Effect of in vitro enzymatic degradation on 3D printed poly(ε-caprolactone) scaffolds: morphological, chemical and mechanical properties

Materials

Linear thermoplastic PCL with a molecular weight (Mw) of 50,000 Da (Capa® 6500; Perstorp Caprolactones, Cheshire, UK) in the form of pellets was employed as received, for the production of 3D scaffolds. Lipase enzyme (Amano Lipase PS, from Burkholderia cepacia [was Pseudomonas cepacia]) was obtained from Sigma Aldrich.

Scaffold design and fabrication

3D structures based on PCL were fabricated using an extrusion-based additive manufacturing (AM) device called the BioExtruder, which was equipped with a nozzle with a 300-μm inner diameter (26). Block samples were initially designed (30 × 30 × 8 mm³) using computer-aided design (CAD) software (Solid Works, Dassault Systemes, SA). Three different lay-down patterns were adopted with a filament distance (FD) fixed at 1,000 µm and a slice thickness (ST) at 280 µm: square (0/90°), triangular (0/60/120°) and complex polygonal (0/45/90/135°) internal pore geometries were obtained, as shown elsewhere (29).

PCL in the form of pellets was directly melted in the material chamber of the BioExtruder and then transferred to the screw-extrusion chamber via compressed air. Based on previous results, the structures were produced considering the optimized processing conditions indicated in Table II (28). After fabrication, the scaffolds destined to further analysis were cut into smaller pieces, taking care to avoid or minimize damage to their architecture.

Degradation assay

Seven specimens of each scaffold topology, after being sterilized with UV radiation for 30 minutes, were cut into smaller samples, measuring 4 × 4 × 8 mm, accurately weighed (w_0), immersed in PBS solution containing 0.5 mg/mL of lipase (Amano Lipase PS, from Burkholderia cepacia [was Pseudomonas Cepacia]) and incubated (37°C, pH 7.4) for 25 days. The enzyme concentration was established based on previous studies reported in the literature (30). At predetermined intervals of 5 days, scaffolds were withdrawn from the degradation medium. The removal of salt excess was carried out by immersion of the scaffolds in ultrapure water (with a Millipore purification system) for 24 hours. Finally, the scaffolds were carefully wiped, stored in a vacuum desiccator for 12 hours at room temperature and quickly weighed (W_T). The weight loss percentage (WL%) was calculated as follows:

Scaffold characterization

Statistical analysis

The data from the measurements were reported as means and standard deviations and analyzed using 1-way analysis of variance (ANOVA) and Tukey post hoc test (using SigmaPlot 13 software). The significance level was set at a p value of <0.05.

Morphological analysis of nondegraded scaffolds

SEM micrographs were used to assess the scaffolds’ architectural features – namely, road width (RW), filament gap (FG) and layer gap (LG) (Fig. 1; Tab. III).

µ-CT results depicted in Figure 2 show that the PCL scaffolds produced presented a repeatable microstructure with precise pore size and shape and interconnectivity. The results obtained supported the SEM observations, and the experimental values sufficiently substantiated the theoretical ones, defined during the scaffold fabrication by the process and instrument parameters. As shown in Figure 2, all scaffolds were characterized by a mean fiber diameter of 360-380 μm, and a center-to-center fiber distance of 610-650 μm between 2 fibers in a common layer was evaluated, independently of the pore geometry. Scaffold characteristics obtained by µ-CT are summarized in Table IV.

Degradation studies

Degraded PCL scaffolds were investigated in terms of WL%, SEM and µ-CT, focusing attention on the possible influence of architectural parameters on the in vitro enzyme-mediated degradation process.

The influence of these parameters has already been analyzed in earlier studies on 3D PCL scaffolds subjected to in vitro hydrolytic degradation in PBS or SBF (24, 25). The results evidenced that PCL scaffolds, incubated in both the media for 6 months, showed a WL% in the range of 0.2%-1%, independently from the geometry of the pores. Similar results (i.e., WL% in the range of 0.2%-1.2%) were achieved by Idaszek et al (31), who investigated the degradation kinetics in SBF for 3 months of composite PCL-based scaffolds obtained by a solvent casting technique.

In the present study, results for the enzymatic degradation of PCL 3D scaffolds revealed that, regardless of the scaffold pore geometry, the weight loss determined 5 days after the beginning of the degradation experiment maintained a linear and uniform trend throughout the entire period, reaching a maximum WL% of 72.2% at day 25 (Fig. 3).

Brugmans et al (32) studied the enzymatic degradation of PCL scaffolds obtained by a conventional electrospinning method. They found a WL% equal to 44% by lipase treatment up to 56 hours. Murray et al (33) investigated the enzymatic degradation of graphene/PCL composites in lipase solution at 37°C for 4 days. The results revealed that both for thin films made of PCL alone and for the composites, a 50%-60% weight loss over the first 24 hours was achieved.

When comparing our data with those from preceding studies, a slower degradation kinetic was found. This observation can be explained considering that PCL systems were subjected to a superficial erosion mechanism with an initial loss of the amorphous zones located on the surface of the scaffold. In particular, Bölgen et al (34) demonstrated that the surface area to volume ratio (SA/Vol) of a material affects the rate of degradation: samples with higher SA/Vol and porosity are expected to degrade faster, due to the greater water penetration into their structures. In contrast, Athanasiou et al (35) showed that scaffolds based on poly(lactide-co-glycolide) with a higher porosity degraded at a slower rate than those with 0% porosity. A possible explanation was that a greater surface area allowed easier removal of acidic breakdown products, thus reducing autocatalytic degradation phenomena. Bosworth and Downes (36) carried out a degradation study investigating the hydrolysis of PCL scaffolds obtained by different techniques (i.e., 3D bundles, 2D fibrous mats and solvent cast films, with SA/Vol of 11.48, 13.23 and 10.26, respectively) in PBS at 37°C, over a 3-month period. Their work confirmed the hypothesis of Bölgen et al (34).

In our case, as shown in Figure 3, no particular trend was observed for WL% as a function of pore geometry. Based on the hypothesis of Bölgen et al, a degradation kinetic 0/90°>0/60/120°>0/45/90/135° should be expected. A possible explanation for these results may rely on the geometry of the pores. As the number of deposition angles increases from 0/90° to 0/45/90/135° the diffusion of acidic breakdown products is reduced, thereby promoting a higher autocatalytic degradation process. This phenomena appears to counterbalance the effect of SA/Vol ratio (which is smaller in the 0/45/90/135° scaffolds) thus promoting a similar weight loss trend among all scaffold geometries.

SEM observations and µ-CT analysis revealed that the physical features strongly changed over time as an effect of the degradation process. Despite the diameter reduction of the fibers, adhesion with the neighboring fibers was still maintained. On the other hand, as the fiber diameter decreased, FD consequently increased. This means that the increasing of FD led to an increased pore size and porosity (Fig. 4). Furthermore, it was also possible to note the appearance of numerous cavities in the filaments (T10, T15 and T20), as a result of the surface erosion mechanism.

Results obtained from µ-CT analysis were consistent with SEM observations, showing an increase of pore size and porosity as a consequence of the filament thinning (Fig. 5).

DSC analysis

DSC results are reported in Table V and suggested that, during the degradation period, there were no significant differences among the architectures investigated, in terms of melting temperature, melting enthalpy and crystallinity degree.

The enzymatic degradation process did not have a significant impact in the melting temperature of PCL, which remained in the range 61°C-64°C over 25 days. The same observations were reported by Murray et al (33). Furthermore, for all of the analyzed architectures, the crystallinity degree generally increased over time, with statistically relevant differences between day 0 and the successive days, except for the 0/60/120° architecture, where no significant differences were found between T0 and T20 (see Fig. 6). However, during the period analyzed, all of the scaffolds showed an increase of the crystallinity degree, which should probably be ascribed to the random hydrolytic bond cleavage of the amorphous regions present in PCL, since it is well recognized that the degradation medium is facilitated more to penetrate into a disordered network than into an ordered one (37). After 5 days, the crystallinity degree seemed to become constant over time (Tab. V; Fig. 6). This observation suggested that the degradation of amorphous and crystalline regions might occur at the same time but at different rates, with crystalline regions presenting a lower degradation profile when compared with amorphous regions. The results also suggested that the overall degradation process, being related to both crystalline and amorphous PCL regions, might lead to alternate increases and decreases in crystallinity percentage throughout the incubation period investigated.

SEC analysis

Table VI and Figure 7 report the changes in Mn and PDI of PCL scaffolds with different pore geometries, which were degraded for 25 days in enzymatic solution. Before degradation, all scaffolds presented a similar Mn of approximately 41,000 Da, with a PDI between 1.41 and 1.45.

Bosworth and Downes (36) found a decrease in Mn for PCL 3D bundles (6.2%), followed by 2D fibrous mats (5.6%) and then solvent cast films (5.0%). Løvdal et al (38) investigated the mechanical properties of an electrospun PCL exposed to in vitro physiological fluids or to accelerated in vitro degradation conditions (i.e., at pH 12 for 29 days), at 37°C. Mw evaluations after 29 days showed a Mw reduction of 48% compared with that of pristine PCL, with an increased PDI (from 1.9 at day 0 to 4.5 at day 29). On the other hand, enzymatic degradation of PCL scaffolds obtained by a conventional electrospinning method by Brugmans et al (32) resulted in a significant mass loss by lipase treatment up to 56 hours, while Mw remained constant over time.

Our results showed that PCL scaffolds produced with different lay-down patterns and comparable porosities presented an almost similar degradation rate, resulting in a very slight, almost negligible Mn decrease from day 0 to day 25. Moreover, the differences observed were within the range of error associated with the SEC measurements. These observations suggested that PCL scaffolds were degraded mainly by surface erosion mechanisms with significant weight loss but almost invariable molecular weight values throughout the degradation period.

Compressive mechanical tests

Figure 8 reports typical stress-strain curves relevant to the 3D bioextruded scaffolds before the degradation, evidencing the effect of lay-down pattern on the mechanical features of the structures.

The results confirmed that, at a selected FD, the lay-down pattern had a significant effect on the mechanical behavior of the 3D scaffolds. Table VII shows that scaffolds with 0/90° lay-down pattern exhibited a greater compressive modulus (52.1 ± 2.8 MPa at T0) than those obtained for 0/60/120° (44.0 ± 6.2 MPa at T0) and 0/45/90/135° (16.2 ± 2.1 MPa at T0) patterns. When the amplitude of the deposition angle between the filaments of adjacent layers decreased – i.e., passing from a 0/90° to a 0/45/90/135° lay-down – a larger contact area occurred. This in turn should lead to a reduction of the local stress endured by the 3D scaffold, with consequent reduction of the overall mechanical performance.

Figure 9 and Table VII summarize the influence of the in vitro enzymatic degradation on the mechanical performances of PCL scaffolds with the proposed lay-down patterns over time. The data present an overall deterioration of mechanical properties as a function of degradation time, as was also observed by other researchers (38).

In the results obtained, an evident reduction in the overall mechanical properties of the samples, independently of the pore geometry, was observed. This may be related to the significant mass loss caused by the enzymatic degradation mechanism. As observed in the morphological analysis, the surface erosion process led to the thinning of the filaments thus increasing the FD and porosity of the structures. In general, for all of the materials, the variation in mechanical properties as a function of FD and porosity was ascribable to the fact that the filament junctions, when subjected to compression, acted as columns. A scaffold having a higher FD, for a defined area presented a smaller number of columns. This in turn led to a decrease in the scaffold compressive modulus and maximum stress (39).