Storage and release of rare earth elements in microsphere-based scaffolds for enhancing osteogenesis

In osteoporosis and diabetes, it is essential to accelerate the bone repair and regeneration process. Trace rare earth elements such as lanthanum (La) ions (La3+) with appropriate concentrations are bioactive and can effectively regulate bone tissue performances. However, few well-established bone tissue engineering scaffolds can precisely and stably release La3+ to promote bone regeneration significantly. Based on the advantages of biodegradable microspheres and microsphere-based scaffolds for controlled drug release, we developed poly(lactide-co-glycolide) (PLGA)-based microsphere-based scaffolds as both three-dimensional (3D) porous scaffolds and La3+ storage and release systems for osteogenesis. So far, there is no study about microsphere-based scaffolds to release trace La3+ to induce osteogenic differentiation of bone marrow mesenchymal stromal cells (BMSCs). PLGA microspheres co-embedded with La-doped mesoporous silica (LMS) with different amounts of doped La were sintered to prepare the LMS/PLGA (LMSP) microsphere-based scaffold. The La3+ release behavior of LMSP can be controlled by adjusting the doping amount of La in mesoporous silica (MS). All these scaffolds possessed a 3D network architecture. With the increase of La doping, LMSP can better compensate for the pH decrease caused by PLGA degradation. The combination of MS and PLGA can avoid the cytotoxicity of MS alone. All prepared LMSP scaffolds were non-cytotoxic. After BMSCs were implanted on scaffolds, LMSP could promote cells adhesion, proliferation, and osteogenic differentiation. Among these microsphere-based scaffolds, LMSP-3 with stable and higher dose La3+ release behavior showed the strongest ability to enhance the osteogenesis of BMSCs. The results showed that microsphere-based scaffolds with the ability to store and stably control the release of La3+ could effectively improve osteogenic performance, which provides a new idea for the construction of bone tissue engineering scaffolds.


Preparation of LMS and LMSP microspheres scaffolds.
Traditional techniques were used to prepare MS 23 . LMS was prepared as follows: DDA was dissolved in EtOH/deionized aqueous solution (pH = 9). Then, 0.5 g, 1.0 g, 1.5 g, or 2.0 g of La(NO 3 ) 3 ·6H 2 O and TEOS were added as sources of La and Si, respectively, and stirred. The molar composition of the reaction mixture is TEOS: 1.0, DDA: 0.27, EtOH: 9.12, H 2 O: 29.6. The mixture was stirred for 18 h, aged at room temperature for 12 h, dried at 90 °C for 6 h, cleaned with ethanol, and calcined to remove DDA. The LMS prepared by adding 0.5 g, 1.0 g, 1.5 g, and 2.0 g La source were labeled separately as LMS-1, LMS-2, LMS-3, and LMS-4.
MSP and LMSP microspheres were all prepared by the following two methods: the single-emulsion solvent volatilization method and the S/O/W method. In short, 1 g PLGA and 0.15 g MS or La-MS particles were dissolved in 5 mL DCM and mixed within 2 min. The synthetic mixture was then injected into a 1.0% PVA solution and stirred for 10 h at 300 rpm to evaporate the solvent completely. Microspheres were separated and washed Scientific Reports | (2022) 12:6383 | https://doi.org/10.1038/s41598-022-10347-0 www.nature.com/scientificreports/ with deionized water 5 times. MSP and LMSP microspheres were poured into a cylindrical mold and sintered at 70 °C for 2 h. The products of MS, LMS-1, LMS-2, LMS-3, and LMS-4 added LMSP microsphere-based scaffolds were labeled as MSP, LMSP-1, LMSP-2, LMSP-3, and LMSP-4, respectively.
SEM analysis. The morphology of scaffolds was characterized by scanning electron microscopy (SEM, 30XLFEG, Philips, The Netherlands). The elemental compositions of the MS powder and microsphere-based scaffolds were analyzed using an energy dispersive spectrometer (EDS) in conjunction with the SEM system.
Density and porosity determination. The density and porosity of sintered microsphere-based scaffolds were tested refer to Ref. 34 . Briefly, EtOH is selected as the liquid phase, and the temperature is kept at 25 °C. The bottle is filled with EtOH (weight W 1 ), the cylindrical scaffolds (diameter 10 mm, height 20 mm, weight W S ) is put into the bottle, and an equal volume of ethanol spills out. The bottle filled with the scaffolds and the remaining ethanol weighs W 2 . ρ is the density of EtOH at 25 °C. Using the following formula to calculate the porosity (P) and density (D) of the scaffold: Static contact angles. 150  Cell culture on LMS and microsphere-based scaffolds. BMSCs were purchased from American type cultured specimens (ATCC, Manassas, VA). For MS and LMS-1-4 powders, the cells were resuspended in a fresh culture medium and then seeded at the density of 2 × 10 3 cell/well on MS or LMS-1-4 for studies on cell proliferation. The concentration of MS and LMS-1-4 in every well was 0 ppm, 50 ppm, 100 ppm, 200 ppm, and 400 ppm, respectively. For studies on cell proliferation on microsphere-based scaffolds, cells were seeded on the scaffolds (diameter = 10 mm, height = 5 mm) at a density of 1 × 10 5 cells/scaffold, and then DMEM containing 10% fetal bovine serum (FBS) was added for culture.
Cell proliferation. The cell proliferation of each scaffold was analyzed quantitatively by the CCK-8 method.
In short, the medium was removed and the cells were washed twice with PBS (pH = 7.2) at specified intervals. CCK-8 solution was added to each well and incubated in an incubator for 2 h. The absorbance was measured at 450 nm wavelength with the microplate reader (Thermo3001, America). Twenty-four hours after the cells were inoculated with the scaffolds, live/dead assay was performed using the calcein-AM/propidium iodide double staining kit (Sigma).
Osteogenic differentiation. The osteogenic differentiation of BMSCs was observed by culturing the cell scaffolds (2 × 10 5 cells/scaffold) with osteogenic medium (OGM). ALP activity was quantified using the p-nitrophenyl phosphate liquid substrate (Sigma Diagnostics) method. In short, the cell scaffolds were prewashed with PBS, and then the adherent cells were removed from the scaffold and lysed at 4 °C for 10 min, then dissolved in 0.5 mL PBS containing 0.1 M glycine, 1 mM MgCl 2 , and 0.05%Triton X-100 for 10 min. Then it was incubated with p-nitrophenyl phosphate (pNPP) solution at 37 °C for 30 min. To extract the cell layers, the cells were treated with sonication twice for 30 s and then centrifuged at 12,300 rpm for 2 min at 4 °C. ALP activity was assayed in the supernatant, as in Ref. 35 . The ALP values (U/µg) were normalized to protein content using the Micro BCA Protein Assay Kit (Pierce). Meanwhile, ALP activity was detected by ALP staining on day 10. In brief, the cell scaffolds were washed with PBS and fixed with 10% neutral formalin solution for 30 min. The cell scaffolds were dyed with BCIP/NBT at 37 °C for 30 min and then washed with distilled water.
On day 21, the collagen secretion of cells on the scaffolds was detected by Sirius Red staining 35 . After PBS washing and fixation, 0.1% Sirius red (Sigma) staining showed collagen. Unbound stains were washed with www.nature.com/scientificreports/ 0.1 M acetic acid before taking photos. Collagen type I (COL-I) were detected using the COL-I enzyme-linked immunosorbent assay (ELISA) kit (Thermo Fisher). The RNAs of BMSCs were extracted after being cultured for 21 days for reverse transcription real-time polymerase chain reaction (qRT-PCR) evaluation. Four genes related to osteogenesis were selected for analysis. That is bone morphogenetic protein 2 (BMP2), osteocalcin (OCN), COL-I, and RunT-associated transcription factor 2 (RUNX-2). The primers used in this study were designed by Primer Premier 6 software (Premier Biosoft, USA) and are listed in Table 1. Cell/scaffold complexes were crushed in a frozen state using liquid nitrogen, and then the Trizol RNA extract kit (Invitrogen, Thermo Fisher) was used to extract the total RNA. According to the primers shown in Table 1, the qTOWER RT-PCR system (Toyobo, Japan) was applied to perform analysis. And data was processed by the ΔΔCt method with 18S-rRNA as the normalization standard.

Statistical analysis.
The experiments were repeated five times, and the results were expressed as mean ± standard deviation. The results were calculated by one-way ANOVA and were statistically significant. Tukey test was used for mean comparison, and the difference was statistically significant as p < 0.05.

Gene
Forward primer (5′-3′) Reverse primer (5′-3′) CCC GAC GCT TCT  CCT GCA TTT GTT CCC GAA AA   OCN  CAA ACA CGG CAA GGT GTG TGA CGA AGG TCT TGT TGT CAT TGCTG   COL-I  GCG GTG GTT ACG ACT TTG GTT  AGT GAG GAG GGT CTC AAT CTG   Runx-2  CCC AAG CAT TTC ATC CCT CACT CAT ACC GAG GGA CAT GCC TGA   18S  GTA ACC CGT TGA ACC CCA TT  CCA TCC AAT CGG TAG (Table 2). After LMS-1-4 was added into PLGA to prepare microsphere scaffolds, the Si content decreased to about 5%, and LA could not be detected. This phenomenon indicates that there were few LMS-1-4 particles distributed on the surface of the scaffolds. The content of C from PLGA is about 45%. The slight changes of Si, O, and C contents in MSP and LMSP1-4 groups were related to the spatial distribution of MS or LMSP1-4 in PLGA microspheres. The contact angles of the LMSP-1-4 composite surface decreased after MS doping La. Compared with the MSP group, with the increase of La doping amount, the surface contact angle decreased significantly, indicating that LMS can improve the surface hydrophilicity of LMSP1-4 composites (Fig. 2B). The contact angle of LMSP-3 was approximately 71.31.
Release behavior of La 3+ and change of pH value. The La 3+ release behaviors of the LMSP scaffolds were studied by soaking LMSP in PBS at 37 °C for 28 days. At the set time points, the ion concentrations were detected by ICP (Fig. 3). La 3+ is released rapidly within the first three days. As the soaking time increased, the concentration of La 3+ began to fluctuate dramatically, except for the group of LMSP-2 and LMSP-4. At all time points, the release amount of La 3+ in the LMSP-4 group was much higher than that in the LMSP-2 group. What's more, the La 3+ concentration range of all groups was obtained: 0.079 ± 0.002-1.212 ± 0.05 ppb, which was in the safe concentration range. Figure 4 shows the pH change trend of PBS during the degradation of MSP and LMSP microsphere scaffolds. Within 4 weeks, with the increase of La content in the composite scaffolds, the pH value of PBS around the scaffolds increased. At 4 weeks, the pH of LMSP-1-4 groups was significantly higher than the MSP group, and the pH of the MSP group was 3.68 ± 0.19 with that of the LMSP-4 group was 6.39 ± 0.22.

Cells proliferation.
Firstly, a CCK-8 assay was used to determine the cytocompatibility of MS and LMS-1-4. According to the detection results of CCK-8, after co-culture with LMS with different amounts of La, the cell viability on day 1 and day 3 showed a similar trend to that on day 7 (data not shown). After 7 days, BMSCs viability in high concentration groups (100, 200, and 400 ppm) decreased significantly compared with  www.nature.com/scientificreports/ the untreated control group (TCPS cultured cells without materials) (Fig. 5A). The inhibition efficiency was enhanced, especially when the La incorporation amounts were increased. The cell viability of the 50 ppm LMS-1-4 groups was close to that of the untreated control group, indicating that the low concentration LMS-1-4 had good cytocompatibility. After 3, 7, and 10 days of culture, the CCK-8 method was used to quantitatively analyze the proliferation of cells on the scaffolds (Fig. 5). The cells seeded on LMSP scaffolds with different La levels continued to grow from day 3 to day 10, indicating that all scaffolds had no cytotoxicity. Compared with the MSP group, the absorbance of the LMSP-1-4 scaffolds is close to that of the control group, and there is no statistical difference. The results showed that the La-doped modification of MS had no adverse effect on the cell proliferation of the scaffold under the experimental conditions. The green fluorescence shown in the live/dead staining images represents living cells, and all scaffolds showed good cell affinity and no cytotoxicity, consistent with the results of the CCK-8 assay (Fig. 6).
Osteogenesis. The ability of microsphere-based scaffolds to induce osteogenic differentiation was determined primarily by the expressions of general marker proteins like ALP and COL-I. The secreted ALP content was analyzed quantitatively after the cells were cultured on the scaffold for 3, 7, and 10 days (Fig. 7A). Cells cultured on LMSP-3 scaffolds showed higher ALP activity at each time point. In the ALP staining results (Fig. 8), the cells showed the same trend after being cultured on the scaffold for 10 days, which was consistent with the quantitative results.
The secreted COL-1 content was analyzed quantitatively after the cells were cultured on the scaffold for 7, 14, and 21 days (Fig. 7B). The COL-1 content gradually increased along with the inductive culture and displayed  www.nature.com/scientificreports/ a similar ascending trend between groups to that observed for the ALP activity. Overall, BMSCs cultured on LMSP-3 scaffolds showed higher COL-1 content at each time point. As shown in Fig. 9, the collagen secreted by the cells is characterized by Sirius red. The results showed that all components secreted collagen at 21 days. The amount of collagen in the control group (MSP group) was the least, while that in the LMSP-3 group was still the most. It was also consistent with the quantitative results. To further identify the ability to induce osteogenic   www.nature.com/scientificreports/ differentiation, qRT-PCR was applied to detect the expression of osteogenic-related genes (Fig. 10). Results showed a better ability of LMSP-3 microsphere-based scaffolds to induce osteogenic differentiation.

Discussion
In this study, LMSP scaffolds with 3D porous structures and non-cytotoxicity were constructed for bone tissue regeneration and repair. Compared to MSP scaffolds, LMSP groups not only can release La 3+ but also can compensate for pH. Moreover, they have better hydrophilicity and osteogenic differentiation ability. The results revealed that the LMSP group could significantly promote the osteogenesis differentiation of the cells. This conclusion can be supported by the biocompatibility, better ALP activity, collagen secretion, and gene expressions of BMP-2, OCN, COL-I, and Runx-2 of BMSCs. Because microspherical-based scaffolds have excellent initial mechanical properties, they have attracted much attention. To construct LMSP bone tissue engineering scaffolds with good performance, there took the following   www.nature.com/scientificreports/ two preparation steps. First, the PLGA microspheres were prepared via a S/O/W emulsion technique containing LMS particles and La 3+ . Then, prepare microsphere-based scaffolds by low-temperature sintering technique (Fig. 1). The MSP and LMSP-1-4 scaffolds showed 3D interconnected macropores with porosity of 30-35% ( Fig. 2A). Compared with 3D printed scaffolds, the porosity of microsphere sintered scaffolds is much smaller. However, in this study, microsphere sintered scaffolds had visible micropores (more than 100 μm in size), and these sizes allowed the ingrowth, proliferation, and differentiation of bone-associated cells 28,36,37 . When combined with PLGA, LMS enhanced the osteogenic differentiation ability of BMSCs without affecting the cellular activity of cells. From the results of the CCK-8 testing, the LMSP scaffolds showed no cytotoxicity (Fig. 5B). The porous structure of the scaffolds facilitates the cells to enter into and promotes communication between cells. BMSCs enter the scaffolds through channels and adhere closely to the surface of the scaffold (Fig. 6). This result is consistent with the findings of others that cell viability was not affected after cells were grown on scaffolds doped with appropriate amounts of La-doped material 18 . However, LMS-1-4 at high concentrations (100, 200, and 400 ppm) have cytotoxicity (Fig. 5A). Only 50 ppm of LMS-1-4 had good cytocompatibility. According to reports, apoptosis may occur when La 3+ concentration exceeds 10 -5 M 15,38 . It may be because when the concentration of LMS-1-4 is higher, they also release a higher concentration of La 3+ , thus reducing the survival rate of BMSCs. In this study, at all the time points, the concentration of La 3+ ions released from the LMSP scaffold was only about 0.1-1.2 ppb (Fig. 3), which was within the safe concentration range 17,39 .
PLGA-based microspheres have been known as drug carriers for decades. The release of drugs from microspheres depends on drugs diffusion through the polymer matrix and the degradation of the polymer 40 . However, compared with drugs, the release kinetics of La 3+ from microspheres is mainly controlled by La-doped material because the diameter of La 3+ is tiny (only about 0.21 nm). The degradation rate of Si-based materials is closely related to their chemical composition 15 . In this study, the residual template in the preparation of LMS was removed by calcination, and the degradation rate of LMS was slow due to crosslinking caused by high temperature 21 . We adjusted the doping amount of La in LMS and incorporated LMS-1-4 into PLGA to obtain an adjustable release behavior of La 3+ . When PLGA was mixed with a relatively small amount of LMS-1-4 to prepare microsphere-based scaffolds, only about 5% Si and La could not be detected on the surface of scaffolds due to the small amount of LMS-1-4 distributed on the surface of scaffolds (Fig. 1, Table 2). When LMS is encapsulated by PLGA, the release of La 3+ from LMS will be inhibited. ICP results showed that the concentration of La 3+ fluctuated during the whole soaking process (Fig. 3). During the first 3 days of soaking, the concentration of La 3+ in the solution increased with the continuous degradation of LMS. When more La sources are added, more La replaced Si, forming more Si-O-La bonds, resulting in higher La content and lower Si content in LMS (Table 2). At the same time, the difference between the diameters of La and Si atoms will lead to more defects in the mesoporous skeleton. This further accelerated the degradation rate of LMS and La 3+ release, in agreement with other studies 16,19 . Compared with LMS-1-3, the addition of more La sources in the preparation process of LMS-4 leads to more lattice defects. This may be the reason for the larger amount of La 3+ release from LMSP-4 than from LMSP-1-3 and the larger fluctuations in the release profile throughout the release process. P in PBS is deposited on LMSP to form a phosphate layer, which will gradually hinder the diffusion of La 3+ , so the concentration of La 3+ in PBS fluctuates 41,42 . Meanwhile, a quarter of PBS was replaced at each time point, which also resulted in a decrease in the concentration of La 3+ .
In essence, multiple factors at the gene and protein levels governed BMSCs osteogenic differentiation. La 3+ plays an important role in improving the ability to induce osteogenic differentiation of scaffolds. As an early marker of bone formation, ALP is mainly expressed on the cell surface or in stromal vesicles and has a significant regulatory effect on phosphate supplementation during bone mineralization 39 . As an early-stage marker that is observed in the preparation stage of osteogenic differentiation, ALP activity was increased significantly in the LMSP-2-3 group especially in the LMSP-3 group (Figs. 7A, 8). The initial stage of collagen biosynthesis plays an important role in the development of mature bone tissue 43,44 . In this study, the osteoblasts in the LMSP-3 group also showed significantly higher collagen gene and protein secretion (Figs. 7B, 9, 10). BMP-2, OCN, and Runx-2 also take key roles in bone formation. Compared with other groups, LMSP-3 microsphere-based scaffolds showed the higher BMP-2, OCN, and Runx-2 gene expression (Fig. 10). This finding showed that compared with MSP scaffolds, LMSP-1-4 scaffolds especially LMSP-3 scaffolds, have higher osteogenic induction. LMSP-3 concentration in the right conditions could steadily release suitably does of La 3+ , which could support proliferation and promote the osteogenesis differentiation of BMSCs effectively. Previous studies have shown that La 3+ at concentrations of 0.1-1.0 ppb had promoted the osteogenic differentiation of BMSCs and MC3T3-E1 at a range of concentrations. So did-containing materials such as La-substituted layered double hydroxide nanohybrid scaffolds 15,18,41,45 . PLGA was approved by the FDA for use in humans several years ago and is ideal for preparing tissue engineering materials and drug carrier materials. It has many advantages, including good biocompatibility, adjustable degradation rate, good mechanical properties, and is easy to shape. However, PLGA degrades into water and CO 2 in vivo, and access CO 2 leads to increased acidity in the vicinity, which is prone to a local aseptic inflammatory reaction and clinical treatment failure. It is a major common obstacle in developing PLGA-based materials for tissue engineering especially bone tissue engineering. As we all know, the autocatalysis of PLGA may accelerate its degradation 46 . In this study, during the degradation of MSP and LMSP, the acidic degradation products produced by PLGA in the scaffolds were gradually dispersed into the medium, which resulted in the decrease of pH value (Fig. 4). For the LMSP-2-4 group, the pH remained above 6 throughout 4 weeks, which was significantly higher than the MSP and LMSP-1 group. According to our previous studies, LMS is more alkaline than MS. With the increase of the doping amount of La, LMS has more strong alkalinity 34 . It is mainly because La-MS could neutralize acid products. And the material has a more pH compensation effect. Therefore, LMS can balance the acidity caused by PLGA degradation and may effectively alleviate the immune response caused by the acidic environment in the clinic. After BMSCs were inoculated onto scaffolds and cultured with OGM, www.nature.com/scientificreports/ they could express osteogenic markers. Compared with MSP scaffolds, the ability of osteogenic differentiation of BMSC is stronger in the La-doped MSP scaffolds, which laid a foundation for the development of La-containing 3D porous microsphere-based scaffolds for bone repair.

Conclusion
Given the vital regulatory role of La in osteogenesis, it is of great significance and value to study the La 3+ release kinetics of biomaterials targeting bone regeneration to achieve stable and controlled release of La 3+ . In the current attempt, the steady release behavior of La 3+ has been controlled successfully by incorporating LMS with different amounts of doped La into bioresorbable PLGA microsphere-based scaffolds. These LMSP microsphere-based scaffolds provided feasibility to regulate the ALP and COL-1 secretion and osteogenic differentiation at a certain concentration of La 3+ . Based on this study, these LMSP microsphere-based scaffolds with La 3+ stable controlled release kinetics can be applied for further research to deepen the understanding of the role of La 3+ in promoting bone regeneration or other possible biological effects, such as bone immune regulation.