Osteogenesis on titanium (Ti) surfaces is a complex process involving cell–substrate and cell–cell interaction of osteoblasts and endothelial cells. The aim of this study was to investigate the osteogenic properties of Ti surfaces on osteoblasts in the presence of endothelial cells (ECs).
Osteoblast-like cells (MG63 cells) and human umbilical vein endothelial cells (HUVECs) were grown in cocultures on four kinds of Ti surfaces: acid-etched (A), coarse-grit-blasted and acid-etched (SLA), hydrophilic A (modA) and hydrophilic SLA (modSLA) surfaces. MG63 cells in single cultures served as controls. Cell ratios and cell types in cocultures were determined and isolated using flow cytometry. Cell numbers were obtained by direct cell counting. In MG63 cells, alkaline phosphatase (ALP) activity was determined and protein levels of osteocalcin (OC) and osteoprotegerin (OPG) were detected with enzyme-linked immunosorbant assay (ELISA). The mRNA levels of ALP, OC and OPG of sorted MG63 cells were determined with real time polymerase chain reaction (PCR).
MG63 cells proliferated in the presence of HUVECs, which showed higher cell numbers on Ti surfaces (A, SLA, modSLA) after 72 h, and lower cell numbers on Ti surfaces (modA, SLA, modSLA) after 120 h in comparison to single cultures. Protein and mRNA levels of ALP and OPG were higher in cocultures than in single cultures, while OC exhibited a lower expression. These three parameters were higher expressed on modA, SLA and modSLA surfaces compared to A surfaces.
Cocultures of osteoblasts and endothelial cells represent the most recently developed research model for investigating osteogenesis and angiogenesis which play both a major role in bone healing. This paper investigates for the first time the osteogenic properties of titanium surfaces used for dental implants with a coculture system with osteoblast-like cells and endothelial cells: (1) In cocultures with ECs (HUVECs) osteoblast-like cells (MG63 cells) show enhanced expression of early differentiation markers and osteogenic factors on Ti surfaces compared to single cultures of MG63 cells. (2) The differentiation and the expression of an osteogenic phenotype of osteoblast-like cells (MG63 cells) in coculture with ECs (HUVECs) is enhanced by both hydrophilicity and roughness of Ti surfaces.
Wound healing around endosseous titanium (Ti) implant surfaces is principally a series of molecular and cellular events involving protein absorption, cytokine release from immune cells, angiogenesis, recruitment and migration of osteogenic cells, cell adhesion, cell–surface interaction, cell–cell interaction, differentiation, extracellular matrix formation and mineralization . Subsequently, bone bonding or osteointegration is formed with contact osteogenesis and distant osteogenesis .Albrektsson et al. emphasized surface structure as one of the six key factors for the success of endosseous implants. In the last 20 years, a huge number of experimental studies have demonstrated that the early bone response is profoundly influenced by surface properties, such as topography, roughness, surface energy and hydrophilicity . A substantial body of evidence has been accumulated with the development of endosseous implant surface investigations resulting in a higher clinical success rate. However, the search for an optimal surface especially for compromised conditions and acceleration of the healing period is still underway. The mechanisms behind the bone response to surface properties remain unclear to a certain degree.
The in vitro cell culture study is a basic strategy to investigate the mechanism of molecular and cellular response to titanium surface properties. In the past 30 years cell migration, adhesion, proliferation and differentiation have been studied with different cell types involved in tissue-integration of titanium surfaces by monoculturing these cells on surfaces with different properties .
However, wound healing around endosseous implants is a complex process involving both cell–surface interaction and cell–cell interaction of the same and different cell types. Bone vascularization plays a pivotal role in osteogenesis . The existence of tight communication between bone forming osteoblasts and vessel forming endothelial cells in osteogenesis is widely accepted . Therefore, it is not reasonable to investigate the mechanism of bone response to surface properties without the participation of endothelial cells.
One appropriate approach to overcome this disadvantage is the use of a coculture model of osteoblasts and endothelial cells was required. In our previous study, a method using fluorescence-activated cell sorting for analyzing the functional properties of osteoblasts grown in direct contact with human umbilical endothelial cells (HUVECs) has been established . The aim of the present study was to investigate the osteogenic properties of titanium surfaces with different roughness and hydrophilicity with this recently developed coculture model. In line with this objective the following null hypotheses were formulated for endothelial/osteoblast-like cells—cocultures on Ti surfaces: (1) the proportion of MG63 cells and HUVECs remains constant during the incubation period. (2) HUVECs do not affect the proliferation of MG63 cells. (3) HUVECs do not affect the differentiation of MG63 cells. (4) Hydrophilicity and roughness of Ti surfaces do not affect proliferation, differentiation and osteogenic properties of MG63 cells in the presence of HUVECs.
Materials and methods
Ti disks were prepared from 1-mm-thick sheets of unalloyed commercially pure titanium grade 2 (Institut Straumann AG, Basel, Switzerland). The disks were punched to be 15 mm in diameter to fit into the well of a 24-well tissue culture plate. The treatments of four kinds of Ti surfaces, acid-etched (A) surfaces, hydrophilic A (modA) surfaces, coarse-grit-blasted and acid-etched (SLA) surfaces and hydrophilic SLA (modSLA) surfaces have been described in previous studies . Surface roughness, chemical composition, water contact angle and surface free energy have been reported previously . In brief, the arithmetic mean deviations of the surface (Sa) were 0.6 ± 0.01 μm on A and modA surfaces, 1.2 ± 0.04 μm and 1.2 ± 0.03 μm on SLA and modSLA surfaces. On all surfaces, the chemical composition contained carbon, oxygen, nitrogen and titanium. Ti and O were mainly in form of TiO 2 and the suboxides Ti 2 O 3 and TiO on all surfaces whereas carbon was mainly in the form of hydrocarbon and carboxides with about 35 at% on A and SLA surfaces, while on modA and modSLA the carbon amount was about 15 at%. A and SLA showed water contact angles of 122.40 ± 7.39° and 139.88 ± 8.69°, respectively, whereas contact angles on modA and modSLA were close to 0°.
Commercially available osteoblast-like MG63 cells (American Type Culture Collection, Rockville, USA) and human umbilical vein endothelial cells (HUVECs, Technoclone, Vienna, Austria) were used in the present study. MG63 cells were cultured in modified Eagle’s minimum essential medium (Gibco ® , Invitrogen, Carlsbad, USA) supplemented with 10% fetal bovine serum, streptomycin (50 μg/ml), and penicillin (100 U/ml). HUVECs were cultured in endothelial cell growth medium (ECM) supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 0.25 μg/ml fungizone, 2 mM l -glutamine, 5 U/ml heparin, 30–50 μg/ml endothelial cell growth supplement (Technoclone, Vienna, Austria), and 20% fetal calf serum (FCS) in 150 cm 2 tissue culture flasks coated with 0.2% gelatine. Both MG63 cells and HUVECs were maintained at 37 °C in a humidified atmosphere containing 5% CO 2 . Three independent experiments were performed using cells between the third and the sixth passage.
Coculture of MG63 cells and HUVECs
A, modA, SLA and modSLA titanium discs were placed in 24-well tissue culture plates (TPP, Switzerland). MG63 cells and HUVECs cells were seeded together in 24-well plates in ECM at a density of 0.5 × 10 4 and 2.5 × 10 4 cells/cm 2 (1:5), respectively. MG63 cells seeded alone in ECM at the same density served as controls. For cell proliferation assays and evaluations of protein expression of osteogenic markers 12 wells were used for each kind of Ti surface. HUVECs seeded alone in ECM at the same density served as controls for protein expression assays. For real time polymerase chain reaction (PCR) analyses 6 wells were seeded. MG63 cells seeded alone on tissue culture plastic surfaces served as a calibrator for the relative mRNA measurements.
The proliferation of individual cell types in cocultures was determined by direct cell counting and measuring proportions of individual cell types by flow cytometry. The proliferation of MG63 cells in controls was determined by direct cell counting.
The total cell number in coculture and control wells was counted after 24 h, 72 h and 120 h of culture. Cells were washed twice with PBS and detached with 0.05% trypsin/EDTA solution. Each well was washed three times with ECM. Cells were centrifuged, resuspended in 200 μl of ECM and 60 μl of cell suspension was used for cell counting. The cells were dyed by Trypan blue and viable cells were counted blindly both with an Auto T4 Cellometer Reader (Nexcelom Bioscience LLC, Lawrence, MA) and manually.
In parallel to cell counting, the proportion of MG63 cells and HUVECs in coculture was determined by flow cytometry (FC) in order to assess proliferation of individual cell types. Cells remaining after cell counting were centrifuged and resuspended in 100 μl of ice-cold PBS containing 3% (w/v) bovine serum albumin (BSA) and 0.1% NaN 3 (FC buffer). Afterwards, 20 μl of fluorescein isothiocyanate (FITC)-conjugated monoclonal anti-human CD31 (eBioscience, San Diego, CA, USA) were added to this cell suspension and incubated on ice in the dark for 30 min. The cells were washed twice after staining, finally resuspended in 0.5 ml of FC buffer, and the proportion of different cell types was measured by flow cytometry (FACSCalibur, Becton Dickinson, CA, USA). Each sample was acquired in triplicate. Each time 20,000 events were measured. For data analyses the software CellQuest 3.3 (Becton Dickinson, Franklin Lakes, NJ, USA) was used. CD31-positive cells were considered as endothelial cells, CD31-negative cells were considered as MG63 cells. The proportion of MG63 cells in coculture was calculated for each sample based on the means of three acquisitions.
Gene expression in MG63 cells sorted by fluorescence-activated cell sorting
The expression levels of various specific osteoblastic differentiation markers in MG63 cells were measured by real time PCR after 120 h of culture. In cocultures MG63 cells were preliminary separated from HUVECs by fluorescence-activated cell sorting. Cells from 6 wells were pooled together, centrifuged and resuspended in 100 μl. This cell suspension was stained with FITC-conjugated anti-human CD31 antibody similar to the method described above, washed and resuspended in the final volume of 2 ml. The population of MG63 cells was gated in the CD31-negative cells quadrant, sorted in the exclusion mode, and collected into sorting tubes coated with 4% BSA. The number of sorting events was 50,000 for each sample. In controls, MG63 cells were also passed through similar staining and sorting procedures.
After collection cells were centrifuged and the total cellular mRNA was isolated and transcribed into cDNA using TaqMan ® Gene Expression Cells-to-CT™ kit (Ambion/Applied Biosystems, CA, USA) according to the manufacturer’s instruction. Real time PCR was performed on an Applied Biosystems Step One Plus real time PCR system (Applied Biosystems, CA, USA) using the Taqman ® gene expression assays with the following ID numbers (all from Applied Biosystems, CA, USA): alkaline phosphatase (ALP), Hs01029144_m1; osteocalcin (OC), Hs00609452_g1; osteoprotegerin (OPG), Hs00171068_m1; β-actin, Hs99999903_m1. Triplicate PCR reactions were prepared for each sample at the following thermocycling conditions: initiation at 95 °C for 10 min, then performing 40 cycles each of them consisting of denaturation at 95 °C for 15 s and hybridization-elongation at 60 °C for 1 min. The point at which the PCR product was first detected above a fixed threshold (termed cycle threshold, C t ), was determined for each sample. Changes in the expression of target genes were calculated using 2 − ΔΔC t , where <SPAN role=presentation tabIndex=0 id=MathJax-Element-1-Frame class=MathJax style="POSITION: relative" data-mathml='ΔΔCt=Cttarget−Ctβ-actinsample−Cttarget−Ctβ-actincalibrator,’>ΔΔCt=(Ctargett−Cβ-actint)sample−(Ctargett−Cβ-actint)calibrator,ΔΔCt=Cttarget−Ctβ-actinsample−Cttarget−Ctβ-actincalibrator,
Δ Δ C t = C t target − C t β -actin sample − C t target − C t β -actin calibrator ,
taking samples of MG63 cells grown in single culture on tissue culture plastic as a calibrator.
Quantitative measurements of alkaline phosphatase activity
The activity of alkaline phosphatase (ALP) was measured similar to the method described in our previous studies after 120 h of culture. In brief, after stimulation cells were divided into two parts. One part of cells was used for determination of the total number of MG63 cells by direct cell counting and flow cytometry as described above. The other part of cells was lysated in 200 μl PBS containing 0.2% Triton X-100 and homogenized by sonification. The ALP activity was assayed using the conversion of a colorless p -nitrophenyl phosphate to a colored p -nitrophenol according to the manufacturer’s protocol (Sigma, St. Louis, MO, USA). The color change was measured spectrophotometrically at 405 nm, and the amount of enzyme released by the cells was quantified by comparison with a standard curve. ALP levels were normalized to the total number of MG63 cells at the end of the experiment. ALP activity experiments were repeated three times. Data were expressed as ratios of nanomoles of inorganic phosphate (Pi) cleaved by the enzyme in 30 min per 100,000 cells.
Quantitative measurements of osteocalcin and osteoprotegerin production in the medium
The level of protein expression was measured in the medium after 120 h of production using commercially available ELISA kits: OC production by MicroVue™ Osteocalcin EIA kit (Quidel, CA, USA); OPG by a sandwich enzyme-linked immunosorbent assay (Biomedica, Vienna, Austria). Measurements were performed according to the manufacturer’s instruction. The detection limits and sample dilutions were: OC, 0.45 ng/ml, 1:1; OPG, 0.14 pmol/l, 1:500. The results were normalized to the total numbers of MG63 cells, which were measured as described above.
Data are expressed as mean ± SD. The normal distribution of all data was tested with Kolmogorov–Smirnov test. After confirming normal distribution, the statistical differences between coculture and single culture groups were analyzed by t -tests. The statistical differences between the effects of the various titanium surfaces on the cultured cells were analyzed by analysis of variance (ANOVA). All statistical analyses were performed using SPSS 14.0. Differences were considered to be statistically significant at P < 0.05.