The Salivary Microbiome of Diabetic and Non-Diabetic Adults with Periodontal
Disease
Amarpreet Sabharwal,* Kevin Ganley,† Jeffrey C. Miecznikowski,† Elaine M. Haase,*
Virginia Barnes,‡ and Frank A. Scannapieco*
Correspondence: Dr. Frank A. Scannapieco, University at Buffalo, Department of
Oral Biology, School of Dental Medicine,109 Foster Hall, 3435 Main St, Buffalo, NY
14214; E-mail fas1@buffalo.edu; fax: (716)829-3946 (email and fax can be
published)
KEYWORDS: Saliva; Microbiome; Diabetes Mellitus, Type 2; Chronic Periodontitis;
RNA, ribosomal, 16S
Word count: 3630; Tables (1); Figures (3); Supplemental Table (1); References (47)
SHORT TITLE: Oral microbiome in Type 2 diabetes and chronic periodontitis
This is the author manuscript accepted for publication and has undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1002/JPER.18-0167.
This article is protected by copyright. All rights reserved.
SUMMARY: Salivary microbiome diversity decreased in diabetics and increased with
progression of periodontal disease compared to periodontally healthy controls.*
*
Department of Oral Biology, School of Dental Medicine, University at Buffalo,
Buffalo, NY
†
Department of Biostatistics, School of Public Health and Health Professions,
University at
‡
Buffalo, Buffalo, NY
Colgate Palmolive Technology Center, Piscataway, NJ; deceased
This article is protected by copyright. All rights reserved.
Abstract
Background: A comparison of the salivary microbiome of non-diabetic and diabetic
cohorts having periodontal health, gingivitis and periodontitis could reveal microbial
signatures unique to each group that will increase understanding of the role of oral
microbiota in the pathogenesis of disease, and to assist with diagnosis and risk
assessment for both periodontal disease and diabetes.
Methods: A group of individuals diagnosed with Type 2 diabetes (T2D) was
compared to a group without T2D. For both the diabetic and non-diabetic cohorts,
three sub-groups were established: periodontal health, gingivitis, and periodontitis.
Salivary DNA was extracted (n=146), PCR was performed to amplify 16S rRNA
hypervariable region V3-V4, and constructed libraries were sequenced and
subjected to bioinformatic and statistical analyses.
Results: MIcrobiome analysis resulted in 88 different genus level operational
taxonomic units (OTUs) for differential abundance testing. Results were largely
described by two trends. Trend 1 showed OTUs that increased in abundance with
increasing periodontal disease, and in diabetics relative to non-diabetics. Trend 1
OTUs comprised a mix of primarily anaerobic commensals and potential
periodontopathogens. Trend 2 was driven primarily by genera that decreased in
diabetics relative to non-diabetics, which included other anaerobes associated with
periodontal disease. Overall, oral microbial diversity decreased in diabetics and
increased with progression of periodontal disease compared to periodontally healthy
controls.
This article is protected by copyright. All rights reserved.
Conclusions: Although select microbiota increased in both diabetes and periodontal
disease progression, these genera decreased in co-existing diabetes and
periodontal disease. These findings suggest that the genera abundance continues to
change with additional stress imposed by co-existing conditions.
This article is protected by copyright. All rights reserved.
INTRODUCTION
Type 2 diabetes mellitus (T2D), a subclass of diabetes mellitus that is not
insulin responsive or dependent, is a metabolic disorder that is characterized initially
by insulin resistance and hyperinsulinemia and eventually by glucose intolerance,
hyperglycemia and overt diabetes.1 The systemic effects and long-term
complications are serious and can result in significant morbidity and increased risk
for mortality.2 Periodontal disease is often more severe in T2D patients, and
hyperglycemia is thought to be more severe in patients with periodontal disease.3
Both diabetes and chronic periodontal disease have been shown to be associated
with microbiome changes compared to health.4, 5
It has been proposed that microbial dysbiosis, a concept earlier proposed in
the context of the gut microbiota and inflammatory bowel disease6, may influence the
initiation and/or progression of a variety of inflammatory states and chronic diseases.
A comprehensive analysis of the oral microbiome in diabetic patients is required to
better understand how the microbiota relates to both diabetes and periodontal
disease status. This information may also lead to defining microbial signatures for
each disease that may be useful for diagnosis and risk assessment.
Previously, several studies have reported that specific periodontal microbes
are associated with diabetes.7-9 In a recent report on the periodontal status and
salivary microbiome of obese children with and without T2D, differences in salivary
microbial diversity was found to be minimal between groups, with only a few
differences in bacterial genus composition noted between groups.10 Further, a
literature review from 2013 concluded that diabetes does not have a significant effect
This article is protected by copyright. All rights reserved.
on the composition of periodontal microbiota, and there is little evidence that it
influences glycemic control.11 However, conclusions drawn from studies that target
select microbiota are inherently restrictive about comprehensive biofilm composition
and system-level changes.
The microbiome is influenced by selective environmental pressures created
by both local and systemic factors.12 Information obtained through various omics
approaches, when effectively integrated, will allow for a better system level
understanding of a disease process 13 Several studies have demonstrated changes
in the oral and gut microbiome in diabetic patients.4, 7, 14, 15
Increased levels of glucose in saliva of patients with diabetes may influence
microbial diversity.16 A recent study of salivary metabolites in controlled diabetic
patients found an increase in glucose and a significant change in markers of
glycemic control and insulin resistance.17 In participants with periodontal disease, the
concentration of other selected metabolites decreased in diabetics and increased in
non-diabetics. These findings suggest an interaction between diabetes and
periodontal disease. Moreover, in the non-diabetic cohort with periodontal disease,
concentrations of salivary metabolites increased with disease progression
demonstrating a biochemical continuum.17 To extend this metabolomics investigation
using the same cohorts, the aim of the present study is to use 16S rRNA
amplification sequencing to determine if a meaningful difference exists in the
diversity of the salivary microbiome between controlled diabetic and non-diabetic
participants with periodontal health, gingivitis or periodontitis.
This article is protected by copyright. All rights reserved.
MATERIALS AND METHODS
Study Population and Selection Criteria
This cross-sectional study, the details of which have been previously described17,
was approved by the Institutional Review Board at the University at Buffalo and was
conducted in accordance with the Helsinki Declaration. The study population
included adult males and females from the Buffalo, New York area recruited over a
period of 1 year. All participants provided a written informed consent before their
enrollment in the study. Inclusion criteria comprised individuals between 18 to 65
years, in good general health or diagnosed with diabetes, and having a minimum of
20 natural teeth (excluding third molars Individuals were excluded if they were
unable or unwilling to sign the informed consent, showed a history of a medical
condition that required pre-medication prior to dental visits, showed 5 or more
decayed teeth, diseases of the hard or soft tissues, impaired salivary function, were
exposed to antimicrobial drugs within 30 days prior to the first study visit, a history of
uncharacterized systemic disease, were pregnant or nursing, participated in any
other clinical study within 1 week prior to enrollment in this study, currently used
tobacco products, were required to have dental treatment during the duration of the
study, were immune compromised, or received periodontal treatment within the last
30 days.
This article is protected by copyright. All rights reserved.
Diabetes Status
Diabetes status was established by self-reported history. Participants who reported a
history of diabetes were asked to confirm diabetic status by providing blood samples
to measure HbA1c values. An individual with T2D was defined as having HbA1c 
6.5%.
Oral Examination and Subgroup Criteria
For both the diabetic and non-diabetic cohorts, three sub-groups were established:
periodontal health, gingivitis and periodontitis. Participants were stratified according
to the gingivitis index, as described previously.17 Briefly, healthy participants had an
average full mouth Modified Gingival Index (MGI) score below 1.0, with fewer than 3
bleeding on probing (BOP) sites and minimal dental plaque present as scored by the
method of Silness and Löe.18 Participants with gingivitis had an average full mouth
MGI score between 1 and 2 with multiple BOP sites and periodontal pockets < 4 mm.
Participants with periodontitis had an average full mouth MGI of 2.0 or greater with
multiple BOP sites, and 2 or more periodontal pockets with probing depths of 5 mm
or more in at least two quadrants. Measurement of oral indices was performed as
per Lobene et al.19 as follows: 0 = normal, 1 = mild inflammation of the gingival unit
(slight change in color, little change in texture of any portion of the gingival unit), 2 =
mild inflammation of the entire gingival unit, 3 = moderate inflammation (moderate
change in texture, redness, edema and hypertrophy of the gingival unit), 4 = severe
inflammation (marked redness and edema/hypertrophy, spontaneous bleeding or
This article is protected by copyright. All rights reserved.
ulceration) of the gingival unit. Plaque at any level was evaluated. Plaque indices
were assessed for 6 surfaces on each tooth and the plaque index was the average
of the scores obtained from all teeth. The amount of plaque observed on each tooth
was scored 0 to 3 as follows: 0 = no plaque noted; 1 = plaque seen only on the tip of
an explorer passed over the tooth surface; 2 = plaque obvious with the naked eye; 3
= gross deposits of plaque present over the entire tooth.
Study Design and Sample Collection
A single dental examiner evaluated all those enrolled for inclusion/exclusion criteria.
Samples were collected as previously described.17 At the first visit, to minimize the
effect of toothpaste on variations in microbiota, participants received a tube of
standard, non-antimicrobial fluoridated toothpaste with instructions for use. A fasting
blood sample was then collected from participants self-identifying as diabetic to
confirm their diabetic status. Plasma was collected by venipuncture in a sterile blood
collection tube containing EDTA. At this visit, participants were given instructions for
saliva collection to be obtained at the next visit within 30 days. They were to refrain
from eating or drinking (excluding water) from 11:00 PM the previous evening, and to
brush their teeth and entire mouth the evening before saliva donation, but not on the
morning of collection. During the second visit, participants provided a completed
medical history questionnaire, and answered questions to assess compliance with
use of toothpaste previously provided. The oral tissues of each individual were
examined to assess periodontal health status. At this visit, a minimum of 0.5 ml of
This article is protected by copyright. All rights reserved.
unstimulated saliva was collected from each participant in a 50 mL sterile
polypropylene tube, which was immediately frozen in a dry ice bath, and stored at 80C.
Salivary Microbiome Analysis
Frozen saliva samples were thawed and centrifuged at 4C for 15 min at 1500 x g.
Saliva samples were processed for microbiome analysis as described previously.10
The pellet obtained following centrifugation was re-suspended in 300 μL of lysis
solution (20 mg/mL lysozyme in 20 mM Tris-HCl, pH 8.0; 2 mM EDTA; 1.2% Triton
X-100) and incubated at 37°C for 1 h. Viscous samples were dissociated by adding
20 μL of 1M DTT to 300 μL of lysozyme solution, vortexed and incubated at 56°C for
15 min. Salivary DNA was obtained using an automated extraction protocol.§
Amplicon PCR was performed for the 16S rRNA hypervariable region V3-V4 using
forward V3 (5' CCTACGGGNGGCWGCAG 3') and reverse V4 primers (5'
GACTACHVGGGTATCTAATCC 3’). Constructed libraries‖ were sequenced to
obtain paired-end (2 x 300 bp) reads.¶ The raw base call files were processed
automatically to obtain two fastq files (one each for forward and reverse reads) per
sample. # DNA extraction, amplicon PCR, library construction and sequencing were
performed at the genomic core facility.** Paired-end fastq reads were processed
using the software pipeline†† as previously described.20 Briefly, taxonomic
assignments were determined by using a naive Bayesian classifier with the Human
Oral Microbiome Database (HOMD) data files (version 14.5)21 and a 98% bootstrap
This article is protected by copyright. All rights reserved.
confidence threshold. Phylotype clustering was carried out to bin sequences and for
taxonomic classification. Shared and consensus taxonomy files at various taxonomic
levels and associated metadata were used for statistical analysis.†
Statistical Analysis
Clinical data analysis. Participants (n=143) were divided into six groups based on
diabetic and periodontal health status: periodontal health and non-diabetic (C);
gingivitis and non-diabetic (G); periodontitis and non-diabetic (P); periodontal health
and diabetic (D); gingivitis and diabetic (DG); and periodontitis and diabetic (DP).
Group comparisons for age, body mass index (BMI), and hemoglobin A1c (HbA1c)
were performed using ANOVA at level 0.05. Gender differences were assessed
using Fisher’s exact test at level 0.05.
Prior to microbiome data analysis, exploratory analysis was performed to
determine outlying samples based on library size at the genus level. Participants with
raw library sizes below 100 (n=3) were considered to be quantitatively inadequate
§
QIAsymphony SP Complex-200_V6_DSP protocol, QIAGEN Sciences,
Germantown, MD
‖
16S Metagenomic Sequencing Library Preparation, Illumina, San Diego, CA
¶
Illumina MiSeq using MiSeq reagent kit V3, Illumina, San Diego, CA
#
MiSeq Reporter Generate FASTQ workflow, Illumina, San Diego, CA
**University at Buffalo Genomics and Bioinformatics Core, Buffalo, NY
††
MiSeq workflow with the MOTHUR pipeline, Illumina, San Diego, CA
This article is protected by copyright. All rights reserved.
and were removed from the analysis. Additionally, exploratory analysis to determine
outlying operational taxonomic units (OTUs) was based on the variance of an OTU.
A variance less than 1, or a low overall read count (less than 100) across all
participants or being absent (read of 0) in more than 100 participants were used as
criteria for exclusion from differential expression analysis but were retained for alpha
and beta diversity analyses.
Statistical analysis of microbiome composition consisted of three stages. The
first stage identified differentially abundant genera (OTUs at the genus level)
significantly associated with disease states via the approach originally taken in
Robinson and Smyth 2007 22, and later expanded to a generalized linear model
framework (GLM) in McCarthy, Chen & Smyth 2012.23 A negative binomial GLM was
fit to the count data and likelihood ratio tests were performed to compare abundance
levels between diabetic status levels. The full GLM model included an intercept, age,
BMI, an oral health status variable (0=periodontal health, 1=gingivitis,
2=periodontitis), a diabetes indicator variable (0=non-diabetic, 1= diabetic), and an
interaction between periodontal disease and diabetic status. The reduced GLM
model included the intercept, age, and BMI. A likelihood ratio test was performed to
compare the reduced model versus the full model for each OTU with a BenjaminiHochberg false discovery rate (FDR) control at level 0.05 to determine significance
across the family of OTUs at the genus level.24 A dendrogram was then used to
cluster the significant OTUs to determine groups of OTUs with similar patterns of
significance.
This article is protected by copyright. All rights reserved.
The second stage used alpha diversity to assess group differences. Several
measures of alpha diversity (Observed, Chao, ACE, Shannon, Simpson, Inverse
Simpson) were calculated for each combination of periodontal disease progression
and diabetic status. An ANOVA F-test at level 0.05 was performed to test for
significant differences between the diabetic/non-diabetics, and significant differences
between the periodontal disease groups.
The third stage analyzed beta diversity via the Bray-Curtis dissimilarity
index.25 Tests at level 0.05 to compare diabetes and periodontal disease status were
implemented using the Multi-Response Permutation Procedure (MRPP)26 where a
non-parametric test compared the observed mean group difference to the estimated
distribution of possible differences through permutations of the data. All data
analyses were performed using the R programming language. Two different software
packages were used to perform the differential abundance analyses‡‡ and the alpha
and beta diversity analyses.§ 27
RESULTS
This study extended investigation of samples from previously published
metabolomics investigation.17 From a total of 174 participants originally enrolled, 146
contained sufficient quantity of DNA for microbiome analysis. Furthermore, after the
‡‡
edgeR, Bioconductor software package, release 3.6
§§
phyloseq, R software package, The R foundation for Statistical Computing,
Vienna, Austria
This article is protected by copyright. All rights reserved.
microbiome data processing stage it was found that 3 participants had sequence
read counts too low for taxonomic assignment, and therefore were removed from
consideration in the statistical analysis. The final population under study thus
consisted of 143 participants (n=79 diabetics, n=64 non-diabetics).
Comparison of clinical parameters yielded significant differences between
groups in age and BMI, while no significant gender differences were noted (Table 1).
HbA1c was not significantly different between diabetic sub-groups. Accordingly, to
control for differences in age and BMI, those variables were included in the reduced
and full GLM models.
Stage 1 Results
The microbiome OTU pruning resulted in 88 different genus level OTUs for
differential abundance testing (see supplementary Table1 in online Journal of
Periodontology). The likelihood ratio test for each OTU examined the significance of
the oral health status variable, the diabetes status variable or their interaction. This
resulted in 30 genus level OTUs that were significant when controlling the false
discovery rate (FDR) at 0.05. A significant OTU indicates significant increasing or
decreasing abundance when comparing groups (Healthy to Gingivitis to
Periodontitis), or that there was a difference in abundance between non-diabetics or
diabetics (diabetic status variable), or that there was a significantly different
progression of increase or decrease in abundance in periodontal disease (Healthy to
Gingivitis to Periodontitis), compared to non-diabetics and diabetics (interaction
variable). The results are summarized in a heat map of these coefficients (Figure 1).
This article is protected by copyright. All rights reserved.
The brightest bands (bright red or bright green) indicate significantly large
coefficients for that variable.
The dendrogram results at the genus level suggest statistically
significant OTUs are largely described by two trends (Figure 1). Trend 1 shows
OTUs that increase in abundance with increasing periodontal disease (red
bands in the oral health variable), and also show increased abundance in
diabetics relative to non-diabetics (red bands in the diabetic variable). In the
interaction variable for this trend, Bifidobacterium and Abiotrophia
demonstrated a decrease in abundance with progressing periodontal disease
in the diabetic group (strongly green in the interaction variable). Trend 2 is
driven primarily by genera that decrease in diabetics relative to non-diabetics,
as shown by green bands in the diabetic variable. In this trend, Fretibacterium,
Peptostreptococcaceae, Filifactor, Treponema increased in abundance with
progressing periodontal disease.
Stage 2 and 3 Results
Microbial diversity at the genus level, both within (alpha) and between (beta) groups
were measured for each participant (n=143). The Observed, Chao1, ACE, and
Shannon metrics of alpha diversity were significant when measured by diabetes
status. Overall, diabetes status was associated with lower alpha diversity (Figure 2).
Similarly, Bray-Curtis dissimilarity, the beta diversity metric used, was significant for
groupings by diabetes status (P=0.002). However, a principle coordinate analysis
(PCoA) plot (Figure 3) showed clusters that differed in variance but with similar
This article is protected by copyright. All rights reserved.
means. For periodontal disease, the Observed, Chao1, and ACE metrics where
significant, indicating that alpha diversity increases with progressing periodontal
disease. The beta diversity metric for oral health status was not significant (p=0.73).
DISCUSSION
This study was undertaken to examine the salivary microbiome of participants with
diabetes and various degrees of severity of periodontal disease. When the clinical
variables age, BMI and gender of the cohorts were compared, diabetic cohorts were
consistently older and had higher BMI than the non-diabetic cohorts (hence age and
BMI were adjusted for in the modeling process). The average age of diabetic cohorts
was 53 years and average BMI was 33.1. These data align with NHANES surveys
(1988-2012), where the largest proportion of T2D were noted in the 45-64 age group.
Similarly, the average BMI was also comparable to previously published NHANES
surveys (1988-2012).28 There were no significant differences in gender between the
diabetic and non-diabetic cohorts.
We compared microbiome diversity within each group (alpha diversity), and
between groups (beta diversity), and by periodontal health status and diabetes
status. In the exploration of OTUs by abundance at the genus level, it was noted that
a distinction between diabetic and non-diabetic participants drove the largest number
of significant group differences. These results, together with the previously published
metabolomic data17, suggest that diabetes reduces the diversity of the oral
microbiome. Similarly, in exploring microbiome alpha diversity, the strongest effect
This article is protected by copyright. All rights reserved.
was imposed by diabetes status, where diabetics tended to have an overall decrease
in diversity compared to non-diabetics. While the beta diversity metric was significant
for groupings by diabetic status, the PCoA plot showed clusterings that were
generally centralized. It is possible that the modest sample size from the same
ecological environment explains the close clustering in the PCoA plot. Similarly, the
significant MRPP tests might be more an indicator of the size of the clusterings
rather than their location.
Oral microbiome diversity is affected by several factors. In this study, we
demonstrate a significant reduction in bacterial diversity in the saliva of adult
diabetics, contrary to our previous study which did not find significant differences in
in adolescents with T2D and obesity compared to periodontally healthy controls.10
These results also differ from those reported by Casarin et al.7, who found that
microbial diversity was greater in participants with diabetes compared to non-diabetic
participants. This could be explained by the fact that only uncontrolled diabetic
participants studied by Casarin at al., while the diabetics in the present study were
controlled by the use of medication.
Periodontal disease is characterized by destruction of the periodontium
secondary to an immune response to periodontopathogenic biofilm in a susceptible
host.29 Various studies have shown that in periodontal disease, there is a
significantly higher alpha diversity in periodontitis patients, supporting our findings.3032
Additionally, there was increased diversity in deeper periodontal pockets and the
difference was statistically significant when compared to shallow sites.33 We show in
the present study an increase in diversity from health to gingivitis to periodontitis.
This article is protected by copyright. All rights reserved.
Our results, together with Ge et al.33, indicate that perhaps the continuum of increase
is linear with progression of disease and severity. However, while the diversity was
higher in periodontitis, there was less inter-individual variation in periodontal disease
than in healthy samples.34
Our previous study found increased glucose concentration in the saliva of
diabetic participants17, an environmental exposure that may alter the composition of
the microbiome. Goodson et al. studied a large population of adolescents and
demonstrated an overall reduction in total salivary bacterial load and in the bacterial
counts of almost every species tested with increasing salivary glucose
concentration.16 Also, dysbiosis of the gut microbiome has been demonstrated in
diabetic patients. 4, 14 In their seminal work on gut microbiota in obesity, Turnbaugh
et al. found a reduction in microbial diversity in obesity, altered metabolic pathways
and alteration in the proportion of Firmicutes to Bacteroidetes.35-37 Here we did not
see a significant alteration in the Firmicutes to Bacteroidetes ratio (data not shown)
in diabetics, presumably due to the difference in ecology of oral and gastrointestinal
microbiome.38 However, we did see alterations in the salivary metabolic pathways in
participants with T2D.17
The oral cavity has one of the most diverse microbiomes when compared to
other microbial niches in the human body. 38 While the oral microbiome in health is
distinct39, there are various local and systemic factors that may play a role in shaping
its diversity and composition.40 In the present study, in Trend 1 an increased
abundance of Olsenella, Mitsukella, Peptidiphaga, Bifidobacterium, Abiotrophia,
Firmicutes, Veillonellaceae, Neisseriaceae, Actinobacterium, and Corynebacterium,
This article is protected by copyright. All rights reserved.
which includes a mix of primarily anaerobic commensals and potential
periodontopathogens, was found in both in the progression to periodontitis and in
diabetics compared to non-diabetics. When these conditions are considered
together, these genera generally decreased in diabetics compared to non-diabetics
with periodontitis. Interestingly, Filifactor and Treponema, genera associated with
periodontal disease, were increased in relation to periodontal disease status, but not
diabetes status. Trend 2 included a mix of other anaerobes that have been
associated with periodontal disease. These findings suggest that the genera
abundance continues to change with additional stress imposed by co-exiting
conditions.
Limitations
These findings also raise a pertinent question: Is the composition of the salivary and
subgingival plaque microbiomes similarly affected in diabetes? Traditionally,
characterization of the oral microbiome of periodontal disease has been done by
sampling subgingival plaque.30, 31, 41 A study of supragingival plaque, as well as tongue
scrapings and saliva samples used metagenomics and 16S rRNA sequencing for a
more comprehensive characterization of the oral microbiome.42 However, due to
ease of sampling, salivary samples provide a distinctive advantage for screening
purposes and have been used in various studies of health and disease. 10, 16, 43-47 In
addition to sampling, the choice of hypervariable regions of the 16S rRNA gene that
have been sequenced have an impact on characterizing the diversity of the oral
microbiome.31 Another potential limitation is lack of sample replicates within our
study design and thus we are not able to assess individual variability. Also, all of the
This article is protected by copyright. All rights reserved.
participants with diabetes in the current study had their blood glucose under control.
Perhaps the addition of individuals with pre-diabetes (i.e. fasting blood sugar 100125 mg/dL) may have provided an additional transitional microbial data.
CONCLUSIONS
Salivary microbiome diversity decreased in diabetics and increased with
progression of periodontal disease compared to periodontally healthy controls.
Select microbiota increased in both diabetes and periodontal disease, but decreased
in co-existing diabetes and periodontal disease. Further studies of the oral
microbiome at the species level may reveal unique signatures associated with both
periodontal and diabetic status.
ACKNOWLEDGMENTS
This study was supported by a grant from Colgate Palmolive Co., New York, NY.
The authors report no conflicts of interest related to this study. At the time of the
study was conducted, Dr. Barnes was an employee of Colgate Palmolive Co., but
had no conflicts of interest related to this study. AS and KG contributed equally to
this work.
This article is protected by copyright. All rights reserved.
REFERENCES
1.
Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes
mellitus and its complications. Part 1: diagnosis and classification of diabetes
mellitus provisional report of a WHO consultation. Diabet Med 1998;15:539553.
2.
Nathan DM. Long-term complications of diabetes mellitus. N Engl J Med
1993;328:1676-1685.
3.
Llambes F, Arias-Herrera S, Caffesse R. Relationship between diabetes and
periodontal infection. World J Diabetes 2015;6:927-935.
4.
Lambeth SM, Carson T, Lowe J, et al. Composition, diversity and abundance
of gut microbiome in prediabetes and type 2 diabetes. J Diabetes Obes
2015;2:1-7.
5.
Teles R, Teles F, Frias-Lopez J, Paster B, Haffajee A. Lessons learned and
unlearned in periodontal microbiology. Periodontol 2000 2013;62:95-162.
6.
Seksik P, Sokol H, Lepage P, et al. Review article: the role of bacteria in
onset and perpetuation of inflammatory bowel disease. Aliment Pharmacol
Ther 2006;24 Suppl 3:11-18.
7.
Casarin RC, Barbagallo A, Meulman T, et al. Subgingival biodiversity in
subjects with uncontrolled type-2 diabetes and chronic periodontitis. J
Periodontal Res 2013;48:30-36.
8.
Lalla E, Kaplan S, Chang SM, et al. Periodontal infection profiles in type 1
diabetes. J Clin Periodontol 2006;33:855-862.
9.
Makiura N, Ojima M, Kou Y, et al. Relationship of Porphyromonas gingivalis
with glycemic level in patients with type 2 diabetes following periodontal
treatment. Oral Microbiol Immunol 2008;23:348-351.
10.
Janem WF, Scannapieco FA, Sabharwal A, et al. Salivary inflammatory
markers and microbiome in normoglycemic lean and obese children
This article is protected by copyright. All rights reserved.
compared to obese children with type 2 diabetes. PLoS One 2017;12;
doi:10.1371/journal.pone.0172647.
11.
Taylor JJ, Preshaw PM, Lalla E. A review of the evidence for pathogenic
mechanisms that may link periodontitis and diabetes. J Periodontol
2013;84:S113-S134.
12.
Filoche S, Wong L, Sissons CH. Oral biofilms: emerging concepts in microbial
ecology. J Dent Res 2010;89:8-18.
13.
McLean JS. Advancements toward a systems level understanding of the
human oral microbiome. Front Cell Infect Microbiol 2014;4;
doi:10.3389/fcimb.2014.00098.
14.
Larsen N, Vogensen FK, van den Berg FW, et al. Gut microbiota in human
adults with type 2 diabetes differs from non-diabetic adults. PLoS One 2010;5;
doi:
15.
Zhou M, Rong R, Munro D, et al. Investigation of the effect of type 2 diabetes
mellitus on subgingival plaque microbiota by high-throughput 16S rDNA
pyrosequencing. PLoS One 2013;8; doi:
16.
Goodson JM, Hartman ML, Shi P, et al. The salivary microbiome is altered in
the presence of a high salivary glucose concentration. PLoS One 2017;12:
doi:
17.
Barnes VM, Kennedy AD, Panagakos F, et al. Global metabolomic analysis of
human saliva and plasma from healthy and diabetic subjects, with and without
periodontal disease. PLoS One 2014;9; doi:
18.
Silness J, Loe H. Periodontal disease in pregnancy. Ii. correlation between
oral hygiene and periodontal condtion. Acta Odontol Scand 1964;22:121-135.
19.
Lobene RR, Weatherford T, Ross NM, Lamm RA, Menaker L. A modified
gingival index for use in clinical trials. Clin Prev Dent 1986;8:3-6.
20.
Schloss PD, Westcott SL, Ryabin T, et al. Introducing mothur: open-source,
platform-independent, community-supported software for describing and
comparing microbial communities. Appl Environ Microbiol 2009;75:7537-7541.
21.
Rundegren J. Calcium-dependent salivary agglutinin with reactivity to various
oral bacterial species. Infect Immun 1986;53:173-178.
22.
Robinson MD, Smyth GK. Moderated statistical tests for assessing
differences in tag abundance. Bioinformatics 2007;23:2881-2887.
This article is protected by copyright. All rights reserved.
23.
McCarthy DJ, Chen Y, Smyth GK. Differential expression analysis of
multifactor RNA-Seq experiments with respect to biological variation. Nucleic
Acids Res 2012;40:4288-4297.
24.
Benjamini Y, Hochberg Y. Controlling the false discovery rate - a practical and
powerful approach to multiple testing. J Roy Stat Soc B Met 1995;57:289-300.
25.
Bray JR, Curtis JT. An ordination of the upland forest communities of southern
Wisconsin. Ecol Monogr 1957;27:326-349.
26.
Mielke P, Berry K. Description of MRPP and additional MRPP applications. In:
Media SSaB, ed. Permutation methods: a distance function approach:
Springer Series in Statistics, 2007:11-123.
27.
McMurdie PJ, Holmes S. phyloseq: an R package for reproducible interactive
analysis and graphics of microbiome census data. PLoS One 2013;8:e61217.
28.
Casagrande SS, Cowie CC. Trends in dietary intake among adults with type 2
diabetes: NHANES 1988-2012. J Hum Nutr Diet 2017;30:479-489.
29.
Kornman KS. Mapping the pathogenesis of periodontitis: a new look. J
Periodontol 2008;79:1560-1568.
30.
Abusleme L, Dupuy AK, Dutzan N, et al. The subgingival microbiome in health
and periodontitis and its relationship with community biomass and
inflammation. ISME J 2013;7:1016-1025.
31.
Griffen AL, Beall CJ, Campbell JH, et al. Distinct and complex bacterial
profiles in human periodontitis and health revealed by 16S pyrosequencing.
ISME J 2012;6:1176-1185.
32.
Li Y, He J, He Z, et al. Phylogenetic and functional gene structure shifts of the
oral microbiomes in periodontitis patients. ISME J 2014;8:1879-1891.
33.
Ge X, Rodriguez R, Trinh M, Gunsolley J, Xu P. Oral microbiome of deep and
shallow dental pockets in chronic periodontitis. PLoS One 2013;8; doi:
34.
Liu B, Faller LL, Klitgord N, et al. Deep sequencing of the oral microbiome
reveals signatures of periodontal disease. PLoS One 2012;7; doi:
35.
Ley RE, Turnbaugh PJ, Klein S, Gordon JI. Microbial ecology: human gut
microbes associated with obesity. Nature 2006;444:1022-1023.
This article is protected by copyright. All rights reserved.
36.
Turnbaugh PJ, Hamady M, Yatsunenko T, et al. A core gut microbiome in
obese and lean twins. Nature 2009;457:480-484.
37.
Turnbaugh PJ, Ley RE, Mahowald MA, Magrini V, Mardis ER, Gordon JI. An
obesity-associated gut microbiome with increased capacity for energy
harvest. Nature 2006;444:1027-1031.
38.
Human Microbiome Project C. Structure, function and diversity of the healthy
human microbiome. Nature 2012;486:207-214.
39.
Zaura E, Keijser BJ, Huse SM, Crielaard W. Defining the healthy "core
microbiome" of oral microbial communities. BMC Microbiol 2009;9; doi:
40.
Wade WG. The oral microbiome in health and disease. Pharmacol Res
2013;69:137-143.
41.
Haffajee AD, Cugini MA, Tanner A, et al. Subgingival microbiota in healthy,
well-maintained elder and periodontitis subjects. J Clin Periodontol
1998;25:346-353.
42.
Galimanas V, Hall MW, Singh N, et al. Bacterial community composition of
chronic periodontitis and novel oral sampling sites for detecting disease
indicators. Microbiome 2014;2; doi:
.
43.
Cameron SJ, Huws SA, Hegarty MJ, Smith DP, Mur LA. The human salivary
microbiome exhibits temporal stability in bacterial diversity. FEMS Microbiol
Ecol 2015;91; doi:
44.
Haririan H, Andrukhov O, Bertl K, et al. Microbial analysis of subgingival
plaque samples compared to that of whole saliva in patients with periodontitis.
J Periodontol 2014;85:819-828.
45.
Ogawa T, Honda-Ogawa M, Ikebe K, et al. Characterizations of oral
microbiota in elderly nursing home residents with diabetes. J Oral Sci
2017;59:549-555.
46.
Stahringer SS, Clemente JC, Corley RP, et al. Nurture trumps nature in a
longitudinal survey of salivary bacterial communities in twins from early
adolescence to early adulthood. Genome Res 2012;22:2146-2152.
47.
Yang F, Zeng X, Ning K, et al. Saliva microbiomes distinguish caries-active
from healthy human populations. ISME J 2012;6:1-10.
This article is protected by copyright. All rights reserved.
Figure Legend
Figure 1. Heat map dendrogram of model coefficients from the generalized
linear model for significant genera. Red/green values indicate an
increase/decrease in genera abundance with the progression to periodontal
disease (Oral Health variable), and as a function of diabetes (Diabetic variable).
The Interaction variable compares the change in trends of genera abundance
(increase or decrease) between progression to periodontal disease (Oral
Health variable) and diabetes (Diabetic variable). Interaction colors indicate the
change in genera abundance with progressing periodontal disease relative to
diabetes: zero (yellow or orange), diabetics = non-diabetics; positive (red),
diabetics > non-diabetics; and negative (green) diabetics < non-diabetics.
Trend 1, genera associated with increasing abundance with periodontal
disease progression and larger abundance in diabetics relative to nondiabetics; Trend 2, genera associated with primarily a decrease in abundance
in diabetics relative to non-diabetics.
This article is protected by copyright. All rights reserved.
Figure 2. Alpha diversity of microbiota. Panel A, oral health status: A significant (p <
0.05) increase in diversity (Observed, Chao1 and ACE) was noted in the progression
to periodontitis. Panel B, diabetes: A significant (p < 0.05) decrease in diversity
(Observed, Chao1, ACE, and Shannon) was noted in diabetics compared to
controls. C, Control; D, Diabetic; G, Gingivitis; P, Periodontitis.
This article is protected by copyright. All rights reserved.
Figure 3. Beta diversity of microbiota. Panel A, oral health status and Panel B,
diabetics. There was a significant difference in beta diversity between
diabetics and non-diabetics (p = 0.002), but not between groups by oral health
status (p = 0.73). C, Control; D, Diabetic; G, Gingivitis; P, Periodontitis.
This article is protected by copyright. All rights reserved.
Table 1.
Clinical
C
G
P
D
DG
DP
variable
(n=19)
(n=22)
(n=23)
(n=26)
(n=26)
(n=27)
Age,
years
(mean±S
D)
(33.8±14.
5)
(35±12.4
)
(49.2±11.
2)
(50.5±10.
9)
(54.1±7.
4)
(54.3±6.
7)
<0.000
1
BMI
(mean±S
D)
(25.7±4.7)
(26.5±5.
5)
(30.6±5.9)
(33.5±9.7)
(32.8±7.
8)
(32.6±6.
8)
0.0003
Gender,
M/F
6/13
8/14
10/13
15/11
11/15
17/10
0.2123
HbA1c,
(mean±SD
)
NA
NA
NA
(7.1±1.1)
(7.3±1.6)
(7.7±1.9)
0.4179
Summary of clinical variables for 143 participants
NA, not applicable
Supplementary Table 1.
This article is protected by copyright. All rights reserved.
P
value