The first systematic analysis of 3D rapid prototyped poly(ε-caprolactone) scaffolds manufactured through BioCell printing: the effect of pore size and geometry on compressive mechanical behaviour and in vitro hMSC viability

Morphological analysis of the 3D rapid prototyped structures was carried out using a scanning electron microscope (SEM, FEI Quanta 600 F) to visualize and evaluate the physical integrity of the material filaments and layers, as well as verifying if the pore size and geometry obtained were consistent with the theoretical values imposed during the fabrication process.

The porosity was evaluated using the following methodology: (1) measuring the weight and volume of each sample, (2) calculating the apparent density of the PCL scaffolds, (3) applying the formula reported below:

$$\begin{equation} {\rm Porosity} = \bigg(1 - \frac{{\rho *}}{{\rho _{{\rm sub}} }}\bigg) \times 100 \end{equation} \tag{ 1 }$$

where ρ* is the apparent density of the cellular structure (scaffold) and ρ_sub is the density of the original substance (ρ_sub = 1.145 g cm⁻³). Five samples were measured for each scaffold morphology [63].

Compression tests were performed on the 3D rapid prototyped scaffolds in order to highlight the effect of the FD, i.e. centre-to-centre distance (550, 650 and 750 µm) and lay-down pattern (0°/90°, 0°/60°/120° and 0°/45°/90°/135°) on the mechanical properties.

Each block-shaped specimen was characterized by a length (l) of 5.0 mm, a width (w) of 5.0 mm and a height (h₀) of 8.0 mm. All the tests were carried out on scaffolds at a rate of 1 mm min⁻¹ up to a strain value of 0.5 mm mm⁻¹, using an INSTRON 5566 testing system.

The 'apparent' stress σ was evaluated as the force F measured by the load cell divided by the total area of the apparent cross section of the scaffold (A = l w):

$$\begin{equation} \sigma = \frac{F}{A} \end{equation} \tag{ 2 }$$

while strain ε was defined as the ratio between the scaffold height variation (Δh) and its initial height (h₀):

$$\begin{equation} \varepsilon = \frac{{\Delta h}}{{h_0 }}. \end{equation} \tag{ 3 }$$

Mechanical results were reported as mean ± standard deviation and were analysed using the analysis of variance followed by Bonferroni post-hoc comparisons. Statistical differences were set at p < 0.05.

Cell-seeding experiments were carried out using hMSCs (Clonetics, Italy) to provide insight into the effect of pore size and shape on cell adhesion and viability/proliferation.

PCL scaffolds were cut into disc shaped blocks with appropriate dimensions in order to fit inside the 12 well plate, sterilized with a 70% ethanol/water solution for 24 h, washed extensively with PBS 0.01 M pH 7.4 and exposed to UV light for 40 min.

Afterwards, 17 × 10³ cells were re-suspended in 200 µL of α-modified Eagle's medium (α-MEM) (Bio-Whittaker, Belgium), statically seeded directly onto the PCL scaffolds and allowed to proliferate for 21 days. At pre-determined intervals (7, 14 and 21 days) the samples were withdrawn from the incubation media, rinsed with PBS (Sigma-Aldrich, Italy) and analysed by means of an Alamar Blue™ assay as described in previous works [26].

Cell morphology was investigated by confocal laser scanning microscopy (CLSM, Zeiss LSM 510/ConfoCor 2). After the permeabilization step with Triton X-100, samples were incubated with a PBS 0.01 M solution of Phalloidin for 45 min at room temperature. After the dyeing incubation, scaffolds were washed with PBS before being mounted on a glass slide and sealed with resin for microscopic observation.

Confocal images obtained from all the cell-scaffold constructs were also suitably analysed through Image J software (National Institutes of Health, USA) to assess the cell morphology using a shape factor.

The shape factor was expressed as follows:

$$\begin{equation} \Phi = \frac{{\Theta A}}{{P^2 }} \end{equation} \tag{ 4 }$$

where Θ represents a geometric constant (4π) related with the footprint area and P is the perimeter of a cell. Considering that circular objects have the greatest area-to-perimeter ratio, a shape factor of 1 represents a perfect circle. Conversely, a thin thread-like object possesses the lowest shape factor approaching zero [64].

PCL scaffolds with different pore sizes and geometries were designed and produced via bioextrusion. The possibility of varying the lay-down pattern allowed us to control the pore geometry, while pore size and porosity were mainly defined by the variation in the FD.

SEM micrographs performed on PCL scaffolds with a single lay-down pattern of 0°/90° and different FD values (550, 650 and 750 µm) allowed us to confirm that the structures present a well defined internal geometry with square pores in the range of 245–433 µm (figure 4).

Zoom In Zoom Out Reset image size

As illustrated in table 2, a significant consistency between the theoretical values and the real values of filament diameter (RW), filament gap (FG) and layer gap (LG) was achieved for all of the rapid prototyped scaffolds. Porosity was calculated based on equation (1); the values ranged from 49% to 57%.

In terms of pore shape, three lay-down patterns were adopted (0°/90°, 0°/60°/120° and 0°/45°/90°/135°) maintaining a constant FD value of 650 µm, resulting in scaffolds with quadrangular triangular and complex polygonal internal geometries, as confirmed by the SEM micrographs (figure 5).

Zoom In Zoom Out Reset image size

By increasing the number of deposition angles from 2 (0°/90°) to 4 (0°/45°/90°/135°) it was possible to generate more complex pore geometries without a significant variation of the structure's porosity, as illustrated in table 3.

Independently of the pore size or geometry, all the scaffolds produced were characterized by a fully open pore structure and good adhesion between adjacent layers.

Typical stress–strain curves obtained for the 3D rapid prototyped scaffolds are reported in figure 6, which graphically highlight the effect of FD and lay-down pattern on the mechanical performances of the constructs.

Zoom In Zoom Out Reset image size

The mechanical behaviour of the 3D rapid prototyped scaffolds results is similar to that of a flexible foam [73, 74]. A linear region is initially well evident, thus suggesting a stiff mechanical response at the onset. This region is then followed by a region with lower stiffness. Finally another stiff region of the stress–strain curve can be observed that resembles the densification region usually reported for flexible foams [73].

As already shown by previous works [41, 63, 74], unlike the typical behaviour of flexible foams [73] and 3D scaffolds manufactured through fused deposition modelling [75], the central region of the stress–strain curve does not evidence a plateau, which means a zero slope region, it is just characterized by a lower stiffness when compared with the other two regions of the graph.

The compressive modulus (E) has been calculated as the slope of the initial linear region of the stress–strain curve. Values of compressive modulus and maximum stress are reported as mean value ± standard deviation (tables 4 and 5).

As expected, at fixed lay-down pattern and processing parameters, the FD strongly affects the porosity and consequently the mechanical behaviour of 3D rapid prototyped structures. In particular, as the FD increases from 550 to 750 µm (porosity increases from 49% to 57%), the compressive modulus and maximum stress decrease from 52.5 ± 4.5 MPa to 23.1 ± 2.8 MPa and from 8.7 ± 1.2 MPa to 2.7 ± 0.5 MPa, respectively (table 4). In terms of the compressive modulus and maximum stress, the observed differences are statistically significant. Our results are in agreement with previous works performed by other researchers and confirm the porosity dependence of 3D scaffolds when tested under compression [61, 73, 76]. A plausible explanation for the variation in mechanical properties as a function of FD, independently of the material, can be attributed to the column-like behaviour of the filament junctions when undergoing compression. In other words, for a specific area, the scaffold with a smaller distance between filaments has a maximum number of columns, hence increasing the stiffness of the scaffold [76].

On the other hand, at selected FD and processing parameters, the lay-down pattern influences the mechanical behaviour of the scaffolds. As reported in table 5, scaffolds characterized by a 0°/90° (2-angle) pattern exhibit a compressive modulus (34.2 ± 3.8 MPa) and maximum stress (5.6 ± 0.2 MPa) which are greater than those obtained for 0°/60°/120° (3-angle) and 0°/45°/90°/135° (4-angle) patterns. In this case, a decrease in the amplitude of the deposition angle (from 0°/90° to 0°/45°/90°/135°) between struts of adjacent layers implied a larger contact area (fused area), leading to a decrease of the local stress experienced by the 3D structure [60]. Thus, when the scaffolds are compressed, the filaments deposited with a smaller lay-down pattern (0°/45°/90°/135°) can easily slide from each other increasing the scaffold deformability [75]. Anyway, only the differences in terms of the compressive modulus are statistically significant between 3D scaffolds with a 0°/45°/90°/135° lay-down pattern and the other architectures analysed in the work.

In the case of weight-bearing applications such as bone, 3D scaffolds are required to withstand mechanical stresses, both in vitro as in vivo, during a specific period of time, providing adequate pore size and interconnectivity for cell adhesion/proliferation, vascularization and tissue ingrowth. Besides, a scaffold that fails to match the mechanical properties of the native tissue may cause tissue resorption and anomalous growth (if the scaffold has higher mechanical properties than those of the natural tissue) or the complete failure of the scaffold (due to insufficient strength to withstand physiological loads) [19]. As reported by Yang , the values of the compressive strength and the young modulus of human cancellous bone are 4–12 MPa and 0.02–0.5 GPa respectively, which suggests that PCL scaffolds produced via BioCell Printing may also have a great potential in bone TE applications [21].

Finally, as suggested by some authors, the mechanical properties of some synthetic polymers may vary greatly when tested under wet conditions (water, PBS, etc) and over a range of temperatures [75, 77, 78]. From our analysis, no relevant change in the mechanical properties was observed when 3D PCL scaffolds pre-conditioned in PBS for 48 h were tested under compression at 37 °C (data not shown). A plausible explanation for these results lies in the fact that the PCL is highly hydrophobic and crystalline, which presents greater resistance to the plasticizing effect caused by PBS or water. Further support to this observation was provided through studies previously conducted by our group, where we found that PCL scaffolds conditioned in PBS or SBF for six months, kept their physical and structural integrity almost intact [79].

The ultimate goal of a TE scaffold is to closely mimic the native tissue environment, providing adequate chemical, physical and biological cues that will guide cell adhesion, proliferation and differentiation. According to the cell type, scaffold requirements, such as pore size, shape and porosity, may change; therefore, the scaffold should be tailored in order to better accommodate the growth of specific cells [19]. Moreover, the cell colonization of 3D porous matrices is strongly dependent on the ability of cells to bridge pores and spread into the scaffold. If the pores are too large or too small, cells will fail to spread and form networks throughout the scaffold. Therefore, it becomes important to define an optimal pore size range for supporting cell ingrowth [65].

Biological experiments were carried out on PCL scaffolds using hMSCs, to provide insight into the effect of pore size and shape on cell viability, morphology and migration. Cells were seeded at an appropriate density (17 × 10³ cells re-suspended in 200 µl) directly onto the PCL scaffolds, regardless of pore size and shape, and allowed to proliferate for 21 days under static conditions. At pre-determined intervals of 7, 14 and 21 days the scaffolds/cells were withdrawn from the incubator and a quantitative evaluation of cell viability was performed by means of Alamar Blue™ assay. The results showed a strong influence of pore size and geometry on cell viability (figure 7).

Zoom In Zoom Out Reset image size

In general, 3D matrices should be characterized by an internal architecture with sufficiently large and fully interconnected pores to enable cell migration, transport of oxygen and nutrients, metabolic waste removal and vascularization, which are essential requirements in terms of cell vitality. The particular mechanism that promotes or blocks cell attachment, proliferation and differentiation of cells in scaffolds with different pore sizes and shape is not yet fully understood. Quantitative biological data depicted in figure 7(a) reveal that, independently of the pore size, all constructs were able to maintain a high level of cell viability during the incubation period (21 days). Nonetheless, substantial differences in terms of Alamar Blue™ reduction verified at day 21 suggest that scaffolds with larger FDs (FD = 750 µm and pore size of 450 µm) are able to accommodate a higher number of viable cells when compared to scaffolds with smaller FDs (FD = 550 µm and pore size of 250 µm). These differences in terms of cell viability may be related with the larger surface area available (higher porosity) for cells to adhere and proliferate inside the scaffolds with a larger pore size. Furthermore, cell activity appears to be enhanced by the more effective transport of nutrients and oxygen that occurs within scaffolds with larger internal channels.

Regarding the influence of pore shape on cell viability, results depicted in figure 7(b) clearly show an increment on cell viability with the decreasing number of deposition angles. Despite this, a difference can be detected throughout the entire incubation period, which becomes more evident after 21 days of cell culture.

Based on SEM images (figure 5), all specimens exhibited a fully open internal pore architecture, independently of the deposition angle, that in itself should promote an equal cell colonization of the structures. Therefore, the differences found in terms of cell viability according to pore shape can be explained by the angle amplitudes between the filaments that geometrically define the pores in the XY plane of the fabrication process. Our observations reveal that by passing from a complex polygonal shape with 45° amplitude between filaments to quadrangular pore geometry with 90° amplitude, the distance between filaments in the same plane is increased thus improving cell accessibility and consequent colonization of the inner part of the scaffolds. Similar work performed by Hoque reports no significant differences on the proliferation of rSMC cells seeded onto PCL and PCL–PEG scaffolds with different pore shapes, after three weeks in culture [80]. The authors attributed this fact to the complete open pore structure of all three patterns that favoured the nutrient flow in and wastes flow out. This apparently contradictory result needs to be interpreted with caution as the findings of the authors were based on experiments performed with different cells (rSMC) and 3D scaffolds designed/produced applying a higher FD that resulted in different pore geometries from those presented in our study.

During the adhesion process, large pores will induce cells to attach only to a single strut hence feeling the stimuli of the surface, while in small pores, cells can attach to several struts, thus resulting in multiple stimuli. Looking at the scaffolds with smaller pore size (FD = 550 µm) it is possible to observe that the struts in the same plane are much closer, thus creating multiple contact points for the cell to attach to and to eventually differentiate. Additionally, scaffolds with larger pores also possess a higher porosity providing cells with more open space to adhere and proliferate.

Parallel investigations of cell morphology were carried out by confocal laser scanning microscopy (CLSM) in order to highlight the effect of pore size and shape on the adhesion and proliferation of cells on the filaments of the scaffolds (figure 8). Based on our observations, it was not possible to detect significant differences in cell morphology between the different pore topologies. Taking into account that both surface material chemistry and topography affect cell adhesion and proliferation, the similar cell morphology may be ascribed to the fact that the fabrication technology and the material employed to manufacture all the structures were the same.

Zoom In Zoom Out Reset image size

The CLSM images allowed us to confirm the quantitative results obtained with Alamar Blue^TM. As it becomes clear from figure 8, the number of viable cells adhering and proliferating onto the filaments of the scaffolds strongly increases from day 7 to day 21. Cell morphology also varied along the incubation period changing from a circular geometry with a few ramifications (figure 8(a)) to a homogeneous filamentary geometry with an increased number of ramifications (figure 8(c)). This variation indicates an increment in terms of cell adhesion through the establishment of a higher number of cell-scaffolds and cell–cell interactions.

Additional studies of cell adhesion and spreading were performed based on CLSM images with the aim of determining the shape factor of the cells. The shape factor was calculated for all pore topologies based on equation (4). Since no relevant variations were observed between the different pore topologies, we decided to report only the values of the shape factor for scaffolds with quadrangular pores (0°/90°) and FD = 750 µm. Typical values of the shape factor at 7, 14 and 21 days after cell seeding time are reported in figure 9 as mean value ± standard deviation.

Zoom In Zoom Out Reset image size

As it shows in figure 9, the mean cell shape factor decreased significantly from 0.31 ± 0.04 at day 7 to 0.08 ± 0.01 at day 21. Based on previous experiments performed by other authors, it is possible conclude that the lower the cell shape factor, the more elongated the cells [64]. In other words, the establishment of multiple cellular extensions (increased total cellular area) implies a reduction in the shape factor (elongated cells) leading to better adhesion and spreading process.

The output of this study reveals a strong influence of pore size and shape on the adhesion, viability/proliferation of hMSCs seeded and cultured on PCL scaffolds over 21 days. Our findings suggest that large quadrangular pores enhance in vitro hMSC proliferation. As such, we envisage the combination of large quadrangular pores with small triangular pores within a single scaffold as a possible solution towards the design of a functional graded structure capable of promoting optimal tissue regeneration. We believe that such structures should have small pores on the outside (where the access to oxygen and nutrients is much easier) and large pores inside (enabling access to oxygen and nutrient supply) thus forcing cells to migrate to the inner part of the scaffolds and promoting a homogeneous colonization.

Besides these findings, it becomes clear that the information available in the literature regarding the optimal pore size and shape for the adhesion and proliferation of specific cell lines, both in vitro and in vivo, remains very inconsistent and sometimes even contradictory. Therefore, we believe that BioCell Printing holds great potential in this matter, increasing the level of structural accuracy and reproducibility in the fabrication of scaffolds, expanding the range of materials possible of being processed and reducing the process variability inherent to other fabrication methods (e.g. conventional techniques).