Salivary Microbiome in Children With Down Syndrome

Down syndrome (DS), a most frequently occurring genetic disorder, is associated with oral morphological abnormalities and higher incidence rates of oral diseases. Recent studies have analyzed the oral microbiome to elucidate their relationships with oral diseases and general health; however, reports on the oral microbiome in individuals with DS are scarce. This study aimed to characterize the oral microbiome in children with DS. The salivary microbiomes of children with DS (DS) and children without DS (ND) aged 1 to 12 years were compared. Results of culture and quantitative polymerase chain reaction (qPCR) detection tests for cariogenic and periodontopathic bacteria showed no signicant differences in the detected bacteria between the DS and ND groups, with the exception of a signicantly higher detection rate of Candida albicans in children with DS with mixed dentition. A comparison of the salivary microbiomes by 16S sequencing showed no signicant difference in α diversity; however, it showed a signicant difference in β diversity. Children with DS had a higher relative abundance of Corynebacterium and Cardiobacterium, and lower relative abundance of TM7. This study provided basic data on the salivary microbiome of children with DS and further identied a characteristic microbiological marker of children with DS.


Introduction
Down syndrome (DS), also known as trisomy 21, is the most common congenital genetic disorder as well as the most common cause of cognitive disorder in humans that affects approximately 12.6 in 10,000 births in the United States (2006-2010) [1], and 12.3 in 10,000 births in England (2013) [2]. The reported incidence in Japan is 22.6 per 10,000 births according to a 2016 survey, which is approximately twice as high as the incidence rates in Western nations [3]. Regarding dental traits, DS is characterized by abnormal shapes or number of teeth, such as congenitally missing, peg-shaped, conic teeth, or teeth with short roots [4][5][6]. Furthermore, it has been reported that subjects with DS develop periodontal diseases early and have a rapid progression, and conversely, they have fewer dental caries [7,8].
The oral cavity is almost sterile at birth. Different oral bacteria are known to colonize and continue to evolve depending on various factors, such as host and environmental factors, including vertical transmission. In particular, S. mutans infection that occurs between 19-31 months after birth, also known as the "window of infection," is associated with the onset of dental caries [17].
Moreover, diversity of the oral microbiome is known to grow rapidly with tooth eruption, and it is reported that the microbiome on the dorsal surface of the tongue or in the subgingival plaque stabilizes in around 2 years [18,19]. However, some other reports have documented that the oral microbiome continues to evolve after 2 years of age [20,21], thereby leaving many unanswered questions. Colonization of strictly anaerobic gram-negative bacteria, which particularly have strong associations with periodontal disease, is believed to increase with age up until puberty. Therefore, elucidating oral microbiome in early infancy to childhood is important in understanding the onset of dental caries and periodontal diseases in children with DS.
This study performed culture and quantitative PCR to detect the pathogenic microorganisms in the saliva of children in the crucial period of oral microbiome formation with the aim to elucidate the characteristics of the oral microbiome of DS, and further performed high throughput sequencing to compare the salivary microbiomes of children with and without DS.

Participants and ethical review
Participants were 27 children with DS living in Tokyo, and 27 age-matched children without DS who were generally healthy and living in Tokyo (ND). The enrolled participants were further divided into children in the PD and MD to form four groups (Table 1).
Children with standard trisomy 21 were selected for this study, and children who took antibiotics within 7 days were excluded; however, complications of DS were not listed in the exclusion criteria.
This study was approved by the ethics committee of the Japanese Society for Oral Health and was conducted according to their guidelines (Approval no.: . This study was conducted from January 2016 to May 2017, and written informed consent was obtained from the parents of all participants.

Salivary collection and questionnaire survey
Saliva in the oral cavity was collected using SalivaBio Infant's Swabs (SIS: Salimetrics, Carlsbad, CA). The SIS was then washed in 2 mL of sterilized saline solution (Otsuka Pharmaceutical Co., Ltd., Tokyo, Japan) and collected, and the aliquot was centrifuged at 16,400g for 3 minutes to obtain the bacterial pellet. Saliva was collected at least 1 hour after tooth brushing, eating or drinking. Before collection of saliva, the parents of the participants were asked to ll out a questionnaire in order to survey the oral hygiene habits of the participants.
3. DNA extraction DNA was extracted from the bacterial pellet using nexttec™ 1-Step DNA Isolation Kit (nexttec Biotechnologie GmbH, Leverkusen, Germany) according to the manufacturer's instructions, and stored at −40°C until it was used for quantitative PCR or next generation sequencing analysis.
Ampli cation in PCR with CFX Connect™Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA) was performed under the following conditions: initial denaturation at 95°C for 30s, followed by 40 cycles at 95°C for 15s, at 58°C for 30s, and at 72°C for 30s.
Results were analyzed with CFX Manager™ Ver.2.1 software (Bio-Rad Laboratories) and the bacterial count was calculated using the calibration curve created with type strains, wherein ≥2 × 10 2 cells/swab was considered positive.
Candida albicans was detected by culture. BD CHROMagar™ Candida medium plates (Becton, Dickinson and Company, Sparks, MD) by a smear of 200 µL of collected and serially diluted SIS, cultured for 3 days in an aerobic condition at 37°C, and the colony forming units were calculated from the count of grown colonies to nd the count of C. albicans per the total bacterial count.

16S Sequencing
16S rRNA sequencing was performed using universal primers (27Fmod and 338R) [49,50]. Speci cally, Ex Taq polymerase (Takara Bio, Shiga, Japan) was used to amplify approximately 20 ng of template DNA with Veriti Thermal Cycler (Life Technologies Japan, Tokyo, Japan) under the following cycling conditions: initial denaturation at 96°C for 2 min, followed by 25 cycles at 96°C for 30 s, at 55°C for 45s, and at 72°C for 1 min.
The PCR product was puri ed with AMPure XP magnetic puri cation beads (Beckman Coulter, CA, USA) and quanti ed with Quant-iT PicoGreen dsDNA Assay Kit (Life Technologies Japan). After quanti cation, mixed samples were prepared by pooling approximately equal amounts of each ampli ed DNA. Samples were sequenced using a MiSeq Reagent Kit V3 (300 × 2 cycles) and a MiSeq sequencer (Illumina, CA, USA), according to the manufacturer's instructions.
An analysis pipeline was used for processing the 16S rRNA gene V1-V2 region, as previously reported [51,52]. Brie y, after multiplexed sequencing of the 16S amplicons, sequences were assigned to samples based on their barcode sequences. Reads with an average quality value <25, inexact matches to both universal primers, and possible chimeric reads were eliminated. Among high-quality reads, 3,000 reads per sample were randomly chosen and used for the comparative microbiome analysis. We sorted selected reads with the average quality value and grouped them into OTUs using the UCLUST (v.5.2.32) algorithm with a 97% identity threshold [53]. Taxonomic assignments for each OTU were made by similarity searching against publicly available 16S database using the GLSEARCH program (v.36.3.8g). The 16S database was constructed from three publicly available databases: [54] Ribosomal Database Project v.10.31, CORE (http://microbiome.osu.edu/ (31January 2017, date last accessed)), and the reference genome sequence database obtained from the NCBI FTP site (ftp://ftp.ncbi.nih.gov/genbank/ (December 2011, date last accessed)). For assignment at the genus and species levels, sequence similarity thresholds of 96% and 97% were applied, respectively [53]. All high-quality 16S V1-V2 sequences were submitted to the DDBJ/EMBL/GenBank database (Accession number DRA012575).

Data analysis
The student's t-test and Fisher's exact test were used for the analysis of the results of the questionnaire survey and detection rates of the pathogenic bacteria. The Mann-Whitney U test was used for comparisons of diversity and ratios of component bacteria at the genus and species levels.
PCoA was used to visualize similarities/dissimilarities in microbiome structures from the UniFrac Distance [55]. PERMANOVA was conducted to compare overall microbiome structures.
Differences at p < 0.05 were considered statistically signi cant. All analyses were performed using R software program (v3.4.3).

Results
We collected saliva from 27 children with DS and age-matched 27 children without DS, and compared cariogenic and periodontopathic bacteria and salivary microbiome with quantitative PCR and high throughput sequencing of 16S rRNA.
As shown in Table 1, the 27 participants comprised 12 participants with primary dentition (PD group) and 15 with mixed dentition (MD group), and DS-ND comparisons were made separately in the PD and MD groups. Children who received any antibiotics within the last 7 days were excluded from this study; however, complications of DS were not included in the exclusion criteria. In the PD and MD groups, 8/12 and 10/15 children, respectively, had complications related to DS; in both groups, the most common complication was heart disease, which affected 6/12 and 7/15 of children with PD and MD, respectively. There were no differences in the participants' characteristics with regards to age, sex, or the number of teeth. No signi cant differences were found in questionnaire survey results on oral hygiene in the PD group. The MD group had similar results, except for the percentage of children having "oral hygiene practiced by the parents," which was signi cantly higher in the DS group (Table 1).
Salivary cariogenic and periodontopathic bacteria were detected by quantitative PCR. In the PD group, the detection rates of these pathogenic bacteria were low; S. mutans, P. intermedia, T. forsythia, and T. denticola were detected in only 1-2 children with DS, while none of the pathogenic bacteria were detected in the ND group. While detection rates of pathogenic bacteria in the MD group were higher than those in the PD group, S. sobrinus and P. gingivalis were not detected. There were no differences between the children with and without DS in terms of other detected pathogenic bacteria. However, among children with DS, the detection rate for C. albicans by culture was signi cantly higher in the MD group (Table 2).
Salivary bacterial microbiomes were compared by high throughput sequencing. No signi cant differences were observed between the children with DS and ND in the detected operational taxonomic unit (OTU) count or Shannon's diversity index in the PD or MD group (Fig. 1).  Among the top 30 genera, signi cant differences were observed for 14 genera in the MD group. In addition to genera Corynebacterium and Cardiobacterium mentioned above, eight total genera including Abiotrophia, Lautropia, and Capnocytophaga were also signi cantly more abundant in DS, whereas a total of six genera including Prevotella, Actinomyces, Veillonella, and TM7 mentioned above were signi cantly less abundant in children with DS in the MD stage. Figure 4 and 5 present the bacterial species of which abundance was ≥ 0.1% and differed signi cantly between children with DS and ND (p < 0.05). Signi cant differences were observed for 16 bacterial species in the PD group and for 48 species in the MD group. Gemella haemolysans, Neisseria elongata, Rothia aeria, Rothia dentocariosa, and Corynebacterium durum were signi cantly more abundant in children with DS in both the PD and MD groups. Meanwhile, Actinomyces sp. and Haemophilus sputorum were less abundant in children with DS.

Discussion
In subjects with DS, early development of periodontal diseases has been reported, while the rate of dental caries is reportedly low [7,8].
This study therefore compared the salivary microbiomes of children with and without DS aged 1 to 12 years, which is considered the crucial period for microbiome formation, with the aim to elucidate the relationship between the onset of dental caries and periodontal disease in DS.
Amano et al. and Sakellari et al. detected periodontopathic bacteria in 2-to 13-year-old and 8-to 28-yearold subjects with DS, respectively, using quantitative PCR and DNA-DNA hybridization, respectively. They reported that these bacteria were detected earlier in subjects with DS than in healthy subjects [9,10]. With regards to cariogenic bacteria, Scalioni et al. tested for S. mutans in the saliva of 3-to 12-year-old children with in situ hybridization, and reported lower detection rates of the species in children with DS than in healthy children [11].
This study collected the saliva of 27 children with DS and age-matched controls to test salivary cariogenic and periodontopathic bacteria in these samples using quantitative PCR, and found low detection rates of these pathogens in children with PD (aged 1-4 years), and no differences in detection rates between DS and ND in the MD stage (aged 6-12 years) ( Table 2).
Studies to date have measured periodontopathic bacteria in the subgingival plaque, whereas this study tested them in saliva; this difference may explain why we did not observe a difference because of a lower detection rate of the red complex, which are composed of strictly anaerobic bacteria. Moreover, rates of colonization of cariogenic bacteria in children in Japan have recently decreased because of the increase in awareness of the importance of oral hygiene [22]; this may have played a role in the difference between the ndings of this study and those of the previous studies.
The comparison of bacterial microbiome by 16S rRNA high throughput sequencing did not reveal any difference between DS and ND in α diversity of the microbiomes in both PD and MD groups; however, a signi cant difference was observed in β diversity as shown on the PCoA PLot (Fig. 1, 2).
In both PD and MD groups, Corynebacterium and Cardiobacterium were dominant and TM7 was less abundant among the bacterial genera composing the microbiome in DS. Moreover, there were more genera with signi cant DS-ND differences in the MD stage than in the PD stage, suggesting that the DS-ND differences in salivary microbiomes may widen with age.
Willis JR et al. collected the oral rinse samples from individuals with DS aged 7-55 years for a comparison of the microbiome with healthy controls, and reported that the genera Kingella, Staphylococcus, Gemella, Cardiobacterium, Rothia, and Actinobacillus were more abundant, and Alloprevotella, Atopobium, Candidatus, and Saccharimonas were less abundant in DS [16].
The results of this study also showed higher abundance of Cardiobacterium in children with DS in both the PD and MD groups, suggesting that the dominance of Cardiobacterium is a characteristic of the oral microbiome of DS from a very early stage.
At birth, the oral cavity is practically sterile, while bacterial ora develops under the in uence of factors and events, such as mode of delivery, breast or bottle feeding, eruption of tooth, introduction of solids, and oral hygiene status [23,24]. Among them, the oral hygiene status is one of the major factors of dental caries and onset of periodontal diseases as well as microbiome formation [25].
There are several limitations to this study related to insu cient data collection. First, we could not obtain clinical indices, such as the Oral Hygiene Index and the Gingival Index, and we were unable to study the relationships with the degree of dental plaque accumulation or gingival in ammation. However, the results of the questionnaire survey showed no major differences in the number of teeth or oral hygiene habits ( Table 1), suggesting that the differences in the oral microbiome between children with and without DS are primarily due to differences in their oral environments, including salivary properties.
Many characteristics of the saliva of individuals with DS have been reported, including low secretion, increased oxidative stress, increased secretory IgA, and abnormality of inorganic salts [26][27][28][29][30], and all these factors are likely to affect the salivary microbiome [31,32].
Effects of decreased salivary secretion on the oral microbiome have been discussed in the context of Sjögren's syndrome and side effects of radiotherapy and drug therapies [33][34][35]. Such studies have commonly reported higher detection rates of Lactobacilli and C. albicans; however, the present study showed a low detection rate of Lactobacilli and no signi cant difference between children with DS and ND group. In DS group, the observed high detection rate of C. albicans was consistent with the ndings of the previous studies.
Overexpression of the SOD gene encoded on chromosome 21 has been reported to cause overproduction of hydrogen peroxide and associated hydroxy radicals in subjects with DS than in healthy individuals [36], and accordingly, oxidative stress marker levels in their saliva have been reported to be higher compared to those in healthy individuals [26,27]. Furthermore, hydrogen peroxide produced by some bacterial species in the genus Streptococcus in the oral cavity are known to affect the surrounding microbiome, and hydrogen peroxide produced by S. sanguinis inhibits S. mutans colonization [37,38]. Such increased oxidative stress in the saliva may indeed have an effect on the microbiome.
Among the bacterial species found at signi cantly higher abundance in children with DS in both PD and MD groups, Neisseria elongate, Rothia aeria, Rothia dentocariosa, and Corynebacterium durum are catalase-positive bacteria, except Gemella haemolysans [39]. These bacteria are known to show resistance to high oxidative stress environments and may be a factor of dysbiosis in the salivary microbiome of the group of children with DS. Moreover, the aforementioned C. albicans also has a catalase gene, which may similarly explain their higher abundance in the DS group.
Dental plaque is formed through cell to cell interactions between bacteria from initial colonizers to the late colonizers in the oral cavity [40], and dental plaque adapted to the environment are known to develop in the supragingival and subgingival plaque [41,42].
Khocht A et al. studied 40 bacterial species in the subgingival plaque of adults with DS using checkerboard DNA-DNA hybridization. They reported that Streptococcus sp. (i.e., S. oralis, S. mitis, and S. gordonii), which are initial colonizers, were more abundant in adults with DS than in healthy individuals [43]. S. sanguinis, an initial colonizer, was also signi cantly more abundant in our participants with DS in the MD stage. These Streptococcus bacteria are known to be particularly important in the initial stage of dental plaque formation, and their differences in saliva are thus likely to in uence the constituents in subsequent dental plaque formation. C Xiao et al. analyzed supragingival dental plaque in adult patients with dental caries using 16S pyrosequencing and have reported the characteristic presence of Cardiobacterium and Corynebacterium bacteria in participants without dental caries [44]. Waleed F Janem et al. studied the salivary microbiome of obese children with and without type 2 diabetes and have shown that Lautropia, Corynebacterium, and Cardiobacterium bacteria were detected in association with gingivitis [45].
In this study, the genera Corynebacterium and Cardiobacterium were dominant in the saliva of children with DS in the PD and MD stages, and the genus Lautropia was dominant in children with DS in the MD stage. The observed differences in salivary microbiomes, including the aforementioned differences in Streptococcus sp. affect the composition of plaque bacteria and may be associated with the onset of dental caries or periodontal diseases.
These ndings suggest that the distinct characteristics of the salivary microbiome in subjects with DS from that in healthy individuals may be attributable to several factors; however, we were not able to measure the amount of salivary secretion or oxidative stress status in this study. Moreover, complications of DS or history of antibiotic administration within 1 week of specimen sampling can also affect the salivary microbiome [46], thus these factors should also be analyzed in future studies to get a better understanding of the dysbiosis of the salivary microbiome in DS. Furthermore, associations between the development of dental caries, periodontal diseases, and bacteria constituting dental plaque in the affected sites in subjects with DS should also be analyzed.
In conclusion, qPCR and high throughput sequencing provided basic data on the salivary microbiome in DS, and revealed dysbiosis in the salivary microbiome in children with DS when compared with that in healthy children. Moreover, the dominance of Corynebacterium, Cardiobacterium, and lower abundance of TM7 were identi ed as characteristic microbial markers of dysbiosis in children with DS in both the PD and MD stages.   Bacterial species found in salivary microbiomes with signi cant differences between the group of children with DS and the ND group (in the PD stage) The mean abundance (± SD) of 16 species that showed ≥0.1% abundance and signi cant differences p < 0.05% in the Mann-Whitney U test was considered statistically signi cant. DS: Down syndrome, ND: Non-Down syndrome, PD: primary dentition