Download MW3610 Orig artice

JAC
Journal of Antimicrobial Chemotherapy (1998) 41, 163–169
Antibiotic-induced release of endotoxin: in-vitro comparison of
meropenem and other antibiotics
M. Trautmanna*, R. Zicka, T. Rukavinab, A. S. Crossc and R. Marrea
a
Department of Medical Microbiology and Hygiene, University of Ulm, Ulm, Germany; bDepartment of
Microbiology, University of Rijeka, Croatia; cDivision of Infectious Diseases and Greenbaum Cancer Center,
Department of Medicine, University of Maryland, Baltimore, MD, USA
The influence of meropenem, a new carbapenem antibiotic, on cell morphology and in-vitro
lipopolysaccharide (LPS) release from Escherichia coli was compared with that of imipenem,
ceftazidime, tobramycin and ciprofloxacin. Free and cell-associated LPS was quantified by
means of a capture ELISA method based on the recognition of E. coli LPS by monoclonal
antibodies. Microscopically, meropenem was found to induce spheroplast formation similar
to that seen with imipenem, while ceftazidime and ciprofloxacin induced filament formation.
Free and cell-associated LPS levels were low in the presence of meropenem, imipenem,
ciprofloxacin and tobramycin, but high in the presence of ceftazidime. Reduced endotoxin
release appears to be a common property of carbapenem antibiotics. Morphological changes
in bacteria in the presence of antibiotics do not predict their LPS-liberating effect since
ciprofloxacin induced low levels of LPS despite causing filament formation.
Introduction
Systemic infections due to Gram-negative bacilli are frequently complicated by the development of septic shock
and multiorgan failure. It is accepted that Gram-negative
bacterial lipopolysaccharide (LPS, endotoxin) plays a significant role in the pathogenesis of sepsis-induced shock
and tissue injury.1 While treatment with antibiotics is
necessary to combat the infective process, antibiotics are
also known to precipitate the release of endotoxins by
disrupting the outer membrane of the organisms.2,3 When
comparing different classes of antibiotics in vitro, Goto &
Nakamura were the first to realize that bacteriostatic antibiotics did not liberate substantial quantities of endotoxin
while bactericidal drugs induced the release of large
amounts of free LPS.4 Later, it was shown that rapidly
bactericidal agents exhibiting an intracellular killing
mechanism such as the aminoglycosides cause a comparatively minor LPS release while cell-wall-active drugs
may stimulate the release of major quantities of LPS.5
The carbapenem antibiotic imipenem differs from most
of the -lactam antibiotics by causing reduced LPS release
in vitro.6,7 Microscopic and electron-microscopic studies
have shown that imipenem induces spheroplast formation
while ceftazidime and other -lactams induce filament
formation.8 This phenomenon has been explained by
preferential binding of imipenem to penicillin-binding 2
(PBP2) in contrast to the third-generation cephalosporins
which primarily bind to PBP3.9 PBP2 appears to be responsible for longitudinal cell growth (primarily inhibited
by imipenem) while PBP3 is responsible for cell septation
(primarily inhibited by the third-generation cephalosporins). Recently, clinical studies have shown that the
differences in the rate and degree of LPS release induced
by imipenem and third-generation cephalosporins may
have clinical relevance.10,11
Meropenem is a new carbapenem which differs from
imipenem in its stability against renal dehydropeptidase I
and an enhanced activity against several species of Gramnegative enteric bacteria and Pseudomonas spp.12,13 The
higher activity against Gram-negative bacilli has been
related to an affinity of meropenem for both PBPs 2 and
3.14,15 Therefore, it was not clear whether the drug would
primarily cause spheroplast or filament formation or
whether meropenem would cause reduced LPS liberation.
In this study, we analysed the influence of meropenem
on LPS release by means of a serotype-specific, highly
reproducible ELISA detection method of LPS based
on O-antigen-specific monoclonal antibodies. Imipenem,
ceftazidime, ciprofloxacin and tobramycin were compared.
*Corresponding address: Department of Medical Microbiology and Hygiene, University of Ulm, Steinhövelstrasse 9, D-89075 Ulm,
Germany. Tel: 49-731-5026951; Fax: 49-731-5026949.
163
© 1998 The British Society for Antimicrobial Chemotherapy
M. Trautmann et al.
Capture ELISA for E. coli LPS
Materials and methods
Bacteria
E. coli strain Bort (O18:K1:H7), a CSF isolate originally
obtained from a newborn, has been used previously in
our laboratories.16,17 MICs of the antibiotics against this
organism were measured by means of broth microdilution
according to National Committee for Clinical Laboratory
Standards (NCCLS) recommendations.18
Antibiotics
Meropenem, imipenem, ceftazidime, ciprofloxacin and
tobramycin were included in the study. Standard powders
of known potency were obtained from the respective
manufacturers. Stock solutions of 2560 mg/L were stored
at 80°C.
Growth conditions and preparation of test
samples
Bacterial cultures were grown in cation-supplemented
Mueller–Hinton broth in a shaking waterbath at 37°C until
they reached mid-log phase (3–4 h). The suspension was
adjusted nephelometrically to a concentration of c. 108
organisms/mL which was confirmed by plate counts. One
hundred microlitres of this suspension were added to
20 mL of fresh Mueller–Hinton broth containing twice or
50 times the MIC of the respective antibiotic or no
antibiotic (growth control). Thus, the starting concentration
of organisms in each vial was c. 15 105 organisms/mL. At
the time points indicated, 100 L of the bacterial suspension were removed for the determination of viable
bacterial cells by subculturing volumes of 100 L on
Mueller–Hinton agar plates after serial ten-fold dilution in
physiological saline. At the same time, a 1 mL sample was
removed from each suspension, transferred to a 1.5 mL
Eppendorf tube and centrifuged at 12,000g for 15 min.
The supernatant was removed and frozen at 80°C until
determination of free LPS. Cell-associated LPS was
released from the cell pellet by the method of Hitchcock &
Brown19 with minor modifications as follows: the pellets
were solubilized in 50 L of lysis buffer and heated at
100°C for 10 min. After cooling, 10 L of DNase and
10 L of RNase (Boehringer Mannheim, Germany) were
added to yield final concentrations of 100 and 25 g/mL,
respectively. After incubation at 37°C for 2 h, 25 g of
proteinase K (Boehringer Mannheim) solubilized in 10 L
of lysis buffer, was added to the lysates which were
incubated at 60°C for 1 h. After performing these lysis
steps, the cell suspensions had completely cleared without
residual debris. Finally, the enzymes were inactivated by
heating the suspensions at 95°C for 10 min, and the
solutions kept frozen at 80°C until determination of cellassociated LPS.
LPS concentrations were determined by means of a monoclonal antibody (mAb)-based ELISA method. Ninety-sixwell flat-bottomed ELISA plates (Greiner, Nürtingen,
Germany) were coated with 10 g/mL of a purified mouse
mAb specific for E. coli O18 LPS (clone 26, mouse
IgG 1 ).20 After overnight incubation at 4°C and removal
of antigen, non-specific binding sites were blocked by the
addition of a 1% casein–albumin solution as described
previously.17 Standards of 0.5–64 ng/mL of purified E. coli
O18 LPS and test samples containing unknown amounts
of LPS were added to duplicate wells of the plate. After
further overnight incubation at 4°C, plates were washed
three times with phosphate-buffered saline (PBS). Bound
LPS was traced with a second mouse mAb specific for
E. coli O18 LPS (clone 141, mouse IgM 20) at a dilution of
1:2000. Plates were incubated at 4°C for 4 h and washed
again. Bound mouse IgM was detected by adding alkaline
phosphatase-conjugated goat anti-mouse IgM( ) antibody
(Sigma, Deisenhofen, Germany) at appropriate dilution,
and reactions developed as described previously.17
Examination of bacterial morphology
Bacteria were grown in antibiotic-containing medium
(twice and 50 times the MIC) as described above. After
4 h, samples were removed from the suspension and
stained with Gram’s stain or methylene blue. Photographs
were taken using a Zeiss microscope (Zeiss, Oberkochen,
Germany).
Statistics
The Mann–Whitney U-test was used. A P value of
was considered significant.
0.05
Results
MIC determination
MICs (mg/L) for E. coli strain Bort were as follows:
ceftazidime, 0.125; meropenem, 0.015; imipenem, 0.125;
tobramycin, 1.0; ciprofloxacin, 0.0078.
Standardization of LPS capture ELISA
The ELISA detection curve was log-linear at LPS
concentrations of 4–64 ng/mL. Mean background optical
density (OD) values 1 S.D. in the absence of LPS were
0.003
0.029 (n
27). The lower detection limit of the
test, defined as the lowest standard concentration yielding
mean ODs of 2 S.D. above background levels, was
2 ng/mL. The interassay coefficient of variation was 5.5%
at 2 ng/mL, 4.9% at 16 ng/mL and 5.0% at 64 ng/mL
(n
27). In order to determine whether the presence of
164
Meropenem-induced release of endotoxin
antibiotics would influence LPS recovery, a 16 ng/mL LPS
solution was mixed with various antibiotics at a concentration of 50
MIC before determining the LPS
content of the mixture. No significant influence on LPS
recovery could be detected (Table I).
Influence of antibiotics on bacterial growth and LPS
production
There was no significant difference in the killing rate of the
drugs with the exception of tobramycin which was very
rapidly bactericidal at twice the MIC (Figure 1a and b).
LPS levels determined during incubation with 2 MIC
of the drugs were significantly different only for cellassociated LPS. Both ciprofloxacin and tobramycin produced significantly less cell-associated endotoxin than the
-lactams (Table II).
At 50
MIC, both free and cell-associated LPS concentrations were significantly lower for the carbapenems
and ciprofloxacin than for ceftazidime (Table III). No
significant differences between meropenem and imipenem
were found.
Morphological changes in bacteria
Meropenem induced spheroplast formation similar to that
induced by imipenem, while ceftazidime and ciprofloxacin
induced filament formation. Filaments induced by
ciprofloxacin, however, had a stouter appearance than
those formed in the presence of ceftazidime (Figure 3).
Discussion
In most studies dealing with antibiotic-induced LPS
release, bioreactive endotoxin has been measured by
means of the Limulus amoebocyte lysate (LAL) assay
which involves an enzymatic reaction triggered by the core
region of LPS.6,11,21,22 Endotoxin levels determined by
means of the LAL assay therefore reflect not only the
Table I. LPS recovery by ELISA in the presence of
antibiotics
Antibiotic added
None (control)
Meropenem
Imipenem
Ceftazidime
Ciprofloxacin
Tobramycin
LPS concentration (ng/mL)
mean 1 S.D.
% recovery
16.7
16.2
15.3
17.1
14.5
17.1
3.1
1.2
1.2
1.6
1.7
3.3
104
101
96
107
91
107
Drugs were tested at 50 MIC. Samples were spiked with LPS at a final
concentration of 16 ng/mL. Values are from three independent experiments.
Figure 1. Growth of E. coli Bort in the presence of 2 MIC (a)
and 50 MIC (b) of the antibiotics. Data are geometric means
of three to five experiments per time point. —— , control; ,
meropenem; , imipenem; - - - - -, ceftazidime; , ciprofloxacin;
, tobramycin.
concentration of LPS but also the accessibility of the inner
core which may be unfolded during exposure to cell-wallactive drugs. 23 Conversely, aminoglycoside antibiotics are
known to suppress the LAL reaction, although the
mechanism of this suppression has not been elucidated.24
In the present study, we therefore used a mAb-based
ELISA method for LPS quantification which is not influenced by the presence of antibiotics (Table I) and which
has the additional advantage of not being prone to
contamination with other LPSs. Using this test system, we
found that both soluble and cell-associated LPS levels
were similarly low during treatment of the cultures with
meropenem and imipenem (Tables II and III). No significant differences between these drugs were found. A
significantly lower release of free LPS and production of
cell-associated LPS was found, at 50
MIC, for both
carbapenems compared with ceftazidime (Table III).
These findings show that, although meropenem is known
to bind to both PBPs 2 and 3, it does not liberate more LPS
than imipenem.
165
M. Trautmann et al.
166
Meropenem-induced release of endotoxin
(a)
(b)
(c)
(d)
Figure 2. Morphological changes in bacteria in a culture of E. coli Bort incubated without antibiotic (control, a) and in the presence
of 50
MIC of meropenem (b), ceftazidime (c) or ciprofloxacin (d). Changes induced by imipenem were identical to those seen
with meropenem. Final magnification, 1000.
In addition, we included a quinolone antibiotic since
this group of drugs is increasingly used as a primary
therapeutic option in Gram-negative sepsis. Earlier investigators found that higher total amounts of LPS were
produced during exposure of bacteria to ciprofloxacin at
50
MIC compared with imipenem.3,9 A substantial
endotoxin release following exposure of E. coli to ciprofloxacin and other quinolones has also been described by
McConnell & Cohen.21 This stands in contrast to the
clinical experience with ciprofloxacin which, for instance,
has never been reported to produce Herxheimer reactions
when used for treatment of typhoid fever in contrast to
other drugs. 25–27 Our findings were at variance with those
of the studies cited since we found that free LPS levels
with ciprofloxacin were as low as those measured in the
presence of the carbapenems at both 2 and 50 MIC,
while total cell-associated LPS levels were even significantly lower at 2 MIC (Table II). The high levels of
free LPS measured in other studies using the LAL test
may have been due to an overestimation resulting from
the peculiar ‘hypersensitivity’ of the LAL assay for free
LPS which has been described by Mattsby-Baltzer et al.28
In their study, an ELISA detection method similar to the
one described here measured LPS values much nearer to
actual LPS levels determined by chemical analysis of LPS
marker molecules. The lower amount of cell-bound LPS
found in the presence of ciprofloxacin compared with
ceftazidime cannot be explained by differences in the rate
of bacterial killing since killing curves were nearly
identical for ciprofloxacin and the -lactams. Rather, it
may be assumed that the production and/or assembly of
cell-bound LPS was inhibited by ciprofloxacin. LPS
synthesis involves a number of enzymatic and transport
processes encoded by the rfb gene cluster, and quinolones
may interfere with coordinate transcription of these genes
by inhibiting bacterial gyrase.
Although animal experimental studies have shown that
the varying potential of antibiotics to cause endotoxin
release has implications for the outcome of experimental
Gram-negative sepsis,29,30 doubts have been raised as to
whether such findings may have clinical significance.
However, two recent studies show that this may in fact be
the case. In a carefully controlled, though small, study
involving patients with urosepsis, Prins et al.11 were able to
demonstrate that the use of imipenem as a primary antibiotic significantly shortened the time to defervescence
compared with ceftazidime. Furthermore, there was a
tendency for lower values of inflammatory parameters
167
M. Trautmann et al.
such as serum tumour necrosis factor (TNF ) and interleukin 1 (IL-1) levels in the imipenem group. In a second
report, Mock et al.10 performed a post hoc analysis of data
from a previously conducted prospective, randomized trial
designed to test the efficacy of -interferon in septic
trauma patients treated with either PBP3-specific or nonPBP3-specific antibiotics. Mortality rates were 17% in the
PBP3 group and 8% in the non-PBP3 group (P
0.02),
supporting the concept that endotoxin-releasing properties of antibiotics affect the clinical outcome of sepsis.10
Our study shows that meropenem and ciprofloxacin also
have low endotoxin-releasing potential which may deserve
attention in future prospective, randomized clinical trials
on this subject.
Acknowledgement
10. Mock, C. N., Jurkovich, G. J., Dries, D. J. & Maier, R. V. (1995).
Clinical significance of antibiotic endotoxin-releasing properties in
trauma patients. Archives of Surgery 130, 1234–41.
11. Prins, J. M., van Deventer, S. J. H., Kuijper, E. J. & Speelman,
P. (1994). Clinical relevance of antibiotic-induced endotoxin
release. Antimicrobial Agents and Chemotherapy 38, 1211–8.
12. Catchpole, R. C., Wise, R., Thornber, D. & Andrews, J. M.
(1992). In vitro activity of L-627, a new carbapenem. Antimicrobial
Agents and Chemotherapy 36, 1928–34.
13. Hoban, D. J., Jones, R. N., Yamane, N., Frei, R., Trilla, A. &
Pignatari, A. C. (1993). In vitro activity of three carbapenem antibiotics. Comparative studies with biapenem (L-627), imipenem, and
meropenem against aerobic pathogens isolated worldwide. Diagnostic Microbiology and Infectious Disease 17, 299–305.
14. Sumita, Y., Fukasawa, M. & Okuda, T. (1990). Comparison of
two carbapenems, SM-7338 and imipenem: affinities for penicillinbinding proteins and morphological changes. Journal of Antibiotics
43, 314–20.
The authors thank Zeneca Ltd, Plankstadt, Germany, for
supporting the study.
15. Pryka, R. D. & Haif, G. M. (1994). Meropenem: a new carbapenem antimicrobial. Annals of Pharmacotherapy 28, 1045–54.
References
16. Cross, A. S., Gemski, P., Sadoff, J. C., Ørskov, F. & Ørskov, I.
(1984). The importance of the K1 capsule in invasive infections
caused by Escherichia coli. Journal of Infectious Diseases 149,
184–93.
1. Cross, A. S. & Opal, S. M. (1995). Endotoxin’s role in Gramnegative bacterial infection. Current Opinion in Infectious Diseases
8, 156–63.
2. Shenep, J. L. & Morgan, K. A. (1984). Kinetics of endotoxin
release during antibiotic therapy for experimental Gram-negative
bacterial sepsis. Journal of Infectious Diseases 150, 380–8.
3. Van den Berg, C., de Neeling, A. J., Schot, C. S., Hustinx, W. N.
M., Wemer, J. & de Wildt, D. J. (1992). Delayed antibiotic-induced
lysis of Escherichia coli in vitro is correlated with enhancement of
LPS release. Scandinavian Journal of Infectious Diseases 24,
619–27.
4. Goto, H. & Nakamura, S. (1980). Liberation of endotoxin from
Escherichia coli by addition of antibiotics. Japanese Journal of
Experimental Medicine 50, 35–43.
5. Shenep, J. L., Barton, R. P. & Morgan, K. A. (1985). Role of
antibiotic class in the rate of liberation of endotoxin during therapy
for experimental Gram-negative bacterial sepsis. Journal of
Infectious Diseases 151, 1012–8.
6. Jackson, J. J. & Kropp, H. (1992). -Lactam antibiotic-induced
release of free endotoxin: in vitro comparison of penicillin-binding
(PBP) 2-specific imipenem and PBP 3-specific ceftazidime. Journal
of Infectious Diseases 165, 1033–41.
7. Bucklin, S. E., Fujihara, Y., Leeson, M. C. & Morrison, D. C.
(1994). Differential antibiotic-induced release of endotoxin from
Gram-negative bacteria. European Journal of Clinical Microbiology
and Infectious Diseases 13, Suppl. 1, 43–51.
8. Dofferhoff, A. S. M., Nijland, J. H., de Vries-Hospers, H. G.,
Mulder, P. O. M., Weits, J. & Bom, V. J. J. (1991). Effects of
different types and combinations of antimicrobial agents on endotoxin release from Gram-negative bacteria: an in vitro and in vivo
study. Scandinavian Journal of Infectious Diseases 23, 745–54.
9. Prins, J. M. (1995). Antimicrobial Therapy of Severe Gramnegative Infections. Dosing of Aminoglycosides and Studies on
Endotoxin Release. Academisch Proefschrift, University of Amsterdam.
17. Trautmann, M., Cross, A. S., Reich, G., Held, H., Podschun, R.
& Marre, R. (1996). Evaluation of a competitive ELISA method for
the determination of Klebsiella O antigens. Journal of Medical
Microbiology 44, 44–51.
18. National Committee for Clinical Laboratory Standards. (1993).
Methods for Dilution Antimicrobial Susceptibility Testing for
Bacteria that Grow Aerobically—Third Edition: Approved Standard
M7-A3. NCCLS, Villanova, PA.
19. Hitchcock, P. J. & Brown, T. M. (1983). Morphological heterogeneity among Salmonella lipopolysaccharide chemotypes in
silver-stained polyacrylamide gels. Journal of Bacteriology 154,
269–77.
20. Kaufman, B. M., Cross, A. S., Futrovsky, S. L., Sidberry, H. F.
& Sadoff, J. C. (1986). Monoclonal antibodies reactive with K1encapsulated Escherichia coli lipopolysaccharide are opsonic and
protect mice against lethal challenge. Infection and Immunity 52,
617–9.
21. McConnell, J. S. & Cohen, J. (1986). Release of endotoxin
from Escherichia coli by quinolones. Journal of Antimicrobial
Chemotherapy 18, 765–6.
22. Eng, R. H. K., Smith, S. M., Fan-Havard, P. & Ogbara, T.
(1993). Effect of antibiotics on endotoxin release from Gramnegative bacteria. Diagnostic Microbiology and Infectious Diseases
16, 185–9.
23. Bhattacharjee, A. K., Opal, S. M., Taylor, R., Naso, R.,
Semenuk, M., Zollinger, W. D. et al. (1996). A noncovalent complex
vaccine prepared with detoxified Escherichia coli J5 (Rc chemotype) lipopolysaccharide and Neisseria meningitidis group B outer
membrane protein produces protective antibodies against Gramnegative bacteremia. Journal of Infectious Diseases 173, 1157–63.
24. Artenstein, A. W. & Cross, A. S. (1989). Inhibition of endotoxin
reactivity by aminoglycosides. Journal of Antimicrobial Chemotherapy 24, 826–8.
25. Ramirez, C. A., Bran, J. L., Mejia, C. R. & Garcia, J. F. (1985).
168
Meropenem-induced release of endotoxin
Open, prospective study of the clinical efficacy of ciprofloxacin.
Antimicrobial Agents and Chemotherapy 28, 128–32.
26. Edelman, R. & Levine, M. M. (1986). Summary of an international workshop on typhoid fever. Reviews of Infectious Diseases
8, 329–49.
27. Dutta, P., Rasaily, R., Saha, M. R., Mitra, U., Bhattacharya, S.
K., Bhattacharya, M. K. & Lahiri, M. (1993). Ciprofloxacin for treatment of severe typhoid fever in children. Antimicrobial Agents and
Chemotherapy 37, 1197–9.
28. Mattsby-Baltzer, I., Lindgren, K., Lindholm, B. & Edebo, L.
(1991). Endotoxin shedding by enterobacteria: free and cell-bound
endotoxin differ in Limulus activity. Infection and Immunity 59,
689–95.
29. Bucklin, S. E. & Morrison, D. C. (1995). Differences in therapeutic efficacy among cell wall-active antibiotics in a mouse model
of Gram-negative sepsis. Journal of Infectious Diseases 172,
1519–27.
30. Jackson, J. J. & Kropp, H. (1995). Carbapenem- and
cephalosporin-induced release of lipopolysaccharide from smooth
and rough Pseudomonas aeruginosa: in vivo relevance. In
Differential Release and Impact of Antibiotic-induced Endotoxin
(Faist, E., Ed.), pp. 21–35. Merck, Whitehouse, NJ.
Received 13 March 1997; returned 6 May 1997; revised 30 May
1997; accepted 15 August 1997
169