0.2-mm thick grade-2 commercially pure titanium (cp-Ti) sheets (TSM-TECH Co. Ltd., Ulsan, Korea) were mechanically polished to obtain a finish surface with Ra ≤ 0.1 µm. Microgrooves were applied on Ti to have 15 and 60 µm width, and 3.5 and 10 µm depth by photolithography (MEMSware Inc., Kwangju, Gyeonggi, Korea), respectively (NE15/3.5 and NE60/10). Details of the photolithography procedures were reported in our previous study.15 Further acid etching was applied to create Ti surfaces with etched microgrooves (E15/3.5 and E60/10). Both smooth- and acid-etched Ti were used as the controls (NE0 and E0) (Fig. 1).

A drop shape analysis system goniometer, EasyDrop® contact angle measuring instrument (KRÜSS GmbH, Hamburg, Germany), was used for contact angle measurement. Distilled water (6 µl per drop) was used as a probe for contact angle calculation. Measurements were taken for each drop after 15 s deposition from 3 independent samples of each control and experiment Ti substrata. The drop images were captured serially 5 times by a video camera in the directions parallel with as well as perpendicular to the surface microgrooves. The contact angles used for data were the averages of the angles of 5 serial captures of each water drop calculated by an image analysis system.

MC3T3-E1 mouse preosteoblasts (MC3T3 cells) were purchased from Korean Cell Line Bank (KCLB, Seoul, Korea). The cells were cultured in α-modified Eagle's medium (α-MEM; WelGene, Daegu, Korea) containing 10% fetal bovine serum (FBS; Sigma-Aldrich Co., St. Louis, MO, USA) and incubated at 37℃ in a humidified atmosphere of 5% CO₂ for 12 h. Non-adherent cells were aspirated, adherent cells were cultured and expanded and the medium was refreshed every 2 days for confluence till the 3rd-5th passages of culture were obtained. The cells were washed in 10 ml phosphate-buffered salines (PBS; Gibco BRL, Grand Island, NY, USA) and removed by trypsin-EDTA solution (0.25% trypsin and 0.1% glucose dissolved in 1 mM of EDTA-saline; Sigma-Aldrich Co., St. Louis, MO, USA). To induce osteogenic activity, MC3T3 cells were cultured in an osteogenic media [DMEM (Dulbecco's modified Eagle's medium, WelGene, Daegu, Korea) supplemented with 10% FBS (Sigma-Aldrich Co., St. Louis, MO, USA), 50 ug/ml of α-ascorbic acid, 10 mM of β-glycerophosphate, 100 nM of dexametasone and antibiotics (Invitrogen, Carlsbad, CA, USA)].

MC3T3 cells were plated on the control and experiment Ti substrata that were previously attached to the bottom of the 96-well tissue culture plates (96-well Ti substrata) at a population density of 3 × 10³ cell/ml and incubated in a humidified incubator at 37℃ with 5% CO₂ in 95% air for 16 and 24 h. 1 ml of bomodeoxyuridine (BrdU) labeling reagent (Roche Diagnostics GmbH, Mannheim, Germany) was added to each well and the cells were reincubated for 2 h at 37℃. Details of the BrdU assay procedure human gingival fibroblasts were reported in our previous study.16 In all groups, the reaction products were transferred to 96-well plates and the absorbance was measured by monitoring the light absorbance of the solution at 370 nm using a microplate reader (Bio-Rad, Hercules, CA, USA).

To estimate the osteogenic activity, the level of alkaline phosphatase (ALP) activity was measured. MC3T3 cells were plated on the 12-well Ti substrata of the control and experiment groups at a density of 5 × 10⁴ cells/ml and cultured for 2 days for confluence. The cells were then incubated in an osteogenic media [DMEM (WelGene, Daegu, Korea) supplemented with 10% FBS (Sigma-Aldrich Co., St. Louis, MO, USA), 50 ug/ml of α-ascorbic acid, 10 mM of β-glycerophosphate, 100 nM of dexametasone and antibiotics (Invitrogen, Carlsbad, CA, USA)] at 37℃, 5% CO₂ for 1, 7 and 14 days. The cultured cells were washed with PBS (Gibco BRL, Grand Island, NY, USA), removed by trypsin-EDTA solution, lysated with 0.1% Triton X-100 buffer and sonicated in ice. Aliquots of 50 µl were incubated with 100 µl of 1 M Tris-HCl (pH 9.0), 5 mM MgCl₂, 20 µl of 5 mM p-nitrophenyl phosphate solution (Sigma-Aldrich, St. Louis, MO, USA) for 30 min at 37℃. The reaction was quenched by adding 250 µl of 1 N NaOH and placing the mixture in ice. The level of p-nitrophenol production in the presence of ALP was measured by monitoring the light absorbance of the solution at 405 nm using a microplate reader (Bio-Rad, Hercules, CA, USA). Measurements were compared using p-nitrophenol standards and normalized using the total protein amounts to account for differences in the number of cells on different titanium substrata at individual time points.

Experiments involving the contact angle determination, BrdU assay and ALP activity test were repeated simultaneously and independently in triplicate and the mean values and standard deviations were calculated. One-way analysis of variance (one-way ANOVA) was used to compare the mean values between the groups of NE0, NE15/3.5, NE60/10, E0, E15/3.5 and E60/10. Pearson's correlation analysis was used to analyze the correlations between the results from the contact angle determination, BrdU assay and ALP activity test. Multiple regression analysis was used to determine the influential factors, among the hydrophilicity of Ti and cell proliferation, on the osteogenic activity. The SPSS 17.0 software program was used for all statistical analyses in this study.

In ANOVA, multiple comparisons of the data from the contact angle determination in parallel direction with the microgrooves showed the mean contact angle value of NE60/10 to be significantly smaller compared to that of NE0 (P < .01) and the mean values of the E-series to be significantly smaller compared to those of the NE-series (P < .01) (Fig. 2). In the perpendicular direction, the mean value of the NE0 was significantly greater compared to those of NE15/3.5, NE60/10, E0, E15/3.5, or E60/10 (P < .01) and the mean value of E15/3.5 was significantly smaller compared to that of NE60/10 (P < .01) (Fig. 2). Independent t-test (Student t-test), comparing the results between parallel and perpendicular contact angles, showed no differences in the mean values in all groups, except for NE15/3.5, where the contact angles measured in the direction perpendicular to the microgrooves were significantly smaller compared to those measured in the parallel direction (P < .01) (Table I). All other comparisons between groups were not statistically significant.

In ANOVA, multiple comparisons of the MC3T3 cell proliferation data from the BrdU assay at 24 h incubation showed the mean OD value of E60/10 to be significantly greater compared to that of NE15/3.5, E0 or E15/3.5 (P < .05) (Fig. 3). All other comparisons between groups, including those at 16 h incubation, were not statistically significant.

In ANOVA, multiple comparisons of the MC3T3 cells' osteogenic activity data from the ALP activity test after 1, 7 and 14 days of osteogenic culture showed no significant differences in the mean ALP activity values (Fig. 4).

The results from seven experiments such as the contact angle determination in the directions parallel with and perpendicular to the microgrooves (Parallel and Perpendicular); the BrdU assay of MC3T3 cells at 16 and 24 h incubation (BrdU 16 h and BrdU 24 h); and the ALP activity test after 1, 7 and 14 days of osteogenic culture (ALP 1 day, ALP 7 days and ALP 14 days) were used as the variables in the Pearson's correlation analysis. Significant correlations were present between Parallel and Perpendicular, Parallel and ALP 14 days, Perpendicular and ALP 14 days and BrdU 16 h and ALP 14 days (P < .01) (Table II and Fig. 5). Using ALP 14 days as the dependent variable and using Parallel, Perpendicular, BrdU 16 h and BrdU 24 h as the independent variables, multiple regression analysis in the enter method showed that all independent variables were determined to influence ALP 14 days (Table III). In multiple stepwise regression analysis using the dependent and independent variables identical to the enter method, both Parallel and BrdU 16 h were determined as the influential factors on ALP 14 days in the regression model 2, whereas Parallel was determined as the greatest influential factor on ALP 14 days in the regression model 1 (Table III).