Considering the Value of 3D Cultures for Enhancing the Understanding of Adhesion, Proliferation, and Osteogenesis on Titanium Dental Implants

Background: Individuals with pathologic conditions and restorative deficiencies might benefit from a combinatorial approach encompassing stem cells and dental implants; however, due to the various surface textures and coatings, the influence of titanium dental implants on cells exhibits extensive, wide variations. Three-dimensional (3D) cultures of stem cells on whole dental implants are superior in testing implant properties and were used to examine their capabilities thoroughly. Materials and methods: The surface micro-topography of five titanium dental implants manufactured by sandblasting with titanium, aluminum, corundum, or laser sintered and laser machined was compared in this study. After characterization, including particle size distribution and roughness, the adhesion, proliferation, and viability of adipose-derived stem cells (ADSCs) cultured on the whole-body implants were tested at three time points (one to seven days). Finally, the capacity of the implant to induce ADSCs’ spontaneous osteoblastic differentiation was examined at the same time points, assessing the gene expression of collagen type 1 (coll-I), osteonectin (osn), alkaline phosphatase (alp), and osteocalcin (osc). Results: Laser-treated (Laser Mach and Laser Sint) implants exhibited the highest adhesion degree; however, limited proliferation was observed, except for Laser Sint implants, while viability differences were seen throughout the three time points, except for Ti Blast implants. Sandblasted surfaces (Al Blast, Cor Blast, and Ti Blast) outpaced the laser-treated ones, inducing higher amounts of coll-I, osn, and alp, but not osc. Among the sandblasted surfaces, Ti Blast showed moderate roughness and the highest superficial texture density, favoring the most significant spontaneous differentiation relative to all the other implant surfaces. Conclusions: The results indicate that 3D cultures of stem cells on whole-body titanium dental implants is a practical and physiologically appropriate way to test the biological characteristics of the implants, revealing peculiar differences in ADSCs’ adhesion, proliferation, and activity toward osteogenic commitment in the absence of specific osteoinductive cues. In addition, the 3D method would allow researchers to test various implant surfaces more thoroughly. Integrating with preconditioned stem cells would inspire a more substantial combinatorial approach to promote a quicker recovery for patients with restorative impairments.


Introduction
Bone regeneration is limited in the elderly and individuals with pathologic disorders such as osteoporosis, diabetes, and metabolically linked diseases after trauma or surgery [1][2][3][4]. The need for restorative dental treatments is anticipated to increase as the administration of biological material, according to the ethical committee of the University of Udine (Par. N. 103/2011/Sper). The donors' lipoaspirate (15 mL) was obtained using the tumescent procedure, which involves infusing Klein's solution and then recovering the adipose tissue. After that, the adipose tissue was centrifuged at 1800× g for 15 min to remove the red blood cells. ADSCs were placed in Joklik modified alpha-MEM medium (Gibco, Carlsbad, CA, USA) supplemented with 400 U/mL collagenase type 2 (Wortington, Columbus, OH, USA) equal to two times the volume of lipoaspirates and incubated for 15 min at 37 • C under moderate shaking every 2-3 min. The solution was then centrifuged for 10 min at 1800× g. After that, the supernatant was discarded, and the pellet was resuspended and filtered through a 70 µm filter. ADSCs were cultured onto 100 mm dishes (2 × 10 6 cells per dish) in the presence of a proliferation medium composed of α-MEM with L-glutamine 2 mmol/L, 10% fetal bovine serum (FBS), insulin 10 mg/mL, dexamethasone 10 −9 mol/L, ascorbic acid 100 mM, epithelial growth factor 10 ng/mL, and gentamicin 50 ng/mL (all from Sigma-Aldrich, Saint Louis, MO, USA). To inhibit cell differentiation, ADSCs were kept semi-confluent, and 80% of the media was changed every three days. ADSCs were maintained at 37 • C, 5% CO 2 , and a humidified atmosphere in the proliferation medium. A bland detaching solution, namely CTC (collagenase 20 U/mL, trypsin 0.75 mg/mL, 2% heat-inactivated dialyzed chicken serum (Gibco) in calcium, and magnesium Hank's balanced salt solution (HBSS)), was used to detach ADSCs. Cells between passages 1 and 6 (P1 and P6) were used in the experiments [24,26].

Characteristics of Dental Implants
Five different titanium dental implants, 1: sandblasted with aluminum oxide and acid etched (Al Blast, N = 9); 2: laser-sintered (Laser Sint, N = 9); 3: sandblasted with titanium dioxide spheres and fluoride treatment (Ti Blast, N = 9); 4: laser-machined (Laser Mach, N = 9); 5: sandblasted with corundum and different acid etchings (Cor Blast, N = 9), have been tested and compared throughout this study. The implants had a tronco-conical shape with dimensions ranging from 3.6 to 4 mm in diameter and 8 to 9 mm in length and were chosen for their wider market representation, which could be attributed to their higher difference in roughness degree and surface texture compared to our previous research [24]. The surface of Al Blast, which is made of grade 4 titanium, has a uniform finish due to a temperature-controlled process of large-grit sandblasting with aluminum oxide (Al 2 O 3 , around 0.5 mm) and acid etching. According to the manufacturer, no surface debris is left from the blasting and etching procedure, which creates a uniform surface with a very high protein-binding capability before being sterilized by a plasma cold reactor and packaged under clean, low-germ conditions. The Laser Sint implant, made of Ti 6 Al 4 V, is created by sintering metal powder nanoparticles with one another and against the implant surface using a selective laser melting technique via a computer-controlled laser beam to produce a complex implant surface. The final texture is an intercommunicating surface that interacts with the host bone and permits osteoblastic cells to enter the implant body. The implant is dried and made ready for use after being cleaned with acetone, ethanol, and distilled water in an ultrasonic cleaner.
The Ti Blast implant is made of Ti 6 Al 4 V alloy and has a moderately rough surface resulting from two subtractive, sequential manufacturing processes. The titanium dioxide sandblasting spheres (0.25-0.75 mm) give a microscale surface roughness, and the subsequent etching with hydrofluoric acid shapes the nanostructure of the implant. The implant is dried and made ready for use after being sterilized and packaged in an unspecified solution.
The Laser Mach implant, made of Ti 6 Al 4 V alloy, has a sequence of finely carved, cell-sized networks manufactured through a laser subtractive process using a high-energy and highly concentrated laser beam from a CO 2 laser source ranging from 10 W up to 30 W. The implant is dried, sterilized by γ irradiation, then packed and made ready for use.
The Cor Blast implant, made of extra-low interstitial Ti 6 Al 4 V alloy, is etched with different acids (HCl and H 2 SO 4 ) at high temperatures and then sandblasted with corundum (Al 2 O 3 , a naturally occurring mineral of aluminum oxide, 0.25-0.5 mm) at high pressure (up to 5 bar) to provide a highly homogeneous surface. The implant has a high level of cleanliness, which is due to the decontamination procedure in a cold plasma reactor.

Dental Implants Surface Characterization
Morphological analyses were carried out using a scanning electron microscope, Quanta250 (SEM; FEI, Hillsboro, OR, USA), in high vacuum and secondary electron mode with a tension of 30 kV; the operational distance was set in order to obtain the proper magnification. Aluminum stubs covered with carbon double-sided tape were used to secure the samples. The samples were then gold sputtered with a Sputter Coater K550X (Emitech; Quorum Technologies Ltd., Laoughton, UK).
The implants were then subjected to acquisitions at higher magnifications (40, 100, 800, and 1500×), and three principal profiles were obtained for each implant perpendicular to the implant longitudinal axis. To evaluate the surface micro-morphology, average size, distance between the particles, and number of particles, SEM photomicrographs (N = 3 per implant surface, magnification 1500×, considering an area of 100 × 100 µm) were taken between two threads and, after binarization and thresholding, were analyzed using ImageJ and the nearest distances (ND) plugin (http://imagej.nih.gov, USA, accessed on 3 March 2023) [27].
Although initially created for different purposes and highlighting the measurement uncertainty and noise that can be caused by errors induced by this method and by the roughness estimation process itself [28][29][30], the plugin SurfCharJ was used for approximately determining the roughness parameters from the SEM microphotographs [20]. Therefore, roughness was assessed according to ISO 4287/2000, measuring the following parameters: Ra (arithmetical mean deviation), Rq (root mean square deviation), Rsk (skewness of the assessed profile), and Rku (kurtosis of the assessed profile) using SEM photomicrographs and the SurfCharJ plugin [20,24]. Ra and Rq specifically describe the topological events' height and amplitude (peaks and valleys). In contrast, Rsk and Rku describe the distribution and are used to specify both the symmetry and the tailness of the peaks and valleys [10]. The plot profiles of the different implants were obtained using the ImageJ plugins, the surface plot profile, and the 3D surface plot [21,24].

Dental Implant Seeding with ADSCs
Each dental implant (N = 9) was placed vertically in a 96-multiwell plate (Gibco), which was contained in a 150 mm plate to maintain sterility. ADSCs between passages 3 and 6, kept semi-confluent to avoid spontaneous differentiation and maintained at 37 • C with 5% CO 2 in a humidified atmosphere in the proliferation medium, were washed in PBS (Sigma), detached using CTC, and centrifuged for 5 min at 1800× g. Following this procedure, ADSCs were resuspended in the proliferation medium to a final concentration of 6.25 × 10 6 /mL and 200 µL of medium was added at time point zero (T0) to each well. ADSCs were allowed to adhere to the implants, and after 24 h (T1), non-adherent cells were removed, transferring the implants into a clean 96-multiwell plate (Gibco) filled with fresh proliferation medium (200 µL) inside a 150mm plate. On day three (T2), the proliferation medium was changed as described for day one and left at 37 • C with 5% CO 2 in a humidified atmosphere until day seven (T3) [24].

Cell Adhesion, Viability and Proliferation
The procedure for cell detachment from the implants involved transferring the implants into 1.5 mL conical tubes containing CTC, followed by 10 min of incubation at 37 • C in a humid environment with 5% CO 2 . Cells were then centrifuged for 5 min at 1800× g and counted in triplicate at each time point (T1-T3) in the Neubauer chamber. ADSCs' adhesion, viability, and proliferation were evaluated in triplicate (N = 3) starting from T0 and throughout the three time points (T1-T3). Adhesion was considered as the number of adhering cells per mm 2 across the time point of interest and the other time points on the respective implants. Viability was considered as the ratio of dead cells across the time point of interest and the other time points on the respective implants. Proliferation was considered as the difference in cells counted across the time point of interest and the other time points on the respective implants. The total number of cells seeded at T0, as well as the surface of the submerged implant, were used to normalize the adhesion, viability, and proliferation assessment results [24,26].

Gene Expression Analysis
After being washed in PBS (Sigma), detached using CTC, and centrifuged for 5 min at 1800× g, total RNA was extracted from ADSCs cultured on each different implant (N = 3) using TRIzol (Gibco). RNA samples were then quantified by spectrophotometer and treated with DNase I (Ambion, Carlsbad, CA, USA). First strand cDNA synthesis was performed with 100 ng of total RNA using M-MLV reverse transcriptase (Invitrogen). PCR amplification was performed in a final volume of 50 µL, using 10 ng of cDNA, 50 pmol of each primer, and 2 U/µL of TaqI polymerase (Amersham, Piscataway, NJ, USA). The following primers were employed to quantify the expression of osteoblastic markers on days one, three, and seven in triplicate: alkaline phosphatase (alp) GCAGGCAGGCAGCTTCAC, TCAGAACAGGACGCTCAGG, 496 bp, 60

Statistical Analysis
The data are presented as mean ± standard deviations (SD) for cell counts and gene expression, implant roughness, and surface particle characteristics. The statistical significance was ascertained using the analysis of variance (ANOVA) and Bonferroni's post hoc tests. A paired t-test was used to compare data from similar groups. Statistical significance was determined by utilizing SPSS 23.0 (SPSS IBM, USA) to analyze the data, and significance is indicated as + Al Blast, * Laser Sint, £ Ti Blast, % Laser Mach, andˆAl Blast (ANOVA followed by Bonferroni's post hoc test), and * T1 and 24 h-3 days,ˆT2 and 24 h-7 days, and % T3 and 3-7days (paired t-test) with p ≤ 0.05 or p ≤ 0.001.

ADSCs' Isolation, Characterization, and Stemness Potential Assessment
ADSCs were isolated from lipoaspirates taken from healthy donors using the tumescent technique as described in our previously published works [24,26].
In summary, flow cytometry was used to show the stemness-related membrane proteins in ADSCs at passage three (P3), demonstrating the presence of CD73, CD90, and CD105, but not the hematopoietic CD34, CD45, and CD106, as previously reported [24]. To validate their multi-lineage capability, ADSCs were then differentiated through the adipocyte, osteoblastic, and chondroblastic lineages. Adipocyte induction was confirmed by the presence of lipid droplets detected using Oil Red O staining and compared with non-induced cells. Osteoblastically differentiated ADSCs differed from undifferentiated cells by displaying a high quantity of calcium, identified using alizarin red. Finally, Safranin O staining was performed to validate and compare the chondroblastic differentiation of induced ADSCs to that of control cells [24,26].

Implants' Morphological Characterization
SEM microphotographs of the implant surfaces revealed their characteristic morphology ( Figure 1). The texture of the Al Blast implant showed a high level of regular roughness and the presence of numerous deep cavities. Considering that pits and holes are easily invaded by bone cells, the geometry and connectivity of these cavities may allow bone cells to infiltrate the implant body deeply, allowing for better integration ( Figure 1A). Small irregularities with a shattered look were more evident on the Cor Blast implant ( Figure 1B). The Laser Sint implant showed the highest heterogeneity with either a sharp or round morphology ( Figure 1C). Less frequent, but larger, irregularities with a sharp look were more evident on the Laser Mach and Ti Blast implants ( Figure 1D,E).

Implants' Morphological Characterization
SEM microphotographs of the implant surfaces revealed their characteristic morphology ( Figure 1). The texture of the Al Blast implant showed a high level of regular roughness and the presence of numerous deep cavities. Considering that pits and holes are easily invaded by bone cells, the geometry and connectivity of these cavities may allow bone cells to infiltrate the implant body deeply, allowing for better integration ( Figure 1A). Small irregularities with a shattered look were more evident on the Cor Blast implant (Figure 1B). The Laser Sint implant showed the highest heterogeneity with either a sharp or round morphology ( Figure 1C). Less frequent, but larger, irregularities with a sharp look were more evident on the Laser Mach and Ti Blast implants ( Figure 1D,E).  The microarchitecture of a biomaterial, including the number of particles and their size, distribution, and density, dramatically affects its interaction with tissues and cells and influences the mechanical properties of the scaffold itself [27]. In particular, it has been observed that the texture of the implant surface significantly affects cells both in vitro and in vivo, encouraging osseointegration [7][8][9]27] and likely extending its duration and long-term stability [8,11]. To establish whether or not particles were evenly distributed throughout the surface and would have an impact on cell adhesion and proliferation, we thus analyzed the surface mean particle area, distance, spacing, and number ( Figure 2A-J, Table 1).
The microarchitecture of a biomaterial, including the number of particles and their size, distribution, and density, dramatically affects its interaction with tissues and cells and influences the mechanical properties of the scaffold itself [27]. In particular, it has been observed that the texture of the implant surface significantly affects cells both in vitro and in vivo, encouraging osseointegration [7][8][9]27] and likely extending its duration and long-term stability [8,11]. To establish whether or not particles were evenly distributed throughout the surface and would have an impact on cell adhesion and proliferation, we thus analyzed the surface mean particle area, distance, spacing, and number (Figure 2A-J, Table 1). Ti Blast implants' binarized and thresholded SEM microphotographs were analyzed to quantify the mean particle area (F), the distance between the particles (G), the nearest particle distance (H) the average wall thickness between the particle (I) and their surface (J). Results are expressed as the mean ± SD of three independent experiments, with + Al Blast, * Laser Sint, £ Ti Blast, % Laser Mach, andˆCor Blast indicating significance (ANOVA followed by Bonferroni's post hoc test). Table 1. Titanium dental implants' surface particle characteristics. Summary of the results obtained from surface particle analysis. The results demonstrated no significant difference in the mean area of the surface particles between the tested implants (Figure 2A-F; Table 1) (p > 0.05). The Ti Blast implant displayed a slightly lower mean distance between particles (4.06 ± 0.3 µm) than the Laser Mach (5.4 ± 0.76 µm) and Cor Blast (6.17 ± 1.17 µm) implants (p ≤ 0.05) (Figure 2A-E,G; Table 1).
The average wall thickness between the particles of Ti Blast (1.44 ± 0.25 µm) was the smallest and was significantly lower than in the Cor Blast (2.93 ± 1.21 µm) and Laser Mach (2.24 ± 0.74 µm) implants, but the difference was not significant with respect to the Al Blast (2.12 ± 0.76 µm) or Laser Sint (1.98 ± 0.3 µm) implants (p ≤ 0.05) (Figure 2A-E,I; Table 1). The results indicated that the surfaces with the greatest and lowest number of surface particles were, respectively, the Ti Blast (506 ± 18) and Cor Blast (174 ± 69 part/mm 2 ) implants (p ≤ 0.05). The number of surface particles on the Al Blast (250 ± 1.5), Laser Sint (304 ± 39), and Cor Blast (174 ± 69) implants were different from each other (p ≤ 0.05) but no significant differences were seen with respect to the Laser Mach implant (273 ± 126) ( Figure 2J, Table 1).
Surface roughness of titanium implants, particularly moderate roughness, allows for improved cell and bone connections [31][32][33] and mastication load transfer, which also promotes more osteoblastic and progenitor cell differentiation [8,34,35] than smoother or rougher surfaces. Consequently, according to ISO 4287/2000, a rough estimation of roughness was considered using the ImageJ plugin SurfCharJ to measure the following surface roughness parameters: Ra, Rq, Rsk, and Rku [20,24]. The 3D plots ( Figure 3A-E) and surface roughness profiles of the implants ( Figure 3F-J) were created using the Al Blast, Laser Sint, Ti Blast, Laser Mach, and Cor Blast SEM microphotographs, which were then merged in Figure 3K to emphasize their differences.
Following analysis of the arithmetical mean deviation (Ra), root mean square deviation (Rq), skewness of the assessed profiles (Rsk), and kurtosis of the assessed profiles (Rku) [20], substantial differences were evidenced and expressed as the mean ± SD of three independent experiments, with p ≤ 0.05 indicating significance (ANOVA followed by Bonferroni's post hoc test) (Figure 4A-D; Table 2).
The surface texture and roughness analyses revealed that the Ti Blast implant had the highest particle density (comprehensive of the highest number of particles, the smallest nearest particle distance, and a lower trend for the mean distance between particles and average wall thickness), even though there were no appreciable differences in their area. Instead, as seen by the corresponding indicators (Ra and Rq), the Al Blast and Cor Blast implants exhibit the greatest degree of roughness, whereas the Laser Mach implant had the lowest values. As expected, the Laser Mach implant also displayed the highest Rsk and Rku values, as opposed to the Al Blast implant.

ADSCs' Adhesion
In order to address any issues related to the seeding strategy, the ratios of adherent cells retrieved at the three time periods were standardized according to the total number of cells at day zero (1.25 × 10 6 cells) and the surface area of the implants (Figure 5A-D). The differences emphasized by the data are shown as the mean ± SD of three separate experiments, with p ≤ 0.001 indicating significance (ANOVA followed by Bonferroni's post hoc test and paired t-test).
At the three-day time point, the number of cells adhering to the Cor Blast implant was significantly different from those on the Ti Blast (p = 0.001) and Al Blast (p ≤ 0.001) implants; similarly, the Laser Sint implant differed from the Ti Blast (p = 0.032) and Al Blast (p = 0.002) implants. The Ti Blast and Al Blast implants differed from the Laser Sint and Cor Blast implants (p = 0.001 and p ≤ 0.001, respectively) ( Figure 5D and Supplementary Table S1).
On the seventh day, the cell number on the Ti Blast implant varied considerably from the Laser Mach (p ≤ 0.001), Laser Sint (p ≤ 0.001), and Cor Blast (p = 0.024) implants ( Figure 5F and Supplementary Table S1 Table S1).
Finally, comparing the adhesion between the 24 h and seven days time points revealed that the number of cells did not differ significantly for the Laser Mach implant, but did for all the other implants, with a decrease on the Ti Blast (p = 0.011), Al Blast (p ≤ 0.001), and Cor Blast (p = 0.006) implants, and an increase on the Laser Sint implant (p = 0.008) (Figure 5B-D; Supplementary Table S1).    Table S2). The cell viability on the Laser Sint implant was significant throughout all the time points, reaching a peak at three days, with a p ≤ 0.001 at 24 h, p = 0.003 at three days, and p ≤ 0.001 at seven days ( Figure 5H; Supplementary Table S2). Finally, the Ti Blast implant showed no discernible fluctuation in cell viability over any given period of time (p > 0.05) ( Figure 5H; Supplementary Table S2).
In summary, a distinctive difference in adhesion was highlighted for laser treated implants (Laser Mach and Laser Sint), which exhibited the highest degrees; however, limited proliferation was observed, specifically for the Laser Sint implant, and viability differences, except for with the Ti Blast implant, were seen throughout the three time points.

Osteoblastic Induction
The efficiency of the experimental seeding approach and the osteoblastic inductive capability of the implant surfaces were assessed by quantifying the expression of ECM and osteoblastic markers and normalizing their values for adherent cell number and surface area. Values are expressed as the mean ± SD of three independent experiments, with p ≤ 0.05 indicating significance (ANOVA followed by Bonferroni's post hoc test). By enhancing the expression of osn, coll-I, and alp in comparison to control cells, all of the treatments and surfaces of the whole-body implants were able to promote osteoblastic differentiation, as shown by the results. The Al Blast (152.8 ± 5.6 fold change (FC)), Ti Blast (120.1 ± 4.9 FC), Cor Blast (106.5 ± 7 FC), Laser Mach (66.2 ± 0.73 FC), and Laser Sint (62.9 ± 2.9 FC) implants triggered the highest coll-I expressions when compared to the control at the 24 h time point (p ≤ 0.05) ( Figure 6A; Supplementary Table S3).
The   Figure 6B; Supplementary Table S3). Figure 6. Summary of the spontaneous ADSCs differentiation resulting from their seeding on tested whole-body dental implants. After 24 h, three days, and seven days, ADSCs cultured on whole-body  Table S3).
It is interesting to note that by day seven, the late osteoblastic differentiation marker osc could not be induced by any surface.
When comparing the expression of the osteoblastic markers with respect to the different time points across the same implant type, it was found that even though the Al Blast implant did not significantly favor alp synthesis at 24 h and three days, it was able to induce its expression at seven days ( Figure 7A,D; Supplementary Table S4). Moreover, despite both Al Blast and Ti Blast implants not inducing the expression of osn at 24 h, they were able to significantly induce an increase between day three and day seven (p ≤ 0.05) ( Figure 7C,F; Supplementary Table S4). In contrast, alp expression was considerably different between all the intervals (p ≤ 0.05). The Cor Blast implant showed significantly increasing alp, osn, and coll-I at the 24 h-three day and three-seven day intervals, but was not able to induce a substantial increase in alp and coll-I during the 24 h-three day interval (Figure 7A-F;  Supplementary Table S4). On the contrary, the production of the osn messenger was always positively and significantly influenced by the Cor Blast surface (p ≤ 0.05) ( Figure 7C Table  S4). The Laser Sint implant successfully induced osn through all the intervals, 24 h-three days, 24 h-seven days, and three-seven days (p ≤ 0.05), but was able to be inductive for alp only during the three-seven day interval (p ≤ 0.05) and coll-I at 24 h-three days (p ≤ 0.05) and three-seven days (p ≤ 0.05) ( Figure 7A-F; Supplementary Table S4).
Overall, the results showed that sandblasted surfaces (Al Blast, Cor Blast, and Ti Blast) outperformed the laser-treated ones, promoting the highest spontaneous osteoblastic induction and synthesizing the highest levels of coll-I, osn, and alp messengers, although not osc. Significantly, among the sandblasted surfaces, the Ti Blast surface promoted, in general, the strongest expression of coll-I, osn, and alp.

Discussion
Since titanium dental implants have a wide range of surface textures and coatings, their impact on cells varies significantly. However, since 3D cell cultures on whole dental implants more closely resemble and mimic physiological cell interactions, they have been proven advantageous in assessing implant capabilities [24,25]. As a fair representation of the micro-morphological roughness of the implant market, we looked at five distinct commercially available dental implants with moderately rough surfaces, testing their potential under 3D conditions.
We found that while roughness is still significant, other factors sensibly influence the adhesive potential of the surfaces of whole-body titanium dental implants. Specifically, the range of surface roughness between Ra 0.8 and 1.2 µm and Rq 1.4 and 1.8 µm does have a beneficial impact on the adhesion of ADSCs.
Implant surfaces within this range of roughness and surface particle texture, such as those found on the Laser Sint and Laser Mach implants, favor a reasonable degree of adhesion compared to the other tested implants, similar to what we previously demonstrated [24]. We also found that Laser Mach and Laser Sint implants had a modest or negligible proliferative phase following the initial adhesion. This observation is also consistent with our previous finding [24] and the cells cultured on the Laser Sint and Cor Blast implants showed the highest viability.
Several studies have demonstrated that the surface characteristics of the dental implants are impacted by manufacturing procedures, changing their composition and affecting the function of tissue and cells [33,36]. Indeed, every process that physically alters the surface of the dental implants also alters their chemistry by introducing new atoms or molecules, which modify protein adsorption and cell adhesion [33,[36][37][38]. It is well known that the development of cytotoxicity and inflammation may result from the buildup of solubilized metallic ions on the surface of sandblasted implants [39,40]. A substantial volume of data suggests that throughout the manufacturing or handling process, impurities, whether metallic or nonmetallic, such as F, C, Mg, Fe, Al, Ca, P, Sr, and F, are purposefully or inexorably (despite stringent control) introduced or retained onto the implant surface [33,36,38,41]. We also previously demonstrated that, in contrast to machined and laser micro-patterned treatments, "sandblasting" is more likely to introduce elemental traces of C, Fe, Al, and O [42].
According to these findings, the implants that had undergone laser-related treatments, such as the Laser Sint implant, and to a lesser extent the Laser Mach implant, as well as a laser-treated implant from a previous study [24], appeared to favor ADSCs' adhesion better, supporting some proliferation as well, or at least were not involved in reduced viability, possibly as a result of higher cleanliness obtained during manufacturing.
The differences in the surfaces and textures showed an influence on ADSCs' spontaneous differentiation; in this case, the Ti Blast implant, but secondarily the Al Blast implant with some deficiencies, demonstrated the highest inductive potential on the osteoblastic markers coll-I and osn. However, in general, sandblasted implant surfaces, such as the Ti Blast, Al Blast, and Cor Blast surfaces, were more osteoinductive than laser-treated ones. Interestingly, the sandblasted surfaces, Al Blast, Cor Blast, and Ti Blast, underwent treatments involving aluminum, corundum, and fluoride, respectively. Aluminum oxide and corundum (the naturally occurring mineral form of Al 2 O 3 ) are deposited during the sandblasting process and seem to have differential effects according to their concentrations. In fact, it was pointed out that high concentrations of residual Al reduce mineralization and osseointegration, while lower concentrations [43,44] have positive effects on cell viability [36] and may be advantageous for bone formation and mineralization [45] and stem cell differentiation [44]. Fluorine surface treatment has been found to clearly affect actin filaments and cell morphology [38,46] and influence cell differentiation, but not proliferation [38,[47][48][49].
The findings also showed that the Ti Blast Implant, created by sandblasting with titanium dioxide spheres and etching with hydrofluoric acid, showed no significant effects on adhesion and viability despite its moderate roughness. However, among the other implants, it presented the highest density of surface particles, also suggesting the importance of the texture density.
As an explanation, the presence of this high-density texture may not have been as appealing in terms of cell adhesion, but along with the presence of fluorine ions derived from the surface treatment, it may have been very important in inducing ADSCs' spontaneous differentiation [19] via the β-integrins, focal adhesion kinase, and Src signaling pathway [50,51]. Indeed, the integrin-mediated interaction between the extracellular matrix proteins, such as fibronectin, vitronectin, Osn, and Coll-I, and the implant surface has been documented to regulate cell adhesion, differentiation, and survival [23,[50][51][52].
Previous effects become less consistent when considering that, from our previous results, the implant manufactured through hydroxyapatite and bland acid etching (HA Blasted) performed considerably better than all other implants despite having a similar texture and roughness [24]. Therefore, when surface roughness and manufacturing processes are comparable or identical, the difference in cell adhesion and differentiation between HA Blasted and the other implants might be attributed to the surface treatment (Table 3). Indeed, both the amplitude of surface roughness and the surface chemistry of HA-coated implants have been shown to increase cell adhesion [22,24,52] and subsequent differentiation [22,24,[52][53][54][55][56]. In fact, the greater wettability and hydrophilicity stimulate the adsorption of adhesion molecules from the media and, thereby, from the blood onto the titanium implant surface, promoting cell adhesion [52][53][54][55]. Moreover, the HA (Ca 10 (PO 4 ) 6 (OH) 2 ) coating's compositional and structural similarity to bone [57] enhances stem cell differentiation [24,57], stimulating extracellular matrix protein secretion and mineralization [56] via integrin-mediated signal transduction [52,58]. On the down side, the primary drawback of using HA is its limited ability to adhere to metal substrates; however, many researchers are making efforts to improve the techniques [57,59,60].

Conclusions
This study's results align with previous findings, demonstrating that stem cells are actively influenced by the topography and surface treatment of whole-body implants and the 3D culture approach is capable of unveiling the potential of both [24,25]. Indeed, the results revealed that, in addition to the roughness of the implants under consideration, which were representative of the implant market, the 3D culture approach highlighted substantial differences, in fact:

1.
The adhesion and proliferation of the laser-treated implants exhibited the highest degree, possibly due to the influence of the post-processing methods.

2.
The sandblasted implant surfaces performed better in terms of osteoinduction, possibly via the chemical treatments the implants underwent during the manufacturing process. 3.
The Ti Blast implant's osteoinductive capability might have benefited from the highest texture density and the fluorine surface treatment.
In light of these results, it is therefore recommended to pursue further research using the 3D culture methodology to broaden and better understand the effects and differences of implant surfaces' antimicrobial activity, wettability, coatings, or treatments with biologically active molecules, as well as various materials such as zirconia, tantalum, or porcelain.
Finally, by implementing this 3D approach and the subsequent advancement of the combinatorial approach in conjunction with pre-clinical contexts, it would be possible to identify the most efficient and appropriate surface to be used in combination with preconditioned stem cells, promoting a more significant and faster recovery while significantly improving the healing process in individuals with restorative deficiencies.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/biom13071048/s1, Table S1: Summary of the adhesion results on the different titanium implants' surfaces; Table S2: Summary of the viability results on the different titanium implants' surfaces; Table S3: Summary table of  Funding: No external funding, except for the authors' institution's support was provided for this study.

Institutional Review Board Statement:
The study protocol for tissue collection was approved by the Ethics Committee of the University of Udine (Par. N. 103/2011/Sper).

Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The datasets generated and/or analyzed during the current study are available upon reasonable request from the corresponding author.

Conflicts of Interest:
The authors declare that they have no conflict of interest to disclose.