In vitro cellular testing of strontium/calcium substituted phosphate glass discs and microspheres shows potential for bone regeneration

Abstract Phosphate‐based glasses (PBGs) are ideal materials for regenerative medicine strategies because their composition, degradation rates, and ion release profiles can easily be controlled. Strontium has previously been found to simultaneously affect bone resorption and deposition. Therefore, by combining the inherent properties of resorbable PBG and therapeutic activity of strontium, these glasses could be used as a delivery device of therapeutic factors for the treatment of orthopaedic diseases such as osteoporosis. This study shows the cytocompatibility and osteogenic potential of PBGs where CaO is gradually replaced by SrO in the near invert glass system 40P2O5·(16‐x)CaO·20Na2O·24MgO·xSrO (x = 0, 4, 8, 12, and 16 mol%). Direct seeding of MG63 cells onto glass discs showed no significant difference in cell metabolic activity and DNA amount measurement across the different formulations studied. Cell attachment and spreading was confirmed via scanning electron microscopy (SEM) imaging at Days 3 and 14. Alkaline phosphatase (ALP) activity was similarly maintained across the glass compositions. Follow‐on studies explored the effect of each glass composition in microsphere conformation (size: 63‐125 μm) on human mesenchymal stem cells (hMSCs) in 3D cultures, and analysis of cell metabolic activity and ALP activity showed no significant differences at Day 14 over the compositional range investigated, in line with the observations from MG63 cell culture studies. Environmental SEM and live cell imaging at Day 14 of hMSCs seeded on the microspheres showed cell attachment and colonisation of the microsphere surfaces, confirming these formulations as promising candidates for regenerative medicine strategies addressing compromised musculoskeletal/orthopaedic diseases.


| INTRODUCTION
Phosphate-based glasses (PBGs) are highly suited biomaterials to deliver therapeutic ions, particularly for bone regeneration, because their main constituents (i.e., calcium, phosphate, sodium, and magnesium) are found to occur in the inorganic phase of natural bone (Jones & Clare, 2012). The ability to control their dissolution rates, and thus the release of ions, serves as a highly attractive route to direct the biological response for the repair and regeneration of both hard and soft tissues (Hoppe, Guldal, & Boccaccini, 2011). Furthermore, the geometry of PBG can be varied by means of fibre drawing (Ahmed, Collins, Lewis, Olsen, & Knowles, 2004;Sharmin, Rudd, Parsons, & Ahmed, 2016), scaffold production (Jones et al., 2010), and microsphere manufacture (Jones et al., 2010;Lakhkar et al., 2012), making them an extremely versatile and tuneable material.
A particular ion of interest is strontium (Sr), especially for use in hard-tissue repair applications; it has been investigated for incorporation and/or substitution into a vast array of materials such as organic polymers, ceramics, glasses, composites, and alloys (Han et al., 2016;Liu, Li, Ji, & Jia, 2013;Mushahary et al., 2016;Park, Kang, & Hanawa, 2016;Ray et al., 2016;Schumacher & Gelinsky, 2015;Wu, Lin, Yang, & Lee, 2015). Strontium is also of clinical importance because it has been shown to promote bone formation via stimulation of pre-osteoblastic cells replication and to reduce bone resorption via inhibition of osteoclast formation and activity (Canalis, Hott, Deloffre, Tsouderos, & Marie, 1996). In the past, strontium ranelate was administrated to osteoporotic patients, due to its anabolic activity and antiresorptive effect on bone tissue (Ammann, 2005), as the imbalance of bone deposition and resorption characterises osteoporosis (Blake & Fogelman, 2006). However, since 2013, the European Medicine Agency has restricted the use of strontium ranelate for the treatment of severe osteoporosis in postmenopausal women and adult men at high risk of fracture due to increased cardiovascular risks (National Institute for Health and Care Excellence (NICE), 2011). Therefore, investigation into alternative administration routes such as controlled and localised delivery of the ion, as opposed to systemic delivery, could be very beneficial.
Substitution of calcium for strontium in biomaterials (e.g., orthopaedic cements and bioactive glasses) has been of interest as both play a similar chemical role because they are divalent cations with a comparable ionic size (Ca 2+ 1.00 Å and Sr 2+ 1.18 Å) and are accumulated in bone. Due to the ease of substitution of these two cations within a glass system, as well as the positive influence strontium can have on bone metabolism, the use of this ion for bone tissue engineering and regenerative medicine strategies constitutes an area of growing interest (Abou Neel et al., 2009;Al Qaysi et al., 2015;Lee, Obata, Brauer, & Kasuga, 2015). Several studies have investigated the behaviour of cells such as osteoblast-like cells and mesenchymal stem cells towards strontium-containing biomaterials (Aina et al., 2013;Lin et al., 2013;Ni, Shu, Huang, Lu, & Pan, 2012;Santocildes-Romero et al., 2015). For instance, Sr-substituted hydroxyapatite nanocrystals were shown to dose dependently promote higher levels of alkaline phosphatase (ALP) activity, collagen type-I, and osteocalcin production in comparison with pure hydroxyapatite in the osteoblastlike MG63 cell line (Capuccini et al., 2009). Investigation into the effects of Sr-containing PBG has also been reported; for example, human osteosarcoma cells showed higher degree of adhesion to PBG with 1 and 3 mol% SrO substitution in comparison with the 5% and SrO-free control (Abou Neel et al., 2009).
The aim of this study was to investigate the cytocompatibility of near invert phosphate glasses where calcium was eventually fully substituted for strontium on a gradual basis in a previously reported glass system 40P 2 O 5 ·(16-x)CaO·20Na 2 O·24MgO·xSrO (x = 0, 4, 8, 12, and 16 mol%;Patel et al., 2017). All formulations were manufactured as discs and microspheres and tested for their ability to support adhesion, proliferation, and osteogenic differentiation using human osteoblast-like MG63 cells and human mesenchymal stem cells (hMSCs), respectively, in order to investigate their potential use for bone repair applications via two clinically relevant cell types.

| PBG sample production
The following precursors were used to fabricate glass rods of the compositions shown in Table 1: P 2 O 5 , NaH 2 PO 4 , CaHPO 4 , MgHPO 4 ·3H 2 O, and SrCO 3 (Sigma Aldrich, UK). The weighed precursors were thoroughly mixed and then heated in a 5% Au/95% Pt crucible at 350°C for 30 min to dehydrate and remove CO 2 before placing the crucible into a furnace heated to 1150°C for 90 min. The melts were poured into preheated graphite moulds (10°C) above Tg of each sample and annealed for 1 hr, followed by slow cooling to room temperature overnight. Each rod was then cut into discs of 9 mm × 2 mm using a diamond saw using methanol as a lubricant agent. Sterilisation of the discs was carried out via UV light exposure for 1 hr on either side.

| Glass microsphere production
The quenched glass samples were ground using a planetary zirconia ball mill (Retsch Planetary Mill PM100), operated for 2 min at 500 rpm.
Glass powder was separated into particle size in the range 63-125 μm using stainless steel sieves (VWR International, UK). The ground particles were fed through a flame spheroidisation apparatus consisting of (a) a powder feeder, (b) an oxy/acetylene thermal spray gun (MK74 Thermal Spray gun Metallisation Ltd., UK), and (c) a series of collection vessels. Microspheres were collected from the collection vessel furthest away from the spray gun. Microspheres of all compositions were sterilised by washing twice in 70% EtOH for 15 min followed by complete evaporation at room temperature in a sterile environment.

| Cell culture
For cell experiments, materials were purchased from ThermoFisher Scientific (UK) unless otherwise stated.  Figure 1a).
Moreover, the quantification of total cell DNA amount showed a significant increase at each time point up to Day 7 (D1 vs. D3: p < 0.01; D7 vs. D3: p < 0.0001), followed by a plateau at Day 14. It is worth noting that no significant differences were observed between all formulations and the TCP control at each time point (p > 0.9; see Figure 1b).

| Evaluation of hMSC cell growth on microsphere PBG formulations
The metabolic activity of hMSCs cultured on microspheres manufactured from each PBG composition was assayed at Days 2, 7, and 14 after seeding. The results showed a significant increase in cell response over time (p < 0.0001), whereas no differences between cell cultures on the different PBG compositions were detected at each time point (p > 0.90; see Figure 2).

| ALP activity
ALP activity was measured as an early marker of osteogenic differentiation in MG63 and hMSCs after 14 days of culture on both PBG discs and microspheres, respectively.

| hMSC cell imaging on microspheres
Live fluorescence images of hMSCs showed a similar level of cell adhesion to PBG microspheres across all formulations at Day 1, whereas the presence of cell-microsphere aggregates was observed at Day 14 (see Figure 5a). These aggregates were further analysed via ESEM imaging, which showed the attachment and spreading of cells on the microsphere surface, as well as the presence of a dense cell-derived matrix surrounding the microspheres (see Figure 5b).

| DISCUSSION
This study aimed to assess the biological properties of a series of near invert PBGs where CaO content was gradually substituted for SrO by 0% to 100% (Patel et al., 2017) and manufactured in disc and microsphere forms. Microspheres in particular provide the advantage of uniformity in size and shape, making delivery to the target tissue possible through simple injection procedures (Kim & Pack, 2006). Spherical material morphologies also enable enhanced versatility for filling even complex defect shapes compared with bulk scaffolds, which tend to have predetermined shapes (Choi, Zhang, Yeh, Wooten, & Xia, 2012). The increased surface area also allows for efficient cell attachment and spreading, whereas the microsphere interparticulate gaps would be beneficial for oxygen and nutrient flow (Cai, Chen, Hong, Liu, & Yuan, 2013). MG63 cells have been widely used as a versatile tool for the evaluation of the osteogenic potential of new materials and compounds (Rodriguez, Saxena, Hixon, Sell, & Bowlin, 2016), whereas hMSCs cells have been included in this study as a clinically relevant cell model, largely used in in vitro and in vivo studies of bone regeneration (Knight & Hankenson, 2013).
As previously demonstrated, the gradual replacement of CaO content for SrO by 0% to 100% in the glass system used in the this study (40P 2 O 5 ·(16-x)CaO·20Na 2 O·24MgO·xSrO, where x was 0, 4, 8, 12, and 16 mol%) did not significantly alter the PBG dissolution rate, providing a means of controlled release of therapeutic ions such as strontium, magnesium, phosphate, and calcium (Patel et al., 2017). The In vivo too, the degradation rate of a material can significantly affect the outcome of the therapeutic strategy by either altering the structural support provided to the endogenous tissue and/or stimulating the inflammatory response (Vishwakarma et al., 2016;Yu, Tang, Gohil, & Laurencin, 2015).
The dissolution behaviour of the PBG formulations studied here has been previously reported, showing a more stable PBG surface with the addition of SrO, with a weight loss ranging between 0.2% and 0.13% in 30 days for the discs (Patel et al., 2017). Evidence also suggests that the degradation rate of the glass increases with spherical geometries (Hossain et al., 2018;Islam, Hossain, Sharmin, Parsons, & Ahmed, 2017); however, in vivo studies would need to The ability of a biomaterial to induce osteogenic differentiation is a desirable property in view of the application for bone repair. In this study, the assay of ALP activity was performed as an early marker of osteogenic commitment. This enzyme is constitutively active at low levels in all the cells, but its activity significantly increases during the early stages of osteogenic differentiation; therefore, it is commonly considered a good marker for the detection of early osteogenic differentiation (Golub & Boesze-Battaglia, 2007;Prins et al., 2014;Stein, Lian, & Owen, 1990). Strontium ranelate, an Sr-based drug previously administrated to osteoporotic patients, has been reported to have a dual role by enhancing differentiation of osteogenic progenitor cells as well as by inhibiting osteoclastogenesis (Brennan et al., 2009;Fonseca & Brandi, 2010 et al., 2015). Incorporation of 5 and 10 at% of SrO into other biomaterials such as hydroxyapatite resulted in significant increase of ALP activity compared with the TCP control in a dose-dependent manner (Boanini et al., 2014). It seems that such osteogenic effect exerted by strontium may be mediated by the activation of the Wnt/β-catenin pathway both in vitro and in vivo (Yang et al., 2011), by stimulating the activity of the ALP enzyme as well as the production of ECM components including collagen type-I and osteopontin. It is conceivable that the observed response may be the result of a combination of factors rather than the sole exposure to strontium including the combination and rate of ions simultaneously released during glass degradation, which have all been described to influence cell behaviour (Maeno et al., 2005).
Overall, the findings of this study complemented previous works, which aimed to characterise the effect of gradually substituting CaO for SrO in a PBG system. More interestingly, our group has recently reported the manufacturing of PBG microspheres in highly porous conformation with larger surface area and increased degradation rate (Hossain et al., 2018). These microspheres have also shown the ability to encapsulate stem cells and will be further investigated for promoting cell growth and pore colonisation, to deliver enhanced advantages for bone tissue growth via a biomimetic environment. Future work will also be aiming to manufacture novel formulation of porous PBG including therapeutic ions such as strontium with the aim to provide new insights for the design of novel materials for minimally invasive orthopaedic repair and regeneration applications. A comprehensive evaluation of the biological potential of these formulations in vitro would provide important additional information for the further progression toward in vivo studies. In particular, in view of the anti-resorptive activity attributed to strontium as well as the importance of appropriate vascularisation to support the new tissue, further studies could evaluate how these glass formulations can affect the activity of osteoclast cells and the formation of vessel-like structures by endothelial cells.

| CONCLUSIONS
This study showed that discs and microspheres manufactured from the near invert glass system 40P 2 O 5 ·(16-x)CaO·20Na 2 O·24MgO·xSrO (x = 0, 4, 8, 12, and 16 mol%) are cytocompatible and promoted adhesion and growth of human MG63 and hMSCs, respectively, up to 14 days. The observations made in this study suggest that these materials are appropriate for the culture, expansion, and cellular activity of human cells constituting as excellent candidates for bone regeneration.