J Antimicrob Chemother 2012; 67: 1198 – 1206

doi:10.1093/jac/dks002 Advance Access publication 1 February 2012

Synergistic antibacterial efficacy of early combination treatment

with tobramycin and quorum-sensing inhibitors against
Pseudomonas aeruginosa in an intraperitoneal foreign-body
infection mouse model
Louise D. Christensen1, Maria van Gennip1, Tim H. Jakobsen1, Morten Alhede1, Hans Petter Hougen2, Niels Høiby1,3,
Thomas Bjarnsholt1,3 and Michael Givskov1,4*

Downloaded from https://academic.oup.com/jac/article/67/5/1198/979906 by guest on 20 August 2020

1
Department of International Health, Immunology and Microbiology, Faculty of Health Sciences, University of Copenhagen, DK-2200
Copenhagen, Denmark; 2Department of Forensic Medicine, Faculty of Health Sciences, University of Copenhagen, DK-2200 Copenhagen,
Denmark; 3Department of Clinical Microbiology, Rigshospitalet, DK-2100 Copenhagen, Denmark; 4Singapore Centre on Environmental
Life Sciences Engineering, Nanyang Technological University, Singapore 639798, Singapore

*Corresponding author. Tel: +45-35-32-78-55; E-mail: mgivskov@sund.ku.dk

Received 26 September 2011; returned 20 November 2011; revised 20 December 2011; accepted 27 December 2011

Objectives: Quorum sensing (QS)-deficient Pseudomonas aeruginosa biofilms formed in vitro are more suscep-
tible to tobramycin than QS-proficient P. aeruginosa biofilms, and combination treatment with a QS inhibitor
(QSI) and tobramycin shows synergistic effects on the killing of in vitro biofilms. We extended these results
to an in vivo P. aeruginosa foreign-body biofilm model. The effect of treatment initiated prophylactically was
compared with treatment initiated 11 days post-insertion.
Methods: Silicone tube implants pre-colonized with wild-type P. aeruginosa were inserted into the peritoneal
cavity of BALB/c mice. Mice were treated with intraperitoneal or subcutaneous injections of the QSIs furanone
C-30, ajoene or horseradish juice extract in combination with tobramycin. Mice were euthanized on day 1, 2, 3
or 14 post-infection for the estimation of quantitative bacteriology, histopathology and cytokine measurements.
Results: Combination treatment of P. aeruginosa resulted in a significantly lower cfu per implant as compared
with the placebo groups for all QSIs tested. For early-initiated treatment, a significant difference in clearing was
also observed between the combination group and the single-treatment groups, and between the placebo
group and the single-treatment groups. In one case a significant difference in clearing was found between
the two single-treatment groups.
Conclusions: Synergistic antimicrobial efficacy could be achieved when treating mice with both a QSI and tobra-
mycin, resulting in an increased clearance of P. aeruginosa in a foreign-body infection model. Our study highlights
the important prospects in developing an early combinatory treatment strategy for chronic infections.

Keywords: biofilm, in vivo, implant model, antibiotics

Introduction catheters.6 – 8 When P. aeruginosa have colonized a foreign

body, it is almost impossible to eradicate the biofilm and gen-
Infections related to the use of foreign bodies are serious erally it is necessary to remove the implant to get rid of the
complications of medical device insertion.1,2 Each year in the infection.9,10 Unfortunately, replacement surgeries have a
USA alone, about 1 million cases of nosocomial infections high failure rate and the replacements are prone to recoloniza-
are associated with indwelling devices,3 and Pseudomonas aer- tion.9 The biofilm mode of growth protects the bacteria from
uginosa is frequently linked with infections of indwelling cathe- antibiotics such as tobramycin, ciprofloxacin and colistin,
ters and foreign-body implants.4,5 It has, on several occasions, and also shields the bacteria from the host immune
been shown that P. aeruginosa grow as biofilms on medically responses.2,11 – 14
implanted foreign bodies, such as prosthetic heart valves, When P. aeruginosa grow as biofilms, the bacteria use quorum
cardiac pacemakers, total joint prostheses and intravenous sensing (QS) to coordinately control expression of tissue-

# The Author 2012. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.
For Permissions, please e-mail: journals.permissions@oup.com

1198
Combination treatment of in vivo Pseudomonas aeruginosa biofilms JAC
damaging virulence factors, including proteases and rhamnoli- components of the immune system. The QSI compounds admi-
pids. The ability to produce rhamnolipids has been shown to be nistered were furanone C-30,28 ajoene (T. H. Jakobsen, M. van
important for P. aeruginosa and we have proposed that rhamno- Gennip, R. K. Phipps, M. Shanmugham, L. D. Christensen,
lipids function as a protective shield against the innate immune M. Alhede, M. Skindersoe, T. B. Rasmussen, K. Friedrich, F. Uthe,
system where contact causes necrosis of polymorphonuclear P. Ø. Jensen, C. Moser, K. F. Nielsen, L. Eberl, T. O. Larsen,
leucocytes (PMNs).12 In contrast, a P. aeruginosa mutant with D. Tanner, N. Høiby, T. Bjarnsholt and M. Givskov, unpublished
an inactive rhlA gene (which disables the production of rhamno- results) and horseradish juice extract, which we found to
lipids) has lost the ability to cause necrosis. Furthermore, the contain QSI activity.25 We demonstrated the importance of initi-
persistence of the rhlA mutant was significantly reduced in two ating the treatment of the mice prophylactically or early in the
different animal models of infection compared with the parent infection, as early combination treatment improved the ability
wild-type.15 to clear the infection and resulted in a better outcome for the
QS in P. aeruginosa involves two acyl homoserine lactone mice.

Downloaded from https://academic.oup.com/jac/article/67/5/1198/979906 by guest on 20 August 2020

(AHL)-based QS systems: the las and rhl systems. These two
systems are organized with the dual-functioning receptors/tran-
scriptional activators LasR and RhlR, and the signal synthetases Materials and methods
encoded by lasI and rhlI. The synthetases synthesize a specific Bacterial strains
signal molecule named N-(3-oxododecanoyl)-L-homoserine
All experiments were performed with wild-type P. aeruginosa strain PAO1,
lactone (3-oxo-C12-HSL) and N-butanoyl-L-homoserine lactone
obtained from Professor Barbara Iglewski (University of Rochester Medical
(C4-HSL), respectively.16 – 19 In addition to the AHL-based QS
Center, NY, USA). The strain was tagged with a plasmid-based mini-Tn7
systems, P. aeruginosa also employs the Pseudomonas quinolone transposon system (pBK-miniTn7-gfp3) constitutively expressing a
signal (PQS) system. The QS systems of P. aeruginosa have been stable green fluorescent protein, according to Koch et al.29
shown to be hierarchically arranged with the las system control-
ling the rhl system,20,21 and the PQS system positioned between
the las and rhl systems.22 However, the rhl system can operate Animals
independently of the las system23 and it has been suggested Female BALB/c mice were purchased from Taconic M&B A/S (Ry,
that the PQS system controls this activation.24 Denmark) at 8 –9 weeks of age, and were maintained on standard
Recently we have identified 1-isothiocyanato-3- mouse chow and water ad libitum for 2 weeks before the challenge.
(methylsulfinyl)propane, also known as iberin, as a QS inhibitor The animal studies were carried out in accordance with the European
(QSI) compound in horseradish. Furthermore, ajoene Convention and Directive for the Protection of Vertebrate Animals used
(4,5,9-trithiadodeca-1,6,11-triene-9-oxide) was found as a QSI for Experimental and Other Scientific Purposes and the Danish law on
present in garlic extract. By real-time PCR and DNA microarray animal experimentation. All experiments were authorized and approved
analysis, iberin was shown to specifically and effectively target by the National Animal Ethics Committee, Denmark (the Animal Experi-
two of the major QS networks in P. aeruginosa, the LasIR and ments Inspectorate, dyreforsoegstilsynet.dk), and given the permit
number 2010/561-1817.
RhlIR systems, and was found to down-regulate QS-controlled
rhamnolipid production in wild-type P. aeruginosa batch cultures.
However, no effect of iberin could be obtained in vivo using a Foreign-body infection model
foreign-body infection model.25 Ajoene was shown in vitro to sig-
Silicone implants were prepared as described previously by Christensen
nificantly inhibit a subclass of QS-regulated P. aeruginosa genes.
et al.,30 with modifications. Silicone tube implants with a size of 4 mm
Additionally, ajoene was found to have an antimicrobial effect on (inner diameter 4 mm/outer diameter 6 mm, Ole Dich) were used
P. aeruginosa in vivo using a pulmonary infection mouse model instead of square implants and inserted into the intraperitoneal cavity.
(T. H. Jakobsen, M. van Gennip, R. K. Phipps, M. Shanmugham, A bacterial pellet from a centrifuged overnight culture was resuspended
L. D. Christensen, M. Alhede, M. Skindersoe, T. B. Rasmussen, in 0.9% NaCl to an optical density at 600 nm of 0.1 and implants were
K. Friedrich, F. Uthe, P. Ø. Jensen, C. Moser, K. F. Nielsen, left in this solution for 20 h at 378C with shaking at 110 rpm. Animals
L. Eberl, T. O. Larsen, D. Tanner, N. Høiby, T. Bjarnsholt and were challenged according to the method of Christensen et al.30 Mice
M. Givskov, unpublished results). were anaesthetized by subcutaneous (sc) injections in the groin area
The reduced susceptibility of P. aeruginosa biofilms to antibio- with hypnorm/midazolam (Roche) [one part hypnorm (0.315 mg/mL fen-
tics is caused by several different mechanisms of resistance and tanyl citrate and 10 mg/mL fluanisone), one part midazolam (5 mg/mL)
tolerance (for a recent review see Høiby et al.26). It has been and two parts sterile water]. At the end of the experiments, mice were
shown that in vitro-grown P. aeruginosa biofilms in the absence euthanized by intraperitoneal (ip) injection of 10 mL/kg body weight
(BW) pentobarbital (200 mg/mL) with lidocaine hydrochloride (20 mg/
of functional QS (either by mutation or by administration of
mL) (DAK).
QSI drugs) are more susceptible to tobramycin, in contrast to
their QS operative counterparts, suggesting that QSI drugs and
tobramycin show synergistic effects on killing bacterial Bacteriology
biofilms.27,28
After removal from the peritoneal cavity of the mice, silicone implants
In the present study, we demonstrate that QSI compounds were placed in centrifuge tubes containing 2 mL of 0.9% NaCl and kept
provide synergistic antimicrobial effects with tobramycin in com- on ice until the tubes were placed in an ultrasound bath (Bransonicw
bination treatments of P. aeruginosa in vivo biofilms. The experi- model 2510, Branson Ultrasonic Corporation, USA) for 10 min (5 min
ments were performed using an intraperitoneal foreign-body in degas followed by 5 min sonic). After the ultrasound treatment, 100 mL
vivo biofilm infection model that, in contrast to our in vitro of the NaCl–bacteria solution was serially diluted and plated on blue
biofilm models, includes the contribution of important agar plates (State Serum Institute, Denmark) for bacterial visualization.

1199
Christensen et al.

Blue agar plates are selective for Gram-negative bacilli.31 The plates were Histopathology
incubated at room temperature for 2 days before determination of cfu
The abscesses surrounding silicone implants inserted in the mice were
per implant.
prepared for histopathological examination by fixation in 4% (w/v) for-
maldehyde solution (Bie & Berntsen, Denmark), embedding in paraffin
wax and cutting into 5 mm thick sections, followed by haematoxylin
Preparation of QSIs and tobramycin for injection
and eosin staining. The slides were scanned at a low magnification
The furanone C-30 solution was prepared as described by Christensen and, from an average evaluation of a minimum of five representative
et al.30 The mice were given 1 mg/kg BW furanone C-30 and were areas at higher magnification (×400), the type of inflammation was esti-
treated with the inhibitor every 8 h (ip injections) for 2 days starting 1 h mated. The inflammatory responses were scored as acute (.90% PMNs),
post-insertion. The placebo group was injected with a 2.4% ethanol solu- chronic [.90% mononuclear cells (MNs)], both types present, no domin-
tion (96% ethanol in 0.9% NaCl), corresponding to the amount of ethanol ating type (PMNs/MNs) or no inflammation, as described previously for
that the furanone C-30-treated group received. Treatment was continued the evaluation of inflammation in lungs.33 The histopathological evalu-
until 8 h before the mice were euthanized. ation was carried out blind.

Downloaded from https://academic.oup.com/jac/article/67/5/1198/979906 by guest on 20 August 2020

Ajoene was dissolved in 96% ethanol to a concentration of 100 mg/
mL followed by a 40× dilution in a 20% (2-hydroxypropyl)-b-cyclodextrin
solution (vehicle) to a final stock solution of 2.5 mg/mL. The mice were Cytokine levels in peritoneal cavity
given 25 mg/kg BW ajoene and were treated with the inhibitor every
To estimate relative cytokine levels in the peritoneal cavity, the cavity was
24 h (sc injections). For the early treatment experiment, the mice were
flushed with 3 mL of 0.9% NaCl. The collected saline was preserved at
treated from day 2 pre-insertion to day 2 post-insertion; for the late
2808C until assay. The cytokine levels were estimated by using a Multi-
treatment experiment, the mice received ajoene from day 11 to day 13
Analyte ELISArray Kit (SABiosciences, Frederick, MD, USA) using a conven-
post-insertion. Thus, QSI treatment was given for 3 days once the mice
tional ELISA protocol, according to the manufacturer’s instructions. Using
were infected for both the early and late treatment experiments. The
two different arrays (the Mouse Common Chemokines and Mouse In-
placebo group was injected with a 2.4% ethanol solution (96% ethanol
flammatory Cytokines & Chemokines kits), we were able to profile the
diluted in vehicle), corresponding to the amount of ethanol that the
levels of the following cytokines: upon activation normal T cell expressed
ajoene-treated group received. Treatment was continued until 24 h
and presumably secreted, monocyte chemotactic protein (MCP)-1,
before the mice were euthanized.
macrophage inflammatory protein (MIP)-1a, MIP-1b, stromal cell-derived
Horseradish juice extract was prepared by using a juice extractor to
factor 1, interferon-inducible protein 10, monokine induced by interferon
obtain the liquid from a 0.1 kg horseradish root. The horseradish juice
(IFN)-g, eotaxin, thymus and activation regulated chemokine,
was centrifuged at 9000 g for 10 min in 15 mL centrifuge tubes and
macrophage-derived chemoattractant, keratinocyte chemoattractant
the supernatant was discarded. The pellet was mixed with 96%
(KC), 6Ckine, interleukin (IL)-1A, IL-1B, IL-2, IL-4, IL-6, IL-10, IL-12,
ethanol and diluted 40 times in vehicle to a final concentration of the
IL-17A, IFN-g, tumour necrosis factor a, granulocyte-macrophage
juice of 50%. This mixture was then sterile-filtered with a 0.2 mm filter
colony-stimulating factor (GM-CSF), and G-CSF. The plates were read at
and stored at 48C. The mice were treated with 0.2 mL of the solution
450 nm in a Wallac 1420 Viktor2 (Perkin Elmer, MA, USA) with wave-
sc (which corresponds to 0.5% horseradish juice per mouse) every
length correction at 570 nm.
24 h, starting 2 days pre-insertion for the early treatment experiment
and 11 days post-insertion for the late treatment experiment; thus, QSI
treatment was given for 3 days once mice were infected for both the Staining procedures of ex vivo implants
early and late treatment experiments. The placebo group was injected
with a 2.4% ethanol solution (96% ethanol diluted in vehicle), corre- To visualize DNA, PMNs and biofilm, silicone implants were cut with a
sponding to the amount of ethanol that the horseradish juice extract- scalpel into four pieces and stained directly after removal from the
treated group received. Treatment was continued until 24 h before the mice. The pieces were placed in a flow cell with a coverslide on top for
mice were euthanized. visualization with a confocal laser scanning microscope (CLSM) (LSM
Tobramycin sulphate (Apodan, Copenhagen, Denmark) was dissolved 710, Zeiss, Germany). Cell viability was assessed using propidium iodide
in 0.9% NaCl to a concentration of 3 g/kg tobramycin. The mice were (PI) (P-4170; Sigma) was used for visualizing dead cells and DNA as red
injected sc with 30 mg/kg BW tobramycin once every 24 h, initiated fluorescence and green fluorescent Syto9 (Invitrogen) was used as a
24 h post-infection for the early treatment experiment and 12 days post- measure for live cells. Images were obtained with a 63×/oil objective.
insertion for the late treatment experiment. Treatment was continued Image scanning was carried out at 488 nm (green) and 543 nm (red)
until 24 h before the mice were euthanized. laser line from an Ar/Kr laser. The Imaris software package (Bitplane
AG) was used to generate pictures of the DNA.

MIC and pharmacokinetics of tobramycin

Statistical analysis
The MIC of tobramycin was determined by Etest (AB Biodisk, Sweden).
The pharmacokinetics of tobramycin in non-infected BALB/c mice was To compare the bacterial counts (cfu) between two groups of mice, the
determined prior to the combination treatment experiments and per- Mann–Whitney U-test was used (analysis of non-parametric data) for
formed as described by van Gennip et al.,32 with modifications. In calculating P values in the statistical program GraphPad Prism version
short, mice were divided into groups of four and tobramycin was admini- 5.0 (GraphPad Software, Inc., San Diego, CA, USA).
strated sc. At different timepoints after administration (0.25, 0.5, 0.75, 1,
1.5, 2, 2.5, 3 and 4 h), mice were euthanized and blood was aseptically
sampled from the heart with a syringe and transferred to a BD Microtai-
Results
ner SST gold tube (ref. 365956). After centrifugation, serum was stored at MIC and pharmacokinetics of tobramycin
2208C until determination of the concentration using fluorescence polar-
ization immunoassay (Abbott AxSYM, Abbott Diagnostics, USA). Tobramy- The MIC of tobramycin was found to be 0.75 mg/L. Before the
cin sulphate (Apodan, Copenhagen, Denmark) was dissolved in 0.9% treatment experiments were performed, the pharmacokinetics
NaCl. of 30 mg/kg BW tobramycin administered sc in non-infected

1200
Combination treatment of in vivo Pseudomonas aeruginosa biofilms JAC
female BALB/c mice was determined. The pharmacokinetic para- P < 0.006 P < 0.01
meters of the free, unbound fraction (f) of tobramycin following P < 0.0003 P < 0.001
the recent pharmacokinetic/pharmacodynamic terminology P < 0.0002
guidelines34 were estimated to be: fCmax ¼ 91 mg/L (peak con- 108 P < 0.0003
centrations were shown in first sample drawn); ft12 ¼ 0.37 h (half-
lives were determined using linear regression analysis on the 107
first five points of the serum concentration– time curve);
fAUC¼ 57 mg . h/L (the areas under the 24 h concentration– 106
time curves were calculated by the trapezoidal/rule integral);

cfu per implant

fAUC/MIC ratio ¼ 76 (the MIC used for the calculations was
105
0.75 mg/L); and fCmax/MIC ratio ¼ 121, which is well above the
recommended ratio for the optimal concentration-dependent
104

Downloaded from https://academic.oup.com/jac/article/67/5/1198/979906 by guest on 20 August 2020

killing pharmacodynamics of aminoglycosides of 8 –10,35 but
similar to what can be obtained, e.g. in the sputum of cystic
fibrosis (CF) patients when tobramycin is administered by 103
nebulization.36
102

Combination treatment 101

The previous findings that an in vitro wild-type P. aeruginosa
biofilm grown in the presence of a QSI compound is more suscep- 0
tible to tobramycin encouraged us to conduct the same combin- Before insertion Placebo C-30 TOB C-30 + TOB
atory treatment experiments in an in vivo P. aeruginosa n=4 n=9 n=8 n = 10 n = 11
foreign-body biofilm model. To do this, mice had silicone implants Day 0 Day 2
pre-colonized with a wild-type P. aeruginosa biofilm inserted into
the peritoneal cavity, after which they were divided into four treat- Figure 1. Clearance of implants pre-colonized with wild-type
ment groups. Two different QSI compounds and a QSI-containing P. aeruginosa inserted in the peritoneal cavity of BALB/c mice treated
with either placebo (open circles), C-30 (filled triangles), tobramycin
herbal extract were tested in combination with tobramycin: fura-
(TOB) (open triangles) or a combination of TOB and C-30 (C-30+TOB)
none C-30, ajoene and horseradish juice extract.
(filled squares). Squares, triangles and circles represent cfu/implant in
individual mice and horizontal bars represent the medians. The cfu/
Furanone C-30 and tobramycin treatment implant on control implants, not inserted into mice, is represented by
filled circles. Mice were euthanized at day 2 post-insertion and the cfu/
After insertion of the implants, the mice were divided into the fol- implant was determined. P values for the experiments and the number
lowing four groups: Group 1 (placebo group) received 0.9% NaCl (n) of mice in each group are shown.
and a 2.4% ethanol solution (n¼ 9); Group 2 (furanone C-30
group) received furanone C-30 and 0.9% NaCl (n¼ 8); Group 3
(tobramycin group) received tobramycin and a 2.4% ethanol so- initial biofilm infection, mice were divided into the following
lution (n¼ 10); and Group 4 (combination group) received fura- four groups 2 days pre-insertion of the implants: Group 1
none C-30 and tobramycin (n¼ 11). (placebo group) received 2.4% ethanol in vehicle and 0.9%
At 48 h post-infection the mice were euthanized and the NaCl (n¼25); Group 2 (ajoene group) received ajoene and 0.9%
foreign-body implants were removed from the mice and the NaCl (n¼ 22); Group 3 (tobramycin group) received tobramycin
cfu per implant was determined. and 2.4% ethanol in vehicle (n¼ 24); and Group 4 (combination
By means of this experimental setup, we found a significant group) received ajoene and tobramycin (n ¼23).
difference in the clearing of the bacteria from the implants Ajoene treatment was initiated 2 days prior to implantation, as
when treating wild-type P. aeruginosa with a combination of fur- done previously with this QSI using a pulmonary infection model
anone C-30 and tobramycin (Group 4) compared with both the (T. H. Jakobsen, M. van Gennip, R. K. Phipps, M. Shanmugham,
placebo group and the single treatments with furanone C-30 L. D. Christensen, M. Alhede, M. Skindersoe, T. B. Rasmussen,
(Group 2) or tobramycin (Group 3) (P,0.0002, P,0.0003 and K. Friedrich, F. Uthe, P. Ø. Jensen, C. Moser, K. F. Nielsen, L. Eberl,
P,0.001, respectively) (Figure 1). A significant difference in clear- T. O. Larsen, D. Tanner, N. Høiby, T. Bjarnsholt and M. Givskov, un-
ing was also observed between the placebo group and the published results), and tobramycin treatment was initiated 24 h
single-treatment groups (Group 2, P,0.006 and Group 3, post-insertion. Both treatments were injected sc every 24 h. At
P,0.0003). Additionally, we found a significant difference in 72 h after insertion, the implants were removed from the mice
clearing between the two single-treatment groups (Group 2 and the cfu per implant was determined. We found a significant
versus Group 3, P,0.01) (Figure 1). difference in the cfu recovered when comparing the treatment
of wild-type P. aeruginosa with a combination of ajoene and
tobramycin (Group 4) with that of both the placebo group
Ajoene and tobramycin treatment (Group 1) and the single-treatment groups (Group 2 or Group 3)
Ajoene and tobramycin combination treatment was investigated (P,0.0001, P,0.02 and P,0.02, respectively)] (Figure 2a). A sig-
both against an initial biofilm infection, as the case with fura- nificant difference in clearing was also observed between the
none C-30, and against a chronic biofilm infection. For the placebo group and the single-treatment groups (Group 2,

1201
Christensen et al.

P,0.006 and Group 3, P,0.0001). We found no significant differ- (a) P < 0.006
ence in clearing between the two single-treatment groups (Group P < 0.0001 P < 0.02
2 versus Group 3, P¼0.30) (Figure 2a). P < 0.0001
To address the efficacy of the combinatory treatment on a 108 P < 0.02
chronic biofilm infection, mice had silicone implants inserted at
107
day 0 and were divided into four groups: Group 1 (placebo
group) received 2.4% ethanol in vehicle and 0.9% NaCl (n¼ 5);
106
Group 2 (ajoene group) received ajoene and 0.9% NaCl (n¼ 5);

cfu per implant

Group 3 (tobramycin group) received tobramycin and 2.4% 105
ethanol in vehicle (n¼ 6); and Group 4 (combination group)
received ajoene and tobramycin (n¼ 6). 104
Treatments with ajoene/vehicle were initiated day 11 post-

Downloaded from https://academic.oup.com/jac/article/67/5/1198/979906 by guest on 20 August 2020

insertion and tobramycin/NaCl treatments at day 12 post- 103
insertion. Treatments were injected sc every 24 h. At day 11
post-insertion, three mice were euthanized to determine the cfu 102
per implant before starting treatment (see Figure 2b). Here we
observed that all implants were embedded in abscesses. Fourteen 101
days after insertion the rest of the mice were euthanized, and the
implants were removed from the mice and the cfu per implant 0
Before insertion Placebo Ajoene TOB A + TOB
was determined. Furthermore, it was observed that all implants, n = 22 n = 24
n=8 n = 25 n = 23
except one from Group 4, were embedded in abscesses that
Day 0 Day 3
were collected for histopathological evaluation where the type
of inflammation was evaluated. No difference in the cellular in- (b) P < 0.02
flammatory response between any of the groups was observed.
108
By means of cfu we found significantly less wild-type P. aeruginosa
on the implants when treating the mice with a combination of
107
ajoene and tobramycin (Group 4) compared with the placebo
group (Group 1) (P,0.02) (Figure 2b), even though the cfu 106
was only reduced by 4-fold (the median cfu count+SEM for
cfu per implant

Group 1 was 37000000+7200000 and for Group 4 was 105

10600000+8400000). No significant difference was observed
between the other groups. 104
No abscess
103

Horseradish juice extract and tobramycin treatment 102

Horseradish juice extract was also tested against both an initial
biofilm infection and a chronic biofilm infection. We used horse- 101
radish juice extract instead of only the main QSI present in the
0
juice extract, iberin, as iberin has no QSI effect in vivo.25 For
Before insertion Control Placebo Ajoene TOB A + TOB
the initial biofilm infection, mice were divided into the following n=4 n=3 n=5 n=5 n=6 n=6
four groups 2 days pre-insertion of the implants, and treatment
Day 0 Day 11 Day 14
with horseradish juice extract and vehicle was initiated: Group
1 (placebo group) received 2.4% ethanol in vehicle and 0.9% Figure 2. Clearance of implants pre-colonized with wild-type
NaCl (n¼ 18); Group 2 (horseradish group) received horseradish P. aeruginosa inserted in the peritoneal cavity of BALB/c mice treated
juice extract and 0.9% NaCl (n¼ 24); Group 3 (tobramycin with either placebo (open circles), ajoene (filled triangles), tobramycin
group) received tobramycin and 2.4% ethanol in vehicle (TOB) (open triangles) or a combination of TOB and ajoene (A+TOB)
(n¼ 24); and Group 4 (combination group) received horseradish (filled squares). Squares, triangles and circles represent cfu/implant in
juice extract and tobramycin (n¼ 24). individual mice and horizontal bars represent the medians. The cfu/
Treatments were injected sc every 24 h, and treatment with implant on control implants removed day 11 post-insertion is
represented by open squares and control implants, not inserted into
tobramycin and 0.9% NaCl was initiated 1 day post-insertion.
mice (day 0), are represented by filled circles. (a) Early treatment
Mice were euthanized at day 3 post-insertion and the cfu per
experiment where the implants were removed 3 days post-insertion
implant was determined. We found that there was a significant and the cfu/implant was determined. (b) Late treatment experiment
difference in the clearance between the placebo group (Group where the implants were removed 14 days post-insertion and the cfu/
1) and the combination treatment group (Group 4) (P,0.0001) implant was determined. P values for the experiments and the number
(Figure 3a). Furthermore, a significant decrease in the number (n) of mice in each group are shown.
of bacteria recovered from the implants was found when treat-
ing with the combination therapy compared with the two single- placebo group and the single-treatment groups (P,0.04 and
treatment groups (Groups 2 and 3) (P,0.0001 for both). P,0.02, respectively), but no difference was found between
Additionally, a significant difference was observed between the the single-treatment groups (P ¼ 0.57) (Figure 3a).

1202
Combination treatment of in vivo Pseudomonas aeruginosa biofilms JAC
(a) P < 0.04 Horseradish Placebo
P < 0.02 P < 0.0001 1.00
108 P < 0.0001
P < 0.0001

Cytokine level (A450–A570)

107
0.75

106
cfu per implant

0.50
105

104 0.25

Downloaded from https://academic.oup.com/jac/article/67/5/1198/979906 by guest on 20 August 2020

103
0.00
102 KC MCP-1 MIP-1b G-CSF IL-6
Cytokine
101
Figure 4. Cytokine levels (KC, MCP-1, MIP-1b, G-CSF and IL-6), measured
0 using a MultiAnalyte ELISArray Kit, in fluid obtained from flushing the
Before insertion Placebo HR TOB HR + TOB peritoneal cavity of mice treated with either horseradish juice extract
n=8 n = 18 n = 24 n = 24 n = 24 (black bars) or placebo (vehicle, white bars).
Day 0 Day 3

(b) 108 juice extract with tobramycin, had cleaner fur and were less
P < 0.02 influenced by the infection compared with the mice treated
107
with vehicle. We speculated that the better state of the mice cor-
responded to a lower inflammation in the peritoneal cavity. To
106 test this we profiled the cytokine levels in the peritoneal cavity
cfu per implant

using a MultiAnalyte ELISArray Kit (SABiosciences). Saline

105 samples collected from three mice were pooled to provide suffi-
cient material to measure the cytokine levels; the experiment
104 was performed twice. We found a clear trend showing that the
levels of important inflammatory cytokines (KC, MCP-1, MIP-1b,
103 G-CSF and IL-6) were lower in the single-treated horseradish
group compared with the placebo group (Figure 4). This differ-
102 ence was not found in non-infected animals.
To address the outcome in a chronic biofilm infection study,
101 mice had silicone implants inserted at day 0 and the mice
were divided into four groups: Group 1 (placebo group) received
0 2.4% ethanol in vehicle and 0.9% NaCl (n¼ 7); Group 2 (horserad-
Before insertion Control Placebo HR TOB HR + TOB ish group) received horseradish juice extract and 0.9% NaCl
n=4 n=3 n=7 n=6 n=6 n=7 (n¼ 6); Group 3 (tobramycin group) received tobramycin and
Day 0 Day 11 Day 14 2.4% ethanol in vehicle (n¼ 6); and Group 4 (combination
group) received horseradish juice extract and tobramycin (n¼ 7).
Figure 3. Clearance of implants pre-colonized with wild-type Treatments were injected sc every 24 h, where the horse-
P. aeruginosa inserted in the peritoneal cavity of BALB/c mice treated radish juice extract/vehicle treatment was initiated at day
with either placebo (open circles), horseradish juice extract (HR) (filled
11 post-insertion and tobramycin/NaCl treatment at day 12
triangles), tobramycin (TOB) (open triangles) or a combination of TOB
post-insertion. At day 11 post-insertion, three mice were
and HR (HR+TOB) (filled squares). Squares, triangles and circles
represent cfu/implant in individual mice and horizontal bars represent
euthanized to determine the cfu per implant before starting
the medians. The cfu/implant on control implants removed day 11 treatment (see Figure 3b) and, as for the case with the
post-insertion is represented by open squares and control implants, not ajoene treatment experiment, we observed that all three
inserted into mice (day 0), are represented by filled circles. (a) Early implants were embedded in abscesses. At 14 days after inser-
treatment experiment where the implants were removed 3 days tion, the rest of the mice were euthanized and the implants
post-insertion and the cfu/implant was determined. (b) Late treatment were removed from the mice and the cfu per implant was
experiment where the implants were removed 14 days post-insertion determined. Here also, all implants were embedded in
and the cfu/implant was determined. P values for the experiments and abscesses. We found that treatment of wild-type P. aeruginosa
the number (n) of mice in each group are shown. with a combination of horseradish juice extract and tobramycin
(Group 4) resulted in a significant reduction in the cfu as com-
Horseradish juice extract not only had an effect on the bacter- pared with the placebo group (Group 1) (P,0.02) (Figure 3b).
ial count on the implants, but it was also clear that mice treated As observed with the 14 day treatment experiment with
with the extract, and especially the combination of horseradish ajoene and tobramycin, the cfu reduction was small; only an

1203
Christensen et al.

1.3-fold decrease in cfu was found (median cfu count+SEM 100- to 150-fold when comparing the control group with the
for Group 1 was 4 500000+700 000 and for Group 4 was combination treatment group; however, only a minor reduction
3 400000+250 000). No significant difference was observed (1.3- to 4-fold) was observed when starting treatment late. At
between the other groups. days 11 and 14 post-insertion, an abscess had formed around
the implants. Only one mouse receiving combination therapy
of ajoene and tobramycin had no abscess surrounding the
implant at day 14, and the bacterial load on this implant was
Discussion
low compared with that on the rest in this treatment group. The
Treating chronic biofilm infections is one of the most difficult formation of an abscess surrounding the implants when the infec-
medical problems, as when bacteria reside in biofilms they are tion has become a chronic condition is in accordance with previ-
extremely resistant to all kinds of antibacterial drugs as well as ous results;30 however, in the mouse model used previously we
to the host immune defences. Even with the highest deliverable only observed abscesses in NMRI mice and not in BALB/c mice.

Downloaded from https://academic.oup.com/jac/article/67/5/1198/979906 by guest on 20 August 2020

doses of antibiotics, treatment of chronic P. aeruginosa foreign- In the present study, however, we use silicone tube implants
body biofilm infections often fails and this is why the only instead of the previous square ones. This provides us with the op-
option often is to remove the foreign body from the patient.9,10 portunity to infect the mice with a higher bacterial load and this
Interestingly, P. aeruginosa in vitro-grown biofilms have been may lead to abscesses in BALB/c mice. The formation of an
shown to be more susceptible to tobramycin when their QS abscess around the implant may reduce the penetration of the
systems are disabled.14,28,37 Despite the successful demonstra- QSI compounds and tobramycin. Furthermore, bacteria may ex-
tion of synergy in vitro, it does not necessarily mean an improved perience anaerobic conditions, which may offer an explanation
clinical outcome, due to the complexity of the in vivo environ- as to why late treatment results in only a minor reduction of
ment during infection. This study, however, demonstrated that the cfu on the implants. The importance of the early treatment
we could extrapolate our previous results into infected rodents. of P. aeruginosa infections has previously been shown by van
The synergistic effect was demonstrated with two QSI com- Gennip et al.,32 where a pulmonary infection model in mice was
pounds with quite different molecular compositions (furanone studied. Here it was shown that the treatment of mucoid P. aeru-
C-30 and ajoene) and a horseradish juice extract that, similar ginosa isolates with tobramycin initiated 1 h post-infection
to garlic extract, is known to contain QSI activity37,38 (in the resulted in a significant reduction of the bacterial count per lung
former case, mainly in the form of iberin).25 These findings compared with treatment initiated 24 h post-infection.
suggest that the combination treatment of P. aeruginosa bio- Previously, we have shown that functional QS systems play a
films, which aims first at disabling the QS system (thereby redu- key role in the ability of P. aeruginosa to persist in a foreign-body
cing the virulence of P. aeruginosa) and second at killing the infection model and a pulmonary infection model.14,30 Treat-
bacteria with an antibiotic, is a promising strategy to prevent ment with the QSI furanone C-30 alone was found to reduce
biofilm infections developing into a chronic state. Recently, the bacterial count in both mouse models of infection.28,30
Brackman et al.39 presented data indicating synergistic actions However, the present finding that furanone C-30 in combination
of QSI compounds and antibiotics; however, the proposed QS in- with tobramycin shows a synergistic effect on the bacterial clear-
hibitory effect of baicalin hydrate (BA) and cinnamaldehyde (CA) ance is important, as patients who acquire chronic P. aeruginosa
against P. aeruginosa is questionable in vitro. Except for the biofilm infections are often immunocompromised. In this situ-
BA-mediated small reduction in the expression of lasB and an ation it is not expected that the host immune system by itself
unrelated gene, luxI, a thorough transcriptomic analysis (or pro- is capable of clearing the QS-inhibited bacteria. It also minimizes
teomics) of the total genome expression of the QSI has not been the risk of exposing the bacteria to subinhibitory concentrations
undertaken. In addition, there are no data presented as to of tobramycin, which in an in vitro study have been shown to
whether BA shows any inhibitory effect on QS-regulated gene ex- induce biofilm formation.40 In concurrence with this, we found
pression and virulence in vivo, as has been demonstrated with in a pilot study with our in vivo foreign-body infection model
furanone C-30 by Hentzer et al.28 Brackman et al.,39 however, that treatment with 1 mg/kg BW or 5 mg/kg BW tobramycin
showed that a P. aeruginosa mutant in which both synthases resulted in a significantly increased cfu per implant 3 days post-
are knocked out was more susceptible to tobramycin compared insertion. Therefore, and because P. aeruginosa biofilm in the
with the wild-type and that BA or CA did not increase the suscep- lungs of CF patients is treated with very high concentrations of
tibility of the mutant strain toward tobramycin. The reported syn- nebulized tobramycin, the tobramycin concentration used in
ergistic effect of BA as an example of a QSI compound per se the combination studies was 30 g/kg BW, which led to an
with tobramycin is therefore to be considered circumstantial. fCmax/MIC ratio of 121. This concentration was chosen to
In the present study we have investigated combinatory treat- enable a small clearing effect on the biofilm infection by tobra-
ment using three well-verified QSI drugs25,28 (T. H. Jakobsen, mycin to be observed, but was low enough to ensure that a sig-
M. van Gennip, R. K. Phipps, M. Shanmugham, L. D. Christensen, nificant synergistic combination effect would be displayed.
M. Alhede, M. Skindersoe, T. B. Rasmussen, K. Friedrich, F. Uthe, The QS inhibitory effect of ajoene and horseradish extract
P. Ø. Jensen, C. Moser, K. F. Nielsen, L. Eberl, T. O. Larsen, against P. aeruginosa has also been demonstrated by different
D. Tanner, N. Høiby, T. Bjarnsholt and M. Givskov, unpublished in vitro assays (T. H. Jakobsen, M. van Gennip, R. K. Phipps,
results). Our investigation pinpoints the necessity of starting M. Shanmugham, L. D. Christensen, M. Alhede, M. Skindersoe,
QSI treatment in combination with tobramycin prophylactically T. B. Rasmussen, K. Friedrich, F. Uthe, P. Ø. Jensen, C. Moser,
or at an early stage of infection as compared with late-initiated K. F. Nielsen, L. Eberl, T. O. Larsen, D. Tanner, N. Høiby,
treatment. When initiating treatment with ajoene or horseradish T. Bjarnsholt and M. Givskov, unpublished results).25 Transcrip-
juice extract prophylactically, we found a reduction in the cfu of tomic analysis of the effect of ajoene has been undertaken

1204
Combination treatment of in vivo Pseudomonas aeruginosa biofilms JAC
is QS regulated and that QS-deficient biofilms are less tolerant of
antibiotics, including tobramycin.14,43 This offers an explanation
as to why we observe synergistic effects of combination treat-
ments. Furthermore, we have previously shown that wild-type
P. aeruginosa are able to lyse PMNs, which subsequently spill
their internal content of DNA.12 The molecular basis for this DNA
spill is the QS-controlled production of rhamnolipid,12 which we
have proposed to function as a PMN shield.44 Furthermore,
in vitro experiments of the exposure of freshly isolated PMNs to
bacterial biofilms results in the up-regulation of rhamnolipid
production, an effect that is likely transmitted through the QS
controllers.44 The QSIs used inhibit rhamnolipid production by

Downloaded from https://academic.oup.com/jac/article/67/5/1198/979906 by guest on 20 August 2020

disabling the QS-mediated control. Recently, using this mouse
model we visualized PMNs on top of a wild-type P. aeruginosa
biofilm when investigating implants with scanning electron
microscopy. Furthermore, eDNA was also observed on the
implants using CLSM (M. van Gennip, L. D. Christensen,
K. Qvortrup, C. Prats, M. Alhede, H. P. Hougen, R. Phipps, P. Ø.
10 mm
Jensen, N. Høiby, M. Givskov and T. Bjarnsholt, unpublished
results). It is therefore likely that DNA from lysed mouse PMNs
and eDNA from the bacteria are both present on top of the
biofilm, and in this way contribute to the neutralization of tobra-
mycin and other aminoglycosides (see Figure 5 for PI staining of
DNA on an implant removed from a mouse 24 h post-insertion).

Figure 5. Single CLSM image (top view) of a silicone implant removed

24 h post-insertion from a mouse. PI staining of dead bacteria, PMNs
and DNA (white arrows) are visualized. The white box shows a higher
Acknowledgements
magnification image of the area. Scale bar: 10 mm. Parts of this work were presented at the Fifth American Society of
Microbiology Conference on Biofilms, 2009 (Abstract 46) and at the
and it was found that ajoene down-regulated 11 genes, where Twentieth European Congress of Clinical Microbiology and Infectious
Diseases, 2010 (Abstract 1772).
10 were shown to be QS regulated. Furthermore, ajoene (at
We thank Margit Bæksted (Bartholin Institute) for excellent technical
25 mg/kg BW) was found to significantly facilitate the ability of
assistance.
mice to clear a P. aeruginosa lung infection (T. H. Jakobsen,
M. van Gennip, R. K. Phipps, M. Shanmugham, L. D. Christensen,
M. Alhede, M. Skindersoe, T. B. Rasmussen, K. Friedrich, F. Uthe,
P. Ø. Jensen, C. Moser, K. F. Nielsen, L. Eberl, T. O. Larsen, Funding
D. Tanner, N. Høiby, T. Bjarnsholt and M. Givskov, unpublished This study was supported by the Danish Strategic Research Council, the
results). The present data clearly demonstrate that combination Svend Andersen Foundation and the Novo Nordisk Foundation (to M. G.)
therapy shows a greater clinical potential due to the synergistic and the Carlsberg Foundation and the Lundbeck Foundation (to T. B.).
antimicrobial efficacy.
Treatment with horseradish juice extract not only had an effect
on the cfu on implants, but also left the mice in a better state, Transparency declarations
which we found corresponded to lower inflammation in the peri- None to declare.
toneal cavity. Treatment with horseradish juice extract resulted in
reduced levels of KC, IL-6, G-CSF, MCP-1 and MIP-1b. All cytokines
are involved in the inflammatory response in mice when exposed References
to bacteria, e.g. G-CSF is the major mobilizer of PMNs from the 1 Gristina AG. Biomaterial-centered infection: microbial adhesion versus
bone marrow and KC is involved in PMN cell activation. Since the tissue integration. Science 1987; 237: 1588– 95.
difference in cytokine levels was not observed in non-infected 2 Costerton JW, Stewart PS, Greenberg EP. Bacterial biofilms: a common
mice, it is not likely that horseradish exerts anti-inflammatory cause of persistent infections. Science 1999; 284: 1318– 22.
effects but rather works through virulence quenching. 3 Darouiche RO. Treatment of infections associated with surgical
The various mechanisms behind biofilm-specific antibiotic re- implants. N Engl J Med 2004; 350: 1422– 9.
sistance are important to understand in order to develop new anti- 4 Brouqui P, Rousseau MC, Stein A et al. Treatment of Pseudomonas
biofilm drugs. Although the mechanisms are not clarified, the aeruginosa-infected orthopedic prostheses with ceftazidime-
development of aminoglycoside tolerance is, in part, regulated ciprofloxacin antibiotic combination. Antimicrob Agents Chemother
by QS.14,28,41 Mulcahy et al.42 found that the presence of extracel- 1995; 39: 2423– 5.
lular DNA (eDNA) can reduce the susceptibility per se to aminogly- 5 Arciola CR, An YH, Campoccia D et al. Etiology of implant orthopedic
cosides by neutralizing the activity of the antibiotic by means of its infections: a survey on 1027 clinical isolates. Int J Artif Organs 2005;
cation-chelation properties. It has been shown that eDNA release 28: 1091– 100.

1205
Christensen et al.

6 Costerton JW, Montanaro L, Arciola CR. Biofilm in implant infections: its 25 Jakobsen TH, Bragason S, Phipps R et al. Food as a source for QS
production and regulation. Int J Artif Organs 2005; 28: 1062 –8. inhibitors: iberin from horseradish revealed as a quorum sensing inhibitor
7 Braxton EE Jr, Ehrlich GD, Hall-Stoodley L et al. Role of biofilms in of Pseudomonas aeruginosa. Appl Environ Microbiol 2012, in press.
neurosurgical device-related infections. Neurosurg Rev 2005; 28: 249–55. 26 Høiby N, Bjarnsholt T, Givskov M et al. Antibiotic resistance of bacterial
8 von Eiff C, Jansen B, Kohnen W et al. Infections associated with biofilms. Int J Antimicrob Agents 2010; 35: 322–32.
medical devices: pathogenesis, management and prophylaxis. Drugs 27 Bjarnsholt T, Givskov M. The role of quorum sensing in the
2005; 65: 179–214. pathogenicity of the cunning aggressor Pseudomonas aeruginosa. Anal
9 Cherney DL, Amstutz HC. Total hip replacement in the previously septic Bioanal Chem 2007; 387: 409–14.
hip. J Bone Joint Surg Am 1983; 65: 1256 –65. 28 Hentzer M, Wu H, Andersen JB et al. Attenuation of Pseudomonas
10 Neut D, Hendriks JG, van Horn JR et al. Pseudomonas aeruginosa aeruginosa virulence by quorum sensing inhibitors. EMBO J 2003; 22:
biofilm formation and slime excretion on antibiotic-loaded bone 3803– 15.
cement. Acta Orthop 2005; 76: 109–14. 29 Koch B, Jensen LE, Nybroe O. A panel of Tn7-based vectors for

Downloaded from https://academic.oup.com/jac/article/67/5/1198/979906 by guest on 20 August 2020

11 Pamp SJ, Gjermansen M, Johansen HK et al. Tolerance to the insertion of the gfp marker gene or for delivery of cloned DNA into
antimicrobial peptide colistin in Pseudomonas aeruginosa biofilms is Gram-negative bacteria at a neutral chromosomal site. J Microbiol
linked to metabolically active cells, and depends on the pmr and Methods 2001; 45: 187– 95.
mexAB-oprM genes. Mol Microbiol 2008; 68: 223–40. 30 Christensen LD, Moser C, Jensen PO et al. Impact of Pseudomonas
12 Jensen PO, Bjarnsholt T, Phipps R et al. Rapid necrotic killing of aeruginosa quorum sensing on biofilm persistence in an in vivo
polymorphonuclear leukocytes is caused by quorum-sensing-controlled intraperitoneal foreign-body infection model. Microbiology 2007; 153:
production of rhamnolipid by Pseudomonas aeruginosa. Microbiology 2312– 20.
2007; 153: 1329– 38. 31 Høiby N. Epidemiological investigations of the respiratory tract
13 Singh PK, Parsek MR, Greenberg EP et al. A component of innate bacteriology in patients with cystic fibrosis. Acta Pathol Microbiol Scand
immunity prevents bacterial biofilm development. Nature 2002; 417: B Microbiol Immunol 1974; 82: 541–50.
552–5. 32 van Gennip M, Moser C, Christensen LD et al. Augmented effect of
14 Bjarnsholt T, Jensen PO, Burmolle M et al. Pseudomonas aeruginosa early antibiotic treatment in mice with experimental lung infections
tolerance to tobramycin, hydrogen peroxide and polymorphonuclear due to sequentially adapted mucoid strains of Pseudomonas
leukocytes is quorum-sensing dependent. Microbiology 2005; 151: aeruginosa. J Antimicrob Chemother 2009; 64: 1241 –50.
373–83. 33 Moser C, Johansen HK, Song Z et al. Chronic Pseudomonas aeruginosa
15 van Gennip M, Christensen LD, Alhede M et al. Inactivation of the rhlA lung infection is more severe in Th2 responding BALB/c mice compared to
gene in Pseudomonas aeruginosa prevents rhamnolipid production, Th1 responding C3H/HeN mice. APMIS 1997; 105: 838– 42.
disabling the protection against polymorphonuclear leukocytes. APMIS 34 Mouton JW, Dudley MN, Cars O et al. Standardization of
2009; 117: 537–46. pharmacokinetic/pharmacodynamic (PK/PD) terminology for anti-
16 Ochsner UA, Reiser J. Autoinducer-mediated regulation of infective drugs: an update. J Antimicrob Chemother 2005; 55: 601–7.
rhamnolipid biosurfactant synthesis in Pseudomonas aeruginosa. Proc 35 Jenh AM, Tamma PD, Milstone AM. Extended-interval aminoglycoside
Natl Acad Sci USA 1995; 92: 6424– 8. dosing in pediatrics. Pediatr Infect Dis J 2011; 30: 338– 9.
17 Passador L, Cook JM, Gambello MJ et al. Expression of Pseudomonas 36 Heijerman H, Westerman E, Conway S et al. Inhaled medication and
aeruginosa virulence genes requires cell-to-cell communication. Science inhalation devices for lung disease in patients with cystic fibrosis: a
1993; 260: 1127– 30. European consensus. J Cyst Fibros 2009; 8: 295–315.
18 Pearson JP, Gray KM, Passador L et al. Structure of the autoinducer 37 Rasmussen TB, Bjarnsholt T, Skindersoe ME et al. Screening for
required for expression of Pseudomonas aeruginosa virulence genes. quorum-sensing inhibitors (QSI) by use of a novel genetic system, the
Proc Natl Acad Sci USA 1994; 91: 197– 201. QSI selector. J Bacteriol 2005; 187: 1799 –814.
19 Pearson JP, Passador L, Iglewski BH et al. A second N-acylhomoserine 38 Bjarnsholt T, Jensen PO, Rasmussen TB et al. Garlic blocks quorum
lactone signal produced by Pseudomonas aeruginosa. Proc Natl Acad Sci sensing and promotes rapid clearing of pulmonary Pseudomonas
USA 1995; 92: 1490– 4. aeruginosa infections. Microbiology 2005; 151: 3873 –80.
20 Pesci EC, Pearson JP, Seed PC et al. Regulation of las and rhl quorum 39 Brackman G, Cos P, Maes L et al. Quorum sensing inhibitors increase
sensing in Pseudomonas aeruginosa. J Bacteriol 1997; 179: 3127 –32. the susceptibility of bacterial biofilms to antibiotics in vitro and in vivo.
21 Latifi A, Foglino M, Tanaka K et al. A hierarchical quorum-sensing Antimicrob Agents Chemother 2011; 55: 2655–61.
cascade in Pseudomonas aeruginosa links the transcriptional activators 40 Hoffman LR, D’Argenio DA, MacCoss MJ et al. Aminoglycoside antibiotics
LasR and RhIR (VsmR) to expression of the stationary-phase sigma induce bacterial biofilm formation. Nature 2005; 436: 1171–5.
factor RpoS. Mol Microbiol 1996; 21: 1137– 46. 41 Shih PC, Huang CT. Effects of quorum-sensing deficiency on
22 Pesci EC, Milbank JB, Pearson JP et al. Quinolone signaling in the Pseudomonas aeruginosa biofilm formation and antibiotic resistance.
cell-to-cell communication system of Pseudomonas aeruginosa. Proc J Antimicrob Chemother 2002; 49: 309–14.
Natl Acad Sci USA 1999; 96: 11229–34. 42 Mulcahy H, Charron-Mazenod L, Lewenza S. Extracellular DNA
23 Skindersø ME, Alhede M, Phipps R et al. Effects of antibiotics on chelates cations and induces antibiotic resistance in Pseudomonas
quorum sensing in Pseudomonas aeruginosa. Antimicrob Agents aeruginosa biofilms. PLoS Pathog 2008; 4: e1000213.
Chemother 2008; 52: 3648 –63. 43 Allesen-Holm M, Barken KB, Yang L et al. A characterization of DNA
24 Diggle SP, Winzer K, Chhabra SR et al. The Pseudomonas aeruginosa release in Pseudomonas aeruginosa cultures and biofilms. Mol Microbiol
quinolone signal molecule overcomes the cell density-dependency of 2006; 59: 1114– 28.
the quorum sensing hierarchy, regulates rhl-dependent genes at the 44 Alhede M, Bjarnsholt T, Jensen PO et al. Pseudomonas aeruginosa
onset of stationary phase and can be produced in the absence of LasR. recognizes and responds aggressively to the presence of
Mol Microbiol 2003; 50: 29– 43. polymorphonuclear leukocytes. Microbiology 2009; 155: 3500 –8.

1206