The effects of boron trioxide on the structure and degradation behaviour of borosilicate bioactive glass

ABSTRACT

Degradation behaviour of bioactive glass (bioglass) is one of the most important factors affecting bone repair, because an excellent and controllable degradation rate can match the rate of new bone formation. In this research, the derived borosilicate bioglasses based on the 6Na₂O-8K₂O-8MgO-22CaO-18B₂O₃-54SiO₂-2P₂O₅ component were synthesized. The effects of B₂O₃ on the structure, degradation behaviour and cytocompatibility of borosilicate bioglasses were systematically studied. The results showed that with B₂O₃ addition, the network-forming units became diversified and part of BO₄ units transformed into BO₃ units. These factors weakened the chemical durability of borosilicate bioglass, thus accelerating the bioglass shedding and altering the ions release, especially B. This study provides a theoretical basis for designing borosilicate bioglass with adjustable degradation rate and ion release behaviour to meet the diverse needs of clinical bone repair.

KEYWORDS:

• Borosilicate bioglass
• structure
• degradation
• ion release

Introduction

Osteoporosis is a common disease for the elder, which is characterized by impaired bone microstructure and decreased bone mineral composition, resulting in bone fragility and fracture and even death in the elder [1]. Bone cement is a promising material used to fill different shapes of bone defects with minimally invasive surgery that promotes bone repair with less pain [2]. Bioactive glass (bioglass), as the main component of bone cement, can degrade and form osteoid hydroxyapatite in human body [3]. Since the initial report on silicate bioglass in 1971 by Hench et al. [4], a large number of bioglasses for bone repair emerged. Owing to outstanding and controllable degradation properties, the bioglasses containing B₂O₃ is particularly prominent [5,6].

Recently, there is an enormous amount of literature discussing the impact of structure on the dissolution kinetics and mechanism of borosilicate glasses [6–8]. These studies established a solid foundation for the subsequent study of borosilicate glasses with more complex components. The 1393 bioglass (1393, 6Na₂O-8K₂O-8MgO-22CaO-18B₂O₃-54SiO₂-2P₂O₅) [9] is known to be biodegradable and also exhibits excellent hydroxyapatite formation property. Hence, 1393 is being widely considered in the field of bone repair. While the B₂O₃ has been added to tailor the degradation behaviour of 1393 [10,11], the relationships between B₂O₃ doping and glass structure and its degradation behaviour have not been studied and reported systematically yet.

In this work, the structure, degradation behaviour and cytocompatibility of 1393 derived borosilicate bioglasses were studied and the relationships between them were discussed. The results from this study, when extended to more complex borosilicate-based glass compositions, will not only add to our understanding the relationship of boron incorporation-structure network-degradation behaviour, but will also form a theoretical basis for designing bioglass with controllable degradation and ion release, so as to meet the various requirements in bone repair.

Experimental procedure

Bioglass preparation

Borosilicate bioglasses based on 6Na₂O-8K₂O-8MgO-22CaO-18xB₂O₃-(54-18x)SiO₂-2P₂O₅(in mol%, x = 0, 1, 2, 3, designated as 0B, 1B, 2B and 3B, respectively) [9] were prepared by melting the required quantities of Na₂CO₃, K₂CO₃, MgCO₃, CaCO₃, H₃BO₃, SiO₂ and NaH₂PO₄·2H₂O (Sinopharm Chemical Reagent Co., Ltd., Shanghai China) in an electric furnace for 1–2 h at 1150–1300°C. The melts were quenched on a copper plate maintained at room temperature.

Bioglass structure analysis

Fourier transform infrared spectroscopy (FTIR, Nicolet 380) was used. One milligram of the bioglass powder was mixed with 100 mg KBr and pressed to form pellet and analysed in transmittance mode in the wavenumber range 400–2000 cm⁻¹.

Raman study was carried out by using LabRAM HR Evolutions equipped with a CCD detector. A multistage laser power attenuator was used to control the laser power. The spectra were recorded in the 300–1600 cm⁻¹ range of Raman shifts at less than 0.65 cm⁻¹ spectral resolutions.

¹¹B magic-angle spinning-nuclear magnetic resonance (MAS-NMR) spectrum has been recorded on a JEOL JNM-ECZ600R with a 3.2 mm MAS-NMR probe. Test conditions: ZrO₂ rotor, 15.0 kHz spinning frequency, 14.1 T magnetic field, 1.0 μs pulse length, 15 s recycle delay, BPO₄(−3.5 ppm) reference standard.

In vitro degradation of bioglass

Simulated body fluid (SBF) was configured according to the method described by Kokubo et al. [12]. Immersion conditions: 0.5 g bioglass particles (74–154 μm) was immersed in 20 mL SBF. Thereafter, the experimental groups were cultured in a chamber (HWS-80B) at 37°C with 100% humidity. After immersion for different days, the pH value of solution was measured using a pH meter (SX620). The bioglass particles were removed from the solution, washed and dried for weight loss (WL) measurement. The ion release concentration was measured using inductively coupled plasma atomic emission spectroscopy (ICP-OES, Optima 2100DV, USA).

In vitro cytocompatibility of bioglass

Bone marrow stromal cells (BMSCs) were approved by Tongji University School of Medicine. Cells culture was refer to the literature [13]. Extracts of bioglass were prepared by immersing 0.1 g bioglass powders in 10 mL dulbecco's modified eagle medium (DMEM, Gibco, China) at 37°C in a humidified atmosphere of 5% CO₂ for 24 h.

Cell Counting Kit-8 (CCK-8) assay [13] was used to evaluate the cytotoxicity of the ionic dissolution products from bioglass. A total of 100 μL of cell suspensions was seeded in 96-well culture plates at a density of 1 × 10⁵ cells/well. After 1-day incubation, the cultured medium was removed, and then the fresh culture media (control) and culture media with bioglass extracts were added, and sample size was n = 5 for each statistical analysis. After incubation for 1, 3 and 5 days, the optical density (OD) was measured at a wavelength of 450 nm using a spectrophotometric microplate reader (Bio-Rad 680, USA). GraphPad Prism 8.4.3 software GraphPad Software, USA was used for statistical analysis and significant differences was verified through a one-way ANOVA with p < 0.05.

4ʹ,6-Diamidino-2-phenylindole (DAPI); Phalloidin-Tetramethylrhodamine B isothiocyanate (Phalloidin-TRITC, Beyotime, China) phalloidin staining fluorescence labelling method was used to directly observe the BMSCs morphology. The cells cultured for 3 days were washed three times with PBS and fixed in 4% paraformaldehyde (ThermoFisher) for 10 min and permeabilized with 0.1% Triton X-100 solution in PBS for 5 min. TRITC-Phalloidin and DAPI diluted in 1% BSA were adopted to stain the cytoskeleton (red) and nuclei (blue). The staining results were captured using a confocal fluorescence microscope (OLYMPUS BX53).

Results and discussion

Structure of bioglass

The FTIR spectra of 0B–3B bioglasses are shown in Figure 1. Four bands around 478, 765, 950 and 1035 cm⁻¹ in the 0B spectrum are all attributed to Si–O vibration in SiO₄ units [14,15]. In 1B, 2B and 3B spectra, the humps at 570 and 720 cm⁻¹ are both associated to bonding vibrations of B–O–B in BO₃ units [16,17]. The intensity of these two bands strengthened due to the higher B₂O₃ doping content, indicating the increase in BO₃ units. Nevertheless, the 765 cm⁻¹ band could not be found again, which may be hidden because of the stronger 720 cm⁻¹ band. The bands in region 850–1200 cm⁻¹ are generally attributed to B–O stretching vibration of BO₄ units in pentaborate, triborate and diborate groups [16]. For 1B and 2B bioglasses, the overlapping bands of silicate and borate units, as well as the formation of Si–O–B linkages are also included in this region [18]. The 1235 and 1410 cm⁻¹ bands are both related to B–O stretching vibration of BO₃ units, which are incorporated in metaborate, pyroborate and orthoborate groups [19,20]. The strength of these two bands also strengthened with B₂O₃ doping, which further proves that the number of BO₃ units in the glass structure increased.

Figure 1. FTIR spectra of 0B, 1B, 2B and 3B bioglasses.

Figure 2 shows the Raman spectra of 0B–3B bioglasses. The 0B spectrum exhibits four bands at 615, 880, 950 and 1070 cm⁻¹, which are all assigned to breathing modes of Si–O bonds in SiO₄ units [21,22]. In 1B, 2B and 3B spectra, with B₂O₃ doping, the silicate band at 615 cm⁻¹ shifted to 630 cm⁻¹ belonging to breathing mode of Si–O–BO₄ [23]. The strength of silicate bands at 630, 880 and 1070 cm⁻¹ decreased and finally ‘disappeared’ as SiO₂ was reduced. The ‘disappeared’ of 630 cm⁻¹ band may also attributed to the enhancement of boron peaks around 580 and 770 cm⁻¹, which correspond to the breathing vibrations of ‘loose’ BO₂Ø units (Ø is nonbridging oxygen atom) [16] and BO₄ units in six-membered borate rings [24], respectively. The peak at 950 cm⁻¹ is assigned to B–O stretching mode of BO₄ units in pentaborate groups. Particularly, for 1B and 2B bioglasses, the silicate units are also included in this region [25]. The 490 cm⁻¹ peak corresponds to B–O–B stretching vibration of BO₄ units [26]. The hump around 1445 cm⁻¹ represents stretching vibrations of BO₂Ø units linked to BO₄ units and BO₃ units [25]. The thin peak at 1510 cm⁻¹ is assigned to stretching modes of BO₃ units in boroxol rings [27]. Notably, the bands of BO₄ units weakened, while the bands of BO₃ units strengthened. It indicates the increase of BO₃ units with B₂O₃ doping and part of BO₃ units are derived from the transformation of BO₄ units. The bands around 850 and 1300 cm⁻¹ manifests the presence of pyroborate groups due to B–O breathing [24,28]. The BO₃ units were most likely to pair up to form pyroborate groups, as reported [24]. This result further confirmed the increase of BO₃ unit with B₂O₃ introduction.

Figure 2. Raman spectra of 0B, 1B, 2B and 3B bioglasses.

Figure 3 displays the ¹¹B MAS-NMR spectra of 1B–3B bioglasses. The three spectra display two salient features: a narrower peak at 0 ppm corresponds to BO₄ units and a broad feature around 10–20 ppm is attributed to BO₃ units [6]. Clearly, B₂O₃ doping significantly enhanced the bands of BO₃ units, revealing the increasing fraction of BO₃ units in the glass structure.

Figure 3. ¹¹B MAS-NMR spectra of 1B, 2B and 3B bioglasses.

Combined with the results of FTIR, Raman and ¹¹B MAS-NMR, with the addition of B₂O₃, the borate bands emerged and became stronger, while the silicate bands decreased and disappeared. B₂O₃doping leads to the diversification of structural groups, thereby reducing the connectivity of glass structure [5]. Besides, some BO₄ units were converted to BO₃ units, increasing the fraction BO₃ units in the glass structure. Here is possible scenario: as B₂O₃ doping increases, numerous boron atoms exist in the glass structure, which need extra oxygen atoms to form units. Thereby, the oxygen ions in BO₄ units before are now serving to new boron atoms to form BO₃ units. As we known, the strength of BO₃ trihedral (2D) is weaker than that of BO₄ and SiO₄ tetrahedra (3D), which further reduces the stability of glass structure [8].

Degradation behaviour of bioglass

Figure 4(a) shows the pH values of the solutions soaked with 0B, 1B, 2B and 3B bioglasses, respectively. The four pH curves present similar trend: firstly increase rapidly and then slow considerably before reaching limiting values. The initial rapid increase of pH value can be attributed to preferential release of Na and K ions, resulting from an ion-exchange reaction mechanism, where Na and K ions from the bioglass consume H ions from the solution for charge compensation [29,30]. Then, the B and Si ions being released from bioglass form acidic species in solution that serve to neutralize the alkalinity [7], thus preventing the pH rising. Nevertheless, the strong basic overwhelms the weak acidic tendency, hence the pH values are always increasing [10]. The pH rising rate increased with B₂O₃ doping, which is closely related to the lower chemical durability of bioglass caused by a higher B₂O₃ content [10]. The results of pH values at day 1 (Figure 4(a₁)) further prove the above point, as Na and K ions are more loosely bound in the bioglass with low stability.

Figure 4. (a) pH values and (b WL percentages of 0B, 1B, 2B and 3B bioglasses after immersion in SBF at 37°C for different days: (a₁) enlarged image of (a) at day 1, (b₁) WL percentages of the four bioglasses after immersion in deionized water at 37°C for 1 day.

The WL of 0B, 1B, 2B and 3B bioglasses as a function of immersion time show an approximately similar trend (Figure 4(b)), which increase gradually at beginning, then rapidly increase and finally slow down again. The slow degradation at beginning is major due to the release of Na, K, Mg and Ca ions from bioglass. Meanwhile, Si–OH layer was formed on the bioglass surface [31,32. Thereafter, the network-forming species were attacked by SBF, resulting in the break of glass structure and a rapid degradation at the middle stage of immersion. Subsequently, the Si–OH layer became thicker, meanwhile, the Ca–P compound precipitated and formed into hydroxyapatite on the bioglass surface [10]. Both factors impeded the bioglass degradation. The highlight is that the degradation rate accelerated with B₂O₃ doping. After 14 days immersion, the WL percentages of 3B bioglasses was six times larger than that of 0B bioglass. This is because of the low connection and weak stability of glass structure caused by B₂O₃ doping. Actually, the glass undergo preferential dissolution of BO₃ units, as compared to BO₄ units, which also accelerates the bioglass degradation [6]. The Si-OH layer [7] and Ca-P compound [32,33] on the bioglass surface may affect the results. Hence, the bioglasses were soaked in deionized water (Figure 4(b₁)). The results further confirm the above point. Practically, the barrier layer reduces the ion release from bioglass, which in turn proves the faster degradation rate of bioglass with higher B₂O₃ content.

The ions released from 0B, 1B, 2B and 3B bioglasses are shown in Figure 5. The release amount of Ca ions is always larger than that of B and Si ions. It is related to the structural sites of ions where Ca is in the interspace, while B and Si are both network builders. Compared to Si ions, B ions are more susceptible to erosion, because the strengths of boron units are weaker than that of silicon units [8]. As shown in Figure 5(a,b), the ions release rate of the four bioglasses increased initially and then decreased significantly, but soon rose again and finally slowed down. Here are possible scenarios: at the beginning, the surface area contacted with solution and the concentration gradient were both large, thereby the ions released instantly in large quantities. Later, the Si–OH layer and/or Ca–P compound formed on the surface prevented further ions release. After the bioglass surface fell off, the ions from inner layer released into solution immediately, so that the ions release rate rose again, but not as much as before. Notably, the release rates of B from 2B and 3B bioglasses were very close to that of Ca, which caused by the poor stability of glass structure with B₂O₃ doping. However, the momentum of B release was not maintained at the later stage of degradation. It could be attributed to the initial ‘burst’ of B that result in a small amount of B remaining. These results indicated that the bioglasses degrade from the surface to inside, layer by layer, and the bioglasses undergo preferential ions dissolution order of Ca＞B＞Si. When B₂O₃ content exceeds the SiO₂ content, the poor chemical durability of bioglass and increase number of BO₃ units will both accelerate the B ions release. The visual models of bioglasses with different B₂O₃ content before and after immersion are shown in Figure 6.

Figure 5. (a–d) Ions release and (a₁–d₁) ions release rate of 0B, 1B, 2B and 3B bioglasses after immersion in SBF at 37°C for different days.

Figure 6. Models of 0B, 1B, 2B and 3B bioglasses before and after degradation.

In vitro cytocompatibility of 0B, 1B, 2B and 3B bioglasses towards BMSCs shows in Figure 7. The quantitative analyses of the cell viability show no statistical difference (P > 0.05), meaning all the four bioglasses are safe to cells. Besides, the fluorescent images of F-actin (red) and cytoblast (blue) of BMSCs after 3-day incubation were collected. It can be observed that, in all the four experimental groups, a decent number of cells presented in the field of vision. Compare to the control group, all cells cultured with bioglass extracts displayed good spreading morphology. This further proves that the bioglasses with different B₂O₃ content have good biocompatibility.

Figure 7. (a) Quantitative analyses of the cell viability of BMSCs after incubation with ionic dissolution products of 0B, 1B, 2B and 3B bioglasses. (b) Fluorescent images of BMSCs morphology after 3 days of incubation with ionic dissolution products of 0B, 1B, 2B and 3B bioglasses.

Conclusion

In this research, the effects of B₂O₃ on the structure and degradation behaviour of borosilicate bioglass were systematically studied and clarified. The addition of B₂O₃ leads to the diversification of structural groups and the transformation between BO₄ units and BO₃ units. The two factors weakened the chemical durability and accelerated the degradation as well as ions release of borosilicate bioglass. In the degradation process, the borosilicate bioglass undergo preferential dissolution of alkali metal species and alkaline earth species, as compared to network-forming species. However, when B₂O₃ content exceeds the SiO₂ content, the B ions release rate significantly increases and can almost catch up Ca ions release rate, while the Si ions always release in a low speed. The predictive models of 1393 derived borosilicate bioglass before and after immersion in biological fluid improves our understanding of the relationship of composition–structure–property. It may also provide a theoretical basis for bioglass design, in order to meet the various requirements of degradation property and ion release in bone repair.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The work was supported by the National Natural Science Foundation of China [51972232] and the National Key Research and Development Projects Program of China [2018YFC1106300].