Preventing root caries development under oral biofilm challenge in an artificial mouth

Objectives: To study the preventive effects of chlorhexidine against root caries under oral biofilm in an artificial mouth. Study Design: Sixteen human tooth-root disks were inoculated with a salivary sample that was produced by mixing the unstimulated saliva of three adults who had no untreated caries. The disks were incubated in an artificial mouth fed with a 5% sucrose solution three times daily for one week. Eight disks received a twice daily rinse of 0.12% chlorhexidine (test group). The other eight disks were rinsed in distilled water (control). The biofilm was then studied with three techniques: colony forming unit (CFU) counting, scanning electron microscopy (SEM) and confocal laser scanning microscopy (CLSM). The changes in the chemical structure of the root surface were studied using Fourier transform infra-Red spectroscopy. Type-I collagen and proteoglycans on the root surface were quantified using immunocytochemical staining. Results: The log CFU for the test and control groups were 4.21 and 8.27, respectively (p<0.001). The CFU count of Streptococci and Lactobacilli were negligible. Both the SEM and the CLSM showed suppressed bacteria growth in the test group. The log [amide-I: HPO42-] of the test and control groups were 1.11 and 1.93, respectively (p=0.02). The mean counts of sound type-I collagen in the test and control groups were 16.8/?m2 and 13.0/?m2, respectively (p<0.001), whereas the mean counts of intact proteoglycans were 5.6/?m2 and 3.5/?m2, respectively (P<0.001). Conclusions: Chlorhexidine suppressed the growth of selected cariogenic bacteria in oral biofilm on the root surface and thus protected tooth-root from cariogenic challenge. Key words:Chlorhexidine, biofilm, caries risk, root, caries, artificial mouth, demineralization, streptococci, lactobacilli, proteoglycans, collagen I.


Introduction
The proportion of older people in the population is gro�ing and their inadequate plaque control procedures ma�e root caries an increasingly common dental health problem (1). Soft tissue recession due to age, traumatic toothbrushing habits, periodontal disease or periodontal treatment �ill unavoidably result in more tooth root surfaces that are at ris� for the development of root caries (2). Restorative treatment of root caries is notoriously difficult. Post�treatment pain and hypersensitivity are very common, and this may contribute to increased tooth loss in many people �ith root caries. Therefore, prevention of root caries is an important issue in clinical practice. Fluoride agents have been used in dental caries management (3). In addition, Clinical experience and popu-lation�based studies have demonstrated that mechanical plaque control methods can maintain good oral hygiene and oral health. Ho�ever, these plaque control methods require good manual dexterity and can be difficult to implement in certain circumstances, such as tooth hypersensitivity or among older patients (4). In these circumstances, antimicrobial agents may serve as a valuable complement to mechanical plaque removal (5). Chlorhexidine is one of the commonly used antimicrobial agents in the management of caries and periodontal diseases. Its popularity is not only due to its broad antimicrobial spectrum, �hich includes Gram�positive and Gram�negative bacteria, but also due to its retention in the mouth, �hich prolongs its antimicrobial effect. The upta�e of chlorhexidine by bacteria has been sho�n to be extremely rapid, �ith a maximum effect occurring in a very short time of around 20 seconds (6). Nevertheless, Autio�Gold (7) revie�ed the chlorhexidine literature and found that its action on dental plaque or oral biofilm is inconclusive. Additional factors such as the strain of bacteria in the oral biofilms (8), the gro�th condition of the oral biofilm (9) and the nature of the bacteria in the substratum (10) might affect the efficacy of chlorhexidine. Further studies are thus necessary to study the anti�caries effect of chlorhexidine. A recent systemic revie� concluded that the development of caries on the root surface is associated �ith the composition and quantity of dental plaque, diet, the composition and flo� of saliva, and exposure to fluoride (2). Laboratory studies �ere conducted to anti�caries effect of chlorhexidine on single species biofilms (11,12) in simplified environment. It is note�orthy that bacteria in biofilms are notably less susceptible to chlorhexidine than �hen gro�n in batch culture. Many bacteria living together in a biofilm experience stress. This implies that their stress defense mechanisms are turned on and thus they are even less susceptible to chlorhexidine. Dental plaque is a complex multi�species biofilm �here bacteria communicate, protect one another and form a strong resistance to anti� microbial. Study using a complex multi�species biofilm generated from oral cavity should be more realistic approach to evaluate anti�microbial effect of chlorhexidine (13). Furthermore, it is also more desirable to study oral biofilms in an artificial mouth that closely simulates oral environment such as salivary flo�, redox potential, acidity and temperature. This experiment �as therefore carried out to study the effects of chlorhexidine on tooth root surface challenged �ith oral biofilm developed from human saliva in a sophisticated artificial mouth.

Material and methods
�Tooth dis� preparation This study �as approved by the Institutional Revie� Board (IRB�UW08�952) and patient consent �as obtained before the study. Diamond trephine �as used to prepare 16 tooth root dis�s �ith 5mm diameters from 16 extracted sound human molars. Half of tooth root dis�s' surfaces �ere covered �ith nail vanish (Clarins, Paris, France) as an internal control. The remaining surfaces of the tooth root dis�s �ere covered �ith varnish. The tooth dis�s �ere sterilized �ith ethylene oxide for 16 hours (14). �Oral bacteria sampling The method of creating an oral bacteria sample �as adapted from Navazesh (15). Three healthy middle aged patients (35 to 44 years old) attending a dental hospital �ere recruited for saliva collection. They had no clinically detectable caries, no periodontal disease �ith poc�et depth 4mm or above, no salivary gland disease or dysfunction, no systemic diseases and they �ere not ta�ing medication. They had abstained from their oral hygiene practice for 24 hours prior to saliva collection. Each patient �as as�ed to spit out unstimulated saliva. Five ml of saliva �as collected from each patient and then the saliva samples collected from the three patients �ere mixed together and centrifuged for 10 min at 3,000 rpm. After centrifugation, cell pellets �ere harvested and �ashed three times �ith 1% phosphate buffered saline (PBS). The �ashed pellets �ere re�suspended in PBS for bacteria adhesion. �Formation of oral biofilms in the artificial mouth Each root dis� �as inoculated �ith a 300µL aliquot of bacteria and �as then inserted into the artificial mouth for bacteria adhesion. The artificial mouth is sho�n in figure 1. A humidified gas mixture of 5% carbon dioxide and 95% nitrogen �as supplied continuously at 60ml/min. The temperature inside the incubator �as maintained at 37ºC. Simulated oral fluid defined medium mucin (DMM) �as continuously supplied at 0.06 mL/min to simulate the salivary flo� (16). A sucrose solution at 5% �as supplied for six minutes �ith a flo� rate of 15 mL/hr monitored by a computer program (LabVIEW ® soft�are Version 2.2). The sucrose supply �as delivered every eight hours to simulate a real life dietary situation. The inoculated bacteria �ere allo�ed to gro� and form a biofilm on the surface of the tooth dis�s in the artificial mouth for seven days. �Group assignment The 16 dis�s �ere randomly divided into test and control groups. Throughout the seven day test period, 0.12% chlorhexidine gluconate �as supplied at a flo� rate of 0.25 mL/min for six minutes every t�elve hours to the test group of tooth root bloc�s in the artificial mouth. Distilled �ater �as used in the control group. �Biofilm study After the seven�day experiment, the oral biofilm �as collected from the tooth root bloc�s. The Gro�th Kinetic of the generated biofilm �as assessed by Colony Forming Unit (CFU) counting. The total CFU �as counted in a blood agar plate and individual species �ere counted in selective medium plates. Selective media agar plates of Mitis Salicarius, Rogosa and Actinomyces �ere used for Streptococci, Lactobaciili and Actinomycetes spp., respectively. Topographical features of the biofilm �ere observed using scanning electron microscopy (SEM) (Leo 1530 Gemini, Ober�ochen, Germany) at 12 �V in high�vacuum mode (17). The viability of the biofilm �as studied using confocal laser scanning microscopy (CLSM) (Fluovie� FV 1000, Olympus, To�yo, Japan). The biofilms �ere labelled using t�o fluorescent probes, specifically propodium iodide (PI) and SYTO�9 (LIVE/ DEAD BacLight Bacterial viability �it, Molecular Probes, Eugene, OR, United States). PI specifically labels the dead cells in red, �hereas live cells are labelled green by STYO�9. Cellular images of the biofilms �ere performed using CLSM (Fluovie� FV 1000, Olympus, To�yo, Japan) (19). Four images of each biofilm specimen �ere obtained using CLSM (Fluovie� FV 1000, Olympus, To�yo, Japan) and examined using special image analysis soft�are (Image J; National Institutes of Health, USA). The red�to�green (dead�to�live bacteria) ratio �as calculated to indicate the anti�microbial effect of the therapeutic agent (18). �Hard tissue study Each tooth root dis� �as sectioned longitudinally to create thin sections that �ere each approximately 120 e560 µm thic�. The changes in the chemical structure of the root surface �ere studied using Fourier transform infra� red spectroscopy (FTIR) (UMA�500 machine, Bio�Rad Laboratories, CA, USA), and the amount of type I collagen and proteoglycans �ere quantified using immunocytochemical staining. In the FTIR analysis, the mineral content �as calculated on the basis of the spectrally derived matrix�to�mineral ratio (the integrated area of protein amide I pea� bet�een 1585�1720cm�1 and phosphate (HPO 4 2-) pea� bet�een 900 and 1200cm�1) from an area of 100 × 100 µm on the surface of the root bloc�s. The varnished part of root surface of the dis� �as used as an internal control. After the FTIR assessment, the six thin sections �ere then ultrasonicated for one minute in de�ionized �ater (pH 7.4) and exposed to 10% citric acid for 15 seconds to decalcify the surfaces of the sections. A double� immunolabelling procedure �as performed �ith t�o monoclonal primary antibodies: an IgG anti�type I collagen and an IgM anti�chondroitin 4/6 sulphate (mouse monoclonal; Sigma Chemical Co., St. Louis, MO, USA). They �ere used for the simultaneous detection of the distribution of antigenically intact collagen fibrils and proteoglycans. Gold labeling �as performed �ith t�o secondary antibodies conjugated �ith gold particles of different sizes: an IgG goat anti�mouse�IgG conjugated �ith 30�nm gold particles (British BioCell International) for type I collagen identification, and an IgG goat anti�mouse�IgM conjugated �ith 15 nm colloidal gold for chondroitin 4/6 sulphate identification (British Bio-Cell International). The specimens �ere prepared and examined under SEM (Leo 1530, Ober�ochen, Germany) at 20 �V. Images �ith the same magnification �ere obtained for each section (n=6 in each group). The labeling index �as calculated as the mean of the gold particle number/μm 2 in the visible organic net�or� obtained from each image (19).

-Statistical Analysis
Analyses �ere performed �ith SPSS 17.0 soft�are (SPSS Inc., Chicago, IL, USA). A parametric t test �as used to detect differences bet�een the treatment and control groups in log CFU value and log [amide I: HPO42�] value, and bet�een the mean number of 30 nm of colloidal labeling type I collagen and the 15 nm particles labeling proteoglycans. A 5% statistical significance level �as used for all analyses. Table 1 sho�s the bacterial count, in log CFU, for both the test and the control group after the seven days of the experiment. Streptococci, Lactobacilli, Actinomycetes and the total bacterial counts �ere significantly lo�er in the test group than in the control group. The SEM images of the roots after seven days of incubation (Fig.  2) also sho�ed differences bet�een the test and control groups. In the test group, no biofilm �as found, only small clusters of bacterial cells �ere observed as iso-  lated groups, �hereas a mono�layer of sparse biofilm �as observed in the control group. In the CLSM images (Fig. 3), red (dead) cells clearly dominated the green (live) cells in the chlorhexidine treatment groups, indicating that the effects of chlorhexidine on bacteria viability �as significant. The red to green (dead to live) ratios of the bacteria in the test group �as 26.21±13.3

�Oral biofilm characteristics
to 0.02±0.01, sho�ing a significant increase of red cells �ith chlorhexidine applications (p=0.02). �Hard tissue characteristics As table 2 sho�s, the log [amide I: HPO42�] of the human root surface under the oral biofilm treated �ith chlorhexidine �as significantly lo�er than that of the control group (p=0.022). (Fig. 4) sho�s an SEM im-   age of the immunocytochemical staining of the type I collagen (30 nm particles) and proteoglycans (15 nm particles); the colloidal gold is seen as bright spots. In the control group, there �ere structural modifications to the collagen net�or� (such as collapse, s�elling, and branching) of the root surfaces that had not received the chlorhexidine treatment. The test group, treated �ith chlorhexidine, sho�ed little damage to the collagen. (Table 2) and (Fig. 4) sho� that there �ere significant differences in the number of 30 nm gold particles representing type I collagen detected in the test and control groups (p<0.05). The distribution of proteoglycans, as represented by the number of 15 nm gold particles, �as higher in the test group than in the control group (p<0.05).

Discussion
When considering the outcome of this study, the limitations of such an in vitro study must be ta�en into account. The artificial mouth used in this study is one of the most sophisticated model systems used for caries research; it mimics the in vivo environment in terms of temperature, humidity, sucrose supply and salivary rate. Ho�ever, it has its o�n stable environment that is different from the in vivo environment. In addition, there are variations in the composition of biofilms bet�een individuals. The variations apply not only to species composition, but also to baseline metabolic activities and to the consortia response to various gro�th substrates or antimicrobials. A study of the temporal changes in bacterial population using denaturing gradient gel electrophoresis found bacterial community diversity in the saliva of different individuals (20). The present study collected bacterial samples from three individuals. The results cannot be interpreted as representative of the bacteria existing in a large population.
The consortia biofilm of four species of oral bacteria on bovine enamel and dentine (21) and human dentine (22) successfully reproduced caries in an artificial mouth. These studies sho�ed that a microcosm can be generated, creating a laboratory subset that evolved from a natural system (23). After five days, the microcosm formed using mixed salivary bacteria is similar to oral biofilms and also varies bet�een individuals (24). A stabilized microcosm is thus one of the better simulations of in vivo oral biofilm. Therefore, the current study used mixed salivary bacteria from individuals to form a microcosm.
The anti�bacterial activity of chlorhexidine has been studied using various models. It has been sho�n that the minimum concentration of chlorhexidine required to �ill bacteria in a biofilm is considerably higher (10� 100 times) than the amount required to �ill the bacteria in a suspension. Theoretically, the inhibition of caries by an antimicrobial agent can be achieved by the in-hibition of acid production or the formation of dental plaque. In the present study, chlorhexidine �as supplied from the beginning of the experiment. It �as found that chlorhexidine had a higher suppression on the activities of selected cariogenic bacteria, including streptococci and Lactobacilli, than on the activities of Actinomyces. This study also found that in the presence of chlorhexidine, the bacteria failed to form biofilm �hen compared �ith the control group, in �hich a mono�layer of biofilm �as formed. Similar results have been reported in previous in vitro studies, �hich found that pulses of chlorhexidine led to a selective long term suppression of Streptococci mutans (25,26). Moreover, it �as also found that chlorhexidine �as more efficacious against streptococci and Lactobacilli than against Actinomyces. In a previous clinical trial, an increase in the number of Actinomyces viscosus/naeslundii �as noted after treatment �ith chlorhexidine varnish (27). This study also found a significantly higher dead/live ratio value �ith chlorhexidine application, suggesting that chlorhexidine could exert an antimicrobial effect on biofilm. Computer soft�are �as used to differentiate colors (red/green) and areas to ma�e quantitative analysis of confocal images possible (28). The results, ho�ever, may vary due to the uneven distribution of bacteria in different thic�nesses of biofilm. Furthermore, the quality of the images could be affected by conditions such as brightness, �hite balance and contrast. Therefore, this method �as only used to support the analysis, but conclusions cannot be dra� solely based on these results. Once the apatite crystallite of tooth tissue is dissolved, collagen fibers in the tooth are exposed and amide I is released as a by�product of collagen brea�do�n. The mineral density of the root surface can be assessed �ith the HPO 4 2band in the FTIR spectrum (29). The log [amide I: HPO 4

2-
] value has been used in previous laboratory studies as an indicator of the extent of demineralization of tooth tissue; the larger log values correspond to a greater extent of demineralization (22,30). The FTIR results of this study sho�ed there �as less demineralization of root surface in the chlorhexidine treatment groups than in groups �ithout treatment. The use of a double�labeling technique permitted an investigation of both the presence and the distribution of type I collagen fibrils and proteoglycans on the root surface (31). These are the main structural components of the root surface net�or� (32). The number of gold particles representing type I collagen and proteoglycans �ere higher in the test group than in the control group. With the chlorhexidine application, the bacteria failed to form a biofilm and the harmful effects of the bacteria on root surfaces �ere thus reduced in the test group and the collagens in the root surface �ere protected. In conclusion, Chlorhexidine suppressed the activities of selected cariogenic bacteria in biofilm generated from human saliva; it also protected the mineral and organic content of human tooth roots from caries attac�s. It can be a useful agent to prevent root caries in particular those patients �ith high caries ris�. Further clinical trials should be performed to substantiate its effectiveness in root caries prevention.