Single nucleotide polymorphisms (SNPs) have been associated with microbial commensal species, and these associations may mean an interplay between overlapping pathophysiological processes. Several SNPs are associated with the gut microbiome (1–6), and among these initially identified associations, some were associated with outcomes such as food intake, hypertension, atopy, chronic obstructive pulmonary disease, body mass index, and lipids (7). Combined analyses of the oral microbiota and genetic variation of the host have been done in humans (8), indicating that genetic variation is an explanation for a stable or recurrent oral microbiome in each individual. These initial analyses unveiled associations between rs1196764 in the locus of the adaptor protein, phosphotyrosine interacting with pleckstrin homology domain and leucine zipper 2 (APPL2) and Prevotella jejuni, Oribacterium uSGB 3339, and Solobacterium uSGB 315 recovered from the tongue dorsum. The analysis also showed that carrying the variant allele of APPL2 rs1196764 and one of the three bacterial species made an individual less likely to have dental calculus.

An opportunity exists to identify disease associations by determining underlying associations between multiple risk factors. Previous work has suggested a number of associations between common genetic variation in humans and dental caries (9, 10), as well as that certain microorganisms are associated with the disease (11). Therefore, we tested for associations between oral microorganisms and genetic variation of a well-characterized cohort of 6- to 19-year-old individuals to unveil unknown associations between oral microbiota and host SNPs.

Data from 92 6- to 19-year-old individuals (12) (Supplemental Material) that had their whole saliva samples characterized using the Ibis Universal Biosensor, which combines polymerase chain reaction (PCR) and mass spectroscopy (MS), to detect species composition in a sample. Dental caries experience data available (DMFT/DMFS scores) referred to caries experience only in the permanent dentition. In brief, individuals were asked to expectorate (spit) 1 ml of unstimulated saliva in a plastic vial. Participants went at least 60 min without eating before sample collection. Saliva samples were immediately placed in a container with ice and then stored at −80°C. Subjects were recruited at the Department of Pediatric Dentistry of the University of Pittsburgh School of Dental Medicine between May and October 2012 as part of the Dental Registry and DNA Repository (DRDR) project. This study was approved by the University of Pittsburgh Institutional Review Board (approval number 0606091). All parents of the participating children provided written informed consent for their participation in the study after they provided written assent for their participation. For DNA extraction, the saliva was placed into a sterile microcentrifuge tube containing ATL lysis buffer (Qiagen, Germantown, MD, cat# 19076) and proteinase K (Qiagen, cat# 19131). Samples were incubated at 56 °C until lysis. One-hundred microliters of a mixture containing 50 μl each of 0.1 mm and 0.7 mm Zirconia beads (Biospec cat# 11079101z, 11079107zx, respectively) were added to the samples which were then homogenized for 10 min at 25 Hz using a Qiagen TissueLyser. Nucleic acid from the lysed sample was then extracted using the Qiagen DNeasy tissue kit (Qiagen cat# 69506). For microbiota analysis, 10 μl of each sample was loaded per well onto the BAC detection PCR plate (Abbott Molecular, cat# PN 05N13-01). The BAC detection plate is a 96-well plate containing 16 primers that survey all bacterial organisms by using the omnipresent loci (e.g., 16S rRNA sequences), while some are targeted to specific pathogens of interest (e.g., the Staphylococcus-specific tufB gene). The plate also includes primers for the detection of Candida species and some antibiotic resistance markers (e.g., mecA, vanA, vanB, and KPC). An internal calibrant of a synthetic nucleic acid template is also included in each assay, controlling for false negatives (e.g., from PCR inhibitors) and enabling a semi-quantitative analysis of the amount of template DNA present. PCR amplification was carried out, and the products were desalted in a 96-well plate format and sequentially electrosprayed into a mass spectrometer. The spectral signals were processed to determine the masses of each of the PCR products present with sufficient accuracy that the base composition of each amplicon could be unambiguously deduced. Using combined base compositions from multiple PCRs, the identities of the pathogens and a semi-quantitative determination of their relative concentrations in the starting sample were established by using a proprietary algorithm to interface with the Ibis database of known organisms. To display the distribution of species per patient, we generated heat map graphics using the “image()” and “grid()” functions in the “graphics” package part of R base distribution (version 3.2.2) (R Foundation for Statistical Computing, Vienna, Austria) (12).

Further, genomic DNA was obtained from these same saliva samples and genotyping data for 43 single nucleotide polymorphisms (SNPs) were generated. These SNPs (Table 1) were selected based on previous data suggesting them to be marking genes that were associated with dental caries (13–20). TaqMan chemistry and endpoint analysis were used to determine genotypes (21). When a particular species was frequent enough in the group of 92 6- to 19-year-old individuals (present in at least 18 individuals), chi-square or Fisher's exact tests were used to test for overrepresentation of alleles in individuals carrying particular species, with an alpha of 0.001 (0.05/43). This cutoff of 20% aimed to avoid multiple comparisons of very small frequencies.

The 92 6- to 19-year-old individuals studied had a mean age of 12.74 years (ages ranged from 6 to 19 years), and 44 individuals were females. Most of them (N = 63) were white, and among the remaining individuals, 21 were Black and 8 were Asian. Twenty individuals were caries-free, and among the remaining individuals that had caries experience, DMFT scores (decayed, missing due to caries, filled teeth) ranged from 1 to 17 and DMFS scores (decayed, missing due to caries, filled surfaces) ranged from 1 to 41 (Supplemental Material). Sixty-seven microbial species were identified in the samples from the 92 6- to 19-year-old individuals studied (Table 2).

Five microbes (Candida albicans, Saccharomyces cerevisiae/paradoxus, Staphylococcus epidermidis, Stomatococcus mucilaginosus, and Streptococcus pneumoniae) that were at least frequent in 20% of the sample were used to create groups for testing for overrepresentation of alleles of the SNPs listed in Table 1, which are located in loci that were previously associated with dental caries (13–20). All genotypes were in Hardy-Weinberg equilibrium. Having Candida albicans was associated with ENAM rs3796704 (p = 0.0006). Half of the individuals colonized with Candida albicans carried the variant allele of ENAM rs3796704, whereas less than 10% of individuals without Candida albicans carried the allele. Being positive for Staphylococcus epidermidis was associated with TUFT rs3828054 (p = 0.001). No individuals colonized by Staphylococcus epidermidis carried the TUFT rs3828054 variant allele. No other comparison showed differences that were statistically significantly different (all data are available as Supplemental Materials). Microorganisms historically associated with caries lesions such as Streptococcus mutans and Lactobacilli were found in just a few individuals and did not provide a chance for testing for associations.

Dental caries can be defined as having a complex mode of inheritance, which includes the influence of more than one gene with small effects and environmental factors. Putative associations between dental caries with certain microbes or with genomic variation have been reported. These results did not translate to new preventive strategies, in part due to the fact that we do not understand how these different contributing factors may interact. Here we report novel associations between microbial species found in the mouth and common genomic variants in the population that have been previously associated with dental caries. These kinds of analyses aim to provide new insight into how the host interacts with microorganisms and how these may impact health and disease.

The success of genome-wide association studies (GWAS) in determining associations between SNPs and disease has motivated the proposal of testing microorganism whole genomes as well. Initial attempts have identified associations between ranges of bacteria, viruses, and protozoa and traits under strong selection such as drug resistance. The challenge of identifying variants under moderate selection still exists, and solutions are needed for instances of increased population stratification in microorganisms due to selection and complex recombination patterns (22). To minimize the chance for spurious associations, we tested SNPs that were previously associated with dental caries, a meaningful oral health microbial-mediated outcome.

Candida albicans is suggested to play a role in the development of dental caries (23), and understanding this would provide useful new insights. We suggested that variation in enamel conformation, depending on genetic variation, may increase the risk of developing dental caries lesions (24). The association between ENAM rs3796704 and the presence of Candida albicans suggests that variation in enamel conformation may increase dental caries susceptibility in the presence of Candida albicans colonization. These possible interactions should be modeled in future studies.

The role of Staphylococcus epidermidis in dental caries is unclear, and it is possible that it exists in the presence of certain genetic variations. The meaning of the many species that can be found in the mouth after molecular technology became available is currently object of study. Interestingly, Streptococcus mutans, which is classically associated with initiation of dental caries lesions, was only detected in one individual out of 92, suggesting that the individuals studied here did not have newly forming dental caries lesions. We showed earlier (12) that when compared, the results between PCR-MS and 16S rRNA sequencing differ significantly. The samples we tested by PCR-MS displayed a range in the number of genera detected from 1 to 4, including both bacteria and fungi. Both PCR-MS and 16s rRNA sequencing detected a prevalence of Streptococci, Neisseria, and Rothia in the samples. Sequencing of the 16S rRNA gene suggests that Prevotella and Veillonella were not detected by PCR-MS; while staphylococci and fungi were not detected by 16S rRNA sequencing. Using 16S rRNA sequencing therefore should provide the opportunity for identification of additional associations.

Larger sample sizes will allow for these analyses to be conducted taking into consideration age, sex, ethnicity, and dental caries severity. These data were available but were not used since initial comparisons were already with small numbers. Sex, age, and ethnicity variations were not accounted for in the comparisons presented here, and undetected population substructures can always unveil spurious associations. Another limitation is that the quantification of microorganisms is relevant data for interpreting associations with diseases and we utilized only representation of species in this report. Future approaches should include measures of quantification of microorganisms. Finally, some SNPs had lower amplification ratios, which also impacted some of the comparisons. In summary, associations between genomic variants that impact dental caries risk may be explained by individual microbial colonization, and genomic and microbiome host data should be considered when testing for associations between dental caries and other risk factors.