Enhanced osteogenic activity of Ti alloy implants by modulating strontium configuration in their surface oxide layers

To guarantee the long-term stability of an orthopaedic implant, non-degradable surface coatings with the ability to selectively release bioactive drugs or ions are especially desirable. In this study, SrO–TiO2 composite coatings were deposited on the surface of Ti alloys, whose release behavior of bioactive Sr ions was modulated by the Sr configurations, either interstitial atoms in solid solution (TiySr2−2yO2) or strontium titanate (SrTiO3). A perfect linear relationship between the amount of the released Sr ions and the Sr content in the coating was observed. Among the SrO-doped TiO2 coatings, the 20% SrO–TiO2 coating where Sr existed in both forms of TiySr2−2yO2 and SrTiO3 not only promoted proliferation of bone cells but also enhanced their osteogenic differentiation, which was proved to be related to its Sr release behavior. However, overdosing with 30% SrO only resulted in one single Sr configuration (SrTiO3) and an inferior osteogenic function. This study suggests that Sr configurations of both interstitial atoms of the solid solution and SrTiO3 can realize the selective release of Sr, but they possibly have different effects on the biological functions and other properties including corrosion resistance.


Introduction
A variety of metal ions have been documented to be essential for the human body and widely used as therapeutic agents in the treatment of many types of diseases, such as anemia, bone, brain and neuron diseases. [1][2][3] In bone tissue engineering, some metal ions, represented by strontium (Sr), 4 zinc (Zn), 5 magnesium (Mg) 6 etc. were proven to promote bone tissue regeneration and have been widely used in the eld of therapeutic tissue engineering. One of the most important advantages of using metal ions is that it does not pose risks of decomposition or instability, which are intrinsic to some drugs and proteins. 3,7 Among those ions reported to be essential for bone metabolism, Sr has been demonstrated to stimulate bone regeneration and inhibit bone resorption, [8][9][10] which has raised great interest in construction of osteoinductive bone biomaterials.
Titanium and its alloy are the most commonly used implant materials for dental and orthopaedic applications due to their good mechanical properties, biocompatibility and corrosion resistance. 11,12 However, they cannot achieve sufficient functional integration (osseointegration) with the surrounding bone to establish a rm and long-lasting anchor, as the implant surface lacks osteoinductivity. Therefore, surface activation is required to obtain a satisfactory long-term performance. 13,14 Beneting from the osteogenic function of Sr ions, they have been used by various methods to activate Ti alloy implant surface. [15][16][17][18] The commonest way is to introduce Sr in a biodegradable material that can be deposited onto the implant surface via a certain of surface techniques. [19][20][21] Upon degradation of the material, the Sr ions release and the resultant biological effects on bone cells can be realized. 16 However, since the bone implant is supposed to permanently stay in human body, the utilization of a biodegradable coating on its surface could risk the interfacial stability and probably cause implant aseptic loosening upon the coating absorption, if its degradation rate cannot match the new bone formation rate. 22 Therefore, to better use of the osteogenic benets of Sr and avoid the risk of implant loosening caused by the coating degradation, a chemically stable surface coating with an ability to release Sr ions would be preferred.
In this study, we choose biocompatible and chemically stable TiO 2 as the main component of the surface coating material, which is supposed to permanently exist during the service time of the implant. Sr ions were incorporated into the TiO 2 promote the osseointegration of the implant. To modulate the ion release behavior, the amount of Sr in the TiO 2 was tailored in order to control the Sr conguration (Sr-TiO 2 ) solid solution and strontium titanate (SrTiO 3 ) in the composite coating. The release of Sr is not associated with the degradation of the whole coating material, thus avoiding the risk of compromising the implant stability resulted from by the coating degradation. The ion release behavior, corrosion resistance and bioactivity of the Sr-doped coatings were evaluated and discussed.

Coating fabrication
TiO 2 and SrO nanopowders were used as starting materials. 10% SrO-TiO 2 , 20% SrO-TiO 2 and 30% SrO-TiO 2 composite powders were produced by mixing using planetary ball mill. Coatings were deposited on biomedical grade Ti alloy (Baoji Junhang Metal Material Co., Ltd. Shanxi, China) with a diameter of 15 mm and 1 mm thickness by an atmospheric plasma spraying system (9M, Sulzer Metco, USA). Before plasma spraying, the Ti alloy substrates were ultrasonically cleaned in absolute ethanol and sandblasted with brown corundum. For cell culture experiments, the Ti alloy discs without any coatings and with pure TiO 2 coating were used as controls. The main parameters used in this study to prepare the TiO 2 , 10% SrO-TiO 2 , 20% SrO-TiO 2 and 30% SrO-TiO 2 coatings are listed as follows: spraying power was 42 kW, Ar ow rate was 40 L min À1 , H 2 ow rate was 12 L min À1 , spraying distance was 100 mm and powder feed rate was 30 g min À1 .
2.2. Coating characterization 2.2.1. X-ray diffraction (XRD) and scanning electron microscopy (SEM). The phase structures of the coatings were conducted by X-ray diffraction (XRD, D/max 2500PC, Rigaku, Japan) with Cu Ka radiation (l ¼ 1.5418 A) in the range of 20-80 (2q). Scanning electron microscopy (SEM, S-3400, Japan) was used to examine the surface morphology.
2.2.2. Ion release prole. The coatings were immersed in 1 mL of a-minimum essential medium (a-MEM, Hyclone, USA) (pH ¼ 7.4) at 37 C. The medium was refreshed every 3 d. At each time point (3, 6 and 9 days), the culture medium was collected for measurement. The ion concentration of Ti and Sr in the culture medium was measured by inductively coupled plasma atomic emission spectroscopy (ICP-OES).

2.2.3.
In vitro cell-free mineralization. The as-sprayed TiO 2 and the SrO doped TiO 2 coatings were cleaned in an ultrasonic bath with ethanol and distilled water. Modied simulated body uid (2Â SBF) with Ca and P ion concentrations double those in the normal simulated body uid (SBF). The samples were soaked in the 2Â SBF at 37 C for 14 d without stirring and the solution were refreshed every 7 days. The 2Â SBF solution containing 142.0 mmol L À1 Na + , 5.0 mmol L À1 K + , 1.5 mmol L À1 Mg 2+ , 5.0 mmol L À1 Ca 2+ , 148.5 mmol L À1 Cl À , 4.2 mmol L À1 HCO 3 À and 2.0 mmol L À1 HPO 4 2À and 0. 5  The corrosion resistance of the coatings were measured by the Autolab electrochemical workstation (PGSTAT 302N, METROHM, Swiss) in SBF solution using a three-electrode conguration comprising an Ag/AgCl electrode as the reference electrode, a platinum rod as the counter electrode, and the sample as the working electrode. The measurements were performed at room temperature with a scanning rate of 5 mV s À1 . The corrosion rate was calculated based on the following equation. 24 2.3. Biological test 2.3.1. Cell culture and seeding. Rats bone marrow mesenchymal stem cells (rBMSCS) was isolated from SD rat bone marrow. SD rats were purchased from Guangdong Medical Laboratory Animal Center (Guangzhou, China). Briey, bone marrow was rinsed out with complete culture medium consisting of a-MEM, 10% fetal bovine serum (FBS, Gibico, USA) and 1% antibiotics (100 mg mL À1 gentamycin and 100 U mL À1 penicillin). When reaching 80-90% conuence, cells were digested by 0.25% trypsin, collected by centrifugation and diluted to the desired density in culture medium. 1 mL of cell suspension with a density of 2 Â 10 4 cells per cm 2 was added onto each samples placed in 24-well cell culture plates. The medium was refreshed every two days.
2.3.2. Initial cellular attachment. To evaluate the cell attachment, cells were cultured on the samples for 24 h, and then were xed with 2.5% glutaraldehyde in 30 min. For SEM, cell were washed with PBS twice and then dehydrated in gradient ethanol solution with nal drying by isoamyl acetate. For immunouorescent staining, the xed cells were permeabilized with 0.1% (v/v) Triton X-100 for 7 min, and the unspecic staining was blocked by incubation in 1% BSA for 30 min. For focal adhesion staining, the specimens were incubated overnight at 4 C with 2 mg mL À1 of mouse anti-mouse vinculin primary antibody (Abcam, UK) diluted in 1% BSA/ PBS solution. Aer three washes, the sample was then incubated for 1 h with 2 mg mL À1 of goat anti-mouse IgG H&L (Alexa Fluor® 647) (Abcam, UK). The actin cytoskeleton was labeled by Phalloidin-FITC (Sigma, USA) for 45 min, and the cell nucleus was stained by 4 0 ,6-diamidino-2-phenylindole dihydro-chloride (DAPI, Sigma, USA) for 5 min. The stained cells were observed under the uorescence microscope (OLYMPUS-BX53, Japan).
2.3.3. Cell proliferation. Cell Viability Kit-8 (CCK-8, Beyotime, China) was used to evaluate the cell proliferation on the coated Ti alloys. Aer incubation for 3 and 7 days, culture medium was removed and replaced by 10% CCK-8 working solution, followed by 2 h incubation. The OD values at 450 nm were read by a multi-well plate reader (Thermo Scientic Multiskan GO, Thermo).
2.3.4. Alkaline phosphates (ALP) activity assay. For alkaline phosphatase (ALP) activity evaluation, aer 14 days' incubation, the medium was removed from the cell culture plates and washed with PBS twice, followed by 30 min incubation in 200 mL of 1% Triton X-100 solution containing 100 mM phenylmethanesulfonyl uoride (PMSF). Then 50 mL of the cell lysates was transferred to a 96-well plate and incubated for 2 h at 37 C with 200 mL of p-nitrophenyl phosphate substrate solution (pNPP, Sigma, USA). ALP activity was quantied according to the absorbance at a wavelength of 405 nm. The total protein content in each cell lysate was determined using a BCA protein assay kit, to which ALP activity was normalized.
2.3.5. Cell differentiation. RBMSCs were seeded onto the Ti6Al4V, TiO 2 and the SrO doped TiO 2 coatings at density of 2 Â 10 4 cells per cm 2 and cultured for 7 and 14 days. The total RNA isolation and qPCR were performed strictly following the instruction from the assay kit provider. The forward and reverse primers of the selected genes are listed in Table 1. Detailed information about the experimental procedure can be found in our previous work. 25

Statistical analysis
For statistical analysis, SPSS 17.0 program was used and the data were expressed as mean AE SD. Levene's test was performed to determine the homogeneity of variance for all the data. Tukey HSD post hoc tests were used for the data with homogeneous variance. Tamhane's T2 post hoc was employed in the case that the tested group did not have a homogeneous variance. A pvalue of less than 0.05 was considered signicant.

Morphology and structural characterization
The SEM images of the surface morphology of the TiO 2 -based coatings are showed in Fig. 2. All the coatings show a typical morphology of a plasma sprayed coating, having a rough surface with a surface roughness of 4-6 mm ( Fig. 2A-D). No signicant difference was found among the TiO 2 , 10% SrO-TiO 2 , 20% SrO-TiO 2 and 30% SrO-TiO 2 coatings.  Table 2. The 20% SrO-TiO 2 coating (À573.7 mV) and the TiO 2 coating (À570.34 mV) have a comparable E corr , which is less negative than those of the 10% SrO-TiO 2 (À599.7 mV) coating. The I corr of the 20% SrO-TiO 2 coating (1.92 mA cm À2 ) is slightly lower than that of the 10% SrO-TiO 2 coating (2.06 mA cm À2 ), but obviously higher than that of the TiO 2 coating (2.70 mA cm À2 ). The corrosion rate calculated based on the I corr has the same tendency. Compared to other coatings, the 30% SrO-TiO 2 coating has the most negative E corr (À1048.5 mV) and the highest I corr (7.90 mA cm À2 )/corrosion rate. These results suggest that the 20% SrO-TiO 2 coating has the best corrosion resistance while the 30% SrO-TiO 2 coating has the worst. It was reported that a lower point of zero charge (PZC) of materials exhibits a higher pitting potential and better corrosion resistance. [26][27][28][29] It was found that the PZC of SrTiO 3 is around 8.5-9.5, higher than that of TiO 2 (5-7), 27 which could be one of the possible reasons for the deteriorated corrosion resistance of the 30% SrO-TiO 2 coating. In addition, the corrosion resistance of the coating is also strongly inuenced by the structural defects, such as microcracks and pores etc. Future work will be carried to investigate into the inuence of the amount of Sr on the microstructures of the SrO-TiO 2 coatings. showing the best ability to induce in vitro mineralization. Among the SrO doped TiO 2 coatings, the amount of precipitation formed on the 10% SrO-TiO 2 coating seems the largest, followed by the 20% SrO-TiO 2 coating. Fig. 5 shows XRD results of the coatings soaked in 2Â SBF for 14 days. The characteristic peaks at 25.8 and 31.6 corresponding to hydroxyapatite were observed in the XRD patterns, indicating that the newly formed precipitate was hydroxyapatite. In addition, the relative intensity ratio of HAp to TiO 2 is also found to decrease inversely with the Sr amount, consistent with what we observed from the SEM images (Fig. 4). These results suggest that the addition of SrO compromises the ability of the TiO 2 coating to induce apatite formation. Further work will be carried to illustrate the underlying mechanisms.

Ion release prole in culture medium
The Sr ion release behavior of the TiO 2 , 10% SrO-TiO 2 , 20% SrO-TiO 2 and 30% SrO-TiO 2 coatings were evaluated over 9 days and the release prole is shown in Fig. 6. It can be seen that Sr is releasable in all the doped TiO 2 coatings, and its concentration in the immersion medium increases as the immersion time extended from 3 days to 9 days (Fig. 6A). As shown in Fig. 6B, the Sr concentration at each times for the SrO doped TiO 2 coating is linearly related to the relative amount of SrO in the starting composite powders.   and 30% SrO-TiO 2 coatings for 24 h. Cells on all the coatings get attened and well attach to their underlying substrates. Among these coatings, the cells cultured on the 10% SrO-TiO 2 coating and the 20% SrO-TiO 2 coating, especially the latter one, exhibit distinct and well-dened stress bers and cytoskeleton. Fig. 7B shows the SEM images of the cells cultured on the samples for 24 h. Most of the cells on the Ti, TiO 2 , 10% SrO-TiO 2 and 20% SrO-TiO 2 coatings exhibit a attened polygonal shapes, while those on the 30% SrO-TiO 2 coatings are less attened.

Cell viability and proliferation
Cell proliferation results are shown in Fig. 8A. At day 3, cells cultured on the 20% SrO-TiO 2 coatings show higher proliferation rates than those cultured on the TiO 2 , 10% SrO-TiO 2 and 30% SrO-TiO 2 coating. Aer culturing for 7 days, the proliferation rate of the cells on the 20% SrO-TiO 2 coating is still the highest, followed by that on the 10% SrO-TiO 2 coating, while the cells on the 30% SrO-TiO 2 coating exhibit the lowest proliferation rate. The average proliferation rate for the cells on the TiO 2 coating is higher than that on the Ti6Al4V, but without signicant statistic difference. These results indicate that Sr doping has great effects on the cell proliferation on the coatings and the 20% SrO-TiO 2 coating has the best positive effect.

Alkaline phosphatase (ALP) activity
Fig . 8B shows the ALP activity of the cells cultured on the Ti6Al4V, the TiO 2 , 10% SrO-TiO 2 , 20% SrO-TiO 2 , 30% SrO-TiO 2 coatings at day 14. It can be seen that the ALP activity of the cells cultured on the 10% SrO-TiO 2 coating and the 20% SrO-TiO 2 coating is comparable, which is signicantly higher than those for the cells on the Ti6Al4V and the TiO 2 coating. The ALP activity of the cells on the 30% SrO-TiO 2 coating is lowest. Fig. 9 displays the expression of the osteogenic-related genes by the cells cultured on the Ti6Al4V, the TiO 2 coating, and the SrOdoped TiO 2 coatings. Among these genes, RUNX-2 and COL-I are early markers, while OCN and OPN are later markers for osteogenic differentiation. As shown in Fig. 9, aer 7 days, the COL-I expression levels by the cells cultured on the TiO 2 coating, the 10% SrO-TiO 2 coating and the 20% SrO-TiO 2 coating are signicantly higher than those on the Ti6Al4V and the 30% SrO-TiO 2 coating. At day 14, the COL-I expression level on the 20% SrO-TiO 2 coating become the highest, which nearly doubles those for the other coatings. For OCN, no difference is found for all the samples at day 7. However, the cells cultured  This journal is © The Royal Society of Chemistry 2018 on the 20% SrO-TiO 2 coating display the highest OCN expression level at day 14 and no signicant difference can be seen between the 10% SrO-TiO 2 coating and the 30% SrO-TiO 2 coating. For OPN, its levels expressed by the cells on the 10% SrO-TiO 2 coating and the 20% SrO-TiO 2 coating are comparable and signicantly higher than those for the other coatings at day 7. Aer culturing for 14 days, no signicant difference can be found anymore between the 10% SrO-TiO 2 coating and 20% SrO-TiO 2 coating, but their expression levels are higher than the control groups. For RUNX-2, at both day 7 and 14, its expression levels by the cells cultured on the 10% SrO-TiO 2 coating and the 20% SrO-TiO 2 coating are comparable and higher than those by the cells on the control groups. Again, the cells cultured on the 30% SrO-TiO 2 coating show the lowest Fig. 6 Concentrations of Sr ions released from TiO 2 , 10% SrO-TiO 2 , 20% SrO-TiO 2 , 30% SrO-TiO 2 coatings after immersion in culture medium for 3, 6 and 9 days. expression level. In general, it can be concluded from these results that the 20% SrO-TiO 2 coating are superior to the others in promoting osteogenic differentiation, while the 30% SrO-TiO 2 coating showed obviously less osteogenic activity compared to the other doped coatings.

Discussion
In this study, we developed SrO-TiO 2 coatings to enhance the osseointegration of the metallic orthopaedic implant. Compared to the commonly used biodegradable Sr-containing biomedical coating, 30-34 the combination of bioactive Sr and non-degradable TiO 2 in this study endowed the coating with an ability to release bioactive Sr (Fig. 6) and thus obviously improving the osteogenic activity (Fig. 9). It should be noted that the release of Sr is not combined with the release of Ti, indicating that the release of Sr in the developed coating system is selective, which will benet the long-term stability of the implant.
In the SrO-doped TiO 2 coatings, the Sr exists in two different congurations: Sr intercalated in TiO 2 lattice and SrTiO 3 . In the 10% SrO-TiO 2 coating, the amount of SrTiO 3 calculated from XRD patterns is around 6.31 wt%. The atomic ratio of the Sr in the SrTiO 3 relative to the total amount of Ti in this coating is 0.017, much less than the theoretic value of 0.052, suggesting that there must be some part of Sr (0.035) existing in other congurations. Based on the solid solution and doping theory, the most possible conguration of Sr is as interstitial atoms of the TiO 2 lattice in SrO-TiO 2 solid solution (Fig. 1). When the SrO amount increases to 20%, the percent of SrTiO 3 is around Fig. 8 The proliferation and ALP activity and of rBMSCs cultured on the Ti6Al4V, TiO 2 , 10% SrO-TiO 2 , 20% SrO-TiO 2 , 30% SrO-TiO 2 coatings at day 14. Fig. 9 Quantitative PCR analysis of the cells cultured on TiO 2 , 10% SrO-TiO 2 , 20% SrO-TiO 2 , 30% SrO-TiO 2 coatings for 7 and 14 days, housekeeping gene GAPDH was used as an internal control. *Statistically significant difference among different samples (p < 0.05).
19.7 wt%, leading to a Sr/Ti of 0.061, which is nearly half of the theoretic value (0.115). Therefore, there are also a large part of Sr existing as an interstitial solute in the SrO-TiO 2 solid solution. 35,36 However, in the 30% SrO-TiO 2 coating, the ratio of Sr in SrTiO 3 relative to the total Ti in the coating (0.193) is very close to the theoretic value (0.198), therefore, it can been concluded that Sr in this coating exists only in the form of SrTiO 3 . Therefore, the ion release of Sr from the SrO-doped TiO 2 coating is contributed by two different Sr congurations. For the 10% SrO-TiO 2 coating and the 20% SrO-TiO 2 coating, the released Sr is from both SrTiO 3 and the Ti y Sr 2À2y O 2 solid solution, whereas that for the 30% SrO-TiO 2 coating is solely from SrTiO 3 . As shown in Fig. 6, the linear release behavior was achieved for all the coatings by the two Sr congurations, suggesting that the Sr release could be precisely controlled by adjusting the amount of Sr and the Sr congurations in the coating. However, it should be noted that there are numerous nanosized crystals precipitated on the surface of the 30% SrO-TiO 2 coating aer incubation with cells for 24 h, as shown in Fig. 10A. Based on the EDS results, the Sr/Ti ratio of the newly formed crystals is around 0.45, nearly doubles the theoretical value of 0.19, implying that the newly formed crystals is a Sr-rich compound. Based on this, it can be deduced that the Sr released from the 30% SrO-TiO 2 coating possibly led to an increased supersaturation degree with respective to a certain of Srcontaining compound, which "recycles" the released Sr ions, and ultimately decreased the ion concentration in their extracts. 37 Therefore, the linear release behavior for the 30% SrO-TiO 2 with respective to the SrO amount is also contributed by the newly formed layer of the crystals.
The TiO 2 powders we used in this study are composed of 85% anatase and 15% rutile TiO 2 . However, it can be seen that the relative amount of the anatase TiO 2 was signicantly reduced aer plasma spaying, indicating that plasma spraying promotes the phase transformation from anatase to rutile (Fig. 10B). With the increase in the amount of Sr in the coating, the relative amount of the anatase TiO 2 increases, applying that the Sr incorporation can suppress the anatase to rutile transformation. It is well-known that rutile and anatase are two most common phases for TiO 2 . 38 V. Sollazzo reported that anatase TiO 2 has better bioactivity compared to the rutile phase. 39 Although no direct evidence is obtained in this study to proof the contribution of the anatase TiO 2 on the bioactivity, the rutile to anatase transformation caused by Sr doping still could be a merit of our coating design, which might provide some guidance for further biomedical coating design based on the TiO 2 material. Based on the biological results, we found that the 20% SrO-TiO 2 coating not only promoted the proliferation of rBMSCs (Fig. 8A), but also enhanced their osteogenic differentiation, as indicated in the ALP activity (Fig. 8B) and gene expression levels ( Fig. 9), which is closely related to the Sr ions released from the coating. To illustrate the dominant effects of the Sr ion on the enhancement of the osteogenic activity, we used the extract of the samples to culture the rBMSCs to evaluate the biological effects of the dissolution products. The expression levels of bone-related genes (OPN, RUNX-2, COL-I and OCN) by the cells cultured for 7 days in the extracts were presented in Fig. 10C. It can be seen that the cells cultured by the extract of the 20% SrO-TiO 2 coating express the highest levels of these genes, whereas those cultured in the extract of the 30% SrO-TiO 2 coating express the lowest levels. The consistency of the gene expression results for the cells directly cultured on the coating surface and incubated with the coating extracts further validates the importance of Sr ions for the osteogenic activity. Similar to the biological drugs, the effects of the bioactive ions are also does-dependent. 10,40,41 Therefore, based on the biological results, we may conclude that the Sr ions released from the 20% SrO-TiO 2 coating is the optimal and the Sr ions from the 30% SrO-TiO 2 coating are possibly overdosed.

Conclusion
In summary, non-degradable bioactive coating was designed based a SrO-TiO 2 system in this work, which is able to selectively release Sr ions. It was proven that the release of Sr ions is dependent on the Sr congurations in the coating, which has great inuence on the osteogenic activity of the bone cells. Sr mainly exists as interstitial atoms in a solid solution of Ti y -Sr 2À2y O 2 when its amount is less than 10 wt%, and only SrTiO 3 conguration appears when the Sr amount is higher than 30%. Both congurations present in the coatings with a Sr amount lying in between. We found that the addition of Sr compromises the in vitro mineralization ability of the TiO 2 coating, inversely proportional to its amount. However, its benecial effects on osteogenesis enhancement are prominent. It was proved that the 20% SrO-TiO 2 coating shows the best capacity of enhancing cellular proliferation and osteogenic differentiation, pointing out its potential application as an orthopaedic implant coating.

Conflicts of interest
There are no conicts to declare.