Nanostructured TiO2 Surfaces Promote Human Bone Marrow Mesenchymal Stem Cells Differentiation to Osteoblasts

Micro- and nano-patterning/modification are emerging strategies to improve surfaces properties that may influence critically cells adherence and differentiation. Aim of this work was to study the in vitro biological reactivity of human bone marrow mesenchymal stem cells (hBMSCs) to a nanostructured titanium dioxide (TiO2) surface in comparison to a coverglass (Glass) in two different culture conditions: with (osteogenic medium (OM)) and without (proliferative medium (PM)) osteogenic factors. To evaluate cell adhesion, hBMSCs phosphorylated focal adhesion kinase (pFAK) foci were analyzed by confocal laser scanning microscopy (CLSM) at 24 h: the TiO2 surface showed a higher number of pFAK foci with respect to Glass. The hBMSCs differentiation to osteoblasts was evaluated in both PM and OM culture conditions by enzyme-linked immunosorbent assay (ELISA), CLSM and real-time quantitative reverse transcription PCR (qRT-PCR) at 28 days. In comparison with Glass, TiO2 surface in combination with OM conditions increased the content of extracellular bone proteins, calcium deposition and alkaline phosphatase activity. The qRT-PCR analysis revealed, both in PM and OM, that TiO2 surface increased at seven and 28 days the expression of osteogenic genes. All together, these results demonstrate the capability of TiO2 nanostructured surface to promote hBMSCs osteoblast differentiation and its potentiality in biomedical applications.


Introduction
Bone tissue engineering is a strategy to replace autologous or heterologous bone grafts with an artificial material (scaffold) that mimics the bone structure [1][2][3]. A bone graft substitute may follow a series of indication and should share a higher number of properties, such as a good material that allows a structural framework for bone growth and the good stem cell to produce new bone [4,5], good cell attachment and the absence of immune reactions. The scaffold is asked to attract osteoprecursor cells, such as bone marrow stem cells (BMSCs), and to induce them to the osteoblastic phenotype (osteoinduction). The subsequent step is the production of new bone by the newly formed osteoblasts [6,7]. To properly allow this last event, materials should be resorbable. The resorption induced by osteoclasts may be counteracted by the production of new bone. This gradual process takes a duration in the order of weeks/months and should ideally lead to the complete dissolution of the scaffold. An ideal scaffold should also allow the development of a vascularized network to provide the proper irroration of the regenerated tissue [8]. In addition, it should consent to cell adhesion and promote cell growth and differentiation. It should facilitate extracellular matrix regeneration and cell spatial distribution through the scaffold to mimic the physiological bone architecture [9].
As regards to these aspects, surface topography and chemistry take significant influences on the biological performance of biomedical scaffolds [10]. In this context, titanium dioxide (TiO 2 ) represents a gold standard in bone tissue engineering, in fact the nanotopography of TiO 2 raised several important properties, which allow the use of this scaffold to mimic the morphology and the hierarchical organization typical of the extra cellular matrix (ECMs) in bone [11,12]. TiO 2 nanostructured surface was obtained by the deposition of a supersonic beam of TiO x clusters. The typical features of this scaffold showed dimensions in the range 1 to 100 nm. In previous studies, the characteristic of this nanostructured surface, such as the porosity and the nanotopography were analyzed and, in addition, a variety of chemical groups and immobilized functional peptides to functionalize the surface for the improvement of cell attachment and proliferation were evaluated [13][14][15].
Bone marrow stromal cells, for more than twenty years, have represented a good source of osteoblast precursor cells [16]. Recent studies have demonstrated the differentiation potential of human BMSC. Under appropriate culture conditions, these human stem cells can differentiate into ligament, tendon [17], muscle [18,19], nerve [20,21], endothelium [22] or hepatic tissue [23]. Moreover, human bone marrow mesenchymal stem cells (hBMSCs) not only contribute structurally to tissue repair but also possess strong immunomodulatory and anti-inflammatory properties that may influence tissue repair by modulation of local environment.
In this study, we evaluated the biocompatibility of Titanium dioxide nanostructured clusters deposited on a coverglass surface (Tethis ® company, Milan, Italy), with respect to a microscopy coverglass (Glass). We performed a detailed investigation, in terms of adhesion, proliferation and differentiation towards bone phenotype of human multipotent stem cells on nanostructured TiO 2 and Glass surfaces. Furthermore, to comprehend the influence of surface nanotopography on hBMSCs adherence and differentiation, the cells were cultivated in the presence (osteogenic medium (OM)) or absence (proliferative medium (PM)) of osteogenic factors. Considering the clinical applications of TiO 2 nanostructured surface in nanomedicine and bone tissue engineering, the main objective of this manuscript was to elucidate the biological mechanisms of the interaction cell-biomaterial surface, in order to improve the use of surface nanotopography for bone grafts.

Morphological Evaluation of Nanostructured TiO 2 Surface
Titanium dioxide surface used in this study was realized by the deposition of a supersonic beam of TiO x clusters [13]. The surface of the Glass (used as control) and of nanostructured TiO 2 were different at SEM (Figure 1): a uniform and particulate structure of the clusters, with diameter under 100 nm of dimension was observed for the TiO 2 surface ( Figure 1D-F) but not for the Glass (Figure 1A-C). On the TiO 2 , it is possible to see the typical nanoclusters that begin to be distinguishable at high magnification 50,000ˆand more evident at 100,000ˆ( Figure 1E,F). Further chemical characterizations were previously reported [13].

Cell Attachment and Cytoskeleton Morphology
Cell attachment and morphology at short (24 h) and long (seven days) time incubation were properly analyzed ( Figure 2). To evaluate cell attachment, hBMSCs were seeded on the different surfaces (Glass and TiO2), cultured for 24 h, then fixed and stained with anti-p-FAK (Y397, green fluorescence).

Cell Attachment and Cytoskeleton Morphology
Cell attachment and morphology at short (24 h) and long (seven days) time incubation were properly analyzed ( Figure 2). To evaluate cell attachment, hBMSCs were seeded on the different surfaces (Glass and TiO 2 ), cultured for 24 h, then fixed and stained with anti-p-FAK (Y397, green fluorescence). high magnification 50,000× and more evident at 100,000× ( Figure 1E,F). Further chemical characterizations were previously reported [13].

Cell Attachment and Cytoskeleton Morphology
Cell attachment and morphology at short (24 h) and long (seven days) time incubation were properly analyzed ( Figure 2). To evaluate cell attachment, hBMSCs were seeded on the different surfaces (Glass and TiO2), cultured for 24 h, then fixed and stained with anti-p-FAK (Y397, green fluorescence).   In Figure 2, CLSM images of focal adhesion of cells cultured on Glass ( Figure 2A) and TiO 2 ( Figure 2B) are shown representative. From these images, we did not note any visible morphological alterations between cells adherent to Glass and TiO 2 nanostructured surface. Furthermore, to evaluate possible surface influences, the area of p-FAK positive foci was normalized to cells area: TiO 2 surface showed an increment of about 5% in foci-positive area ( Figure 2C) and the increment was statistical significant (p < 0.05). Finally, in order to evaluate the effect of both surfaces on cytoskeleton morphology, cells were observed with CLSM at 24 h and seven days of culture. At 24 h in proliferative conditions, hBMSCs were observed through immunofluorescence of actin and tubulin filaments, as shown in Figure 2D-G. No specific differences in morphology of adherent cells to the different surfaces were observed at 20ˆand 40ˆmagnification. At seven days, hBMSCs were observed in both PM ( Figure 3A,C) and OM ( Figure 3B,D) culture conditions. In particular, on Day 7 in proliferative conditions, the cytoskeleton morphology of the seeded cells showed some slight differences. As observed in Figure 3A,C, hBMSC cells cultivated on TiO 2 showed lamellipodia more prolonged and often divided in two or more branches if compared to Glass surface. This difference was again observed more manifest on TiO 2 if compared to Glass even if at lower extent in OM culture conditions ( Figure 3C,D). The overall shape of the cells remained fibroblast-like. An increment in alpha-tubulin signal was also observed in hBMSCs seeded on TiO 2 nanostructured surface with respect to Glass ( Figure 3D). Notably, it was demonstrated that alpha-tubulin decreased in osteogenic conditions, especially in cells seeded on TiO 2 surface, with a concomitant increase of actin. In Figure 2, CLSM images of focal adhesion of cells cultured on Glass ( Figure 2A) and TiO2 ( Figure 2B) are shown representative. From these images, we did not note any visible morphological alterations between cells adherent to Glass and TiO2 nanostructured surface. Furthermore, to evaluate possible surface influences, the area of p-FAK positive foci was normalized to cells area: TiO2 surface showed an increment of about 5% in foci-positive area ( Figure 2C) and the increment was statistical significant (p < 0.05). Finally, in order to evaluate the effect of both surfaces on cytoskeleton morphology, cells were observed with CLSM at 24 h and seven days of culture. At 24 h in proliferative conditions, hBMSCs were observed through immunofluorescence of actin and tubulin filaments, as shown in Figure 2D-G. No specific differences in morphology of adherent cells to the different surfaces were observed at 20× and 40× magnification. At seven days, hBMSCs were observed in both PM ( Figure 3A,C) and OM ( Figure 3B,D) culture conditions. In particular, on Day 7 in proliferative conditions, the cytoskeleton morphology of the seeded cells showed some slight differences. As observed in Figure 3A,C, hBMSC cells cultivated on TiO2 showed lamellipodia more prolonged and often divided in two or more branches if compared to Glass surface. This difference was again observed more manifest on TiO2 if compared to Glass even if at lower extent in OM culture conditions ( Figure 3C,D). The overall shape of the cells remained fibroblast-like. An increment in alpha-tubulin signal was also observed in hBMSCs seeded on TiO2 nanostructured surface with respect to Glass ( Figure 3D). Notably, it was demonstrated that alpha-tubulin decreased in osteogenic conditions, especially in cells seeded on TiO2 surface, with a concomitant increase of actin.

Effects of the Nanostructured TiO2 Surface on Proliferation and Differentiation of hBMSC Cells to Osteoblasts
To evaluate the effects exerted by Glass and TiO2 surfaces on hBMSCs proliferation and differentiation to an osteoblast phenotype, cell viability assay (Alamar blue), apoptosis, gene

Effects of the Nanostructured TiO 2 Surface on Proliferation and Differentiation of hBMSC Cells to Osteoblasts
To evaluate the effects exerted by Glass and TiO 2 surfaces on hBMSCs proliferation and differentiation to an osteoblast phenotype, cell viability assay (Alamar blue), apoptosis, gene expression  Statistical significance values are indicated as *: p < 0.05, **: p < 0.001.

Cell Viability
hBMSCs cultured on TiO2 nanostructured surface in PM for seven, 14 and 28 days displayed the same viability measurements at each time point with respect to cells seeded on Glass. In fact, at seven days, the values of cell proliferation were 1.45 × 10 5 ± 0.04 and 1.47 × 10 5 ± 0.03 for Glass and TiO2, respectively. At 14 days, hBMSCs reached the values of 2.1 × 10 5 ± 0.04 on Glass and 2.5 × 10 5 ± 0.04 on TiO2, with a slight increase in cell number on TiO2 surface with respect to Glass. At 28 days, proliferation values were 4.1 × 10 5 ± 0.04 and 4.61 × 10 5 ± 0.04 for Glass and TiO2, respectively, with always a limited increase for cells cultured on TiO2 surface. Nevertheless, no large differences in proliferation were highlighted by Alamar blue assay, and, in addition, no apoptosis was revealed on the two different surfaces (data not shown).

Gene Expression Analyses
In order to characterize cell genotype, gene expression analyses were performed on hBMSC cells cultivated in PM and OM culture conditions on Glass and TiO2 surfaces by qRT-PCR technique at seven and 28 days of culture ( Figure 4), respectively. In Table 1, the gene primers used for qRT-PCR are reported. Figure 4 shows the fold induction of gene expressed in arbitrary units setting the expressions of the indicated genes in cells grown in Glass as equal to 1. At seven days, gene expression results of ALP, RUNX-2, COL 1 and FN in proliferative and osteogenic culture conditions are presented in Figure 4A. In PM, hBMSCs seeded on TiO2 showed no remarkable differences in gene expression of all selected genes that, however, were highlighted in OM. In fact, in OM, the expression of the same genes on TiO2 surface were significant higher (between 1.3 and 2 fold higher) with respect to Glass (for RUNX-2 and COL 1 p < 0.01, for ALP and FN p < 0.05), indicating an osteoinductive effect exerted by the combination of the culture medium with the nanostructured surface. At 28 days, the results of qRT-PCR for BOSP, Osterix, OSC, BMP-2, ALP and DCN genes are shown in Figure 4B. In PM there was an interesting up-regulation of the expression of OSC, BMP-2 and DCN genes in cells seeded on TiO2 nanostructured surface with respect to Glass (3 fold higher with p < 0.001 for OSC, 0.5 fold higher with p < 0.05 for BMP-2 and 0.4 fold higher with p < 0.05 for .04 for Glass and TiO 2 , respectively, with always a limited increase for cells cultured on TiO 2 surface. Nevertheless, no large differences in proliferation were highlighted by Alamar blue assay, and, in addition, no apoptosis was revealed on the two different surfaces (data not shown).

Gene Expression Analyses
In order to characterize cell genotype, gene expression analyses were performed on hBMSC cells cultivated in PM and OM culture conditions on Glass and TiO 2 surfaces by qRT-PCR technique at seven and 28 days of culture (Figure 4), respectively. In Table 1, the gene primers used for qRT-PCR are reported. Figure 4 shows the fold induction of gene expressed in arbitrary units setting the expressions of the indicated genes in cells grown in Glass as equal to 1. At seven days, gene expression results of ALP, RUNX-2, COL 1 and FN in proliferative and osteogenic culture conditions are presented in Figure 4A. In PM, hBMSCs seeded on TiO 2 showed no remarkable differences in gene expression of all selected genes that, however, were highlighted in OM. In fact, in OM, the expression of the same genes on TiO 2 surface were significant higher (between 1.3 and 2 fold higher) with respect to Glass (for RUNX-2 and COL 1 p < 0.01, for ALP and FN p < 0.05), indicating an osteoinductive effect exerted by the combination of the culture medium with the nanostructured surface. At 28 days, the results of qRT-PCR for BOSP, Osterix, OSC, BMP-2, ALP and DCN genes are shown in Figure 4B. In PM there was an interesting up-regulation of the expression of OSC, BMP-2 and DCN genes in cells seeded on TiO 2 nanostructured surface with respect to Glass (3 fold higher with p < 0.001 for OSC, 0.5 fold higher with p < 0.05 for BMP-2 and 0.4 fold higher with p < 0.05 for DCN). This genes up-regulation was maintained  Figure 4B), where TiO 2 nanostructured surface enhanced the expression of OSC, BMP-2 and DCN (0.3 fold higher for OSC, 0.2 fold for BMP-2 and 2.5 fold for DCN, with all p-value < 0.01) with respect to Glass. All together, these results confirmed that TiO 2 nanostructured surface, in comparison to Glass, did not alter gene expression of hBMSC cells; at the same time, in combination with OM conditions, TiO 2 surface promoted genotype bone differentiation of mesenchymal cells increasing the expression of some key genes involved in osteoblasts development. * GAPDH is the housekeeping gene.

Bone Matrix Deposition: Quantification and Immunolocalization Analyses
In order to evaluate the amount of extracellular matrix constituents produced by hBMSCs seeded on Glass and TiO 2 surfaces, an ELISA assay was performed. In Table 2a,b the protein content results are presented for hBMSCs seeded on the two different surfaces in PM (2a) and OM (2b) at 28 days, as pg/(cellsˆdisk). There was a significant enhancement of protein deposition for ALP, type-I collagen, type-III collagen, OP, ON, OSC, FN and DCN at 28 days in OM for cells seeded on TiO 2 nanostructured surface with respect to ECM deposition ratio on Glass (Table 2b). Differently, in PM, there was a slightly greater increase of ECM protein deposition only for FN, ON and OP for cells seeded on TiO 2 surface with respect to Glass surface (Table 2a). For the other proteins, no peculiar differences were noted. These results are almost in synchrony with gene expression analysis, highlighting the properties of nanostructured surface in combination with OM to direct terminal bone differentiation of hBMSC cells by the translation of specific osteogenic proteins.
At the end of cell culture (28 days), the immunolocalization of bone proteins were evaluated for TiO 2 and Glass surfaces by CLSM analysis (Figure 5). In particular, immunolocalization of type-I collagen, osteopontin and osteocalcin are reported in Figure 5 ( Figure 5A-F for PM, and Figure 5G-N for OM). Immunological studies confirmed that the nanostructured TiO 2 surface triggered osteogenic differentiation of hBMSC cells both in PM ( Figure 5D-F) and OM ( Figure 5L-N), with respect to the immunolocalization of osteogenic markers on Glass surface. In fact, the presence of these putative osteogenic proteins, in parallel with previous analyses of gene expression and protein deposition, seemed to be maintained or even slightly potentiated by TiO 2 surface, especially when combined with OM culture conditions ( Figure 5L-N). There was a significant enhancement of protein deposition for ALP, type-I collagen, type-III collagen, OP, ON, OSC, FN and DCN at 28 days in OM for cells seeded on TiO2 nanostructured surface with respect to ECM deposition ratio on Glass (Table 2b). Differently, in PM, there was a slightly greater increase of ECM protein deposition only for FN, ON and OP for cells seeded on TiO2 surface with respect to Glass surface (Table 2a). For the other proteins, no peculiar differences were noted. These results are almost in synchrony with gene expression analysis, highlighting the properties of nanostructured surface in combination with OM to direct terminal bone differentiation of hBMSC cells by the translation of specific osteogenic proteins.
At the end of cell culture (28 days), the immunolocalization of bone proteins were evaluated for TiO2 and Glass surfaces by CLSM analysis (Figure 5). In particular, immunolocalization of type-I collagen, osteopontin and osteocalcin are reported in Figure 5 ( Figure 5A-F for PM, and Figure 5G-N for OM). Immunological studies confirmed that the nanostructured TiO2 surface triggered osteogenic differentiation of hBMSC cells both in PM ( Figure 5D-F) and OM ( Figure 5L-N), with respect to the immunolocalization of osteogenic markers on Glass surface. In fact, the presence of these putative osteogenic proteins, in parallel with previous analyses of gene expression and protein deposition, seemed to be maintained or even slightly potentiated by TiO2 surface, especially when combined with OM culture conditions ( Figure 5L-N).

Alkaline Phosphatase Activity and Immunolocalization
On Day 28 of culture, ALP activity was measured as reported in Materials and Methods section ( Figure 6E). The ALP activity produced per min per mg of protein was higher in osteogenic condition both for cells seeded on Glass and TiO 2 nanostructured surface with respect to proliferative condition ( Figure 6E, p < 0.01). Nevertheless, no differences appeared between the two different surfaces in OM.
On Day 28 of culture, ALP activity was measured as reported in Materials and Methods section ( Figure 6E). The ALP activity produced per min per mg of protein was higher in osteogenic condition both for cells seeded on Glass and TiO2 nanostructured surface with respect to proliferative condition ( Figure 6E, p < 0.01). Nevertheless, no differences appeared between the two different surfaces in OM. ALP immunolocalization was then performed using CLSM ( Figure 6A-D). Density and morphology of ALP structures were quite similar on Glass both in PM and OM ( Figure 6A,C), with no appreciable visible differences. On the contrary, on TiO2 nanostructured surface, ALP immunolocalization was greater in OM with respect to PM culture conditions ( Figure 6B,D).

Localization of Calcium Deposits
On Day 28 (the end of the culture condition), hBMSC cells cultured in PM and OM on the two different surfaces were observed at CLSM for the detection of calcium deposits after staining with calcein ( Figure 7). Cells cultured on Glass ( Figure 7A) and on TiO2 ( Figure 7B) in OM resulted abundant in calcium deposits, even if they were more visible on TiO2 nanostructured surface compared to Glass. No calcium deposits were observed on either surface when cells were cultured in PM (data not shown). Complexone analysis ( Figure 7C) performed on cells cultured in osteogenic conditions revealed similar values between the surfaces, at about 2 pg/cell. However, on TiO2 surface, a slightly increase in calcium deposits with respect to Glass surface were observed ( Figure 7C, p < 0.05).  ALP immunolocalization was then performed using CLSM ( Figure 6A-D). Density and morphology of ALP structures were quite similar on Glass both in PM and OM ( Figure 6A,C), with no appreciable visible differences. On the contrary, on TiO 2 nanostructured surface, ALP immunolocalization was greater in OM with respect to PM culture conditions ( Figure 6B,D).

Localization of Calcium Deposits
On Day 28 (the end of the culture condition), hBMSC cells cultured in PM and OM on the two different surfaces were observed at CLSM for the detection of calcium deposits after staining with calcein ( Figure 7). Cells cultured on Glass ( Figure 7A) and on TiO 2 ( Figure 7B) in OM resulted abundant in calcium deposits, even if they were more visible on TiO 2 nanostructured surfacecompared to Glass. No calcium deposits were observed on either surface when cells were cultured in PM (data not shown). Complexone analysis ( Figure 7C) performed on cells cultured in osteogenic conditions revealed similar values between the surfaces, at about 2 pg/cell. However, on TiO 2 surface, a slightly increase in calcium deposits with respect to Glass surface were observed ( Figure 7C, p < 0.05). On Day 28 of culture, ALP activity was measured as reported in Materials and Methods section ( Figure 6E). The ALP activity produced per min per mg of protein was higher in osteogenic condition both for cells seeded on Glass and TiO2 nanostructured surface with respect to proliferative condition ( Figure 6E, p < 0.01). Nevertheless, no differences appeared between the two different surfaces in OM. ALP immunolocalization was then performed using CLSM ( Figure 6A-D). Density and morphology of ALP structures were quite similar on Glass both in PM and OM ( Figure 6A,C), with no appreciable visible differences. On the contrary, on TiO2 nanostructured surface, ALP immunolocalization was greater in OM with respect to PM culture conditions ( Figure 6B,D).

Localization of Calcium Deposits
On Day 28 (the end of the culture condition), hBMSC cells cultured in PM and OM on the two different surfaces were observed at CLSM for the detection of calcium deposits after staining with calcein ( Figure 7). Cells cultured on Glass ( Figure 7A) and on TiO2 ( Figure 7B) in OM resulted abundant in calcium deposits, even if they were more visible on TiO2 nanostructured surface compared to Glass. No calcium deposits were observed on either surface when cells were cultured in PM (data not shown). Complexone analysis ( Figure 7C) performed on cells cultured in osteogenic conditions revealed similar values between the surfaces, at about 2 pg/cell. However, on TiO2 surface, a slightly increase in calcium deposits with respect to Glass surface were observed ( Figure 7C, p < 0.05).

Discussion
The main goal of this manuscript was to elucidate, with experimental studies, the biocompatibility of a particular nanostructured surface based on Titanium dioxide deposited on a coverglass (TiO 2 ) with respect to a Glass surface. The biocompatibility was investigated in terms of adherence, proliferation and differentiation of hBMSC cells to both types of surface. Bone marrow mesenchymal stem cells have already demonstrated the capacity to spontaneously differentiate into bone on supports/scaffolds [24] with chemical or mechanical stimulation [25,26]. Nevertheless, in this study, we focused on the osteoinductive and proliferative effects of nanostructured TiO 2 surface.
Nanomaterials represent a particular sub-field of tissue engineering and they have become an innovative technology to be applied in regenerative medicine. Several studies were related to the design of various types of nanostructured scaffolds that could allow the regeneration of different type of tissues [27][28][29]. In addition, it was reported that these materials and matrices have been compared and evaluated for their better biocompatibility and efficacy in supporting the damaged tissue [30][31][32].
Titanium dioxide surface used in this study was produced by the deposition of a supersonic beam of TiO x clusters (http://www.tethis-lab.com) [13]. This method produced a biocompatible substrate composed by films with a homogeneous nanoscale porosity and roughness [13,14]. This particular nanoscale (with clusters under 100 nm of dimension) may be tuned to modulate specific cell-biomaterial surface interactions. For this reason, we deeply investigated how this nanotopography could interfere with stem cell adhesion, proliferation and differentiation to ameliorate autologous bone grafts. We used human bone marrow mesenchymal stem cells for their plasticity in supporting healing processes and for their large and undeniable use in regenerative medicine.
First of all, we analyzed at 24 h and seven days cell attachment and morphology by performing a careful comparison with cells seeded on TiO 2 nanostructured surface and Glass (Figures 2 and 3). The evaluation of p-FAK foci indicates unequivocally cellular behavior in response to different matrices [33,34]. In our study, TiO 2 nanostructured surface promoted in a short time (24 h) the amount of hBMSCs pFAK foci with respect to common Glass (Figure 2). In fact, cells modify focal adhesion in response to modification in extracellular matrix environment, like molecular/biochemical composition or modification in 2D or 3D structure. In this case, the surface of Titanium dioxide changed hBMSCs microenvironment and ECM and provoked a phosphorylation of focal adhesion kinase (FAK) at Y397 that stimulate cell proliferation. As regards morphology, no visible morphological alterations were observed between cells seeded on the two different surfaces, they maintained fibroblast-like morphology, but with an increase of alpha-actin presence and a down-regulation of alpha-tubulin for hBMSCs seeded on TiO 2 nanostructured surface in OM ( Figure 3). These data seems to confirm previous results reported by Khang et al. (2012), showing that the roughness of the modified surface may induce osteoblasts differentiation with an increase of actin expression in cells [35]. The down-regulation of the alpha-tubulin could also be a response of hBMSC cells to the nano-topography of the surface and to the OM that enhanced bone differentiation. In addition, no cell apoptosis was revealed on different surfaces and cell viability analyses at seven, 14 and 28 days both in PM and OM showed no large differences of cell proliferation. We may conclude that the surface of TiO 2 appeared not toxic for the cells: moreover, these results seems to follow the trends obtained by other groups by culturing cancer cell lines on the same type of biomaterial surface [13].
Secondly, a temporal (at seven and 28 days) and functional pattern of osteogenesis gene expression was analyzed for cells seeded on the two different surfaces (TiO 2 , and Glass). We focused our attention on the expression of different genes at each time point. In particular, at seven days we evaluated the expression of RUNX-2, type-I collagen, ALP and FN, that are genes related to the early and intermediate bone development; in fact RUNX-2 is an important transcription factor associated with osteoblast differentiation [36], and ALP and FN represented early bone formation [37]. Instead, at 28 days, BOSP, Osterix, OSC, BMP-2, ALP and DCN expression were investigated (Figure 4). These genes are osteoblast-specific markers and they are not present in undifferentiated hBMSCs. They could be detected from day 14 in OM, composed by dexamethasone, ascorbic acid, and β-glycerophosphate [38].
Gene expression results confirmed that TiO 2 surface , in comparison with Glass, did not interfere on gene expression of hBMSC cells. At seven days, the nanostructured surface enhanced osteoblast differentiation by the up-regulation of the genes implicated in early bone development, whereas at 28 days, it promoted genotype bone differentiation by the up-regulation of three important genes, like DCN, BMP-2 and OSC, still in PM and OM (Figure 4b). Surprisingly, TiO 2 surface induced an overexpression of OSC just in PM, even 3 fold higher with respect to Glass. It was shown by previous studies that, in fact, systemic and local effects of OSC potentially link bone remodeling, vascular calcification, and energy metabolism [39]. Furthermore, the bone-morphogenetic protein 2 (BMP-2) is associated to participate in bone healing and regeneration. Previous studies demonstrated that BMP-2 protein is one of the most potent growth factors that induce mesenchymal stem cell and osteoprogenitor cell differentiation into osteoblasts [40]. Probably, the effect of the TiO 2 nanostructured surface, in combination with OM, provoked a positive effect on cell bone differentiation by moderate high levels of BMP-2 and an up-regulation of OSC gene.
In order to corroborate gene expression data, we calculated the amounts of extracellular matrix constituents produced by cells at 28 days on different surfaces (Table 2a,b, ELISA assay). For this purpose, we measured the amount of the fundamental bone matrix constituents such as osteopontin, osteocalcin, osteonectin and fibronectin. No appreciable differences in the amounts of proteins in PM conditions were observed, both on Glass and on TiO 2 nanostructured surface. Otherwise, there was a specific enhancement induced by TiO 2 surface of osteogenic protein deposition in OM. All these proteins produced by cells represent important markers of bone development: osteopontin is known to play an important role in cell adhesion [41] and calcification of mineralized tissue, whereas DCN represents an osteoblasts terminal differentiation marker [42]. Osteonectin is a calcium and collagen binding ECM glycoprotein and modulates cell-matrix interactions [43]. In summary, based on molecular and protein results, we may assume that the main effect of TiO 2 nanostructured surface was to facilitate hBMSC cells adherence and, at the same time, to promote protein ECM deposition and, therefore, cell differentiation.
To reinforce molecular and protein results, immunolocalization of osteogenic markers at 28 days confirmed the osteoconductive role of titanium dioxide surface: the presence of type-I collagen, osteopontin and osteocalcin seems to be more evident on TiO 2 nanostructured surface in OM with respect to Glass ( Figure 5). In addition, quantitative analysis of the calcium mineral content showed that TiO 2 surface promoted deposition of newly mineral matrix (Figure 7), indicating that, more likely, the physical-chemical clusters of the nanosurface promoted Ca 2++ deposition. Accordingly, to these data, also the increase of ALP activity was consistent to mineral matrix deposition ( Figure 6), underlying the mechanism that osteoblastic marker proteins such as alkaline phosphatase and the neo-formed mineralized extracellular matrix (ECM) were a consequence of osteoblastic differentiation supported by the nanostructured material.
In summary, this study should be considered a preliminary in vitro investigation to setup further analyses of the effects of nanostructured surfaces on human multipotent cells and subsequently to be translated to animal and human trials. We demonstrated that the nanostructure of TiO 2 surface could be successfully employed for in vitro studies of biocompatibility and we speculated that this biomaterial surface might be a good promising surface for bone tissue engineering applications.
The slides obtained were sterilized under the UV light of our sterile cabinet for 1-2 h and then left overnight with complete medium to assess sterility. This step lets also the proteins of the serum to attach to the surface. This approach reproduces an in vivo situation in which many proteins will adsorb on the surface until complete colonization by the cells. However, studies with the crude surface were also done. The slides were further characterized with SEM imaging prior to any use, to evaluate the topography and the microstructural configuration. As a control surface, the same microscopy coverglass used for the coating deposition of the film of cluster-assembled TiO 2 , was employed for the study.

Reagents
Unless otherwise specified, all reagents were from Sigma Aldrich (St Louis, MO, USA). Dr. Larry W. Fisher (National Institutes of Health, Bethesda, MD, USA) provided us with the rabbit polyclonal anti type-I collagen, anti-osteopontin, anti-osteocalcin and anti-alkaline phosphatase.

Scanning Electron Microscopy Analysis
We analyzed the microstructural surface of TiO 2 nanostructured surface with scanning electron microscopy (SEM). The biomaterial surface was treated for SEM analysis with 2.5% (v/v) glutaraldehyde solution in 0.1 M sodium cacodylate buffer (pH = 7.2) for 1 h at 4˝C. Then, as previously described, samples coated with gold were observed at 50 and 1000ˆwith a Stereoscan 440 microscope (Leica Miicrosystems, Bensheim, Germany) at 8 kV [44,45].

Cell Culture
The human mesenchymal stem cells (hBMSCs) were expanded from discarded residues (3 mL) of bone marrow harvested for hematopoietic stem cell transplantation, (3 mL) after obtaining written informed consent from the healthy donors.
Successively, hBMSCs were phenotypically analyzed in order to evaluate their mesenchymal properties, according to the International Society for Cellular Therapy [46].

Adhesion and Morphological Analysis
For adhesion (24 h in PM) and morphological analyses (24 h in PM and seven days in PM and OM), paraformaldehyde fixed hBMSC cells on Glass and TiO 2 were permeabilized with 0.1% Triton X-100 for 10 min at room temperature (RT).
For morphological studies: Cells either at 24 h in PM or at seven days in PM and OM, were incubated with phalloidin (Alexa-Fluor-488 phalloidin, Invitrogen) for 20 min and then with anti-tubulin (Alexa-Fluor 633, Invitrogen) for 20 min at RT.
All the previous samples were mounted and nuclei were counterstained with Hoechst (Sigma Aldrich). The images were taken by the TCS SPII confocal microscope (Leica Microsystems, Bensheim, Germany) equipped with a digital image capture system at 20ˆand 40ˆmagnification. In particular, for quantification of foci, more than 10 images were collected and analyzed with ImageJ (http://rsb.info.nih.gov/ij/; National Institutes of Health, Bethesda, MD, USA) to count positive foci of cells seeded on TiO 2 and on Glass surfaces in each photographed field.

Immunological Studies
Cells at 28 days of culture on Glass and TiO 2 surfaces in different medium (PM and OM) were stained with primary antibodies against anti-type-I collagen, anti-osteopontin, anti-osteocalcin, anti-alkaline phosphatase (anti-ALP) at 4˝C ON. We used as secondary antobodies Alexa-Fluor-488 goat anti-rabbit IgG (HþL; Invitrogen). Pictures were taken with confocal microscope (Leica Microsystems, Bensheim, Germany) at 40ˆmagnification.

Cell Viability Assay
To evaluate proliferation of hBMSCs at seven and 14 days on the two different surfaces (Glass and TiO 2 ), we used Alamar Blue reagent (Sigma Aldrich). According to manufacturer instructions 1:10 dilution in serum free Dulbecco's modified Eagle's medium (DMEM) low glucose was added to cells and left for 3 h at 37˝C in 5% CO 2 . Then, absorbance of 100 µL at 595 nm was read and converted to cell numbers with a conversion curve. The cell viability was also performed at the end of culture conditions (28 days in PM and OM culture conditions).

Apoptosis
Apoptosis of hBMSCs was performed by Annexin V technique as previously described [47]. The Annexin V-FITC Apoptosis Detection Kit (Bender Medsystems, Vienna, Austria) was used according to the manufacturer's instructions to evaluate apoptosis of hBMSCs seeded and incubated for 24 h at 37˝C on the two different surfaces.

Gene Expression Analyses
Total RNA from hBMSC cells seeded on Glass and TiO 2 surfaces at 7 and 28 days in PM and OM was extracted and retrotranscribed into cDNA as previously reported [48]. Gene expression analyses were performed by qRT-PCR using oligonucleotide primers exemplified in Table 2. The fold expression of each sample was normalized to the GAPDH housekeeping gene and analyzed in triplicate.

Purified Proteins and Polyclonal Antisera
All the proteins analyzed in this manuscript were purified as described previously [49]. Dr. Larry Fisher (National Institutes of Health, Bethesda, MD, USA) provided us with the rabbit polyclonal anti-human antibodies against type-I and -III collagen, decorin, osteopontin, osteocalcin, osteonectin, and ALP [49].

Bone ECM Proteins Extraction and ELISA Assays
The evaluation of ECM produced by cells seeded on Glass and TiO 2 in PM and OM culture conditions was performed by an enzyme-linked immunosorbent assay (ELISA) as previously reported [50,51]. The total protein concentration was evaluated with the BCA Protein Assay Kit (Pierce Biotechnology, Inc., Rockford, IL, USA). The total protein concentration was in PM 90.1˘7.1 µg/mL for the hBMSCs seeded on Glass and 118.5˘8.2 µg/mL for the hBMSCs seeded on TiO 2 nanostructured surface. In OM, the protein concentration was 85.7˘2.5 µg/mL for cells on Glass and 105.0˘3.3 µg/mL for cells on TiO 2 surface.

ALP activity
ALP activity of cells was estimated by a colorimetric assay in PM and OM conditions, as previously reported [49,50].

Calcium deposition
To evaluate the calcium deposition, fluorescent calcein detection and calcium-cresolphthalein complexone methods were performed on hBMSCs seeded on both surfaces as described previously [50][51][52]. 4.12.1. Calcein detection hBMSCs seeded on Glass and TiO 2 at 28 days of culture were stained with a calcein solution 5 mM in PBS (Invitrogen, Carlsbad, CA, USA) for 30 min at 22˝C. The pictures were taken by a confocal microscope at 40ˆmagnification.

Calcium-Cresolphthalein Complexone Method
In order to evaluate calcium deposition, the calcium-cresolphthalein complexone method was performed on hBMSCs seeded on both surfaces at 28 days in PM and OM conditions as previously reported [49,52].

Statistics
Each experiment reported in the Results Section was done in triplicates and at least in 3 separated experiments. Results are expressed as the mean˘standard deviation. Statistical significance between Glass and TiO 2 surfaces was evaluated by the one-way analysis of variance (ANOVA) with post hoc Bonferroni test.