**3. Results and Discussion**

The BET analysis evidenced a SSA of 11.6 m2/g, the particle size analysis showed a non-symmetric distribution with a mode grain size of 2 μm; characteristic particle sizes are *d*<sup>50</sup> = 2.0 μm and *d*<sup>90</sup> = 4.4 μm, with 94% of grain sizes below 5 μm and the residual 4% below 9 μm.

Figure 1 shows the morphology (a, b) and the compositional analysis (c) of SBA2 glass powders. The powders display the typical sharp-cornered morphology of ball-milled glass. Moreover, as previously mentioned, the majority of SBA2 powders showed a grain size <5 μm. In addition, the performed DTA analysis evidenced a glass transition temperature of about 550 ◦C, the crystallization onset at 600 ◦C, a crystallization peak at 655 ◦C and a melting temperature of 1220 ◦C.

**Figure 1.** Scanning electron microscopy (SEM) (**a**,**b**) and energy dispersive x-ray spectrometry (EDS) (**c**) analysis of SBA2 powders after milling process. Scale bars: 40 μm (**a**), 10 μm (**b**).

The electrospinning process of PLLA-SBA2 is carried out by adjusting the voltage bias as well as the solution flow rate to obtain almost bead-free fibres. The morphology of neat PLLA and composite fibres is shown in SEM micrographs in Figure 2. Composite fibres exhibit a quite homogeneous distribution of embedded SBA2 particles (arrows in Figure 2b).

The SBA2 incorporation also leads to an increase of the overall roughness in the electrospun mats. Interestingly, the average diameter of PLLA-SBA2 filaments is around one half of fibres without SBA2 (Table 1), similarly to previous findings on composite fibres made by PCL and commercially available Bioglass®(45S5) particles [42]. This effect might be explained due to the change of intrinsic solution properties, including rheology and conductivity, of the polymeric solutions after the addition of SBA2 particles [44]. In addition, the physical effect of particles in the electrified solution is associated to the onset of enhanced whipping and varicose instabilities, which affect the morphology of the ultimately

deposited fibres [45], as shown by the local increase of filament radius in the composite fibres close to the particles. The resulting excess mass due to local polymer accumulation leads to the here found decrease of diameter along the rest of the fibre length due to the overall polymer mass conservation. Furthermore, we find that both the types of fibres display cylindrical shape and high surface porosity, with pore size around 100 nm (Figure 2c,d), which is characteristic of PLLA electrospun with highly volatile solvents [28] and expected to favour cell attachment. Indeed, it is known that both microand nanoporosity play an important role in protein adhesion and cell function [46]. In addition, the incorporation of the bioactive glass particles in PLLA-SBA2 fibres is confirmed by SEM-EDS analysis as shown in Figure 3 and Figure S1.

**Figure 2.** SEM micrographs of (**a**) neat polylactic acid (PLLA) fibres and (**b**) PLLA-SBA2 fibres (scale bar = 25 μm). Neat PLLA fibres (**c**) and composite fibres (**d**) at high magnification, evidencing their highly porous surface. Scale bar: 2 μm.

**Table 1.** Average diameters, minimum and maximum values measured for the transversal size of neat polymeric and composite fibres.


**Figure 3.** EDS spectrum of electrospun PLLA-SBA2 fibres (**a**) and its comparative SEM-EDS layered image (**b**) showing peaks of SBA2 constituents such as Ca (yellow), Na (orange) and Si (cyan). O (bright green) and C (red) are mainly related to PLLA.

The mechanical properties of the polymeric and composite fibres mats, determined from their stress–strain curves, are summarized in Figure 4, in terms of Young's modulus (Figure 4a), elongation at break (Figure 4b) and tensile strength (Figure 4c) for either randomly oriented and aligned electrospun samples. Particularly, in uniaxially aligned arrays, in which the fibres and to some extent polymeric chains [47–51] are oriented along the traction axis, mechanical differences between composite and neat PLLA are more pronounced [49]. The elastic modulus of PLLA-SBA2 fibres is found to be (34 ± 5) MPa, significantly higher than the value measured for pristine PLLA fibres, (20 ± 2) MPa, just for the aligned fibres. Correspondingly, PLLA-SBA2 exhibits a relatively lower elongation at break (15 ± 2%) compared to the value of PLLA samples in the same alignment conditions (20 ± 2%), as well as higher tensile strength (11 ± 1 MPa), compared to (7 ± 1) MPa. The mechanical behaviour of the composites tightly depends on the interaction at the polymer-bioactive glass interface and on the homogeneity of the dispersion of glass particles in the fibres, which if poor would lead to agglomeration in clusters [51]. The present findings in terms of increased elasticity and concomitant reduction of the elongation at break and tensile strength are indicative of a reinforced fibre system where inorganic fillers lead to embrittlement of the fibres [39,52].

**Figure 4.** Young's modulus (**a**), elongation at break (**b**) and tensile strength (**c**) of random and aligned fibres. Results are expressed as (mean ± standard deviation). Bars show statistically significant differences (*p* < 0.05). In the inset of (**a**) and (**c**) a zoom view of the properties of randomly oriented fibres is reported.

The analysis of both PLLA-SBA2 and PLLA fibres, performed by collecting SEM micrographs of samples soaked in SBF solution, evidences polymer degradation starting from the 7th day following incubation, resulting in cracks and points of break in the surface of fibres. As highlighted in SEM micrographs in the top of Figure 5, in PLLA-SBA2 fibres this degradation leads to a progressively increasing exposure of glass particles to the surrounding microenvironment. In order to investigate the resulting bioactive glass dissolution from PLLA cracks propagating along the nanopores on the fibre surface, we perform further in vitro degradation studies, incubating the composite scaffolds in DPBS in physiological conditions (at pH = 7.4 and 37 ◦C), as showed in Figure S2c. During the immersion in DPBS for 21 days, the corresponding changes of the solution pH was monitored, as displayed in Figure 5 together with corresponding SEM micrographs at each time point. PLLA fibres clearly undergo acid hydrolysis in the solution which leads to a decrease of pH, whereas for the composite fibres this trend is overcome by the release of SBA2 dissolution products into the solution. The pH increase related to bioactive glass dissolution products have been already investigated in several medium by Cerutti et al. [52]. This increased pH associated with the dissolution-precipitation of bioactive glass is known to affect various cellular processes, being correlated with increased metabolic activity and proliferation rate in mammalian cells [53]. Our data on pH variations are also supported by the corresponding measurements of sample weight loss, performed after 21 days, which is found to be 2.2% of the initial weight for neat PLLA fibres and 9.7% for PLLA-SBA2 fibres.

**Figure 5.** In vitro degradation studies and dissolution of PLLA and PLLA-SBA2 electrospun fibres. Change in the pH of DPBS solution for PLLA and PLLA-SBA2 fibres, which is associated to the acid hydrolysis of the polymer components and to the release of SBA2 into the solutions at 0, 1, 7, 14, 21 days from incubation and the relatives SEM micrographs at high magnification of composite and polymeric fibres along soaking experiments in SBF at the same time point (scale bar = 2 μm).

The acellular bioactivity of the electrospun fibres is evaluated by exploring the formation of HCA on their surfaces upon soaking in SBF, at 37 ◦C for different time periods. The bioactivity of PLLA-SBA2 fibres can be clearly appreciated when a large quantity of HCA with typical cauliflower-like morphology is found nearby the electrospun filaments (rightmost SEM micrographs in the top of Figure 5). Neat fibres, instead, do not lead to HCA formation. These findings are supported by EDS, detecting high Ca and P peaks belonging to HCA (Figure 6) with Ca/P ratio of 1.9, as well as by FTIR analysis (Figure 7). Indeed, FTIR spectra of PLLA-SBA2 fibres following incubation feature two new peaks at 600 cm−<sup>1</sup> and 560 cm−<sup>1</sup> corresponding to HCA-associated P–O groups asymmetric bending [54] and the increase in intensity of the peak centred at 960 cm−<sup>1</sup> ascribable to the contribution of Si-OH symmetric stretching characteristic of the bioactive glass after immersion in SBF [54,55].

The formation of HCA following immersion in SBF solution confirms that the incorporation of the bioactive glass particles in the PLLA filaments is not preventing the characteristic bioactivity of this specific composition of bioactive glass to be highlighted [32]. In addition, the formation of HCA after 21 days suggests this class of scaffolds not only as useful systems for bone tissue engineering but also for soft tissue repair applications, because release of ions is detected since the earliest time point, as evidenced by the pH variation in the degradation studies. In particular, these fibrous architectures would be highly suitable for the release of therapeutic ions embedded in the bioactive glass structure, occurring without simultaneous damage in the fibrous structure of the scaffolds. Indeed, the double-scale temporal dynamics exhibited by the two mechanisms might be highly advantageous for wound healing processes, where the composite can still provide mechanical support while serving as reservoir of therapeutic ions [56]. The ability to modulate ion release through bioactive glass composition engineering, in relation to the addressed cell type, might be especially important in this respect. This relevance is also suggested by previous works that emphasized the influence of ion release in stimulating vascularization and triggering the production of angiogenic growth factors during soft tissue repair [57,58]. Other application fields potentially benefiting from interplaying ionic release and gradual morphological changes in composite scaffolds include the development of biomedical materials able to modulate the inflammatory response [59], to control cell

proliferation [11], to support the regeneration of the peripheral nerve and the treatment of relevant pathological conditions such as chronic osteomyelitis [60–62]. The smart combination of biologically active ion release and nanotopography has been highlighted for tissue engineering [63].

**Figure 6.** EDS analysis of PLLA-SBA2 samples after 7 days (**a**,**b**) and 21 days (**c**,**d**) of soaking in SBF. EDS spectra of PLLA-SBA2 (**a**,**c**) and comparative EDS layered images (**b**,**d**) reveal an increment of Ca (yellow) and P (purple) content. SBA2 constituents: Na (orange), Al (green) and Si (cyan). O (bright green) and C (red) mainly relate to PLLA.

**Figure 7.** Fourier transform infrared (FTIR) spectra of neat PLLA (PLLA\_d0) and composite scaffolds (PLLA-SBA2\_d0) before incubation in simulated body fluid (SBF) and of composite fibres after 14 (PLLA-SBA2\_d14) and 21 days (PLLA-SBA2\_d21) of incubation in SBF. The inset better highlights the new peaks ascribable to hydroxycarbonate apatite (HCA) formation mentioned in the text.

Finally, we investigated the biocompatibility of PLLA-SBA2 fibres using a murine-derived ST2 stromal cell line. The morphology of cells, inspected by SEM, was found to be affected by the fibrous structure of the scaffold, with significant cells elongation along the directions of supporting fibres as shown in Figure 8. The cytoskeleton actin and the nuclei of cultured cells were also analysed by immunofluorescence microscopy (Figure 9). Overall, cells adhesion was efficient on the nanocomposite filaments embedding bioactive glass particles. In previous works, in vitro cytotoxicity tests of SBA2 in PMMA-based cement were accomplished using the indirect contact method [33]. Here, the cell viability is directly evaluated at 1 and 7 days after seeding. WST-8 assays highlight, for absorbance detected at 450 nm at day 1 for cultures on PLLA fibres, an almost double value compared to those measured for cultures on PLLA-SBA2, results shown in Figure 10. This result correlates well with the increase of the pH value found during degradation studies, namely with the release of ions from the bioactive glass particles. Indeed, it is known that the alkalinisation of the medium can influence significantly cell metabolism [64] and can consequently induce distinct phases in cell cycling on composite and neat fibres, respectively, as also suggested by previous reports focused on the behaviour of human osteoblasts exposed to the ionic dissolution products of bioactive glass [65,66].

At day 7 after seeding the values of absorbance measured on PLLA fibres and on PLLA-SBA2 fibres become comparable, as shown in Figure 10, evidencing that the initial conditions of pH changes do not affect cell viability at longer timescales. ST2 cells cultured on composite fibres generally exhibit a relatively lower cell density following seeding, then reaching cell densities comparable to those on PLLA fibres after 7 days. Indeed, despite a significant difference in cell density immediately after cell seeding, these results evidence the ability of PLLA-SBA2 fibres to providing an effective and viable environment for subsequent scaffold colonization.

**Figure 8.** SEM micrographs of dehydrated ST2 cells, cultured on neat (**a**,**b**) and composite (**c**,**d**) fibres for 1 (**a**,**c**) and 7 (**b**,**d**) days. Scale bar: 20 μm.

**Figure 9.** Fluorescent labelled actin filaments (red) and nuclei (blue) in ST2 cells cultured on neat PLLA (**a**,**b**) and PLLA-SBA2 (**c**,**d**) fibres, after 1 day (**a**,**c**) and 7 days from cell seeding. Scale bar: 50 μm.

**Figure 10.** WST-8 assay: histograms of absorbance at 450 nm for PLLA and PLLA-SBA2, one (d1) and seven (d7) days after the seeding.
