|Home | About | Journals | Submit | Contact Us | Français|
Fluoride is the mainstay of dental caries prevention, and yet current applications offer incomplete protection and may not effectively address the infectious character of the disease. Therefore, we evaluated the effectiveness of a novel combination therapy (CT; 2 mM myricetin, 4 mM tt-farnesol, 250 ppm of fluoride) that supplements fluoride with naturally occurring, food-derived, antibiofilm compounds. Treatment regimens simulating those experienced clinically (twice daily for ≤60 s) were used both in vitro over a saliva-coated hydroxyapatite biofilm model and in vivo with a rodent model of dental caries. The effectiveness of CT was evaluated based on the incidence and severity of carious lesions (compared to fluoride or vehicle control). We found that CT was superior to fluoride (positive control, P < 0.05); topical applications dramatically reduced caries development in Sprague-Dawley rats, all without altering the Streptococcus mutans or total populations within the plaque. We subsequently identified the underlying mechanisms through which applications of CT modulate biofilm virulence. CT targets expression of key Streptococcus mutans genes during biofilm formation in vitro and in vivo. These are associated with exopolysaccharide matrix synthesis (gtfB) and the ability to tolerate exogenous stress (e.g., sloA), which are essential for cariogenic biofilm assembly. We also identified a unique gene (SMU.940) that was severely repressed and may represent a potentially novel target; its inactivation disrupted exopolysaccharide accumulation and matrix development. Altogether, CT may be clinically more effective than current anticaries modalities, targeting expression of bacterial virulence associated with pathogenesis of the disease. These observations may have relevance for development of enhanced therapies against other biofilm-dependent infections.
Many diseases in humans are caused by and/or exacerbated by biofilm formation, including those occurring in the mouth (e.g., dental caries) (9). A distinctive feature of biofilm-related illnesses is that upon biofilm establishment, the resident microorganisms frequently become recalcitrant to antimicrobial therapies, making them difficult to remove, especially without disturbing the normal flora (6, 9, 17).
Dental caries represents one of the most prevalent and costly biofilm-dependent diseases worldwide, which afflicts children and adults alike (14, 41). This ubiquitous disease results from the complex interactions that occur within the oral cavity among virulent microorganisms, their products, host salivary constituents, and dietary carbohydrates (2, 41, 45). These interactions, in addition to other factors (e.g., frequency of sugar intake and poor oral hygiene), help to modulate the transition from a healthy caries-free condition to a disease state. The establishment of pathogenic plaque biofilms eventually causes the decay of susceptible tooth surfaces (2, 41, 45). Novel therapies that compromise the ability of virulent species to assemble and maintain biofilms on tooth surfaces may represent potentially effective alternatives to the use of broad-spectrum microbicides (e.g., chlorhexidine), which can indiscriminately eradicate species present in the mouth, even those beneficial to oral health (21, 26, 43).
In the mouth the (frequent) consumption of sucrose can serve as the catalyst for caries development, since it is a substrate for both the production of acid (through fermentation) and exopolysaccharides (EPS). EPS are key structural and protective matrix components of virulent dental biofilms that act as a supportive framework and barrier to diffusion (2, 41, 45, 63). Streptococcus mutans is considered one of the prime etiologic agents of dental caries, although additional organisms may be involved (2, 41, 45, 56). This bacterium possesses multiple exoenzymes (e.g., glucosyltransferases [Gtfs]) that make it a chief producer of EPS, while it is also both highly acidogenic and aciduric (2, 45). EPS synthesis makes S. mutans a formidable opponent to oral health, since the glucans produced by secreted Gtfs that are incorporated into the pellicle provide bacterial binding sites in situ to promote tight adherence and microbial accumulation on tooth surfaces (2, 58, 61). Gtfs also bind many oral bacteria, such that glucans are formed on the surfaces of S. mutans and other oral microorganisms, which progressively accumulate to enmesh these microorganisms in the matrix (2, 63, 64). This ultimately creates a highly adherent and cohesive biofilm that shelters resident organisms from antimicrobials and other inimical influences (2, 63, 64). The metabolic activity of acidogenic bacteria embedded in the EPS-rich and diffusion-limited matrix leads to acidification of the milieu, allowing only acid-tolerant bacteria to prosper, which in turns ensures continued localized acid production (2, 41, 45). The accumulation of acids eventually leads to the dissolution of the adjacent tooth enamel (expressed clinically as carious lesions or cavities) (2, 41). Therefore, EPS-mediated biofilm construction offers great advantages to the establishment, survival, and persistence of virulent organisms within the oral cavity.
Fluoride is regarded as the gold standard among anticaries agents (8). However, its widespread use, although effective, does not offer complete protection against the disease (8). Fluoride exerts its major effect by reducing demineralization and enhancing remineralization of early carious lesions, and yet it does have other biological activities (57). Fluoride's use alone may not adequately address the infectious character of the disease, although studies indicate that fluoride may have some effects on bacterial metabolism (e.g., inhibition of enolase and acidification of the cytoplasm) and gtf expression (33, 40, 55, 59). Chlorhexidine is a potent broad-spectrum microbicidal agent that suppresses mutans streptococci levels in saliva. However, it is not compatible with fluoride, has adverse side effects (tooth staining and calculus formation) and is not suitable for daily preventive or therapeutic use (4, 26). Therefore, inclusion of other bioactive agents that affect virulence and/or the ability of cariogenic bacteria to form biofilms may offer an attractive and possibly superior anticaries therapy than fluoride alone. Unfortunately, most of the compounds tested thus far are broad-spectrum antimicrobials (4, 26). Thus, we sought to supplement the activities of fluoride to specifically target the virulence of S. mutans through the inclusion of naturally occurring antibiofilm molecules.
The use of natural products in the treatment and prevention of dental caries is an attractive approach, especially considering that natural products may possess a wide array of activities and functions and have a rich history of use in traditional medicine (21). In the present study, we tested a novel nonmicrobicidal approach, using selected naturally occurring food-derived agents (myricetin and tt-farnesol) that inhibit the ability of S. mutans to assemble biofilms, all without altering cell viability (19, 21). Myricetin is an effective inhibitor of Gtf activity and also reduces expression of the gtfBC gene cluster, whereas tt-farnesol disrupts acid production, acid tolerance, and polysaccharide synthesis by affecting the permeability of the S. mutans membrane (19, 20, 26, 28, 29, 31, 32). Such properties may complement the cariostatic properties of fluoride by augmenting the overall biological activity against cariogenic biofilm formation while keeping its proven physicochemical effects (19, 21).
We previously demonstrated that the combination of natural agents with fluoride is effective in disrupting S. mutans biofilm formation by reducing the amount of insoluble glucan and the overall acidogenicity of these biofilms (19). Preliminary studies have shown that the combination of agents is more effective in reducing biofilm formation than any single compound or either of these natural agents alone with fluoride. However, further analyses were required to evaluate the potential effectiveness of this combination in vivo and to elucidate the underlying mechanisms of action using a clinically relevant treatment regimen (brief topical exposures). If effective, this therapy could bridge the gap between the current chemical modalities (e.g., fluoride and chlorhexidine) used to prevent or treat biofilm-dependent oral diseases. Using molecular techniques in tandem with an in vivo model of dental caries, we demonstrate that our combination therapy is highly effective in vivo, since it targets specific virulence traits of S. mutans associated with the pathogenesis of dental caries disease. At the same time, we elucidate the mechanisms of action, which could be used to further improve this chemotherapy to better disrupt the therapeutic targets identified in the present study.
Animal experiments were performed as described previously (3). Eighteen litters of eight female Sprague-Dawley rats aged 15 days were purchased with their dams from Harlan Laboratories (Madison, WI) and screened for infection with S. mutans. Any animals infected with S. mutans prior to inoculation were removed from the study. Then, the dams, while still nursing, were infected by mouth using an actively growing culture of S. mutans UA159 (3), which is transmitted to the pups. Pups were also directly infected with S. mutans and, at weaning, pups aged 21 days were checked for the relative level of S. mutans infection. Infected pups were then placed into three groups (animals from each litter were selected for each group at random), and their teeth were treated topically for 30 s using a camel hair brush twice daily using, as follows, (i) combination therapy (CT; 2 mM myricetin, 4 mM tt-farnesol, 250 ppm of fluoride), (ii) fluoride therapy (F; 250 ppm fluoride), and (iii) vehicle control (V; 25% ethanol in 2.5 mM phosphate buffer [pH 6.5]). The combination and concentrations of the natural agents were selected based on previously published (29, 30) and unpublished dose-response studies and solubility in the vehicle system. Each group was provided the National Institutes of Health diet 2000 (22) and 5% sucrose water ad libitum. The experiment proceeded for 3 weeks (19 animals/group) and 5 weeks (15 animals/group); all animals were weighed weekly, and their physical appearance was noted daily. At the end of the experimental period, the animals were sacrificed on the day following the final day of treatment. The jaws were aseptically dissected and processed for RNA extraction (see below) and microbiological analysis of the animal's plaque as detailed in Klein et al. (25). For microbiological analysis, the left jaws were sonicated in 5 ml of 154 mM sterile NaCl solution for plaque removal. The suspensions obtained were serially diluted and plated on mitis salivarius agar plus bacitracin to estimate the S. mutans population and on blood agar to determine the total cultivable aerotolerant flora in the plaque (3). All of the jaws were defleshed, and the teeth were prepared for caries scoring according to Larson's modification of Keyes' system (35). This study was reviewed and approved by the University of Rochester Committee on Animal Resources (UCAR 2007-128).
S. mutans strain UA159 biofilms were grown on saliva-coated hydroxyapatite discs (surface area, 2.7 ± 0.2 cm2; Clarkson Chromatography Products, Inc., South Williamsport, PA) vertically suspended in 24-well plates using a custom-made wire disc holder as detailed previously (28). Each disc was inoculated with approximately 2 × 106 CFU of S. mutans/ml in ultrafiltered (10-kDa cutoff; Millipore, Billerica, MA) yeast tryptone extract broth containing 1% (30 mM) sucrose at 37°C and 5% CO2. During the first 19 h, the organisms were grown undisturbed to allow initial biofilm formation; the biofilms (19 h old) were then treated twice daily (at 10 a.m. and 4 p.m.) up to 76 h of the experimental period with the CT, F, and V treatments. The biofilms were exposed to the treatments for 60 s, dip-washed twice in sterile saline solution (to remove excess agents or vehicle control), and then transferred to fresh culture medium. The biofilms were removed at 46 h for microarray and 46, 52, and 76 h for reverse transcription-quantitative PCR (RT-qPCR) studies.
RNA was extracted and purified using protocols optimized for biofilms formed in vitro (10) and in vivo (25). Briefly, disc sets and rodent jaws were incubated in RNALater (Applied Biosystems/Ambion, Austin, TX), and the biofilm material was removed from the discs and teeth. Acid phenol-chloroform extractions were then performed, as described earlier (12). The RNAs were purified and DNase treated on the column using the Qiagen RNeasy micro kit (Qiagen, Valencia, CA). The RNAs were then subjected to a second DNase I treatment with Turbo DNase (Applied Biosystems/Ambion) and purified using an RNeasy MinElute cleanup kit (Qiagen). The RNAs were quantified using the NanoDrop ND1000 spectrophotometer (Thermo Scientific, Wilmington, DE). The RNA quality was evaluated using an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA), and all RNAs used to prepare cDNAs for microarray analysis were determined to have RNA integrity number of ≥9.0.
Whole-genome profiling was conducted using version 1 microarrays for S. mutans strain UA159 provided by the J. Craig Venter Institute (JCVI). A detailed description of these slides could be found elsewhere (http://pfgrc.jcvi.org). Reference RNAs were prepared as previously described (23), and cDNAs for both experimental and reference samples were synthesized following protocols provided by JCVI (http://pfgrc.jcvi.org/index.php/microarray/protocols.html). Experimental cDNAs were labeled with indocarbocyanine (Cy3)-dUTP, while reference cDNAs were labeled with indodicarbocyanine (Cy5)-dUTP (Amersham Biosciences, Piscataway, NJ). Hybridizations were performed using the MAUI hybridization system (BioMicro Systems, Salt Lake City, UT), and slides were then washed and scanned using a GenePix scanner (Axon Instruments, Inc., Union City, CA) according to JCVI protocols. After scanning, single-channel images were simultaneously uploaded into JCVI Spotfinder321 (http://www.tm4.org/spotfinder.html), creating an overlay image with both Cy5 and Cy3 channels. A Spotfinder grid file (http://pfgrc.jcvi.org/index.php/microarray/array_description/streptococcus_mutans/version1.html) was used and adjusted to remove empty spots from the analysis. Spotfinder 3.2.1. was used to identify spots and assign relative spot intensities in a .mev file for uploading into JCVI MIDAS (http://www.tm4.org/midas.html). MIDAS was used to perform a Lowess normalization and flag and remove spots with low-intensity values (<1,000). Statistical analysis (unpaired class comparison, P cutoff value of 0.001) was performed using BRB-Array Tools, which is available as freeware (http://linus.nci.nih.gov/BRB-ArrayTools.html). MDV (available from LANL Oralgen [http://www.oralgen.lanl.gov/]) was used to assign gene names and functional classes to identified genes, as described previously (23). The genes were then sorted in Microsoft Excel to identify those with an absolute fold change of ≥1.8. These genes were organized into functional categories using a set of defined gene of interest (GOI) selection criteria, which was used to evaluate the impact of treatment by identifying the major themes affected.
We attempted to devise standard criteria to select biologically meaningful and significant GOIs from S. mutans transcriptome data that may be associated with the pathogenesis of dental caries. We focused on genes associated with bacterial adhesion, extracellular polysaccharide (EPS) metabolism, biofilm formation, and fitness, including survival/adaptation mechanisms against environmental assaults. Therefore, the transcriptome data was sorted into the following functional categories: EPS (genes involved in biofilm matrix production), IPS (genes involved in intracellular polysaccharide storage), biofilm formation/adhesion (genes involved in biofilm formation that do not play a specific role in matrix production), the glycolytic pathway (genes with known functions in sugar metabolism), stress (genes with known functions in the stress response, including oxidative stress and acid tolerance response), regulators, or two component systems (TCS; genes with known regulatory functions), hypothetical (genes lacking an identified function), and other. Genes with unknown functions were grouped into the hypothetical and other (genes with functions not associated with the aforementioned functional categories) categories. Furthermore, we also took into account the potency of the transcriptional change by sorting genes according to their magnitude of fold change (treatment/vehicle). Thus, the criteria would select not only genes with relevant functions associated with the disease process but also those most dramatically affected (repressed or induced) by our treatment.
We performed RT-qPCR to validate and further characterize specific GOIs selected from microarray data analysis (32). Using methods described previously, the RNAs purified from both in vitro (32) and in vivo (25) samples were used as templates for cDNA synthesis. For a list of the primers used in this study, see Table S1 in the supplemental material. Samples in which no reverse transcriptase was added served as negative controls. cDNAs were amplified using a MyiQ real-time PCR detection system with iQ SYBR green Supermix (Bio-Rad). Standard curves were used to determine the relative number of cDNA molecules, which were normalized to the relative number of 16S rRNA cDNA molecules in each sample, as previously described (32). These values were used to determine the fold of change between each treated sample and the vehicle control.
The SMU.940 gene was functionally inactivated through allelic replacement with a nonpolar spectinomycin marker (spt). Briefly, a 1-kb fragment flanking the 5′ end of SMU.940 was amplified from strain UA159 using the primers NTsmu940-F (CTGGTGGACATGCTTATG) and NTsmu940-R (CTTAGTTTCAGCATGCGTGTTGGAC), and a 1-kb fragment flanking the 3′ end of SMU.940 was amplified using primers CTsmu940-F (CAATATATCGGCATGCTCTACTATATG) and CTsmu940-R (CCAGATTATGATTTAGCC). The underlined bases correspond to the SphI restriction sites that were included for cloning purposes. After amplification, the two PCR fragments were digested with SphI and ligated to a spectinomycin resistance marker that was obtained as an SphI fragment. The ligation mixture was used to transform S. mutans UA159 followed by plating onto BHI containing spectinomycin (1 mg ml−1). The deletion was confirmed by PCR sequencing of the insertion site and flanking regions. Based on the genetic organization, introduction of the Spt marker in SMU.940 has no polar effect. The gene immediately downstream of SMU.940 is transcribed in the opposite orientation and SMU.940 is the last (downstream) gene in its operon. There is one gene (SMU.941) upstream of SMU.940, although it is not known whether these genes are cotranscribed.
We examined the three-dimensional (3D) architecture and quantified the amount of EPS present within intact biofilms of S. mutans UA159 ΔSMU.940::spt and its parental wild type through the use of confocal fluorescence imaging (23, 24). EPS was labeled via incorporation of the Alexa Fluor 647 dye, while S. mutans cells were stained with SYTO 9 (23, 24). Imaging was performed using an Olympus FV 1000 two-photon laser scanning microscope (Olympus, Tokyo, Japan) equipped with a ×10 (0.45 numerical aperture) or ×25 LPlan N (1.05 numerical aperture) water immersion objective lens. The excitation wavelength was 810 nm, and the emission wavelength filter for SYTO 9 was a 495/540 OlyMPFC1 filter, while the filter for Alexa Fluor 647 was an HQ655/40M-2P filter (34). Each biofilm was scanned at five positions randomly selected at the microscope stage (64) for a total of 19 stacks per sample (ΔSMU.940::spt and parental wild type) from three independent experiments, and confocal image series were generated by optical sectioning at each of these positions. Amira 5.4.1 software (Visage Imaging, San Diego, CA) was used to create 3D renderings of each biofilm structural component (EPS and bacteria). COMSTAT and newly developed DUOSTAT (http://www.imageanalysis.dk) were used to calculate the biomass, average thickness, number and size of microcolonies, and the colocalization between the bacterial and EPS channels for each image.
For the in vitro studies, treatments were compared using regression models to obtain overall tests of equality and pairwise comparisons. The significance level was set at 5%, and no adjustments were made for multiple comparisons. For the animal study, an analysis of outcome measures was done with transformed values of the measures in order to stabilize variances, as detailed in Raubertas et al. (50); smooth-surface and sulcal caries scores were expressed as proportions of their maximum possible value (124 and 56, respectively), and the arcsine transformation was applied. The data were then subjected to analysis of variance (ANOVA) in the Tukey-Kramer honest standard deviation test for all pairs. The statistical software JMP (version 3.1; SAS Institute, Cary, NC) was used to perform the analyses. The level of significance was set at 5%.
We first assessed the impact of our combination therapy (CT) on caries development using a rodent model to evaluate its effectiveness as a potentially novel and useful anticaries treatment. We simulated the conditions that might be experienced clinically in humans by applying our agents topically twice daily for brief exposures (30 s). Using this model, we determined that the incidence and severity of smooth-surface lesions in CT-treated animals was reduced compared to the vehicle control after both 3 weeks (Table 1) and 5 weeks (Table 2) of infection. The extent of this reduction was significantly greater than that of treatment with fluoride (positive control; P < 0.05). CT treatment also reduced the development of sulcal surface caries, showing significant differences compared to the vehicle control and to some extent fluoride (e.g., at the Dx level in the 3-week caries study; see Table S2 in the supplemental material). In tandem with caries scoring, we also observed that CT did not affect the viability of the total flora or S. mutans populations obtained from plaque samples, as determined by CFU counting (Tables 1 and and2).2). These observations were congruent with a lack of biocidal activity against S. mutans biofilms in vitro (19).
After demonstrating that CT treatment was highly effective in reducing the incidence and severity of dental caries in our in vivo model, we sought to better characterize the mechanisms by which CT affects the molecular biology of S. mutans. We used microarrays and RT-qPCR to examine the impact of treatment on 46 h in vitro biofilms under a similar treatment regimen. RT-qPCR was necessary to assess the transcriptional profiles of gtfB, gtfC, and gtfD, as microarrays have a low confidence in detecting the mRNA levels of these genes, particularly if they are repressed (23). Using this approach, we identified 186 genes that were differentially regulated in response to treatment with either one or both therapies tested (CT or F; see Table S3 in the supplemental material). Roughly 8% of the genome was differentially regulated in response to treatment, demonstrating that CT induces a significant response in S. mutans.
The majority of targets were affected by both CT and F (162 genes). However, there was a subset of genes that was either only or more affected by CT. Of these genes, we identified two that were affected by CT, but not fluoride. Thus, the combination therapy does not appear to largely affect genes through a mechanism that is independent of the action of fluoride, as anticipated. However, many of the genes are uniquely affected by CT in that the level of induction or repression (fold change relative to vehicle control) is greater for CT versus fluoride alone (total of 53 genes).
To better evaluate the volume of data generated during S. mutans transcriptome studies, we developed a new classification system to organize genes into categories relevant to S. mutans biofilm formation, fitness, and virulence expression (refer to Materials and Methods). We then examined the distribution of genes either up- or downregulated by CT compared to the vehicle control (162 genes; see Fig. 1). This approach allowed us to reliably and readily identify the major molecular themes affected by treatment, which can be applied to downstream studies. Overall, categories with known functions that were most affected were regulators, genes involved in the glycolytic pathway, and genes involved in the stress response. However, we did not identify any genes previously implicated in biofilm formation/adhesion, other than EPS-related genes.
We selected specific targets for microarray validation and subsequent characterization based on their predicted functions (representatives from each functional category) and their magnitude of repression or induction. We chose genes that were either highly induced or repressed by CT, particularly those that appeared to be more greatly affected by CT than fluoride alone. Using RT-qPCR, we validated the expression profiles of 10 genes that we identified as differentially regulated via microarray (Fig. 2). Six of these genes (sloA, sodA, lrgA, SMU.940, gtfB, and copY) were highly induced or repressed. All six genes have predicted functions that could help to account for an attenuation in biofilm formation in vitro (19) and/or a reduction in the incidence and severity of caries after treatment (the present study).
To better understand the overall kinetics of expression in vitro and to in part assess the potential substantivity of short-term topical exposures to CT, we evaluated the transcriptional profiles of selected genes at additional biofilm developmental stages. We focused on several genes that were especially highly induced or repressed by CT, including the following: gtfB, sloA, sodA, copY, lrgA, and SMU.940 (see Table S1 in the supplemental material for a list of the primers used).
We tested expression of these genes at the middle (52 h) and later (76 h) stages of biofilm formation, while also varying the length of the treatment-free interval before RNA collection. We found that the expression patterns of these genes (Fig. 3) were similar to those observed in 46-h biofilm samples (early stage) tested via microarray and RT-qPCR (Fig. 2). SMU.940, copY, sloA, sodA, and gtfB were all severely repressed by the combination therapy at 52 and 76 h with either a 1-h or a 4-h posttreatment incubation (Fig. 3). There was no apparent difference in the level of repression for SMU.940, copY, sloA, or sodA when we varied the length of the posttreatment incubation (1 h versus 4 h). This demonstrates that brief topical applications of our agents can sustain their effects on the transcription of these genes hours after treatment. However, the kinetics of expression were not the same for gtfB, since the degree of repression after the 1 h of incubation was greater than after 4 h, indicating that the ability of CT to repress gtfB transcription may dissipate with time. Even then, gtfB was significantly repressed at all of the time points tested (Fig. 3). We also tested one upregulated gene, lrgA, the expression of which was equally induced at all time points tested. Altogether, these data demonstrate that CT is effective in altering the transcriptional profiles of several genes at various stages of biofilm development, and it sustains this activity despite using only brief/periodic applications of the agents.
We also evaluated the expression of selected genes from in vivo RNA samples collected during our animal study. To our knowledge, this is the first time that the transcriptional profiles of S. mutans have been examined in rodent plaque samples (10, 25). RNA yields from in vivo samples are typically lower than those obtained from in vitro samples. As a result, fewer targets may be selected for RT-qPCR analysis from in vivo samples. We selected two of the most severely CT-repressed targets for in vivo analysis, gtfB and sloA, which are directly involved in virulence (2, 46, 52). We found that gtfB and sloA were also strongly repressed in vivo in response to CT (Fig. 4), which agrees well with our in vitro studies. gtfB was significantly repressed by CT in the 3-week, but not in the 5-week study, while sloA was significantly repressed in the 5-week study but not in the 3-week study. This suggests that the timing of gtfB and sloA repression is distinctive.
In our initial microarray analysis, we identified SMU.940 as one of two genes that was highly and uniquely repressed by the combination therapy (see Table S2 in the supplemental material). SMU.940 is annotated as a putative hemolysin, although no studies have directly assessed its function in S. mutans. Therefore, to better understand its function and potential role in cariogenic biofilm formation, we constructed an SMU.940 deletion mutant (ΔSMU.940), which we then compared to the parental wild-type strain. Using confocal microscopy combined with quantitative computational analysis (COMSTAT and DUOSTAT), we determined that the ΔSMU.940 strain is significantly attenuated for biofilm formation (Fig. 5). In comparison to the parental wild type, the mutant strain has no apparent growth defect (data not shown), which indicates that the attenuation in biofilm formation is likely due to mutation of the SMU.940 locus, rather than a secondary effect.
Overall, the mutant forms a particularly defective biofilm displaying significantly less EPS and fewer/smaller microcolonies (see Table S4 in the supplemental material). The data show a dramatic disruption in the EPS content, with an ~3-fold reduction in the total EPS biomass in SMU.940 mutant biofilms. As a result, these biofilms have a reduced EPS/cell ratio (a >2-fold difference compared to UA159 biofilm), and a lower percentage of the cells was found colocalized with the EPS (44.4 ± 10.2 for ΔSMU.940 versus 74.7 ± 9.7 for the parental strain). Interestingly, the biomass of S. mutans cells within the biofilms are only slightly reduced compared to the parental strain (30.49 ± 6.49 versus 38.64 ± 8.46), confirming that the mutation was not detrimental to overall growth.
The ability to effectively disrupt pathways specific and essential to the biofilm lifestyle of bacterial pathogens, all without affecting the viability of the normal flora, is an attractive approach for the prevention and/or reduction of biofilm-related illnesses, especially those that occur in complex microenvironments, such as the human mouth (21). Likewise, our CT might represent an effective modality for the prevention of (oral) biofilm-related diseases in humans. Our design is in line with this emerging philosophy, which promotes the development of infectious disease therapies that target specific virulence properties, rather than causing general defects in growth or viability (5). The ability of the combination therapy to significantly reduce the incidence and severity of caries in rodents at both 3 and 5 weeks, often more effectively than fluoride, and without biocidal activity, is clinically relevant. Fluoride is currently the most effective anticaries agent, although it does not completely eliminate or prevent the disease in its present state (8). New strategies for oral care will likely not replace the use of fluoride but would rather seek to augment its effects (21). Thus, a combination approach may be most advantageous (19, 26, 43). CT of course has additional physicochemical effects, due to the presence of fluoride. However, our study has focused on identifying its molecular targets within S. mutans, which may ultimately repress the development of virulent-cariogenic biofilms in a specific manner.
There are definite challenges to the development of new anticaries therapies, since the concentration of oral therapeutic agents is rapidly reduced during swallowing and expectoration, such that effectiveness necessitates substantivity (21). To be amenable to a common clinical situation (21, 26), we chose to deliver the combination therapy topically for brief exposures twice daily. The combination therapy has potent anticaries effects when locally applied to rodent dentition, which indicates that the concentration of these agents is efficacious under a sucrose-rich, caries-promoting diet. Persistence of such an effect is highly desirable, considering that most topical applications in humans are typically sustained no more than a couple minutes at best during daily oral hygiene practices (21). Furthermore, the in vivo model more accurately simulates the conditions encountered in the mouth, including exposure to salivary and host cellular components in the presence of both hydrodynamic and abrasive forces (20, 27). In addition, the rat harbors a complex and mixed oral flora akin to the human mouth, even when it has been intraorally infected with S. mutans (20, 27).
To better understand how the combination therapy affects the virulence of S. mutans, we focused on identifying the biological and molecular processes affected by CT within biofilms, which may help to explain the reduction in cariogenicity in vivo. We subsequently determined that (i) CT disrupts specific genes generally associated with EPS synthesis and/or stress tolerance (e.g., gtfB, sloA, sodA, and copY), some of which are proven virulence factors in S. mutans (gtfB and sloA), (ii) CT has sustained effects on the expression of these virulence genes in vitro, corresponding to its effectiveness in vivo, and (iii) a previously unknown gene (SMU.940) that is highly repressed by CT may be a novel factor associated with EPS accumulation, matrix assembly, and biofilm formation. We propose that CT is effective in reducing the development of cariogenic biofilms through several potentially complementary and/or overlapping mechanisms that primarily target S. mutans EPS-rich matrix production, while also compromising the overall fitness of the organism by repressing stress defense and/or altering bacterial membrane physiology (summarized in Fig. 6).
Significant alterations in the biochemical and physiological profiles of CT-treated biofilms (19) supports the hypothesis that CT-induced expression changes may account for a decrease in the incidence and severity carious lesions in the rodent model of the disease. Our transcriptome data confirms that the gtf genes (specifically gtfB) are major targets for CT, while these data simultaneously point to other genes/mechanisms (e.g., sloA and copY), repression of which may also impair EPS production. SloA is part of a manganese/iron transport system, while CopY is also predicted to be involved in copper transport (42). Thus, reduced expression of sloA and copY may further suppress transcription of the gtf genes, since copper and manganese act as effector molecules that modulate their expression (1, 7). This may help explain why this therapy is particularly effective in impairing gtfB gene expression and glucosyltransferase activity within S. mutans biofilms (19).
It has been widely recognized that the synthesis of exopolysaccharides, particularly those generated by GtfB, is necessary for the assembly and maintenance of plaque-biofilms, and is a major virulence factor associated with the pathogenesis of dental caries disease (2). GtfB binds to the membrane of S. mutans and other oral microorganisms in an active form, producing large amounts of insoluble glucan in situ (16, 18, 58). These biopolymers allow S. mutans to persist within the oral cavity by facilitating the formation of highly cohesive and adherent biofilms. EPS also acts as a protective and diffusion-limiting physical barrier and its disruption may leave biofilm cells exposed to inimical influences (2, 15, 38, 54). Insoluble glucans produced by GtfB are required for the formation of complex 3D biofilm structures on saliva-coated hydroxyapatite surfaces, since deletion of the gtfB gene results in a phenotype devoid of microcolonies in both single-species and multispecies biofilms (34, 63). A recent study showed a clear link between the formation of such microcolony complexes and the development of acidic microenvironments in close proximity to apatite surface; pH at the surface of attachment is inversely proportional to microcolony size (63). Such acidic niches provide an ideal milieu for the survival of acidogenic and aciduric bacteria, helping to ensure continued localized acid production (63), and subsequent development of carious lesions through demineralization of the tooth enamel (41, 45). Thus, it stands to reason that this would be a powerful target for any anticaries therapy aimed at impeding the ability of S. mutans to form the EPS-rich, acidic microenvironment that is associated with the clinical onset of caries disease.
Our data also show that CT may affect the ability of this pathogen to thrive and become dominant within the biofilm milieu, which is congruent with the reduced acidurance and acidogenicity observed in CT-treated biofilms (19). Treatment with the combination therapy likely impairs the fitness of S. mutans in biofilms by repressing the transcription of key genes involved in acidurance (e.g., dnaK) and surviving osmotic (e.g., potE) and oxidative stresses (1, 36, 37, 42, 46, 52, 60). Genes implicated in oxidative stress tolerance were especially highly repressed by CT, including sloA, sodA, and copY. SloA is required for virulence and oxidative stress tolerance by S. mutans (46, 52), whereas CopY may play a role in biofilm detachment (42). Metal accumulation in bacteria is not only important for its potential antioxidant benefit, but metals also frequently serve as cofactors for critical enzymatic and metabolic reactions, as is the case for superoxide dismutase (SodA) (13, 44, 53). SodA is an established virulence factor that reduces superoxide and is essential in vivo, the repression/dysfunction of which may be linked to repression of sloA, since manganese and iron are key cofactors for this enzyme (13).
Upregulation of lrgA and/or changes in the expression of fatty acid synthesis genes (e.g., upregulation of fabK) following CT treatment may be another indicator that the combination therapy is compromising the fitness of S. mutans, since it may correlate with alterations in the membrane physiology. LrgA is a membrane-associated protein related to the bacteriophage holin/anti-holin family (49) and has been implicated in cell lysis and genomic DNA release in Staphylococcus aureus (51), while the fab genes have established roles in fatty acid synthesis and membrane remodeling in S. mutans (48). tt-Farnesol and fluoride can increase proton permeability and affect the ΔpH across the membrane of S. mutans (20, 47); we have collected additional data that indicate CT treatment induces a global change in the membrane fatty acid profiles of S. mutans (see Table S5 in the supplemental material). Shifts in the fatty acids profiles can affect the function of F-ATPase and the overall permeability of the membrane, altering the ability of S. mutans to maintain intracellular ΔpH, greatly impairing acid tolerance (19, 47, 48).
One of the unique aspects of our work was the examination of the transcriptional profiles of selected S. mutans genes (e.g., gtfB and sloA) from in vivo plaque samples. First, we found potent and sustained effects on transcription in biofilms formed in vitro (up to 4 h) following brief exposures to CT. We then found that these genes were also repressed in vivo, which supports the notion that sustained alterations in the transcription of these targets may help to account for the reduction in biofilm virulence in the rodent model. gtfB expression was especially reduced at an early stage of caries development (3 weeks), whereas sloA was significantly repressed at a later stage (5 weeks). This pattern of expression might speak to the kinetics of gene repression in vivo, since the expression of gtfB is probably more critical in early biofilm formation when the matrix is constructed and the biofilm is initially established (2). On the other hand, the expression of sloA may be more critical in later stages, after which the biofilm has been continuously exposed to environmental stresses (36, 39). We propose that gtfB and sloA may represent key targets for CT. Enhanced repression of these targets could serve as the focal point of any subsequent formulations to further improve this therapy; this could possibly be accomplished by adjusting the concentration(s) of constituent agents.
In addition to disrupting established virulence mechanisms, CT affects a previously uncharacterized gene that appears to have a significant impact on biofilm formation and 3D architecture. Deletion of the SMU.940 gene dramatically alters biofilm formation; biofilms formed by the mutant possess significantly less EPS, are thinner, and have fewer and smaller microcolonies. A reduction in the amount of EPS and microcolony assembly would disrupt the creation of acidic microenvironments localized at the surface of biofilm attachment (63), potentially slowing the progression of the disease. Although the precise function of SMU.940 is currently unknown, it is possible that is has some relevance to the mevalonate pathway, since its location in the S. mutans chromosome is proximal to isoprenoid synthesis genes involved in cell wall biosynthesis (62). A defect in isoprenoid synthesis could potentially affect the topography and/or structure of the membrane. Changes in the membrane structure may affect GtfB binding to S. mutans and/or the function of glucan-binding proteins (Gbps), which could potentially reduce glucan production on the bacterial surface and its adhesive interactions with the EPS-matrix. If so, this could help to account for the defect in EPS accumulation and EPS-enmeshed microcolony development observed in biofilms formed by the ΔSMU.940 strain. Further studies focusing on GtfB binding capacity and enzymatic activity in situ, as well as Gbp expression/function may elucidate the potential functional role of this gene.
In summary, our chemotherapeutic strategy effectively bridges the gap between the current chemical modalities (e.g., fluoride and chlorhexidine) used to prevent or treat dental caries disease. It is not targeting the viability of the microorganism(s) per se, as with many broad-spectrum antiseptic approaches currently in use (20). Rather, it specifically impairs the ability of S. mutans to assemble cariogenic biofilms and to adapt and thrive within this milieu while augmenting the overall cariostatic properties of fluoride. However, additional studies are needed to determine the effects of CT on other species present in plaque biofilms. Fluoride does of course help to reduce demineralization and subsequently enhance remineralization (8, 47, 48, 57), which should complement the microbiological effects of CT. Whether CT actually enhances the bioavailability of fluoride remains to be elucidated. It is apparent that EPS in human plaque can influence the concentration of inorganic ions, such as fluoride (11).
Clearly, targeting the expression of genes involved in glucan synthesis and the stress response would be a highly effective strategy for impairing the ability of S. mutans to colonize the host, establish virulent biofilms in vivo, and ultimately elicit disease (37). These biofilms should be exquisitely attenuated, as they likely cannot mount a robust stress response and do not have a fully developed EPS-rich matrix. The therapy also disrupts a novel factor (SMU.940) associated with biofilm formation and EPS accumulation. Moreover, the effects on S. mutans transcription are sustained for hours after topical exposure to CT, suggesting that it may be clinically effective in humans, where significant intervals of time (hours) will pass between applications of oral hygiene products (21). CT may ultimately be successful in preventing and/or treating caries in humans; it is more effective than fluoride using a clinically relevant treatment strategy in a well-established biological model of the disease. Our data provide the scientific basis for proceeding to the next step toward developing a rationale for future clinical studies. At the same time, we have identified putative molecular targets that could be used as focal points for elucidation of the exact regulatory mechanisms and further improvement of the effectiveness of our therapy.
We acknowledge the National Institutes of Health for research funding through grants R01 DE16139, R01 DE18023, and T90DE021985.
Published ahead of print 17 September 2012
Supplemental material for this article may be found at http://aac.asm.org/.