In vitro sl:udy of the penetration of three bacterial strains into root dentine

Fabienne Perez, DCD, Tarik Rochd,b Jean-Philippe and Georges Michel,e Toulouse, France
UFR ODONTOLOGIE AND UFR SC. PHARMACEUTIQUES.

Lodter, DSO, Paul Calas, DSO,d

Thle purpose of this study was to assess in vitro migration of the following three bacterial species into dentinal tubules: Streptococcws sanguis, Actlnomyces naeslundii and Prevotella intermedia. Twenty-seven bovine incisors were randomly divided into three groups. Only the root dentin was used for the experiment. Within each group, the nine incisors were sectioned into 36 dentin blocks that were incubated with one of the tested bacterial suspensions. Twelve samples were taken after 10, 20, and 28 days. Half the samples were observed with scanning electron microscopy and the others with light microscopy after standard histologic procedures, and the depth of bacterial penetration was measured. Analysis of the examinations showed that whatever the technique used, only one strain on this experimental model migrated into dentinal tubules. Ssnguis was observed at a depth of 792 pm, but no migration was observed for either A.naes/uncfi or P.intermedia. The differences in migration of the bacteria tested appear to be related to their morphologic factors and cellular arrangement.
(ORAL SURC ORAL MED ORAL PATHOL 1993;76:97-103)

The bacteria responsible for endodontic infections have been well known since the works of Sundqvist and Bystriim.le4 Even if the bacteria considered to be the most pathogenic are gram-negative bacilli, particularly black-pigmented Bacteroides,5-8 other bacterial species such as Actinomyces seem to be found frequently in failed endodontic treatment cases.gylo These failures are mainly linked to the persistence of bacteria inside the root canal system. Besides the main root canal, bacteria can lodge in cementum crypts, secondary canals, or dentinal tubules. Previous studies have revealed bacterial penetration in the root dentin, but the bacterial species studied were different from the root canal flora, l1 or were only observed by a single method, light microscopy,r2 I3 thus giving random results. Consequently, the aim of this study was to observe in vitro migration of the foIllowing three nonmotile bacterial species that belong to conventional root canal flora into root dentinal tubules: a gra-m-positive COCCUS, Streptococcus sanguis; a gram-positive bacillus, Actinomyces naeslundii; and a gram-negative bacillus, Prevotella
aAssociate Assistant, Endodontics Department, Laboratoire de Biomateriaux, UFR Odontologie. bLaboratoire de Virologie et Microbiologic Industrielle, UFR SC. Pharmaceutiques. Professor, Head of Research Department, Laboratoire de Biologie Buccale, UFR Odontologie. dProfessor, Head of Endodontics Department, Laboratoire de Biomattriaux, UFR Odontologie. eProfessor, Head of Research Department, Laboratoire de Virologie et Microbiologic Industrielle, UFR SC. Pharmaceutiques. Copyright @ 1993 by Mosby-Year Book, Inc. 0030-4220/93/$1.00 + .lO 7/15/46623

intermedia. Two observation techniques were used: scanning electronic microscopy (SEM) and light microscopy.
MATERIAL AND METHODS

The following bacterial strains were used for this experiment: Streptococcus sanguis NCTC 7853 cultured in trypticase soy agar (TS) and incubated in TS broth; Actinomyces naeslundii ATCC 10039 cultured in chocolate-polyvitex agar and incubated in Schaedler broth with Vit K3; and Prevotella intermedia NCTC 93336 cultured in schaedler agar and incubated in Schaedler broth. P.intermedia is an obligate anaerobe, A.naeslundii, and S.sanguis are facultative anaerobes, and the three test organisms were cultured under anaerobic conditions. Twenty-seven freshly extracted bovineincisors were used for these experiments. They were stored in a 0.5% sodium hypochlorite solution (NaOCl) (Prolabo, Paris, France) for surface disinfection. In accordance with the technique described by Haapasalo and Orstavik,14 the apical 5 mm and the crown were sectioned with the use of a rotating diamond saw and water irrigation (Komet 925 P, Brasseler Gmbh, Lemgo, Germany) (Fig. 1). The root canal was widened to a diameter of 2 mm with a tungsten carbide round bur (Komet HlS IS0 023, Brasseler Gmbh). Root cementum was completely removed by a tungsten carbide bur (Bisico 55 1O-60, Lan9on/Provence, France) to obtain a 15 mm dentin cylinder. Four equally thick slices were sectioned with the use of the rotating diamond saw and water irrigation (Komet 925 P, Brasseler Gmbh). Each slice sample had a mean diameter of 6 mm and a height of 4 mm. The
97

98 Perezet al.

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY

July I993

Fig. I. Preparation of bovine dentin specimens.

samples were then treated in an ultrasonic bath (Bransonic 2200 El-50 KHz) in a 17%disodium ethylenediamine tetraacetate (EDTA) solution (Prolabo, Paris, France) for 10 minutes, followed by a 2.5% NaOCl solution for 10 minutes. They were finally rinsed in a phosphate-buffered saline solution (pH = 7.2) to remove all traces of EDTA and NaOCl. A few random samples were taken at this stage with a SEM to check the apertures of the dentinal tubules. The sampleswere then sterilized for 30 minutes in an autoclave at 120 C. Sterilization was checked by placing them in a TS broth (Bio-Merieux, Marcy lttoile, France) for 10 days at 37 C in anaerobic conditions. This also allowed the culture medium to penetrate the tubules, thus favoring bacterial growth. Such penetration was further increased by placing a 2 mm diameter sterile ultrasonic titanium probe in the medium for 10 minutes before inoculating (Vibracell, Bioblock scientific-600W, 20KHz-Sonics and materials, Danbury, Conn.). The first 36 blocks of dentin were incubated in tubes that contained a S.sanguis NCTC 7863 suspensionin 2 ml TS (1 O9cells/ml). Another 36 blocks were then incubated with a A.naeslundii suspension in Schaedler broth and the last 36 samples with a suspension of P.intermedia in Schaedler broth. Each tube contained two samples.The medium was changed every 4 days. Culture purity was systematically checked whenever the medium and suspensionwere changed. Such handling necessitated highly rigorous asepsis and was performed in a laminar flow hood. Twelve blocks were removed for each experiment after 10 days, 20 days, and 28 days of incubation respectively. Development of infection, hence of the migration of the bacterial strain inside the tubules, was revealed by examining samples with the SEM and by using light microscopy after a Brown and Brenn stain.

For SEM examination, sampleswere fixed in 2.5% glutaraldehyde for 30 minutes, dehydrated in increasingly concentrated ethanol solutions, and longitudinally split. They were then gold plated to a thickness of 200 nm and observedwith a SEM (CAMBRIDGE S 250 MK3, Cambridge, U.K.). After fixing (10% neutral formol), demineralization (5% nitric acid), and dehydration in ethanol solutions, the preparations were embedded in paraffin for histologic examination. Five micron thick sections were then cut. Each section was stained using the Brown and Brenn technique for routine light microscopicexamination. l5 On average; 11sectionswere extracted from each dentin block in different areas after reorientation of the dentin piece. These sections were representative of the entire thickness of the dentin block. Observations were carried out using two magnification strengths (X40 and X100), and the migration distance from the pulpal surface to the outer dentin was estimated with a graduated, precalibrated microscope eyepiece (Leitz, Hamburg, Germany, X10).
RESULTS

Over the three experiments, 27 bovine incisors provided 108 dentin blocks of which 54 were observedby SEM and 54 by light microscopy. Only light microscopy revealed maximum penetration of the bacteria. SEM observation was impossible as SEM examination did not display all of the bacteria migrating through the tubules with sufficient certainty. Indeed, the defects in the slices or the actual anatomy of the dentinal tubules and their three-dimensionally winding paths restrict observation of the bacteria and hence the detection of their maximum depth of penetration. SEM gives less complete results but does confirm the data obtained with the use of a light microscope.
Migration of Streptococcus sang&

SEM examination. After 10 days S.sangwis cells are shown at a depth of 200 pm in samples. Many of the tubules are empty. Penetration increases after 20 then 28 days. The canal wall is entirely colonized by S.sanguis that assuresa chain-like organization and bacteria are located more than 600 pm inside the tubules (Fig. 2). Light microscopy examination. As with the SEM examination, it could be seenafter 10 days that there were a great number of empty tubules among the sections. Bacteria was collected on the canal wall on tubule apertures and scattered throughout the circumference of the canal. After 20 days, not all tubules were infected, but

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY

Volume 76. Number 1

Perez et al.

99

Fig. 2. SEM of Ssanguis nification X50008.)

at

28 days.S: Cells of S.sanguis 666 pm deep,many tubulesare empty. (mag-

migration had increased to reach a maximum of 792 w. After 28 days, a fringe of dentinal infection of approximately 150 bbrncould be located around the canal and the slices observed revealed a major penetration of S.sanguis (Fig. 3). Data on penetration minimum and maximum are given in Table I. In light microscopy, one sample had to be discarded because the observations did not provide a sufficiently reliable reading and measurement of bacterial penetration.
Migration of Actinomyces naeslundii

Table

I. Lengths of penetration of S. sanguis cells into root dentine

Extreme

Meanpenetration (s.4 PM

values(minimummaximum)pm
150-792 150-648 180-792 300-737

All specimens
10 days 20 days 28 days I

458.8 (+ 109.3) 4;::s I: 478.6 (t Kj 101.0)

After 10 days, A.naeslundii cells covered the canal wall but no tubules were infected. After 20 days, areas of intense bacterial concentration alternated with these where no cells were observed.After 28 days, bacteria were arranged in clusters, but there was no penetration into the dentinal tubules. Light microscopy examination. No bacterial migration was observed on histologic slices. Bacteria stained blue with the Brown and Brenn technique created a fringe concentrated on the canal periphery but had not penetrated the tubules. After 28 days, enormous bands of bacteria were occupying the canal wall and blocked tubule apertures (Fig. 4).
SEM examination.
Migration of Prevotella intermedia

pletely empty without any traces of P.intermedia penetration. Light microscopy examination. The pink staining of the gram-negative bacteria made it difficult to examine the tubules themselves, which were stained yellow. We could, however, observe that there were few bacteria visible in the lumen of the canal and that no P.intermedia cells had penetrated the dentinal tubules. Some clusters were located on the canal wall (Fig. 5).
DISCUSSION

Bacterial colonization was very slight on the wall after 10 days, and no bacilli were detected in the dentin. Bacilli had accumulated on the wall over time, but the tubules had remained comSEMexamination.

The experiment revealed that ofithe three bacterial speciestested, only one S.sanguis migrated into the tubules of bovine dentin. However, can this be correlated to any classification? S.sanguis is a gram-positive coccus,but belonging to any particular gram category does not seemto affect migration becauseA.naeslundii, which is also

%I

Perez et al.

ORAL SURGERY ORAL MEDICINEORAL PATHOLOGY July 1993

Fig. 3. Light microscopy of S.sanguis at 28 days. Chain of S.sanguis in dentinal tubules. (Brown and Brenn stain; original magnification X100.)

Fig. 4. Light microscopy of A.naeslundii at 28 days. Bacterial fringe of A.naeslundii cells: A, covers canal wall; D is dentin. (Brown and Brenn stain; original magnification ~25.) a gram-positive bacillus, does not migrate any more than P.interPnedia, which is gram-negative. If we consider the classification from the oxygen tolerance point of view, this experiment studied the migration of two facultative anaerobes: S.sanguis, A. naeslundii and of one obligate anaerobe: P.intermedia. According to the results, this factor doesnt seem to be the cause of the different migratory behavior of the three bacterial strains. This corroborates the works of Akpata and Blechman12 who studied the migration of two facultative and two obligate anaerobes. More logically, the intervening factor would appear to be morphologic both from a qualitative and quantitative viewpoint. It is a qualitative factor because cocci penetrate the tubules and the two bacilli do no% migrate, but it is also a quantitative factor because bacteria aggregate form clusters, and it is this morphologic structure that seems to condition bacteria! penetration into dentinal tubules. Measurements of bovine dentinal tubules made in iight microscopy or in SEM have revealed varying diameters: 4 to 7 pm after dentinal pretreatment. It is obvious that S.sanguis, the size of which is about 1

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY

Volume 76, Number I

Perez et al.

101

Fig. 5. Light microscopy of Pintermedia at 28 days. Clusters of P.intermedia cells: Pi located on canal surface, D is dentin. (Brown and Brenn stain; original magnification X100.)

w,

l6 can penetrate the tubules. Yet, the mode of interbacterial grouping seemsto be responsible for the behavior of each species. S.sanguis cells grow by elongation along main cell axis and are organized in chain.j This organization is like a train that slips into and slides around the tubules without the chain being a geometric obstacle to its progression. This was very clearly observed by SEM and light microscopy. Moreover, the S.sanguis strain is not mobile, and given the depth of penetration, it is remarkable that such a bacterial strain should travel such a distance. We can therefore ask how and why bacteria penetrate the dentin? Two phenomenamay occur: first, bacteria that grow under favorable conditions multiply in the culture medium and as this medium penetrates dentinal tubules, S.sanguis can penetrate deeply by a passive process, and secondthe culture medium cannot be renewed in the tubules but is enough nutrition to permit growth and multiplication of bacteria inside dentinal tubules and to increase their penetration. As for the A.naesZundii migration, tubule diameter is the sameas in the first experiment, but in the results we have seen th[at A.naeslundii cells were only concentrated on the canal wall. Yet the size of these filamentous bacteria, approximately 1 pm X 5 pm, and their elongated forms would be propitious in making their way through the tubules.17 Consequently, their grouping mode should be examined to try to explain this lack of penetration. After 24 hours of incubation, A.naeslundii microcolonies had a densecenter composedof diphteroidal cells and of partly fragmented filaments surrounded

by long, branched, irregularly curved threads, projecting out in all directions.18The morphologic structure of Actinomyces may vary considerably according to media, incubation conditions, and the age of inoculum. Consequently, curved filaments can be straighter and of varying length. Others have bulging or budding ends.A.naeslundii cells may produce long filamentous projections that radiate from the cell wall surface. SEM was used to describe Actinomyces cells and mode of growth more precisely. Surface fibrils may play a major role in the attachment of these bacteria to epithelial cells, tooth surfaces, or other bacteria. The works of Ellen et a1.20 have in fact shown that when pili or fimbriae are removed from A.naeslundii cells, their attachment is significantly reduced. In addition to pili, A.naeslundii cells possessdevices for aggregation through the production of dextrans and levans.21Culture media that contains carbohydrates and consequently has low pHs also allow an accumulation of A.naeslundii through the synthesis of elements required for their aggregation. So both their fibril structure and wall composition can agglutinate A.naesZundii cells and form the entangled clusters that we have clearly observed. Although the initial size of bacteria is much lower than tubule diameter, this standard grouping mode prevents all cells from penetrating dentinal tubules. As we did not find any isolated bacteria whatever the observation techniques used, it is highly likely that aggregation be produced very quickly, and it is consequently unlikely that single cells can be observedin the tubules.

702 Perez et al.

ORAL SURGERY ORAL MEDlClNE ORAL PATHOLOGY

July 1993

In SEM, clusters of A.naesEundii frequenlly have a nonhomogeneousdisposition. Perfectly clean dentin zones are close to zones covered with bacteria. This preferential location in someareas could be the result of constituent dentin differences. But, as all samples are taken from bovine dentin and have undergone the same dentinal treatment, it is difficult to explain this observation. In light microscopy, histologic sections reveal a great cell density, but the canal wall is partially covered. We also found the same results in the SEM examination As was seen earlier, the absence of intratubular P.intermedia migration was more diflicult to observe becauseof the pink tint of the cells after Brown and Brenn staining. The SEM study shows P.intermedia cells on the canal wall, which is the sign of a bacterial growth, but without major penetration into the dentin. Bacteria have progressively accumulated on tubule apertures over time, which prevents any progression into the tubules. In this same experiment, the size of P,intermedia cells did not initially seemto be an obstacle to their penetration because the mean size was 0.7 pm 23 wide by 2 pm long. 22, But we are forced to admit that no bacteria were detected in the dentinal tubules. P. intermedia cells are little bacilli with central or terminal swellings. Culture conditions may also modify the morphologic aspectof the cells; they are longer when stained from broth cultures and larger when fermentable carbohydrates are present. The cell surface of P.intermedia also contains fibrillar structures or pili that allow coaggregation.24 Another obstacle to penetration could also be the persistence of the odontoblastic prolongment inside the tubules that could act as a physical barrier. But SEM images did not show any traces of these structures, and we consider that they were totally eliminated from the samples before bacterial inoculating. As there are a large number of bacteria on the wall after 28 days, it is plausible to think that if they could have migrated into the tubules, they would have already done so.
ONCL~SION

or of bacteria that are similar in cellular morphologic factors but differ in growth pattern,
REFERENCES 1. Sundqvist G. Bacteriological studies of necrotic dental pulps.

[Dissertation no. 71. Umea University, Umea, Sweden: 1976. 2. Bystrom A, Sundqvist 6. Bacteriologic evaluation of the et%cacy of mechanical root canal instrumentation in endodontic therapy. Stand J Dent Res 1981;89:321-5. 3. Bvstriim A. ClaessonR. Sundovist G. The antibacterial effect of camphorated paramonochldrophenol, camphorated phenol, and calcium hydroxide in the treatment of infected root canals. Endod Dent Traumatol 1985;1:170-5. 4. Bystriim A, Sundqvist 6. Bacteriologic evaluation of the effect of 0.5 percent sodium hypochlorite in endodontic therapy.
ORALSURGORALMEDORALPATHOL 1983;55:307-12.

5. Sundqvist GV; Eckerbom MI, Larsson AP, Sjijgren UT. Capacity of anaerobic bacteria from necrotic dental pulps to induce purulent infections. Infect Immun 1979;25:585-93. 6. Griffee MB, Patterson SS, Miller CH, Kafrawy AH, Newton CW. The relationship of Bacteroides melaninogenicus to symptoms associated with pulpal necrosis. ORAL SURG ORAL
MED ORAL PATHOL 1980;50:457-61.

The differences in penetration observed in this study among the three bacterial speciesshow that the processof migration under in vitro conditions could depend not only on morphologic criteria but also on interbacterial aggregation criteria. A possible evolution of this work will be to study the migratory behavior of isogenic strains of the same bacteria that differ only in intercellular arrangement

I. Van Winkelhoff AJ, Carlee AW, De Graaff J. Bacteroides endodontalis and other black pigmented Bacteroides speciesin odontogenic abscesses. Infect Immun 1985;49:494-7. 8. Sundqvist G, Johansson E, SjBgren U. Prevalence of black pigmented bacteroides speciesin root canal infections. J Endod 1989;1:13-9. 9. Sjogren U, Happonen RP, Kahnberg KE, Sundqvist G. Survival of Arachnia propionica in periapical tissue. Int Endod 3 1988;21:277-82. 10. Nair RPN, Sjogren U, Grey G, Kahnberg KE, Sundqvist G. IntFaFadiCUkir bacteria and fungi in root-filled, asymptomatic human teeth with therapy-resistant periapical lesions: a longterm light and electron microscopy follow-up study. J of Endod 1990;16:580-8. 11. Orstavik D, Haapasalo M. Disinfection by endodontic irrigants and dressingsof experimentally infected dentinal tubtrIes. Endod Dent Traumatol 1990;6:142-9. 12. Aknata ES. Blechman H. Bacterial invasion of ouloal dentin wail in vitro. J Dent Res 1982;61:435-8. z A 13. Safavi KE, Spangberg LSW, Langeland K. Root canal dentinal tubule disinfection. J Endod 1990;16:207-10. 14. Haapasalo M, Orstavik D. In vitro infection and disinfection of dentinal tubules. J Dent Res 1987;66:1375-9. 15. Brown JH, Brenn L. A method for the differential staining of gram-positive and gram-negative bacteria in tissue sections. Bull Johns Hopkins Hosp 1931;48:69-73. 16. Hardie JM. Oral steptococci, gram-positive cocci-genusstreptococcus.In: Sneath PHA, Mair NS, Sharpe E, eds. Bergeys manual of systematic bacteriology. Baltimore: Williams and Wilkins, 1986;1043-71. 17. Schaal KP. Irrigular, nonsporing, gram-positive rods, Genus Actinomyces. In: Krieg NR, Holt JG, eds. Bergeys manual of systematic bacteriology, Vol. 2. Baltimore: Williams and Wilkins, 1984:1383-1418. 18. Schaal KP, Pulverer G. The genera Actinomyces, Agromyces, Arachnia, Bacterionema, and Rothia. In: Starr, Stolp, Triiper, Balows, Schlegel, eds. The Prokaryotes: a handbook on habitats, isolation, and identification of bacteria. Berlin: SpringerVerlag, 1981:1923-50. 19. M&tire FC, Vatter AE, Baros JB, Arnold J. Mechanism of coaggregation between Actinomyces viscosus T14 V and Streptococcus sang& 34. Infect Immun 1978;21:978-88. 20. Ellen RP, Walker DL, Chan KH. Association of long surface appendages with adherence-related functions of the grampositive species Actinomyces naeslundii. J Bacterial 1978; 134:1171-5.

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 76, Number I

2 1. Bourgeau G, McBride BC. Dextran-mediated interbacterial aggregation between dextran synthetizing streptococci and Actinomyces viscosus. Infect Immun 1976; 13: 1228-34. 22. Johnson JL, Holdeman LV. Bacteroides intermedius comb nov and descriptions of Bacteroides corporis sp nov and Bacteroides levii sp nov. Jnt J Syst Bacterial 1983;33:15-25, 675. 23. Holdeman LV, KelYy KW, Moore WEC. Anaerobic gramnegative straight, curved and helical rods Genus Bacteroides. In: Krieg NR, Holt JG, eds. Bergeys manual of systematic bacteriology, vol. 1. Baltimore: Williams and Wilkins, 1984; 604-3 1.

Perez et al.

103

24. Gibbons RJ, Van Home J. Selective bacterial adherence to oral epithelial surfaces and its role as an ecological determinant. Infect Immun 1971;3:567-73. Reprint requests: Fabienne Perez, DCD Service dOdontologie conservatrice-Endodontie Faculti: de Chirurgie Dentaire 3 Chemin des Maraichers 3 1062-Toulouse Cedex France

CORRECTION In the article Narcotic receptor blockade and its effect on the analgesic response to placebo and ibuprofen after oral surgery by Hersch et al., which appeared in the May 1993 issue (Oral Surg Oral Med Oral Path01 1993;75:539-46), the name, Andrei Barasch, DMD, should have been included as a coau-