The Effect of Different Cleaning Protocols of Polymer-Based Prosthetic Materials on the Behavior of Human Gingival Fibroblasts

Dental implant abutment and prosthetic materials, their surface treatment, and cleaning modalities are important factors for the formation of a peri-implant soft tissue seal and long-term stability of bone around the implant. This study aimed to investigate the influence of a polymeric material surface cleaning method on the surface roughness, water contact angle, and human gingival fibroblasts (HGF) proliferation. Polymeric materials tested: two types of milled polymethylmethacrylate (PMMA-Ker and PMMA-Bre), three-dimensionally (3D) printed polymethylmethacrylate (PMMA-3D), polyetheretherketone (PEEK), and polyetherketoneketone (PEKK). Titanium (Ti) and zirconia oxide ceramics (ZrO-HT) were used as positive controls. A conventional surface cleaning protocol (CCP) was compared to a multi-step research cleaning method (RCP). Application of the RCP method allowed to reduce Sa values in all groups from 0.14–0.28 µm to 0.08–0.17 µm (p < 0.05 in PMMA-Ker and PEEK groups). Moreover, the water contact angle increased in all groups from 74–91° to 83–101° (p < 0.05 in the PEKK group), except ZrO-HT—it was reduced from 98.7 ± 4.5° to 69.9 ± 6.4° (p < 0.05). CCP resulted in higher variability of HGF viability after 48 and 72 h. RCP application led to higher HGF viability in PMMA-3D and PEKK groups after 48 h, but lower for the PMMA-Ker group (p < 0.05). After 72 h, no significant differences in HGF viability between both cleaning methods were observed. It can be concluded that the cleaning method of the polymeric materials affected surface roughness, contact angle, and HGF viability at 48 h.


Introduction
Dental implants have become a highly widespread and predictable treatment method for rehabilitating partially or completely edentulous patients. The soft tissues surrounding dental implants play a significant role in ensuring esthetics, preventing the invasion of the microorganisms and other hazardous agents, and avoiding crestal bone loss, thus ensuring the longevity of implant-supported of human gingival fibroblast (HGF) cells using titanium and zirconia as positive controls. The null hypothesis tested was that there are no significant differences in surface roughness, hydrophilicity, and HGF proliferation between surface cleaning protocols and types of polymeric materials.

Surface Polishing
Each side of the specimens was polished according to the manufacturer's protocol ( Table 2). Specimens were repolished following the same protocol before every separate experiment was used in the study.
After washing, specimens were air-dried at room temperature for 24 h. The same protocol was repeated after every specimen polishing session just before the next experiment.

Profilometry
Surface mean roughness (Sa) was measured using a 3D optical profiler system (PLµ 2300, Sensofar, Sensofar Group, Barcelona, Spain) with a confocal objective 50×/0.8 A with FOV 255 × 191 µm (Nikon Lu Plan, Nikon Metrology NV, Leuven, Belgium). Five specimens from every material group were selected randomly, and 3 images of surface areas (two areas were randomly chosen on a surface periphery and one in the center) on every selected specimen were made. The images were processed, and Sa values measured using Gwyddion Software (Czech Metrology Institute, Jihlava, Czech Republic).

Water Contact Angle Measurements
To evaluate surface hydrophilicity, the mean water contact angle was measured for each material group. After the surface cleaning samples were subsequently placed in a Krüss EasyDrop system (KRÜSS GmbH, Hamburg, Germany) and deionized water droplets (16 Ω, 2 µL) were placed on the samples, pictures were taken after 10 s and finally analyzed using Krüss software (KRÜSS GmbH, Hamburg, Germany). Two measurements (one on each side of the droplet) were obtained, and the mean value was calculated. The sample chamber temperature was kept constant at 21 • C using a LabTech H50-500 water chiller (LabTech Srl, Sorisole BG, Italy). Five specimen surfaces were randomly selected from each material group resulting in 5 measurements.

The Assessment of Fibroblast Proliferation on Different Dental Materials
Before the proliferation experiment, all the specimens were treated with UV-C light (Sylvania G15W T8 lamps, Feilo Sylvania Group, Shanghai Feilo Acoustics Co., Budapest, Hungary) with a peak wavelength of 253.7 nm for 7 min at a distance of 12 cm, resulting in irradiance of around 3.49 mW/cm 2 .
For the evaluation of cell proliferation, a suspension of HGF (30 × 10 3 cells/mL) was prepared and poured into the plate wells with the specimens, 200 µL into each. Three specimens of every dental material were used per each time point. The amount of the cells grown on the specimens and on the control plastic surface was registered at 24, 48 and 72 h. The experiment was repeated 3 times.
Amount of living cells in the well was registered using MTT test. At every check point cell growth medium was carefully removed from the well and 100 µL of MTT (3-(4,5-dimetiltiazol-2-il)-2,5-diphenyltetrazolium bromide; Merck Chemicals, Merck KGaA, Darmstadt, Germany) (1 mg/mL prepared in phosphate-buffered saline) was poured into every well. After 1 h of incubation at 37 • C, 5% CO 2 ) MTT was removed and the formed formazan crystals were dissolved in 100 µL ethyl alcohol (96%). Then, 50 µL of the developed solution was transferred to the clean wells. Later, the optical density (OD) of the solvent and specimens was measured at wavelength of 570 nm using spectral scanning microplate reader (Varioskan Flash, Thermo Scientific, Waltham, MA, USA). The difference between every specimen under observation and the mean of solvent OD was subsequently calculated. The obtained OD is proportional to the number of live cells on the observed specimen. To compare the assays, the OD, equal to the count of live cells grown on every specimen, was described in ratio with the negative control group (control). The negative control group was considered to be the OD, which described the number of live cells grown on a plate well surface.
Ratio OD of the specimen/OD of the control group at 24 h shows if the surface of a specific material inhibits the cell growth (the ratio is below 1) or, on the contrary, promotes cell proliferation (the ratio is higher than 1). A value of 1 is considered the cell growth on the control group surface at 24 h.

Statistical Analysis
Statistical analysis was performed using R i386 4.0.0 (Lucent Technologies, Auckland, New Zealand). The graphs were plotted using the ggplot2 plugin (Lucent Technologies, Auckland, New Zealand). Data normality was tested, and parametric methods were used in case of normal data distribution; otherwise, a non-parametric analysis was performed. In the case of parametric analysis, the equality of variances was tested using two variances F-test (two-sided) for two groups and Levene's test for homogeneity of variance (center-mean) for more than two groups. In the case of unequal variances, parametric tests were adjusted accordingly. To compare the means of two groups two-tailed independent samples t-test was used as a parametric test and two-sample Wilcoxon test (two-tailed with a normal approximation with continuity correction) as a non-parametric option. In the case of parametric means comparison for more than two groups, one-way ANOVA and subsequent Tukey's post hoc tests. For non-parametric multiple-group means comparison Kruskal-Wallis rank-sum test and pairwise comparisons using the Wilcoxon exact rank-sum test with p value adjustment using Benjamini and Hochberg method were performed. The statistical significance level was set at p < 0.05.

Surface Roughness
The results of surface profilometry are presented in Table 3. In the case of the CCP, no statistically significant differences between material groups were observed. After cleaning the specimens using the RCP, ZrO-HT group surface roughness was significantly lower (Kruskal-Wallis and pairwise Wilcoxon, p = 0.042) compared to Ti, PMMA-Ker, PMMA-Bre, PEEK, and PEKK groups. Significant surface roughness differences for each material group comparing two surface cleaning protocols are presented in Figure 1. Examples of material surfaces under a confocal microscope (PLµ 2300, Sensofar, Sensofar Group, Barcelona, Spain) are shown in Figure 2.  Statistically significant (Kruskal-Wallis and pairwise Wilcoxon, p-adjusted < 0.05) differences comparing material group means within each cleaning protocol are marked with a bracketed notation on the respective side of the table. CCP-Conventional cleaning protocol; RCP-Research cleaning protocol.

Contact Angle
Results of water contact angle (WCA) measurements are presented in Table 4. In case of CCP, ZrO-HT group showed significantly higher (ANOVA and post hoc Tukey's Contrasts) contact angle compared to Ti (p < 0.001), PMMA-Bre (p = 0.00163), PMMA-3D (p < 0.001), and PEKK (p < 0.001) groups. Morever, the Ti group showed a significantly lower contact angle compared to PMMA-Ker (p = 0.00181) and PEEK (p < 0.001) groups. Finally, the PMMA-3D group contact angle was significantly lower than the PEEK group (p = 0.03175). In case of surface preparation using the RCP, ZrO-HT group showed significantly lower contact angle compared to PMMA-Ker (p = 0.0107), PMMA-Bre (p = 0.0183), PMMA-3D (p = 0.0131), PEEK (p < 0.001), and PEKK (p < 0.001) groups. Furthermore, in the Ti group, the contact angle was significantly lower than the PEKK group (p = 0.036). Significant contact angle differences for each material group comparing two surface cleaning

Contact Angle
Results of water contact angle (WCA) measurements are presented in Table 4. In case of CCP, ZrO-HT group showed significantly higher (ANOVA and post hoc Tukey's Contrasts) contact angle compared to Ti (p < 0.001), PMMA-Bre (p = 0.00163), PMMA-3D (p < 0.001), and PEKK (p < 0.001) groups. Morever, the Ti group showed a significantly lower contact angle compared to PMMA-Ker (p = 0.00181) and PEEK (p < 0.001) groups. Finally, the PMMA-3D group contact angle was significantly lower than the PEEK group (p = 0.03175). In case of surface preparation using the RCP, ZrO-HT group showed significantly lower contact angle compared to PMMA-Ker (p = 0.0107), PMMA-Bre (p = 0.0183), PMMA-3D (p = 0.0131), PEEK (p < 0.001), and PEKK (p < 0.001) groups. Furthermore, in the Ti group, the contact angle was significantly lower than the PEKK group (p = 0.036). Significant contact angle differences for each material group comparing two surface cleaning protocols are presented in Figure 3. Images of water droplets used for contact angle measurement on different surfaces are presented in Figure 4. Table 4. Water contact angle (degrees) average values for each material group and cleaning protocol presented as averages +/− standard deviations.
Results of water contact angle (WCA) measurements are presented in Table 4. In case of CCP, ZrO-HT group showed significantly higher (ANOVA and post hoc Tukey's Contrasts) contact angle compared to Ti (p < 0.001), PMMA-Bre (p = 0.00163), PMMA-3D (p < 0.001), and PEKK (p < 0.001) groups. Morever, the Ti group showed a significantly lower contact angle compared to PMMA-Ker (p = 0.00181) and PEEK (p < 0.001) groups. Finally, the PMMA-3D group contact angle was significantly lower than the PEEK group (p = 0.03175). In case of surface preparation using the RCP, ZrO-HT group showed significantly lower contact angle compared to PMMA-Ker (p = 0.0107), PMMA-Bre (p = 0.0183), PMMA-3D (p = 0.0131), PEEK (p < 0.001), and PEKK (p < 0.001) groups. Furthermore, in the Ti group, the contact angle was significantly lower than the PEKK group (p = 0.036). Significant contact angle differences for each material group comparing two surface cleaning protocols are presented in Figure 3. Images of water droplets used for contact angle measurement on different surfaces are presented in Figure 4.

HGF Proliferation
Cell viability data is presented in Tables 5 and 6 for each cleaning protocol. HGF proliferation showed a tendency to increase over time for both cleaning protocols ( Figure 5). Significant differences comparing both periods (t-test, p < 0.05) were detected for CCP in PMMA-3D, PEKK groups. In the case of RCP, these differences were observed in Ti, PMMA-Ker groups. In the case of the CCP, data indicates more variability in HGF proliferation. After 48 h, polymeric materials (except for PMMA-Ker) cleaned by multistep RCP showed a tendency for higher cell viability (PMMA-3D, PEKK p < 0.05). After 72 h, cell viability tended to be higher for conventionally cleaned materials. However, due to the high dispersion of the values, the difference between two cleaning protocols at this time point was not statistically significant.

HGF Proliferation
Cell viability data is presented in Tables 5 and 6 for each cleaning protocol. HGF proliferation showed a tendency to increase over time for both cleaning protocols ( Figure 5). Significant differences comparing both periods (t-test, p < 0.05) were detected for CCP in PMMA-3D, PEKK groups. In the case of RCP, these differences were observed in Ti, PMMA-Ker groups. In the case of the CCP, data indicates more variability in HGF proliferation. After 48 h, polymeric materials (except for PMMA-Ker) cleaned by multistep RCP showed a tendency for higher cell viability (PMMA-3D, PEKK p < 0.05). After 72 h, cell viability tended to be higher for conventionally cleaned materials. However, due to the high dispersion of the values, the difference between two cleaning protocols at this time point was not statistically significant. The results are presented as averages +/− standard deviations. Statistically significant (ANOVA and post hoc Tukey's Contrasts, p < 0.05) differences comparing material group means within each time period are marked with a bracketed notation on the respective side of the figure. The results are presented as averages +/− standard deviations. Statistically significant (ANOVA and post hoc Tukey's Contrasts, p < 0.05) differences comparing material group means within each time period are marked with a bracketed notation on the respective side of the figure. The results are presented as averages +/− standard deviations. Statistically significant (ANOVA and post hoc Tukey's Contrasts, p < 0.05) differences comparing material group means within each time period are marked with a bracketed notation on the respective side of the figure. The results are presented as averages +/− standard deviations. Statistically significant (ANOVA and post hoc Tukey's Contrasts, p < 0.05) differences comparing material group means within each time period are marked with a bracketed notation on the respective side of the figure.
The results are presented as averages +/− standard deviations. Statistically significant (ANOVA and post hoc Tukey's Contrasts, p < 0.05) differences comparing material group means within each time period are marked with a bracketed notation on the respective side of the figure.

Discussion
This study has evaluated the effect of two cleaning protocols applied to five types of polymeric materials. The surface roughness, hydrophilicity, and fibroblast cell culture response were analyzed. In clinical practice, polymers are used as temporary or permanent prosthetic materials [38,39]. Temporary abutments are an important part of the treatment course as they form and condition soft tissues during the sensitive initial healing phase [40]. Moreover, in some cases of immediate implant placement or surgical soft tissue management procedures (especially in the esthetic area), temporary abutments might serve for the entire healing period, which could last from weeks up to a few months [41][42][43]. With increase applications of one-stage surgery and immediate or early loading, permanent prosthetic materials can also be used during the healing phase [44][45][46]. The ultimate attempt during these early healing stages is to guide the cellular response of soft tissues to form the architecture and sealing around dental implant abutment similar to that around healthy natural dentition.
The aim of this in vitro study was to evaluate the effect of five different polymer-based materials and their surface characteristics using two different cleaning protocols on HGF proliferation. The results revealed significant differences in surface roughness, water contact angle, and HGF proliferation between the groups. Therefore, the null hypothesis was rejected.
Previous findings indicated that a threshold of 0.2 µm average roughness value reduces bacterial adhesion significantly [26,[47][48][49]. In this study, the roughness values of tested polymeric materials

Discussion
This study has evaluated the effect of two cleaning protocols applied to five types of polymeric materials. The surface roughness, hydrophilicity, and fibroblast cell culture response were analyzed. In clinical practice, polymers are used as temporary or permanent prosthetic materials [38,39]. Temporary abutments are an important part of the treatment course as they form and condition soft tissues during the sensitive initial healing phase [40]. Moreover, in some cases of immediate implant placement or surgical soft tissue management procedures (especially in the esthetic area), temporary abutments might serve for the entire healing period, which could last from weeks up to a few months [41][42][43]. With increase applications of one-stage surgery and immediate or early loading, permanent prosthetic materials can also be used during the healing phase [44][45][46]. The ultimate attempt during these early healing stages is to guide the cellular response of soft tissues to form the architecture and sealing around dental implant abutment similar to that around healthy natural dentition.
The aim of this in vitro study was to evaluate the effect of five different polymer-based materials and their surface characteristics using two different cleaning protocols on HGF proliferation. The results revealed significant differences in surface roughness, water contact angle, and HGF proliferation between the groups. Therefore, the null hypothesis was rejected.
Previous findings indicated that a threshold of 0.2 µm average roughness value reduces bacterial adhesion significantly [26,[47][48][49]. In this study, the roughness values of tested polymeric materials after CCP tended to be higher than the threshold, but lower after RCP. Recent research has indicated that titanium surfaces smoother than 0.1 µm might have a negative impact on fibroblast function [50]. Surface roughness in the range of 0.1-0.2 µm showed higher fibroblast adhesion than smoother or rougher surfaces for titanium, zirconium dioxide, and lithium disilicate materials [51]. In the case of CCP, only Ti and ZrO were in this range, while after RCP, all tested materials except for ZrO (0.079 +/− 0.017 µm) were in this range.
There is limited data regarding the influence of the cleaning method applied to the surface roughness parameters. Heimer et al. found that an ultrasonic bath for 6.3 min resulted in similar material surface roughness (Ra) values (0.033 µm for PEEK and 0.066 µm for PMMA) compared to other laboratory cleaning systems operating for 15-20 min [19]. The results contradict our findings, most likely due to differences in study design and methods used. In the present study, polishing protocols recommended by manufacturers were applied to each material. Moreover, in the case of CCP ultrasonic bath was used for 3 min, and finally, a non-contact surface profilometry method was used to evaluate roughness (Sa).
Hydrophilic surfaces are known to be favorable for eucaryotic cells [30,31]. The hydrophobicity of the implant abutment surface is also reported to influence the adhesion of certain bacteria [63]. Wassmann et al. conducted an experiment demonstrating that hydrophobic surfaces are more attractive to Staphylococcus epidermidis, which causes a cytotoxic effect on human fibroblasts and therefore interferes with osseointegration and soft tissue healing [64][65][66]. Using CCP lead to hydrophobic surfaces of ZrO and PEEK, while in the case of RCP, only polymeric materials surfaces (PMMA, PEEK, and PEKK) were hydrophobic.
A tendency for increased contact angle after application of RCP was observed in all groups except for ZrO, with significant differences in Ti and PEKK groups. Changes in the surface hydrophilicity after using different cleaning protocols were also demonstrated in previous studies: ultrasonic bath alone resulted in the least hydrophilic surfaces for PMMA and PEEK compared to other laboratory cleaning methods [19].
Different water contact angle measurements were reported by other studies: PMMA 72-99° [20,21], 3D printed PMMA (SLA) 71-79° [58], PEEK 10-114° [15,20,24,54], and PEKK 64-83° [57]. The findings of the current study for PMMA and PEEK are similar to those mentioned above, but PMMA-3D and PEKK showed higher contact angles after RCP. This research showed a tendency of a more thorough cleaning method (RCP) to result in a higher water contact angle for tested materials compared to CCP.
The CCP resulted in higher variability of cell proliferation and was less predictable in the outcome. Gheisarifar et al., (2020) used ultrasonic cleaning of specimens and concluded that PEEK plasma treatment (Sa 0.68-2.14 µm, WCA 10-12°) increased HGF proliferation [15]. The present study has demonstrated that the ultrasonic cleaning based method (CCP) of PEEK surface (Sa 0.28 ± 0.1, WCA 91 ± 9°) tends to be less favorable for HGF proliferation compared to RCP (Sa 0.17 ± 0.06 µm, WCA 97 ± 12°). This tendency was not evident at 72 h. Similar results were observed with PMMA-3D and PEKK materials. PMMA-Ker demonstrated opposite outcomes in terms of HGF proliferation, favoring CCP at 48 h. Methacrylate polymers were shown to have different HGF cytotoxicity and attachment properties, depending on their composition and fabrication method [21]. This study showed similar results comparing fibroblast proliferation over time under different cleaning protocols on PMMA-Ker, PMMA-Bre, and PMMA-3D surfaces.
Another study evaluated immortalized human gingival epithelial keratinocytes (iHGEK) behavior on smooth Ti, rough ZrO, and medium PEEK surfaces and concluded similar cellular responses to all three materials [24]. This research provides similar results with HGF culture, as there were no significant differences in terms of cell proliferation between Ti, ZrO, and PEEK groups.
Though the average proliferation values tended to be higher for the CCP after 72 h, due to high variability, these differences were not significant in any of the groups (including both positive controls). This study backs the importance of multiple repetitions (at least three independent experiments) of cell proliferation experiments to evaluate the effect of the material surface.
Furthermore, limitations of the study design must be taken into account as it was tested under the sterile, well-controlled conditions with HGF monoculture. Under clinical conditions, immediately after the placement of implant abutment, its surface becomes a subject of fibroblasts, epithelial cells, and microbial adhesion, as well as inflammatory tissue reaction. Human histology research shows that long epithelial junction predominates in contact with the transmucosal implant component [12]. Furthermore, the collagen fibers tend to be oriented parallel to the surface [12]. This can be influenced by many factors, including the implant-abutment connection, microgap, and "micro-trauma" due to connections and disconnections of the implant transmucosal component during the treatment [12].
The design of the current study did not allow to evaluate these circumstances. As results can be influenced by different types of materials, polishing and cleaning protocols, cell and bacteria types used in the studies, further research is needed to provide clinical recommendations.

Conclusions
Considering the limitations of the current study, the following conclusions can be drawn: • Polymer-based material surface cleaning protocol can significantly influence roughness, contact angle, and fibroblast proliferation of polymer-based materials; • Lower surface roughness (Ra < 0.2 µm) resulted using an RCP and was higher (Ra > 0.2 µm) when the CCP was applied; • RCP showed a tendency to reduce hydrophilicity of polymer-based material surfaces; • CCP resulted in more variability in surface characteristics, and the cellular response was less predictable. RCP significantly favored HGF proliferation on PMMA-3D and PEKK surfaces after 48 h.