Novel Yttria-Stabilized Zirconium Oxide and Lithium Disilicate Coatings on Titanium Alloy Substrate for Implant Abutments and Biomedical Application

This study aimed to create novel bioceramic coatings on a titanium alloy and evaluate their surface properties in comparison with conventional prosthetic materials. The highly polished titanium alloy Ti6Al4V (Ti) was used as a substrate for yttria-stabilized zirconium oxide (3YSZ) and lithium disilicate (LS2) coatings. They were generated using sol-gel strategies. In comparison, highly polished surfaces of Ti, yttria-stabilized zirconium oxide (ZrO2), polyether ether ketone (PEEK) composite, and poly(methyl methacrylate) (PMMA) were utilized. Novel coatings were characterized by an X-ray diffractometer (XRD) and scanning electron microscope (SEM). The roughness by atomic force microscope (AFM), water contact angle (WCA), and surface free energy (SFE) were determined. Additionally, biocompatibility and human gingival fibroblast (HGF) adhesion processes (using a confocal laser scanning microscope (CLSM)) were observed. The deposition of 3YSZ and LS2 coatings changed the physicochemical properties of the Ti. Both coatings were biocompatible, while Ti-3YSZ demonstrated the most significant cell area of 2630 μm2 (p ≤ 0.05) and the significantly highest, 66.75 ± 4.91, focal adhesions (FAs) per cell after 24 h (p ≤ 0.05). By contrast, PEEK and PMMA demonstrated the highest roughness and WCA and the lowest results for cellular response. Thus, Ti-3YSZ and Ti-LS2 surfaces might be promising for biomedical applications.


Introduction
Stable and healthy surrounding tissues define the success of a dental implant [1,2]. The importance of bone tissue is not questionable from a biomechanical point of view [3][4][5], but the soft tissue is of biological significance [6]. Currently, the prevention of peri-implant diseases is attracting increasing attention, and according to tendencies of the latest studies, the relevance of soft tissue is becoming more popular [7,8]. The function of the soft tissue barrier is protective and prevents infectious pathogens from entering and spreading, so one of the essential criteria for implant success is peri-implant soft tissue sealing [9]. Moreover, soft tissue acts not only as a physical but also as a biological barrier, which has immune response capacity against microorganisms [10,11].

Surface Preparation
To achieve identical conditions, polishing was performed using a rotary machine (Holzmann metal lathe ED3000ECO, Maschinenhandel Gronau Inc., Gehren, Germany). The surface of all specimens was polished with decreasing coarseness of water-resistant silicon carbide (SiC) abrasive paper P2000, P2500, P3000, P4000 (Starcke GmbH and Co. KG, Melle, Germany) and later with SiC abrasive pad P5000 (Trizact™, 3M Company, St. Paul, MN, USA) under water cooling/washing. The final polishing was completed with a diamond polishing paste and a natural brush (Zirkopol, Feguramed GmbH, Odenwald, Germany). The polishing cycle for each coarseness was performed for 60 s at 3000 rpm.

Preparation of Stabilized Zirconium Oxide (3YSZ) and Lithium Disilicate (LS2) Sols
The coatings were prepared using the sol-gel process, initially preparing an alkoxide solution of the precursors, followed by spin-coating them on the polished titanium alloy substrate. For the preparation of 3YSZ (by mol% Y 2 O 3 ) sol, 0.0073 mol of yttrium nitrate hexahydrate and 0.115 mol of ZIP were dissolved in dry 2-propanol solution stabilized with 0.0345 mol of AcAc. To initiate a hydrolysis reaction, 0.23 mol of H 2 O was added in a final step.
For the preparation of lithium disilicate, LiOMe sol 0.01 mol and 0.01 mol TMOS were dissolved in dry methanol, then 0.015 mol of H 2 O was added in a final step. Both solutions were kept at room temperature (RT) for 24 h, and after that, obtained sols were ready for the spin-coating procedure.

Deposition of 3YSZ and LS2 Coatings
Prior to the coating procedure, prepared sols were filtered through a 0.2 µm nylon membrane filter. Ti substrates were cleaned with a non-ionic surfactant solution (RBS Neutral T, Carl-Roth) in an ultrasonic bath (Sonorex, BANDELIN electronic GmbH and Co., Berlin, Germany), then thoroughly washed with deionized H 2 O and 2-propanol and allowed to dry in air. To prepare the desired coating, 50 µL of the corresponding sol was placed on a Ti substrate and spin-coated up to 2000 rpm. Prepared 3YSZ and LS2 coatings were placed in a muffle furnace (SNOL 13/1100, Umega group, Utena, Lithuania) and heated at 600 • C for 2 h. After this step, prepared coatings were evaluated using x-ray diffraction (XRD) and ready for other experiments. A commercial x-ray diffractometer (Rigaku SmartLab, Rigaku Corporation, Tokyo, Japan) was used for the XRD analysis of deposited coatings. To achieve higher sensitivity, spectra were recorded by employing grazing angle geometry, 0.01 deg. step size, and 1 deg/min scan speed. For evaluation of coating thickness, several coatings were applied to a 10 × 10 mm silicon (100) wafer. All coating procedures were the same as for the Ti.

Surface Morphology
Surface morphology and roughness parameters of specimens were characterized by using a scanning electron microscope (SEM) (Hitachi SU-70, Hitachi, Ltd., Chiyoda, Tokyo, Japan) and atomic force microscope (AFM) (Agilent 5500 AFM/SPM, Agilent Technologies, Palo Alto, CA, USA). Data on the means of the surface topography and roughness average (S a ) were collected by randomly scanning each sample in three areas. AFM topographic images were recorded in contact mode using a silicon probe with a 10 nm tip radius. Each image was recorded in 256 × 256 pixel size. For evaluation of coating thickness, Si-coated samples were broken in half, and the thickness of sample cross-sections was measured by SEM.

Water Contact Angle and Surface Free Energy
Specimens were all cleaned in 2-propanol and bi-distilled water using sonication. They were then dried for 2 h in a vacuum oven at 50 • C. An optical tensiometer (CAM 200, KSV Instruments Ltd., Helsinki, Finland) was used to measure the WCA and SFE. For measurement purposes, bi-distilled water was used in a heavy phase. To measure the WCA of the solvent droplet, a modest drop (4-7 µL) of a particular liquid was placed on a tested substrate at RT. The SFE was calculated using the Owens-Wendt method [30].

Biocompatibility Evaluation
To determine the biocompatibility of specimens, 3-[4,5-dimethylthiazole-2-yl]-2,5-diphenyltetrazolium bromide (MTT) assay was used. It was conducted by seeding HGF-1s (3 × 10 4 cell/cm 2 ) in tissue culture plates. After 24 h, samples were placed on an HGF-1 monolayer and were incubated for a further 24 h. After reaching a predetermined time point, the growth media was removed, and specimens were rinsed once with phosphate-buffered saline (PBS). Furthermore, in each sample, 0.2 mg/mL MTT (MTT, Sigma-Aldrich Inc., St. Louis, MO, USA) solution was added in PBS and incubated for 1 h at 37 • C in a humidified atmosphere consisting of 5% CO 2. MTT solution was subsequently discarded, and formed formazan was dissolved in dimethyl sulfoxide (DMSO) (dimethyl sulfoxide, Sigma Aldrich, Dorset, UK) by incubating with mild shaking (25 rpm) at RT for 10 min. Using a microplate spectrophotometer (Varioskan Flash, Thermo Fisher Scientific, Inc., Waltham, MA, USA, the absorbance was evaluated at 570 nm. A specimen without contact with cells was attributed as a background value.

Cellular Adhesion
A quantitative and qualitative assessment was used to determine cell adhesion efficiency and adhesion strength on tested specimens by applying F-actin staining and FA visualization. To achieve this, HGF-1s were cultivated (15 × 10 3 cell/cm 2 ) for 2 and 24 h on tested surfaces. After specific time points that were previously determined, specimens were fixed with 4% paraformaldehyde (Carl Roth, GmbH, Karlsruhe, Germany) in PBS (Gibco, Thermo Fisher Scientific, Waltham, MA, USA). This was achieved at RT for 15 min with mild agitation (25 rpm). Samples were then rinsed twice in PBS with 0.05% Tween-20 (Sigma-Aldrich Co., St. Louis, MO, USA), permeabilized with 0.2% Triton X-100 (Sigma-Aldrich Co., St. Louis, MO, USA) in PBS for 5 min at RT with 25 rpm agitation, and then washed twice with 0.05% Tween-20 solution. After that, specimens were blocked for 30 min with 3% bovine serum albumin (BSA; AppliChem GmbH, Darmstadt, Germany) and 10% FBS prepared in PBS. After the subsequent blocking procedure, samples were incubated with primary mouse anti-vinculin antibody (1:50; Merck Millipore, Burlington, MA, USA) at RT for 1 h in blocking solution while using gentle shaking at 25 rpm. Soon after, specimens were washed three times for 5 min with a 0.05% Tween-20 solution. They were then incubated with secondary goat anti-mouse Alexa Fluor 488-conjugated antibodies (Invitrogen, Carlsbad, CA, USA) and tetramethyl rhodamine iso-thiocyanate (TRITC)-conjugated phalloidin (1:500; Merck Millipore, Carlsbad, CA, USA) in PBS in a darkened environment at RT for 1 h with 25 rpm agitation. After that, the samples were washed three more times with PBS for 5 min at RT and stained with 12.5 µg/mL 4 ,6-diamidino-2-phenylindole (DAPI; Merck Millipore, Carlsbad, CA, USA) solution in PBS for 5 min in the dark at RT with 25 rpm agitation. Then, specimens were washed three times with PBS for 5 min at RT and visualized using a confocal laser scanning microscope (CLSM) (Leica SP5 TCS, Leica Microsystems, Wetlzer, Germany). Using the image processing program ImageJ (1.8.0_112) (Wayne Rasband, National Institute of Mental Health, Bethesda, MD, USA), measurement of the cells' surface area and counting of FAs within the cells on different samples were performed. Then, the quantitative differences in cells' adhesion efficiency and strength were determined.

Statistical Analysis
The specimen size was calculated using G*Power 3.1 software (Heinrich Heine University, Dusseldorf, Germany) before the study. The data were processed using the GraphPad Prism 8 software package (San Diego, CA, USA). The roughness, WCA, SFE, relative HGF-1 cell counts, and adhesion area measurements were analyzed based on standard deviations and means. All data sets were checked for their normality using the D'Agostino-Pearson omnibus test and then analyzed by one-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test. Additionally, the relationship between experiments was evaluated based on Pearson correlations. A p-value of ≤0.05 was considered statistically significant.

Phase Composition
The formation of ceramic 3YSZ and LS2 glass-ceramic coatings was confirmed by XRD ( Figure 1). In both cases, single-phase crystalline coatings were obtained. Some additional diffraction peaks of the Ti alloy substrate were visible in both patterns. In the case of the 3YSZ coating, observable peaks in the XRD pattern correspond to the tetragonal ZrO 2 phase. Visible diffraction peaks on XRD patterns of the LS2 coating can be associated with the orthorhombic lithium disilicate phase. The LiSiO 3 phase commonly found in lithium silicate glass-ceramics was not detected in this case. It is important to note that the diffraction peak on LS2-coated Ti corresponding to the (040) plane was more intense than expected, which could indicate the formation of slightly texturized lithium disilicate crystal orientation. These findings on both surfaces established that coatings of monophase 3YSZ and LS2 were achieved by the sol-gel method. Other groups were excluded due to lack of chemical modification of polished surfaces.

Surface Characterization
The surface roughness of each group was measured after preparation. The Sa means were used to express the measurements. The Sa of all surfaces was below the micro level (Table 1). After polishing, the lowest roughness was found on the ZrO2 surface (Sa = 5.53 ± 0.21 nm), and the highest on PMMA (Sa = 65.23 ± 2.41 nm). Sa values were significantly different between all groups (p < 0.0001) ( Table 2). The coatings of both bioceramics reduced the roughness of the polished titanium alloy substrate surface. Compared to the surface of the polished titanium alloy (Sa = 17.67 ± 0.35 nm), after the deposition of the 3YSZ coating (Sa = 16.61 ± 0.52 nm), the decrease in surface roughness was not statistically significant (p > 0.05), but the LS2 coating (Sa = 9.61 ± 0.66 nm) decreased the roughness significantly (p < 0.0001). Table 1. Descriptive statistics of all groups' surface physicochemical experiments (mean ± standard deviation): roughness (nm), water contact angle (WCA) (deg.), dispersive surface free energy (SFE_dispersive) (mN/m), polar surface free energy (SFE_polar) (mN/m), total surface free energy (SFE_total) (mN/m). Data of titanium alloy (Ti), yttria-stabilized zirconium oxide coating on titanium alloy (Ti-3YSZ), lithium disilicate coating on titanium alloy (Ti-LS2), yttria-stabilized zirconium oxide, polyether ether ketone composite (PEEK) and poly(methyl methacrylate) (PMMA) groups.

Surface Characterization
The surface roughness of each group was measured after preparation. The S a means were used to express the measurements. The S a of all surfaces was below the micro level (Table 1). After polishing, the lowest roughness was found on the ZrO 2 surface (S a = 5.53 ± 0.21 nm), and the highest on PMMA (S a = 65.23 ± 2.41 nm). S a values were significantly different between all groups (p < 0.0001) ( Table 2). The coatings of both bioceramics reduced the roughness of the polished titanium alloy substrate surface. Compared to the surface of the polished titanium alloy (S a = 17.67 ± 0.35 nm), after the deposition of the 3YSZ coating (S a = 16.61 ± 0.52 nm), the decrease in surface roughness was not statistically significant (p > 0.05), but the LS2 coating (S a = 9.61 ± 0.66 nm) decreased the roughness significantly (p < 0.0001). Table 1. Descriptive statistics of all groups' surface physicochemical experiments (mean ± standard deviation): roughness (nm), water contact angle (WCA) (deg.), dispersive surface free energy (SFE_dispersive) (mN/m), polar surface free energy (SFE_polar) (mN/m), total surface free energy (SFE_total) (mN/m). Data of titanium alloy (Ti), yttria-stabilized zirconium oxide coating on titanium alloy (Ti-3YSZ), lithium disilicate coating on titanium alloy (Ti-LS2), yttria-stabilized zirconium oxide, polyether ether ketone composite (PEEK) and poly(methyl methacrylate) (PMMA) groups.  The morphology of bioceramic coatings in Ti-3YSZ and Ti-LS2 was evaluated by SEM. In both images, a uniform surface structure was observed. In the image (Figure 2a), small cracks in the 3YSZ coating were detected, while the LS2 layer was without cracks and regular ( Figure 2b). All detected cracks were less than 1 µm in size. 3YSZ coatings consist of a uniform layer of nano-sized grains, whereas LS2 contains micron-sized domains of differently orientated nano-sized grains and/or glassy phase. The grain size of both ceramic coatings was less than 10 nm. The determination of the exact crystallite size was limited by instrument resolution. The determined coating thickness of 3YSZ and LS2 was 182 nm and 159 nm, respectively (Figure 2c,d). The morphology of bioceramic coatings in Ti-3YSZ and Ti-LS2 was evaluated by SEM. In both images, a uniform surface structure was observed. In the image (Figure 2a), small cracks in the 3YSZ coating were detected, while the LS2 layer was without cracks and regular ( Figure 2b). All detected cracks were less than 1 µm in size. 3YSZ coatings consist of a uniform layer of nano-sized grains, whereas LS2 contains micron-sized domains of differently orientated nano-sized grains and/or glassy phase. The grain size of both ceramic coatings was less than 10 nm. The determination of the exact crystallite size was limited by instrument resolution. The determined coating thickness of 3YSZ and LS2 was 182 nm and 159 nm, respectively (Figure 2c,d).

Water Contact Angle and Surface Free Energy
All polished surfaces of commercially available abutment materials demonstrated relatively high WCA (Table 3), which means higher hydrophobicity. The WCA means of all groups were significantly different (p ≤ 0.05). The lowest difference was between the Ti (60.21 ± 1.58 • ) and ZrO 2 (57.32 ± 1.86 • ) groups (p = 0.0264), while the difference among the remaining groups was even higher (p < 0.0001). WCA measurements established the influence of coatings because of this increased hydrophilicity of the titanium alloy surface. The WCA values of Ti-3YSZ (32.58 ± 1.94 • ) and Ti-LS2 (25.71 ± 1.33 • ) were statistically lower than the Ti alloy itself (60.21 ± 1.58 • ), as well as others (p < 0.0001). The highest hydrophilicity was detected in the Ti-LS2 group.  (Table S2). A strong negative correlation between WCA and total SFE was established (r = −0.9348) (Table S3). Table 4. Multiple comparisons of surface total SFE (mN/m) among groups (n = 10 for each group) by Tukey's test and significant difference of means at the p ≤ 0.05 level: **** p < 0.0001; ** p = 0.007; ** # p = 0.0071. Non-significant (ns) p > 0.05.

Biocompatibility Evaluation
The biocompatibility of the newly formed coatings on the Ti alloy surface was evaluated and compared with the biocompatibility of conventional implant prosthetic materials. The results were standardized by the number of cells obtained on the polished Ti alloy control surface. After evaluating the growth of relative HGF-1 cell count (RCC), it was found that all materials used for the assay were biocompatible (Figure 3a). The highest RCC (i.e., the best biocompatibility) was observed for the ZrO 2 (1.08 ± 0.05) surfaces. Moreover, no statistically significant differences were found between polished Ti (1.00 ± 0.17) and Ti-3YSZ (0.94 ± 0.12) (p > 0.05) or Ti (1.00 ± 0.17) and Ti-LS2 (0.87 ± 0.11) (p > 0.05) specimens. The lowest biocompatibility was found in the PMMA (0.65 ± 0.11) group, which was significantly different from others (p < 0.0001), except PEEK (0.79 ± 0.06) (p > 0.05).

Biocompatibility Evaluation
The biocompatibility of the newly formed coatings on the Ti alloy surface was evaluated and compared with the biocompatibility of conventional implant prosthetic materials. The results were standardized by the number of cells obtained on the polished Ti alloy control surface. After evaluating the growth of relative HGF-1 cell count (RCC), it was found that all materials used for the assay were biocompatible (Figure 3a). The highest RCC (i.e., the best biocompatibility) was observed for the ZrO2 (1.08 ± 0.05) surfaces. Moreover, no statistically significant differences were found between polished Ti (1.00 ± 0.17) and Ti-3YSZ (0.94 ± 0.12) (p > 0.05) or Ti (1.00 ± 0.17) and Ti-LS2 (0.87 ± 0.11) (p > 0.05) specimens. The lowest biocompatibility was found in the PMMA (0.65 ± 0.11) group, which was significantly different from others (p < 0.0001), except PEEK (0.79 ± 0.06) (p > 0.05).

Cell Adhesion Area
HGF-1 adhesion on tested specimens was determined by calculating cell surface areas 2 and 24 h post-seeding. The HGF-1 surface area was visualized by cell F-actin staining. To evaluate cell adhesion differences that occurred on tested specimens, a qualitative and quantitative assessment of the images obtained by CLSM was performed. The results of HGF-1 adhesion showed that all tested

Cell Adhesion Area
HGF-1 adhesion on tested specimens was determined by calculating cell surface areas 2 and 24 h post-seeding. The HGF-1 surface area was visualized by cell F-actin staining. To evaluate cell adhesion differences that occurred on tested specimens, a qualitative and quantitative assessment of the images obtained by CLSM was performed. The results of HGF-1 adhesion showed that all tested surfaces were attractive for cell attachment. Changes in the HGF-1 surface area (2 and 24 h after seeding) showed that the cells were growing on all specimens, as the surface areas increased during the time (Figures 3b and 4).  Moreover, 2 h after seeding, the largest HGF-1 surface area (the best cell adhesion) was found on the Ti-3YSZ specimens (1789 μm 2 ) (p < 0.0001), and the lowest cell area was determined on both polymeric surfaces (PEEK (581.8 μm 2 ) and PMMA (648.6 μm 2 )), but these two were not significantly different. Furthermore, after 24 h, the highest cell adhesion area was observed on the Ti-3YSZ (2630 μm 2 ) specimens, and the lowest was determined on polymeric PEEK (1534 μm 2 ) and PMMA (1847 μm 2 ) surfaces as well as Ti-LS2 specimens (1888 μm 2 ).
It is also important to mention that compared to 2 h, after 24 h, the cell attachment area increased in all groups except the Ti-LS2 group (1507 μm 2 after 2 h and 1888 μm 2 after 24 h), where the increment of adhesive cell area was not statistically significant (p > 0.05). Moreover, even though the increase of cell surface area on PEEK and PMMA samples after 24 h was statistically significant (p < 0.0001), it reached only a similar surface area size as on Ti-3YSZ after 2 h (Figure 3b).

Cell Focal Adhesion (Density Distributions)
Immunohistochemically stained FAs were visualized ( Figure 4) and quantified ( Figure 5) to evaluate cell adhesion strength. Obtained quantitative FA results were expressed as the density distribution function of cells and number of focal adhesions 2 and 24 h after seeding. Also, means of FAs per cell were provided for statistical data comparison (Tables S4,S5 and Table 5). Moreover, 2 h after seeding, the largest HGF-1 surface area (the best cell adhesion) was found on the Ti-3YSZ specimens (1789 µm 2 ) (p < 0.0001), and the lowest cell area was determined on both polymeric surfaces (PEEK (581.8 µm 2 ) and PMMA (648.6 µm 2 ), but these two were not significantly different. Furthermore, after 24 h, the highest cell adhesion area was observed on the Ti-3YSZ (2630 µm 2 ) specimens, and the lowest was determined on polymeric PEEK (1534 µm 2 ) and PMMA (1847 µm 2 ) surfaces as well as Ti-LS2 specimens (1888 µm 2 ).
It is also important to mention that compared to 2 h, after 24 h, the cell attachment area increased in all groups except the Ti-LS2 group (1507 µm 2 after 2 h and 1888 µm 2 after 24 h), where the increment of adhesive cell area was not statistically significant (p > 0.05). Moreover, even though the increase of cell surface area on PEEK and PMMA samples after 24 h was statistically significant (p < 0.0001), it reached only a similar surface area size as on Ti-3YSZ after 2 h (Figure 3b).

Cell Focal Adhesion (Density Distributions)
Immunohistochemically stained FAs were visualized ( Figure 4) and quantified ( Figure 5) to evaluate cell adhesion strength. Obtained quantitative FA results were expressed as the density distribution function of cells and number of focal adhesions 2 and 24 h after seeding. Also, means of FAs per cell were provided for statistical data comparison (Tables S4 and S5 and Table 5).  The images obtained by CLSM showed that 2 h after seeding, cells began to form FAs on all examined surfaces ( Figure 4); moreover, the quantitative results revealed the differences between them ( Figure 5). It was observed that on polymeric PEEK and PMMA surfaces, most of the cells did not form FAs (Figure 5d,f) or formed only a small amount of them: PEEK-5.15 ± 0.74 FAs/cell, PMMA-5.76 ± 0.93 FAs/cell. The projections of both histograms had similar configurations and revealed that the surfaces of both polymers were not attractive for an early cell attachment. A wide range of FA distribution was observed on the Ti-3YSZ surface (Figure 5c). In this group, cells were relatively uniformly distributed with a similar quantity of 0 to 100 FAs. Additionally, the concentration of cells without FAs on Ti-3YSZ was lower than on polymers but higher than on other surfaces, while cells with the highest FA counts after 2 h were also the highest. No trend in cellular concentration by FAs was observed in this group, while its mean was 28.00 ± 3.76 FAs/cell. In the remaining groups, FA numbers in cells increased respectively: Ti-LS2 FAs ranged from 15 to 20 (Figure 5e), ZrO 2 FAs~35 (Figure 5b), and Ti FAs~50 (Figure 5a). However, both the ZrO 2 and Ti alloy histograms replicated the portion of the Ti-3YSZ plot showing similar concentrations of the most FA-rich cells, and the concentration of the FAs within the cells in all three groups were very similar. Comparing FAs/cell, the statistical difference did not exist between these three groups (p > 0.05) (Table S4).
Twenty four-hour results revealed that the highest number of FAs were formed in HGF-1 grown on the Ti-3YSZ surface (Figure 5c). Two peaks of cell concentrations were detected in this group. The highest cell concentration was~45 FAs per cell, but even a wave of 75-80 FAs per cell was observed. The mean on Ti-3YSZ group was 66.75 ± 4.91 FAs/cell, and it was statistically the highest (p ≤ 0.05) (Table S5). By contrast, the PEEK group showed persistent high concentrations of cells that formed comparable weak adhesions after 24 h (Figure 5d). However, compared to the histogram after 2 h, two peaks of FA concentrations were already observed after 24 h; cells that began to attach to the surface (cells with a relatively small number of FAs) and cells with relatively strong adhesions (cells with a high number of FAs (~60)) were found. On the PEEK surface, cells behaved similarly as on the PMMA group: Higher concentrations of cells formed weak adhesions (~20 FAs per cell), but another, relatively lower, portion of cells demonstrated strong attachment (~75-80 per cell) (Figure 5f). The FA numbers in cells grown on Ti-LS2 also increased after 24 h, and the maximum cell concentration was found containing~25 FAs (Figure 5e). Thus, PEEK and PMMA sample groups induced weak and slow HGF-1 adhesion and comparing FAs/cell, the statistical difference did not exist between these three groups (p > 0.05) (Table S5). On the ZrO 2 surface, cell adhesion during the time also improved: The maximum FA number determined in HGF-1, 24 h after seeding, was~45 (Figure 5b). Comparing both periods (2 and 24 h), all tested groups showed an increment of FAs, except the Ti alloy surface group. There were no changes in FA numbers on the latter sample, comparing 2 and 24 h.

Discussion
This study reveals new scope for Ti alloy prosthetic part improvements. Titanium alloys are widespread in implant dentistry and its prosthetics, as well as in orthopedic medicine, because of their relatively bioinert and mechanical properties [31]. Despite that, the usage of metal alloys is limited mainly because of expected corrosion, allergies [32], and aesthetic issues [33]. Moreover, the titanium metal-based surface is sensitive to mechanical damage [34], and the abutment can be scratched during oral hygiene, resulting in retention for plaque accumulation [35]. The latest findings of ceramic prosthetic materials revealed their better bioinertia and peri-implant soft tissue integration [15] and higher surface hardness [36], but their use is relatively limited due to bending sensitivity [37] and brittleness [38,39].
The conventional layering ceramics were successfully applied for porcelain fused to metal restorations for aesthetic reasons. Unfortunately, the fusing on titanium alloys is still quite delicate on a count of lower adhesion strength between the titanium core and ceramic [40]. The weak fusing between these surfaces is mostly related to entrapped sand particles on titanium after sandblasting [41] and thick oxide due to alloy reactivity [42]. Surface improvements by thin films or coatings open up new fields of material improvements and applications. In this study, newly generated bioceramic coatings were expected to improve cell-surface interactions. Similar studies have achieved promising results with coatings of other bioceramics [43,44]. For this study, unique coatings of yttria-stabilized zirconium oxide and lithium disilicate bioceramics using the sol-gel method were selected for surface improvements of a titanium alloy substrate. The bioceramic coating on Ti alloys might be a solution, which allows for combining the mechanical properties of the Ti substrate and the advantages of bioceramics. The substrate isolation by a coating of bioceramics reduces its corrosion capacity and establishes ceramic-tissue contact. The chosen ceramics are widely used for prosthetic purposes, and their biocompatibility and durability have already been proven [45]. The sol-gel method was chosen for coating because of its simplicity, cost-effectiveness, and successful application with other substrates and bioceramic materials [46].
By contrast with conventional semi-soft YSZ milling blanks, the high temperature sintering to stabilize the crystalline structure by yttrium was unnecessary in a sol-gel strategy. The crystalline form of zirconium particles is not used for sols-gels, where elements are mixed homogeneously at the molecular level, therefore phase formation temperatures are lower. In this synthesis, the temperature was used for a few reasons: to evaporate solvents, to burn-out the organic components, and to open the pores of the substrate surface for molecules of the coating. There are a number of studies showing that tetragonal YSZ can be obtained from as low as 500 • C [47,48], and Li 2 Si 2 O 5 at 550 • C [49][50][51]. Therefore, the temperature was lower than for sintering of conventional powder-based YSZ ceramic and was selected regarding the transformation phases of the titanium alloy.
The success of both coatings was confirmed by XRD, which showed chemical deposition of 3YSZ and LS2 on the surface of the polished titanium alloy. Additionally, morphological findings established the distribution of coatings. The layer of 3YSZ was not uniform and showed submicron cracks, whereas the LS2 coatings did not contain any cracking. The nature of the phase composition might influence such behavior: 3YSZ coating is purely crystalline in nature, whereas LS2 generally is glass-ceramic. Therefore, the crystallization mechanism and kinetics are different for these coatings. However, the cracking of the 3YSZ coating could not be related to phase transformation from tetragonal to monoclinic [52], as it was in another study where a different methodology was generated for YSZ coating. In this study, the tetragonal phase was confirmed by XRD, and the only explanation of volume expansion during the cooling might be accepted as a probable factor. It is also important to note that XRD revealed no titanium oxide interlayer; therefore, in both cases, no unwanted oxidation of the Ti alloy substrate took place, which in some cases induces cracking of the upper ceramic film. In fact, this shows that deposited coatings effectively isolate the Ti alloy surface, thus decreasing the probability of releasing allergy-promoting ions such as Al or V.
The AFM was selected as one of the most suitable instruments for describing the surface morphology of biomaterials [53], and the roughness of surface was expressed by a commonly used Sa parameter [54]. Results of surface roughness showed tendencies that additional coatings decreases the roughness by covering and filling remaining scratches and irregularities of the surface. This was statistically confirmed in the Ti-LS2 group (roughness significantly from S a = 17.67 ± 0.35 nm to S a = 9.61 ± 0.66 nm (p < 0.0001)), whereas Ti-3YSZ revealed only a slight decrease (to S a = 16.61 ± 0.52 nm (p > 0.05)), which could also be related to different crystallization of the coating. Unfortunately, both coatings did not demonstrate such a low surface roughness as on the highly polished ZrO 2 surface (S a = 5.53 ± 0.21 nm). In this study, the established different roughness between Ti and ZrO 2 might be related to the origin of the material, and the surface hardness influenced polishing efficiency. The idea of coating might be a prospective possibility to decrease the surface roughness of metal alloys and improve their usage for implant prosthetics and other biomedical implants. However, the thickness of coatings limited their effectiveness to change the substrate-dependent surface roughness.
The experimental coatings used in this study significantly changed the WCA of the surface. After coating, the water contact angle decreased significantly from 60.21 ± 1.58 • to 32.58 ± 1.94 • in the Ti-3YSZ group and to 25.71 ± 1.33 • in the Ti-LS2 group (p < 0.0001), thus surfaces became significantly hydrophilic after the deposition of both coatings. Also, in both groups, the total surface free energy increased from 41.81 ± 2.01 mN/m to 60.03 ± 0.57 mN/m in the Ti-3YSZ group and 65.70 ± 1.00 in the Ti-LS2 group (p < 0.0001). The polar surface free energy increased from 7.37 ± 1.62 mN/m to 25.14 ± 0.41 mN/m in the Ti-3YSZ group and 33.39 ± 0.59 mN/m in the Ti-LS2 group (p < 0.0001). This is be related to changes in surface chemistry and morphology. These coatings are composed of nanosized crystallites that increase the total surface area of the substrate, and in the case of Ti-LS2, Li+ ions were exposed to the surface, which could significantly increase surface polarity as well as hydrophilicity. However, it also must be biocompatible.
The high biocompatibility of experimental coatings was confirmed. It was similar to that of conventional prosthetic materials or even higher and enabled the favorable application of this technology for biomaterials and biomedical applications for future studies. No cytotoxic effects that could be attributed to the new materials and their chemical composition or morphology have been identified. On the other hand, no improvement in biocompatibility was observed with coatings, but it also established the stability of the coatings' composition. Nevertheless, according to the results of biocompatibility polymeric materials, PEEK and PMMA showed the lowest rates, establishing that these types of materials were less favorable for direct contact with living tissues comparing to inorganic materials. This might be related to the origin of materials and their chemical composition, for example, the remaining free monomer in PMMA [55]. However, it might also be determined by the hydrophobic character of these materials. Thus, materials of inorganic origin exhibited better biocompatibility.
Surface morphology is another relevant factor in determining cell behavior [56]. Also, a negative correlation between the biocompatibility and surface roughness was confirmed (r = −0.7036; p < 0.0001) in this study. In all groups with commercially available prosthetic materials, an obvious relationship between the surface roughness and the relative cell count was visible. When the S a decreased, the RCC increased, so the increase of surface smoothness possibly alters the direct interface between the cells and the substrate. By contrast, this tendency was not confirmed with coated surfaces, where the roughness was comparably low. This confirmed that nanoparticles change surface properties as well as interact with living cells [57].
Furthermore, coatings' improved surface attraction for cell adhesion was also confirmed. The increased surface area occupied by a cell showed the dynamics of cells' attachment on surfaces. It is known that the faster a cell occupies a larger surface area, the more attractive this substrate is for cell attachment [58]. The largest surface area of HGFs after 2 and 24 h was observed on the Ti-3YSZ surface (1789 µm 2 and 2630 µm 2 , respectively), so this surface promoted the best and fastest cell adhesion compared to other tested materials. The rapid quantitative cell response and comparable high surface area were also established on Ti-LS2 surface after 2 h: 1507 µm 2 . Unfortunately, the cell expansion on this surface was insignificant after 24 h (1888 µm 2 ), despite the lowest roughness and WCA values. However, the qualitative examination of the stained cell on Ti-LS2 after 24 h provided more promising results. The obvious elongation of HGFs and concentrated actin and vinculin at both opposite tails of the spindle was visible. This cellular elongation enables the assumption that on this substrate, cells proliferated faster compared to others [59,60], and because of intensive proliferation, they de-attached from the surface, showing a lower adhesive area as well as a lower amount of FAs. However, for a more comprehensive evaluation, the motility of cells should be observed. Thus, promising results of the adhesive process on the Ti-LS2 substrate were established only after 2 h.
It is interesting that PEEK and PMMA surfaces were less attractive for cells. The results of cell surface area measurements showed that HGF-1 adhesion was slowest on these samples, and the lowest number of formed FAs within the cell was determined in these two polymeric groups. This may be due to the relatively hydrophobic character of surfaces, and hydrophobicity may decrease the attractiveness of surfaces for cell adhesion [61] and soft tissue integration [62].
Cell FA formation during the adhesion process also indicates that the substrate is attractive for cell attachment, and the number of FAs formed in the cell indicates the strength of cell-surface interaction [19,58,63]. In this study, HGF-1, grown on Ti-3YSZ samples (which demonstrated a smooth surface (S a = 16.61 ± 0.52 nm) and hydrophilic character (32.58 ± 1.94 • WCA)), formed the highest numbers of FAs compared to other tested specimens after 24 h (66.75 ± 4.91 FAs/cell (p ≤ 0.05)). In contrast, FA densities in cells were exceptionally low on polymeric materials, PEEK and PMMA. The latter samples demonstrated the highest surface roughness; thus, the low density of FAs might be a result of that. Other studies have shown that the increase of FAs coincided with decreasing roughness [64][65][66]. On smooth surfaces, a closer cell-substrate contact could be achieved than on rough ones [67]. Furthermore, fibroblasts are rugophobic cells, which means they have adverse reactions to the surface roughness [68]. Moreover, PEEK and PMMA demonstrated more hydrophobic character compared to Ti-3YSZ. Surface hydrophobicity alters the adsorption of extracellular proteins, which in turn could result in weaker FAs [69,70]. Cell attachment on the surface is promoted only during cell response to the extracellular matrix, which is mainly formed by proteins [19]. Therefore, protein adsorption on the surface could be another critical factor for FA formation, and it is closely related to the surface WCA [67].
Thus, surface coatings of titanium alloys allow the physicochemical properties of the surface to be improved, leading to improved behavior of HGFs. Moreover, surface coatings can reduce the accumulation of bacterial plaque, increase surface hardness, and increase the resistance of implant abutment surface scratching [71]. However, these require additional studies. Moreover, the organization of HGFs' rich connective tissue around a regular titanium surface might be delayed 4-8 weeks at least [6,72,73]. The findings of this study established promising results of 3YSZ and LS2 coatings, which might be useful for in vivo studies for a faster sealing of the peri-implant mucosa. However, the results of this study remain limited, and additional studies of protein adsorption are necessary for a better explanation of cell behavior. Additionally, these novel coatings should be investigated with in vitro bacterial models as well as applied in vivo to determine their influence on clinical outcomes.

Conclusions
In this study, the novel coatings of 3YSZ and LS2 were achieved by the sol-gel method, while the LS2 coating was used for the first time on titanium alloy. Both of the novel coating materials might be useful for the improvement of the physicochemical properties of titanium alloys. All the investigated materials were biocompatible. The Ti-YSZ was the most suitable for cellular adhesion responses compared to other groups, as this surface demonstrated the most significant cell area (the best cell adhesion) 2630 µm 2 (p ≤ 0.05), and the significantly highest 66.75 ± 4.91 (p ≤ 0.05) focal adhesions (FAs) per cell after 24 h post-seeding. PMMA and PEEK showed the worst cellular behavior, demonstrating the highest roughness (S a = 65.23 ± 2.41 nm and S a = 38.00 ± 0.78 nm, respectively, p < 0.0001) and WCA (68.84 ± 2.65 • and 75.62 ± 2.46 • , respectively, p < 0.0001). Further studies are needed, but 3YSZ and LS2 coatings might be promising for biomedical applications.