Clin Oral Invest DOI 10.1007/s00784-014-1200-y
ORIGINAL ARTICLE
Relationship of children’s salivary microbiota with their caries status: a pyrosequencing study S. Gomar-Vercher & R. Cabrera-Rubio & A. Mira & J. M. Montiel-Company & J. M. Almerich-Silla
Received: 16 April 2013 / Accepted: 30 January 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Different dental caries status could be related with alterations in oral microbiota. Previous studies have collected saliva as a representative medium of the oral ecosystem. The purpose of this study was to assess the composition of oral microbiota and its relation to the presence of dental caries at different degrees of severity. Materials and methods One hundred ten saliva samples from 12-year-old children were taken and divided into six groups defined in strict accordance with their dental caries prevalence according to the International Caries Detection and Assessment System II criteria. These samples were studied by pyrosequencing PCR products of the 16S ribosomal RNA gene. Results The results showed statistically significant intergroup differences at the class and genus taxonomic levels. Streptococcus is the most frequent genus in all groups; although it did not show intergroup statistical differences. In patients with cavities, Porphyromonas and Prevotella showed an increasing percentage compared to healthy individuals. Bacterial diversity diminished as the severity of the disease increased, so those patients with more advanced stages of caries presented less bacterial diversity than healthy subjects. Conclusion Although microbial composition tended to be different, the intragroup variation is large, as evidenced by the lack of clear intragroup clustering in principal component analyses. Thus, no clear differences were found, indicating that using bacterial composition as the sole source of S. Gomar-Vercher : J. M. Montiel-Company : J. M. Almerich-Silla (*) Stomatology Department, University of Valencia, C/Gascó Oliag 1, 46010 Valencia, Spain e-mail: [email protected] R. Cabrera-Rubio : A. Mira Department of Genomics and Health, Centre for Advanced Research in Public Health, Valencia, Spain
biomarkers for dental caries may not be reliable in the unstimulated saliva samples used in the current study. Keywords Dental caries . 16S ribosomal RNA . ICDAS II . Saliva . Oral health . Childhood
Introduction Oral health plays an important role in overall health and is strongly influenced by the oral microbiota. Some species promote healthy conditions while others contribute to disease [11]. The composition of the salivary microbiota has been used as a diagnostic marker for oral cancer [16], periodontal disease [8], and dental caries [25]. The makeup of the saliva plays a part not only in susceptibility to caries and in demineralization of the enamel but also in remineralization and resistance to dental caries [19]. Changes in saliva composition can cause parallel destabilization of the oral microbiota, and vice versa, during caries development [20]. Mutans streptococci (MS) are considered one of the main causal agents of dental caries [21]. Analyzing the bacterial plaque by surfaces is not an entirely reliable method for predicting MS prevalence in relation to the total oral bacteria, as a wide variety of bacterial colonies occurs between the surfaces. For this reason, some authors [12] have collected saliva as the medium for analysis. Moreover, other species unrelated to MS such as Veillonella, Lactobacillus, Bifidobacterium, Propionobacterium, low pH non-Streptococcus mutans streptococci, Actinomyces spp., and Atopobium spp. have also been associated with different stages of caries [1, 4], making it necessary to use samples which are representative of the entire oral cavity. Saliva is an easily obtained, non-invasive way to obtain samples of oral bacteria from various sites such as the mucous
Clin Oral Invest
membranes and the supra- and subgingival plaque [3, 7, 17]. Previous studies [9, 21] have taken unstimulated saliva samples as representative of the entire oral ecosystem. Recent studies based on polymerase chain reactiondenaturing gradient gel electrophoresis (PCR-DGGE) profiles show similar microbiota predominance in unstimulated saliva and in supragingival dental plaque [13]. However, this method is very limited as regards the microbial diversity that it can detect and the quantification of differences between samples. Pyrosequencing with the 16S RNA gene is an automated and refined method that assesses the microbiological variety in the oral cavity faster, at a lower cost, and more sensitively than can be achieved with conventional PCR followed by DGGE or cloning, as hundreds or thousands of 16S gene sequences can be obtained from each sample [10]. The purpose of this study was to assess the composition of oral microbiota in unstimulated saliva samples by pyrosequencing PCR products of the 16S ribosomal (r)RNA gene and studying their relation to the presence of dental caries.
Materials and methods Sample size and selection The 2010 oral health survey of schoolchildren in the Valencia region of Spain, which was carried out in accordance with the International Caries Detection and Assessment System II (ICDAS II) criteria [22], examined a random sample of 456 12-year-old children. Following the examination, samples of unstimulated saliva were collected. The intraoral examinations were carried out at the children’s schools and the saliva samples were collected after obtaining the informed consent of the children’s parents. The saliva was collected in the morning, approximately 2 h after food intake followed by brushing, using three sterilepacked ISO 50 diameter paper points which were placed on the floor of the mouth for 30 s. The points were then removed and put into sterile Eppendorf tubes at −20 °C for storage. A total 456 unstimulated saliva samples were distributed in the six groups defined by their dental history of caries according to the different stages of caries found during exploration, based on the ICDAS II criteria. After that, a computer random selection was done in order to obtain a maximum of 20 samples in each group until a total of 110 samples. The final size and distribution of the groups were as follows: Group I (n=20): caries-free without fillings Group II (n=20): caries-free with fillings Group III (n=19): initial caries lesions (ICDAS II codes 1 and 2) without fillings
Group IV (n=18): initial caries lesions (ICDAS II codes 1 and 2) with fillings Group V (n=17): caries lesions with breakdown (ICDAS II codes 3 to 6) with fillings Group VI (n=16): caries lesions with breakdown (ICDAS II codes 3 to 6) without fillings.
Sample processing DNA extraction The DNA from the saliva collected on the paper points was extracted using the RTP® Bacteria DNA Mini Kit (STRATEC Molecular, Berlin, Germany), following the manufacturer’s instructions. DNA amplification and pyrosequencing PCR amplifications were performed (25 cycles of 94 °C— 2 min, 94 °C—10 min, 52 °C—30 s, 68 °C—26 s, 68 °C— 7 min) using universal bacterial primers for the 16S RNA gene (primers 27F and 533R), following Cabrera-Rubio et al. (2012) [4]. Roche 454 FLX Titanium pyrosequencing system adapters A and B had previously been added to the primers [18], which allowed the amplicons to bind to the capture microbeads for sequencing. Each sample was amplified using a forward primer containing a different identification sequence of eight base pairs (bp), which serve as a “barcode” to distinguish the different samples [2]. The PCR products obtained were run on 1.6 % agarose gel and the PCR product was then purified using NucleoFast 96 PCR plates (Macherey-Nagel). “Barcode” primer allowed multiple amplified samples to be mixed in a single pyrosequencing reaction on eighths of a plate. Each eighth produced an average of 68,000 sequences about 400 bp in length, so the combination of 20 samples per well, mixed in equimolar amounts, obtained about 3,500 sequences per sample. Data analysis The sequences were separated, using the “barcodes” to distinguish the samples to which they belonged. Sequences under 250 bp were eliminated from the analysis, as were those that did not meet an average minimum quality (a value of 20). Chimeras were removed before downstream analyses and the sequences were assigned taxonomically using the Ribosomal Database Project (RDP) classifier [5]. The classification was based on the RDP database, with a confidence interval of 80 % [24]. Sequences assigned to photosynthetic bacteria such as Cyanobacteria were removed as they appeared to correspond
Clin Oral Invest
to chloroplasts from vegetable remnants, which are typically amplified by universal bacterial primers for rRNA genes. Rarefaction curves were calculated at the operational taxonomic unit (OTU) level using the RDP classifier, and the Chao1 and Shannon indexes were used to calculate the total number of species-level phylotypes, normalizing samples to the same number of sequences by random sampling of the reads and grouping sequences at 97 % sequence identity with the aRarefactWin program. Univariant statistical analysis was then performed to compare the proportional means, using the Kruskal-Wallis nonparametric test, and for comparison between groups U MannWhitney with the SPSS v.19.0 program. The significance level applied was p
ORIGINAL ARTICLE
Relationship of children’s salivary microbiota with their caries status: a pyrosequencing study S. Gomar-Vercher & R. Cabrera-Rubio & A. Mira & J. M. Montiel-Company & J. M. Almerich-Silla
Received: 16 April 2013 / Accepted: 30 January 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Different dental caries status could be related with alterations in oral microbiota. Previous studies have collected saliva as a representative medium of the oral ecosystem. The purpose of this study was to assess the composition of oral microbiota and its relation to the presence of dental caries at different degrees of severity. Materials and methods One hundred ten saliva samples from 12-year-old children were taken and divided into six groups defined in strict accordance with their dental caries prevalence according to the International Caries Detection and Assessment System II criteria. These samples were studied by pyrosequencing PCR products of the 16S ribosomal RNA gene. Results The results showed statistically significant intergroup differences at the class and genus taxonomic levels. Streptococcus is the most frequent genus in all groups; although it did not show intergroup statistical differences. In patients with cavities, Porphyromonas and Prevotella showed an increasing percentage compared to healthy individuals. Bacterial diversity diminished as the severity of the disease increased, so those patients with more advanced stages of caries presented less bacterial diversity than healthy subjects. Conclusion Although microbial composition tended to be different, the intragroup variation is large, as evidenced by the lack of clear intragroup clustering in principal component analyses. Thus, no clear differences were found, indicating that using bacterial composition as the sole source of S. Gomar-Vercher : J. M. Montiel-Company : J. M. Almerich-Silla (*) Stomatology Department, University of Valencia, C/Gascó Oliag 1, 46010 Valencia, Spain e-mail: [email protected] R. Cabrera-Rubio : A. Mira Department of Genomics and Health, Centre for Advanced Research in Public Health, Valencia, Spain
biomarkers for dental caries may not be reliable in the unstimulated saliva samples used in the current study. Keywords Dental caries . 16S ribosomal RNA . ICDAS II . Saliva . Oral health . Childhood
Introduction Oral health plays an important role in overall health and is strongly influenced by the oral microbiota. Some species promote healthy conditions while others contribute to disease [11]. The composition of the salivary microbiota has been used as a diagnostic marker for oral cancer [16], periodontal disease [8], and dental caries [25]. The makeup of the saliva plays a part not only in susceptibility to caries and in demineralization of the enamel but also in remineralization and resistance to dental caries [19]. Changes in saliva composition can cause parallel destabilization of the oral microbiota, and vice versa, during caries development [20]. Mutans streptococci (MS) are considered one of the main causal agents of dental caries [21]. Analyzing the bacterial plaque by surfaces is not an entirely reliable method for predicting MS prevalence in relation to the total oral bacteria, as a wide variety of bacterial colonies occurs between the surfaces. For this reason, some authors [12] have collected saliva as the medium for analysis. Moreover, other species unrelated to MS such as Veillonella, Lactobacillus, Bifidobacterium, Propionobacterium, low pH non-Streptococcus mutans streptococci, Actinomyces spp., and Atopobium spp. have also been associated with different stages of caries [1, 4], making it necessary to use samples which are representative of the entire oral cavity. Saliva is an easily obtained, non-invasive way to obtain samples of oral bacteria from various sites such as the mucous
Clin Oral Invest
membranes and the supra- and subgingival plaque [3, 7, 17]. Previous studies [9, 21] have taken unstimulated saliva samples as representative of the entire oral ecosystem. Recent studies based on polymerase chain reactiondenaturing gradient gel electrophoresis (PCR-DGGE) profiles show similar microbiota predominance in unstimulated saliva and in supragingival dental plaque [13]. However, this method is very limited as regards the microbial diversity that it can detect and the quantification of differences between samples. Pyrosequencing with the 16S RNA gene is an automated and refined method that assesses the microbiological variety in the oral cavity faster, at a lower cost, and more sensitively than can be achieved with conventional PCR followed by DGGE or cloning, as hundreds or thousands of 16S gene sequences can be obtained from each sample [10]. The purpose of this study was to assess the composition of oral microbiota in unstimulated saliva samples by pyrosequencing PCR products of the 16S ribosomal (r)RNA gene and studying their relation to the presence of dental caries.
Materials and methods Sample size and selection The 2010 oral health survey of schoolchildren in the Valencia region of Spain, which was carried out in accordance with the International Caries Detection and Assessment System II (ICDAS II) criteria [22], examined a random sample of 456 12-year-old children. Following the examination, samples of unstimulated saliva were collected. The intraoral examinations were carried out at the children’s schools and the saliva samples were collected after obtaining the informed consent of the children’s parents. The saliva was collected in the morning, approximately 2 h after food intake followed by brushing, using three sterilepacked ISO 50 diameter paper points which were placed on the floor of the mouth for 30 s. The points were then removed and put into sterile Eppendorf tubes at −20 °C for storage. A total 456 unstimulated saliva samples were distributed in the six groups defined by their dental history of caries according to the different stages of caries found during exploration, based on the ICDAS II criteria. After that, a computer random selection was done in order to obtain a maximum of 20 samples in each group until a total of 110 samples. The final size and distribution of the groups were as follows: Group I (n=20): caries-free without fillings Group II (n=20): caries-free with fillings Group III (n=19): initial caries lesions (ICDAS II codes 1 and 2) without fillings
Group IV (n=18): initial caries lesions (ICDAS II codes 1 and 2) with fillings Group V (n=17): caries lesions with breakdown (ICDAS II codes 3 to 6) with fillings Group VI (n=16): caries lesions with breakdown (ICDAS II codes 3 to 6) without fillings.
Sample processing DNA extraction The DNA from the saliva collected on the paper points was extracted using the RTP® Bacteria DNA Mini Kit (STRATEC Molecular, Berlin, Germany), following the manufacturer’s instructions. DNA amplification and pyrosequencing PCR amplifications were performed (25 cycles of 94 °C— 2 min, 94 °C—10 min, 52 °C—30 s, 68 °C—26 s, 68 °C— 7 min) using universal bacterial primers for the 16S RNA gene (primers 27F and 533R), following Cabrera-Rubio et al. (2012) [4]. Roche 454 FLX Titanium pyrosequencing system adapters A and B had previously been added to the primers [18], which allowed the amplicons to bind to the capture microbeads for sequencing. Each sample was amplified using a forward primer containing a different identification sequence of eight base pairs (bp), which serve as a “barcode” to distinguish the different samples [2]. The PCR products obtained were run on 1.6 % agarose gel and the PCR product was then purified using NucleoFast 96 PCR plates (Macherey-Nagel). “Barcode” primer allowed multiple amplified samples to be mixed in a single pyrosequencing reaction on eighths of a plate. Each eighth produced an average of 68,000 sequences about 400 bp in length, so the combination of 20 samples per well, mixed in equimolar amounts, obtained about 3,500 sequences per sample. Data analysis The sequences were separated, using the “barcodes” to distinguish the samples to which they belonged. Sequences under 250 bp were eliminated from the analysis, as were those that did not meet an average minimum quality (a value of 20). Chimeras were removed before downstream analyses and the sequences were assigned taxonomically using the Ribosomal Database Project (RDP) classifier [5]. The classification was based on the RDP database, with a confidence interval of 80 % [24]. Sequences assigned to photosynthetic bacteria such as Cyanobacteria were removed as they appeared to correspond
Clin Oral Invest
to chloroplasts from vegetable remnants, which are typically amplified by universal bacterial primers for rRNA genes. Rarefaction curves were calculated at the operational taxonomic unit (OTU) level using the RDP classifier, and the Chao1 and Shannon indexes were used to calculate the total number of species-level phylotypes, normalizing samples to the same number of sequences by random sampling of the reads and grouping sequences at 97 % sequence identity with the aRarefactWin program. Univariant statistical analysis was then performed to compare the proportional means, using the Kruskal-Wallis nonparametric test, and for comparison between groups U MannWhitney with the SPSS v.19.0 program. The significance level applied was p
