined: 10 min, 2 h and 24 h. The resulting samples were removed from the nonsolvent solution, taken off the metal matrix, and allowed to dry at room temperature (RT) at high humidity. The scheme of the manufacturing process is presented in
Figure 13.
4.3. Porogen Addition
In order to increase the number and size of surface pores, prostheses with the addition of a porogen were fabricated. Three porogens, namely NaCl (5% w/v, 10% w/v), PVA (5% w/v, 10% w/v) and hexane (5% v/v, 10% v/v) were selected. Porogen was added in given concentrations to the polymer solution. After porogen addition, the polymer solution was thoroughly mixed to distribute the porogen evenly throughout the polyurethane volume. Then, the prostheses were manufactured as described above.
4.4. Surface Characterization
The morphology of the obtained structures was examined with a scanning electron microscope (SEM, Phenom G1, Phenom World, Eindhoven, The Netherlands). Rectangular fragments were cut out from each cylindrical structure (
n = 4). The internal and external surfaces were analyzed. Additionally, surface pore sizes were measured manually (
n = 100 per variant) and wall thicknesses (
n = 5, in 3 different spots) for materials considered to be most advantageous in terms of morphology. The measurements were performed based on SEM images with Fiji software [
74].
For selected the materials, the surfaces and cross-sections were analyzed with a stereoscopic microscope (Leica, Wetzlar, Germany). The internal diameter was measurement manually using ImageJ.
4.5. Mechanical Testing
Mechanical properties were tested for the selected materials. Cylindrical structures (4 mm inner diameter, 60 mm length; n = 5) were subjected to a uniaxial stretching test according to protocols based on ASTM standards (D 638-02a and 882-02). The analyses were carried out using an Instron 3345 model with 5 mm∙min−1 head speed at RT and ambient humidity.
4.6. Porosity
Material porosity was determined using the gravimetric method [
51,
67]. The material porosity was calculated on the basis of its apparent density (ρ
app) and known density of the polymer (ρ
p = 1.2 g/cm
3 [
75]), according to the following formula:
where values of ρ
app were measured from the dimensions and weights of the materials (
n = 5).
4.7. Cytotoxicity Evaluation
Cytotoxicity testing, MTT cell proliferation assay (Thiazolyl Blue Tetrazolium Bromide, Sigma-Aldrich, Poznań, Poland), was performed according to ISO 10993-5 [
41]. L929 cells were cultured with Dulbecco’s Modified Eagle Medium (DMEM, ThermoFisher, Waltham, MA, USA) supplemented with bovine serum (10%
v/
v, ThermoFisher) and a mixture of penicillin–streptomycin antibiotics (1%
v/
v, ThermoFisher) in an incubator (37 °C, 5% CO
2).
On the first day, cells were seeded at a density of 1·104/per well and were grown for 24 h. Sterilized materials (using 70% ethanol, washed three times with sterile PBS) (n = 5) were incubated in DMEM medium for 24 h. After this time, material extracts were added to the cells and incubated for another 24 h. The negative control was cells with DMEM medium, and the positive control was cells treated with 0.1% Triton X. After removing the extracts from the wells, MTT solution (1 mg/mL DMEM, ThermoFisher, Waltham, MA, USA) was added. The plates were placed in an incubator for 4 h. After this time, the solution was removed and isopropanol was added to dissolve the formazan crystals. Absorbance was measured at a wavelength of 570 nm.
Cell viability was calculated using the following formula:
where A
S is the mean absorbance value of the sample and A
C is the mean absorbance value of the negative control. Cell viabilities are presented against negative control (which is assumed to be 100%).
4.8. Statistical Analysis
The results of the surface pore diameters, wall thicknesses, porosity, mechanical testing and cell viabilities were expressed as means ± SD. Statistically significant differences were investigated with a single-factor analysis of variance (ANOVA) for p < 0.05 with a post hoc Tukey’s test (OriginPRO 2020b, OriginLab Corporation, Northampton, MA, USA).