Biphasic Osteogenic Characteristics of HumanMesenchymal Stem Cells Cultured on TiO 2 Nanotubes of Different Diameters

We cultured human mesenchymal stem cells (hMSCs) on TiO2 nanotubes with diameters of 30–100 nm to assess the size-effect of TiO2 nanotubes on the behavior and osteogenic functionality of hMSCs. Most studies of the expression of genes encoding alkaline phosphatase (ALP), osteocalcin (OCN), osteopontin (OPN), and integrin-β (INT-B), after 1 week of incubation, supported the results of cell viability and MTT assays at 48 hrs of plating. However, after 2 weeks of incubation, expression of ALP, OPN, INT-B, and protein kinase R-like ER kinase (PERK) genes were significantly higher in cells cultured on 70 nm TiO2 nanotubes than that in cells cultured on other TiO2 nanotubes and Ti. This biphasic osteogenic characteristic of hMSCs is supposed to relating to the nature of the hMSCs adhering to the substrate at the beginning of incubation, and the nanostructural stimulation caused by the topography of TiO2 nanotubes at a later stage of incubation. The discovery of these biphasic characteristics of hMSCs cultured on different-sized TiO2 nanotubes may contribute to resolving the discrepant results relating to the size-effect of TiO2 nanotubes on the adhesion, proliferation, and functionality of cells.

However, there have been some inconsistencies in the literature regarding the optimal size of TiO 2 nanotubes for eliciting maximal adhesion, proliferation, and functionality of cells [5,6,15,[20][21][22]. Furthermore, several researchers have reported that TiO 2 nanotubular structure inhibited the behavior and functionality of vascular smooth muscle cells, compared to the effect of pure Ti, regardless of the size of the nanotubes [23].Therefore, identifying factors besides the size of the TiO 2 nanotubes that play a critical role in determining the adhesion, proliferation, and functionality of cells is crucial for resolving the incongruity regarding the size effect of TiO 2 nanotubes on cell behavior and functionality.
We have previously shown that different sterilization methods (wet autoclave versus dry autoclave) of TiO 2 nanotubes induced different osteogenic differentiation of human mesenchymal stem cells, due to the residual air trapped in the nanotubes [24].Similarly, Zhao et al. compared several types of sterilization methods, such as autoclaving, ultraviolet (UV) radiation, and ethanol sterilization, to assess the cytocompatibility of TiO 2 nanotubes sterilized by each method and reported that UV sterilization resulted in the best cell functionality, including adhesion and proliferation [25].However, because many experimental factors (e.g., electrolyte solution, anodization voltage, anodization time, temperature, and specimen shape) are involved in preparation of TiO 2 nanotube arrays on a Ti surface, there are many variables that need to be considered in such in vitro and in vivo experiments.
Surface roughness is known to be a key factor determining the morphogenic effects of implant materials on adhesion, proliferation, and functionality of various types of cells [26][27][28][29][30], and most researchers have focused on the effect of nanotopography and morphology on the behavior and functionality of cells [31][32][33][34][35][36].TiO 2 nanotubes manufactured on commercial Ti are mainly anodized by electrolyte solution.Commercial Ti plate is prepared by a machining and rolling process, which can lead to the formation of microand macroscale roughness at the surface of Ti, which can, in turn, distort the nanostructural features in the absence of further treatment.On the other hand, electropolishing of Ti leads to reduced micro-and nanoroughness (<5 nm) and a uniform flat surface [37].TiO 2 nanotube arrays fabricated on electropolished Ti indeed have true nanotubular structure and can more precisely indicate the effect of TiO 2 nanotubes on cell behavior and functionality.
In this study, we prepared TiO 2 nanotubes with diameters of 30, 50, 70, and 100 nm on electropolished Ti, thereby minimizing factors that may interfere with nanotubular features.We used these to assess the size effect of TiO 2 nanotubes on the behavior and osteogenic functionality of hMSCs, as measured by cell viability, MTT(3-(4,5dimethylthiazole-2-yl)-2,5-diphenyl tetrazolium bromide) assay, real-time PCR assay, and alizarin red assay.

Materials and Methods
2.1.TiO 2 Nanotubes Fabrication.Commercial Ti sheet (0.2 mm thick, 99.5%, Hyundai Ti, South Korea) was electropolished by electrochemical etching technique as published previously [37].As shown Figure 1(a), electropolished Ti showed mirror surface.Electropolished Ti sheet was cleaned by acetone and D.I. (deionized) water.The nanotubes were prepared in 0.5 w/v % hydrofluoric acid (Merck, 48 w/v %, NJ, USA) in water with acetic acid (JT Baker, 98 w/v %, NJ, USA; Volumetric Ratio = 7 : 1) at 5, 10, 15, and 20 V for 20 min.A platinum electrode (DSM, 99.99%,South Korea) served as the counterpart.The samples were then rinsed with D.I. water, dried at 60 • C for 24 hrs, and heat-treated at 500 • C for 2 hrs to crystallize amorphous TiO 2 nanotubes into anatase structure.The morphology of TiO 2 nanotube arrays was observed by field emission scanning electron microscope (FE-SEM; S4800, Hitachi, Japan).The crystalline phases of TiO 2 nanotube arrays were determined by X-ray diffractometer (XRD; X'Pert PRO MRD, PANalytical B.V., Netherlands) with Ni-filtered Cu-Kα ray.The glancing angle of the specimen was fixed at 5 • against the incident beam enabling the detection of XRD patterns to be the depth of less than 5 μm from the top surface of the substrate.
Anatase TiO 2 nanotube specimens were cut into 1.27 * 1.27 cm 2 and sterilized by wet autoclaving method as previously reported [24].And then, wet-autoclaved samples were sterilized by using ultraviolet (UV) radiation.An electropolished Ti sheet cut into identical sized pieces was used as a control and was autoclaved by the same methods.

hMSCs Cell Culture.
Human mesenchymal stem cells (hMSCs) were obtained from Lonza Corporation (Poietics hMSCs, Switzerland).The cell growth media were composed of α-MEM (Invitrogen, CA, USA), 10% Fetal Bovine Serum (FBS) (Invitrogen), 1% penicillin-streptomycin (Invitrogen).The cell suspension was plated in a cell culture dish and incubated under 37 • C, 5% CO 2 environment.All experiments of hMSCs were conducted with cultures at passage 4-5.When hMSCs showed high confluence, they were seeded onto TiO 2 nanotube experimental substrate placed on a 12well plate with the density of 25,000 cells/well and stored in a CO 2 incubator for each incubation time.Also, osteogenic induction media (cell growth media including 10 mM βglycerol phosphate (Sigma Co., MO, USA), 150 μg/mL ascorbic acid (Sigma), and 10 nM dexamethasone (Sigma) were added to promote the osteogenic differentiation of hMSCs after 3 days of incubation.Osteogenic induction media were changed every 2-3 days.

Cell Viability
Test.Fluorescein diacetate (FDA; Sigma, MO, USA) staining technique was performed to count viable hMSCs adhered to the specimen.At 2, 24, and 48 hrs after plating, hMSCs on the substrates were rinsed with phosphate-buffered saline solution (PBS) (Invitrogen, CA, USA) and incubated with FDA working solution (50 μg FDA dissolved in 10 mL PBS solution) for 30 seconds and then washed three times by PBS solution.The washed specimens were pictured by an inverted fluorescence microscope (CKX41, Olympus Co., Japan).We tested 2 samples per experimental group for cell adhesion and proliferation test and counted the number of cells adhered at 4 corners of a specimen and a center of a specimen.

MTT Assay.
After the selected incubation periods, the samples were washed by PBS solution and transferred to a new 12-well plate. 1 mL of MTT dye agent (Sigma) was added to each well.After 3 hours of incubation in 5% CO 2 incubator, 1 mL of isopropanol was added to each well and 12-well plate was shaken for 30 minutes.The absorbance of each solution was measured at the wavelength of 570 nm by microplate ELISA reader (Spectra Max 250, Thermo Electron Co., USA).

2.5.
Real-Time PCR Assay.After 1 and 2 weeks of culture, total RNA of the cells cultured on the experimental substrates were extracted by TRI agent (Invitrogen) and reversetranscribed into cDNA by cDNA Synthesis Kit (SuperScript VILO; Invitrogen).Real-time PCR was performed by SYBR Green Dye-Based Gene Expression Assays (Invitrogen), and the information of PCR primer used in this study is listed in Table 1.Real-time PCR was carried out by using StepOne Real-Time PCR System (Applied Biosystems, USA).cDNA samples (1 μL for total volume of 20 μL) were analyzed  for gene of interest and for house-keeping gene GAPDH.Cycle-threshold point was compared to quantify the gene expression level of each sample.All levels of expression of experimental group were normalized by the level of expression of control group (hMSCs cultured on pure Ti).
2.6.Alizarin Red Assay.After 3 weeks of incubation, the samples were rinsed with PBS solution and fixed with 4% paraformaldehyde solution (Sigma) at room temperature for 20 min.After fixation, the samples were stained with alizarin red solution (Sigma) for 20 min and then washed twice with D.I. water.The stained sample was photographed by inverted microscope (CKX 41, Olympus, Japan) to observe the formation of bone nodule.Then, the sample was reacted with 10% acetic acid (JT Baker, USA) for 30 min and stored at 85 • C for 10 min.After heating, solution was collected from the sample and centrifuged at 11,000 rpm for 15 min.Supernatant was collected from the solution after centrifugation.Same volume of 10% ammonium hydroxide (Sigma) was added to the solution in order to neutralize the supernatant.The absorbance of each solution was measured at the wavelength of 405 nm by microplate ELISA reader (Spectra Max 250, Thermo Electron Co., USA).x The data of control group (Ti) was significantly higher than that of experimental (TiO 2 nanotubes) group (P < 0.05).

Data Analysis.
In terms of cell adhesion, MTT, and alizarin red assay, all data were expressed as mean ± standard deviation and analyzed statistically by one-way ANOVA (SPSS 12.0, SPSS GmbH, Germany) and Duncan's multiple range test as a post hoc test.Significant difference was determined at P values at least less than 0.05.-B) after 1 week and 2 weeks of incubation time.The graph shows the average ± standard deviation bars.* The data of experimental group was significantly higher than that of control group (P < 0.05).X The data of control group was significantly higher than that of experimental group (P < 0.05).* * The data was significantly the highest among all groups (P < 0.05).ordered nanotubes with 4 different pore sizes, between 30 and 100 nm, created by controlling the anodization voltage across a range from 5 to 20 V, as previously reported [15].

Results and Discussion
Figure 2 indicates X-ray diffraction (XRD) patterns of heat-treated 30 and 100 nm TiO 2 nanotubes.As shown XRD results, anatase TiO 2 and Ti crystalline phases were mainly detected.Also, there was no dramatic difference of crystallinity between 30 and 100 nm TiO 2 nanotubes after heat treatment.Therefore, we did not consider the effect of nanotube crystallinity on the cell behavior and osteogenic functionality of hMSCs in this research.
Figure 3 represents images of live hMSCs stained with FDA solution and shows the number of cells adhered to the substrate.As shown in Figure 3(a), hMSCs had not yet fully spread after 2 hrs of incubation but took on an elongated appearance after 24 hrs of incubation.In Figure 3(b), it can be seen that the cell number increased over time in both the experimental and control groups; there was no significant difference between any of the TiO 2 nanotube groups after 2 hrs or after 24 hrs of incubation (P > 0.05).
However, by 48 hrs after plating, the number of cells adhering to 30 nm TiO 2 nanotubes was significantly higher than that adhering to 50, 70, and 100 nm TiO 2 nanotubes (P < 0.05; Figure 3(b)); this data showed trends similar to those of previously published data.However, in previous studies, the number of cells adhering to pure Ti within 48 hrs of incubation was significantly lower than that adhering to any of the TiO 2 nanotubes [15,19].In contrast, in this study, we found that the number of adherent cells cultured on electropolished Ti was significantly higher than that of cells cultured on 50, 70, and 100 nm TiO 2 nanotubes (Figure 3(b)).
Figure 4 shows the results of the MTT assay for assessing the proliferative ability of hMSCs.The data indicated a similar trend to that of the FDA-staining test.Therefore, the results of both the cell viability test and the MTT assay indicated clearly that electropolishing of Ti promotes the adhesion and proliferation of hMSCs.
In order to assess the osteogenic functionality of the cells, we performed real-time PCR analysis for alkaline phosphatase (ALP), osteocalcin (OCN), osteopontin (OPN), and integrin-β1 (INT-B) expression after 1 and 2 weeks of incubation (Figure 5).After 1 week of incubation, expression of OCN, OPN, and INT-B genes was not significantly different between the TiO 2 nanotubes of 30, 50, and 70 nm diameter and electropolished Ti; however, the expression of these genes was significantly different between these and the 100 nm TiO 2 nanotubes.Thus, the 1-week incubation realtime PCR results were mostly similar to the results of the cell viability test and the MTT assay.However, ALP gene expression in cells cultured on 30 and 50 nm TiO 2 nanotubes was significantly higher than that for cells cultured on pure Ti (P < 0.05; Figure 5).
Two weeks after plating, ALP, OPN, and INT-B gene expression was highest in cells cultured on 70 nm TiO 2 nanotubes (P < 0.05).In additions, OCN gene expression The data of experimental group was significantly higher than that of control group (P < 0.05).
was significantly higher in cells grown on 30 and 70 nm TiO 2 nanotubes than in cells cultured on other TiO 2 nanotubes and Ti (Figure 5).Typically, after incubation commences, osteogenic characteristics should be manifested only by hMSCs adhering to the substrate.However, the results of real-time PCR analysis indicated that, after the early stages of incubation, hMSCs reacted to the nanostructural stimulation of the TiO 2 nanotubes and altered their osteogenic functionality according to the size of TiO 2 nanotubes.
We also investigated expression of protein kinase Rlike ER kinase (PERK) and activating transcription factor 6 (ATF6), which are known to promote osteoblast differentiation of hMSCs, by real-time PCR (Figure 6).PERK gene expression was slightly, although significantly, higher in cells cultured on 70 nm TiO 2 nanotubes than in those grown on other TiO 2 nanotubes or Ti.Additionally, the expression of ATF6 was significantly higher in cells grown on any of the TiO 2 nanotubes than in cells grown on Ti (Figure 6).
Many studies have reported that bone morphogenetic protein 2 (BMP2) activates unfolded protein response (UPR) transducers, such as PERK, ATF6, and OASIS, as BMP2 treatment dramatically promotes the expression of ATF6, with a concomitant increase in ALP and OCN expression [38][39][40][41].Even though we did not use BMP2 to enhance osteogenic differentiation of hMSCs in this study, the results of PERK and ATF6 gene expression lead us to speculate that PERK and ATF6 regulate ALP, OCN, and OPN transcription.In future studies, we intend to resolve the relationship between the promotion of osteogenic differentiation in hMSCs by UPR transducers and physical-mechanical stimuli, such as those arising from the nanotopography of TiO 2 nanotubes.
The result of the alizarin red assay, as shown in Figure 7, indicated that the calcific deposition in cells grown on 50, 70, and 100 nm TiO 2 nanotubes was significantly higher than that in cells grown on 30 nm TiO 2 nanotubes and Ti.These results supported those of the real-time PCR assays showing that cells grown on 70 nm TiO 2 nanotube have higher ALP, OCN, PERK, and ATF6 gene expression than cells grown on other TiO 2 nanotubes.In terms of showing high value of 100 nm TiO 2 nanotubes after 3 weeks of incubation, we are undergoing to resolve this phenomenon.
Therefore, we can surmise that the biphasic osteogenic characteristics of hMSCs might arise from the interplay of the osteogenic functionality of hMSCs that adhere to the substrate in the early stage of incubation, as well as that caused by nanostructural stimulation that varies according to the size of TiO 2 nanotubes.In order to more thoroughly investigate the effects of geometrical or topographical parameters of TiO 2 nanotubes on cell behavior as a function of incubation time, external factors, such as cell-ECM interaction, cellcell communication, and physicochemical and mechanical stimuli, should be carefully considered.

Summary
The biphasic osteogenic characteristics of hMSCs cultured on different sizes of TiO 2 nanotubes have been investigated here.We found that, after 48 hrs of incubation, hMSCs cultured on Ti and 30 nm diameter TiO 2 nanotubes showed higher adhesion and proliferation than those cultured on 50, 70, and 100 nm diameter TiO 2 nanotubes.Most of the gene expression studies performed after 1 week of incubation supported results from cell viability and MTT assays at 48 hrs of plating.However, after 2 weeks of incubation, ALP, OPM, INT-B, and PERK gene expression was higher in cells cultured on 70 nm TiO 2 nanotubes than in cells cultured on other TiO 2 nanotubes or Ti.We propose that  this was due to the interplay of osteogenic characteristics of hMSCs adhered to the substrate at the beginning of incubation and those caused by nanostructural stimulation in response to the stimulation provided by the different-sized TiO 2 nanotubes at a later stage of incubation.We conclude that these biphasic features of hMSCs cultured on different sizes of TiO 2 nanotubes may help resolve conflicts relating to the effect of the size of TiO 2 nanotubes on the adhesion, proliferation, and functionality of cells.

Figure 1 :
Figure 1: (a) The photograph of commercial Ti (a) and electropolished Ti (b), (b) SEM micrographs of heat-treated TiO 2 nanotubes with different diameters.The images show highly ordered nanotubes with four different pore sizes between 30 and 100 nm.

Figure 1 (
Figure 1(a) shows the difference in appearance between commercial Ti and electropolished Ti surfaces, whereas Figure 1(b) indicates the scanning electron microscopy (SEM) images of heat-treated 30, 50, 70, and 100 nm TiO 2 nanotubes, respectively.The micrographs show highly

Figure 5 :
Figure5: Quantitative PCR analysis for alkaline phosphatase (ALP), osteocalcin (OCN), osteopontin (OPN), and integrin-beta (INT-B) after 1 week and 2 weeks of incubation time.The graph shows the average ± standard deviation bars.* The data of experimental group was significantly higher than that of control group (P < 0.05).X The data of control group was significantly higher than that of experimental group (P < 0.05).* * The data was significantly the highest among all groups (P < 0.05).

Figure 6 :
Figure6: Quantitative PCR analysis for protein kinase R-like ER kinase (PERK) and activating transcription factor 6 (ATF6) after 2 weeks of incubation time.The graph shows the average ± standard deviation bars.* The data of experimental group was significantly higher than that of control group (P < 0.05).

Figure 7 :
Figure 7: Alizarin red assay of hMSCs cultured on flat Ti and various sized TiO 2 nanotubes after 3 weeks of incubation.* The data of experimental group was significantly higher than that of control group (P < 0.05).