Conducting colloidal dispersions
Colloidal dispersions of CP stabilized with polysaccharides are smart formulations helping to solve the problem of the poor solubility of CP in aqueous environment, and improve the cytocompatibility of the final composite 27. Another, not previously mentioned advantage of such composites appears in the case when CPs are mixed with other hydrophilic polymer matrices – for example, when polymer-based scaffolds are fabricated. Here, the better polymer-polymer compatibility of HaPANI or HaPPy colloids containing HA combined with the bulk polymer of the scaffold (also based on HA) is expected in comparison with pristine PANI or PPy.
The size and polydispersity index of colloidal particles used for scaffold fabrication were determined on thoroughly dialysed samples, purified of possible impurities which might cause cytotoxic effects 28. The hydrodynamic diameters of particles were different, with micrometer sizes observed for HaPANI (2200 ± 260 nm) and smaller sizes determined for HaPPy (447 ± 8 nm). Both samples showed, however, similar polydispersity indexes (~0.2). The morphology of particles observed by transmission electron microscopy (Figure 1A and 1B) conformed with scattering measurements (Figure 1C and 1D), showing in both cases regular, spherical particles with a smooth surface, and, in the case of HaPANI, the interconnection of such particles to form bigger, loose clusters.
Preparation and characterization of scaffolds
The CP-based colloidal dispersions HaPANI and HaPPy were used, each incorporated into the used HA matrix and in three different concentrations. Two methods of scaffold preparation were employed, each followed by an identical freeze-drying procedure. The first route included chemical crosslinking with EDC and NHS, the latter used as the activator of carboxylic groups; the second procedure involved crosslinking by the freeze–thaw method using PVA, hereinafter referred to as the physical route. The question, hence, arose whether – and if so, how – these two preparation methods together with scaffold composition affect the material and biological properties of the given scaffolds.
Mechanical properties defined in terms of the modulus of elasticity (E) were significantly different for samples of HaPPyPh crosslinked by freeze-thawing in the presence of PVA when compared with chemically crosslinked HaPPyCh and scaffolds containing HaPANI colloidal particles (Figure 2). Their E values were the highest (17 000 – 50 000 Pa) and strongly depended on the content of PPy particles in the samples, with the highest value observed for HaPPyPh_108. A corresponding dependence of the elasticity modulus on the content of colloidal particles was seen for chemically crosslinked HaPPyCh, though with significantly lower E of only 5 000 – 17 000 Pa. This indicates a stronger network for the samples containing PVA and good mutual compatibility between all the involved components, HA, PVA, and CP-based colloidal particles, of which the last component contributed to the good level of compatibility through the acidic nature of the particle dispersion 29. Differences between physically and chemically crosslinked samples containing HaPANI (Figure 2) were only minor in comparison with the corresponding scaffolds prepared with HaPPy. The modulus for physically crosslinked HaPANIPh ranged from 5500 to 15000 Pa, with the lowest value observed for the sample with 0.054% (w/w) colloid (HaPANIPh_54). This E range is actually similar to the E range observed for chemically crosslinked samples (HaPANICh), i.e., from 5000 to 15000 Pa; however, the correlation between the E values and the amount of colloidal particles is different in these samples. In summary, the scaffolds incorporating HaPANI colloidal particles exhibited a low elasticity modulus, which moreover lacked correlation with the colloid content in the sample. In contrast, the behaviour of all scaffolds containing HaPPy colloid was more predictable in terms of E vs particle content. This is surprising, as one would expect more similar behaviour among scaffolds containing PVA irrespective of the type of colloidal particles used, as PVA contribute to a rubbery and elastic texture and higher mechanical strength of scaffolds thanks to the better distribution of the mechanical load along the crystallites of the three-dimensional scaffold structure 30. The mechanical properties of the scaffolds hence result from the crosslinking routes and the content/type of colloidal particles used for their fabrication.
Surface topography and electrical properties
Lyophilized samples with a medium concentration of the conducting component (0.054% w/w) were analysed using AFM to determine their surface properties and surface conductivity (Figure 3).
The surface roughness (Sa) and maximum surface height (Sz) of the samples as presented in figure 3 shows that the surface characteristics of the scaffolds are rather similar. Their surfaces are irregular with embedded pores, these seen mainly in physically crosslinked samples. Chemical crosslinking gave rise to scaffolds with a flake-like surface structure. The second and third columns of Figure 3 show current maps expressed as TUNA and Peak currents. These unambiguously prove that the scaffolds are conductive and that their average values of TUNA current are similar, lying in the range of 2.1 – 3.8 pA, irrespective of the type of conducting particles used and of the fact that the percentage of the conducting component in the chemically crosslinked samples was higher with respect to the total mass of the scaffold than in the scaffolds crosslinked physically. A bigger difference in the measured values can be seen for maximum currents (max TUNA, max Peak current), which were lower for physically crosslinked samples, with a maximum current of about 12 pA. In the case of samples crosslinked chemically, values of 15 and 19 pA for HaPANICh _54 and HaPPyCh _54, respectively, were measured. This minor difference is a natural consequence of the higher ratio of CP-based colloids in the chemically crosslinked samples. In this respect, both HaPANI and HaPPy colloids performed, within the same type of crosslinking mechanism, similarly. Obviously, as the samples contain particles of intrinsically conducting polymer (CP) with mixed ionic and electron conductivity, and as the surroundings within hydrated scaffolds are also ionically conductive, the scaffolds thus form a suitable environment which can facilitate the adhesion and proliferation of electrically conductive cells and tissues 20.
Porosity and morphology
Morphology, including the size of pores in lyophilized scaffolds, was assessed by means of the visual observation of SEM images (Figure 4). The images were used for the measurements of pore size and the calculation of porosity, which were both quantified according to the chord length method given by the protocol of ASTM E-112 (Figure 5).
According to the results of the image analysis, the physically crosslinked HaPPyPh samples with the two lower concentrations of colloidal particles (0.036, 0.054%) exhibited the biggest pores in the scaffolds. The pore sizes were the smallest in the formulation in which the colloid concentration was 0.108% (HaPPyPh_108).
In numerical values, the porosity ranged from 26 – 37%, with the lowest value for the highest content of colloidal particles, as observed for HaPPyPh_108. The HaPPyCh samples, prepared with chemical crosslinking, yielded a lower porosity of 22 to 28%, which was lowest for the scaffold with the medium content of colloidal particles (HaPPyCh_54). The physically crosslinked HaPANIPh samples were, with respect to pore sizes, similar to the above-discussed samples containing PPy colloidal particles, and showed pore sizes of 271, 288, and 177 µm for scaffolds with increasing concentrations of CP-based colloid. Calculated porosity values for these formulations ranged between 27 and 32%. HaPANICh scaffolds, prepared by chemical crosslinking, exhibited porosity values comparable to those of scaffolds crosslinked physically, with one exception: the sample HaPANICh_36. This outlying sample contained the lowest concentration of HaPANI colloidal particles and exhibited a porosity of more than 50% and an average pore size of 547 ± 126 µm. The porosities of the other two HaPANI-based samples were 30% and 35% for the medium and highest colloid contents, respectively. According to the image analysis supported by visual evaluation, the bigger pores were always seen in samples with the lowest content of colloidal particles and, in general, the sizes of pores in most of the scaffolds were roughly estimated to be within the range of 200 – 300 µm, which is sufficient for the ingrowth of cells into the structure and for nutrient supply and metabolite removal.
The course of swelling was similar for all samples, and the initial increase in swelling degree (SD), observed within the first approx. 100 min, was followed by equilibrium (Figure 6). At equilibrium, the SD of physically crosslinked HaPPyPh scaffolds ranged from 92 to 96% and decreased with growing colloid content in the sample. In comparison, differences in the swelling of all chemically crosslinked HaPPyCh samples were only minor and their equilibrium SD was 93 – 95%. Scaffolds containing HaPANI exhibited lower SD for physically crosslinked samples in comparison with samples that were crosslinked chemically. With short swelling times, the lowest SD of 87% was recorded for HaPANIPh_54, while the highest equilibrium SD of 93% was observed for the sample containing the highest concentration of particles. Chemically crosslinked HaPANICh scaffolds swelled better, their SD ranging from 95 to 97%. In this respect, SD is influenced by PVA in the hydrogel matrix; in chemically crosslinked samples with PVA absent, the naturally high ability of HA to take up water prevailed and the SD was, therefore, higher. When in a mixture, the crystallinity of PVA affects the hydrogel structure, which is less loose compared with samples containing solely HA. As a result, physically crosslinked gels absorb less water, this resulting in lower SD. Nevertheless, regardless of the observed differences, the SD values were similar and all scaffolds showed good swelling characteristics.
The absence of cytotoxicity is one of the fundamental requirements that biomaterials must meet. The cytotoxicity of a full set of samples was therefore tested. To show representative information, the protocol of ISO 10993-5 standard was employed, including the use of an appropriate cell line; i.e., NIH/3T3 fibroblasts. The cytotoxicity was determined on native samples prior to lyophilisation and on lyophilised scaffolds, which are marked with the superscript L.
The summary of cytotoxicity data recorded for chemically crosslinked scaffolds and presented in Figure 7 unambiguously demonstrates that all the native scaffolds prepared by chemical crosslinking, HaPANICh or HaPPyCh, can be classified as non-cytotoxic, with cell viabilities higher than 70%. The situation was, however different for lyophilised scaffolds. After lyophilisation, the cytotoxicity increased, especially in the case of HaPPy containing scaffolds, the lyophilised HaPPyCh_L scaffolds showing cytotoxic effects with concentrations of extract higher than 25%, irrespective of the concentration of the colloidal particles used. In the case of HaPANICh_L containing HaPANI colloid, the cytotoxicity was lower and only the two highest concentrations of extracts (70 and 100%) exceeded the threshold for cytotoxicity, and moreover only slightly. The effect of lyophilisation on cytotoxicity can be related to the protocol of testing given by the ISO standard. The procedure defines the ratio of material/extraction medium according to the mass. In the case of lyophilised samples with water removed, the ratio of the mass of dry matter (which in fact contains impurities) to the amount of extraction medium is higher. This situation can lead to the higher cytotoxicity of lyophilized samples, despite the fact that the procedure itself should not produce additional cytotoxic compounds.
The behaviour of physically crosslinked samples (Figure 8) prior to lyophilisation was similar to that observed for scaffolds crosslinked chemically, and only in a few cases did the samples show a cytotoxic effect, which was, moreover, observed at only the highest extract concentration. Only the highest extract concentration of native HaPANIPh scaffold significantly exceeded the cytotoxicity threshold, and the effect correlated with the concentration of HaPANI colloidal particles in the sample. Contrary to the situation observed for scaffolds prepared by chemical crosslinking, here lyophilisation decreased the negative impact of the scaffolds on cell viability. This could actually be observed in the case of all samples, especially in the case of HaPANIPh_L, where none of the lyophilised hydrogel samples was cytotoxic.
The difference in the behaviours of lyophilized scaffolds prepared by chemical and physical crosslinking could, to some extent, be expected, as physically crosslinked scaffolds lack the possible residues of cytotoxic crosslinking agents, which can be released during the lyophlization process. We can also hypothesise that the difference is related to the different pore structures and the type of bound cytotoxic residual compounds within the scaffold structure. It seems that the CP-based colloids presented here do not substantially contribute to the overall cytotoxicity of the scaffolds, either in the case of HaPANI or HaPPy colloidal particles. Indeed, this is not surprising, as previous studies by Humpolicek et al have shown that the cytotoxicities of both CPs are similar 31 .
Cell adhesion, growth and ingrowth
The cytotoxicity of scaffolds, which can be related to the composition of the material and the possible leaching of residual precursors or reagents under extraction was low. The lyophilized samples were therefore subjected to another cytocompatibility study, this involving the determination of bio-interface properties represented by the ability of cells to adhere onto the scaffold surface and subsequently grow. The lyophilized samples were chosen, as their internal architecture was more appropriate for cell ingrowth than that of native scaffolds. The cells were seeded on the surfaces of scaffolds and cultivated for one week. After this time elapsed, the cells were able to adhere and grow on all surfaces. Representative pictures are presented in Figure 9.
Scaffolds suitable for tissue engineering must not only allow the attachment and growth of cells on their surfaces, but also be able to facilitate cell ingrowth into the structure of the scaffold. As mentioned above, an increasing concentration of HaPANI or HaPPy colloidal particles within the scaffold resulted in the higher cytotoxicity of the scaffold. The ingrowth of cells, allowed by their cultivation in a bioreactor mimicking in vivo conditions, was therefore tested only on samples with the lowest concentration (0.036%) of HaPANI and HaPPy (Figure 10). According to the results, it is obvious that cells were able to efficiently ingrow into the scaffold structure. This test showed not only that the scaffolds prepared here were cytocompatible with respect to basic parameters such as absence of cytotoxicity, but also that their bulk architecture created a friendly environment for cell ingrowth; thus, such scaffolds can be suitable for use in tissue engineering. Furthermore, their inherent conductivity, as a cell-instructive property, assured by the presence of CP-based colloids, opens up their potential use in a wide range of applications relating to the tissue engineering of electro-sensitive tissues.