Enterococcus faecalis Demonstrates Pathogenicity throughIncreased Attachment in an Ex Vivo Polymicrobial PulpalInfection

Wayne Nishio Ayre,a Genevieve Melling,a Camille Cuveillier,a Madhan Natarajan,a Jessica L. Roberts,a* Lucy L. Marsh,a

Christopher D. Lynch,b Jean-Yves Maillard,c Stephen P. Denyer,d Alastair J. Sloana

aSchool of Dentistry, Cardiff University, Cardiff, United KingdombUniversity Dental School and Hospital, University College Cork, Cork, IrelandcSchool of Pharmacy and Pharmaceutical Sciences, Cardiff University, Cardiff, United KingdomdSchool of Pharmacy and Biomolecular Sciences, University of Brighton, Brighton, United Kingdom

ABSTRACT This study investigated the host response to a polymicrobial pulpal in-fection consisting of Streptococcus anginosus and Enterococcus faecalis, bacteria com-monly implicated in dental abscesses and endodontic failure, using a validated exvivo rat tooth model. Tooth slices were inoculated with planktonic cultures of S. an-ginosus or E. faecalis alone or in coculture at S. anginosus/E. faecalis ratios of 50:50and 90:10. Attachment was semiquantified by measuring the area covered by fluo-rescently labeled bacteria. Host response was established by viable histologicalcell counts, and inflammatory response was measured using reverse transcription-quantitative PCR (RT-qPCR) and immunohistochemistry. A significant reduction incell viability was observed for single and polymicrobial infections, with no significantdifferences between infection types (�2,000 cells/mm2 for infected pulps comparedto �4,000 cells/mm2 for uninfected pulps). E. faecalis demonstrated significantlyhigher levels of attachment (6.5%) than S. anginosus alone (2.3%) and mixed-speciesinfections (3.4% for 50:50 and 2.3% for 90:10), with a remarkable affinity for the pul-pal vasculature. Infections with E. faecalis demonstrated the greatest increase in tu-mor necrosis factor alpha (TNF-�) (47.1-fold for E. faecalis, 14.6-fold for S. anginosus,60.1-fold for 50:50, and 25.0-fold for 90:10) and interleukin 1� (IL-1�) expression(54.8-fold for E. faecalis, 8.8-fold for S. anginosus, 54.5-fold for 50:50, and 39.9-fold for90:10) compared to uninfected samples. Immunohistochemistry confirmed this, withthe majority of inflammation localized to the pulpal vasculature and odontoblast re-gions. Interestingly, E. faecalis supernatant and heat-killed E. faecalis treatments wereunable to induce the same inflammatory response, suggesting E. faecalis pathoge-nicity in pulpitis is linked to its greater ability to attach to the pulpal vasculature.

KEYWORDS coculture, Enterococcus faecalis, ex vivo, host, infection, model,polymicrobial, pulpal infection, pulpitis, Streptococcus anginosus

The dental pulp is a complex environment composed of soft connective tissue,nerves, blood vessels, and a variety of cells, such as dental pulp stem cells,

fibroblasts, and odontoblasts (1). When the pulp becomes inflamed in response tobacterial infection or other stimuli, this is known as pulpitis. Early stages are considered“reversible,” and treatment involves removal of the stimulus, such as carious lesions, inorder to maintain pulp vitality. If untreated, however, the microbial invasion mayprogress into the deeper dentine and subsequently the pulpal chamber, resulting insevere tissue degradation and necrosis. This condition, known as “irreversible pulpitis,”requires a challenging and difficult endodontic or root canal treatment, which involvesthe removal of the pulp and obturation with an inert material. The success rate of root

Received 17 December 2017 Returned formodification 19 January 2018 Accepted 21February 2018

Accepted manuscript posted online 26February 2018

Citation Nishio Ayre W, Melling G, Cuveillier C,Natarajan M, Roberts JL, Marsh LL, Lynch CD,Maillard J-Y, Denyer SP, Sloan AJ. 2018.Enterococcus faecalis demonstratespathogenicity through increased attachmentin an ex vivo polymicrobial pulpal infection.Infect Immun 86:e00871-17. https://doi.org/10.1128/IAI.00871-17.

Editor Nancy E. Freitag, University of Illinois atChicago

Copyright © 2018 American Society forMicrobiology. All Rights Reserved.

Address correspondence to Wayne NishioAyre, ayrewn@cardiff.ac.uk.

* Present address: Jessica L. Roberts, NorthWales Centre for Primary Care Research, BangorUniversity, Bangor, United Kingdom.

HOST RESPONSE AND INFLAMMATION

crossm

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 1Infection and Immunity

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

canal treatments is highly variable, ranging from 31% to 96% depending on clinicalconsiderations (2), and studies across a range of countries have shown that a highpercentage (up to 67.9%) of patients who have undergone this treatment subsequentlydevelop apical periodontitis (3, 4). An alternative endodontic treatment is vital pulpo-tomy, which involves removal of the coronal pulp, leaving the radicular pulp vital andfree of any pathological alterations (5). Although this procedure is thought to requireshorter appointment times and can be accomplished in one visit, the efficacy of thetechnique is debated, with success rates of clinical studies ranging from 70% to 96% (6).Accurate models to better understand the process of pulpal infection and to test theefficacy of novel therapeutics will aid in the development of more effective vital pulptreatments. In vitro monolayer cell culture models lack the complexity of the pulpalmatrix, while in vivo studies suffer from systemic factors, high costs, and ethicalconsiderations. To overcome these limitations, Roberts et al. (7) developed an ex vivococulture system to model pulpal infections on rat tooth slices. This study focusedpredominantly on the Streptococcus anginosus group (SAG), consisting of S. anginosus,Streptococcus constellatus and Streptococcus intermedius, Gram-positive cocci that arepart of the body’s commensal flora. The members of this group are known to beprimary colonizers of the oral cavity due to their ability to attach to the salivary pellicleand other oral bacteria (8). They are considered opportunistic pathogens and havebeen reported to form dental abscesses (9). The study by Roberts et al. demonstrateda significant reduction in viable pulp cells and an increase in cytokine expression andbacterial attachment over 24 h as a result of S. anginosus infections (7).

Although Roberts et al. demonstrated invasion of the dental pulp by S. anginosusgroup species, the number of microbial species encountered in the oral cavity is farmore diverse, with studies identifying between 100 and 300 different species fromdifferent regions of the oral cavity in healthy individuals (10). It is therefore unsurprisingthat complex mixed-species microbiomes are often detected in cases of pulpitis (11). Aslesions progress into the tooth, a shift in microbial species due to environmental andnutritional changes has been well documented (12). Of particular interest is the speciesEnterococcus faecalis, a Gram-positive facultative anaerobic coccus that is also part ofthe normal human commensal flora (13). E. faecalis has been shown to be pathogenic,particularly in endodontic failure (14), with prevalence in such infections ranging from24% up to 77% (15). Although highly implicated in persistent endodontic failure,molecular studies have recently revealed that the species is frequently present innecrotic pulps, highlighting its potential role in late-stage pulpitis (16, 17).

This study aimed to use a validated ex vivo coculture model to quantify and betterunderstand the host tissue response to mixed-species pulpal infections caused by S.anginosus and E. faecalis. Understanding the mechanism of complex pulpal infectionsand the host inflammatory response may elucidate potential targets for more effectivevital pulp therapies.

RESULTSMixed-species culture does not significantly influence S. anginosus and E.

faecalis growth rates. The growth characteristics of a simple mixed-species planktonicbroth culture were investigated to ensure potential competitive growth between S.anginosus and E. faecalis would not influence the ex vivo experiments investigating hosttissue response.

Clinical isolates of the species S. anginosus and E. faecalis were selected from theculture collection of the Oral Microbiology Unit, School of Dentistry at Cardiff Univer-sity. Species identities were confirmed by standard microbial identification tests and16S rRNA sequencing, as described in Materials and Methods and the supplementalmaterial (see Fig. S1 and S2 in the supplemental material).

Figure 1 shows the planktonic growth curves for S. anginosus and E. faecalis aloneand in combination at ratios of 50:50 and 90:10 over 24 h in brain heart infusion (BHI)broth. E. faecalis reached mid-log phase earlier than S. anginosus (8 h for E. faecaliscompared to 10 h for S. anginosus). When cultured at a ratio of 50:50, however, S.

Nishio Ayre et al. Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 2

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

anginosus reached mid-log phase at a time similar to that for E. faecalis (10 h). When thebacteria were cultured at an S. anginosus/E. faecalis ratio of 90:10, S. anginosus reachedmid-log phase at approximately 8 h and E. faecalis at approximately 12 h. Growth ratecalculations during the log phase demonstrated no significant differences between E.faecalis and S. anginosus under all culture conditions (P � 0.05) (Table 1).

E. faecalis demonstrates greater levels of attachment to dental pulp than S.anginosus at 24 h, with particular affinity for the pulpal vasculature. To assessdifferences in bacterial attachment to the dental pulp, the ex vivo rat tooth model wasinfected with planktonic cultures of S. anginosus and E. faecalis individually or asmixed-species infections. Gram staining and fluorescent labeling of bacteria wereundertaken to localize and semiquantify bacterial attachment.

High levels of bacterial attachment to the pulp were detected for tooth slicesincubated with E. faecalis (Fig. 2A) and mixed species (S. anginosus and E. faecalis) (Fig.2B and C). Attachment was predominantly observed in intercellular spaces within thepulpal matrix and around the pulpal vasculature. Bacteria were also observed attachedto soft tissue surrounding the tooth and within dentinal tubules (Fig. 2D and E).

FIG 1 Growth curves of E. faecalis (A), S. anginosus (B), and E. faecalis and S. anginosus combined at ratios of 50:50 (C) and 90:10 (D). Mean values of threeexperimental repeats are shown, with error bars indicating standard deviations.

TABLE 1 Growth rates of S. anginosus and E. faecalis during log phase, alone and incombination at ratios of 50:50 and 90:10

SpeciesRatio (S. anginosusto E. faecalis)

Avg growth rate duringlog phase (CFU/ml/h) SD

E. faecalis Alone 1.51 0.20S. anginosus 2.00 0.25E. faecalis 50:50 1.62 0.10S. anginosus 1.53 0.14E. faecalis 90:10 1.98 0.12S. anginosus 1.57 0.12

E. faecalis Pulpal Pathogenicity and Attachment Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 3

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

Attachment of bacteria was not detected using Gram staining on tooth slices incubatedwith S. anginosus alone.

Control samples demonstrated low levels of background fluorescence (Fig. 3A).Infections consisting of E. faecalis alone had the greatest fluorescent signal, in particularcentered near the pulpal vasculature (Fig. 3B). S. anginosus demonstrated low bacterialattachment spread evenly across the pulp (Fig. 3C). When, E. faecalis and S. anginosuswere combined, higher levels of attachment were observed than with S. anginosusalone (Fig. 3D and E), with attachment again localized predominantly to the pulpalvasculature. When the percent bacterial coverage was semiquantified (Fig. 3F), thesingle-species E. faecalis infection had significantly higher levels of bacterial attachmentthan S. anginosus alone (approximately 6.5% compared to 2%; P � 0.00021) and themixed-species infections (50:50, P � 0.0235; 90:10, P � 0.0032).

S. anginosus and E. faecalis infections significantly reduce pulp cell viability,with E. faecalis infections inducing a significantly greater inflammatory response.To establish the dental pulp host responses to S. anginosus and E. faecalis infectionsalone and as mixed-species infections, histomorphometric analysis was performedalongside reverse transcription-quantitative PCR (RT-qPCR) and immunohistochemistryfor tumor necrosis factor alpha (TNF-�) and interleukin 1� (IL-1�) expression.

Histological cell counts of the infected tooth sections demonstrated a significantreduction (P � 0.05) in viable cells due to infection by both E. faecalis and S. anginosusalone and in combination (Fig. 4A). There were no significant differences in cellnumbers between single-species infections and multispecies infections.

All infected samples had significantly higher proinflammatory cytokine expression(TNF-� [Fig. 4B] and IL-1� [Fig. 4C]) than the control samples (P � 0.05). The single-species infection with E. faecalis resulted in significantly higher levels of TNF-� andIL-1� expression than S. anginosus infection (P � 0.0276 and P � 0.0234 for TNF-� andIL-1�, respectively). Combining E. faecalis and S. anginosus did not result in a signifi-cantly higher inflammatory response from the pulp than E. faecalis alone (TNF-�, P �

0.493 and P � 0.096 for 50:50 and 90:10, respectively; IL-1�, P � 0.988 and P � 0.400for 50:50 and 90:10, respectively).

Negative controls replacing the primary TNF-� antibody with a nonimmune immu-noglobulin G (IgG) control showed no immunopositivity (see Fig. S3 in the supplemen-tal material). Similarly, primary exclusion controls were negative for staining, indicating

FIG 2 Gram stains of tooth slices infected with E. faecalis (A), 50:50 S. anginosus-E. faecalis (B), and 90:10S. anginosus-E. faecalis (C to E). The arrows indicate areas of bacterial attachment. P, dental pulp; D,dentine; S, soft tissue surrounding the tooth. Representative images of three experimental repeats areshown.

Nishio Ayre et al. Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 4

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

specific binding of the secondary antibody (see Fig. S3 in the supplemental material).Control samples demonstrated low expression of TNF-�, and interestingly, S. anginosusalone did not induce a high TNF-� response (Fig. 4D). Samples incubated with E. faecalisalone or in combination with S. anginosus had the most pronounced staining withinboth the pulp (around the vasculature) and the odontoblast layer. The levels of TNF-�staining in these samples were similar to those encountered in the rat lung positivecontrol (see Fig. S3 in the supplemental material).

Immunohistochemistry staining for IL-1� showed no positive signal for IgG and theprimary exclusion controls (see Fig. S3 in the supplemental material). Similar to theTNF-� immunohistochemistry, the control sample and the sample incubated with S.anginosus alone had few positively stained cells, while samples incubated with E.faecalis alone and in combination with S. anginosus had more positively stained cells(Fig. 4D). Although the level of staining was not as pronounced as that observed withTNF-�, the positive cells were again located adjacent to the pulpal vasculature andsimilar in staining to the positive lung control (see Fig. S3 in the supplemental material).

Greater host inflammatory response to E. faecalis is not due to differences inwater-soluble cell wall proteins or culture supernatants. To establish whether theincreased host inflammatory response to E. faecalis was due to specific water-solublecell proteins or components of the culture supernatant, SDS-PAGE was performed toidentify proteins in water-soluble cell wall proteins and culture supernatants. Similarly,heat-killed E. faecalis and the E. faecalis supernatant were used to stimulate the pulp inorder to assess the host response.

Few differences were observed between the water-soluble cell wall proteins of S.anginosus and E. faecalis when cultured alone and in combination with each other (see

FIG 3 (A to E) Localization of bacterial attachment by fluorescence microscopy for tooth slices infectedwith no-bacteria control (A), E. faecalis (B), S. anginosus (C), 50:50 S. anginosus-E. faecalis (D), and 90:10S. anginosus-E. faecalis (E). P, dental pulp; O, odontoblast region; D, dentine. Representative images ofthree experimental repeats are shown. (F) Bacterial coverage as quantified by the area of fluorescencerelative to the total pulp area (*, P � 0.05; **, P � 0.01; ***, P � 0.001). Mean values of three experimentalrepeats are shown, with error bars indicating standard errors of the mean.

E. faecalis Pulpal Pathogenicity and Attachment Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 5

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

Fig. S4A in the supplemental material). In terms of the culture supernatant, there wasone band at approximately 35 kDa observed with the E. faecalis cultures that was notobserved with S. anginosus (see Fig. S4B in the supplemental material).

When culturing the rat tooth slices with the E. faecalis supernatant or heat-killed E.faecalis, no significant differences were observed in TNF-� expression compared to theuntreated controls (Fig. 5A) (P � 0.196 and P � 0.152 for supernatant and heat-killed

FIG 4 (A) Viable cells counted per square millimeter of pulp. Tooth slices infected with E. faecalis and S. anginosus, both alone and incombination, after 24 h all resulted in a significant reduction in viable-cell numbers in the pulp compared to the noninfected control (*, P � 0.05).Mean values of three experimental repeats are shown, with error bars indicating standard errors of the mean. (B and C) Fold change in TNF-�(B) and IL-1� (C) gene expression as a result of E. faecalis and S. anginosus infections, alone and in combination (*, P � 0.05, and **, P � 0.01compared to control samples; �, P � 0.05, and ��, P � 0.01). Mean values of three experimental repeats are shown, with error bars indicatingstandard errors of the mean. (D) Immunohistochemistry of TNF-� and IL-1� for control samples and tooth slices infected with S. anginosus, E.faecalis, 50:50 S. anginosus-E. faecalis, and 90:10 S. anginosus-E. faecalis. Representative images of three experimental repeats are shown.

Nishio Ayre et al. Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 6

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

E. faecalis, respectively). A significant increase was observed in IL-1� expression for thetooth slices cultured with heat-killed E. faecalis compared to the untreated controls (Fig.5B) (P � 0.041), but not for E. faecalis supernatant (P � 0.148).

The negative controls (IgG and primary exclusion controls) and the control samplefor the TNF-� immunohistochemistry did not show staining (see Fig. S5 in the supple-mental material). The tooth slices incubated with E. faecalis supernatant had few cellsstained positive for TNF-�, the majority of which were concentrated at the pulpalvasculature and the odontoblast layer (Fig. 5C). Similarly, heat-killed E. faecalis had fewcells expressing TNF-� (Fig. 5C), while the lung positive control stained positive forTNF-� (see Fig. S5 in the supplemental material).

The IgG control, the primary exclusion control, and the untreated sample (see Fig.S5 in the supplemental material) did not stain positive for IL-1�. Fewer cells werepositive for IL-1� than for TNF-� (Fig. 5C). Samples treated with E. faecalis supernatantshowed some cells stained positive within the pulpal vasculature, while heat-killed E.faecalis showed few positively stained cells. The positive lung control demonstratedcells stained positive for IL-1� expression (see Fig. S5 in the supplemental material).

DISCUSSION

This study successfully employed an existing ex vivo rat tooth infection model tostudy the effect of mixed-species E. faecalis and S. anginosus pulpal infections on cellviability, bacterial attachment, and host inflammatory response.

By studying simple planktonic growth kinetics, it was established that E. faecaliscaused the S. anginosus bacteria to reach log phase at a higher rate. This concept ofpolymicrobial synergy has been highlighted in recent work, which investigated metab-olite cross-feeding, where metabolic end products produced by one bacterium are

FIG 5 (A and B) Fold change in TNF-� (A) and IL-1� (B) gene expression relative to �–actin as a result of treatingtooth slices with E. faecalis supernatant and heat-killed E. faecalis (*, P � 0.05 compared to control samples). Meanvalues of three experimental repeats are shown, with error bars indicating standard errors of the mean. (C)Immunohistochemistry of TNF-� and IL-1� for control samples and tooth slices infected with E. faecalis supernatantand heat-killed E. faecalis. Representative images of three experimental repeats are shown.

E. faecalis Pulpal Pathogenicity and Attachment Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 7

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

consumed by a second community member (18–20). In particular, this has beendemonstrated for a similar oral pathogen, Streptococcus gordonii. Lactate produced byS. gordonii as the primary metabolite during catabolism of carbohydrates was found tosupport the growth of Aggregatibacter actinomycetemcomitans (20). Interestingly, in astudy using a primate model, the addition of E. faecalis to a four-strain mixed-speciesculture resulted in higher levels of survival of all four strains of bacteria than in theabsence of E. faecalis (21). Another mechanism for coordinating activities and commu-nicating between microbial species is quorum sensing, which has been shown to occurbetween different groups of streptococci (22). Although the rate of growth during thelog phase was not altered during mixed-species planktonic culture in this study, it isimportant to appreciate that under mixed-species biofilm conditions, alterations ingrowth are likely to occur.

The mixed-species infection did not result in higher levels of bacterial attachmentthan with E. faecalis alone. The data suggest that E. faecalis is capable of attaching tothe dental pulp to a greater extent than S. anginosus, with a particular affinity for thepulpal vasculature. This was not attributed to a higher rate of growth or a highernumber of bacteria, as a similar number of S. anginosus cells were counted after 24 hin planktonic broth culture. Similarly, in the mixed-species culture, where S. anginosusachieved log phase at an earlier time point, attachment was not as high as for E. faecalisalone. The increased attachment may therefore be due to differences between thespecies in terms of motility, sensing, or cell surface adhesins. E. faecalis and S. anginosusare classified as groups D and F, respectively, using Lancefield grouping (23), a methodof grouping based on the carbohydrate antigens on the cell wall. These differences insurface carbohydrates could mediate changes in attachment to epithelial cells, asdemonstrated by Guzmàn et al. (24). A review by Fisher and Phillips (25) highlighted E.faecalis-specific cell wall components that play a vital role in pathogenic adhesion.Aggregation substance (Agg) increases hydrophobicity and aids adhesion to eukaryoticand prokaryotic surfaces and also encourages the formation of mixed-species biofilmsthrough adherence to other bacteria. Extracellular surface protein (ESP) promotes adhesion,antibiotic resistance, and biofilm formation. Adhesin to collagen of E. faecalis (ACE) is acollagen binding protein belonging to the microbial surface components recognizingadhesive matrix molecules (MSCRAMM) family. ACE plays a role in the pathogenesis ofendocarditis, and E. faecalis mutants that do not express ACE have been shown to havesignificantly reduced attachment to collagen types I and IV but not fibrinogen (26, 27).While S. anginosus has been shown to adhere to the extracellular matrix componentsfibronectin, fibrinogen, and laminin, binding to collagen types I and IV was much lessprominent (28). This is of particular interest in explaining differences in pulpal adher-ence and the tendency of E. faecalis to localize near the pulpal vasculature, as collagenfibers are often found at higher density around blood vessels and nerves (29).

Although the levels of cell death were the same in the groups tested, infectionsconsisting of E. faecalis alone produced a greater inflammatory response than S.anginosus and mixed-species infections. This increase in inflammation was not due tosupernatant or water-soluble cell wall virulence factors of E. faecalis, as treatment of thedental pulp with these isolated factors did not yield high levels of TNF-� and IL-1�

expression at gene and protein levels. Basic analysis of supernatant and water-solublecell wall proteins by SDS-PAGE showed similar bands; however, this might have beendue to the absence of serum or collagen (present in the coculture model), which havebeen shown to influence the production of virulence factors, such as ACE (27). Theseresults indicate the pulpal inflammation caused by E. faecalis is likely due to the higherlevels of attachment to the dental pulp. Similar pathogenic traits have been establishedfor E. faecalis in urinary tract infections and endocarditis (30). Increased attachment tothe dental pulp would allow direct contact between cells and cell wall components,such as lipoteichoic acid (LTA), which induces activation of cluster of differentiation 14(CD-14) and Toll-like receptor 2 (TLR-2) (31). An in vivo study that infected canine pulpwith lipopolysaccharides (LPS) from Escherichia coli and LTA from E. faecalis demon-strated that LTA treatment led to pulp destruction, albeit to a lesser extent than LPA

Nishio Ayre et al. Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 8

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

(32). In vitro studies investigating macrophage responses to E. faecalis LTA found thatTNF-� expression was significantly increased in a dose-dependent manner (33), withone study attributing it to the NF-�B and p38 MAPK signaling pathways (34). Thesestudies, however, were performed using monolayer cultures, allowing easy access forLTA to activate Toll-like receptors, whereas the presence of extracellular matrix wouldlimit the penetration of virulence factors into the dental pulp in vivo. Furthermore,macrophages are normally present as monocytes in normal healthy pulp and require astimulus to become activated (35). Studies using immunohistochemistry have shownthese monocytes, as well as dendritic cells, to be located predominantly around bloodvessels, with few distributed throughout the pulp (36, 37).

High levels of TNF-� expression were also observed in the odontoblast region usingimmunohistochemistry. Due to their anatomical location, odontoblasts are the first cellsto encounter foreign antigens through either infiltration of virulence factors throughdentinal tubules or the breakdown of enamel and dentine. By Gram staining in thisstudy, E. faecalis was observed within the dentinal tubules of the infected tooth slices.This phenomenon has been previously reported in human teeth (38). Odontoblasts,which line the dentine, have been shown to express TLRs and to play a role in the pulp’simmune response, in particular to bacterial exotoxins (39–41). This explains the highinflammatory response observed for both infections and supernatant treatments whenassessed using immunohistochemistry. Cytokine gene expression using RT-qPCR, how-ever, did not demonstrate higher levels when the dental pulp was treated withsupernatants or heat-killed bacteria. This may be attributed to the fact that themethods employed for pulp extraction would be unlikely to fully remove the odonto-blast cells.

Although the host response to a mixed-species infection consisting of S. anginosusand E. faecalis was established and the potential pathogenicity of E. faecalis in pulpalinfections was elucidated, there are several limitations to this study. The methodsemployed to fluorescently localize the bacteria could potentially result in diffusion-related artifacts. More specific postprocessing techniques, such as fluorescent in situhybridization (FISH) probes, may allow more specific identification, quantification, andlocalization of mixed-species pulpal infections. While the ex vivo model offers a three-dimensional (3D) organotypic culture setting, the static nature, which lacks blood flow,does not allow full observation of the systemic immune response. Potential methods toovercome this may involve addition of monocytes directly to the culture media andprolonged incubation times to stimulate repair mechanisms. Closer examination ofattachment mechanisms using ACE-negative E. faecalis mutants and purified LTA wouldalso help fully establish the pathogenicity of E. faecalis in pulpal infections. This willallow the model to be used to develop more effective treatments for pulpitis byassessing the efficacy of antimicrobial and anti-inflammatory treatments in inhibitingbacterial colonization.

In conclusion, this study modeled a mixed-species pulpal infection consisting of S.anginosus and E. faecalis using a validated ex vivo rat tooth model. Although E. faecaliscaused S. anginosus to reach the logarithmic growth phase more rapidly, the mixed-species infection did not result in higher levels of cell death, attachment, or inflamma-tory response from the dental pulp. E. faecalis was found to elicit a much greaterinflammatory response, which was due to higher levels of attachment to the dentalpulp, with a particular affinity for the pulpal vasculature. Future work will focus onassessing the mechanisms and attachment kinetics in order to elucidate the molecularprocess and the rate at which E. faecalis colonizes the pulp.

MATERIALS AND METHODSMaterials. All reagents, including culture media, broths, and agars, were purchased from Thermo

Scientific (Leicestershire, UK) unless otherwise stated.Bacterial identification. The S. anginosus and E. faecalis strains studied were clinical isolates selected

from the culture collection of the Oral Microbiology Unit, School of Dentistry at Cardiff University. Toconfirm the identity of the species, standard microbial identification tests were performed by assessing

E. faecalis Pulpal Pathogenicity and Attachment Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 9

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

the colony appearance on blood agar, Gram staining, hemolysis, presence of catalase, lactose fermen-tation (MacConkey agar), Lancefield grouping, and bile esculin agar growth.

16S rRNA sequencing was also performed on the S. anginosus and E. faecalis clinical isolates tovalidate species identity. S. anginosus and E. faecalis were cultured overnight in BHI broth at 37°C and 5%CO2. DNA was extracted using a QIAamp DNA minikit (Qiagen, Manchester, UK), according to themanufacturer’s instructions. The DNA was used in a PCR with 16S rRNA bacterial universal primers D88(F primer; 5=-GAGAGTTTGATYMTGGCTCAG-3=) and E94 (R primer; 5=-GAAGGAGGTGWTCCARCCGCA-3=)(42), and sequencing of the products was performed by Central Biotechnology Services (Cardiff Univer-sity) using a 3130xl genetic analyzer (Applied Biosystems). DNA sequences were aligned with GenBanksequences using BLAST (NCBI) to establish the percent sequence identity.

Growth curves. Overnight cultures of S. anginosus and E. faecalis in BHI broth were prepared anddiluted to 108 CFU/ml (absorbance at 600 nm, 0.08 to 0.1). The inoculum was diluted in BHI broth to givea starting concentration of 102 CFU/ml. Mixed-species planktonic cultures with a total of 102 CFU/mlconsisting of 50% S. anginosus and 50% E. faecalis (referred to here as 50:50) and 90% S. anginosus and10% E. faecalis (referred to here as 90:10) were prepared. The broths were incubated at 37°C and 5% CO2,and 1-ml aliquots were removed every 4 h for 24 h. The absorbance of the aliquots was measured at 600nm using an Implen (Munich, Germany) OD600 DiluPhotometer, and 50 �l was spiral plated on trypticsoya agar using a Don Whitley (West Yorkshire, UK) automated spiral plater. The remaining aliquot wasthen heat treated at 60°C for 30 min prior to spiral plating on bile esculin agar containing 6.5% (wt/wt)sodium chloride. Heat treatment and the presence of high concentrations of bile and sodium chloridepermit the growth of only E. faecalis and not S. anginosus (43). The plates were incubated at 37°C and5% CO2 for 24 h prior to counting. The E. faecalis counts were subtracted from the total counts to givethe number of S. anginosus bacteria. The specific growth rate was calculated using the log phase of eachgrowth curve and the following equation: � � ln(x � x0)/t, where � is the growth rate in CFU per milliliterper hour, x is the number of CFU per milliliter at the end of the log phase, x0 is the number of CFU permilliliter at the start of the log phase, and t is the duration of the log phase in hours.

Coculture model. The coculture rat tooth infection model was prepared as described by Robertset al. (7). Twenty-eight-day-old male Wistar rats were sacrificed under schedule 1 of the UK AnimalsScientific Procedures Act, 1986, by a qualified technician at the Joint Biological Services Unit, CardiffUniversity, for harvesting of tissue. Upper and lower incisors were extracted, and the incisors were cutinto 2-mm-thick transverse sections using a diamond-edged rotary bone saw (TAAB, Berkshire, UK). Thesections were transferred to fresh sterile Dulbecco’s modified Eagle medium (DMEM) for no more than20 min before being cultured in 2 ml DMEM supplemented with 10% (vol/vol) heat-inactivated fetal calfserum, 0.15 mg/ml vitamin C, 200 mmol/liter L-glutamine, 100 U/ml penicillin, 100 �g/ml streptomycinsulfate, and 250 ng/ml amphotericin B at 37°C and 5% CO2 for 24 h. The tooth slices were then washedin 2 ml of phosphate-buffered saline (PBS), transferred to supplemented DMEM without antibiotics, andincubated overnight to remove traces of antibiotic. S. anginosus and E. faecalis were cultured to log phasein BHI broth for 8 to 12 h before dilution to 102 CFU/ml in BHI broth. The bacteria were then used aloneor combined for mixed-species infections (S. anginosus/E. faecalis ratios of 50:50 and 90:10). Fortymicroliters of 1% (wt/vol) fluorescein diacetate (FDA) in acetone was added to 2 ml of the bacterialsuspension and incubated for 30 min at 37°C and 5% CO2 before being passed through a 0.22-�msyringe-driven filter unit (Millipore, Oxford, UK). Bacteria captured on the filter were then resuspended in2 ml sterile supplemented DMEM without antibiotics and with 10% (vol/vol) BHI (referred to here asDMEM-BHI) and used to inoculate one tooth slice. The tooth slices were incubated with the bacteria at37°C and 5% CO2 for 24 h under constant agitation at 60 rpm in the dark. Sterile DMEM-BHI was usedas a control. After incubation, the tooth slices were processed for histology in the dark. The tooth sliceswere fixed in 10% (wt/vol) neutral buffered formalin at room temperature for 24 h. The slices weredemineralized in 10% (wt/vol) formic acid at room temperature for 72 h; dehydrated through a series of50% (vol/vol), 70% (vol/vol), 95% (vol/vol), and 100% (vol/vol) ethanol, followed by 100% (vol/vol) xylene,for 5 min each; and embedded in paraffin wax. Sections 5 �m thick were cut and viewed under afluorescence microscope with a fluorescein isothiocyanate (FITC) filter, with images captured using aNikon digital camera and ACT-1 imaging software (Nikon UK Ltd., Surrey, UK). To quantify cell viabilityand structural degradation, sections were stained with hematoxylin and eosin (H&E) prior to capturingimages with a light microscope.

Gram stain of tissue sections. Gram stains of tooth slices were performed using a modified Brownand Brenn method (44). Paraffin-embedded tooth slices were cut, using a microtome, into 5-�m sectionsand rehydrated through a series of xylene and 100, 95, and 70% (vol/vol) ethanol for 5 min each. Thesections were immersed in 0.2% (wt/vol) crystal violet for 1 min, rinsed with distilled water, immersed inGram’s iodine for 1 min, rinsed with distilled water, decolorized with acetone for 5 s, and counterstainedfor 1 min with basic fuchsin solution prior to washing with distilled water and mounting. Lightmicroscopy images were captured at �100 magnification using a Nikon digital camera and ACT-1imaging software (Nikon UK Ltd., Surrey, UK).

Semiquantification of cell viability by cell counts. ImageJ (National Institutes of Health, Bethesda,MD, USA) was used to count the nuclei per pulp on stained histological sections. For each time point,sections were cut from 5 tooth slices. Images were captured at �20 magnification and combined usingImageJ software (see Fig. S6 in the supplemental material). The blue field was extracted from the images,and the moments threshold method was applied to separate the pulp cells. The watershed function wasapplied to split adjacent cell nuclei, and the particles ranging in size from 3 to 100 �m2 were counted.The data were normalized to the pulpal area, and standard errors of the mean were calculated.

Nishio Ayre et al. Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 10

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

Semiquantification of bacterial coverage. ImageJ was used to quantify the area of the pulpinoculated with fluorescent bacteria. The green field of the fluorescent image was extracted, and theimage was converted into binary form using the moments threshold method. The pulpal area wasmanually selected, and the total area of the pulp was measured. The area covered by fluorescent bacteriawas then measured and calculated as a percentage of the selected pulp area (see Fig. S7 in thesupplemental material).

RT-qPCR of cytokines. Four-millimeter-thick tooth slices were cultured as previously described for24 h with either sterile DMEM-BHI as a control, DMEM-BHI inoculated with 102 CFU/ml S. anginosus or E.faecalis, or DMEM-BHI with mixed species (S. anginosus and E. faecalis at 50:50 and 90:10 ratios). Afterincubation, the tooth slices were transferred to sterile PBS, and the pulp was removed by flushing thepulpal cavity with PBS using a 0.1-mm needle and syringe. RNA was extracted using TRIzol reagent(ThermoFisher Scientific, Loughborough, UK), followed by RNase treatment (Promega, Southampton, UK)according to the manufacturers’ instructions.

Analysis of gene expression was performed in accordance with the Minimum Information forPublication of Quantitative Real-Time PCR Experiments (MIQE) guidelines (45). RNA concentrations weredetermined using a NanoVue spectrophotometer (GE Healthcare Life Sciences, Buckinghamshire, UK).RNA purity was determined by ensuring the ratio of absorbance at 260/280 nm was above 1.8, and RNAquality was checked by separating 1 �g of RNA electrophoretically on a 2% agarose gel containingSafeView (NBS Biologicals, Cambridgeshire, UK) in Tris-borate-EDTA buffer to ensure intact 28S and 18SrRNA bands using a Gel Doc EZ system (Bio-Rad, Hertfordshire, UK). Figure S8 in the supplementalmaterial shows RNA integrity following extraction for the samples tested.

cDNA was synthesized by reverse transcription using Promega (Southampton, UK) reagents in aG-Storm (Somerton, UK) GS1 thermocycler. One microgram extracted RNA was combined with 1 �lrandom primer in a 15-�l reaction mixture in nuclease-free water at 70°C for 5 min. The suspension wasadded to 5 �l Moloney murine leukemia virus (MMLV) reaction buffer (1.25 �l deoxyribonucleotidetriphosphates [dNTPs] [10 mM stock dNTPs], 0.6 �l RNasin, 1 �l MMLV enzyme, and 2.15 �l nuclease-freewater) and incubated at 37°C for 1 h.

The resultant cDNA was diluted 1:10 in nuclease-free water (25 ng cDNA). The forward and reverseprimers used are listed in Table 2. Ten microliters of PrecisionFast qPCR SYBR green MasterMix with lowROX (PrimerDesign, Chandler’s Ford, United Kingdom) was combined with 2 �l of forward and 2 �l ofreverse primers (3 �M) with 1 �l nuclease-free water prior to addition of 5 �l cDNA in BrightWhitereal-time PCR fast 96-well plates (PrimerDesign, Chandler’s Ford, United Kingdom). The plates weresubsequently heated to 95°C for 20 s, followed by 40 cycles of 95°C for 1 s and 55°C for 20 s andmelting-curve analysis at 95°C for 15 s, 60°C for 60 s, and 95°C for 15 s in a QuantStudio 6 Flex real-timePCR system with QuantStudio real-time PCR software (ThermoFisher Scientific, Loughborough, UK).Relative TNF-� and IL-1� gene expression was calculated with the �-actin gene as the reference geneand uninfected samples as the control using the method of Livak and Schmittgen (46).

Primer specificity was ensured by the presence of single melting-curve peaks (see Fig. S9 in thesupplemental material) and by running products on agarose gels, as previously described, to confirmsingle bands and correct product lengths (see Fig. S10 in the supplemental material). Primer efficiencywas between 90 and 110% for all the primers used (see Fig. S11 in the supplemental material) and wasdetermined using total rat RNA converted to cDNA, as previously described, and serially diluted 1:4 innuclease-free water. Reference gene validation was performed by comparing gene stabilities across allthe samples using NormFinder software (47). The �-actin gene was found to be the most stable referencegene (see Fig. S12 in the supplemental material).

TNF-� and IL-1� immunohistochemistry. Immunohistochemical staining of the tooth slices forTNF-� and IL-1� was performed based on methods used by Smith et al. (48). Rat lung was used as a

TABLE 2 Sequences of primers used for qPCR analysis

Gene Accession no.

PrimerProductlength (bp)

Meltingtemp (°C) Efficiency (%) ReferenceOrientation Sequence (5=-3=)

GAPDHa NM_017008.4 Forward GCA AGA GAG AGG CCC TCA G 74 61.0 106.37 48Reverse TGT GAG GGA GAT GCT CAG TG 59.4

�-Actin NM_031144.3 Forward TGA AGA TCA AGA TCA TTG CTC CTC C 155 60.69 108.56 49Reverse CTA GAA GCA TTT GCG GTG GAC GAT G 64.37

HPRT-1b NM_012583.2 Forward TGT TTG TGT CAT CAG CGA AAG TG 66 60.24 91.71 50Reverse ATT CAA CTT GCC GCT GTC TTT TA 59.43

RPL13ac NM_173340.2 Forward GGA TCC CTC CAC CCT ATG ACA 131 61.8 99.99 51Reverse CTG GTA CTT CCA CCC GAC CTC 63.7

18S rRNA V01270 Forward AAA CGG CTA CCA CAT CCA AG 159 57.3 90.22 52Reverse TTG CCC TCC AAT GGA TCC T 56.7

TNF-� NM_012675.3 Forward AAA TGG GCT CCC TCT CAT CAG TTC 111 62.7 90.28 53Reverse TCT GCT TGG TGG TTT GCT ACG AC 62.4

IL-1� NM_031512.2 Forward ATG CCT CGT GCT GTC TGA CCC ATG TGA G 135 70.06 94.80 54Reverse CCC AAG GCC ACA GGG ATT TTG TCG TTG C 70.16

aGAPDH, glyceraldehyde 3-phosphate dehydrogenase.bHPRT-1, hypoxanthine phosphoribosyltransferase 1.cRPL13, ribosomal protein L13a.

E. faecalis Pulpal Pathogenicity and Attachment Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 11

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

positive control for TNF-� and IL-1� following fixation in 10% (wt/vol) neutral buffered formalin at roomtemperature for 24 h; dehydration through a series of 50% (vol/vol), 70% (vol/vol), 95% (vol/vol), and100% (vol/vol) ethanol, followed by 100% (vol/vol) xylene, for 5 min each; and embedding in paraffinwax. The paraffin-embedded tooth slices and lung samples were cut, using a microtome, into 5-�msections and incubated on glass slides at 65°C for 1 h. The samples were subsequently rehydratedthrough a series of xylene and 100% (vol/vol), 95% (vol/vol), and 70% (vol/vol) ethanol and double-distilled water for 5 min each. Endogenous peroxidase activity within the tissue sections was quenchedby incubation in 3% (wt/vol) hydrogen peroxide for 10 min, followed by 2 washes for 2 min inTris-buffered saline (TBS). Nonspecific binding was blocked with 3% (vol/vol) normal horse serum (VectorLaboratories, Peterborough, UK) in TBS for 30 min. The sections were incubated for 1 h with primaryantibodies for TNF-� and IL-1� (Santa Cruz Biotechnology, Heidelberg, Germany) diluted 1:50 in TBScontaining 1% (wt/vol) bovine serum albumin (Sigma-Aldrich, Gillingham, UK). An immunoreactivityassay was then performed using a Vectastain ABC peroxidase detection kit (Vector Laboratories,Peterborough, UK). Negative controls included omission of the primary antibody and replacement of theprimary antibody with immunoglobulin G isotype diluted to the working concentration of the primaryantibody. The sections were counterstained with 0.05% light green for 30 s, dehydrated with 100%ethanol and xylene for 10 min each, and mounted using VectaMount permanent mounting medium(Vector Laboratories, Peterborough, UK) prior to imaging using a Nikon digital camera and ACT-1imaging software (Nikon UK Ltd., Surrey, UK).

SDS-PAGE of bacterial proteins. An overnight culture of S. anginosus and E. faecalis in BHI broth wasprepared and diluted to 102 CFU/ml. S. anginosus and E. faecalis were cultured at 37°C and 5% CO2 for24 h alone or in combination at ratios of 50:50 and 90:10. The suspensions were centrifuged at 5,000 �g for 5 min, and the supernatant was used for analysis of supernatant proteins. The pellet was lysed inradioimmunoprecipitation assay (RIPA) buffer by vortexing for 30 s followed by 30 s ultrasonication at 50J using a Branson (Connecticut, USA) SLPe sonifier. Protein concentrations in the supernatant and thebacterial pellet were quantified using a bicinchoninic acid (BCA) assay (ThermoFisher Scientific, Lough-borough, UK) and 20 �g of protein in Laemmli buffer (Bio-Rad, Hertfordshire, UK) separated by SDS-PAGEat 200 V for 40 min. Gels were stained using a Bio-Rad Silver Stain Plus kit according to the manufacturer’sinstructions and imaged using a Gel Doc EZ System (Bio-Rad, Hertfordshire, UK).

E. faecalis supernatant and heat-killed E. faecalis treatments. An overnight culture of E. faecaliswas diluted in 20 ml DMEM-BHI medium to give a starting inoculum of 102 CFU/ml, as previouslydescribed. After incubation for an additional 24 h at 37°C and 5% CO2, the suspension was centrifugedat 5,000 � g for 5 min. The supernatant was filtered through a 0.22-�m syringe filter and frozenovernight at �20°C before freeze-drying for 24 h using a ScanVac CoolSafe freeze-dryer (LaboGene,Lynge, Denmark). The pellet of bacteria was resuspended in 20 ml of PBS and centrifuged at 5,000 � gfor 5 min. This step was repeated to ensure minimal carryover of the culture supernatant. The pellet wasthen resuspended in 20 ml DMEM-BHI and heated to 100°C for 1 h. The solution was then frozenovernight at �20°C before freeze-drying as previously described. Twenty milliliters of sterile DMEM-BHIwas also frozen and freeze-dried as a control. All freeze-dried samples were individually resuspended in20 ml of sterile DMEM-BHI and used to culture rat tooth slices for RT-qPCR of cytokines and immuno-histochemistry of TNF-� and IL-1�, as previously described.

Statistical analysis. A one-way analysis of variance (ANOVA) was performed using the data analysispackage in Excel (Microsoft, Reading, UK) to determine the relative significance of the differencesbetween the infected groups and the controls in terms of cell counts, bacterial coverage, and cytokineexpression. The Tukey-Kramer test was used in conjunction with ANOVA to compare the significantdifferences between all possible pairs of means. A P value of �0.05 was considered significant.

SUPPLEMENTAL MATERIAL

Supplemental material for this article may be found at https://doi.org/10.1128/IAI.00871-17.

SUPPLEMENTAL FILE 1, PDF file, 1.4 MB.

ACKNOWLEDGMENTThis work was supported by the Dunhill Medical Trust (grant number R232/1111).

REFERENCES1. Nanci A. 2013. Dentin-pulp complex, p 183. In Nanci A (ed.), Ten cate’s

oral histology: development, structure, and function, 8th ed. MosbyElsevier, St. Louis, MO.

2. Ng YL, Mann V, Rahbaran S, Lewsey J, Gulabivala K. 2007. Outcome ofprimary root canal treatment: systematic review of the literature. Part 1.Effects of study characteristics on probability of success. Int Endod J40:921–939.

3. Mukhaimer R, Hussein E, Orafi I. 2012. Prevalence of apical periodontitis andquality of root canal treatment in an adult Palestinian sub-population. SaudiDent J 24:149–155. https://doi.org/10.1016/j.sdentj.2012.02.001.

4. Berlinck T, Tinoco JMM, Carvalho FLF, Sassone LM, Tinoco EMB. 2015.

Epidemiological evaluation of apical periodontitis prevalence in an ur-ban Brazilian population. Braz Oral Res 29:1–7. https://doi.org/10.1590/1807-3107BOR-2015.vol29.0051.

5. Anonymous. 2016. Guideline on pulp therapy for primary and immaturepermanent teeth. Pediatr Dent 38:280 –288.

6. Demarco FF, Rosa MS, Tarquínio SBC, Piva E. 2005. Influence of therestoration quality on the success of pulpotomy treatment: a preliminaryretrospective study. J Appl Oral Sci 13:72–77. https://doi.org/10.1590/S1678-77572005000100015.

7. Roberts JL, Maillard JY, Waddington RJ, Denyer SP, Lynch CD, Sloan AJ.2013. Development of an ex vivo coculture system to model pulpal

Nishio Ayre et al. Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 12

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

infection by Streptococcus anginosus group bacteria. J Endod 39:49 –56.https://doi.org/10.1016/j.joen.2012.09.005.

8. Jenkinson HF, Demuth DR. 1997. Structure, function and immunogenic-ity of streptococcal antigen I/II polypeptides. Mol Microbiol 23:183–190.https://doi.org/10.1046/j.1365-2958.1997.2021577.x.

9. Shweta S, Prakash SK. 2013. Dental abscess: a microbiological review.Dent Res J 10:585–591.

10. Bik EM, Long CD, Armitage GC, Loomer P, Emerson J, Mongodin EF,Nelson KE, Gill SR, Fraser-Liggett CM, Relman DA. 2010. Bacterial diver-sity in the oral cavity of ten healthy individuals. ISME J 4:962–974.https://doi.org/10.1038/ismej.2010.30.

11. Rôças IN, Alves FRF, Rachid CT, Lima KC, Assunção IV, Gomes PN, SiqueiraJF, Jr. 2016. Microbiome of deep dentinal caries lesions in teeth withsymptomatic irreversible pulpitis. PLoS One 11:e0154653. https://doi.org/10.1371/journal.pone.0154653.

12. Hahn CL, Liewehr FR. 2007. Relationships between caries bacteria, hostresponses, and clinical signs and symptoms of pulpitis. J Endod 33:213–219. https://doi.org/10.1016/j.joen.2006.11.008.

13. Suchitra U, Kundabala M. 2006. Enterococcus faecalis: an endodonticpathogen. Endodontology 18:11–13.

14. Sundqvist G, Figdor D, Persson S, Sjogren U. 1998. Microbiologic analysisof teeth with failed endodontic treatment and the outcome of conser-vative re-treatment. Oral Surg Oral Med Oral Pathol Oral Radiol Endod85:86 –93. https://doi.org/10.1016/S1079-2104(98)90404-8.

15. Stuart CH, Schwartz SA, Beeson TJ, Owatz CB. 2006. Enterococcus faecalis:its role in root canal treatment failure and current concepts in retreat-ment. J Endod 32:93–98.

16. Salah R, Dar-Odeh N, Abu Hammad O, Shehabi AA. 2008. Prevalence ofputative virulence factors and antimicrobial susceptibility of Enterococ-cus faecalis isolates from patients with dental diseases. BMC Oral Health8:17. https://doi.org/10.1186/1472-6831-8-17.

17. Hegde A, Lakshmi P. 2013. Prevalence of selected microorganisms in thepulp space of human deciduous teeth with irreversible pulpitis. Endod-ontology 25:107–111.

18. Brown SA, Whiteley M. 2007. A novel exclusion mechanism for carbonresource partitioning in Aggregatibacter actinomycetemcomitans. J Bac-teriol 189:6407– 6414. https://doi.org/10.1128/JB.00554-07.

19. Ramsey MM, Whiteley M. 2009. Polymicrobial interactions stimulateresistance to host innate immunity through metabolite perception.Proc Natl Acad Sci U S A 106:1578 –1583. https://doi.org/10.1073/pnas.0809533106.

20. Ramsey MM, Rumbaugh KP, Whiteley M. 2011. Metabolite cross-feedingenhances virulence in a model polymicrobial infection. PLoS Pathog7:e1002012. https://doi.org/10.1371/journal.ppat.1002012.

21. Fabricius L, Dahlen G, Sundqvist G, Happonen RP, Moller AJ. 2006.Influence of residual bacteria on periapical tissue healing after chemo-mechanical treatment and root filling of experimentally infected mon-key teeth. Eur J Oral Sci 114:278 –285. https://doi.org/10.1111/j.1600-0722.2006.00380.x.

22. Cook LC, LaSarre B, Federle MJ. 2013. Interspecies communicationamong commensal and pathogenic streptococci. mBio 4:e00382-13.https://doi.org/10.1128/mBio.00382-13.

23. Lancefield RC. 1933. A serological differentiation of human and othergroups of hemolytic streptococci. J Exp Med 57:571–595. https://doi.org/10.1084/jem.57.4.571.

24. Guzmàn CA, Pruzzo C, Platé M, Guardati MC, Calegari L. 1991. Serumdependent expression of Enterococcus faecalis adhesins involved in thecolonization of heart cells. Microb Pathog 11:399 – 409. https://doi.org/10.1016/0882-4010(91)90036-A.

25. Fisher K, Phillips C. 2009. The ecology, epidemiology and virulence ofenterococcus. Microbiology 155:1749 –1757. https://doi.org/10.1099/mic.0.026385-0.

26. Nallapareddy SR, Qin X, Weinstock GM, Höök M, Murray BE. 2000. Entero-coccus faecalis adhesin, ace, mediates attachment to extracellular matrixproteins collagen type iv and laminin as well as collagen type i. InfectImmun 68:5218–5224. https://doi.org/10.1128/IAI.68.9.5218-5224.2000.

27. Singh KV, Nallapareddy SR, Sillanpaa J, Murray BE. 2010. Importance ofthe collagen adhesin ace in pathogenesis and protection against Entero-coccus faecalis experimental endocarditis. PLoS Pathog 6:e1000716.https://doi.org/10.1371/journal.ppat.1000716.

28. Allen BL, Katz B, Hook M. 2002. Streptococcus anginosus adheres tovascular endothelium basement membrane and purified extracellularmatrix proteins. Microb Pathog 32:191–204. https://doi.org/10.1006/mpat.2002.0496.

29. Gong Q, He L, Liu Y, Zhong J, Wang S, Xie M, Sun S, Zheng J, Xiang L,Ricupero CL, Nie H, Ling J, Mao JJ. 2017. Biomaterials selection for dentalpulp regeneration, p 159 –173. In Ducheyne P (ed), Comprehensivebiomaterials II, 2nd ed. Elsevier, Oxford, United Kingdom.

30. Guzmàn CA, Pruzzo C, Li Pira G, Calegari L. 1989. Role of adherence inpathogenesis of Enterococcus faecalis urinary tract infection and endo-carditis. Infect Immun 57:1834 –1838.

31. Park OJ, Han JY, Baik JE, Jeon JH, Kang SS, Yun CH, Oh JW, Seo HS, HanSH. 2013. Lipoteichoic acid of Enterococcus faecalis induces the expres-sion of chemokines via TLR2 and PAFR signaling pathways. J Leukoc Biol94:1275–1284. https://doi.org/10.1189/jlb.1012522.

32. de Oliveira LA, Barbosa SV. 2003. The reaction of dental pulp to Esche-richia coli lipopolysaccharide and Enterococcus faecalis lipoteichoic acid.Braz J Microbiol 34:179 –181.

33. Baik JE, Ryu YH, Han JY, Im J, Kum KY, Yun CH, Lee K, Han SH. 2008.Lipoteichoic acid partially contributes to the inflammatory responses toEnterococcus faecalis. J Endod 34:975–982. https://doi.org/10.1016/j.joen.2008.05.005.

34. Wang S, Liu KUN, Seneviratne CJ, Li X, Cheung GSP, Jin L, Chu CH, ZhangC. 2015. Lipoteichoic acid from an Enterococcus faecalis clinical strainpromotes TNF-� expression through the NF-�b and p38 MAPK signalingpathways in differentiated thp-1 macrophages. Biomed Rep 3:697–702.https://doi.org/10.3892/br.2015.495.

35. Jontell M, Gunraj MN, Bergenholtz G. 1987. Immunocompetent cells inthe normal dental pulp. J Dent Res 66:1149 –1153. https://doi.org/10.1177/00220345870660061101.

36. Okiji T, Jontell M, Belichenko P, Dahlgren U, Bergenholtz G, Dahlstrom A.1997. Structural and functional association between substance p- andcalcitonin gene-related peptide-immunoreactive nerves and accessorycells in the rat dental pulp. J Dent Res 76:1818 –1824. https://doi.org/10.1177/00220345970760120301.

37. Jontell M, Okiji T, Dahlgren U, Bergenholtz G. 1998. Immune defensemechanisms of the dental pulp. Crit Rev Oral Biol Med 9:179 –200.https://doi.org/10.1177/10454411980090020301.

38. Chivatxaranukul P, Dashper SG, Messer HH. 2008. Dentinal tubule inva-sion and adherence by Enterococcus faecalis. Int Endod J 41:873– 882.https://doi.org/10.1111/j.1365-2591.2008.01445.x.

39. Veerayutthwilai O, Byers MR, Pham TT, Darveau RP, Dale BA. 2007. Differ-ential regulation of immune responses by odontoblasts. Oral MicrobiolImmunol 22:5–13. https://doi.org/10.1111/j.1399-302X.2007.00310.x.

40. Durand SH, Flacher V, Romeas A, Carrouel F, Colomb E, Vincent C,Magloire H, Couble ML, Bleicher F, Staquet MJ, Lebecque S, Farges JC.2006. Lipoteichoic acid increases TLR and functional chemokine ex-pression while reducing dentin formation in in vitro differentiatedhuman odontoblasts. J Immunol 176:2880 –2887. https://doi.org/10.4049/jimmunol.176.5.2880.

41. Jiang HW, Zhang W, Ren BP, Zeng JF, Ling JQ. 2006. Expression of Tolllike receptor 4 in normal human odontoblasts and dental pulp tissue. JEndod 32:747–751. https://doi.org/10.1016/j.joen.2006.01.010.

42. Paster BJ, Boches SK, Galvin JL, Ericson RE, Lau CN, Levanos VA, Sahas-rabudhe A, Dewhirst FE. 2001. Bacterial diversity in human subgingivalplaque. J Bacteriol 183:3770–3783. https://doi.org/10.1128/JB.183.12.3770-3783.2001.

43. Kumar S. 2012. Textbook of microbiology, 1st ed, p 251. Jaypee BrothersMedical Publishers, London, United Kingdom.

44. Bancroft JD, Gamble M. 2008. Microorganisms, p 312. In ●●● (ed), Theoryand practice of histological techniques, 8th ed. Churchill Livingstone,London, United Kingdom.

45. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, MuellerR, Nolan T, Pfaffl MW, Shipley GL, Vandesompele J, Wittwer CT. 2009. TheMIQE guidelines: minimum information for publication of quantitativereal-time PCR experiments. Clin Chem 55:611– 622. https://doi.org/10.1373/clinchem.2008.112797.

46. Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expression datausing real-time quantitative PCR and the 2(-delta delta C(T)) method.Methods 25:402– 408. https://doi.org/10.1006/meth.2001.1262.

47. Andersen CL, Jensen JL, Orntoft TF. 2004. Normalization of real-timequantitative reverse transcription-PCR data: a model-based varianceestimation approach to identify genes suited for normalization, appliedto bladder and colon cancer data sets. Cancer Res 64:5245–5250. https://doi.org/10.1158/0008-5472.CAN-04-0496.

48. Smith EL, Locke M, Waddington RJ, Sloan AJ. 2010. An ex vivo rodentmandible culture model for bone repair. Tissue Eng Part C Methods16:1287–1296. https://doi.org/10.1089/ten.tec.2009.0698.

E. faecalis Pulpal Pathogenicity and Attachment Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 13

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

49. Xing W, Deng M, Zhang J, Huang H, Dirsch O, Dahmen U. 2009. Quan-titative evaluation and selection of reference genes in a rat model ofextended liver resection. J Biomol Tech 20:109 –115.

50. Harrington J, Sloan AJ, Waddington RJ. 2014. Quantification of clonalheterogeneity of mesenchymal progenitor cells in dental pulp andbone marrow. Connect Tissue Res 55:62– 67. https://doi.org/10.3109/03008207.2014.923859.

51. Seol D, Choe H, Zheng H, Jang K, Ramakrishnan PS, Lim T-H, Martin JA.2011. Selection of reference genes for normalization of quantitativereal-time PCR in organ culture of the rat and rabbit intervertebral disc.BMC Res Notes 4:162. https://doi.org/10.1186/1756-0500-4-162.

52. Langnaese K, John R, Schweizer H, Ebmeyer U, Keilhoff G. 2008. Selection ofreference genes for quantitative real-time PCR in a rat asphyxial cardiacarrest model. BMC Mol Biol 9:53. https://doi.org/10.1186/1471-2199-9-53.

53. Lardizábal MN, Nocito AL, Daniele SM, Ornella LA, Palatnik JF, Veggi LM.2012. Reference genes for real-time PCR quantification of micrornas andmessenger RNAs in rat models of hepatotoxicity. PLoS One 7:e36323.https://doi.org/10.1371/journal.pone.0036323.

54. Qiang L, Lin HV, Kim-Muller JY, Welch CL, Gu W, Accili D. 2011.Proatherogenic abnormalities of lipid metabolism in SirT1 transgenicmice are mediated through CREB deacetylation. Cell Metab 14:758 –767.https://doi.org/10.1016/j.cmet.2011.10.007.

Nishio Ayre et al. Infection and Immunity

May 2018 Volume 86 Issue 5 e00871-17 iai.asm.org 14

on October 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from