Treponema denticola as a prognostic biomarker for periodontitis in dogs

Periodontal disease is one of the most common disorders in the oral cavity of dogs and humans. Periodontitis, the irreversible periodontal disease, arises progressively from gingivitis, the reversible inflammatory condition caused by dental plaque. Although the etiology of periodontitis has been widely studied in humans, it is still insufficient for the etiological studies on periodontitis in dogs. Many studies have reported that human periodontitis-related bacteria are putative pathogens responsible for periodontitis in dogs. However, most of these studies have focused on the appearance of a specific microbiome, and most of the cohort studies have insufficient sample sizes to generalize their results. In the present study, subgingival samples collected from 336 teeth were categorized into three groups at first, based on clinical outcomes (healthy, gingivitis, periodontitis). Subsequently, the periodontitis samples were further divided into three subgroups (early, moderate, and advanced periodontitis) according to the degree of periodontal attachment loss. Healthy and gingivitis were grouped as a reversible group, and the three subgroups were grouped as an irreversible group. To investigate trends of periodontopathic bacteria in the samples of dogs, a quantitative real-time polymerase chain reaction (PCR) was performed for quantification of 11 human periodontopathic bacteria as follows: Aggregatibacter actinomycetemcomitans (Aa), Porphyromonas gingivalis (Pg), Tannerella forsythia, Treponema denticola (Td), Fusobacterium nucleatum, Prevotella nigrescens, Prevotella intermedia, Parvimonas micra, Eubacterium nodatum, Campylobacter rectus, and Eikenella corrodens. The PCR results showed that Aa and Pg, the representative periodontopathic bacteria, were not significantly correlated or associated with the periodontitis cases in dogs. However, interestingly, Td was strongly associated with the irreversible periodontal disease in dogs, in that it was the most prevalent bacterium detected from the dog samples. These findings indicate that the presence and numbers of Td could be used as a prognostic biomarker in predicting the irreversible periodontal disease and the disease severity in dogs.


Introduction
Periodontal disease is one of the most common disorders in the oral cavity of dogs and humans [1][2][3][4][5]. Periodontal disease is a group of illness occurring in the supportive tissues of was performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering. The owners of the dogs were informed about the aim of the study and gave permission for sampling of their dogs. During the period January 2019 to February 2020, a total of 336 subgingival clinical samples were collected from 176 dogs. These dogs were aged over 1 year. Dogs were evaluated via a serum biochemical panel and CBC were performed on the day prior to scheduled periodontal treatment. They had not received any antibiotics over the past 3 months. Their body weights were less than 20 kg, a common finding in the veterinary field in South Korea.

Sample collection
Clinical sample collections were performed under general anesthesia. The anesthesia protocol was briefly as follow: the premedication for the general anesthesia was glycopyrrolate (0.01 mg/kg, Mobinul; Myungmoon, Gyeonggi, Korea), subcutaneously, butorphanol (0.1 mg/Kg, Bu; Myungmoon, Gyeonggi, Korea) and midazolam (0.2 mg/Kg, Midacom; Myungmoon, Geyonggi, Korea), intravenously. Propofol (4 mg/Kg, Probio; Myungmoon, Geyonggi, Korea) was administered intravenously for the induction and the anesthesia was maintained with isoflurane f at 1.5~2% and O 2 at 2 L/min followed by the placement of cuffed endotracheal tube. Lactated Ringer solution was administered through IV at a rate of 10 ml/Kg/hr through the procedure. Conductive-fabric patient warming system was placed under the dogs and they were monitored using combination monitoring equipment. Sample collection sites were divided into rostral and caudal teeth of the oral cavity. Target teeth were mandibular and maxillary canine teeth known to be functionally important rostral teeth. The major masticatory teeth, maxillary fourth premolars teeth and mandibular first molars teeth, were also targeted as caudal teeth. Two samples were collected from each dog (one from the rostral target teeth and one from the caudal target teeth). One sample was collected if target teeth were missing in the oral cavity for some reasons.
All samples were obtained using a sterilized paper point International Organization for Standardization #30, which was gently inserted into the subgingival pocket at 6 points around each tooth for 30 seconds and transferred immediately into a sterilized transport tube ( Fig  1A). Six paper points obtained subgingival plaque from six points around a tooth, and all six paper points were placed in a sterilized sample container, then a unique barcode number was assigned to the sample container and analyzed as one sample. Teeth were examined and evaluated by intraoral dental radiography and periodontal probing.

Sample grouping by five stages of periodontal disease (PD)
Intraoral dental radiographs were obtained under general anesthesia and evaluated by the same veterinarian (DH-K) with a standard approach under the same conditions. According to the classification criteria of the American Veterinary Dental College [17], samples were divided into five PD groups based on clinical conditions, such as healthy (PD0), gingivitis (PD1), early periodontitis (PD2), moderate periodontitis (PD3), and advanced periodontitis (PD4). These five-stage PD grouping was based on the degree of attachment loss and gum condition. PD0 and PD1 included subjects without attachment loss. PD2, PD3, and PD4 included subjects with < 25%, 25%~50%, and > 50% attachment loss, respectively. A reversible group comprised PD0 and PD1, and an irreversible was group comprised PD2, PD3, and PD4 ( Fig 1B).

DNA extraction
DNA extraction was performed using an Exgene TM Cell SV kit (GeneAll Biotechnologies, Seoul, South Korea) according to the manufacturer's instructions. The paper point was treated with 180 μL lysozyme at 30 mg/mL and incubated at 37˚C for 30 minutes. Proteinase K solution (20 μL of 20 mg/mL) and 200 μL of buffer BL were added to each sample followed by incubation at 56˚C for 30 minutes and 95˚C for 15 minutes. Subsequently, 200 μL of absolute ethanol was added, and the mixture was transferred into a column and centrifuged at 14,000 rpm for 1 minute. After washing the column with 600 μL of buffer BW, 700 μL of buffer TW was added. Next, 100 μL of buffer AE was used to elute DNA. DNA was quantified with a NanoDrop Spectrophotometer (Thermo Fisher Scientific, Inc., Waltham, MA, USA). DNA samples were stored at -20˚C before use.

Quantitative real-time polymerase chain reaction assay
Targeted oral bacteria and primer/probe set sequence used for quantitative real-time PCR in this study are listed in Table 1. PCR amplification was performed in a reaction volume of 20 μL (Bioneer, Inc., Daejeon, South Korea). PCR cycling was performed using a CFX96™ Real-Time System (Bio-Rad Laboratories Inc., Hercules, CA, USA). Cycling conditions consisted of an initial denaturation step at 95˚C for 5 minutes, followed by 40 cycles of denaturation at 95˚C for 30 seconds, primer annealing at 60˚C for 40 seconds, and primer extension at 72˚C for 30 seconds. After completion of the cycling steps, a final extension step at 72˚C for 5 minutes was performed. The normalized value of expression for each species was calculated as the ratio of the relative copy number of the reference species. The qPCR performed with the primer-probe sets used in this study are as shown in the Table 1, and the data for assessment of specificity and sensitivity of the primer-probe sets detecting 11 bacteria from the subgingival plaques of dogs can be found in Supporting Information section.

Statistical methods
All statistical analyses were performed using the MedCalc Statistical Software version 18.5 (MedCalc Software, Ostend, Belgium). The independence of characteristics of samples was analyzed by independent t-tests. The correlation between the number of bacteria and the severity of the periodontal condition was evaluated by Pearson's correlation coefficient. Associations of each bacterium or combination of bacteria between reversible and irreversible groups were analyzed using logistic regression. The significance level was set at p < 0.05.

Characteristic profiles of dogs that donated the subgingival plaque samples
Characteristic profiles of dogs that donated the subgingival plaque samples are summarized in Table 2. The mean ages were 9.36 and 9.69 years for the reversible and irreversible groups, respectively. Males accounted for 54.7% and 58.3% in the reversible and irreversible groups, respectively. The mean body weights were 6.44 kg and 5.79 kg for the reversible and irreversible groups, respectively. An independent t-test was performed to show a statistical independence between these characteristics of the dogs and the reversible and irreversible groupings based on the dog's periodontal disease conditions. The statistical results showed that these dog's characteristics were not significantly associated with their periodontal disease conditions. Of the total 336 samples, 190 samples (PD0, 70; PD1 120 samples) were classified into the reversible group, and 146 samples (PD2, 77; PD3, 34; PD4, 35 samples) were classified as the irreversible group.

Prevalence of the putative periodontopathic bacteria in dogs
Overall bacterial prevalence. The overall prevalence of the tested 11 bacteria detected in the reversible and irreversible groups is shown in Fig 2. Fn, Pm, and Ec showed a high prevalence over 50%, whereas Aa, Pg, and Pn showed a low prevalence less than 10%.
Prevalence of the putative periodontopathic bacteria in the reversible and irreversible groups. The prevalence of each bacterium was compared between the reversible and irreversible groups (Fig 3A). The prevalence of all bacterial species in the irreversible group was higher than that in the reversible group. Among the tested 11 bacterial species, Aa, Tf, Td, Pn, Pi, Pm, En, and Cr were detected two times more in the irreversible group compared with the PLOS ONE reversible group. Prevotella intermedia showed the biggest difference in prevalence among the reversible and irreversible groups at 5.95 times. Treponema denticola showed the second biggest difference at 5.69 times. Prevalence of bacteria in the reversible and periodontal disease (PD) groups. This study also compared the prevalence of each bacterium between the reversible group and each PD group (Fig 3B). Although the severity of periodontal disease was getting worse from PD2 to PD4, the prevalence of Aa and Pg showed a gradually decreasing trend. In contrast, the prevalence of Fn, Pn, Pi, and Pm continually increased along the increase of PD severity. Bacteria that showed increases over two times in their prevalence for all PD groups were Tf, Td, Pn, Pi, En, and Cr. were Tf, Td, Pn, Pi, En, and Cr showed more than double the difference in the prevalence of all PD groups. The fold change of T. denticola, the biggest change, was 7.68 times between the reversible and PD3 groups. P. intermedia showed the biggest difference between the reversible and PD4 groups at 9.31 times. The bacterium that showed the biggest increase in its prevalence between the reversible group and PD3 group (moderate periodontitis) was T. denticola. P. intermedia showed the biggest difference between the reversible group and PD4 group (advanced periodontitis).

Mean number of bacteria
Overall mean number of bacteria. This study analyzed the mean number of all bacterial species in all groups (Fig 4). The mean number of F. nucleatum was 2.76E+07, which was the largest, whereas the mean number of A. actinomycetemcomitans was 6.21E+02, which was the smallest.
Mean  Mean numbers of all bacterial species were increased in the irreversible group. Numbers of Aa, Pg, Td, Pm, En, and Cr in the irreversible group were higher over five times compared with those in the reversible group. However, Aa might not be associated with irreversible periodontal condition in dogs because of its extremely low overall bacterial number. The bacterium that showed the biggest difference in its number between the two groups was T. denticola (25.14 times). This result indicated a significant correlation between the number of bacteria and periodontitis, an irreversible condition.
Mean numbers of bacteria in the reversible and PD groups. Mean numbers of bacteria in the reversible group and each PD group were compared. Results are shown in Fig 5B. In the comparison between the reversible and PD2 groups, Aa and Td showed significant increases

PLOS ONE
(24.31 times and 22.95 times, respectively). The difference of Td, 32.06 times, was unrivaled and overwhelming in comparison with the PD3 group. In the comparison of the reversible group and the PD4 group, Td and En showed the most overwhelming differences at 23.23 times and 21.62 times, respectively. Aa, Pg, and Tf showed significant differences in their numbers between the reversible and PD2 groups. However, the present study found that the number of bacteria continuously decreased as the severity progressed from PD3 to PD4. Specifically, Aa was not detected in the PD4 group. Conversely, Pm, En, and Cr increased continuously as severity progressed. Compared with the reversible group, the significant difference of Td in its number was observed in all PD groups. These results suggest that the number of bacteria except for a few could be correlated with the severity of periodontal condition, especially in the case of T. denticola.

Correlation between the severity in the irreversible group and the number of bacterial species
Correlation analysis between the severity in the irreversible group and bacterial species was performed. All species, except Aa and Pg, showed reliable positive correlations with the irreversible group. In particular, Td, Pm, and Cr showed moderate correlations with Pearson's r > 0.4 (Table 3).

Association between the number of each bacterial species and periodontal condition
Statistical methods were used to analyze associations between periodontal conditions and the number of bacteria using various comparisons (Table 4). Table 4 represents data that were p > 0.05. As shown in Table 4, the number of individual bacterial species was relatively increased, except for Aa and Pg, in the comparison between the reversible and irreversible

PLOS ONE
groups. Similarly, Aa and Pg showed no significant reliability in associations between the reversible and each PD group. These facts indicated that the number of Aa and Pg had no association with periodontal conditions. Based on this, A. actinomycetemcomitans and P. gingivalis were excluded from further analysis.

Association between the combinatory coexistence and numbers of bacterial species with the PD groups
With exception of Aa and Pg, as shown in Table 5, the combinatory coexistence and numbers of bacterial species among the remaining nine putative periodontopathic bacteria in the five-

PLOS ONE
stage PD groups showed significant associations with the irreversible group and with the respective PD groups. For the comparison between the reversible and irreversible groups, bacterial numbers of Tf, Td, Pi, En, and Ec showed significant associations (p = 0.001, p < 0.001, p = 0.006, p = 0.009, and p = 0.004, respectively). In the case of comparison between the reversible group and the PD2 group, numbers of Tf, Td, En, and Ec (p < 0.001, p = 0.001, p = 0.002, and p = 0.002, respectively) showed significant associations with the PD2 group. For the comparison between the reversible and PD3 groups, numbers of Td, Fn, Pi, Pm, and Ec also showed significant associations with the PD3 group (p < 0.001, p = 0.019, p = 0.001, p = 0.020, and p = 0.043, respectively). Numbers of Td, Pi, and Pm also showed significant associations with the PD4 group in the comparison between the reversible and PD4 groups (p = 0.005, p < 0.001, and p = 0.014, respectively). Numbers of Td, Fn, Pi, and Pm showed significant associations with PD3 + PD4 group in the comparison between the reversible and PD3 + PD4 groups (p < 0.001, p = 0.005, p < 0.001, and p = 0.004, respectively). P. intermedia showed a high significance in the comparison between the reversible and PD3 groups, between the reversible and PD4 groups, and between the reversible and PD3 + PD4 groups. However, T. forsythia showed a significance in the comparison between the reversible and PD2 groups only. Based on the above results, T. denticola seemed to be a superior species to others in all different comparisons analyzed statistically in this study.
This study also statistically analyzed the association between each PD group. Only T. denticola was significantly associated with the PD3 group in the comparison between the PD2 and PD3 groups. P. intermedia was only significantly associated with the PD4 group in the comparison between the PD2 and PD4 groups. However, no bacteria showed significance in the comparison between the PD3 and PD4 groups.

Discussion
Periodontitis is the most prevalent oral disease in humans. It also appears in approximately 44-63.6% of dogs [2]. The severity and prevalence of periodontitis in dogs increase with age [3]. This inflammatory oral disease is observed in approximately 80% of dogs aged 4 years [4], 82% of dogs aged 6-8 years, and 96% of dogs aged 12-14 years [5].
Periodontitis is an irreversible inflammatory periodontal condition that is always accompanied by anatomical destruction in the periodontium. It is classified into three grades as early, moderate, and advanced periodontitis according to the severity of attachment loss [17]. Since the irreversibly destroyed periodontium cannot be reestablished, understanding the etiology of periodontitis is significantly important to predict and prevent the expected irreversible condition. Therefore, this study analyzed the trends of 11 bacterial species in dogs with reversible or irreversible periodontal conditions. Dewhirst et al. [20] have obtained 353 canine oral bacterial taxa from 51 dogs and analyzed them. They found that only 16.4% of oral bacterial taxa were shared between dogs and humans and that 83.6% of identified taxa in dogs were novel. The 11 bacterial species focused in this study were included in 16.4% of oral bacterial taxa sharing with humans in the study. Other studies have also confirmed that canine microbiome is significantly different from human microbiome, revealing that bacteria not found in human subgingival plaque or did not have a high rate of prevalence were detected in dog subgingival plaque with significantly high percentages [11,21]. Wallis et al. [12] have longitudinally researched changes that occur in subgingival bacterial communities during the transition from health through gingivitis to the early stage of periodontitis in dogs. Their study suggested that a reduction in the proportion of health-associated species was a more evident indicator than the appearance of disease-associated species. Although the microbiome and environmental factors of dogs are relatively different from those of humans, most human periodontitis-related bacterial species are suspected as putative periodontitis-related pathogens in dogs [13,14].
Kumar et al. [22] have reported that higher number of species lost or gained is closely related to changed clinical status of the periodontal condition. They also suggested that the appearance of previously undetected bacterial species was not related to periodontitis. Rather, it was the result of microbial succession through fluctuation of newly abundant and/or meager bacterial species in subgingival space [23]. According to Griffen's study, it is significantly compelling to understand changed patterns of the number of putative periodontal pathogenic bacterial species and their associations through the periodontitis process in dogs. However, only a few studies have assessed periodontal pathogenic bacterial species and their associations in dogs. The majority of previous studies have analyzed only the difference in the prevalence of putative pathogens between limited severity of the disease such as health, gingivitis, and early periodontitis. Based on findings of the study by Griffen et al. [23], the prevalence of the 11 bacterial species in dogs, trends in their number changes, and their correlations and associations with irreversible periodontal condition with different severity were evaluated in this study for the first time in the veterinary field.
Nishiyama et al. [14] have assessed the presence of putative periodontal pathogens based on human periodontitis-related bacterial species in 40 dogs. It showed that Pg, Cr, Aa, Pi, Tf, Fn, and Ec were detected in 64%, 36%, 24%, 20%, 20%, 16%, and 12% of 25 dogs with periodontitis, respectively. T. denticola and D. pneumosintes were not detected in that study. Another study has analyzed 73 subgingival samples taken from dogs to assess the associations of red complex bacteria species in dogs [24]. It indicated that dogs with gingivitis or periodontitis were more likely to be infected with T. forsythia and P. gingivalis. Recently, Ö zavci et al. [25] have investigated the periodontal disease-related pathogens of dental plaque from 50 cats and 51 dogs with periodontal disease by PCR analysis. Their findings indicated that the periodontal disease-related pathogens, which showed high prevalence in the periodontal disease of cats and dogs, were P. gingivalis (cats 96%, dogs 88%), P. nigrescens (cats 90%, dogs 57%), and P. gulae (cats 70%, dogs 39%).
However, in this study, A. actinomycetemcomitans and P. gingivalis did not show any reliable correlations or associations with periodontitis. Numbers of all individual bacterial species, except Aa and Pg, in the irreversible group and each PD group were higher than those in the reversible group. Specifically, Aa showed extremely low overall prevalence and bacterial count. This study found that Aa and Pg did not have any reliable association or contribution to periodontal disease in dogs. Except Aa and Pg, the other nine bacteria appeared to contribute to the periodontal process, although their mutual associations remained unclear.
The etiological findings from the periodontitis cases after analysis with the subgingival plaque samples in dogs were different from those that are widely studied in humans. Previous studies have described ecological succession and colonization of putative periodontopathic bacterial complexes in the subgingival plaque in a color-coded manner to demonstrate their abundance and the degree of association with periodontitis, including non-spore-forming anaerobes that usually form bacterial complexes in the subgingival plaque [28,29]. Subsequently, the keystone hypothesis has suggested that certain low-abundance pathogens could orchestrate inflammatory disease by remodeling a normally nonpathogenic microbiota into a dysbiotic one and altering host immune protective mechanisms [30,31]. According to these studies, Pg plays the strongest key role in the initiation and progression of human periodontitis, different from our findings. Aa is also strongly associated with human periodontitis by overwhelming the natural resistance of local host innate defense response and producing inflammatory cytokines that can result in connective tissue and bone loss [32].
One study has suggested that periodontal disease is not caused by a few isolated bacteria but by a community-wide effect of bacteria [12]. To confirm this, bacterial associations were analyzed using a combination of nine bacteria excluding Ag and Pg in this study. In the statistical evaluation of putative symbiotic associations of the combination of these nine bacteria between the reversible and irreversible groups, Tf, Td, Pi, En, and Ec showed significant associations with the irreversible group. Specifically, T. denticola showed a significantly strong reliability.
This study also evaluated the association of the combination of nine bacteria excluding Aa and Pg in the comparison between the reversible and PD groups. Tf, Td, En, and Ec were highly associated with the PD2 group, suggesting that the periosteum showed initial anatomical destruction. In the comparison between the reversible group and the PD3 group involving moderate periodontal destruction, Td, Fn, Pi, Pm, and Ec showed relatively high associations with the PD3 group. In the comparison between the reversible group and the PD4 group, the terminal stage of periodontitis, Td, Pi, and Pm were reliably associated with the PD4 group. Td, Fn, Pi, and Pm also showed strong associations with the PD3 + PD4 group in the comparison between the reversible group and the PD3 + PD4 group, suggesting a clinically more than moderate stage of periodontitis. Tf and En were related to the initial stage of irreversible periodontal condition. Pi and Pm were related to more than moderate stage of irreversible periodontal condition. However, T. denticola had a superior significance in contributing to periodontitis at different severity levels in dogs.
To understand whether any bacteria had strong associations with the progress of the irreversible periodontal condition, this study also analyzed the association between a combination of nine bacteria and progression between the PD groups. T. denticola showed strong association with progression between PD2 and PD3. P. intermedia also showed a reliable association with progression between PD2 and PD4. However, no bacteria showed significant association with progression between PD3 and PD4. Therefore, T. denticola could be a predictable biomarker for the deterioration from early periodontitis to moderate periodontitis. In the aggravation from early periodontitis to advanced periodontitis, P. intermedia can be suggested as a predictable biomarker.
A previous study has investigated Treponema spp. in a gingival plaque from both healthy and periodontitis affecting 11 dogs to describe their occurrence and diversity. According to this phylogenetic analysis, Td and Treponema maltophilum (Tm) were mainly detected in the two groups in the diversity of treponema population. In addition, it indicated that Td and Tm are common in dogs regardless of periodontal status [33]. However, this finding is significantly different from that of the present study, which investigated 176 dogs. Although the overall prevalence of Td was 25.60%, Td showed the second biggest difference at 5.69 times in comparison of the prevalence between the reversible and irreversible group in this study. The mean number of Td in the irreversible group was also significantly higher than that in the reversible group, 25.14 times. Furthermore, this study reveals that T. denticola is most significantly correlated and associated with the irreversible group (periodontitis) in the above various statistical comparisons.
Several studies have revealed that Porphyromonas gulae is predominant in periodontal disease in dogs instead of P. gingivalis. Previous studies have noted that P. gulae is rarely found in humans but is commonly detected in animals with active periodontitis [7,34]. P. gulae is distinct from other related genera of P. gingivalis, which is known as a keystone pathogen in human periodontitis. Recently, it has been reported that P. gulae might be associated with periodontitis in dogs [35,36]. However, it is not clearly understood how P. gulae contributes to periodontal disease progression or how it cooperates with other bacterial species. Other studies have also shown that bacteria not found in human subgingival plaque or do not have a high rate of prevalence are significantly highly detected in subgingival plaque of dogs [11,12]. According to these studies, Porphyromonas cangingivalis, Porphyromonas gingivicanis, P. gulae, Porphyromonas canoris, and Fusobacterium sp. are highly prevalent in subgingival plaque of dogs, unlike in that of human beings. Although they might not contribute to the initiation or the progression of periodontitis in dogs because of their high prevalence through health to early periodontitis, we need to further evaluate the correlations and associations of these highly prevalent bacteria in dogs with periodontal disease, along with nine bacterial species analyzed in the present study.
Out of the putative 11 periodontopathic bacteria tested in this study, A. actinomycetemcomitans and P. gingivalis were not significantly associated in periodontal disease of dogs. However, the remaining nine others were significantly associated with periodontal diseases in dogs. Interestingly, among the nine PD-associated bacteria, T. denticola was the remarkable bacterium found in the irreversible group, which showed a correlation with the disease severity of periodontitis. Therefore, these findings suggest that T. denticola could be used as a prognostic biomarker in predicting the irreversible periodontal disease and the disease severity in dogs.  Table. Test results of qPCR specificity with the primer/probe set. (DOCX)