1 INTRODUCTION

Peri-implantitis is a plaque-associated pathological condition in tissues surrounding dental implants.1 Treatment of peri-implantitis should therefore include anti-microbial procedure aiming at eliminating or disrupting the bacterial biofilm. A clean implant surface with sufficient biocompatible properties is a prerequisite for soft tissue integration and potential re-osseointegration. Surgical techniques are commonly required for proper access to affected implant sites.2 Following the surgical exposure of the contaminated implant surface, a number of different methods of mechanical instrumentation with or without chemical agents have been proposed. These include wiping with gauze soaked in saline,3-5 hand instruments,5 use of ultrasonic devices,6 lasers,7 rotating titanium brushes,6 air-polishing devices,8 and implantoplasty.7, 8 Clinical studies demonstrated that successful treatment outcomes, as evidenced by reduction of probing pocket depth and absence of bleeding on probing, can be achieved and maintained in the long term.3, 4 However, there is no consensus regarding the optimal protocol for the implant surface decontamination procedure.9 In addition, clinical studies revealed that resolution of peri-implantitis lesions is influenced by implant surface characteristics.3, 4

Furthermore, there is limited knowledge on the influence of titanium surface properties after decontamination procedures on cytocompatibility10 and epigenetic mechanisms.11 Previous studies have shown that surface structure and surface energy influence epigenetic mechanisms.11 Epigenetic modifications are associated with the configuration of chromatin, which are dependent on DNA methylation and histone acetylation. DNA methylation is commonly associated with gene silencing, while histone acetylation interferes with chromatin rearrangement and gene transcription. Vital factors influencing epigenetic changes in cells are stiffness and structure of the surface onto which cells adhere. Interestingly, it has been reported that cells were more affected by changes in surface topography at the nanometer scale as compared to micro- or macroscales.12 Mechanical instrumentation induces alterations of surface topography also at the nanometer scale. Thus, the assessment of osteogenic and epigenetic changes in cells attached on the treated surfaces would provide new knowledge on the influence of implant surface characteristics on expression of osteogenic factors contributing to tissue regeneration.

Well-controlled in vitro studies may provide a further understanding on the cleaning capacity of different methods of mechanical instrumentation at titanium surfaces, potentially improving clinical protocols. The use of a single-species bacterial biofilm has been shown to provide a well-controlled situation in experimental models and Streptococcus gordonii is a known initial colonizer of the oral biofilm and routinely identified in peri-implantitis patients.13, 14 In the present study, titanium discs with two different surface characteristics were experimentally covered with a S. gordonii biofilm and the cleaning potential of four different mechanical cleaning procedures was evaluated. In addition, cytocompatibility of the instrumented titanium discs for osteoblast-like cells and their epigenetic pattern were evaluated. The null-hypothesis was the lack of any differences in terms of cleaning capacity and post-cleaning cytocompatibility of implant surfaces between treatment groups.

2 MATERIALS AND METHODS 2.1 Titanium discs and surface characterization

Sterile, non-modified [Ti(s)] and shot-blasted [Ti(r)] titanium discs (Dentsply Sirona Implants) with a diameter of 12 mm and a thickness of 2 mm were used. Surface roughness was evaluated using white light interferometry (MicroXAM, ADE Phase Shift Technology, United States). Thus, at three randomly selected locations on each disc, arithmetical mean height (Sa), developed interfacial area ratio (Sdr), and density of summits (Sds) were assessed. Ti(s) discs demonstrated the following surface topography: Sa 0.187 ± 0.011 μm; Sdr 3.631 ± 0.283%; Sds 0.131 ± 0.007 summits/μm2. The corresponding values for Ti(r) discs were: Sa 1.546 ± 0.017 μm; Sdr 66.711 ± 2.843%; Sds 0.409 ± 0.002 summits/μm2 (Figure 1).

Surface characterization of titanium discs. Ti(s), non-modified titanium surface; Ti(r), modified (shot-blasted) titanium surface

2.2 Bacterial strain and culture conditions

S. gordonii ATCC 33399T were cultured on horse blood agar plates under aerobic conditions (N2: CO2/90%: 10%, respectively) at 37°C for 3 days. Single colonies of S. gordonii were then transferred from the agar plates to a brain heart infusion liquid culture medium. Turbidity was calibrated by measuring the absorption at OD660: 0.5.

2.3 Biofilm formation on titanium discs

The bacteria suspension (1.0 ml) was transferred to 24-well polystyrene culture plates containing the titanium discs previously coated with saliva for 4 h at 37°C (see Supporting Information). The plates were incubated under aerobic conditions (N2: CO2/90%: 10%, respectively) at 37°C for 48 h to allow biofilm formation on the titanium discs. In order to remove all suspended bacteria and preserve only the attached biofilm, all contaminated titanium discs were gently immersed in PBS. The discs were then transferred into new sterile 24-well plates for the subsequent experimental steps.

2.4 Surface decontamination procedures The biofilm-contaminated titanium discs were cleaned using (i) gauze soaked in saline (GS), (ii) an ultra-sonic device (US), (iii) a rotating nickel-titanium brush (TiB) or (iv) an air-polishing device (AP). The cleaning procedures were carried out with a constant and overlapping linear movement for 30 s by one trained operator (Y.I.). GS group: sterile 10 × 10 mm gauzes soaked in saline were used. The gauze was replaced after 15 s of use. US group: an ultra-sonic device (Airflow Prophylaxis Master®, EMS, Nyon, Switzerland) with a high-tech Polyether Ether Ketone (PEEK) fiber tip was used. The device was set at the recommended power of 5 bar static pressure and maximum irrigation with saline. Cleaning was performed with a constant light pressure. TiB group: a rotating nickel-titanium brush (NiTi brush, Nano model, HANS Korea, South Korea) was used at a rotation speed of 1200 rpm under irrigation with saline. The TiB was used at a variety of angles (45°–60°) and cleaning was performed with a constant light pressure. AP group: an air-polishing device (Airflow Prophylaxis Master®, EMS) with AIR-FLOW® powder PLUS (EMS) containing erythritol (sugar alcohol, 14 μm), amorphous silica and 0.3% chlorhexidine (CHX) was used. The handpiece was kept at a distance of 2.0 mm with an angulation of 90° to the titanium surface. The device was set at a power of 5 bar static pressure and maximum irrigation with saline.

Contaminated, untreated discs were used as negative controls (Bio) and non-contaminated, untreated titanium discs were used as positive controls (C). Following treatment, the discs in all four treatment groups were gently rinsed with 10 ml sterile saline to remove any potential deposits and transferred into a new sterile 24-well plate for further analysis.

2.5 Evaluation of residual deposits

Scanning electron microscopy (SEM; GeminiSEM 450, ZEISS, Germany) was used to identify residual bacteria and deposits as a result of the instrumentation. Discs (3 smooth and 3 moderately rough in each) from the controls (C and Bio) and the four treatment groups were placed into a 24-well plate, gently washed twice with PBS and covered in 2.5% glutaraldehyde buffered in 0.1 M PIPES buffer for 60 min. The disks were then washed for 5 × 5 min with 0.1 M PIPES buffer and incubated in 1% osmium tetroxide in 0.1 M PIPES buffer for 30 min in a cold and dark room. After washing for 5 × 5 min, the discs were dehydrated in increasing concentrations of ethanol (30%, 50%, 70%, 85%, and 95%) 5 min in each and for 20 min in 100% ethanol. The discs were dried with hexamethyldisilazane (HMDS), and finally sputtered with a gold coater (Quorum Sputter Coater Q150T E, Quorum Technologies Ltd, United Kingdom). Qualitative patterns of residual biofilm or any material deposits were evaluated at three different magnifications (500×, 5000×, and 25,000×) using a SE2 detector. In addition, a backscattered electrons (BSEs) detector was used for the identification of light elements, such as polyether ether ketone. These were shown as darker structures relative to the titanium background. To avoid bias, observation sites were selected by applying a grid with five cross points to the image of the entire disc with the inbuilt microscope SEM software at 19× magnification. At each cross point, a micrograph (magnification 500×, 5000×, and 25,000×) was captured, resulting in a total of five micrographs per disc for each magnification. Micrographs obtained with the SE2 and BSEs detectors at 5000× magnification were analyzed with an image analysis software (Fiji, National Institutes of Health, Bethesda, MD, United States) to assess the number of residual bacteria and the area of foreign materials. Micrographs obtained at 25,000× magnification were used for analyzing spot-like deposits using the Matlab® software (MathWork Inc., Natrick, MA, United States). Each image was analyzed as follows: The software pre-processed the image with a DoG filter (sigma 1.5 in units of pixels) to enhance spot-like features and remove the heterogeneous and complex background. Then the software used a GLRT15 for spot detection (window size 41 × 41 pixels, FWHM of 1.5 pixels, and chi-square of 100 for constant false alarm rate test) and reported the number of spots detected in the image.

2.6 Culture of osteoblast-like cell

Osteoblast-like cells (human osteosarcoma cell; MG63; Sigma-Aldrich Sweden AB, Stockholm, Sweden) were cultured in MEME supplemented with 10% NCS, 1% of non-essential amino acids solution, 1% of GlutaMAX, and 1% of antibiotic-antimycotic (100 units/ml penicillin, 100 μg/ml streptomycin, and 250 ng/ml amphotericin B). The cells were incubated in a humidified atmosphere of 95% air and 5% carbon dioxide at 37°C. The culture medium was renewed every 3 days. At ~80% confluency, cells were detached using 0.05% trypsin-0.53 mM EDTA-4 Na (Invitrogen, Ltd., Paisley, United Kingdom) at 37°C for 5 min and, immediately after surface decontamination, seeded onto the control and the treated titanium discs in 24-wells plate. The cells were then cultured using MEME supplemented with 10% NCS without any antibiotics.

2.7 Immunofluorescent staining of attached cells and analyses

A fluorescent microscope (FM; Leitz DM-RXA, Leica, Wetzlar, Germany) was used to evaluate the number of cells attached to the titanium surfaces and to assess cell morphology. Discs from the control and the four treatment groups (3 smooth and 3 moderately rough in each) were placed into a 24-wells plate immediately after surface decontamination and cells were seeded at a density of 4.0 × 104 cells/ml, corresponding to 1.0 × 104 cells/cm2. After 24-h incubation in a humidified atmosphere of 95% air and 5% carbon dioxide at 37°C, attached cells were fixed in 4% formaldehyde for 10 min at room temperature and washed three times with PBS. The cells were then incubated in 0.1% Triton-X-100 in TBS (Tris buffered saline, SIGMA-ALDRICH, USA) for 15 min and subsequently in 1% bovine serum albumin for 60 min. After washing with TBS, cells were stained using fluorescent dyes mouse anti-vinculin monoclonal antibody (abcam ab129002, 1:100, Cambridge, United Kingdom) followed by secondary antibodies conjugated with Alexa Fluor 488 (Invitrogen, Fischer Scientific, Göteborg, Sweden) and rhodamine phalloidin (abcam ab176757, 1:1000, Cambridge, United Kingdom). To visualize DNA content, cells were stained for 5 min in the dark using Hoechst (Invitrogen, Fischer Scientific) before being washed three times with PBS followed by double rinsing with distilled water. Five pictures from each disc at 100× and 400× magnifications were obtained. Images at 100× magnification were used for assessment of the number of attached cells using the Cell Counter plugin system of the Fiji software. Individual cell area was quantified at 400× magnification.

2.8 Evaluation of cell adherence and residual bacteria

SEM was used to assess presence of cells and bacteria attached to the titanium surfaces. Immediately after surface decontamination, discs from control groups and the four different treatment groups (3 smooth and 3 moderately rough in each) were placed into a 24-wells plate and cells were seeded at a density of 2.0 × 104 cells/ml, corresponding to 0.5 × 104 cells/cm2. After a 24-h incubation, the discs were processed in accordance with the previously described methods. At the pre-selected five cross points, micrographs were obtained at 2000× magnification and the % area of cells and bacteria was evaluated (Fiji software).

2.9 Osteogenic and epigenetic changes

Gene expression was analyzed using reverse transcriptase polymerase chain reaction (RT-PCR). After 24-h incubation with a concentration of 8.0 × 104 cells/ml, corresponding to 2.0 × 104 cells/cm2, total RNA from the attached cells on control and treated titanium discs (4 smooth and 4 moderately rough in each) was extracted using Rneasy Plus Mini Kit (Qiagen AB, Solna, Sweden). The RNA concentration was measured with the Qubit® 2.0 Fluorometer (Life Technologies Stockholm, Sweden) using the Qubit® RNA BR Assay Kit (Life Technologies Stockholm). One hundred nanograms of RNA per sample were reverse-transcribed into cDNA in a 20 μl reaction using iScript reverse transcription supermix for RT-qPCR (Bio-Rad Laboratories AB). The reaction was run on the MJ Mini 48-well Personal Thermal Cycler (Bio-Rad Laboratories AB) for 5 min at 25°C, 20 min at 46°C, 1 min at 95°C and hold at 4°C. cDNA was diluted 1:10 with nuclease free water. All qPCR-reactions were performed using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad Laboratories AB) according to the manufacturer's instructions. The information on the primers analyzed in this study described in Supporting Information. Stability comparisons of suitable reference genes were performed using the full TATAA Human Reference Panel (TATAA Biocenter). The software geNorm (Bio-Rad Laboratories AB) based on the method described by Vandesompele et al.16 was used to determine the best reference genes. The most suitable reference genes for normalization in this material were TBP and RRN18n. The 2−ΔΔCT method was used to analyze the relative difference in gene expression from the real-time quantitative PCR experiments.17

2.10 Data analysis

Data analysis was performed with Stata (version 16, Stata Corporation, College Station, TX, United States). Multiple linear regression was used to assess the effect of treatment method and surface category on outcomes related to (i) residual deposits, (ii) cytocompatibility, and (iii) relative gene expression. Interaction between treatment method and surface category was considered. Statistical significance was set at p < .05.

3 RESULTS 3.1 Surface analysis

Photographic images of the titanium discs representing the different treatment groups and controls are shown in Figure 2. Scratch marks after use of TiB on both types of discs could be observed with the naked eye. Discs instrumented with US demonstrated minor alterations in surface topography with a continuous scratch mark from the PEEK tip of the ultra-sonic device. Unlike US- and TiB-treated discs, discs instrumented with GS and AP showed no visual surface damage. To the naked eye, all four cleaning procedures seemed to have successfully removed S. gordonii biofilm. SEM analysis demonstrated that instrumentation with GS were effective in removing bacteria on Ti(s) discs without any obvious surface damages (Figure 2). Bacteria and gauze fiber deposits, however, were detected particularly on Ti(r) discs after instrumentation with GS. The discs instrumented with US did not show any residual bacteria. A dark material corresponding to PEEK was identified by the BSEs detector. PEEK remnants were detected on both Ti(s) and Ti(r) discs, but were more frequently found in Ti(r). When US-treated discs were observed at a higher magnification (25,000×), it was revealed that PEEK remnants were mixed with residual bacteria. Instrumentation of Ti(s) discs with TiB resulted in the elimination of bacteria, while the use of TiB on Ti(r) discs failed to achieve complete removal. Scratch marks on Ti(s) discs and “flattening” of Ti(r) discs were evident following cleaning with TiB. SEM images of discs after cleaning by AP demonstrated the absence of residual bacteria and surface topographical alterations on both Ti(s) and Ti(r) discs. Small powder-like deposits were observed on Ti(s) discs on SEM images obtained at 25,000× magnification, while no such deposits were detected on Ti(r) discs.

Photographic images and examples of scanning electron microscopy (SEM) images of the two different titanium discs following the application of four different decontamination methods, control and biofilm-contaminated surfaces at magnifications of 500×, 5000×, and 25,000×. Backscattered electrons (BSEs) detector was used for the identification of light elements, such as polyether ether ketone, that is shown as a darker color than titanium

As shown in Figure 3, the number of residual bacteria on Ti(s) discs in all treatment groups were close to 0. Following instrumentation of Ti(r) discs with GS, a significantly higher mean number of residual bacteria (152.7 ± 75.7) was observed when compared to the other treatment groups. Residual bacteria were detected on Ti(r) discs following cleaning with TiB (33.5 ± 22.2) at significantly higher number than following the use of US (0) or AP (0). Foreign material deposits were detected on SEM images of GS and US groups at the magnification of 5000×. The mean area% of foreign material (gauze fiber) on GS-treated Ti(r) discs was 1.1 ± 1.3%. GS-treated Ti(s) discs did not show any foreign material deposits. The mean area% of foreign material (PEEK deposits) on Ti(s) and Ti(r) were 3.0 ± 3.6% and 10.8 ± 9.6%, respectively. The numbers of spot-like deposits detected in the US and TiB groups were significantly higher when compared to the GS group for both Ti(s) and Ti(r) discs. Spot-like deposits on AP-treated Ti(s) discs were significantly more common than on GS- and US-treated Ti(s) discs, while the number of deposits on AP-treated Ti(r) discs was not significantly different from GS-treated Ti(r) discs.

(A) Number of residual bacteria per micrographs at a magnification of 5000×. (B) Area% of foreign material per micrographs at a magnification of 5000×. (C) Number of spot-like deposits per micrographs at a magnification of 25,000×. Data are shown as box plots overlapped with scatter plots

3.2 Cytocompatibility

Morphological observations from the fluorescent microscopic images (400×) of cells stained with rhodamine phalloidin and vinculin showed that cells on Ti(s) discs had a more elongated form compared to cells grown on Ti(r) discs (Figure 4). Cells grown on both type of discs treated with the TiB and AP were enlarged with a clear stretch of lamellipodia-like actin protrusions, similar to cells in the control group. The majority of cells on Ti(r) discs treated with the GS and US had a smaller cytoplasm and a spherical shape than the TiB and AP. While expression of vinculin was observed in cells on all surfaces, the localization of vinculin at the tip of stretching cytoplasmic projections were more evident in cells in the TiB and AP groups than in the GS and US groups. SEM images also indicated that cells grown on both types of discs treated with either TiB or AP were enlarged when compared to those of the GS and US groups.

Representative fluorescent microscope and scanning electron microscopy (SEM) images of the two different titanium discs following the application of four different decontamination methods and control

Results from the assessment of cell number are presented in Figure 5A. It was revealed that the number of cells that were attached to the titanium surfaces in all treatment groups (GS, US, TiB, and AP) was significantly lower than that of the control group. This was seen for both Ti(s) and Ti(r) discs. No significant differences in cell numbers were found between any of the treatment groups at Ti(s) discs. For Ti(r) discs, however, the number of cells in the TiB and AP groups was significantly larger than in the GS and US groups. There were no significant differences between Ti(s) and Ti(r) discs within any of the treatment groups. Cytomorphometric evaluation of the area stained with actin within the individual cells is shown in Figure 5B. It was revealed that cells in the GS and US groups were significantly smaller than cells in the TiB and AP groups of both Ti(s) and Ti(r) discs. There were no significant differences in actin area among the C and AP groups. At Ti(s) discs, the largest mean actin area of cells was observed following treatment with AP (1391.9 ± 116.4 μm2), while the smallest values were measured for US (1030.2 ± 61.7 μm2) followed by GS (1107.6 ± 114.0 μm2). For Ti(r) discs, the largest mean actin area of cells was observed following treatment with TiB (1081.8 ± 93.7 μm2), while the smallest mean actin areas of cells were measured on discs treated with GS (701.5 ± 27.0 μm2) followed by US (713.6 ± 93.7 μm2). In all four treatment groups, significant differences between Ti(s) and Ti(r) discs were observed.

(A) Number of cells attached to the titanium discs after 24 h incubation. (B) Cell actin area on the titanium discs. Mean values and ±SD; * indicates a significant difference (p < .05) between the (sub)groups and # indicates a significant difference (p < .05) between disc categories. All p values are based on multiple regression analysis

The results from the assessment of cell adherence and residual bacteria are illustrated in Table 1. For Ti(s) discs, the mean %area of cells in the control group was approximately three times greater than in the GS and US groups (p = .001). The mean %area of cells in the AP group was significantly greater than the mean %area of cells in the GS and US groups. The highest mean %area of bacteria coverage was found in the GS group followed by the TiB group. For Ti(r) discs, the mean %area of cells in the control group was about four times greater than in the GS and US groups (p = .001). The mean %area of cells on the TiB- and AP-treated surfaces was significantly greater than GS- and US-treated surfaces. Similar to Ti(s), the highest percentage of bacterial coverage for Ti(r) discs was found in the GS group followed by the TiB group. For both Ti(s) and Ti(r) discs, decontamination with AP showed superior results in terms of cell coverage and removal of bacteria. Significant differences between the disc types were found for bacterial coverage on GS- and TiB-treated surfaces.

TABLE 1. Mean %area of cells and bacteria on titanium surfaces per SEM images at 2000× magnification Group Surface Cells (%area) Versus C (p value) Versus GS (p value) Versus US (p value) Versus TiB (p value) Bacteria (%area) Versus GS (p value) Versus US (p value) Versus TiB (p value) C Ti(s) 15.1 ± 5.1 - - - - - - - - Ti(r) 12.3 ± 2.1 - - - - - - - - GS Ti(s) 5.1 ± 2.5 .001 - - - 0.153 ± 0.142a - - - Ti(r) 2.9 ± 0.3 .001 - - - 0.785 ± 0.213a - - - US Ti(s) 5.6 ± 0.8 .001 .835 - - 0.005 ± 0.006 .090 - - Ti(r) 3.2 ± 1.3 .001 .916 - - 0.007 ± 0.013 .000 - - TiB Ti(s) 10.6 ± 5.0 .054 .052 .079 - 0.027 ± 0.021a .144 .791 - Ti(r) 8.5 ± 2.3 .133 .032 .040 - 0.242 ± 0.118a .000 .011 - Air Ti(s) 10.9 ± 3.3 .102 .026 .041 .742 0.001 ± 0.002 .082 .961 .754 Ti(r) 10.0 ± 3.0 .339 .009 .011 .564 0 (not detected) - - - a Indicates a significant difference (p < .05) between disc categories. All p values are based on multiple regression analysis. 3.3 Osteogenic and epigenetic changes

No significant differences were observed in the expression of ALP among treatment groups and between types of surfaces (Figure 6). Cells grown on Ti(s) discs instrumented with TiB and AP had significantly higher OC expression compared to C, GS, and US, while no differences in OC expression were found among treatment groups at Ti(r) discs. In the control group, OC expression was significantly up-regulated in cells on Ti(r) discs when compared to Ti(s) discs. DNMT1 expression in cells on US-treated Ti(s) discs was significantly lower than on AP-treated Ti(s) discs and control discs. No significant differences among groups were found for Ti(r) discs. The expression level of DNMT3a in cells grown on Ti(r) discs were significantly higher for controls than GS-, US-, and TiB-treated discs, while no significant differences among groups were found for Ti(s) discs. HDAC1 expression in cells grown on Ti(s) discs treated with GS and US were significantly lower when compared to the controls and TiB- and AP-treated surfaces. No significant differences were found in the expressions of HDAC1 at Ti(r) discs. The HDAC2 expression in cells grown on Ti(s) control discs was significantly higher than the expression observed in all the other groups. Expression of HDAC2 in cells grown on Ti(r) discs were lower for TiB when compared to the control discs and discs treated with US.

Relative gene expressions in cells attached to non-contaminated and untreated titanium discs and to treated titanium discs. Mean values and ±SD; * indicates a significant difference (p < .05) between (sub)groups and # indicates a significant difference (p < .05) between disc categories. All p values based on multiple regression analysis

4 DISCUSSION

The goal of surface decontamination of implants affected by peri-implantitis is to eliminate or minimize bacterial plaque without critically damaging or otherwise negatively affecting the implant surface. Thus, the aim of this in vitro study was to investigate surface cleanness and cytocompatibility of titanium discs with different surface characteristics that was exposed to four different methods of mechanical instrumentation. It was demonstrated that mechanical decontamination of experimentally contaminated titanium discs was effective in reducing S. gordonii biofilm and air-polishing was the most effective method. No treatment modality restored cytocompatibility to the level of non-contaminated and untreated controls. The use of a titanium brush and air-polishing showed superior results in restoring cytocompatibility. Thus, the null-hypothesis of no differences between treatment groups was rejected.

The number of residual bacteria, area of foreign materials and number of spot-like deposits were used as indicators for evaluating the cleaning capacity of the four different treatment methods in the present study. Overall, biofilm removal was more effective at Ti(s) discs than at Ti(r) discs. At Ti(s) discs, all treatment modalities were equally effective, while bacteria on Ti(r) discs remained, particularly after instrumentation with GS and TiB. Cleaning of contaminated implants using gauze soaked in saline is a commonly used technique documented in clinical3-5 and pre-clinical studies.18-22 Available data, however, indicate that treatment outcomes were heavily influenced by implant surface characteristics.3, 4, 18, 19 Several in vitro studies23-25 also reported on superior results at devices with non-modified surfaces than those with modified surfaces. Charalampakis et al.25 examined the effect of mechanical instrumentation with GS on a 4-day biofilm grown intra-orally on titanium discs with four different surface characteristics. It was reported that titanium discs with a moderately rough surface harbored a firmly attached biofilm after GS cleaning, while titanium discs with a smooth surface presented with few biofilm clusters. The findings by Charalampakis et al.25 are in agreement with the results from the present study. Ultrasonic scalers with coated tips are commonly used for non-surgical instrumentation at implants. The procedure has been reported to be effective in removing biofilm from contaminated titanium discs.26-29 The present study confirmed these findings. However, it should be emphasized that deposits consisting of a mixture of PEEK and bacteria were identified on the titanium discs following instrumentation, particularly on Ti(r) discs. The presence of PEEK deposits following the application of ultrasonic scaler with coated tips has been described previously. Thus, Rühling et al.30 examined non-contaminated implant surfaces after instrumentation with a Teflon-coated US tip. While no deposits were observed on smooth implant surfaces, remnants of Teflon were observed on rough surfaces. Similar findings were also described in studies by Cha et al.31 and Schwarz et al.32 The rotating titanium brush used in the present study was narrow and had a conical shape. In previous in vitro studies24, 33-37 evaluating the effectiveness of a titanium brush on the removal of biofilm, a brush designed with horizontal bristles was used. Toma et al.33 assessed the residual amount of S. gordonii biofilm on rough titanium discs after decontamination. It was reported that the titanium brush performed better than plastic curettes and similar to air-polishing and implantoplasty. In the present study, AP was superior in terms of cleaning capacity but also resulted in spot-like deposits on Ti(s) discs. The observation that AP was effective in biofilm removal is in agreement with findings presented in two systematic reviews38, 39 which further indicated that the effectiveness of AP is less dependent of surface characteristics but more dependent on the type and size of powders as well as device settings.

Supracrestal tissue re-attachment and re-osseointegration following treatment of peri-implantitis depend on cell adhesion to the titanium oxide layer.40, 41 There is some evidence that bacterial plaque may alter the composition of the titanium oxide layer and that such alterations may be difficult to restore through decontamination.34, 42-44 Cytocompatibility and cellar response to treated titanium discs were therefore evaluated in this study. The present results indicated that GS treatment was inferior in restoring cytocompatibility of Ti(r) with when compared to the other methods of decontamination. The relatively high number of residual bacteria and foreign material deposits observed in the GS group, especially on Ti(r) discs, may have negatively influenced cytocompatibility. The reduced cytocompatibility of US-treated discs in the present study may be explained by the residual material deposits consisting of a mixture of PEEK and bacteria, especially on Ti(r) discs. The effect of residual material deposits on cell viability has been described in a number of studies.28, 32, 45 Thus, Schwarz et al.32 evaluated mitochondrial activity of human SaOs-2 osteoblasts grown on titanium discs after biofilm removal by ultrasonic scaling with a PEEK tip. Mitochondrial activity was significantly lower when compared to the non-contaminated and untreated titanium discs. The authors suggested that the biocompatibility of implant discs was impaired by debris from the PEEK fiber. The rotating titanium brush has the potential to expose a new TiO2 layer following the abrasive instrumentation as indicated by Al-Hashedi et al.34 The authors demonstrated that the titanium brush reduced organic impurities and increased the exposure of titanium surface. Furthermore, extensive grinding on titanium implants resulted in the exposure of a titanium surface that favored cellular attachment.46 These observation may explain the superior results of TiB in terms of cytocompatibility. On Ti(r) discs, cell density and spread following AP treatment were greater than GS and US treatment. The findings may be explained by the effective removal of bacteria by AP. It should be noted, however, that cytocompatibility of AP-treated titanium discs was not completely restored to the level of pristine discs. In a previous study,42 it was shown that although complete bacteria elimination was achieved by AP,