smoking decreases structural and functional resilience in ...€¦ · college and research...

Smoking decreases structuraland functional resilience in thesubgingival ecosystemJoshi V, Matthews C, Aspiras M, de Jager M, Ward M, Kumar P. Smoking decreasesstructural and functional resilience in the subgingival ecosystem. JClin Periodontol 2014;41: 1037–1047. doi: 10.1111/jcpe.12300.

AbstractAims: Dysbiotic microbial communities underlie the aetiology of several oral dis-eases, especially in smokers. The ability of an ecosystem to rebound from the dys-biotic state and re-establish a health-compatible community, a characteristicknown as resilience, plays an important role in susceptibility to future disease.The present investigation was undertaken to examine the effects of smoking oncolonization dynamics and resilience in marginal and subgingival biofilms.Materials and methods: Marginal and subgingival plaque and gingival crevicularfluid samples were collected from 25 current and 25 never smokers with pre-exist-ing gingivitis at baseline, following resolution, after 1, 2 4, 7, 14 and 21 days ofundisturbed plaque formation and following resolution. 16S cloning and sequenc-ing was used for bacterial identification and multiplexed bead-based flow cytome-try was used to quantify the levels of 27 immune mediators.Results: Smokers demonstrated an early pathogenic colonization that led to sus-tained pathogen enrichment with periodontal and respiratory pathogens, elicitinga florid immune response. Smokers also demonstrated greater abundance of path-ogenic species, poor compositional correlation between marginal and subgingivalecosystems, and significantly greater pro-inflammatory responses following resolu-tion of the second episode of disease.Conclusions: The ability of the subgingival microbiome to “reset” itself followingepisodes of disease is decreased in smokers, thereby lowering the resilience of theecosystem and decreasing its resistance to future disease.

Vinayak Joshi1,*, Chad Matthews1,†,Marcelo Aspiras2, Marko de Jager2,‡,

Marilyn Ward2 and Purnima Kumar1

1Division of Periodontology, College of

Dentistry, The Ohio State University,

Columbus, OH, USA, 2Philips Oral

Healthcare, Bothell, WA, USA

Present addressess: *Maratha Mandal Dental

College and Research Institute, Belgaum,

India: †Carolina Periodontics and

Endodontics, Rockhill, SC, USA: ‡Philips

AVENT, Eindhoven, The Netherlands

Key words: 16S; bacteria; DNA;

experimental; gingivitis; host-bacterial

interactions

Accepted for publication 11 August 2014

Bacteria colonize the tooth soonafter eruption and establish commu-nities in two distinct, yet geographi-cally connected habitats – thesupragingival area and the subgingi-val sulcus (Listgarten 1976). It is

known that the subgingival and thesupragingival (marginal) biofilms arecompositionally different – a featureknown as community structure; andthat these compositional differenceslead to different functionalities (Page1986). One function is maintainingequilibrium with the host immuneresponse; dysbiosis within this com-munity leads to disruption of thisequilibrium and results in disease(Hajishengallis & Lamont 2012). Theability of an ecosystem to reboundfrom the dysbiotic state andre-establish a health-compatible

community, a characteristic knownas resilience, plays an important rolein susceptibility to future disease(Wardwell et al. 2011).

One of the most common resultsof dysbiosis in the oral cavity is gin-givitis (Sreenivasan et al. Ash Jr.et al. 1964, Listgarten 1976, Listgar-ten et al. 1985). Although this is areversible condition, repeated epi-sodes of gingivitis and persistent gin-gival infection predispose theindividual to irreversible periodontaldisease (periodontitis) and eventualtooth loss (Lang et al. 2009), sug-

Conflict of interest and source of

funding statement

The study was supported by aresearch grant from Philips OralHealthCare. Marcelo Aspiras, MarilynWard and Marko de Jager areemployees of the funding agency.

© 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd 1037

J Clin Periodontol 2014; 41: 1037–1047 doi: 10.1111/jcpe.12300

gesting an aetiological role for lossof resilience in disease causation. Inaddition, gingivitis is a necessaryprecursor to periodontitis (Page &Schroeder 1976, Page 1986). It isestablished that smokers are at high-risk for periodontal disease and thatone in two smokers are likely todevelop extensive and severe disease(Bergstrom 1989, Bergstrom et al.2000, Tomar & Asma 2000). It ispossible that smoking increases therisk for disease by promoting dysbi-osis and decreasing community resil-ience, thereby shifting theequilibrium of host-bacterial interac-tions. However, the effects on smok-ing on these parameters of ecologicalhealth are largely unknown.

We and others have previouslydemonstrated that smoking promotesacquisition and stable colonization ofpathogens in a nascent ecosystem(Kumar et al. 2011) and that disease-associated microbiomes are patho-gen-enriched communities in thesehigh-risk individuals (Shchipkovaet al. 2010, Bizzarro et al. 2013).Therefore, the purpose of this inves-tigation was to extend the timeline ofthe previous investigation to examinethe effects of smoking on structuralresilience (as measured by commu-nity diversity, succession and res-ponse to disturbance) and functionalresilience (as measured by the hostresponse) following episodes of per-turbation and stress.

Methods

Subject selection

Approval for this study was obtainedfrom the Office of ResponsibleResearch Practices at The Ohio StateUniversity (Protocol number:2007H0259). Twenty-five currentsmokers (defined as those whosmoked more than 100 cigarettes intheir lifetime and are currently smok-ing) and 25 never smokers (definedas those who had never smoked asingle cigarette in their lifetime) withpre-existing gingivitis (mean baselinegingival index greater than 2.0 [Loeand Silness (Loe & Silness 1963)],attachment loss ≤1 mm and probedepths ≤4 mm at all sites) wererecruited and written informed con-sent obtained. Smoking status wasassessed by questionnaire and sali-vary cotinine levels. Exclusion crite-

ria included diabetes, medicalconditions that required use of pro-phylactic antibiotics, current orplanned pregnancy, HIV infection,long-term (greater than 3 months)use of medications known to causegingival changes, (e.g. immunosup-pressants, phenytoin, calcium chan-nel blockers, aspirin, NSAIDS,bisphosphonates or steroids), antibi-otic therapy or oral prophylactic pro-cedures within the last 3 months andless than 20 teeth in the dentition.Subjects were discontinued from thestudy if dictated by a contributorychange in medical or dental status,use of antibiotics and antibacterialmouthrinses, or non-compliance withwearing the stent while brushing.

Clinical study design and procedures

Baseline clinical data, marginal pla-que, subgingival plaque and gingivalcrevicular fluid (GCF) samples werecollected from those qualifying forthe study, following which profes-sional dental cleaning (oral prophy-laxis) was administered and oralhygiene instructions were provided(Fig. 1). Subjects returned 7–10 daysafter this initial visit, when gingivalhealth was confirmed (mean Gingi-val Index <0.25) and clinical data,plaque and gingival crevicular fluidsamples were collected; followingwhich professional oral prophylaxiswas administered once again. Acrylicstents were fabricated to cover three

teeth and 2 mm of marginal gingivain two contra-lateral quadrants (atotal of six teeth) and subjects wereinstructed to wear this while brush-ing for the duration of the study.Subjects returned 24 h later and clin-ical data, marginal plaque, subgingi-val plaque and GCF samples werecollected; following which oral pro-phylaxis was administered to zeroplaque. Subjects returned after 48 h;when 2-day old data and sampleswere collected, followed by oral pro-phylaxis. A similar protocol was fol-lowed to collect 4-day, 7-day, 14-dayand 21-day samples. Professionaloral prophylaxis was provided at thisvisit and subjects returned 7–10 dayslater for clinical data and samplecollection.

Examiner calibration

All subjects were examined by oneof two periodontists (CRM, PSK)calibrated to a Gold Standard exam-iner. The two examiners were foundto have an agreement coefficient (j-statistic) of 0.94 for PD and 0.96 forCAL, PI and GI with the GoldStandard examiner.

Data collection

Smoking status and tobacco expo-sure was assessed by questionnaireand salivary cotinine measurements.PD, BOP and CAL were recordedthroughout the mouth on six sites

Fig. 1. Clinical study design.

© 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

1038 Joshi et al.

per tooth using a PCP-UNC 15probe. Turesky modification of theQuigley Hein plaque index (Tureskyet al. 1970) and Loe and Silness gin-gival index (Silness & Loe 1964)were used to record plaque levelsand gingival health status at sixsites/tooth respectively.

Sample collection

Marginal plaque samples were col-lected and pooled from all test teethby scraping the cervical 2 mm ofeach mesial inter-proximal surface(buccal and lingual) with a 204S sca-ler and wiping onto paper points.Gingival crevicular fluid (GCF) sam-ples were collected from 12 sites byinserting Periopaper (Oraflow Inc.,Smithtown, NY, USA) into themesial sulcus for 30 s. GCF volumewas measured with a Periotron 8000(Harco Electronics Limited, Winni-peg, Canada) and samples immedi-ately frozen at �20°C. Subgingivalplaque samples were collected fromthe same sites by inserting one sterileendodontic paper point (DENTS-PLY-Caulk, Milford, DE, USA) intoeach mesial sulcus for 10 s, followedby scraping with a curette. Sampleswere placed in 1.5 ml microcentri-fuge tubes and frozen at �20°C untilfurther analysis.

DNA isolation

Subgingival bacteria were separatedfrom the paper points by adding200 ll of phosphate-buffered saline(PBS) to the tubes and vortexing.Marginal bacteria were separatedfrom the points by bead beating for60 s. DNA was isolated with a Qia-gen DNA MiniAmp kit (Qiagen,Valencia, CA, USA).

Amplification of 16S rDNA

Bacterial 16S rRNA genes wereamplified previously described (Ku-mar et al. 2005). The PCR productswere purified with the Qiaquick PCRpurification kit (Qiagen).

Cloning and sequencing

The 16S amplicons generated byPCR were cloned into E.coli using acommercially available kit (TOPOTA cloning kit, Invitrogen, SanDiego, CA, USA). Inserts were puri-

fied with a Millipore kit (Millipore,Billerica, MA, USA) and sequencedwith an ABI Prism cycle sequencingkit (BigDye Terminator CycleSequencing kit) using an ABI 3730instrument.

Sequence analysis

Partial sequences of 1200–1400 bpwere obtained from each amplicon.Sequences were clustered into spe-cies-level operational taxonomicunits (s-OTUs) at 97% sequencesimilarity and assigned a taxonomicidentity by alignment to locallyhosted version of the Greengenesdatabase (DeSantis et al. 2006) usingthe RDP Classifier. Unifrac andcommunity diversity metrics werecomputed as previously described(Lozupone et al. 2007). All analyseswere conducted using the QIIMEpipeline (Caporaso et al. 2010).

Cytokine assay

Periopaper strips were thawed on iceand GCF was eluted as previouslydescribed (Kumar et al. 2011).Cytokine analysis was done using acommercially available mutiplexedbead-based assay designed to quanti-tate multiple cytokines. A panel of27 cytokines was selected. Briefly, 27distinct sets of fluorescently dyedbeads (Bio-rad laboratories, Inc.,Hercules, CA, USA) were conju-gated with monoclonal antibodiesand incubated with 50 ll of GCF.25 ll of biotinylated detection anti-body and 50 ll of reporter weresequentially added. The level of eachcytokine was analysed by measuringthe fluorescence of each bead type aswell as the fluorescent signal fromthe reporter on a Bio-Plex 200 detec-tion system.

Data normalization and statistical analysis

A minimum of 100 sequences wasidentified from each sample. A vari-ance stabilizing transformation wasused to create normal distribution ofthe species and genus level data(Shchipkova et al. 2010). Diversityand equitability estimates were madeusing s-OUT data. Repeated Mea-sures ANOVA was used to comparethe levels of each species andimmune mediator over the five visits.A “cytokine score” was computed to

reduce inter-subject variability inimmune mediator levels (Bauer et al.2006). To achieve this, the absoluteconcentration of each analyte permicrolitre of GCF was normalizedacross all visits for each subject, suchthat the maximum value was 1.Thus, irrespective of inter-subjectvariations in immune mediator lev-els, the values ranged from zero toone for each subject. A “bacterialscore” was similarly computed bynormalizing the proportion of eachgenus across all visits for each sub-ject. Multivariate analysis was usedto examine the correlation betweenbacterial and cytokine scores duringinduction and resolution of gingivi-tis. Reported p-values correspond toSpearman’s correlation coefficients.All statistical analyses were carriedout with JMP (SAS Institute Inc.,Cary, NC, USA).

Results

Subject demographics and clinical

characteristics of natural and

experimental disease

The clinical and demographic char-acteristics are shown in Table 1.Clinically, there were no significantdifferences between current andnever smokers in mean gingivalinflammation or plaque accumula-tion (Fig. 2a,b). However, clusteranalysis of the individual subjectsrevealed that more smokers devel-oped gingival inflammation earlierthan never smokers (Fig. 2c). Basedon these clinical findings, the sam-ples were re-distributed into the fol-lowing groups for further analysis:Natural gingivitis (GI>2), Resolution1 (GI<1), Healthy phase (GI<1),Transition phase (2 > GI>1), Experi-mental gingivitis (GI>2) and Resolu-tion 2 (GI<1). There were nodifferences in gingival or plaque indi-ces between natural and experimen-tal gingivitis or between resolutionsof each disease state.

Bacterial shifts in marginal and

subgingival biofilms

A total of 66,068 near full-length,chimera-depleted sequences wereused in all analyses. These sequencesrepresented 164 and 293 specieslevel operational taxonomic units(s-OTUs) in the marginal biofilm;


Smoking and subgingival resilience 1039

and 235 and 311 s-OTUs in the sub-gingival biofilm of never and currentsmokers respectively. The transitionfrom disease to health was accompa-nied by a decrease in diversity inboth current and never smokers; thediversity increased during transitionto disease and decreased followingresolution of disease (Fig. 3a,b).Smokers demonstrated greater sub-gingival bacterial diversity than non-smokers during naturally occurringgingivitis and during resolution fromboth disease states (Fig. 3b). Smok-ers also demonstrated greater bacte-rial diversity than non-smokers inmarginal plaque during resolutionfollowing experimental gingivitis(Fig. 3a). These differences werestatistically significant (p < 0.05, rep-eated measures ANOVA).

Microbial-mucosal characteristics of

health and disease in never smokers

In never smokers, health (Resolution1, Healthy phase, Resolution 2) wascharacterized by low diversity bio-film communities (Fig 3a,b) thatwere dominated by species belongingto Veillonella, non-mutans Strepto-coccous, Neisseria, Actinomyces,Haemophilus and Abiotrophia inmarginal plaque and Actinomyces,Abiotrophia, Selenomonas, Neisseria,non-mutans Streptococcus and Veillo-nella in subgingival plaque (Fig. 3C).Species belonging to these generaconstituted 75% of the communitiesin these habitats. Transition to dis-

ease was accompanied by a slightbut insignificant increase in diversity,with decrease in levels of Veillonella,Streptococcus, Neisseria and Abiotro-phia and increase in the levels ofSelenonomas, Dialister, Campylobac-ter, Eubacterium, Lachnospira andParvimonas in both subgingival andmarginal biofilms. The establishmentof gingivitis was characterized byfurther decreases in levels of Veillo-nella, Streptococcus, Neisseria andAbiotrophia and increases in thespecies belonging to the generaCorynebacterium, Actinomyces, Hae-mophilus, Capnocytophaga, Dialister,Campylobacter, Eubacterium andSelenomonas in both habitats. In thesubgingival biofilm, gingivitis wasaccompanied by an increase in Trep-onema in addition to the other spe-cies. The microbial communities innaturally occurring and experimentalgingivitis were remarkably similar aswere the communities following reso-lution from these two disease states(p > 0.05, Repeated Measures ANO-

VA). Significant correlations wereobserved in the composition andabundance of species in the marginaland subgingival communities duringeach phase of colonization (Fig. 4a).Host immune mediators exhibited adistinct clustering in response to thiscolonization (p < 0.0001, factoranalysis of Principal ComponentsAnalysis, Fig. 5). Two distinct clus-ters were observed-cluster 1 wascomposed of IL-5, IL-6, IL-7, IL-8,IL-10, IL-13, IL-15, IL-17, IL-1ra,

FGF, PDGF, MIP-1a and Eotaxin;whereas cluster 2 was formed byRANTES, GCSF, GM-CSF, VEGF,TNF-a, IFN-c, IL-1b, IL-2, IL-4,IL-9 and IL-12. Health was associ-ated with low levels of Cluster 1cytokines and high levels of Cluster2 cytokines. Transition to diseasewas accompanied by high levels ofboth Cluster 1 and Cluster 2 cyto-kines as well as low levels of MIP-1b, MCP-1 and IP-10. Establishedgingivitis was associated with highlevels of Cluster 2 cytokines. Strongpositive correlations were observedbetween Abiotrophia, Veillonella,Streptococcus, Gemella and Neisseriaand the cytokines IP-10, IL-10, IL-6,as well as between Selenomonas, Par-vimonas, Megasphaera, Lachnospira,Eubacterium, Dialister and Campylo-bacter and RANTES, IL-4, VEGF,G-CSF, IFN- c, IL-1B, IL-2, GM-CSF, TNF- a, IL-9 (Fig. 6a). Strongnegative correlations were observedbetween these cytokines and Abiotro-phia, Veillonella, Streptococcus, Gem-ella and Neisseria.

Microbial-mucosal characteristics of

health and disease in current smokers

The three health-compatible mar-ginal communities were similar incomposition and diversity; however,in contrast to the never smokers,Resolution 1 and Resolution 2 insmokers exhibited greater subgingi-val diversity than the Healthy phase(Fig. 3b). In addition, these twostates were characterized by signifi-cantly greater subgingival diversitythan in non-smokers. The health-compatible subgingival communitiesof smokers were dominated by speciesbelonging to Parvimonas, Gemella,Treponema, Eubacterium, Pseudomo-nas, Dialister, Selenomonas, Veillonel-la and Streptococcus, whereas themarginal biofilm was dominated byStreptococcus, Veillonella, Selenomon-as, Pseudomonas and Haemophilus(Fig. 3c). Transition to disease wasaccompanied by a slight but insignifi-cant increase in diversity, withdecreases in levels of Veillonella andStreptococcus, and increase in the lev-els of Selenonomas and Dialister inmarginal biofilms and Treponema,Prevotella, Campylobacter, Eubacteri-um, Dialister and Selenomonas in sub-gingival biofilms. The establishmentof gingivitis was characterized by fur-

Table 1. Demographic features of study population

Never smokers Current smokers p-value

Mean age (years) 26.5 � 5.5 25.5 � 3.5 p = 0.20Gender – female (%) 77.6 56.4 p = 0.18Race (%)

Caucasian 80.0 85.0 p = 0.03African-American 6.0 8.0 p = 0.06Asian 6.0 3.0 p = 0.035Other 8.0 4.0 p = 0.013

Education (average years/individual) 13 � 3 14 � 2 p = 0.25Dental prophylaxis (number of annualvisits/individual)

0.96 0.97 p = 0.23

Brushing frequency(times per day/individual)

1.57 1.28 p = 0.23

Frequency of flossing (times per day) 1.0 1.0 p = 0.23Mean number of teeth 28.0 28.0 p = 1Mean probing pocket depth (mm) 3.7 � 0.8 4.3 � 1.2 p = 0.25Mean clinical attachment loss (mm) 1.1 � 0.3 0.8 � 0.7 p = 0.18Tobacco exposurePack years 0 5.6 � 2.4 p = 0.0015Salivary cotinine (ng/ml) 0.0 23.3 � 3.7 p = 0.0001


1040 Joshi et al.

(a)

(c)

(b)

Fig. 2. Gingival and Plaque indices of 25 current and 25 never smokers. (a) Shows the Gingival Index and (b) the Plaque Index.The inset demonstrates the same data for the two episodes of disease and resolution only. Both groups demonstrated an increase inplaque and gingival indices over the observation period. (c) Shows the clustering of the subjects by the extent of gingival inflamma-tion (GI<1 – Red, 1 < GI<2 – Green and GI>2 – Blue). Smokers demonstrated an earlier onset of clinically visible inflammation(1 < GI<2 and GI>2) than non-smokers.



ther decreases in levels of Veillonella,Pseudomonas and Streptococcus, andincreases in the species belonging tothe genera Parvimonas, Megasphaera,Treponema, Capnocytophaga, Dialis-ter, Campylobacter and Selenomonasin both habitats. The microbial com-munities in naturally occurring andexperimental gingivitis were signifi-cantly different, with greater abun-dances of Eubacterium, Treponema,Parvimonas and Campylobacter andlower levels of Veillonella and Strep-tococcus. The composition of thecommunities also differed signifi-cantly during resolutions from thesetwo disease states (p < 0.05, repeatedmeasures ANOVA). Significant correla-tions were observed in the composi-tion and abundance of species in thesupragingival and subgingival com-munities during all phases of coloni-zation except during Resolution 2

(Fig. 4b). Host immune mediatorsexhibited a distinct clustering inresponse to this colonization(p < 0.0001, factor analysis of Princi-pal Components Analysis, Fig. 5).Two distinct clusters were observed-cluster 1 was composed of IL-17, IP-10, FGF, MIP-1a, Eotaxin, MCP-1and MIP-1b, whereas cluster 2 wasformed by IL-1b, IL-2, IL-4, IL-5,IL-6, IL-7, IL-8, IL-12, IL-13, IL-1ra,IFN-c, TNF-a, PDGF and VEGF.Health was associated with low levelsof Cluster 1 cytokines and high levelsof Cluster 2 cytokines. Establishedgingivitis was associated with highlevels of Cluster 2 cytokines. Theimmune profiles of natural and exper-imental gingivitis were significantlydifferent (p < 0.05, repeated measuresANOVA). Both Cluster 1 and Cluster 2cytokines remained at high levels fol-lowing resolution of experimental

gingivitis. Strong positive correlationswere observed between Streptococcus,Selenomonas, Rothia, Pseudomonasand Propionibacterium and the cyto-kines IL-12, IL-17, RANTES, MIP-1A, IL-13, Eotaxin, IL-1b, IL-2, IL-4,IL-6, IL-7, IL-9; as well as betweenVeillonella, Gemella, Lautropia andthe cytokines IL-12, IL-17, RANTES(Fig. 6b).

Discussion

Experimental gingivitis as a model system

of microbial-mucosal interactions

The human body plays host to bio-films in several mucosal niches; andthese communities have an integralrole in maintaining health and caus-ing disease in these environments(Rakoff-Nahoum et al. 2004). It isoften difficult to examine the shifts

(a)

(c)

(b)

Fig. 3. Shannon Diversity Index and genera in current and never smokers. Diversity of marginal biofilm is shown in (a) and subgin-gival biofilm in (b). Smokers demonstrated an increased diversity between the two episodes of gingivitis as well as the two resolu-tion states ($p < 0.05, $$p < 0.01 repeated measures ANOVA), and showed a greater diversity than never smokers (**p < 0.01).Distribution of sequences by genus in marginal and subgingival biofilms of current (red) and never smokers (green) is shown in Fig-ure 3c. Circles are sized by relative abundance of each genus. Genera accounting for >0.5% of total sequences are shown. Signifi-cant differences were observed in the levels of several genera during the different periods of observation (repeated measures ANOVA

on transformed variable, *p < 0.05, **p < 0.01, ***p < 0.001).


1042 Joshi et al.

that occur in these ecosystems, eitherbecause they are not easily accessibleor because transitions from health todisease cannot be easily monitored.The subgingival microenvironmentprovides an easily accessible nichewith a defined microbial populationand therefore serves as an idealmodel to examine mucosal-microbialinteractions. Furthermore, gingivitisis a reversible, bacterially inducedinflammation that occurs both natu-rally and can be induced andresolved in a controlled manner (Loeet al. 1965, Theilade et al. 1966). In

addition, numerous studies indicatethat there are no differences in eitherthe clinical characteristics or bacte-rial profiles between natural andexperimental gingivitis (Loesche &Syed 1978, Syed & Loesche 1978,Moore et al. 1987). Hence, a longitu-dinal examination of the clinical,microbiological and immunologicalchanges that occur during onset andresolution of naturally occurring andexperimentally induced gingivitisallowed us to examine communityresilience following two subsequentepisodes of bacterially induced

inflammation. To the best of ourknowledge, this is the first evidenceon the effect of smoking onaltering community resistance andresilience in any ecological niche inthe human body.

Smoking and parameters of ecosystem

health

Three important parameters thataffect the health of an ecosystem arebiodiversity, disturbance and succes-sion. In non-smokers, marginal andsubgingival biofilms of Resolution 1and Resolution 2 closely resembledthat of Healthy phase; with bacterialprofiles similar to health-compatiblecommunities described in previousstudies (Tanner et al. 1998, Kumaret al. 2006), indicating that thesecommunities reach steady-statewithin 48 h and continue to existstably in this health-compatiblestate. Several examples of syntrophyand mutualism were observed inthese communities (Fig. 4a,b). It haspreviously been shown that Veillo-nella and Streptococcus share anutritional syntrophy, in that Veillo-nella are reliant on the lactate that isproduced by the Streptococci as afood source (Kuramitsu et al. 2007).In this study, Streptococci andVeillonella dominated the marginaland subgingival ecosystems inhealth. An increase in plaque bio-mass (Fig. 2) was accompanied bydecline in levels of Streptococci by50% on day 4 (transition phase) andon day 7, the levels of Veillonelladeclined to comparable levels(Fig. 3c). Another interestinginstance of mutualism was apparentbetween Streptococcus, Actinomycesand Selenomonas. Oral Streptococciappear to be important for Actino-myces colonization, and Selenomonasare poor colonizers unless Actinomy-ces naeslundii is present (Kolenbran-der et al. 1989, Palmer et al. 2001).In the present investigation, surpris-ingly high levels of Selenomonaswere found within 24 h in certainsubjects, and correlated stronglywith the levels of Streptococcus andActinomyces. These subjects alsodeveloped gingivitis earlier (Fig. 2c).Selenomonas are acknowledged asperiodontal pathogens and areimportant in gingivitis (Tanner et al.1989). Taken together, the data sug-gest that high levels of Streptococci

(a)

(b)

Fig. 4. Correlations between species abundances in marginal and subgingival biofilmsduring the different observation periods. Non-smokers are shown in (a) and smokersin (b). Smokers exhibited significantly lower correlation between species in marginaland subgingival plaque during Resolution 2 (*p < 0.05, repeated measures ANOVA).



and Actinomyces may predispose toearly colonization with pathogensand may increase susceptibility todisease. The immune response to thiscommunity was predominantly anti-

inflammatory, with high levels of IP-10, IL-1ra, IL-10 and IL-15. IP-10 apowerful inhibitor of VEGF and isan anti-angiogenic factor (Bodnaret al. 2006), and IL-1ra, IL-10 and

IL-15 are anti-inflammatory factorspreviously found in high levels inhealthy gingiva (Johnson & Serio2007). High levels of Streptococcus,Neisseria, Veillonella, Gemella and

Fig. 5. Principal component analysis (PCA) of cytokine scores during each observation period. Cytokines are coloured by lowestscore (green) to highest score (red), based on their concentrations in each subject during each observation period. Host immunemediators exhibited a distinct clustering in response to colonization in both current and never smokers (p < 0.0001, factor analysisof Principal Components Analysis).


1044 Joshi et al.

Abiotrophia correlated strongly withthese anti-inflammatory mediators.Thus, the central characteristics of ahealth-compatible community innon-smokers appear to be cooper-ativity, low diversity and low immu-nogenicity. The switch from healthto disease is triggered by a changefrom an aerobic to anaerobic envi-ronment (due to an increase in pla-que mass and biofilm thickness),which then triggers a nutritionally

driven cascade of events within thebiofilm and then extended to thehost.

In contrast, the development ofthe marginal and subgingival ecosys-tems in smokers was characterized byrapid microbial succession with highlevels of periodontal and systemicpathogens, and was accompanied byhigh levels of pro-inflammatory cyto-kines (Figs 3 and 5). We have previ-ously reported that smokers acquire

large numbers of periodontalpathogens belonging to the generaFusobacterium, Selenomonas, Dialis-ter, Treponema and Parvimonas andsystemic pathogens including Pseudo-monas and Haemophilus during earlybiofilm formation and that there is alarge amount of fluctuation in thesecommunities (Kumar et al. 2011).The present investigation demon-strates that this early pathogen acqui-sition leads to earlier pathogenenrichment in the community. It iswell documented that smokingdecreases oxygen tension (Haniokaet al. 2000), increases subgingivaltemperature (Dinsdale et al. 1997),reduces the Eh (Kenney et al. 1975)and increases the amount of free iron(Weinberg 1999). It is also wellknown that these factors promoteparasitic symbiosis (Wilson et al.1992), preferential colonization withperiodontal pathogens (Dinsdaleet al. 1997) as well as colonization byspecies with poor iron acquisitionabilities (Weinberg 2000). Takentogether with previous research, thisstudy demonstrates that that smokingcreates an environment that does notsupport niche saturation by earlycolonizers, leading to an highlydiverse, pathogen-enriched, unstablecommunity that is susceptible todisruption and therefore highlypro-inflammatory. This is also sup-ported by the clinical data (Fig. 2C)that more smokers developed gingivalinflammation earlier than non-smok-ers.

Smoking and community resilience

A fundamental requirement formaintaining a healthy oral ecosystemis an understanding of the commu-nity’s response to stress and pertur-bation. This study investigated theeffects of smoking on the structuraland functional resilience of marginaland subgingival biofilms followingtwo episodes of disease, using diver-sity, compositional succession andpro-inflammatory potential as mark-ers of resilience. In non-smokers, aremarkable similarity was observedbetween Resolution 1 and Resolu-tion 2 clinically, microbiologicallyand immunologically (Figs 3, 4 and5). Both communities were 90% sim-ilar in composition, exhibited lowdiversity and elicited a low pro-inflammatory response, suggesting

(a)

(b)

Fig. 6. Hierarchical cluster analysis of correlations between bacterial genera andimmune mediators. Non-smokers are shown in (a) and smokers in (b). Distinct pat-terns of positive and negative correlations were observed in host–bacterial interactionsof non-smokers, but not smokers.



that these communities are highlyresilient following perturbation.However, in smokers, the commu-nity in Resolution 2 demonstratedhigher levels of the pathogens Trepo-nema, Selenomonas, Parvimonas,Eubacterium, Pseudomonas and Lac-tobacillus when compared with Reso-lution 1. The community was alsomore diverse compared to Resolu-tion 1. In addition, the lowest corre-lations were observed between themarginal and subgingival OTUs dur-ing Resolution 2. It has previouslybeen demonstrated that subgingivalcolonization is significantly affectedby the composition of marginal pla-que (Gomes et al. 2008, Meulmanet al. 2012). Thus, the present inves-tigation indicates that with each epi-sode of disease, the interactionsbetween these two geographicallyconnected habitats decreases. Thepro-inflammatory response to thiscommunity was also significantlyhigher as compared to Resolution 1.The clinical implications of thisfinding are enormous. Our data sug-gests that smoking increases suscep-tibility to disease by lowering theability of the ecosystem to “reset”itself, thereby lowering its resilience.Given that sustained and repeatedepisodes of gingivitis predispose toperiodontitis (Lang et al. 2009), thisreduction in resilience could a mech-anism by which smoking increasesthe risk for periodontitis and war-rants further investigation.

In summary, smokers demon-strated an early pathogenic coloni-zation that leads to sustainedpathogen enrichment, eliciting aflorid immune response. Oral micro-bial communities in smokers alsodemonstrate a decreased resilienceto repeated episodes of disease; withgreater abundance of pathogenicspecies, poor compositional correla-tion between the marginal and thesubgingival ecosystems and signifi-cantly greater pro-inflammatoryresponses. Our data suggest thatrepeated episodes of such insultsmight increase susceptibility tofuture disease in smokers. Furtherstudies examining the patterns ofgene expressions within these com-munities are necessary to under-stand the effect of smoking onaetiopathogenesis of periodontal dis-eases.

References

Ash, M. M. Jr., Gitlin, B. N., Smith, W. A. (1964)Correlation between plaque and gingivitis.Journal of Periodontology 35, 424–429.

Bauer, J. W., Baechler, E. C., Petri, M., Batliwal-la, F. M., Crawford, D., Ortmann, W. A.,Espe, K. J., Li, W., Patel, D. D., Gregersen, P.K. & Behrens, T. W. (2006) Elevated serumlevels of interferon-regulated chemokines arebiomarkers for active human systemic lupuserythematosus. PLoS Medicine 3, e491.

Bergstrom, J. (1989) Cigarette smoking as riskfactor in chronic periodontal disease. Commu-nity Dentistry and Oral Epidemiology 17, 245–247.

Bergstrom, J., Eliasson, S. & Dock, J. (2000)Exposure to tobacco smoking and periodontalhealth. Journal of Clinical Periodontology 27,61–68.

Bizzarro, S., Loos, B. G., Laine, M. L., Crielaard,W. & Zaura, E. (2013) Subgingival microbiomein smokers and non-smokers in periodontitis:an exploratory study using traditional targetedtechniques and a next-generation sequencing.Journal of Clinical Periodontology 40, 483–492.

Bodnar, R. J., Yates, C. C. & Wells, A. (2006)IP-10 blocks vascular endothelial growth fac-tor-induced endothelial cell motility and tubeformation via inhibition of calpain. CirculationResearch 98, 617–625.

Caporaso, J. G., Kuczynski, J., Stombaugh, J.,Bittinger, K., Bushman, F. D., Costello, E. K.,Fierer, N., Pena, A. G., Goodrich, J. K., Gor-don, J. I., Huttley, G. A., Kelley, S. T.,Knights, D., Koenig, J. E., Ley, R. E., Lozu-pone, C. A., McDonald, D., Muegge, B. D.,Pirrung, M., Reeder, J., Sevinsky, J. R., Turnb-augh, P. J., Walters, W. A., Widmann, J.,Yatsunenko, T., Zaneveld, J. & Knight, R.(2010) QIIME allows analysis of high-through-put community sequencing data. Nature Meth-ods 7, 335–336.

DeSantis, T. Z., Hugenholtz, P., Larsen, N.,Rojas, M., Brodie, E. L., Keller, K., Huber, T.,Dalevi, D., Hu, P. & Andersen, G. L. (2006)Greengenes, a chimera-checked 16S rRNA genedatabase and workbench compatible withARB. Applied and Environment Microbiology72, 5069–5072.

Dinsdale, C. R., Rawlinson, A. & Walsh, T. F.(1997) Subgingival temperature in smokers andnon-smokers with periodontal disease. Journalof Clinical Periodontology 24, 761–766.

Gomes, S. C., Nonnenmacher, C., Susin, C., Op-permann, R. V., Mutters, R. & Marcantonio,R. A. (2008) The effect of a supragingival pla-que-control regimen on the subgingival micro-biota in smokers and never-smokers: evaluationby real-time polymerase chain reaction. Journalof Periodontology 79, 2297–2304.

Hajishengallis, G. & Lamont, R. J. (2012) Beyondthe red complex and into more complexity: thepolymicrobial synergy and dysbiosis (PSD)model of periodontal disease etiology. Molecu-lar Oral Microbiology 27, 409–419.

Hanioka, T., Tanaka, M., Takaya, K., Matsu-mori, Y. & Shizukuishi, S. (2000) Pocket oxy-gen tension in smokers and non-smokers withperiodontal disease. Journal of Periodontology71, 550–554.

Johnson, R. B. & Serio, F. G. (2007) The contri-bution of interleukin-13 and -15 to the cytokinenetwork within normal and diseased gingiva.Journal of Periodontology 78, 691–695.

Kenney, E. B., Saxe, S. R. & Bowles, R. D.(1975) The effect of cigarette smoking on

anaerobiosis in the oral cavity. Journal of Peri-odontology 46, 82–85.

Kolenbrander, P. E., Andersen, R. N. & Moore,L. V. (1989) Coaggregation of Fusobacteriumnucleatum, Selenomonas flueggei, Selenomonasinfelix, Selenomonas noxia, and Selenomonassputigena with strains from 11 genera of oralbacteria. Infection and Immunity 57, 3194–3203.

Kumar, P. S., Griffen, A. L., Moeschberger, M.L. & Leys, E. J. (2005) Identification of candi-date periodontal pathogens and beneficial spe-cies by quantitative 16S clonal analysis. Journalof Clinical Microbiology 43, 3944–3955.

Kumar, P. S., Leys, E. J., Bryk, J. M., Martinez,F. J., Moeschberger, M. L. & Griffen, A. L.(2006) Changes in periodontal health status areassociated with bacterial community shifts asassessed by quantitative 16S cloning andsequencing. Journal of Clinical Microbiology44, 3665–3673.

Kumar, P. S., Matthews, C. R., Joshi, V., de Jag-er, M. & Aspiras, M. (2011) Tobacco smokingaffects bacterial acquisition and colonization inoral biofilms. Infection and Immunity 79, 4730–4738.

Kuramitsu, H. K., He, X., Lux, R., Anderson,M. H. & Shi, W. (2007) Interspecies interac-tions within oral microbial communities.Microbiology and Molecular Biology Reviews71, 653–670.

Lang, N. P., Schatzle, M. A. & Loe, H. (2009)Gingivitis as a risk factor in periodontal dis-ease. Journal of Clinical Periodontology 36(Suppl. 10), 3–8.

Listgarten, M. A. (1976) Structure of the micro-bial flora associated with periodontal healthand disease in man. A light and electron micro-scopic study. Journal of Periodontology 47,1–18.

Listgarten, M. A., Schifter, C. C. & Laster, L.(1985) 3-year longitudinal study of the peri-odontal status of an adult population with gin-givitis. Journal of Clinical Periodontology 12,225–238.

Loe, H. & Silness, J. (1963) Periodontal disease inpregnancy. I. Prevalence and severity. Acta Od-ontologica Scandinavica 21, 533–551.

Loe, H., Theilade, E. & Jensen, S. B. (1965)Experimental gingivitis in man. Journal of Peri-odontology 36, 177–187.

Loesche, W. J. & Syed, S. A. (1978) Bacteriologyof human experimental gingivitis: effect ofplaque and gingivitis score. Infection and Immu-nity 21, 830–839.

Lozupone, C., Lladser, M. E., Knights, D.,Stombaugh, J. & Knight, R. (2007) UniFrac:An effective distance metric for microbial com-munity comparison. ISME Journal 5, 169–172.

Meulman, T., Casarin, R. C., Peruzzo, D. C.,Giorgetti, A. P., Barbagallo, A., Casati, M. Z.,Sallum, E. A., Goncalves, R. B. & Nociti, F.H. Jr (2012) Impact of supragingival therapyon subgingival microbial profile in smokers ver-sus non-smokers with severe chronic periodon-titis. Journal of Oral Microbiology 4,doi:10.3402/jom.v4i0.8640.

Moore, L. V., Moore, W. E., Cato, E. P., Smibert,R. M., Burmeister, J. A., Best, A. M. & Ranney,R. R. (1987) Bacteriology of human gingivitis.Journal of Dental Research 66, 989–995.

Page, R. C. (1986) Gingivitis. Journal of ClinicalPeriodontology 13, 345–359.

Page, R. C. & Schroeder, H. E. (1976) Pathogene-sis of inflammatory periodontal disease. A sum-mary of current work. LaboratoryInvestigation; A Journal of Technical Methodsand Pathology 34, 235–249.


1046 Joshi et al.

Palmer, R. J. Jr, Kazmerzak, K., Hansen, M. C.& Kolenbrander, P. E. (2001) Mutualism ver-sus independence: strategies of mixed-speciesoral biofilms in vitro using saliva as the solenutrient source. Infection and Immunity 69,5794–5804.

Rakoff-Nahoum, S., Paglino, J., Eslami-Varza-neh, F., Edberg, S. & Medzhitov, R. (2004)Recognition of commensal microflora by toll-like receptors is required for intestinal homeo-stasis. Cell 118, 229–241.

Shchipkova, A. Y., Nagaraja, H. N. & Kumar, P.S. (2010) Subgingival microbial profiles ofsmokers with periodontitis. Journal of DentalResearch 89, 1247–1253.

Silness, J. & Loe, H. (1964) Periodontal disease inpregnancy. II. Correlation between oral hygieneand periodontal condtion. Acta OdontologicaScandinavica 22, 121–135.

Sreenivasan, P. K., DeVizio, W., Prasad, K. V.,Patil, S., Chhabra, K. G., Rajesh, G., Javali, S.B. & Kulkarni, R. D. (2010) Regional differ-ences within the dentition for plaque, gingivitis,and anaerobic bacteria. Journal of Clinical Den-tistry 21, 13–19.

Syed, S. A. & Loesche, W. J. (1978) Bacteriologyof human experimental gingivitis: effect of pla-que age. Infection and Immunity 21, 821–829.

Tanner, A., Bouldin, H. D. & Maiden, M. F.(1989) Newly delineated periodontal pathogenswith special reference to selenomonas species.Infection 17, 182–187.

Tanner, A., Maiden, M. F., Macuch, P. J., Mur-ray, L. L. & Kent, R. L. Jr (1998) Microbiotaof health, gingivitis, and initial periodontitis.Journal of Clinical Periodontology 25, 85–98.

Theilade, E., Wright, W. H., Jensen, S. B. & Loe,H. (1966) Experimental gingivitis in man. II. Alongitudinal clinical and bacteriological investi-gation. Journal of Periodontal Research 1, 1–13.

Tomar, S. L. & Asma, S. (2000) Smoking-attrib-utable periodontitis in the United States: find-ings from NHANES III. National Health andNutrition Examination Survey. Journal of Peri-odontology 71, 743–751.

Turesky, S., Gilmore, N. D. & Glickman, I.(1970) Reduced plaque formation by the chlo-romethyl analogue of victamine C. Journal ofPeriodontology 41, 41–43.

Wardwell, L. H., Huttenhower, C. & Garrett, W.S. (2011) Current concepts of the intestinal

microbiota and the pathogenesis of infection.Current Infectious Disease Reports 13, 28–34.

Weinberg, E. D. (1999) The development ofawareness of the carcinogenic hazard of inhalediron. Oncology Research 11, 109–113.

Weinberg, E. D. (2000) Microbial pathogens withimpaired ability to acquire host iron. BioMetals13, 85–89.

Wilson, M., Gibson, M., Strahan, D. & Harvey,W. (1992) A preliminary evaluation of the useof a redox agent in the treatment of chronicperiodontitis. Journal of Periodontal Research27, 522–527.

Address:Purnima Kumar4111 Postle Hall305, W 12th AvenueColumbus, OH 43210USAE-mail: [email protected]

Clinical Relevance

Scientific rationale for the study: Itis known that smoking increasesthe risk for periodontitis; however,the mechanisms are not fully eluci-dated.Principal findings: Our data indi-cate that smokers have a lower

resilience in their subgingival micro-bial community, therefore, this com-munity does not fully return to ahealth-compatible state followingresolution form gingivitis. This lossof microbial resilience leads to agreater pro-inflammatory hostresponse, which increases with each

episode of gingival inflammation.This appears to be a mechanism bywhich repeated episodes of gingivi-tis may predispose a smoker todestructive periodontal disease.Practical implications: Preventionof gingivitis is extremely critical ina smoker.



smoking decreases structural and functional resilience in ...€¦ · college and research...

Documents