For figures, tables and references we refer the reader to the original paper.
The composition of the oral microbiota is determined by a variety of synergistic interactions, such as food webs and intergeneric coaggregation, that facilitate their persistence in a dynamic environment.1 Likewise, interbacterial antagonism is an evolutionary mechanism that arms certain bacterial populations against elimination by other microorganisms. The summation of the antagonistic effects caused by so-called beneficial oral microbiota presents a substantial prevention of colonization by exogenous and opportunistic endogenous pathogens. Any disruption of this harmonic relationship between the host and commensal microorganisms is therefore considered an important factor for the development of oral pathologies, such as tooth decay and periodontal diseases.2,3 Some in vitro studies have already been performed characterizing potential probiotic bacteria in the oral cavity.4,5 Conversely, comparison of the prevalence of commensal species with beneficial properties for periodontal health between healthy and diseased individuals is only sparsely investigated.
Clear differences were observed between healthy and diseased individuals in the prevalence of bacterial strains that are antagonistic toward oral pathogens.6,7 Subgingival plaque samples of healthy patients contained organisms that could inhibit the growth of the—at that time—presumed periodontopathogens.6,7 Subgingival plaque samples from diseased sites from patients with aggressive periodontitis and refractory periodontitis almost invariably lacked such inhibitory bacteria. Interestingly, subgingival plaque samples from clinically healthy sites in these patients with periodontitis contained inhibitory bacteria in proportions similar to those in subgingival plaque taken from healthy control patients.6,7 This observation was also validated for other potential beneficial species, such as Lactobacillus gasseri,8 Bifidobacteria,9 and Streptococcus sanguinis.10
With the emergence of widespread antibiotic resistance, an alternative for conventional periodontal therapy is necessary. The importance of beneficial commensal bacteria has already been shown with regard to the immune response and periodontal colonization.11,12 Because the presence of these commensal bacteria in healthy compared with diseased individuals has only been investigated for a small study population and with a limited study design, no firm conclusions can be drawn. The aim of the current study is to investigate whether oral bacteria can cause antagonism toward periodontopathogens using a larger study population and comparing different assay techniques.
Additionally, the inhibitory effect of some commercial dietary probiotics on periodontopathogens was evaluated and compared with the inhibitory effect of orally derived beneficial bacteria toward a panel of periodontopathogens.
Bacterial Strains and Culturing Conditions
The periodontopathogens Aggregatibacter actinomycetemcomitans,‡ Porphyromonas gingivalis,§ Prevotella intermedia,‖ and Fusobacterium nucleatum¶ were maintained on blood agar,# supplemented with hemin (5 mg/mL), menadione (1 mg/mL), and 5% sterile horse blood.** The probiotic Lactobacillus strains Lactobacillus rhamnosus,††‡‡ Lactobacillus casei,§§‖‖¶¶ Lactobacillus fermentum,## and Lactobacillus paracasei L07-21*** were maintained on agar plates.††† All incubations of broth cultures‡‡‡ were performed in anaerobic conditions (80% N2, 10% H2, and 10% CO2) at 37°C unless stated otherwise.
Participant Selection and Sample Collection
Subgingival plaque samples were obtained from 35 patients attending the Department of Periodontology of the University Hospital Leuven. The project received full approval from the Ethical Committee of the University Hospital of the Catholic University of Leuven, Leuven, Belgium. Patients were well informed of the study protocol and objectives and gave their written consent before participation. The patients were included if they fulfilled the following inclusion and exclusion criteria: 1) ≥35 years of age; 2) ≥14 teeth with ≥3 teeth in every quadrant; 3) no dental implants or prosthetic devices of any kind present in the mouth; 4) no medication usage (including antibiotics); 5) no antiseptic mouthrinse usage or professional dental cleaning within 6 months before sample taking; 6) no systemic diseases in the past and no common diseases (e.g., flu or cold) within 6 months before sampling; 7) no oral lesions or signs of necrotizing oral diseases at the time of examination; 8) no pregnancy; 9) no smoking; 10) no periodontal treatment (root planing) in the past; and 11) a Turesky plaque index > 1. This population was divided into a group of patients with periodontitis (later referred to as “periodontitis”) and a group of individuals that did not have clinical symptoms of periodontitis (later referred to as “healthy”). Periodontal health or disease was evaluated by clinical examination. Patients with periodontitis exhibited ≥8 teeth with a probing depth (PD) and clinical attachment level (CAL) ≥ 6 mm with ≥1 of these teeth in each quadrant. Patients without periodontitis (healthy) exhibited no teeth with a PD >3 mm and a CAL >4 mm. Clinical examination and sample taking were performed by one single calibrated examiner (WT). Two weeks after baseline examination, plaque samples were taken with eight sterile paper points§§§ in a gingival sulcus (healthy volunteers) or from the apex of the deepest periodontal pocket (patients with periodontitis) in each quadrant after having removed all supragingival plaque with a sterile curet and having isolated and dried the sampling area with cotton rolls and airflow. In each individual, the four quadrants were sampled, and the paper points were pooled in 1.5 mL of reduced transport fluid until processing. Processing was initiated within a 2-hour timeframe.
Bacterial Isolation Procedure
Plaque samples were dispersed by vortexing and used for anaerobic culturing on blood agar plates as described previously.13 After 7 days of incubation, 50 random colonies of each patient were checked for their ability to produce growth inhibition of the periodontal pathogens P. gingivalis, P. intermedia, F. nucleatum, and A. actinomycetemcomitans. For this purpose, the agar overlay technique was used as described previously.14 Plates were incubated at 37°C in an anaerobic atmosphere. Test strains that exhibited growth inhibition of the indicator strain were selected and retrieved from the backup culture plate and streaked on blood agar plates until pure cultures were obtained. Strains were stored at −80°C until further use.
Quantification of Inhibition Mechanisms of Oral Isolates and Probiotic Strains
The quantification of the inhibitory mechanism was performed with the agar well diffusion method.15 Next to the test strains, broth cultures of the test strains were filter sterilized by passing them through 0.2-μm syringe filters‖‖‖ to obtain a cell-free supernatant. A fraction of the cell-free supernatant was, before filtration, neutralized with 1N NaOH to pH 7.0. A blank was also included in each assay, which was 50 μL sterile broth.¶¶¶ After application of the four different fractions of the test culture to the respective blood agar plates, the plates were incubated anaerobically at 37°C. The zones of inhibition (the distance between the edge of the well and the edge of the lawn of the indicator strains) were measured when the indicator bacteria formed a confluent lawn on the plates. Triplicate experiments were performed on separate days.
Identification of Oral Isolates
The 15 oral isolates that on average demonstrated the most pronounced inhibition of the indicator strains were identified with a combination of phenotypical and molecular tests. The phenotypical identification included Gram staining, colony morphology, and biochemical tests.### Molecular identification was performed on purified DNA that was extracted**** from liquid†††† overnight cultures of the isolates. According to the results of the phenotypic identification tests, a procedure was selected for 16S ribosomal RNA (rRNA) gene sequence analysis.
Amplification of Streptococcus 16S rRNA gene sequences.
The majority of the variation between 16S rRNA gene sequences of streptococcal species is found in the 5′ terminal part of the gene, especially in the v1 and v2 regions corresponding to bases 63 to 92 and 193 to 222 of the Streptococcus pneumoniae R6 16S rRNA gene (GenBank accession no. NC_003098).16 This region was amplified by polymerase chain reaction (PCR) using primers Streptococcus v1 forward (5′-AGTTTGATCCTGGCTCAGGACG-3′ and Streptococcus v2 reverse (5′-CAACTAGCTAATACAACGCAGGTC-3′).17
Amplification of coryneform 16S rRNA gene sequences.
The 16S rRNA gene sequence of the coryneform bacteria was amplified using primers TPU-1 (5-AGAGTTTGATCMTGGCTCAG-3′) and 1492RPL (5-GGTTACCTTGTTACGACTT-3′).18 PCRs (25 μL) contained 0.4 μM each primer,‡‡‡‡ 0.025 U/μL DNA polymerase,§§§§ 1× PCR buffer, 1.5 mM MgCl2, 0.2 mM concentrations of deoxynucleoside triphosphates, and 2.5 μL DNA. Thermal cycling consisted of an initial denaturation at 95°C for 2 minutes, followed by 35 cycles of 45 seconds at 95°C, 30 seconds at 50°C, and 2 minutes at 72°C. A final elongation step of 10 minutes at 72°C was included. The presence of the PCR product was verified on a 1% agarose gel.
Cloning and sequencing of the amplified 16S sequences.
Cloning was performed as described previously.19 The insert was cycle sequenced‖‖‖‖ according to the instructions of the manufacturer. Sequencing primers T7 (5′- TAATACGACTCACTATAGGG-3′) and SP6 (5′-TATTTAGGTGACACTATAG-3′) flanking the multiple cloning site of the vector produced sequencing reads in two directions. After purification¶¶¶¶ of the sequencing product, the samples were loaded on a capillary electrophoresis system. The obtained sequence reads were verified and assembled.#### The consensus sequences were used for homology search with the basic local alignment search tool (BLAST) and for comparison to the Ribosomal Database Project.20 Strains were identified to species level if the 16S rRNA gene sequence of the individual strain was ≥97.0% homologous to the type strain of a species20 and when phenotypic testing did not indicate any aberrant reactions regarding the published data for this particular species.
The difference in the prevalence of the inhibiting strains among healthy and diseased individuals was evaluated by performing a χ2 test. To compare the extent of the growth inhibition caused by the isolated strains, a mixed linear model was built. The growth inhibition of the different isolated strains could be compared for each periodontopathogen. Also an evaluation between the extent of growth inhibition between different periodontopathogens was made for the isolated strain. The agreement between the results obtained with the agar overlay technique and the agar well diffusion assay was evaluated with a Fisher exact test. All statistical analyses for this study were performed using software***** with the level of significance set at P <0.05.
Isolation of Antagonistic Strains From Subgingival Plaque Samples: Agar Overlay Assay
Subgingival plaque samples from 16 healthy volunteers and 19 patients with periodontitis were collected. Patient characteristics are shown in Table 1. In total, 1,750 individual colonies were screened with the agar overlay assay for their ability to induce growth inhibition of A. actinomycetemcomitans, P. gingivalis, P. intermedia, and F. nucleatum. With this technique, inhibitory colonies could be isolated from nine of the 35 sampled individuals with no significant difference (P >0.05) between healthy patients and patients with periodontitis (Table 1). From these, 74 inhibitory colonies could be isolated, which represent 4.23% of the total number of isolated colonies.
P. intermedia was the most frequently inhibited periodontopathogen, followed by P. gingivalis, A. actinomycetemcomitans, and F. nucleatum. Of the 39 colonies that inhibited P. intermedia, only 15 were from healthy volunteers. In contrast, of the 23 colonies that inhibited P. gingivalis, 20 were from healthy volunteers. The number of colonies that inhibited either P. intermedia or P. gingivalis was significantly different between healthy volunteers and patients with periodontitis (P <0.05). Although tendencies were observed, no significant differences (P >0.05) were found in the number of colonies that inhibited A. actinomycetemcomitans or F. nucleatum between patients with periodontitis and healthy individuals.
Agar Well Diffusion Assay: Spectrum of Antagonistic Activity
Although the agar well diffusion assay (Fig. 1) was used to quantify the amount of growth inhibition on each periodontopathogen for the 74 strains that were isolated with the agar overlay assay (see below), this assay also provided data on the antagonistic spectrum of these strains, albeit analyzed with a different technique. Of the 74 isolates that showed antagonistic activity toward ≥1 periodontopathogen via the agar overlay assay, only 48 isolates also showed antagonistic activity toward ≥1 periodontopathogen via the agar well diffusion assay. These 48 strains were derived in equal proportions from healthy individuals and patients with periodontitis (Table 1).
The agar well diffusion assay showed that P. gingivalis was the most frequently inhibited periodontopathogen, followed by F. nucleatum, P. intermedia, and A. actinomycetemcomitans. No significant differences (P >0.05) were observed in the number of colonies that inhibited P. gingivalis, P. intermedia, A. actinomycetemcomitans, or F. nucleatum between patients with periodontitis and healthy individuals. When looking at the seven commercially available probiotic bacteria (Fig. 2), they all inhibited the growth of P. gingivalis and P. intermedia, whereas A. actinomycetemcomitans was only inhibited by five strains and F. nucleatum only by two strains.
Agar Well Diffusion Assay: Quantification of Antagonistic Activity
Using the agar well diffusion assay (Fig. 3), the amount of growth inhibition toward each periodontopathogen was evaluated for the 74 isolates. With this assay, 21 isolates demonstrated an inhibition zone of ≥2 mm toward P. gingivalis. For the periodontopathogens F. nucleatum and A. actinomycetemcomitans, six oral isolates demonstrated an inhibition zone of ≥2 mm. Inhibition of the growth of P. intermedia by >2 mm could not be recorded with this assay. As shown in Table 2, P. gingivalis was quantitatively most clearly inhibited with an average inhibition zone for all 74 oral isolates of 1.15 ± 0.73 mm. The average inhibition zones for A. actinomycetemcomitans, F. nucleatum, and P. intermedia were 0.39 ± 0.41, 0.38 ± 0.36, and 0.30 ± 0.24 mm, respectively. No significant differences (P >0.05) were found for oral isolates originating from healthy patients versus patients with periodontitis.
For the commercial probiotic bacteria, P. gingivalis (average zone of inhibition, 4.07 ± 0.84 mm) and P. intermedia (average zone of inhibition, 1.71 ± 0.39 mm) were much more inhibited in their growth than A. actinomycetemcomitans (average zone of inhibition, 0.47 ± 0.24 mm) and F. nucleatum (average zone of inhibition, 0.14 ± 0.15 mm). Neutralizing the pH of the cell-free supernatant with NaOH completely eliminated the inhibition of periodontopathogenic growth.
If these results are compared with the average antagonistic activity of the isolated strains, the following observations are made: the probiotic strains cause 1) a 3.54 times stronger inhibition of P. gingivalis, 2) a 5.70 times stronger inhibition of P. intermedia, and 3) a 1.21 times stronger inhibition of A. actinomycetemcomitans. The inverse was observed for F. nucleatum, in which it was observed that the isolated strains on average caused a 2.71 times stronger inhibition compared with the probiotic strains.
The 15 isolated strains that on average caused the most pronounced inhibition were identified as follows: 1) one strain of Streptococcus salivarius; 2) three strains of Streptococcus mitis; 3) two strains of Streptococcus oralis; 4) three strains of Bifidobacterium dentium; 5) one strain of Actinomyces viscosus; and 6) five strains of Actinomyces naeslundii (Table 3). The growth inhibition on the four periodontopathogens for these strains was only present if the complete overnight cultures were used (Fig. 3). No growth inhibition could be observed in the negative control, with cell-free supernatants or with the pH neutralized, cell-free supernatants.
If the results from the seven commercially available probiotic bacteria are compared with the results from the 15 strains just mentioned, the differences between the groups were much less pronounced (Figs. 2 and 3). However, the commercially available probiotics still showed a stronger inhibition of P. gingivalis and P. intermedia, and the oral isolated strains showed a clearly stronger inhibition of F. nucleatum and A. actinomycetemcomitans.
Certain indigenous bacterial species and their products are beneficial for a healthy periodontium.21 When the homeostasis within the oral microbiota is disrupted, diseases such as dental caries and periodontitis can develop. These indigenous bacterial species are equipped with a number of specific or non-specific mechanisms to antagonize pathogenic species. Mechanisms, such as competition for and exclusion of adhesion receptors, competition for nutrients, and production of antimicrobials or surfactants, account for the protective effects of beneficial bacteria.11,22 Hillman et al.6,7 demonstrated that the prevalence of species capable of antagonizing oral pathogens is higher in periodontally healthy individuals. Especially antagonistic strains that inhibited the growth of P. intermedia were more prevalent in healthy individuals than in those with periodontitis (0% versus 37%, respectively). Similar observations were made for other periodontopathogens by this group of researchers. In the current study, the study of Hillman et al.6,7 is repeated, although a substantially larger study population was sampled and screened for the prevalence of antagonistic strains by means of the agar overlay assay. P. gingivalis, A. actinomycetemcomitans, F. nucleatum, and P. intermedia were selected as the model periodontopathogens, representing key pathogens from the red and orange complexes.23
When comparing the results of both studies, some similarities were observed. Although not always statistically significant in the present study, the prevalence of strains antagonistic toward P. gingivalis, A. actinomycetemcomitans, and F. nucleatum was found to be higher in healthy individuals than in individuals with periodontitis. This suggests the involvement of the beneficial microbiota in the protection of oral health. An exception to this finding is that, in the current study, the prevalence of strains that inhibit P. intermedia is lower in the healthy population (2.67% in the diseased population compared with 1.76% in the healthy population). At this time, to the best of our knowledge, a conclusive explanation for this finding is unavailable. However, because the study included 35 patients, a chance factor can still be involved.
The substantially higher concentration of antagonizing strains by Hillman et al.7 compared with our observations brings into focus the importance of the study methodology used. The observed discrepancy with our results could at least partially be explained by the fact that the incubations in the study of Hillman et al. were all performed in a microaerophilic atmosphere. This is in contrast to our methodology, in which all incubations were performed in complete absence of oxygen.7 In anaerobiosis, it has been shown that there is a strong downregulation of peroxidogenesis by oral streptococci.24,25 Eliminating the effect of hydrogen peroxide from the antagonistic interactions between beneficial and pathogenic oral species could (partially) account for the substantial difference between the detection rates of antagonizing strains in this study compared with the former study. However, it has to be stressed that the study of Hillman et al.7 sampled only one healthy individual.
The 74 isolates that inhibited growth of the periodontopathogens in the agar overlay assay were subsequently investigated with the agar well diffusion assay. The comparison between the two antagonism assays generally yielded a poor agreement for these 74 strains. This poor agreement between the two tests shows that, for the evaluation of microbial antagonisms among dental plaque bacteria, study results obtained by different techniques are incomparable. A factor that can be involved in the discrepancy between the two assays is the fact that the sequence of inoculation of test and indicator strains differs in the two assays: in the agar overlay assay, the test bacteria are allowed to grow to a stationary phase before being exposed to the pathogen as indicator strain. Conversely, in the agar well diffusion assay, both test and indicator strains are inoculated simultaneously on the assay culture plate. This aspect has been shown to considerably influence antagonistic interactions among oral bacteria in vitro.26 It seems most likely that the earlier inoculated strain has an evident advantage toward the later introduced strain because of an abundance of nutrients, although the authors speculate that interspecies communications are involved in this interaction. This is analogous with our observations because the total number of colonies that were able to antagonize ≥1 pathogen was found to be lower in the agar well diffusion assay than in the agar overlay assay (Table 1).
Our results suggest that the mechanism by which the isolated strains suppress growth of the periodontopathogens was caused by the competition for nutrients: although the overnight cultures of most of these strains induced growth inhibition, no detectable inhibitory substance was present in the supernatants of any of the isolates. The 15 strains that on average caused the most explicit growth inhibition were identified as typical representatives of the oral microbiota. All of these strains are species of viridans streptococci, Actinomyces, or Bifidobacteria. Although competition for nutrients seems the most probable antagonistic mechanism observed in our experiments, these genera have also been shown to be capable of causing broad-spectrum antagonism toward other strains by the production of organic acids, such as formic, lactic, and succinic acids.22,27-29 Lactic acid is also known to function as a permeabilizer of the Gram-negative bacterial membrane30 and even to have some iron-chelating capacity.31 Although in most interbacterial relationships the concomitant lowering of the environment pH32 is considered to be of negligible importance compared with the direct cellular effect of the organic acids, this may not be so in respect to the periodontopathogens that prefer basic environments. This is especially true for P. gingivalis, which stops growing if the environmental pH drops below 6.5.33 Although this is a suitable explanation for the fact that P. gingivalis is the most strongly inhibited strain among the four tested periodontopathogens, this clearly only partially accounts for the antagonistic activity because the amount of growth inhibition in the wells containing cell-free supernatants was below the detection limit.
A wide range of bacterial species have been shown to produce bacteriocins, narrow-spectrum antimicrobial substances that are produced to withstand competition by similar strains. The absence of such bacteriocin-producing strains in our collection of 1,750 examined strains is in contrast to a number of publications that describe the discovery of bacteriocin-producing Gram-negative oral bacteria. Again, this discrepancy is most likely caused by a different study methodology because, in some studies, the antimicrobial compounds can only be isolated from the concentrated intracellular fraction of bacterial cultures.34-36 The physiologic function of an intracellular antimicrobial substance is at least peculiar, and the poor characterization of the antagonistic compounds cannot exclude a biologic function different from bacteriocin.35
The microbial antagonism of probiotic Lactobacillus strains toward oral pathogens has been described recently.8 The authors describe a slightly different antagonistic spectrum compared with that of the probiotic strains that are used in this study. Although the four tested pathogens were all susceptible to antagonisms of the Lactobacillus strains, in our observation, P. gingivalis is substantially more inhibited than A. actinomycetemcomitans, P. intermedia, and F. nucleatum. Kõll-Klais et al.,8 however, found that A. actinomycetemcomitans was most pronouncedly inhibited. Because in the latter study two different assays were used for the evaluation of the antagonisms of A. actinomycetemcomitans and P. gingivalis, respectively, the comparison between antagonistic potentials may not be appropriate.
Our data suggest that the growth inhibition caused by the lactic acid bacteria toward the pathogens was mainly caused by the production of large amounts of organic acids because the neutralization of supernatants pH completely eliminated the antagonistic activity. This effect is in line with observations in other reports in which it has been shown that Lactobacillus strains antagonize many Gram-negative bacteria, mostly by the antimicrobial effect of lactic acid.37
In a study by Stamatova et al.,38 the adhesion of Lactobacillus strains to saliva-coated surfaces was evaluated in vitro. The results showed significant variations in the adhesion capacity of the studied Lactobacillus strains. Adhesion to oral surfaces is of primary importance for colonization in the oral cavity, and, in that aspect, the use of the commensal oral microbiota as beneficial strains circumvents this problem.38
In general, it can be concluded from the current study that the investigation of microbial antagonisms among dental plaque bacteria can be performed using different approaches. However, different methodologies yield results that are difficult to compare. Results suggested that some oral bacteria can cause antagonism toward periodontopathogens by means of competition for nutrients. The effects described here, however, probably only represent part of the entire beneficial effect of such strains on oral health: it is important to mention the involvement of hydrogen peroxide as an antimicrobial substance that was not included in our tests. It is generally accepted that hydrogen peroxide is produced by viridans streptococci25 and Lactobacillus species,39 and it is therefore likely that this augments their antimicrobial effect in vivo.
The commensal oral microbiota is considered to induce a beneficial oral immune response11or to interfere with periodontopathogen colonization.40 The interference with pathogen recolonization was confirmed in vivo as an addition to the standard therapy of periodontitis.12 These observations underline the therapeutic and prophylactic potential of applications that stimulate oral health by the application of beneficial effector strains.
This study was supported by Catholic University of Leuven Grant OT 07/057 and Research Foundation Flanders Grants G.0772.09 and 220.127.116.11. Dr. Teughels was also supported as a postdoctoral fellow of the Research Foundation Flanders. The authors report no conflicts of interest related to this study.