Analysis of Oral Microbiome on Temporary Anchorage Devices Under Different Oral Hygiene Conditions

Background: Temporary anchorage devices (TADs) are maximum anchorages that have been widely used in orthodontic treatment. Poor oral hygiene might result in inammation and decreased stability of the TADs. In this study, we aimed to unveil the differences in the microbiome between TADs under different oral hygiene conditions. Methods: Oral hygiene condition was stratied by Oral Hygiene Index- Simplied (OHI-S), Plaque Index (PLI), and Gingival Inammation. Scanning electron microscopy (SEM) was used to analyze the existence of biolm on the surface of 8 TADs, Ten TADs from the good oral hygiene group (GOH), and 10 TADs from the poor oral hygiene group (POH) were analyzed by 16S rRNA gene sequencing. Results: Principal coordinate analysis (PCoA) based on β diversity revealed differential sample clusters depending on oral hygiene conditions. When comparing specic genera, Veillonella, Streptococcus, Neisseria, were more enriched in the poor oral hygiene group. Conversely, Fusobacterium, Porphyromonas exhibited more richness in the good oral hygiene group. TADs in the good oral hygiene group demonstrated enriched microbial activities involved with signal transduction, cell mobility and energy metabolism. TADs in poor oral hygiene demonstrated enriched functions in membrane transport, transcription and signaling molecules and interactions. Conclusions: In summary, this analysis elucidated the difference in total composition and function of TADs oral microorganisms between patients with good oral hygiene and patients with poor oral hygiene, which highlighted the importance of maintaining good oral hygiene in TADs treatment.


Background
Anchorage control has been a major concern for orthodontists for decades. The effectiveness of traditional extraoral anchorage relies greatly on patients' obedience, while intermaxillary and intramaxillary anchorage might cause unwanted movement of the anchorage tooth. As "maximum anchorage", the temporary anchorage devices (TADs) have been well received since their appearance. Not only does this device facilitate anchorage reinforcement compared to conventional anchorage control, but it also has great advantages in exibility, versatility, control, and minimal invasiveness [1,2]. It is particularly indicated in anterior en-masse retraction, molar protraction, as well as the intrusion of supraerupted teeth, and midline correction [3].
The success rate of TADs is generally over 80% [4]. However, on some occasions, TADs present excessive mobility which eventually leads to loss of anchorage. Factors affecting the success rate of TADs may be divided into location-related factors (bone quality, mucosa, position), orthodontic-related factors (force application), and implant-maintenance factors (the presence of in ammation) [5]. Oral hygiene condition is one of the decisive factors of in ammation and re ects implant-maintenance condition [5][6][7]. Poor oral hygiene induces changes in the surrounding environment, which eventually in uences the colonization of TADs surrounding bacteria. Studies targeting peri-implantitis have unveiled the pathogenic role of bacteria when attached to the surface of a dental implant if oral hygiene methods were not properly introduced [8]. TADs and dental implants share a common denominator in being titanium implants in the oral cavity. When a TAD is inserted trans-gingivally and exposed to all sorts of oral bacteria, an arti cial sulcus is created, which favors microorganisms migration and bio lm formation [9]. Until now, a few epidemiological surveys have listed oral hygiene as the most signi cant factor affecting TADs success [10,11]. Even though other clinical trials have not con rmed the direct effect of poor oral hygiene on the success rate, they have highlighted the strong role of local in ammation in increasing the risk of failure [5][6][7]12].
Current studies concerning the microbiome effects on TADs have concentrated on the observation of attached micro ora and the identi cation of certain species. Ferreira et al utilized a scanning electron microscope (SEM) to observe the microbiota attached to the surface of the TADs, revealing bio lm existence on the head, transmucosal, and body segments of the TADs. With the utilization of uorescence imaging, Garcez et al observed higher uorescent intensity in in amed TADs with redness on surrounding tissues [13]. When exploring the differences of microbiome between stable TADs and unstable TADs, nevertheless, polymerise chain reaction (PCR) and DNA-DNA hybridization revealed no major differences in the testi ed microbiome. These studies partly uncovered the role of microbiomes in generating TADs surrounding diseases but failed to build up the complete connection. Moreover, these studies did not demonstrate the overall structure of bio lm and neglected the immune response from the human host.
The emergence of next-generation sequencing (NGS) technology has created a new way for microbiome research, which can be used to detect population diversity, identify the structure of microbiome and predict functional roles on certain occasions [14]. Therefore, 16S rRNA gene sequencing would be an effective tool in assessing the microbiome on the TADs surface.
Therefore, here, we brought up the hypothesis that oral hygiene condition in uences microbiota composition on the TADs surface. We conducted a case-control study to investigate the structural and compositional changes in oral microbiota on TADs depending on different oral hygiene conditions.

Participant Selection and Sample Collection
All of the subjects in this study participated in orthodontics treatments in Peking University Hospital of Stomatology. In all of them, the use of orthodontic anchorage was indicated. Each participant signed an informed consent form to enroll in the trial. This analysis was rati ed by the Ethics Committee of the Peking University Hospital of Stomatology under PKUSSIRB-202060204. All methods were carried out in accordance with relevant guidelines and regulations.
All subjects selected met the following criteria: 1) aged 12-45 years; 2) periodontally healthy or received systematic periodontal treatment before orthodontic treatment. Before placement of the appliance, patients with a periodontal probing depth of less than 3 mm were included; 3) non-smokers; 4) without systematic disease; 5) not pregnant; 6) no antibiotics used up to three months before removal.
Self-drilling titanium orthodontic TADs (diameter, 1.5 mm; length, 7 mm or 8 mm; Zhongbang Medical Treatment Appliance, Xi'an, China) were inserted in the maxilla, between tooth roots of anterior or posterior teeth, between the buccal or palatal surface. All TADs were inserted by one experienced orthodontist. No damage to the adjacent tooth roots was observed. All patients received oral hygiene instructions to brush TADs and the surrounding tissues when adopting oral hygiene methods. All TADs were activated 1 month after placement. In total, 8 TADs from 8 individuals, 4 from the good oral hygiene group (GOH), and 4 from the poor oral hygiene group (POH) were included for SEM. Twenty TADs from 20 patients, 10 from GOH, and 10 from POH were included for 16S rRNA gene sequencing. All of the included TADs remained stable during treatment and were removed until attaining the desired result.

Oral Hygiene strati cation
Clinical parameters of the oral hygiene condition were obtained when the TADs were removed. Oral hygiene conditions were measured by Oral hygiene Index-Simpli ed (OHI-S) [15], Plaque Index (PLI) [16] as well as gingival in ammation. The Index teeth of the OHI-S and PLI were 16,11,26,31,36, and 46. The clinical parameter score of each patient was determined by the average measured value of the designated tooth. Gingival in ammation was evaluated by calculating the number of swollen dental papillae in the whole oral cavity, as follows: 0, normal gingiva; 1, slight edema in no more than two dental papillae; 2, redness, edema, and glazing in no more than two dental papillae; 3, redness, edema, and glazing in more than two dental papillae and no more than ve dental papillae; 4, redness, edema, and glazing in more than ve dental papillae and no more than eight dental papillae; and 5, redness, edema, and glazing in more than eight dental papillae.
GOH group was de ned as a tiny amount of dental plaque on the index teeth and only minor signs of gingival in ammation. POH group was de ned as a massive amount of dental plaque on the index teeth and marked signs of gingival in ammation.
To be more speci c, the GOH group referred to OHI-S no more than 1, PLI less than 1, and Gingival In ammation no more than 2. Conversely OHI-S more than 1, PLI no less than 2, and Gingival In ammation no less than 3 was de ned as POH group.

DNA Extraction
The TADs were placed in nonpyrogenic microcentrifuge tubes containing 0.5 mL normal saline solution and stored at − 20°C refrigerator temporarily. Before DNA extraction, the tubes were agitated in an ultrasound cleaner (SB-3200DTN, Scientz, Ningbo, China) for 20 minutes. The tubes were then centrifuged at 8000 rpm for 15 minutes to remove the supernatant. The precipitate was then sent for DNA extraction.
QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany) was used to extract genome DNA. Extraction procedures were performed according to the kit instructions. Before extraction, 180 µL lysozyme (Solarbio, Beijing, China) was added to the reaction system. The system was then incubated at 37°C for 30 minutes. NanoDrop 2000 Spectrophotometer (Thermo Fisher Scienti c, Carlsbad, CA, United States) and 1% agarose gel electrophoresis were used to determine the purity and integrity of DNA.

16S rRNA Gene Sequencing
Primer 341F (5'-CCTACGGGRSGCAGCAG-3') and 806R (5'-GGACTACVVGGGTATCTAATC-3')were added to amplify the 16S rRNA hypervariable V3-V4 region. Unique index and adapter sequence were added to the 5' to distinguish each sample. Next, the KAPA HiFi Hotstart ReadyMix PCR kit (Kapa Biosystems, Wilmington, MA, USA) was utilized to perform PCR. Then, AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) was used to purify the amplicon. The library was quanti ed by Qubit (Thermo Fisher Scienti c, Carlsbad, CA, United States) and real-time PCR with the same proportion. Miseq PE250 platform (Illumina, San Diego, CA, USA) was adopted for sequencing.

Data Processing
Raw data were deposited at Sequence Read Archive under project no. PRJNA690665. Preprocessing of data was performed under guidance [17].  [20,21] were both employed as databases in sequence annotation.
To start the downstream analysis, random rare cation procedures were taken for each pre-processing sequence to mitigate the effect of varying sequencing depths. α-diversity indices (the Chao1 richness estimator, Shannon index) were calculated as metrics to the microbial diversity within each sample. Bray Curtis distance and Unifrac distance were assessed as representations of the overall microbiome dissimilarities or β-diversity. Principal coordinate analysis (PCoA) was implemented to re ect the βdiversity through R. Then, ASVs were classi ed into microbial taxa (phylum, class, order, family, and genus). The phylogenetic tree was constructed on ITOL (https://itol.embl.de/). Linear discriminant analysis Effect Size (LEfSe) was used to identify differentially taxa between groups [22]. Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) [23] (version 1.1.3) tool was adopted to predict functional roles based on the Kyoto Encyclopedia of Gene and Genomes (KEGG) pathway.

Statistical Analysis
An independent sample Student's t-test and nonparametric Wilcoxon's test were used to evaluating demographic features and clinical parameters between the two groups. The difference between αdiversity was calculated by analysis of variance (ANOVA). The pairwise permutational multivariate analysis of variance (PERMANOVA) procedure tested the signi cance of β-diversity between TADs with good or poor oral hygiene. This was realized by the Adonis function of the R package vegan 2.5-6, allowing for 9999 permutations. Edge R test was used to determine differences between ASVs. Bar plots and cladogram graphs served as visualizations of LEfSe results (http://huttenhower.sph.harvard.edu/galaxy/). Spearman's correlation coe cients were calculated to measure the correlation between bacteria. Cytoscape (ver. 3.5.1) was used to visualize the correlation network between genera. LEfse test was also employed to identify statistically different KEGG pathways between the two groups.

Data Availability Statements
Raw data were deposited at Sequence Read Archive under project no. PRJNA690665.

Overview of Subjects and Samples
In this study, we performed SEM observation on 8 samples from 8 individuals. Then, we conducted 16S rRNA gene sequencing of 20 TADs samples from 20 individuals.
The individuals' demographic features and clinical parameters were presented in Table 1 for 16S rRNA sequencing analysis. Similarly, those features for samples involving SEM were also listed in Additional File 1. Signi cance was observed in OHI-S, PLI, and Gingival In ammation in the two tables. To explore the microbiome on the TADs surface under different oral hygiene conditions, we rst performed SEM to prove the existence of the microbiome on TADs (Fig. 1). SEM demonstrated the existence of micro ora on of the surface of observed TADs, both in GOH and POH. Rods and coccoid bacteria were all seen in this region. Besides, tissue remnants containing bers and red blood cells were also observed. This testi ed microbiome colonization on the surface of TADs.

Phylogenetic Alterations under Different Oral Hygiene Conditions
During 16S rRNA gene sequencing processing, a total of 491600 clean reads were acquired. The average sequences for each sample were 24580, eliciting 960 ASVs.
To characterize the microbiome of the individuals with poor oral hygiene, α-diversity and β-diversity were rst evaluated as re ections of the overall structural features and composition (Fig. 2). Although TADs in individuals with poor oral hygiene presented higher α-diversity re ected by Chao1 and Shannon, no statistical difference was observed (p = 0.874, p = 0.685, respectively) (Fig. 2a, Fig. 2b). However, PCoA based on weighted UniFrac distances revealed a statistically signi cant discrepancy in phylogenetic structures between the GOH and POH (p = 0.004) (Fig. 2c). Different clusters were formed, indicating separation in microbiome composition between the groups. Similarly, PCoA based on Bray Curtis distances revealed a clear separation between the groups (p = 0.017) (Fig. 2d).
To better characterize the differences, the LEfSe test was used. The cladograms revealed the hierarchical relationship of those bacteria (Fig. 4a). The bar plot showed the signi cance of the differential microbiome (Fig. 4b). At the genus level, Veillonella, Streptococcus, and Neisseria were more enriched in the poor oral hygiene group. Conversely, Fusobacterium, Porphyromonas exhibited more richness in the good oral hygiene group (Fig. 4c). When comparing speci c species, however, Veillonella dispar and Streptococcus mutans showed higher abundance in the POH group, while Fusobacterium nucleatum and Porphyromonas gingivalis did not differ signi cantly between groups (Additional File 3).

Correlation Networks of Microbiome on the TADs
Bacteria showed strong interactions with each other. In order to explore the connection between the most abundant genera, we then generated abundance-based correlation networks (Fig. 5)

Microbiota Involvement in Functional Variation
To study the functional changes in TADs in individuals with poor oral hygiene, the PICRUSt algorithm was employed to predict the path of microbiota derivation based on the KEGG database. Differences in functional abundance between TADs with good oral hygiene and TADs with poor oral hygiene were evaluated (Fig. 6). TADs in the good oral hygiene group demonstrated enriched microbial activities involved with signal transduction, cell mobility and energy metabolism. TADs in poor oral hygiene demonstrated enriched functions in membrane transport, transcription and signaling molecules and interactions.

Discussion
Our study characterized alterations in microbial community pro les on the TADs surface depending on the oral hygiene condition. We identi ed the compositional and phylogenetic changes in the microbiome on the surface of TADs in relation to the oral hygiene condition. We also predicted the correlation between microbiome colonization and functional involvement. To the best of our knowledge, this is pioneering research in understanding how oral hygiene in uences microbiome colonization on the surface of TADs.
This is also the rst article that characterizes the microbiome on TADs with next-generation sequencing.
Microbial will colonize on the surface of TADs. When a TAD is inserted, a new site is created, which is de ned as the gingival sulcus between the surrounding gingiva and the TAD cervical [24]. In a previous study utilizing SEM, Ferreira discovered bacteria colonization on the head, transmucosal surface, and body segment of TADs [25]. Similarly, in our study, the existence of micro ora was observed on the surface of TADs. Previous studies also observed the adhesion, aggregation, and development of the microbial colonization process in TADs using cell growth methods or uorescence images [13,24]. The interactions between the microorganisms and the host maintain the microecological balance around the TADs [9].
There is strong evidence that patients who failed to control plaque at a low level after implant surgery have an increased risk of peri-implantitis. Resembling this process, if oral hygiene methods could not be implemented properly, in ammation around TADs is generated. As local in ammation contributes to the increased risk of TADs failure, the longevity of the TADs is eventually compromised [5][6][7]12]. In our study, TADs in the good oral hygiene and poor oral hygiene groups displayed compositional and phylogenetic differences. Genera Veillonella, Streptococcus, Neisseria were signi cantly elevated in the POH group. Streptococcus species, especially acidogenic and aciduric species, such as Streptococcus mutans, are linked to the progression of dental caries [26,27]. It can adhere to the solid surface and promote interactions with other microorganisms. Veillonella has been frequently detected in the micro ora of individuals with poor oral hygiene conditions, and it has also been elevated in the microbiota of patients with dental caries and periodontitis [28]. In the pathogenesis of dental caries, the presence of Veillonella can promote the formation of bio lms of Streptococcus mutans, Streptococcus gordon, and Streptococcus salivarius [29]. In periodontitis, the adhesion of Veillonella and Porphyromonas gingivalis in the process of bio lm formation can produce a synergistic effect [30]. In summary, a higher proportion of Streptococcus, Veillonella, and Neisseria may be an indicator of poor oral hygiene conditions. Although, in the good oral hygiene group, genera Fusobacterium, Porphyromonas, were enriched. Fusobacterium nucleatum and Porphyromonas gingivalis did not differ signi cantly between groups. These species were gram-negative species that populate the subgingival crevice of the mouth. We speculated that in good oral hygiene, supragingival bacteria that habit on the exposed solid surfaces were wiped off promptly. This led to a higher proportion of subgingival bacteria. Moreover, previous studies targeting stable and unstable TADs unveiled a slight elevation, though not signi cant, in the prevalence and quanti cation of these periodontopathic pathogens in successful TADs [31], which indicate the presence of periodontal pathogens alone might not induce signi cant changes to the local environment towards TADs.
Characterizing microbiome function is necessary to broaden our knowledge of how oral hygiene affects the microbiome on the TADs' surface. We used PICRUSt as a substitution method to characterize functional changes, which has been implemented in other sequencing studies [32][33][34]. Microbial activities involved with membrane transport, transcription and signaling molecules and interactions were abundant in the POH group. In dental caries, vigorous microbial metabolism was indicated in the oral bacterial community [34]. In periodontal disease, genes of pathogenic activities were detected in functional prediction, including bacterial motility, energy metabolism, and lipopolysaccharide biosynthesis [32,33]. Nevertheless, in our study, the POH group demonstrated vigorous metabolic activities. No difference among these clearly de ned pathogenic functions was found, indicating a rather mild change in microbial function in poor oral hygiene situations. Demonstrating the role of the key bacteria that encode these functions and setting up the link between these functions and the mobility of TADs will be crucial in future research.
Our study discussed the relationship between oral hygiene and microbiomes on the TADs surface using next-generation sequencing. Meanwhile, our study had a few limitations. The surgical technique applied in the extraction of TADs would unavoidably contact the adjacent soft tissue. This contact might lead to the induction of a small amount of bacterial DNA from other sites of the oral cavity. Besides, the TADs samples were di cult to acquire, and the sample size was relatively limited. Last, each individual also exhibited individual variance in the oral microbiome composition. Considering that oral hygiene affects TADs surrounding microenvironment and could potentially in uence the success of TADs, further researches are needed to explain the mechanism of the oral microbiome and its relations with immobility.

Conclusions
This analysis elucidated the difference in total composition and function of TADs oral microorganisms between patients with good oral hygiene and patients with poor oral hygiene, which highlighted the importance of maintaining good oral hygiene in TADs treatment.    Differential genera between GOH and POH based on LEfSe test. a. The cladogram depicting the differential microbiome between GOH and POH. b. The bar plot depicting differential genera between GOH and POH. c. Violin plots of the differential genera Veillonella, Streptococcus, Neisseria, Fusobacterium and Porphyromonas.

Figure 5
Correlations between the microbiomes. a. Correlation networks of the abundant genera of microbial communities on the TADs. Spearman correlation coe cients > 0.6 and P-values < 0.05 are shown in the network. Green nodes represent GOH-enriched genera and pink nodes represent POH-enriched genera.
Lines in blue between the nodes show positive correlations. Lines in red shows negative correlations.