Tuning of ion-release capability from bio-ceramic-polymer composites for enhancing cellular activity

In our previous study, we investigated the synergetic effects of inorganic ions, such as silicate, Mg2+ and Ca2+ ions on the osteoblast-like cell behaviour. Mg2+ ions play an important role in cell adhesion. In the present study, we designed a new composite that releases a high concentration of Mg2+ ions during the early stage of the bone-forming process, and silicate and Ca2+ ions continuously throughout this process. Here, 40SiO2–40MgO–20Na2O glass (G) with high solubility and vaterite-based calcium carbonate (V) were selected as the source of silicate and Mg2+ and Ca2+ ions, respectively. These particles were mixed with poly(lactic-co-glycolic acid) (PLGA) using a kneading method at 110°C to prepare the composite (G-V/PLGA, G/V/PLGA = 4/56/40 (in weight ratio)). Most of the Mg2+ ions were released within 3 days of immersion at an important stage for cell adhesion, and silicate and Ca2+ ions were released continuously at rates of 70–80 and 180 ppm d−1, respectively, throughout the experiment (until day 7). Mouse-derived osteoblast-like MC3T3-E1 proliferated more vigorously on G-V/PLGA in comparison with V-containing PLGA without G particles; it is possible to control the ion-release behaviour by incorporating a small amount of glass particles.


Introduction
Tissue engineering combining three elements: namely, cells, scaffolds and growth factors, has been attracting attention in recent years as a method for repairing tissues damaged from injury or disease [1]. Examples of the growth factor include 2. Material and methods

Preparation of SiO 2 -MgO-Na 2 O glass
A batch mixture of nominal compositions of 40SiO 2 -40MgO-20Na 2 O in mol% was prepared using reagent-grade chemicals such as SiO 2 , MgO and Na 2 CO 3 (Kishida Chemical, Japan). The mixture was melted at 1500°C for 30 min in a platinum crucible in air and then quenched on stainless steel using an iron-pressing method, resulting in the formation of plate-likes glass pieces. The resulting glasses are denoted as 'G'. The glasses were crushed roughly using an alumina mortar and then pulverized using ball-milling for 24 h with 400 g of zirconia balls of 3 mm in diameter and 30 g of glass with 100 ml of methanol. Figure 1a shows an image of the G particles using scanning electron microscopy (SEM) (JSM-6301F, JEOL, Japan). A conductive coating treatment was carried out prior to the observation using an osmium coater (NEOC Neo Osmium Coater, Meiwafosis Co., Ltd, Japan). The glass particle sizes were approximately 1 µm or less. As shown in figure 1b, halo peaks were observed in an X-ray diffraction (XRD; X'pert, Philips, the Netherlands: CuKα, 45 kV, 40 mA) pattern of the particles; G was concluded to be glassy particles. In this pattern, two halo peaks were shown; this glass may contain two different networks.
The structure of the prepared glass sample was measured using laser Raman spectroscopy with an Nd: YAG laser (NRS-3300, 532 nm, JASCO Co., Japan). The structure of silicon in the glass was examined through magic-angle-spinning nuclear magnetic resonance spectroscopy (MAS-NMR) (JNM-ECA600II, JEOL, Japan) at a resonance frequency of 119.24 MHz, operating at pulses of 5.0 µs and at a recycle delay of 120 s using an 8.0 mm zirconia rotor. The sample spinning speed was 6 kHz. In addition, 4-dimethyl 4-silapentane sulfonate sodium (1.534 ppm) was used as a reference material of 29 Si.

Preparation of vaterite particles
Calcium carbonates consisting predominantly of vaterite were prepared using a carbonation process in methanol [37]. CO 2 gas was blown for 3 h at a flow rate of 300 ml min −1 into a suspension consisting of 7.0 g of Ca(OH) 2 in 180 ml of methanol at 0°C in a Pyrex beaker. The resulting slurry was dried at 70°C in air, resulting in the formation of fine-sized powders. Figure 2 shows an SEM photograph of the powders. The surface area of the calcium carbonates obtained was determined to be 40 m 2 g −1 through a nitrogen gas sorption analysis. Secondary particles of 0.5-1 mm in diameter were formed, and an agglomeration of primary particles of 20-100 nm in diameter was observed. XRD analysis (CuKα, 45 kV, 40 mA) showed that the calcium carbonate consists predominantly of vaterite with a trace amount of calcite (denoted as V hereafter).

Preparation of PLGA composites
To prepare a composite containing G and V, PLGA (75% lactide/25% glycolide) (PURASORB ® , Corbion Purac, The Netherlands) with an average molecular weight (M w ) of 102 kDa was used. PLGA was melted  at 135°C in a kneader, and then mechanically blended with G and V particles for 10 min. The resulting composite is denoted as G-V/PLGA. For comparison, composites consisting of V and PLGA (denoted by V/PLGA) and consisting of β-TCP and PLGA (denoted by TCP/PLGA) were also prepared using the same procedure as G-V/PLGA. The composition of each sample is shown in table 1.
Each composite was dissolved in chloroform. The composite slurry was cast in a Teflon ® vessel and dried at room temperature to form a film. Film-like samples were then used for the following evaluation tests.

Measurement of ion release from composites
The amounts of ions released from the G-V/PLGA films (14 mm in diameter and 160 ± 20 µm in thickness) were evaluated by immersing them in 1 ml of a culture medium (α-MEM) containing 10% fetal bovine serum (FBS) and incubating at 37°C in a humidified atmosphere of 95% air with 5% CO 2 for 7 days. The medium was changed after 1 day of culturing and then changed every other day.
An XRD analysis (CuKα, 45 kV, 40 mA) of G-V/PLGA was conducted to examine the change in crystal phase before and after immersion in α-MEM.
The supernatant solution after immersion was diluted 10 times with distilled water, and the concentration of each ion in the solution was measured using inductively coupled plasma atomic emission spectrometry (ICP-AES) (ICPS-7000, Shimadzu, Japan).
To examine the charge of the ion distribution in G-V/PLGA, the composite was immersed in α-MEM for 7 days, and to prepare the sample for observation with a scanning transmission electron microscope (STEM) (JEM-2100F, JEOL, Japan) equipped with an energy dispersive X-ray spectrometer, the composite was processed using a focused ion beam processing observation apparatus (FIB) (EM-9320FIB, JEOL, Japan). An element mapping image of a section of the sample was obtained.

Cell culture test
PLGA, G-V/PLGA, V/PLGA and TCP/PLGA films of 14 mm in diameter and 160 ± 20 μm in thickness were prepared for the cell culture tests. They were sterilized using ethylene oxide gas. Mouse osteoblastlike cells (MC3T3-E1 cells) were seeded onto the films in 24-well plates at a density of 30 000 cells well −1 . The α-MEM containing 10% FBS was used as the culture medium, and the cells were cultured at 37°C in 5% CO 2 for 7 days. The medium was changed after 1 day of culturing and then changed every other day.
The number of MC3T3-E1 cells was evaluated after treatment using a Cell Counting Kit-8 (Dojindo, Japan). One reagent used in the kit is a water-soluble tetrazolium salt, which is reduced by 1 µm Figure 2. SEM image of V particles.   [38][39][40]. The Q Si 3 peak was unclear. Figure 3b shows the spectrum of 29Si MAS-NMR and the deconvoluted peaks using a Gaussian function. The peaks derived from Q Si 1 , Q Si 2 and Q Si 3 , which were confirmed to appear at −73, −82 and −91 ppm, respectively, were observed [39].
The peak-integrated ratio of Q Si 1 /Q Si 2 was estimated to be approximately 73/27. As a result, the Q Si 1 unit was concluded to be a dominant structure without the Q Si 3 unit. Figure 4 shows the XRD patterns of G-V/PLGA before and after immersion in α-MEM. The peaks derived from vaterite were confirmed. A weak peak corresponding to calcite appearing in the pattern was confirmed. Figure 5 shows the ion release behaviour from G-V/PLGA in α-MEM. The Mg 2+ and Ca 2+ ions in the media were 28 and 160 ppm, respectively. Silicate ions were released continuously (approx. 70-80 ppm) from the composite during the 7-day period. Approximately 74% of the Mg 2+ ions were released within 1 day, and approximately 90% were released within 3 days. Ca 2+ ions were released continuously, similarly as the silicate ions. Figure 6 shows element mapping images of the cross-section of G-V/PLGA around its surface before and after the 7 days of immersion. In the images prior to immersion, V-derived Ca and G-derived Si were observed. Note that the amount of G was only 4%. The G particles were sparsely embedded in the PLGA matrix phase. In the images taken after 7 days of immersion, V-derived Ca was clearly observed; almost no change in the distribution state was observed before or after immersion. By contrast, from the images taken after 7 days of immersion, no G particles were observed inside the material, and Si and P appeared within the vicinity of the sample surface on the V particles.         The following three factors are considered reasons for these ions-release behaviours.

Discussion
The first factor is the solubility of the glass particles consisting of Q Si 1 (73%) and Q Si 2 (27%). The short chain structure is easily broken by H 2 O through a mechanism close to the reaction of a Bioglass ® surface [42]. To form a silanol group on the surface, the glass releases Mg 2+ and Na + ions through an exchange with protons (or H 3 O + ) in the medium, and a loss of soluble silica in the form of Si(OH) 4 will then be induced. Figure 6 indicates that, after the immersion of G-V/PLGA in the medium, the G particles are dissolved, and a thin Si-rich layer around the composite surface is formed. The layer will be a silica gel phase formed through the condensation and repolymerization of the silanols. As shown in figure 6, phosphate ions appeared within the vicinity of the sample surface after immersion in α-MEM. The silica gel layer could enhance the formation of calcium phosphate phase around the sample surface. The difference in ion-releasing behaviours between G-V/PLGA and V/PLGA in figure 8 is related to the formation of the silica layer, which inhibits the initial burst release of Ca 2+ ions from V/PLGA. The second factor is the water-uptake capability of PLGA. PLGA has been reported to show a much higher water-uptake ability than PLLA when it contains 20% of poly(glycolic acid) and when PLLA is immersed in a phosphate buffer solution [23]. Owing to this ability of PLGA, the glass particles will be dissolved by water invading inside the material. Pathways through which the ions are released are created by this entrapped water. As a result, Mg 2+ and Na + ions are quickly released. Silicate ions are released slowly, however, owing to their low solubility, and some remain around the surface of the composite. V particles are also dissolved in the uptake water, and Ca 2+ ions are continuously released.
The third factor is the high content of the filler, namely, 60 wt%, in the composite. In our previous study, the ion-release behaviours of the PLLA composites containing 30 and 60 wt% of SiV in a Tris buffer solution were examined [32]. The composite containing 30 wt% of SiV released approximately 60% of the silicate ions within 1 day of immersion, whereas the composite containing 60 wt% of SiV released one almost completely within 1 day. This burst release is considered to be from the percolation effect of the filler, namely SiV, owing to its contact in the PLLA matrix phase. Such an effect occurs because G-V/PLGA contains 60 wt% (approx. 50 vol%) of filler (G and V particles).
The combination of these three factors indicates that water can easily infiltrate the G-V/PLGA. The high solubility of glass implies that the infiltrated water will easily dissolve the Mg 2+ and silicate ions, even if present in a small amount, namely, only 4 wt%. The large amount of V particles, namely, 50 wt%, indicates a continuous supply of Ca 2+ ions. The proliferation of MC3T3-E1 on G-V/PLGA showed significantly higher values during each culture period. As shown in figure 8, the release of Ca 2+ ions from V/PLGA, and the release of silicate, Mg 2+ and Ca 2+ ions from G-V/PLGA, was observed.
The combination of Mg 2+ , Ca 2+ and silicate ions released from G-V/PLGA might enhance the proliferation of the cells. It is difficult at this stage to conclude the enhancing effect and its mechanism. However, we believe that the present study demonstrates that the inclusion of a small amount of ion-releasing glass particles might have a positive effect on the stimulation of osteoblast-like cells.

Conclusion
In this study, we aimed at the preparation of biodegradable polymer composites for tuning the release of silicate, Mg 2+ and Ca 2+ ions using highly soluble glass and calcium carbonate particles. A 40SiO 2 -40MgO-20Na 2 O glass with high solubility was newly prepared as the releasing source of Mg 2+ and silicate ions. The glass consisted predominantly of a Q Si 1 group with a portion of a Q Si 2 group; this structure was shown to ease the dissolution in an aqueous solution.
A total of 4 wt% of glass particles and 56% calcium carbonate (vaterite) particles containing a trace amount of calcite were embedded in a PLGA matrix phase. The resulting composite showed a unique ion-releasing behaviour in a culture medium, demonstrating a burst release of Mg 2+ ions immediately after immersion in the medium and the continuous release of Ca 2+ and silicate ions.
Such an ion dissolution behaviour is thought to be caused by the following three factors: (i) the high solubility of the glass particles, (ii) the high water-uptake capability of PLGA and (iii) the percolation effect of the filler (60 wt%) embedded in the polymer matrix.
The continuous release of silicate ions was considered to originate from the balance between the amounts of soluble silica and the gelating one derived from the hydrolysis of the glass particles. The continuous dissolution of Ca 2+ ions originated from the high solubility of vaterite and a percolation effect with a large amount of filler, namely, 60 wt%, in addition to the hydrous capability of PLGA.
A tuning of the ion-release behaviour when using a small amount of glass particles demonstrated the possibility of enhancing the proliferation of osteoblast-like cells.