Journal of Dental Research

http://jdr.sagepub.com/

Validation of an in vitro Biofilm Model of Supragingival Plaque

B. Guggenheim, E. Giertsen, P. Schüpbach and S. Shapiro
J DENT RES 2001 80: 363
DOI: 10.1177/00220345010800011201

The online version of this article can be found at:

http://jdr.sagepub.com/content/80/1/363

Published by:

http://www.sagepublications.com

On behalf of:
International and American Associations for Dental Research

Additional services and information for Journal of Dental Research can be found at:

Email Alerts: http://jdr.sagepub.com/cgi/alerts

Subscriptions: http://jdr.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://jdr.sagepub.com/content/80/1/363.refs.html

Downloaded from jdr.sagepub.com by guest on May 23, 2011 For personal use only. No other uses without permission.
B. Guggenheimrl*, E. Giertsen2,
P. Schupbachl, and S. Shapirol Validation of an in vitro Biofilm
lInstitute for Oral Microbiology and General Immunology,
Center for Dentistry, Oral Medicine, and Maxillofacial Model of Supragingival Plaque
Surgery, University of Zurich, Plattenstrasse 11, Postfach,
CH-8028 Ziirich 7, Switzerland; and 2Department of
Odontology-Cariology, Faculty of Dentistry, University of
Bergen, Norway; *corresponding author,
guggenhe(zzmk.unizh.ch
J Dent Res 80(1 ):363-370, 2001

ABSTRACT INTRODUCTION
The study of biofilm structure and function The existence of micro-organisms as the polyspecies consortium known as
mandates the use of model systems for which a I "dental plaque" has profound implications in the etiology of caries,
host of environmental variables can be rigorously gingivitis, and periodontal diseases. For more than a century, dental plaque
controlled. We describe a model of supragingival has been a major target of chemoprophylaxis (Miller, 1890) and, with regard
plaque containing Actinomyces naeslundii, to periodontal diseases, of chemotherapy (van Winkelhoff and Winkel,
Veillonella dispar, Fusobacterium nucleatum, 1997). In view of the species heterogeneity and structural complexity of
Streptococcus sobrinus, and Streptococcus oralis dental plaque, simplified models reflecting the most salient features of these
wherein cells are cultivated anaerobically in a ecosystems are required to clarify the physicochemical and metabolic
saliva-based medium on hydroxyapatite discs interactions underlying the development and maintenance of this biofilm.
coated with a salivary pellicle, with material and Various multispecies models of dental plaque have been described and
pieces of apparatus common to all microbiology applied to problems of clinical relevance, most notably biofilm permeability
laboratories. After 0.5 hr, 16.5 hrs, 40.5 hrs, and and chemical control of plaque. These systems usually consist either of flow
64.5 hrs, the composition of adherent biofilms was cells (Christersson et al., 1987; Sjollema et al., 1989; Larsen and Fiehn,
analyzed by culture techniques, live/dead 1995) or of chemostats modified to allow for insertion and removal of
fluorescence staining, and confocal laser scanning colonizable surfaces (Herles et al., 1994; Bradshaw et al., 1996; Kinniment
microscopy. Repeated independent trials et al., 1996; Bowden, 1999). While these devices have contributed to our
demonstrated the repeatability of biofilm understanding of microbial adhesion and biofilm formation, their use has
formation after 40.5 hrs and 64.5 hrs. Brief certain drawbacks. They can be cumbersome to construct and/or difficult to
exposures of biofilms to chlorhexidine or maintain over long periods of time. Since clearance of pulsed substances is a
Triclosan produced losses in viability similar to function of flow rate and volume, chemostats operating with low flow rates
those observed in vivo. This biofilm model should and relatively large volumes can have quite long mean residence times,
prove very useful for pre-clinical testing of rendering them impractical for studies of selected compounds with short-
prospective anti-plaque agents at clinically term exposures, as is common in oral hygiene procedures. Moreover,
relevant concentrations. systems with working volumes of more than a few milliliters preclude the
use of media constituted from natural substrates such as saliva.
KEY WORDS: biofilm model, dental plaque To circumvent the aforementioned problems, we have developed a
bacteria, confocal laser scanning microscopy, model of supragingival plaque which is technically simple to prepare,
chlorhexidine, triclosan. maintain, and analyze, and which allows for the short-term exposure of the
biofilm to selected substances. The aim of this study was to demonstrate the
repeatability of this dental plaque model and its applicability for the testing
of antimicrobial compounds.

MATERIALS & METHODS

Strains and Inoculum Preparation
Actinomyces naeslundii OMZ745, Veillonella dispar OMZ493, Fusobacterium
nucleatum OMZ596, Streptococcus sobrinus OMZ176 (serotype d), and
Streptococcus oralis OMZ607 were obtained from the culture collection of the
Institute for Oral Microbiology and General Immunology, University of Zurich.
Strains were maintained by anaerobic incubation at 37°C on plates of Columbia
blood agar base (Oxoid, Ltd., Basingstoke, Hamps., UK) supplemented with 5%
(v/v) hemolyzed human blood (CBA). Loopfuls of CBA-grown cells were
inoculated into filter-sterilized fluid universal medium (Gmiir and Guggenheim,
1983) supplemented with 67 mmol/L Sorensen's buffer, pH 7.2 ("modified fluid
Received May 17, 2000; Last revision August 17, 2000; universal medium", mFUM). For cultivation of V dispar, mFUM containing
Accepted September 13, 2000 0.10% (w/v) Na lactate was used.
Downloaded from jdr.sagepub.com by guest on May 23, 2011 For personal use only. No other uses without permission. 363
364 Guggenheim et al. J Dent Res 80(1) 2001
Pre-cultures (10 mL) in mFUM were incubated anaerobically Geigy AG, Basel, Switzerland) was added. The suspension was
at 37°C for 15 hrs; seed cultures of each species were prepared by shaken at room temperature for 1 hr, then steamed for 2 hrs. The
inoculation of mFUM (10 mL) with aliquots (200 FL, or 500 jL resulting mixture was shaken ovemight at room temperature, then
for F. nucleatum) of pre-culture and anaerobic incubation at 37°C sonified (60'C) in a water bath. Shaking was resumed, and after 6
for ca. 7 hrs. Following microscopic verification of purity, seed days the suspension was diluted with distilled water (- 75 mL) and
cultures were adjusted independently to OD550 1.0 ± 0.05 (' 107 passed through a 0.3-,um cellulose nitrate filter (Sartorius AG,
cells/mL) by dilution with fresh mFUM (or, in the case of F. Gottingen, Germany). The filtrate was shell-frozen and
nucleatum, by enrichment with mFUM-grown cells). Aliquots (1 lyophilized. By 'H-NMR (300 MHz), the lyophilizate was found to
mL) of each density-adjusted culture were combined and stored on contain 44 mol% of Triclosan (data not shown).
ice for ca. 6 hrs (vide infra).
Treatnent of Biofilms with Antimicrobial Agents
Processed Saliva To evaluate the inhibitory action of chlorhexidine digluconate
Whole, unstimulated saliva was obtained for 1 hr per day over [CHX; 20% (w/v) stock solution, Sigma Chemical Co., St. Louis,
several days from volunteers (with their informed consent) at least MO, USA] and Triclosan toward biofilms, we transferred HA discs
1.5 hrs after eating, drinking, or toothcleaning. At all times, saliva from their wells at 16.5 hrs to new wells containing either
samples, collected in sterile 50-mL polypropylene tubes, were suspensions of Triclosan in 10% (v/v) aqueous ethanol or aqueous
either chilled in an ice bath or frozen at -20°C. When a total of ca. solutions of CHX or Triclosanl2-HP-B3-CyD (1 mL, room
500 mL saliva had been collected, it was pooled and centrifuged temperature); physiological saline was used as a negative control.
(30 min, 4°C, 27,000 x g), and the supemate was pasteurized Following a one-minute exposure, discs were rinsed by being
(60'C, 30 min) and re-centrifuged in sterile bottles; the resulting double-dipped sequentially in 3 x 2-mL portions of fresh
supemate was dispensed into sterile 50-mL polypropylene tubes physiological saline (immersion time per dip, 10 sec), transferred
and stored at -200C. We assessed the efficacy of pasteurization by to fresh saliva (800 gL) + mFUM (800 ,uL) containing 0.15%
plating processed saliva samples onto CBA; after 72 hrs at 370C, (w/v) glucose + 0.15% (w/v) sucrose, and incubated anaerobically
no CFUs were observed on either aerobically or anaerobically at 37°C. This procedure was repeated 4 hrs and 8 hrs later, except
incubated plates. that at these later times incubation media were not replaced.
Following the third one-minute exposure, discs were incubated
Preparation of Biofilms undisturbed for 16 hrs. At the end of this time (biofilm age, 40.5
Biofilms were grown on sintered circular hydroxyapatite (HA) hrs), the three-fold exposure of discs to antimicrobial agent at four-
discs (Hy-Apatite discs; Sudco B.V., Landgraaf, The Netherlands). hour intervals was repeated, the medium being replaced following
New discs were 10.6 mm in diameter and 1.3 mm high, and could the first immersion. Final exposure to antimicrobial agent occurred
be re-used after being washed in tap water, with their surfaces when the biofilms were 48.5 hrs old; they were incubated an
refreshed by grinding off of the upper 20 gm with silicon carbide additional 16 hrs and harvested at the age of 64.5 hrs.
paper (500-grit size; Struers A/S, R0dovre, Denmark), then
autoclaved for 20 min. To allow for formation of a salivary pellicle, Harvesting the Biofilms
we placed each HA disc in a well of a sterile 24-well polystyrene At the end of their designated incubation times, HA discs not
cell culture plate (Nunc A/S, Roskilde, Denmark), incubated with destined for confocal laser scanning microscopy (CLSM) were dip-
saliva (for 4 hrs, gently shaken, at room temperature). Saliva was washed twice, then swirled gently in physiological saline to
aspirated from each well and replaced with a mixture of saliva (800 remove loosely adherent bacteria. To harvest adherent cells, we
gL) + mFUM (800 gL) containing 0.3% (w/v) glucose [final used either of two procedures:
glucose concentration, 0.15% (w/v)], and the wells were inoculated (1) Each disc was placed in a sterile plastic Petri dish, and
with the pooled oral bacterial species (200 gL). disc surfaces were scraped by means of sterile Perio Soft-Scalers®
Cell culture plates were incubated anaerobically at 37°C for [Hawe Neos Dental Dr. H. v. Weissenfluh AG, Bioggio (TI),
either 0.5 hr, 16.5 hrs, 40.5 hrs, or 64.5 hrs. Medium was renewed Switzerland]. The surface of the scraped disc and the Petri plate
after 16.5 hrs at 24-hour intervals (one medium replacement for were rinsed with physiological saline (10 x 100 gL), the pooled
40.5-hour biofilms, two for 64.5-hour biofilms) by the aspiration of washings were adjusted to a final volume of 1 mL, and the cell
spent medium from wells and the addition of fresh medium suspension was vortexed vigorously for 2 min and sonified (5 sec,
composed of saliva (800 gL) + mFUM (800 gL) containing 0.15% 30 W, room temperature).
(w/v) glucose + 0.15% (w/v) sucrose in lieu of 0.3% (w/v) glucose (2) Each disc was transferred to a sterile 50-mL
[final sugar concentrations: 0.075% (w/v) glucose + 0.075% (w/v) polypropylene tube containing physiological saline (1 mL, room
sucrose]. Following medium replacement, plates were retumed to temperature) and vortexed vigorously for 2 min, following which
the anaerobic incubator. fluids were transferred to sterile 6-mL polystyrene tubes and
sonified.
Triclosan I 2-Hydroxypropyl-B-Cyclodextrin
Inclusion Complex Examination of Harvested Cells
2-Hydroxypropyl-B-cyclodextrin (2-HP-B-CyD; 500 mg, 0.36 Aliquots (25 gL) of sonified cells were deposited onto glass
mmol; C*Cavitron 82005, Cerestar USA, Inc., Hammond, IN, microscope slides containing LIVE/DEAD® BacLightTm Bacterial
USA; ds = 4.5) and (hydroxypropyl)methyl cellulose (Loftsson et Viability Kit solution (0.5 gL; Molecular Probes B. V., Leiden,
al., 1994; Loftsson, 1998) [25 mg; Fluka Chemie AG, Buchs (SG), The Netherlands) and incubated in the dark for 15 min at room
Switzerland; cat. no. 56340, mw 9 x 104] were dissolved in
-
temperature. Three counts of 100 bacteria each at different sites on
distilled water (10 mL), and Triclosan (500 mg, 1.73 mmol; CIBA- the slide were obtained (eyepiece, xlO; objective, xlOO; tubus
Downloaded from jdr.sagepub.com by guest on May 23, 2011 For personal use only. No other uses without permission.
J Dent Res 80(1) 2001 An in vitro Dental Plaque Model 365
factor, 12.5) with the use of a Leitz Dialux 22 microscope [Leica procedure for the detection of statistically deviating results well-
Mikroskopie Systeme AG, Glattbrugg (ZH), Switzerland] suited to problems of replicability involving small numbers of
equipped with filters for fluorescein isothiocyanate fluorescence observations; further details on this method are to be found in
and an Osram 50W high-pressure Hg lamp. Kelly (1990) and Burke (1998). Coefficients of variation (CVs) for
In addition to live/dead staining, serial dilutions (10-2_10-5) of each bacterial population at 40.5 hrs and 64.5 hrs also were
sonified cells were prepared in physiological saline and aliquots calculated.
(50 gL) spirally plated (Spiral System., Model D, Spiral Systems,
Inc., Cincinnati, OH, USA) onto plates of CBA, Mitis Salivarius RESULTS
Agar (MSA; Difco Laboratories, Inc., Detroit, MI, USA) + 0.001%
(w/v) Na tellurite, and Fastidious Anaerobe Agar (FAA; Medium
Biologische Arbeitsgemeinschaft GmbH, Lich, Germany) Preliminary experiments indicated that supplemental phosphate
supplemented with erythromycin (1 mg/L; Sigma), norfloxacin (1 was required to prevent growth-inhibitory declines in pH. In
mg/L; Sigma), and vancomycin [4 mg/L; Vancocing CG, Eli Lilly the absence of added phosphate, the pH of the biofilm
(Suisse) S. A., Geneve, Switzerland]. CBA and FAA plates were incubations fell below 5.0 within a matter of hours,
incubated anaerobically at 37°C, while MSA plates were incubated necessitating frequent changes of medium; whereas in the
at 37°C in an anaerobic environment enriched with 10% (v/v) CO2. presence of 67 mmol/L phosphate, the pH never fell below 6.5
After 72 hrs, colony-forming units (CFUs) were counted with the during this same period, despite the presence of sugar in the
aid of a stereomicroscope. Differentiation of the five species was medium.
achieved by observation of colonial morphology in conjunction
with microscopic examination of cells from selected colonies. CLSM of Intact Biofilms
CBA plates were used to obtain total bacterial counts and to By 30 min, saliva-coated surfaces of HA discs were only
enumerate A. naeslundii and V dispar; differential counting of S. sparsely populated by bacteria, mostly single cells or short
sobrinus and S. oralis was accomplished with the use of MSA, streptococcal chains attached directly to the pellicle, with the
whereas selective growth of F. nucleatum was achieved with FAA. occasional small clump (Fig. la). By 16.5 hrs, the biofilm
Confocal Laser Scanning Microscopy (CLSM) contained a substantially greater cell population, arranged as
diffuse galaxies of microcolonies against a background of
Non-invasive confocal imaging of fully hydrated biofilms was adherent single cells (Fig. lb). Sagittal (xz-, yz-) sections (Fig.
accomplished by means of an Axioplan light microscope (Carl lb1) revealed a very shallow biofilm extending 15-20 gm
Zeiss AG, Oberkochen, Germany) fitted with water-immersion above the disc surface, with narrow prominences and plateaux
"dipping lenses" (x25, x63) and an Ar-Kr laser (MRC-600, Bio- of cells permeated by unstained regions which may correspond
Rad Microscopy Divison, Hemel Hempstead, UK). Specimens to cell-free regions of exopolysaccharide and/or the
were stained with LIVE/DEAD® BacLightTm Bacterial Viability canalization reportedly typical of biofilms (Stoodley et al.,
Kit solution, following which they were stored in physiological 1999). By 40.5 hrs, there was a nearly confluent presence of
saline for a maximum of 15 min before being covered with cells across the surfaces of HA discs (Fig. lc), composed
distilled water and viewed in fluorescence mode. An excitation mostly of discrete cells and small clusterings of cells
wavelength of 488 nm was used, and all light emitted above 500 interspersed with lacunae. Sagittal sections (Fig. Ic1) show
nm was collected. Biofilm structure was analyzed as a series of cellular prominences separated by gaps of 1-1.5 jm. The 40.5-
horizontal optodigital sections, each 1.5 gm thick; intervening gaps hour biofilm consisted of a cell layer (15-25 jm high) firmly
between horizontal sections ranged from 0.5 jim to 5.0 jtm over adherent to the substratum, overlaid by a mass (35-55 jim high)
the entire height (z-axis) of the biofilm. Digital images were of loosely attached bacteria. By 64.5 hrs, the biofilm was a
processed with the use of Imaris® 3.0 software (Bitplane AG, confluent mixture of some single cells and an abundance of
Zurich, Switzerland). galaxies of cells, within which was seen the occasional small
lacuna (Fig. Id). The firmly adherent stratum of the biofilm had
Statistics achieved a height of 20-30 jm, while the overlying loosely
The repeatability of biofilm populations at 40.5 hrs and 64.5 hrs arranged cell mass had an irregular appearance, with
was examined in six different populations per biofilm, viz., that of prominences extending up to 140 jm above the disc surface.
each constituent species (A. naeslundii, V. dispar, F. nucleatum, S. Sagittal views showed the biofilm still punctuated with
sobrinus, S. oralis) plus that of the biofilm as a whole (total viable unstained regions (Fig. 1di).
counts on CBA). CFUs per population for triplicate discs
inoculated with identical polyspecies suspensions were averaged Harvesting the Biofilms
(cf. Dean and Dixon, 1951) and subjected to logarithmic Cell recoveries on CBA for discs harvested at 64.5 hrs were 5.3
transformation. Approximately normal distributions of logl0 CFU (± 2.7) x 108 CFUs per scratched disc and 12.3 (± 2.5) x 108
values per population per time point were confirmed by means of CFUs per vortexed disc (averages of triplicate harvestings ±
the Ryan-Joiner test (Ryan and Joiner, 1976), as implemented in SD). Therefore, either of these harvesting procedures is
MINITAB release 11.21. Each population per time point was adequate for near-quantitative recovery of biofilm cells from
checked for the presence of outliers at the 0.05 significance level HA discs.
by the method of Grubbs (1969, 1972), such that the null Aliquots of cells recovered from different biofilms were
hypothesis (H.) of no outliers per population per time point was live/dead-stained and examined by fluorescence microscopy; in
tested against the one-sided alternative (HA) of there being at least this way, the population viability at 40.5 hrs was found to be
one outlier per population. Grubbs' test is a sensitive univariate 84.9% ± 3.6% (biofilm sample size, n = 27; range = 70%-
Downloaded from jdr.sagepub.com by guest on May 23, 2011 For personal use only. No other uses without permission.
366 Guggenheim et a/. J Dent Res 80(1) 2001
94%); and at 64.5 hrs, 71.1% i 11.8% (biofilm sample size, n = 27;
range = 45%-87%).
Biofilm Formation and Biofilm Growth
Growth profiles for each species comprising the biofiln (Fig. 2a)
resembled those observed in batch cultures, wherein CFUs
increased most rapidly between 0.5 hr and 16.5 hrs, decelerated
between 16.5 hrs and 40.5 hrs, and remained stable or declined
slightly between 40.5 hrs and 64.5 hrs. Average CFU values on
CBA matched very closely [r2(d.) - 0.976] the sums of the five
component species cultured on differential media at each of the
four time points (Fig. 2b), indicating that the differential media
used distinguish both qualitatively and quantitatively between
the structurally and physiologically disparate bacteria
constituting the biofilm.
The repeatability of the biofilm-generating procedure was
assessed for 40.5-hour and 64.5-hour biofilms. In six different
trials (on six different occasions), trios of biofilms were formed,
each time with use of a freshly prepared five-species inoculum.
At the designated times, biofilms were harvested, diluted, and
plated, and average log1o CFUs per population were calculated
for each trial. The resulting sets of six values per biofilm
population per time point were subjected to Grubbs' test (Grubbs,
1969; Grubbs and Beck, 1972), and in all cases the calculated G-
values were found to be less than the critical G-values for n = 6 at
the 95% confidence level (Table). Since no outliers could be
detected per population per time point, each population
constituted a homogeneous cluster, in which case the procedure
used for generating 40.5-hour and 64.5-hour biofilms is
repeatable at the 0.05 level of significance. This conclusion is
reinforced by the mean value of the CVs for all biofilm
populations (average CV = 4.5%; range = 1.1-I 1.0%), which
compares favorably with CVs obtained for other CFU-based
studies of methodological repeatability (Fowler et al., 1978; Chen
et al., 1989; Power et al., 1998).
Effects of Antimicrobial Agents on Biofilm Cells
Responses of the biofilm model to treatments with Triclosan and
CHX are shown in Figs. 3 and 4, respectively. Repeated brief
exposures of discs to 0.2% (w/v) Triclosan (6.9 mmol/L)
suspended in 10% aqueous ethanol resulted in decreases in
recovered CFUs by >~2 orders of magnitude. Triclosan 2-HP-B-
CyD, which is completely soluble in water without recourse to
detergents or emulsifiers, and which has a pleasant, sliglhtly sweet
taste, was tested against biofilms at a concentration corresponding
to 2.5% (w/v) of Triclosan ( 86 mmol/L). At this concentration,
Triclosanl2-HP-B-CyD reduced CFUs by 3 to 4 orders of
magnitude. F. nucleatun2 was most sensitive to the inhibitory
action of Triclosan; treatment of biofilins with either "free" or
complexed Triclosan completely eradicated F. nucleatunm from the
bacterial population (limit of detectability, 20 CFU per species per
disc). Treatment of biofilms with either IO% aqueous ethanol or 2-
HP-13-CyD had no effect on the bacterial populations of the
biofilm compared with controls treated with physiological saline
(data not shown).

Figure 1. Representative CLSM images of biofilms at (a) 0.5 hrs, (b) 16.5
hrs, (c) 40.5 hrs, and (d) 64.5 hrs. (a-d) Optodigital horizontal (xy) sections
through biofilms. (b1, cl, d,) Optodigital sagittal (xz) sections. To enhance
resolution, we stretched the sagittal images vertically by a factor of 4. Length
barForinpersonal
Downloaded from jdr.sagepub.com by guest on May 23, 2011 (a) represents 10uses
use only. No other length
pm;without bars in (b-d) represent 100 pm.
permission.
J Dent Res 80(1) 2001 An in vitro Dental Plaque Model 367

A 9

-
-- -
-
8-
7
EE
b~
6
U
5-
'. 'M

..S'
y

ML
0 4- EdS
co 3-
2-
1
o 10 20 30 40 50 60 70 An Vd Fn Ss So tCFU
incubation time (h)
9-
Figure 2a. Average CFUs recovered per species per time point per disc.
0 A. naeslundii; + V. dispar; * F. nucleatum; o, S. sobrinus; A S. oralis. 8-
E3~~
E,
.)
rj 7-
B 6-
0
5-
*>D 8-
4-
'%

SF6-
0.
7- / 0 3-
2-7
1-
LFLZ ~X w

D: 5- If
An Vd Fn Ss So tCFU
UO 4
Figure 3. Boxplots depicting the effect of a 0.2% (w/v) suspension of
-3 I3 Triclosan (upper Fig.) and an aqueous solution of Triclosan 2-HP-13-
0 10 20 30 40 50 60 70 CyD inclusion complex containing 2.5%c (w/v) of Triclosan (lower
Fig.) on species comprising the biofilm. Hollow rectangles,
incubation time (h) physiological saline controls; grey rectangles, Triclosan-treated. An,
A. naeslundii; Vd, V. dispar; Fn, F. nucleatum; Ss, S. sobrinus; So,
Figure 2b. Comparison of CFUs obtained by summing individual counts S. oralis; tCFU, total CBA CFUs. Each box represents a distribution
of all five biofilm species (o) with total CFUs by plating on non-selective for 3 pairs of independent measurements; horizontal bars within
medium (0) as a function of incubation time. boxes are median values.

There was a clear dose-response effect of CHX toward the tooth surfaces in the absence of high amounts of sucrose.
biofilim (Fig. 4). A reduction in CFUs for each species was However, recovered CFUs for V. dispar of- 104/disc after
apparent at 0.1% (w/v) CHX (1.1 mmol/L), but the range of a 30-minute exposure to the mixed inoculumn were similar
recovered CFUs tended to be broad. Slightly raising the CHX to those found for A. naeslundii and S. oralis, whereas
concentration to 0.12% (w/v) (1.3 mmol/L)
brought about dramatic reductions (- 5 orders Table. G-values Obtained per Population per Time Point
of magnitude) in recovered CFUs, with much
narrower ranges of CFU values. When the Bacteria 40.5 hrs 64.5 hrs
concentration of CHX was increased to 0.2% Gla G2 G3 GI G2 G3
(w/v) (2.2 mmol/L), no viable cells were
recoverable from the discs. A. naeslundii 1.732 3.00 0.0677 1.175 2.26 0.4877
V. dispar 1.519 2.89 0.1634 1.285 2.56 0.3445
DISCUSSION F. nucleatum 1.508 2.16 0.4217 1.165 2.16 0.5296
The bacteria chosen for incorporation into S. sobrinus 1.527 2.91 0.1502 1.280 2.31 0.4586
our biofilm model were five species S. oralis 1.635 2.82 0.1731 1.406 2.39 0.4008
commonly encountered in supragingival Total CBA CFUs 1.503 2.45 0.3531 1.443 2.37 0.3970
plaque. It has been claimed (Marsh and a The larger of the polar values has been listed for each population. For an inter-trial
Martin, 1992) that Veillonella spp. and G,
size n = 6, the critical values of G1, G2, and G3 at the 95% confidence level are 1.822,
mutans streptococci have low adherence to 3.01, and 0.9436, respectively (Burke, 1 998).
Downloaded from jdr.sagepub.com by guest on May 23, 2011 For personal use only. No other uses without permission.
368 Guggenheim et al. J Dent Res 80(1) 2001

9 without recourse to quantitative obfuscation.

Besides its simplicity of design, a great advantage of our
88 biofilm model lies in the fact that additives can be applied to
and removed from the system virtually instantaneously.
While pulsing chemostats operating at low dilution rates
leads to skewed impulse functions with very steep rise times
and much more shallow settlement times, dipping disc-bound
biofilms into test substances and then into rinse media
produces exposure kinetics akin to a Heaviside unit-step
function. Thus, our biofilm model allows for much more
controlled substance exposure times (issues of substantivity
notwithstanding) resembling those encountered during
mouthrinsing. The utility of this feature of our biofilm model
is illustrated in experiments with Triclosan and CHX.
An Vd Fn Ss So tCFU Minimal inhibitory concentrations obtained by
conventional procedures with the use of monospecies
Figure 4. Boxplots depicting the effects of increasing concentrations of cultures in which the test substance is continuously present at
CHX on species comprising the biofilm. Each biofilm was subjected to a fixed concentration from the time of inoculation are of
four treatments: 1 st (hollow) rectangles, physiological saline control; minimal clinical relevance, since sessile bacteria are
2nd (grey) rectangles, 0.10% (w/v) CHX; 3rd (black) rectangles, 0.1 2%
(w/v) CHX; and 4th rectangles, 0.2% (w/v) CHX. Abscissa markedly more resistant to antimicrobials (l10-104-fold) than
abbreviations are as per Fig. 3. Each box represents a distribution for planktonic cells (Shapiro and Guggenheim, 1998); moreover,
3 pairs of independent measurements; horizontal bars within boxes are exposure times during oral hygiene procedures are very
median values. short. The concentrations of Triclosan and CHX inhibiting
bacterial growth in the biofilm model were of the same order
of magnitude as those required for plaque inhibition in VivlO
recovered CFUs for S. sobrinus were I O0/disc. The latter
-
(vide infra).
observation supports earlier work by van Houte et al. It is clear (Fig. 3) that microcrystalline suspensions of
(1976) that sucrose is not required for adherence to and Triclosan are potent growth inhibitors of all of the species
colonization of tooth surfaces by mutans streptococci. The constituting the biofilm model. The antibacterial effect of
recovery of total CFUs from biofilms was similar to that Triclosan toward the biofilm was less pronounced than that
observed for one-week-old sucrose-stimulated buccal of CHX (Fig. 4), in agreement with clinical short-term
plaques formed on enamel slabs in situ. The average log,( plaque inhibition studies (Jenkins et al., 1991, 1994).
CFUs per mm2 of HA surface were 6.30 after 40.5 hrs and Mouthrinses with 0.2% Triclosan exerted a modest plaque-
5.89 after 64.5 hrs, compared with 6.43 per mm2 of enamel inhibitory effect (Jenkins et al., 1991). The antibacterial
surface for the in situ appliances (Giertsen et al., 2000). effects of cyclodextrin-solubilized Triclosan were somewhat
Devices commonly used for the study of biofilm greater than those of the Triclosan suspension, though the
formation-e.g., parallel plate flow chambers (van Kooten effective concentration of Triclosan in Triclosanl2-HP-8-
et al., 1992), Rototorques (Trulear and Characklis, 1982), or CyD was about an order of magnitude higher than that in the
constant depth film fermenters (Peters and Wimpenny, suspension. Complexes of hydrophobic species with
1988) utilize fluid flow or scraper blades to generate cyclodextrins reputedly have enhanced bioavailabilities
continuous detachment forces; whereas biofilms cultivated (Aithal et al., 1995; Ashall et al., 1995), though preliminary
in cell culture plates were subjected only intermittently to experiments with planktonic cells of A. naesllunidii and S.
such forces in the course of dip-washing or gentle swirling. sobrinus (cf Shapiro et al., 1994) indicated that MICs for the
It has been estimated (Bos et al., 1999) that the detachment Triclosanl2-HP-B-CyD inclusion complex are about an order
force exerted by gentle swirling in dilute aqueous milieux is of magnitude higher than those for uncomplexed Triclosan
10-5- 10` dyne/cell, while that exerted by dip-washing (S. Shapiro, unpublished observations).
(passage of an HA disc through liquid-air interface)
a is 10- Commercial preparations of CHX for oral disinfection
2 dyne/cell. Christersson et al. (1988), partitioning micro- contain between 0.1% and 0.2% CHX. There was a clear
organisms between saliva and germanium surfaces, reported dose-response effect of 0. 1%, 0. 12%, and 0.2% CHX toward
significant detachment at drag and lift forces of 10-' dyne/cell the biofilm (Fig. 4), comparable with plaque inhibition
and 10-'' dyne/cell, respectively. However, under our obtained in short-term mouthrinsing experiments with
experimental conditions, the detachment forces exerted by 0.01%-0.2% CHX (Jenkins et al., 1994). At 0.2%, CHX
gentle swirling and dip-washing sufficed to dislodge only the reduced the viability of the in vitro biofilm to zero,
loose halo of bacteria surrounding the biofilm but not the concordant with the report that a single mouthrinsing with
firmly attached cells from disc surfaces. While descriptions 0.14% chlorhexidine diacetate (2.2 mmol/L) reduced total
such as "gentle rinsing to remove non-adherent or loosely CFUs in plaque by 90% (compared with a 22% reduction for
adherent micro-organisms" have been criticized on the a single mouthrinsing with 0.030% chlorhexidine diacetate)
grounds of quantitative imprecision (Bos et al., 1999), the (Giertsen and Scheie, 1995).
fact remains that such descriptions convey methodological This underscores one of the great strengths of this new
details in a concise and readily understandable manner oral biofilm model, viz., that its simplicity allows for rapid
Downloaded from jdr.sagepub.com by guest on May 23, 2011 For personal use only. No other uses without permission.
J Dent Res 80(1) 2001 An in vitro Dental Plaque Model 369
and convenient manipulation of numerous intrinsic variation in doing the standard plate count described in
(species, medium) and extrinsic (feeding regimens, Standard Methods for the Examination of Dairy Products. J
additives) parameters such that the biofilm can be fine- Food Protect 41:4-7.
tuned to suit particular experimental needs. Despite obvious Giertsen E, Scheie AAa (1995). Effects of chlorhexidine-fluoride
differences between our in vitro model and supergingival mouthrinses on viability, acidogenic potential, and glycolytic
plaque in vivo in terms of species composition, exposure to profile of established dental plaque. Caries Res 29:181-187.
cellular and salivary defense mechanisms, feeding regimen Giertsen E, Guggenheim B, Thurnheer T, Gmuir R (2000).
and consequent growth conditions, hydrodynamic and Microbiological aspects of an in situ model to study effects of
abrasive forces, etc., the model still was able to simulate antimicrobial agents on dental plaque ecology. Eur J Oral Sci
reasonably the antibacterial effects of chlorhexidine and 108:403-411.
Triclosan observed in clinical studies. Therefore, this Gmur R, Guggenheim B (1983). Antigenic heterogeneity of
versatile model of oral biofilm should prove extremely Bacteroides intermedius as recognized by monoclonal
useful for studies of microbial adhesion and co-aggregation, antibodies. Infect Immun 42:459-470.
biofilm spatial organization, gene expression, and metabolic Grubbs FE (1969). Procedure for detecting outlying observations in
interplay, as well as for pre-clinical testing of prospective samples. Technometrics 1 1: 1-2 1.
anti-plaque agents. Grubbs FE, Beck G (1972). Extension of sample size and
percentage points for significance tests of outlying
ACKNOWLEDGMENTS observations. Technometrics 14:847-854.
We thank A. Meier for excellent technical assistance. The help Herles S, Olsen S, Afflitto J, Gaffar A (1994). Chemostat flow cell
of Dr. M. Hochli (Elektronenmikroskopisches Zentralla- system: an in vitro model for the evaluation of antiplaque
boratorium der Universitat Zurich, Switzerland) with CLSM is agents. J Dent Res 73:1748-1755.
gratefully acknowledged. This study was supported in part by Jenkins S, Addy M, Newcombe R (1991). Triclosan and sodium
Patentmedelsfonden for Odontologisk Profylaxforskning lauryl sulphate mouthrinses (II). Effects of 4-day plaque
(Sweden), Colgates Forskningsfond (Norway), and A/S Norsk regrowth. J Clin Periodontol 18:145-148.
Dental Depots Fond for Odontologisk Forskning (Bergen). Jenkins S, Addy M, Newcombe RG (1994). Dose response of
chlorhexidine against plaque and comparison with triclosan. J
Clin Periodontol 21:250-255.
REFERENCES Kelly PC (1990). Outlier detection in collaborative studies. JAssoc
Aithal KS, Udupa N, Varma BRR, Kurian R (1995). Antibacterial Off Anal Chem 73:58-64.
activity of ciprofloxacin-betacyclodextrin complex and its Kinniment SL, Wimpenny JW, Adams D, Marsh PD (1996).
efficacy in dental implants. Indian JDent Res 6:123-128. Development of a steady-state oral microbial biofilm
Ashall PM, Kiernan BM, Russell NR (1995). Complexation of the community using the constant-depth film fermenter.
bacteriostat, triclosan, with -y-cyclodextrin. 1st World Meet. Microbiology 142:631-638.
Pharm., Biopharm. Pharm. Technol. Chatenay Malabry, France: Larsen T, Fiehn NE (1995). Development of a flow method for
APGI, pp. 605-606. susceptibility testing of oral biofilms in vitro. APMIS 103:339-
Bos R, van der Mei HC, Busscher HJ (1999). Physico-chemistry of 344.
initial microbial adhesive interactions its mechanisms and Loftsson T (1998). Increasing the cyclodextrin complexation of
methods for study. FEMS Microbiol Rev 23:179-230. drugs and drug bioavailability through addition of water-
Bowden GH (1999). Controlled environment model for soluble polymers. Die Pharmazie 53:733-740.
accumulation of biofilms of oral bacteria. Methods Enzymol Loftsson T, Fri6riksd6ttir H, Sigur6ard6ttir AM, Ueda H (1994).
310:216-224. The effect of water-soluble polymers on drug-cyclodextrin
Bradshaw DJ, Marsh PD, Schilling KM, Cummins D (1996). A complexation. Int JPharm 1 10:169-177.
modified chemostat system to study the ecology of oral Marsh P, Martin M (1992). Oral microbiology. 3rd ed. London:
biofilms. JAppl Bacteriol 80:124-130. Chapman & Hall, pp. 72-74.
Burke S (1998). Missing values, outliers, robust statistics and non- Miller WD. The micro-organisms of the human mouth. [Unaltered
parametric methods. VAM Bull (1 9):22-27. reprint (1973) of the 1890 edition.] Basel, Switzerland: S.
Chen ZR, Cheng SG, An YQ, Cao XH (1989). Experimental Karger, pp. 225-237.
research on utilization of bioluminescent technique to Peters AC, Wimpenny JWT (1988). A constant depth laboratory
substitute the current viability count for BCG vaccine. Chin model film fermentor. Biotechnol Bioengin 32:263-270.
MedJ 102:906-910. Power CA, McEwen SA, Johnson RP, Shoukri MM, Rahn K,
Christersson CE, Fornalik MS, Baier RE, Glantz P-0 (1987). In Griffiths MW, et al. (1998). Repeatability of the PetrifilMTM
vitro attachment of oral microorganisms to solid surfaces: HEC test and agreement with a hydrophobic grid membrane
evaluation of a controlled flow method. Scand J Dent Res filtration method for the enumeration of Escherichia coli
95: 151-158. 01 57:H7 on beef carcasses. J Food Prot 61:402-408.
Christersson CE, Glantz P-O, Baier RE (1988). Role of Ryan TA Jr, Joiner BL (1976). Normal probability plots and tests for
temperature and shear forces on microbial detachment. normality. Technical Report. University Park, PA: Department
ScandJDentRes 96:91-98. of Statistics, The Pennsylvania State University, 12 pp.
Dean RB, Dixon WJ (1951). Simplified statistics for small numbers Shapiro S, Guggenheim B (1998). Chemoprophylaxis in the oral
of observations. Anal Chem 23:636-638. cavity: <<plus on change les choses, plus elles devraient
Fowler JL, Clark WS Jr, Foster JF, Hopkins A (1978). Analyst rester les memes>>. In: Oral biology at the turn of the

Downloaded from jdr.sagepub.com by guest on May 23, 2011 For personal use only. No other uses without permission.
370 Guggenheim et al. J Dent Res 80(1) 2001
century. Misconceptions, truths, challenges and prospects. Trulear MG, Characklis WG (1982). Dynamics of biofilm
Guggenheim B, Shapiro S, editors. Basel, Switzerland: S. formation. J Wat Pollut Contr Fed 54:1288-1301.
Karger, pp. 226-238. van Houte J, Upeslacis VN, Jordan HV, Skobe Z, Green DB (1976).
Shapiro S, Meier A, Guggenheim B (1994). The antimicrobial Role of sucrose in colonization of Streptococcus mutans in
activity of essential oils and essential oil components towards conventional Sprague-Dawley rats. JDent Res 55:202-215.
oral bacteria. Oral Microbiol Immunol 9:202-208. van Kooten TG, Schakenraad JM, van der Mei HC, Busscher HJ
Sjollema J, Busscher HJ, Weerkamp AH (1989). Real-time
enumeration of adhering microorganisms in a parallel plate (1992). Development and use of a parallel-plate flow chamber
flow cell using automated image analysis. J Microbiol Meth for studying cellular adhesion to solid surfaces. J Biomed
9:73-78. Mater Res 26:725-738.
Stoodley P, Boyle JD, DeBeer D, Lappin-Scott HM (1999). van Winkelhoff AJ, Winkel EG (1997). Systemic antibiotic therapy
Evolving perspectives of biofilm structure. Biofouling 14:75-90. in severe periodontitis. Curr Opin Periodontol 4:35-40.

Downloaded from jdr.sagepub.com by guest on May 23, 2011 For personal use only. No other uses without permission.