Smoking-induced subgingival dysbiosis precedes clinical signs of periodontal disease

doi:10.21203/rs.3.rs-1805059/v1

Smoking accelerates periodontal disease and alters the subgingival microbiome. However, the relationship between smoking-associated subgingival dysbiosis and progression of periodontal disease is not well understood. Here, we sampled 233 subgingival sites longitudinally from 8 smokers and 9 non-smokers over 6-12 months, analyzing 804 subgingival plaque samples using 16 rRNA sequencing. At equal probing depths, the microbial richness and diversity of the subgingival microbiome was higher in smokers compared to non-smokers, but these differences decreased as probing depths increased. The overall subgingival microbiome of smokers differed significantly from non-smokers at equal probing depths, which was characterized by colonization of novel minority microbes and a shift in abundant members of the microbiome to resemble periodontally diseased communities enriched with pathogenic bacteria. Temporal analysis showed that microbiome in shallow sites were less stable than deeper sites, but temporal stability was not significantly affected by smoking status. We identified 5 taxa – Olsenella sp. 807, Streptococcus cristatus, Atopobium rimae, Prevotella sp. 301 and 308 that were significantly associated with progression of periodontal disease. Taken together, these results suggest that subgingival dysbiosis in smokers precedes clinical signs of periodontal disease, and support the hypothesis that smoking accelerates subgingival dysbiosis to facilitate periodontal disease progression.

Subgingival microbiome

smoking

periodontal disease

temporal dynamics

dysbiosis

Periodontitis is a polymicrobial infection of the gums and teeth-supporting bone affecting nearly 750 million people worldwide ¹. The etiology of periodontitis is multifactorial, but dysbiosis of the subgingival microbiome plays an important role ^2–5. Healthy subgingival space hosts a microbial community generally dominated by Streptococci and other commensal organisms ^4–7. In periodontal disease, the subgingival microbiome shifts towards a more diverse community characterized by putative periodontal pathogens and other gram-negative organisms ^2,4−7. Colonization by these organisms elicits an immune response and inflammation, leading to destruction of the surrounding tissue and tooth loss if left untreated ³. While poor oral hygiene is a major contributor, systemic diseases and other behavioral lifestyle may also disrupt the ecosystem equilibrium, leading to subgingival dysbiosis ^8–10.

Cigarette smoking has wide ranging adverse health consequences. Smokers are at least 50% more likely to develop periodontal disease than non-smokers, and smoking is associated with disease severity in a dose-dependent manner ^11,12. Smoking drastically alters the microbial ecology of the oral cavity through nutrient deprivation, impairment of the immune system, oxygen depletion, and anti-microbial effects ^13,14, thereby shifting the subgingival microbial composition and structure. Cigarette smoke increases the diversity of subgingival communities, reduces the abundance of beneficial bacteria, and favors colonization of pathogenic species ^10,15−20. Thus, smoking may accelerate the development of periodontal disease by altering the subgingival microbiome.

Subgingival microbiome associated with and without periodontal disease in smokers have been investigated in several studies ^10,15−20. However, the longitudinal nature of the subgingival microbiome has not been well characterized. Most published studies have been cross-sectional ^4–6, and thus the organisms associated with clinical progression of periodontal disease in smokers have not been clearly defined. In the present study, we analyzed 804 subgingival plaque samples from 233 unique subgingival sites from 8 smokers and 9 non-smokers at 3–4 time points over 6–12 months. This large longitudinal dataset allowed us to compare and contrast the temporal dynamics of subgingival microbiome in smokers and non-smokers, and identify microbes associated with progression of periodontal disease. To our knowledge, this is one of the most extensive survey of longitudinal subgingival microbiome in smokers to date.

Baseline characteristics of 8 smokers and 9 non-smokers are shown in Table 1. The average mean probing depth, clinical attachment loss, and plaque index were higher in smokers compared to non-smokers (3.88 vs 3.17; 3.58 vs 1.49; 0.926 vs 0.468, respectively). For each subject, subgingival plaque samples from the same sites were collected 3-4 times over 6 to 12 months. A total of 804 samples were sequenced, generating 19,615,615 reads with a mean of 24,398 reads per sample (range = 1,964-740,235 reads per sample). We identified 633 unique OTUs belonging to 10 phyla and 133 different genera. Samples from non-smokers were dominated by five phyla: Firmicutes (30.5%), Bacteroidetes (18.4%), Actinobacteria (17.6%), Fusobacteria (15.5%) and Proteobacteria (14.5%), and the remaining five phyla each comprised less than 2% of the total community. Subgingival microbiome from smokers were dominated by three phyla: Bacteroidetes (31.4%), Fusobacteria (24.5%), and Firmicutes (21.1%). For the subsequent alpha and beta diversity analysis, samples were rarefied to 8,000 reads, which excluded 31 (3.9%) of 804 total samples due to shallow sequencing depths.

We first examined how probing depth and smoking status influenced the richness (Faith’s phylogenetic diversity) and diversity (Shannon diversity) of the baseline subgingival microbiome using linear mixed models (Figure 1). With smokers as the reference group, smoking status (b = -2.610, p = 0.005) and the interaction between smoking status and probing depth (b = 0.430, p = 0.002) were significant predictors of phylogenetic diversity. Smoking status (b = -0.741, p = 0.011) and the interaction between probe depth and smoking status (b = 0.218, p = 0.001) were also significant predictors of Shannon diversity. The probing depth itself was not associated with richness (b = -0.006, p = 0.937) or diversity (b = -0.064, p = 0.065). At shallow probing depths, the phylogenetic diversity of the subgingival microbiome was higher in smokers compared to non-smokers (Figure 1). However, smokers and non-smokers were comparable at greater probing depths.

The differences in alpha diversity in shallow sites prompted us to ask whether there were also differences in the overall subgingival microbial communities between smokers and non-smokers. Using weighted and unweighted UniFrac distance metrics, we observed a modest separation between smokers and non-smokers along the first principal axis at equal probing depths from 2 to 5 mm (Figure 2 and Supplementary Figure 1). Compared to non-smokers, the subgingival communities in shallow sites of smokers shifted (or overlapped) with communities in deeper sites of non-smokers, indicating changes in microbial community structure (or dysbiosis) in smokers at equal probing depths. Due to sparse datasets, comparisons between smokers and non-smokers could not be made for 1 mm and ³ 6 mm sites. The separation between smokers and non-smokers was more pronounced in unweighted UniFrac compared to weighted Unifrac (Figures 2 and Supplementary Figures 1 and 2), suggesting a stronger effect on the minority members of the microbiome. Taken together, these results indicate that, at equal probing depths, the subgingival microbiome of smokers is altered and differs from non-smokers.

To investigate the association between subgingival microbiome, probing depth, smoking status, subject identity, and different subgingival sites, we used distance-based redundancy analysis (db-RDA) on both unweighted and weighted UniFrac distances (Table 2). The db-RDA analysis showed that all the variables examined were significantly associated with microbiome differences for both UniFrac metrics. Smoking and probing depth had strong effects on the subgingival microbiome, separating samples almost exclusively along the first db-RDA axis (Supplementary Figure 3). The effect of subject identity on microbiome was more modest as it separated samples along the second db-RDA axis. Overall, the model explained more variation in the community membership than community structure (Table 2), suggesting that the combination of smoking status, probing depth, subject, and site identity has a greater influence on minority OTUs compared to the dominant OTUs.

To identify specific phyla and OTUs associated with altered subgingival microbiome in smokers, we used the Linear discriminant analysis Effect Size approach (LEfSe). The LEfSe analysis identified seven phyla that were differentially abundant between smokers and non-smokers (Figure 3). Actinobacteria, Firmicutes, and Proteobacteria were overrepresented in non-smokers, and Bacteroidetes, Fusobacteria, Synergistetes, and Spirochaetes were more abundant in smokers. In addition, a total of 99 differentially abundant OTUs with an LDA score greater than 2 were identified (Supplementary Figure 4), including 41 OTUs overrepresented in smokers and 58 OTUs more abundant in non-smokers. The 20 OTUs with the largest LDA scores are shown in Figure 3. The association between smoking and the identified taxa was consistent across both shallow (probing depth < 4mm) and deep (probing depth ≥ 4mm) sites.

The longitudinal study design allowed us to examine factors associated with temporal stability of subgingival microbiome. First, longitudinal data of each subgingival site was collapsed into a single measure by taking the average weighted UniFrac dispersion of each site (i.e. the distance from a sample to the centroid of all samples taken from the same site). Using this approach, a small mean dispersion of UniFrac distance metrics for a given site suggest temporal stability. Linear mixed models were then used to assess the effects of smoking, mean probing depth over time, and standard deviation of the probing depth over time on mean dispersion (or temporal stability). Subject identity was included as a random effect to account for non-independence of multiple sites sampled from a single individual. Linear mixed model analysis showed that mean probing depth (b = -0.014, p = <0.001) was associated with stability of the community structure, but smoking status (b = -0.018, p = 0.101) and the standard deviation of probing depth (b = -0.002, p = 0.743) were not (Figure 4). Thus, subgingival microbiome in shallow sites were more variable, whereas the microbiome in deep sites were more stable over time. However, temporal stability was not affected significantly by smoking status.

To identify specific taxa associated with changes in probing depths over time, subgingival sites were classified into three groups based on changes in probing depth over six months: clinically progressing (∆probing depth ≥ 2 mm), clinically stable (-1 mm ≤ Δprobing depth ≤ 1 mm), and clinically improving (∆probing depth ≤ -2 mm). Of the 233 sites sampled longitudinally, 9 sites (3.9%) progressed by 2 mm or greater, 18 sites (7.7%) improved, and 206 sites (88.4%) were stable (Supplementary Figure 5). The vast majority of baseline sites had a probing depth between 2 mm to 5 mm. Notably, 7 of the 9 progressing sites and 12 of the 18 clinically improved sites were from smokers. Interestingly, many of the progressing sites, irrespective of the baseline probing depth, clustered with sites that had deeper probing depths (Supplementary Figure 6). Next, we used LEfSe to compare sites that progressed to sites that were stable by matching progressed sites with stable sites that had similar baseline probing depths (+/- 1 mm) within each subject. LEfSe analysis identified five OTUs (Olsenella sp., Streptococcus cristatus, two Prevotella sp., and Atopobium rimae) associated with clinical progression, and a Selenomonas species associated with clinical stability (Figure 5). The prevalence (% of samples containing the OTU) and median abundance of these six OTUs was greatest in progressing sites compared to their matched stable sites and all baseline samples (Figure 6). Of the 6 OTUs, Streptococcus cristatus was the most prevalent (72.6% of samples) and abundant (0.16% of reads/sample) phylotype in all baseline sites, and Prevotella species 308DO039 was the least prevalent (30.4% of samples) and least abundant (0.03% of reads/sample). Interestingly, the mean abundance of Selenomonas sp. was greater in stable sites compared to progressing sites, and was more abundant in non-smokers (Supplementary figure 4). These remaining 5 identified taxa were not associated with smoking status except for A. rimae, which was enriched in non-smokers (Supplementary Figure 4).

The adverse impacts of smoking on human health are well documented ²¹. Here, we showed that subgingival dysbiosis is likely another consequence of cigarette smoking. In non-smokers, subgingival microbial communities in shallow sites were considerably less diverse than deep sites. In contrast, shallow sites in smokers had similar diversity as deep sites. Notably, subgingival microbiome in shallow sites of smokers resembled the microbiome dysbiosis in deeper sites of non-smokers. Differential abundance analysis revealed that many taxa associated with smokers have been previously implicated in periodontal disease. However, none of these OTUs were associated with clinical progression of periodontal disease. Longitudinal analysis showed that subgingival microbiome in shallow sites were less stable compared to deeper sites, but the temporal variability was not affected by smoking status. Taken together, our results support the hypothesis that smoking facilitates the development of subgingival dysbiosis associated with periodontal disease.

Consistent with previous studies, we showed that species richness and diversity differ between smokers and non-smokers in shallow sites ¹⁸ but not in deep sites ¹⁷. The subgingival microbiomes of smokers share many similarities to the communities of periodontally diseased individuals. In most host-associated microbiomes, a reduction in microbial diversity is often associated with disease and dysbiosis ²², as organisms are lost and key metabolic pathways are disrupted. Subgingival microbiome differs in that periodontal disease is associated with an increase in microbial richness and diversity ^4,5,7,19. As communities with increased diversity tend to withstand environmental perturbations and pathogen invasion ^23,24, the higher microbial diversity in smokers may withstand dental hygiene practices and commensal colonization, thereby facilitating the development of periodontal disease. Similarly, smoking and periodontal disease may have antagonistic effects on community structure where the impact of smoking decreases as subgingival sites deepen. UniFrac analysis showed that subgingivial microbial communities in shallow sites of smokers resembles the communities in deep sites of non-smokers, which was more pronounced in unweighted compared to weighted analysis. This suggests a disproportionate impact on minority members of the subgingival microbiome, leading to dysbiotic communities that are resilient and more stable over time.

Microbial biomarkers associated with periodontal disease have been well described. Putative periodontal pathogens include members of the red complex (Porphyromonas gingivalis, Tannerella forsythia, and Treponema denticola) and the orange complex (Fusobacterium nucleatum, Fusobacterium periodonticum, Eubacterium nodatum, Parviomonas micra, Prevotella intermedia, Prevotella nigrescens, and several Campylobacter species) ². Among them, Fusobacterium periodonticum was the only organism associated with non-smokers in our study, whereas P. gingivalis, T. forsythia, several F. nucleatum subspecies, and P. nigrescens were associated with smokers (Fig. 3 and Supplementary Fig. 4). Several organisms have been associated with healthy subgingival microbiome ^4,6, many of which were associated with non-smokers in our study, including Granulicatella adjacens, Streptococcus mitis, Streptococcus sanguinis, Streptococcus intermedius, and Veillonella parvula. Recent work has implicated putative periodontal pathogens in systemic diseases. For instance, Fusobacterium nucleatum has been associated with colorectal cancer and adverse pregnancy outcomes (reviewed in ²⁵), and Porphyromonas gingivalis has been associated with different types of cancers (reviewed in ²⁶), Alzheimer’s disease ²⁷, and rheumatoid arthritis ²⁸. Thus, smoking may facilitate or support a microenvironment that favors putative periodontal pathogens, leading to far-reaching effects on the health of the host.

Previous work showed a high degree of inter-individual variations of the healthy subgingival microbiome but relatively low inter-individual variations in the diseased microbiome ⁶. Our longitudinal analysis showed that healthy subgingival microbiome was also characterized by high temporal variation, whereas diseased communities were less variable over time. In our study, temporal microbiome variation did not differ across smoking status, but the probing depth was negatively correlated with temporal variability. After accounting for probing depth, the variation in probing depth was not associated with variation in the microbiome. Interestingly, many of the sites that progressed to disease clustered with deep sites, irrespective of the probing depth at baseline. Thus, these results support the hypothesis that subgingival dysbiosis in smokers precedes clinical signs of periodontal disease, rather than occurring in concert.

Most studies that characterize differences between healthy and diseased subgingival microbiome have been cross-sectional ^4–7. Thus, the causal relationships between microbiome and progression of disease could not be evaluated. Our longitudinal design allowed us to identify specific taxa associated with clinical progression of periodontitis. Despite sampling 233 sites repeatedly from 17 subjects over 6–12 months, only 9 sites progressed by 2 mm or greater. At baseline, these sites varied in probing depths, and some sites progressed from health to disease whereas other sites had periodontal disease at baseline and progressed during the study. We identified several OTUs associated with progression of periodontal disease. Of the five OTUs identified, two OTUs fall within the genus Prevotella, whose members are often associated with periodontitis ^2,4,6. Conversely, Streptococcus cristatus and Atopobium rimae have been shown to be overabundant in the healthy subgingiva ^6,29,30. S. cristatus is a primary adherence point for F. nucleatum and has been shown to suppress immune response to F. nucleatum infection ³¹. Fusobacterium nucleatum can serve as a “bridge species” that aids in the transition from a healthy, commensal-dominated community to a pathogenic one ^31,32. Late colonizers, many of which are pathogenic, cannot incorporate themselves into the subgingival biofilm in the absence of F. nucleatum ³². Thus, a high level of S. cristatus may contribute to disease progression through recruitment and maintenance of F. nucleatum. We note that due to the small number of sites that progressed, we could not distinguish between markers associated progression from early dysbiosis to periodontitis and markers for progression of periodontal disease severity.

There has been considerable debate as to whether subgingival dysbiosis is a local (site-specific) or a global (whole-mouth) event. Earlier studies argued for local changes ^4,33 but later studies suggested a more global process ^5,7,10. Our extensive sampling approach allowed us to compare deep and shallow sites within individuals, and our results suggest that the discrepancies in the literature may reflect methodological rather than biological differences. For instance, PCoA on weighted UniFrac distances separated samples primarily by probing depth, whereas PCoA on unweighted UniFrac distances separate samples by subject identity and smoking status (Fig. 2). This suggests that periodontal disease is associated with shifts in the overall community structure rather than the presence or absence of certain specific bacteria. Thus, unweighted distances that quantify differences in community membership may be imperfect measures for detecting differences across healthy and diseased sites within an individual, and the results of Ganesan et al. ¹⁰ may reflect a strong subject effect rather than the lack of a disease effect. Abusleme et al. ⁵ found that within-subject matched sites that only differed in bleeding on probing did not differ. As a result, probing depth may be a better indicator of subgingival dysbiosis than bleeding on probing. Altogether, subgingival dysbiosis may be site-specific, resulting from local changes in the abundance rather than the presence of different bacteria as the probing depth increases.

This study has several limitations. First, smoking greatly alters the oral environment ^13,14, but whether the microenvironment allow pathogens to outcompete commensals or directly eliminate commensals remains unknown. Second, mechanistic understanding is inherently limited in observational human studies. Third, this study lacked a sufficient number of subgingival sites that progressed clinically, and the sites that progressed primarily came from smokers and were clinically heterogeneous at baseline. Future studies will require full-mouth subgingival sampling in a larger number of periodontally healthy smokers and non-smokers with a much longer follow-up to uncover the successional pattern of dysbiosis and the organisms contributing to or initiating the pathogenic process.

Periodontal disease is a major public health concern. Cigarette smoking disrupts the oral environment and pre-disposes individuals to periodontitis through dysbiosis of the subgingival microbiome. Subgingival communities of smokers are diverse, pathogen-rich, and commensal-poor, but have a similar level of temporal variability as non-smokers. Temporal stability of the subgingival microbiome is modulated by periodontal disease severity. Most notably, subgingival dysbiosis in smokers precedes clinical signs of periodontal disease, supporting the hypothesis that smoking creates a microenvironment that promotes the development of subgingival dysbiosis contributing to periodontal disease. Thus, our study underscores the complex nature of subgingival microbiome and its interaction with environmental gradients. The approach described here should facilitate the design of a larger prospective cohort to further elucidate the transition of microbiome from health to disease.

Subject Recruitment and Sample Collection

Subjects were recruited from the Periodontology Clinic and the DMD Student Dental Clinic at University of Florida College of Dentistry, Gainesville, Florida from March 2012 to April 2013. All subjects were over 18 years of age and had a minimum of 20 natural teeth. Smoking history was obtained by self-report. Smokers were defined as individuals who smoked ≥ 10 cigarettes per day for at least 5 years, and those who have never smoked were non-smokers. Former smokers were not included in this study. Exclusion criteria included diabetes, pregnancy, lactation, systemic antibiotic use within the previous 6 months, periodontal treatment within the previous 12 months, known congenital or acquired immunodeficiency, and use of any immunosuppressive medications.

Clinical measurements (i.e. probing depth, clinical attachment loss, plaque index) were assessed at each visit. Biofilm on the supragingival surface was removed using sterile gauze, and subgingival sites were sampled using sterile endodontic points. For each subject, subgingival sites (range: 7 to 17) were sampled at baseline and the same sites were sampled again at three and six months. Nine of 17 subjects (53%) had additional plaque samples collected from the same sites at 12 months (Table 1). Each sample was transferred to a sterile tube containing storage buffer (MO BIO, Carlsbad, CA), placed immediately on ice, and stored at -80°C until DNA extraction.

DNA extraction, PCR amplification, and Illumina sequencing

Genomic DNA was extracted using the MO BIO PowerSoil DNA extraction kit (Carlsbad, CA) according to manufacturer’s instructions. For each sample, the V1-V3 hypervariable region of the 16S rRNA gene was amplified using composite 27F (5′ AGAGTTTGATCCTGGCTCAG 3′) and 534R (5′ ATTACCGCGGCTGCTGG 3′) primers. PCR reaction mixtures contained 4 microliters of extracted DNA, 100 nM of the forward primer, 100 nM of the reverse primer, and 10 µL of SuperFi PCR master mix (Invitrogen, Carlsbad, CA, USA). 16S rRNA amplicons were analyzed on 1% SYBR Safe agarose gel. Gel slices containing the amplicons were extracted and purified using Qiagen gel extraction kit (Qiagen, Valencia, CA, USA). Purified PCR products were quantified using Qubit HS DNA quantification kit (Invitrogen, Carlsbad, CA, USA) and pooled in equimolar concentration. qPCR was used to quantify the DNA concentration of the pool and prepare the library for sequencing. The use of barcodes allowed for multiplexing and bidirectional sequencing on the Illumina MiSeq platform (Illumina, San Diego, CA, USA).

Data processing

Paired-end reads of 300 nt each covering the V1–V3 hypervariable region of the 16S rRNA gene were processed using custom scripts written in R ³⁴. The reads were filtered based on exact matches to the barcode/primer and an average quality score of 30. Samples were de-multiplexed according to the combination of their unique barcodes (4 to 8 nt long) on each paired end. The barcodes and primers (27F and 534R) were trimmed, and paired-end reads were joined using FLASh ³⁵, with a minimum overlap of 10 bp, to reconstruct the original contiguous amplicon. Reads were assigned reference OTUs using the Human Oral Microbiome Database (HOMD) v.10.1³⁶ and USEARCH alignment with a 97% identity and 80% aligned query threshold. Reads that did not meet the filtering criteria or reference assignment were excluded from subsequent analysis.

OTU tables from multiple sequencing runs were merged, singletons were filtered out, and 11 samples with fewer than 20 total reads were excluded. Relative abundances were calculated from the unrarefied OTU table. The OTU table was then subsampled down to an even sequencing depth of 8,000 reads. Alpha diversity and beta diversity metrics were estimated in QIIME2 (version 2018.8, https://qiime2.org/) using the core-metrics-phylogenetic pipeline. Alpha diversity was measured with species richness (Faith’s phylogenetic diversity) and species diversity (Shannon diversity). Community differences between samples were measured using unweighted and weighted UniFrac distances. Temporal stability of the microbial communities was estimated by the dispersion of weighted UniFrac distances of samples from the same site (mean pocket dispersion) calculated using the betadisper() function in vegan v.2.5-3 ³⁷, grouping by individual sites. Sites with greater mean pocket dispersion indicate more variability over time, whereas sites with lower dispersion are more stable.

Statistical Analyses

Statistical analyses were performed in R v.3.4.2 ³⁴ unless otherwise noted. Differences in alpha diversity across smoking status and probing depth were analyzed with linear mixed models using the lmer() function in lme4 v.1.1–19 ³⁸, with site identity and time point nested within subject identity as random effects. Marginal effects for linear mixed models were calculated using ggpredict() in the ggeffects package v.0.6.0 ³⁹. Principal coordinates analysis (PCoA) of Unifrac distances were used to examine clustering of samples. Statistical significance of environmental variables were tested with distance-based Redundancy analysis (db-RDA) using the dbrda() function in vegan v.2.5-3 ³⁷ with smoking status, probing depth, subject identity, and site identity included as predictors. Linear mixed models were then used to compare temporal stability of the subgingival community structure (mean weighted UniFrac dispersion) across smoking status, mean probing depth, and standard deviation of the probing depth over time. Subject identity was included as a random effect in this analysis to account for multiple sites in the same subject. Differentially abundance analysis was conducted on the unrarefied OTU table with LEfSe (Galaxy version 1.0) ⁴⁰, excluding samples with fewer than 8,000 reads. First, smokers were compared to non-smokers, where probing depth (shallow vs deep sites) and subject identity were accounted for. Then, LEfSe was used to compare clinically progressing sites (∆probing depth ≥ 2) to within-subject matched stable sites (-1 ≤ ∆probing depth ≤ 1) that were within +/- 1 mm of baseline probing depths. All differentially abundant OTUs met the minimum LDA score of 2.

OTU

Operational taxonomic unit

PCoA

Principal Coordinates Analysis

db-RDA

distance-based redundancy analysis

LDA

Linear discriminant analysis

Ethics approval and consent to participate

All subjects provided written informed consent for study participation and procedures. The study was approved by the Institutional Review Board at the University of Florida under project #444-2011. All research was performed in accordance with relevant guidelines/regulations and in accordance with the Declaration of Helsinki.

Consent for Publication

Not applicable.

Availability of Data and Materials

All sequence data are available in DANS under DOI: https://doi.org/10.5281/zenodo.6803209.

Competing Interests

The authors declare that they have no competing interests.

Funding

This work was supported by James & Esther King Biomedical Research Program of Florida Department of Health grant 2KN08, NIH/NIDCR R56 DE02556, and the Gatorade Trust through funds distributed by the University of Florida, Department of Medicine.

Author Contributions

GPW and CW conceived and designed the study. LS and HG collected the clinical specimens. LS, HG, EH, and BH collected the clinical data. KS, JS, LZ, and MK performed the PCR and sequencing. RAT, ECL, GS, and GPW performed and interpreted the analysis. RAT, LS, GPW wrote the manuscript. All authors read and approved the final manuscript.

Acknowledgements

We thank members of Wang laboratory for helpful discussion.

Kassebaum, N. J. et al. Global burden of severe periodontitis in 1990–2010: a systematic review and meta-regression. J Dent Res 93, 1045–1053 (2014).
Socransky, S. S., Haffajee, A. D., Cugini, M. A., Smith, C. & Kent Jr, R. L. Microbial complexes in subgingival plaque. J Clin Periodontol 25, 134–144 (1998).
Kumar, P. S. et al. Changes in periodontal health status are associated with bacterial community shifts as assessed by quantitative 16S cloning and sequencing. J Clin Microbiol 44, 3665–3673 (2006).
Griffen, A. L. et al. Distinct and complex bacterial profiles in human periodontitis and health revealed by 16S pyrosequencing. ISME J 6, 1176 (2012).
Abusleme, L. et al. The subgingival microbiome in health and periodontitis and its relationship with community biomass and inflammation. ISME J 7, 1016–1025 (2013).
Kirst, M. E. et al. Dysbiosis and Alterations in Predicted Functions of the Subgingival Microbiome in Chronic Periodontitis. Applied and Environmental Microbiology 81, 783–793 (2015).
Dabdoub, S. M., Ganesan, S. M. & Kumar, P. S. Comparative metagenomics reveals taxonomically idiosyncratic yet functionally congruent communities in periodontitis. Sci Rep 6, 38993 (2016).
Walker, C. B. et al. The effects of sustained release doxycycline on the anaerobic flora and antibiotic-resistant patterns in subgingival plaque and saliva. J Periodontol 71, 768–774 (2000).
Adriaens, L. M., Alessandri, R., Spörri, S., Lang, N. P. & Persson, G. R. Does pregnancy have an impact on the subgingival microbiota? J Periodontol 80, 72–81 (2009).
Ganesan, S. M. et al. A tale of two risks: smoking, diabetes and the subgingival microbiome. ISME J 11, 2075 (2017).
Tomar, S. L. & Asma, S. Smoking-attributable periodontitis in the United States: findings from NHANES III. J Periodontol 71, 743–751 (2000).
Eke, P. I. et al. Risk indicators for periodontitis in US adults: NHANES 2009 to 2012. J Periodontol 87, 1174–1185 (2016).
Macgregor, I. D. Effects of smoking on oral ecology. A review of the literature. Clin Prev Dent 11, 3–7 (1989).
Palmer, R. M., Wilson, R. F., Hasan, A. S. & Scott, D. A. Mechanisms of action of environmental factors–tobacco smoking. J Clin Periodontol 32, 180–195 (2005).
Haffajee, A. D. & Socransky, S. S. Relationship of cigarette smoking to the subgingival microbiota. J Clin Periodontol 28, 377–388 (2001).
Kumar, P. S., Matthews, C. R., Joshi, V., de Jager, M. & Aspiras, M. Tobacco smoking affects bacterial acquisition and colonization in oral biofilms. Infect Immun 79, 4730–4738 (2011).
Bizzarro, S., Loos, B. G., Laine, M. L., Crielaard, W. & Zaura, E. Subgingival microbiome in smokers and non-smokers in periodontitis: an exploratory study using traditional targeted techniques and a next‐generation sequencing. J Clin Periodontol 40, 483–492 (2013).
Mason, M. R. et al. The subgingival microbiome of clinically healthy current and never smokers. ISME J 9, 268 (2015).
Camelo-Castillo, A. J. et al. Subgingival microbiota in health compared to periodontitis and the influence of smoking. Frontiers in Microbiology 6, 119 (2015).
Karasneh, J. A., Al Habashneh, R. A., Marzouka, N. A. S. & Thornhill, M. H. Effect of cigarette smoking on subgingival bacteria in healthy subjects and patients with chronic periodontitis. BMC Oral Health 17, 64 (2017).
Services, U. S. D. of H. and H. The health consequences of smoking—50 years of progress: a report of the Surgeon General. (2014).
Ma, Z. S., Li, L. & Gotelli, N. J. Diversity-disease relationships and shared species analyses for human microbiome-associated diseases. ISME J 1 (2019).
McCann, K. S. The diversity–stability debate. Nature 405, 228 (2000).
Elmqvist, T. et al. Response diversity, ecosystem change, and resilience. Frontiers in Ecology and the Environment 1, 488–494 (2003).
Brennan, C. A. & Garrett, W. S. Fusobacterium nucleatum—symbiont, opportunist and oncobacterium. Nature Reviews Microbiology 1 (2018).
Atanasova, K. R. & Yilmaz, Ö. Looking in the Porphyromonas gingivalis cabinet of curiosities: the microbium, the host and cancer association. Mol Oral Microbiol 29, 55–66 (2014).
Dominy, S. S. et al. Porphyromonas gingivalis in Alzheimer’s disease brains: Evidence for disease causation and treatment with small-molecule inhibitors. Sci Adv 5, eaau3333 (2019).
Totaro, M. C. et al. Porphyromonas gingivalis and the pathogenesis of rheumatoid arthritis: analysis of various compartments including the synovial tissue. Arthritis Res Ther 15, R66 (2013).
Olsen, I., Johnson, J. L., Moore, L. V. H. & Moore, W. E. C. Lactobacillus uli sp. nov. and Lactobacillus rimae sp. nov. from the human gingival crevice and emended descriptions of Lactobacillus minutus and Streptococcus parvulus. International Journal of Systematic and Evolutionary Microbiology 41, 261–266 (1991).
Kumar, P. S. et al. New bacterial species associated with chronic periodontitis. J Dent Res 82, 338–344 (2003).
Kreth, J., Merritt, J. & Qi, F. Bacterial and host interactions of oral streptococci. DNA Cell Biol 28, 397–403 (2009).
Bradshaw, D. J., Marsh, P. D., Watson, G. K. & Allison, C. Role of Fusobacterium nucleatum and coaggregation in anaerobe survival in planktonic and biofilm oral microbial communities during aeration. Infect Immun 66, 4729–4732 (1998).
Kumar, P. S., Griffen, A. L., Moeschberger, M. L. & Leys, E. J. Identification of candidate periodontal pathogens and beneficial species by quantitative 16S clonal analysis. J Clin Microbiol 43, 3944–3955 (2005).
Team, R. C. R: A Language and Environment for Statistical Computing. (2017).
Magoč, T. & Salzberg, S. L. FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinformatics 27, 2957–2963 (2011).
Chen, T. et al. The Human Oral Microbiome Database: a web accessible resource for investigating oral microbe taxonomic and genomic information. Database 2010, (2010).
Oksanen, J. et al. vegan: Community Ecology Package. (2018).
Bates, D., Maechler, M., Bolker, B. & Walker, S. Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software 67, 1–48 (2015).
Lüdecke, D. ggeffects: Tidy Data Frames of Marginal Effects from Regression Models. J. Open Source Software 3, 772 (2018).
Segata, N. et al. Metagenomic biomarker discovery and explanation. Genome Biol 12, R60 (2011).

Table 1. Baseline characteristics of study participants. The months when samples were taken are shown in parenthesis separated by slashes. For clinical measurements, mean ± standard deviation is shown.

Subject	Smoking Status	Number of Time Points (month sampled)	Number of sites	Probe Depth (mm)	Clinical Attachment Loss (mm)	Plaque Index	Proportion of samples with probe depth > 3 mm (%)
AB	Non-smoker	4 (0/3/6/12)	16	2.6 ± 0.5	0.0 ± 0.0	0.1 ± 0.3	2%
AC	Non-smoker	3 (0/3/6)	13	3.0 ± 0.4	0.1 ± 0.2	0.3 ± 0.7	5%
AD	Non-smoker	4 (0/3/7/14)	7	3.9 ± 1.2	4.5 ± 3.8	0.3 ± 0.5	50%
AH	Non-smoker	4 (0/3/6/12)	12	3.1 ± 0.5	0.0 ± 0.2	0.5 ± 0.6	17%
AJ	Non-smoker	4 (0/3/6/12)	16	3.2 ± 0.6	0.4 ± 0.6	0.9 ± 0.8	25%
AL	Non-smoker	4 (0/3/6/12)	17	3.3 ± 0.8	2.1 ± 1.4	0.8 ± 0.9	30%
AM	Non-smoker	4 (0/3/6/15)	15	3.2 ± 0.7	0.5 ± 0.5	0.8 ± 1.0	33%
AX	Non-smoker	3 (0/6/9)	15	2.7 ± 0.5	1.9 ± 2.4	0.5 ± 0.8	0%
CC	Non-smoker	3 (0/4/8)	15	3.1 ± 0.9	1.0 ± 1.2	0.6 ± 0.9	24%
AA	Smoker	4 (0/?/?/?)	14	3.9 ± 1.1	2.7 ± 2.3	0.6 ± 0.7	60%
AR	Smoker	4 (0/3/6/14)	16	4.4 ± 1.9	3.9 ± 2.2	1.2 ± 0.8	67%
AS	Smoker	4 (0/3/7/13)	9	4.3 ± 1.1	2.9 ± 1.8	1.1 ± 0.8	83%
AT	Smoker	3 (0/4/8)	15	4.9 ± 1.4	5.3 ± 2.2	0.8 ± 0.7	80%
AU	Smoker	4 (0/3/6/12)	10	3.4 ± 1.0	0.7 ± 0.9	0.9 ± 1.0	45%
CE	Smoker	3 (0/5/8)	15	2.1 ± 0.6	0.7 ± 0.6	0.7 ± 0.8	0%
VJ	Smoker	3 (0/4/8)	13	3.4 ± 1.3	4.4 ± 2.4	2.1 ± 0.8	44%
WH	Smoker	3 (0/4/7)	15	4.3 ± 1.1	6.0 ± 2.0	1.0 ± 0.9	84%

Table 2. Distance-based redundancy analysis (db-RDA) on UniFrac distances of subgingival microbiome samples. db-RDA was constrained by smoking status, probing depth, and subject identity. F statistics and p values were generated through ANOVA like permutation tests using 999 permutations.

	Predictors	F statistic	P value	Adjusted R²
Community membership (Unweighted UniFrac)	Probe depth	219.84	0.001	0.553
	Smoking status	69.85	0.001
	Subject identity	34.14	0.001
	Site identity	1.79	0.001
Community structure (Weighted UniFrac)	Probe depth	181.78	0.001	0.493
	Smoking status	83.78	0.001
	Subject identity	19.31	0.001
	Site identity	1.98	0.001

No competing interests reported.

SupplementaryInformation.pdf

Smoking-induced subgingival dysbiosis precedes clinical signs of periodontal disease

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Methods

Subject Recruitment and Sample Collection

DNA extraction, PCR amplification, and Illumina sequencing

Data processing

Statistical Analyses

Abbreviations

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1