Download Antibiotic-induced Bacterial Toxin Release – Inhibition by

Comprehensive Summaries of Uppsala Dissertations
from the Faculty of Medicine 1393
Antibiotic-induced Bacterial
Toxin Release – Inhibition by
Protein Synthesis Inhibitors
BY
GUNILLA HJERDT-GOSCINSKI
ACTA UNIVERSITATIS UPSALIENSIS
UPPSALA 2004
!" #$$% !&'! ( ) ( (
*)) + ( ,-. /) 0 10).
23456 5. #$$%. 74 8
/ 9 :) * 1) :). 7
. !&&. ! . . :18;
!4%4!$4&
/ ) ( ) 4
( )
4 ) ( ) (
) . : )
(
0) ) (
) . 7 4(
)
)0 ((
)
) . /)( 0
( () )
4 ( ( ( ) ( )
( ( ) ) ) *8* &4( <4
.
/) 0 (
( 4
0)
) ( ) *8* &4( <4
( ) ) ) 0
0) ( . :
(
) 6 . : 6 ) ( 0
)) ( ) ( (.
;) 0
( ( .
; ( 0
) . : 4(
(( ( (
= 6 ) 0
.
/) ) ) 7 +17- ( 7 )
( )
. /) 17 0
) ) ( . 2) ( 17.
/) ) ( ) 0) . 7 ( <4
) ) (
() .
> ?*1 17 4 <4
)6
! "#$!%& & '%% #%
& & $()*+) & @ 5
23456 #$$%
:11; $#A#4"%"
:18; !4%4!$4&
''''
4%# +)'.6.BC''''
4%#-
List of papers
This thesis is based on the following papers, which will be referred to in the
text by their Roman numerals:
I
II
III
IV
V
Sjölin J., Goscinski G., Lundholm M., Bring J. and I. Odenholt.
Endotoxin release from Escherichia coli after exposure to tobramycin: dose-dependency and reduction in cefuroxime-induced
endotoxin release. Clinical Microbiology and Infection. 2000;
6:74-81.
Goscinski G., Lundholm M., Odenholt I. and J. Sjölin. Variation
in the propensity to release endotoxin after cefuroxime exposure
in different Gram-negative bacteria. Uniform and dosedependent reduction by the addition of tobramycin. Scandinavian
Journal of Infectious Diseases. 2003; 35:40-46.
Goscinski G., Lipcsey M., Eriksson M., Larsson A., Tano E. and
J. Sjölin. Endotoxin neutralization and anti-inflammatory effects
of tobramycin and ceftazidime in porcine endotoxin shock. Critical Care. 2004; 8:35-41.
Goscinski G., Tano E., Löwdin E. and J. Sjölin. Propensity to
release endotoxin after two repeated doses of cefuroxime in an in
vitro kinetic model; higher release after the second dose. Submitted.
Goscinski G., Tano E., Thulin P., Norrby-Teglund A. and J.
Sjölin. Release of streptococcal SpeA after exposure to penicillin; dependency on dose and inhibition by clindamycin. In manuscript.
Contents
Introduction and background ..........................................................................7
Pathogenesis of severe sepsis and septic shock..........................................7
Endotoxins .............................................................................................9
Exotoxins .............................................................................................11
Endotoxin neutralisation and immunomodulation as treatment strategies
..................................................................................................................11
Release of toxin after exposure to antibiotics...........................................12
Antibiotic-induced release of endotoxin from gram-negative bacteria12
Release of exotoxin from gram-positive bacteria ................................13
Effects of antibiotics on the immune system............................................14
Aims of the study ..........................................................................................15
Materials and methods ..................................................................................16
Bacterial cultures......................................................................................16
Media........................................................................................................16
Antibiotics ................................................................................................16
Determination of minimum inhibitory concentrations (MIC)..................17
Determination of bacterial killing by viable count...................................17
In vitro kinetic model and experimental procedure..................................18
Morphology..............................................................................................18
Determination of endotoxin in the in vitro experiments .........................19
Determination of possible mechanisms for a tobramycin-caused reduction
in cefuroxime-induced endotoxin release ................................................19
Determination of SpeA concentration .....................................................19
Animals, anaesthesia and preparatory procedures....................................20
Administration of endotoxin and antibiotics ............................................21
Measurements ..........................................................................................22
Determination of plasma endotoxin, TNF-D and IL-6 concentrations.....22
Calculations and statistical methods.........................................................22
Results and discussion ..................................................................................24
MIC values ...............................................................................................24
Killing rates ..............................................................................................24
Antibiotic-induced endotoxin release.......................................................25
Endotoxin release in relation to the number of killed bacteria.................27
Possible mechanisms for reduction in endotoxin release by tobramycin .30
Possible anti-inflammatory effects by antibiotics ...................................33
Effect of repeated dosing on endotoxin release and morphology ...........33
Exotoxin release ......................................................................................36
Clinical implications .....................................................................................39
Conclusions...................................................................................................42
Acknowledgements.......................................................................................43
References.....................................................................................................45
Introduction and background
Sepsis is defined as a condition with fever, takycardia, takypnoe and leukocytosis or leukopenia caused by an infection (4). If sepsis leads to organ
dysfunction, hypoperfusion or hypotension, it is classified as severe sepsis.
When the hypotension persists despite adequate fluid resuscitation and hypoperfusion or organ dysfunction is present the term septic shock is used (4).
Bacteria are the most common agents causing septic shock but fungi, mainly
Candida albicans, as well as parasites and viruses may be responsible in rare
cases (101). Among bacterial agents E. coli, group B streptococci and Listeria predominate in the neonates whereas pneumococci, meningococci and
S. aureus are the most common causative agents in children and younger
adults (86). Septic shock in elderly patients is more frequently caused by
gram-negative bacteria from the gastrointestinal, urinary and respiratory
tracts. Since the 1980s there has been a resurgence of severe group A streptococcal (GAS) infections with rapid progression to toxic shock in patients
of all ages, also including otherwise healthy young persons (41,88).
Despite adequate antibiotic treatment and optimal intensive care efforts,
severe infections caused by gram-negative, as well as gram-positive bacteria
still often proceed to shock and multiorgan failure (48,92). Thus, in studies
performed on patients with sepsis and septic shock the mortality rate has
ranged from 30 to 80 %, depending on severity, duration, underlying conditions and on the definitions used (58,70,80,83).
Since the beginning of the antibiotic era, experimental studies and discussions have been pursued to find the reasons for and to prevent therapeutic
failures in severe sepsis where antibiotic treatment even has been suggested
by some authors to be responsible for the aggravating conditions. (25,69,85).
Pathogenesis of severe sepsis and septic shock
It is well known that bacterial products, like endotoxin from the gramnegative and exotoxins from the gram-positive bacteria, in different ways
7
initiate the inflammatory host response in sepsis and that this may further
progress to severe sepsis and septic shock (19,43,84). With a high virulence
or number of bacteria in the circulation or locally, the proinflammatory reaction thus initiated cannot be inhibited by the anti-inflammatory systems that
occur somewhat later. Besides bacterial properties, the degree of response in
the host also depends on individual factors like underlying conditions, medications and probably also genetic factors.
The initial response to toxin release includes activation of the macrophages
and monocytes to produce IL-1, TNF and other cytokines and, through a
complicated interaction with the activated coagulation, fibrinolytic, kallikrein-kinin and complement systems, an activation of the neutrophil granulocytes takes place. Through adhesion of the neutrophils to the endothelium in
multiple organs there is an increased permeability that might lead to a leakage from the blood vessels. Simultaneous activation of the coagulation and
anti-coagulation may lead to imbalance resulting in microthrombosis or,
occasionally, bleeding (86, Fig.1).
Figure 1. Initial activation of inflammatory response.
In septic shock, a general vasodilatation, caused by released substances, such
as TNF, PAF, bradykinins, prostaglandins and nitrous oxide, takes place.
8
This, together with tissue oedema and microthrombosis, leads to a lowered
oxygen uptake by different organs. Hypovolaemia occurs because of the
vasodilatation, as well as dehydration, and the increased capillary leakage.
Initially, the circulation is hyperdynamic with a high cardiac output and a
low vascular resistence. If the hypovolaemia is not corrected, the cardiac
output decreases and a hypodynamic septic shock develops.
A serious inflammatory reaction may also lead to development of acute respiratory distress syndrome (ARDS) in the lungs which further lowers the
oxygen delivery to the organs. Decreased oxygen delivery and organ uptake
causes anaerobic metabolism and metabolic acidosis which, together with
the generalised inflammation results in multiple organ dysfunction.
Endotoxins
The term endotoxin refers to the biological activity of the lipopolysacharide
molecule (LPS) in the cell wall of the gram-negative bacteria. The LPS
molecule consists of 3 parts: the outer polysaccharide hydrophilic O-side
chain, a centre core oligosaccharide and an inner hydrophobic lipid A portion which is the mediator of toxicity (43,58). It has been shown in animals
and humans that administration of purified endotoxin can induce the same
pathophysiological effects as those in gram-negative sepsis (68).
Endotoxin release from the bacterial cell wall occurs mainly as a consequence of death and lysis of the cell and of growth by cell division (46).
When the endotoxin is released and appears as free endotoxin, its biological
activity increases considerably as compared with the bacterial cell-bound
endotoxin (57).
Once released, endotoxin binds to several receptors and carrier proteins, the
most important being the LPS binding protein (LBP) (99, Fig.2).
9
Figure 2. Release of free endotoxin from gram-negative bacteria.
The LBP-endotoxin complex interacts with the CD14-receptor on the
macrophage where a Toll-like receptor-4 (TLR-4) transduces the endotoxin
signal to the cell which responds by cytokine production. LBP is crucial for
the interaction with the CD14 receptor and it has been demonstrated that
TNF-production is reduced 1000-fold in LBP-depleted plasma (64). Besides
binding to the CD14 receptor, endotoxin may also activate the macrophage
via other receptors, such as the E2 integrin CD11/CD18 receptor (45). Membrane bound CD14 may be released from the cell surface and this soluble
CD14 (sCD14) may bind to the LBP-endotoxin complex. LBP is freed and
the endotoxin-sCD14 complex may then activate cells that do not express
CD14, such as endothelial cells (29).
The effect of endotoxin may be inhibited by binding to a LBP-like protein,
the bactericidal/permeability increasing protein (BPI) which is synthesised
by the neutrophil granulocytes upon endotoxin stimulation (93). Already at
the onset of endotoxin release, LPS can bind to specific antibodies that bind
to the O-chain and represent the result of previous endotoxin exposures
(28,35). In addition, low- and high-density lipoproteins and apolipoprotein A
have been shown to inhibit endotoxin-induced cytokine release (26).
10
Exotoxins
Gram-positive bacteria like staphylococci and streptococci produce exotoxins with varying biological activities that can be released without bacterial
growth or lysis. (92). As an example, in severe group A streptococcal infections the pyrogenic exotoxins together with the phagocytosis-inhibiting Mproteins on the streptococcal surface are the most important virulence factors. In similarity to endotoxin, the exotoxins stimulate monocytes and
macrophages to release TNF and other cytokines (37). In addition, the exotoxins have superantigenic properties, meaning that they are mitogenic for
certain T-cell subsets but do not require processing by antigen presenting
cells (52). Consequently, release of excessive amounts of inflammatory cytokines, such as TNF-E, IL-1 and IL-6 can take place. Exotoxins of major
importance are the superantigens streptococcal pyrogenic exotoxin A (speA),
B , C and F (72).
Endotoxin neutralisation and immunomodulation as
treatment strategies
The cornerstones in the treatment of severe sepsis and septic shock include
antibiotics and, if necessary, surgery for the killing and removal of causative
bacteria, fluid replacement for reversal of hypovolemia, oxygen therapy for
the treatment of hypoxia and support of deteriorating vital organs. Despite
this, mortality is still high and therefore numerous attempts have been made
to interfere with the proinflammatory cascades in severe sepsis and septic
shock.
As endotoxin is the major activator in severe gram-negative sepsis, several
clinical trials have studied the effect of polyclonal or monoclonal antibodies
directed against the inner biological part of the LPS molecule (36,102,103).
Positive effects, especially in the most severely sick patients, were seen but
in the confirmatory trials all-cause mortality was not significantly affected.
Other studies have demonstrated a binding and neutralization of LPS after
addition of recombinant BPI that shows higher affinity for endotoxin than
LBP (30,98). In a clinical study in patients with meningococcemia, known to
have high endotoxin concentrations, there was a trend towards an improved
clinical outcome in the group that was treated with rBPI but no statistically
significant effects could be demonstrated (59).
Polyclonal intravenous immunoglobulin (IVIG) has been shown to neutralise
11
streptococcal toxins (73). However, clinical experience is limited to one
retrospective comparative study demonstrating a beneficial effect on mortality in patients with streptococcal toxic shock syndrome (50) and one small
prospective study showing a significant decrease in sepsis-related organ
failure (20). In patients with severe sepsis or septic shock caused by other
bacteria, no beneficial effects have been demonstrated.
Clinical trials with antibodies or other molecules antagonising the effect of
several proinflammatory cytokines, such as TNF-Į, IL-1Eplatelet activating
factor, or coagulation factors, have shown positive results in some subgroups
but no significant effect in all-cause mortality (1,2,3,16,22,74,81,97) and
therefore these products have not been licensed for clinical use. Recently,
beneficial effects on mortality have been demonstrated for corticosteroids (6)
and activated protein C (10). However, with these treatments there are always difficulties in establishing the optimal dose because, if doses are too
high, the beneficial effect may proceed into an adverse one leading to immune paralysis or bleeding.
Release of toxin after exposure to antibiotics
Above their primary bactericidal and bacteriostatic activity, antibiotics have
been shown to exert other effects that may be of importance for treatment of
infections. One of these is the impact on the liberation of bacterial toxins.
Antibiotic-induced release of endotoxin from gram-negative
bacteria
A number of in vitro and in vivo studies have shown an increase in endotoxin release after exposure to different antibiotics (5,18,23,44,77,85).
In general, bactericidal antibiotics liberate initially more endotoxin than bacteriostatic antibiotics and antibiotics active on the cell wall, such as penicillins and cephalosporins, release more than antibiotics with other modes of
action, such as protein synthesis inhibitors. However, there are wide variations between different antibiotics and even among the E-lactam antibiotics,
there are great differences in the propensity to free endotoxin (11,24,85,95).
Penicillin-binding proteins (PBP) are enzymes that are located in the bacterial cell wall and responsible for the cell wall synthesis. They are also the
12
primary targets for the E-lactam antibiotics and, depending on the affinity to
these PBPs, varying amounts of endotoxin are liberated from the gramnegative bacteria at exposure to these antibiotics (78). Beta-lactam antibiotics with affinity for PBP 1, lead to rapid killing without additional release of
endotoxin whereas antibiotics with selective affinity for PBP 2, lead to conversion of the bacteria to round cells, spheroplasts, with loss of viability but
without cell wall destruction and excessive endotoxin release. Binding to
PBP 3, causes selective inhibition of septation and continuing bacterial elongation with formation of long filaments and a subsequent increased endotoxin production.
Thus, release of high amounts of endotoxin is mainly associated with PBP 3
binding. Cefuroxime, cefotaxime, piperacillin and aztreonam bind to PBP 3
and are associated with antibiotic-induced endotoxin release (24,78). Ceftazidime, at high concentrations, binds to PBP 1 and the carbapenems to
PBP 2 (24,49,78). At lower concentrations, ceftazidime and meropenem bind
predominantly to PBP3 resulting in higher release of endotoxin at lower than
at higher doses (49,71,78,94). Simultaneous inhibition of PBP 1a and 3, that
may be caused by ceftazidime at mid-concentration levels, has also been
demonstrated to result in formation of spheroplasts. These findings have
been explained as the sum of inhibitory effects (39).
Among the bactericidal antibiotics, aminoglycosides inhibit protein synthesis
by binding to 16S rRNA (32), which results in rapid killing without excessive endotoxin release (8,24,27). Limited data also suggest that aminoglycosides may even reduce the endotoxin liberation induced by E-lactam antibiotics (24). The quinolones inhibit protein synthesis via inhibition of bacterial
DNA synthesis (42). Despite this mode of action, quinolones have in some
studies been associated with a substantial endotoxin release, whereas in others this has not been observed (58,78).
Antibiotic-induced endotoxin release has been shown to be biologically active and to induce a macrophage TNF response in vitro (23). In animal models, there is strong evidence that soluble endotoxin released from gramnegative bacteria by antibiotics contributes to pathogenesis and mortality in
experimental sepsis (70). However, the clinical significance of antibioticinduced endotoxin release has still not been established.
Release of exotoxin from gram-positive bacteria
Data on exotoxin release from gram-positive bacteria are more limited than
those on endotoxin release from gram-negative bacteria but recently it has
13
been shown the liberation of streptococcal toxin may be dependent on the
antibiotic treatment (17,62,89). It has been demonstrated in vitro that liberation of SpeA is lower after treatment with protein synthesis inhibitors, like
clindamycin and linezolid, than the E-lactam antibiotic penicillin (17,62).
The proposed mechanism has been inhibition of toxin production. In severe
group A streptococcal myositis in mice, these results have been supported
and in the experiments it has been shown that clindamycin significantly increases survival compared with penicillin treatment (91). Inhibition of toxin
synthesis as well as inoculum independency and a long post-antibiotic effect
of clindamycin were suggested as explanations for the result.
Effects of antibiotics on the immune system
In a considerable number of in vitro and ex vivo experiments, antibiotics
have also been found to modulate host inflammatory mechanisms
(31,38,51). Graded doses of various antibiotics have been tested for their
ability to affect functions such as chemotaxis, phagocytosis as well as bacterial killing with the generation of reactive oxygen molecules. This oxidative
stress induced by polymorphonuclear activation during septic shock has been
shown to trigger an inflammatory response with later cytokine production
(55). Several studies have demonstrated that certain E-lactam antibiotics and
aminoglycosides provide protection from oxygen radicals with subsequent
reduction of epithelial damage (15,76). While immune modulation in these
in vitro and ex vivo experiments appear to be marked, in vivo data are limited and clinically decisive effects have not been demonstrated (38).
14
Aims of the study
As the endo- and exotoxins are primary important mediators of septic shock
and since antitoxic and most anti-inflammatory agents have shown no clear
efficacy, an attractive alternative seems to be prevention or minimisation of
their release in order to diminish the effects on the inflammatory response.
Therefore, it was of interest to further study the antibiotic-induced release of
toxins after exposure to antibiotics used for the treatment of the most severe
infections, severe sepsis and septic shock and especially if protein synthesis
inhibitors could reduce the release induced by PBP 3-specific E-lactam antibiotics.
The aims were:
x To investigate the inhibitory effects of varying doses of a protein synthesis inhibitor, tobramycin, on endotoxin release in vitro from various
gram-negative bacteria induced by a PBP 3-specific E-lactam antibiotic,
cefuroxime.
x
To study the relationship between endotoxin release and bacterial killing.
x
To study if tobramycin binds unspecifically to endotoxin in vitro or in
vivo or affects the binding of the LBP-endotoxin complex to the macrophages,in vivo
x
To study the effects on endotoxin release by repeated cefuroxime doses
in an in vitro pharmacokinetic model simulating the human concentration time-profile.
x
To study if tobramycin or ceftazidime can reduce the cytokine and
physiological response to endotoxin in vivo.
x
To investigate the release of SpeA from a GAS strain after exposure to a
E-lactam antibiotic, penicillin G, a protein synthesis inhibitor, clindamycin or a combination of the two and to relate the toxin release to the initial bacterial concentration, the antibiotic doses and the killing rate.
15
Materials and methods
Bacterial cultures
Clinical isolates of Escherichia coli, B 049-3036 (I,IV), B 940-3012 (II), B
942-1035 (II) and Klebsiella pneumoniae B 541-1003 (II) obtained from the
Department of Clinical Microbiology, Uppsala, Sweden, were used. In addition, a clinical isolate of Streptococcus pyogenes group A, 8004 (V) obtained
from Mount Sinai Hospital, Toronto, Canada was studied.
The reference strains of Escherichia coli CCUG 12620* (II), Salmonella
enteritidis CCUG 35852 (II) and Neisseria meningitidis CCUG 15117 (II)
purchased from the Department of Clinical Microbiology, Gothenburg,
Sweden were used, as well as the reference strain of Escherichia coli ATCC
25922 (IV).
* The CCUG number was discovered to be wrong in the lab. storage. The strain should have
been denominated CCUG 17620 which is identical with ATCC 25922.
Media (I, II, IV, V)
The GAS strain (V) was cultured in Todd-Hewitt broth supplemented with
1.5% yeast extract. The Neisseria meningitidis (II) was grown in a serum
containing broth (Fildes and horse serum) made from pyrogen-free water
and all the other gram-negative strains were cultured in Brain Heart Infusion
(BHI) which was also made with pyrogen-free water.
Antibiotics
The antibiotics were obtained as reference powders with known potencies.
Cefuroxime and ceftazidime were purchased from Glaxo Wellcome AB,
16
Gothenburg, Sweden, bensylpenicillin from Astra Läkemedel AB, Södertälje, Sweden, clindamycin from Pharmacia Upjohn, Kalamazoo, Michigan, USA and tobramycin from Eli Lilly Sweden AB, Stockholm, Sweden.
Determination of minimum inhibitory concentrations
(MIC)
MICs were determined in triplicate by two-fold microdilution in broth with
an inoculum of approximately 105 cfu of the test strain per ml according to
NCCL´s standard (I), by E-test (Biodisk, Solna, Sweden) (II,IV), and the
microdilution method (V).
Determination of bacterial killing by viable count (I, II,
V)
Strains in exponential growth phase were diluted and transferred into pyrogen-free glass vials to obtain a starting inoculum of approximately 106
cfu/ml (I,II). The bacteria were then exposed to 0.1 (I), 2 (I,II), 10 (I,II) and
50 (I,II) x MIC of cefuroxime, tobramycin or a combination of the two and
further incubated at 35oC. In each experiment, one glass vial served as a
control. Samples for viable counts were drawn and diluted in phosphate
buffered saline (PBS) immediately before and after addition of antibiotics
and at 1, 2, 6, 24 h (I) and 2, 4 h (II). Dilutions of N. meningitidis were
seeded on chocolate agar plates and the remaining gram-negative strains on
blood-agar plates.
The GAS strain (V) in a log phase culture of approximately 109 cfu/ml was
centrifuged and diluted to around 7.0 log10 cfu/ml in the low and 8.6 log10
cfu/ml in the high inoculum experiment. The strain was then exposed to 2
and 1000xMIC of bensylpenicillin, 2 and 32xMIC of clindamycin and combinations of the two at 2xMIC of both and 1000xMIC and 32xMIC, respectively. These concentrations were chosen to mimic low and high concentrations in the clinical setting. One glass vial without antibiotic served as a control. Samples for viable counts were drawn immediately before and after
addition of antibiotics, and, after further incubation at 37oC, at 3, 6, and 24 h.
Dilutions were spread on blood-agar plates.
17
In vitro kinetic model and experimental procedure (IV)
The in vitro kinetic model has been described in detail elsewhere (61). It
consists of a spinner flask with a total volume of 110 ml with a filter membrane (0.45Pm) fitted in between the upper and bottom part, impeding elimination of bacteria. A magnetic stirrer ensures homogeneous mixing of the
culture and prevents membrane pore blockage. In one of the side arms of the
culture vessel, a silicon membrane is inserted to enable repeated sampling. A
thin plastic tubing from a vessel containing fresh medium is connected to the
other arm. The medium is drawn from the flask at a constant rate by a pump
(type P-500, Pharmacia Biotech. Norden, Sollentuna, Sweden), while fresh,
sterile medium is sucked into the flask at the same rate by the negative pressure built up inside the culture vessel. The flow-rate of the pump was adjusted to obtain the desired half-life of the drug. The apparatus was placed in
a thermostatic room at 37oC during the experiments.
The bacteria were added to the antibiotic suspension immediately before it
was transferred to the culture vessel at time zero, resulting in a starting inoculum of 5x106 cfu/ml . The initial cefuroxime concentration in the vessel
was 60 Pg/ml. The half-time of 1.5 h for cefuroxime was chosen in order to
simulate the human pharmacokinetics. The same dose was placed into the
vessel after about 12 h when the inoculum had reached a level approximating that of the starting inoculum. Six experiments were performed with each
strain.
In an additional experiment, the same initial dose was given, but at the time
of the second dose, a combination of cefuroxime and tobramycin to result in
final concentrations of 60 Pg/ml and 20 Pg/ml, respectively, was given. Two
experiments with each strain were performed.
Samples for viable counts were drawn before addition of antibiotics and at 2
h, 4 h, 14 h and 16 h. After dilution in PBS, the samples were seeded on
Columbia agar plates and incubated at 35oC.
Morphology (IV)
The morphology of E. coli was examined with scanning electron microscopy. For this purpose a separate experiment was performed with the ATCC
25922 strain given the same antibiotic concentrations as those used in the
preceding experiments and with samples drawn before the antibiotic doses
18
and at 1 h and 2 h after each dose. The examinations were made with a LEO
Gemini 1530 scanning electron microscope at an accelerating voltage of
2kV.
Determination of endotoxin in the in vitro experiments
(I, II, IV)
For all endotoxin assays endotoxin-free glass tubes and vials, preheated at
180o for 4 h, were used. Sterile, pyrogen-free plastic syringes and pipettes
were used, and to analyse the free endotoxin, non-pyrogenic filters of 0.45
Pm were employed. Samples for endotoxin analysis were drawn at the same
time points as those for viable counts. The samples were kept on ice, diluted
in pyrogen-free water, filtered and frozen at –20o pending analysis.
Analyses of endotoxin were performed in duplicate with the chromogenic
limulus amoebocyte (LAL) assay (Bio Whittaker, Walkersville, MD, USA).
Determination of possible mechanisms for a
tobramycin-caused reduction in cefuroxime-induced
endotoxin release (I)
Tobramycin at concentrations of 0 (control), 10, 50 and 100xMIC was added
to BHI spiked with endotoxin from E. coli O55:B5 to concentrations of 500
and 2000 EU/mL, respectively. Samples for endotoxin analysis were taken
immediately before addition of tobramycin and after 0.5 h, 2 h, and 4 h of
incubation at 35oC.
In another experiment, tobramycin at similar concentrations as above was
added to bacteria pre-exposed to cefuroxime 10xMIC for 2 h. Samples for
endotoxin analyses were taken before and after addition of tobramycin and
the reduction in the cefuroxime-induced endotoxin release was analyzed and
compared with the control after an additional 2 h.
Determination of SpeA concentration (V)
Samples for SpeA analysis were drawn concomitantly with those for viable
19
counts. At each time point, centrifugation was performed and 5 ml of the
supernatants were transferred to Falcon tubes containing 15 ml of ethanol
and kept at –20oC pending analysis.
The levels of the SpeA present in the concentrated bacterial culture supernatants were quantified by a competitive ELISA described in detail elsewhere (100).
Animals, anaesthesia and preparatory procedures (III)
The study included 13 domestic breed piglets with a median weight of 24 kg.
The animals were 12-14 weeks old and without evidence of illness. Furthermore, the animals had to have a PaO2 >10 kPa (75 mmHg) in arterial blood
and a mean pulmonary arterial pressure (MPAP) of <2.7 kPa (20 mmHg) at
baseline 20 minutes after completed preparatory procedure (see below). The
Ethics committee of Uppsala University, Sweden, approved the experiment.
The piglets were treated according to the European Convention on Animal
Care.
Anaesthesia was induced with the mixture of 6 mg/kg tilétamin-zolazepam
(Zoletil forte vet•, Boehringer Ingelheim Vetmedica, Ingelheim, Germany),
mixed with 2.2 mg/kg of xylazin (Rompun Vet•, Bayer, Leverkusen, Germany) and atropine 0.04 mg/kg (Atropin•10 Pharma, Stockholm, Sweden). Anesthesia was maintained by a continuous infusion of sodium pentobarbital 8 mg/kg per hour (Pentobarbitalnatrium•Apoteket, Umeå, Sweden), pancuronium bromide 0.26 mg/kg per hour (Pavulon•, Organon, Oss,
the Netherlands) and morphine 0.48mg /kg per hour (Morphine•, Pharmacia, Uppsala, Sweden) dissolved in a 2.5% glucose-electrolyte solution.
During the experiment, sodium chloride infusion was given resulting in a
total fluid administration rate of 30 mL/kg per hour.
A bolus dose of 20 mg morphine was given intravenously before the performance of a tracheotomy, which was done in order to secure a free airway
during the experiment and to administer nitrous oxide (N2O) during surgical
preparation. The animals were artificially ventilated throughout the experimental procedure (Servo 900CTM, Siemens Elema, Stockholm, Sweden).
During catheter insertions, 30% oxygen in N2O was given. The gas mixture
was switched to 30% oxygen in air during the rest of the experiment. The
ventilation after preparation was set to yield a PaCO2 between 5.0 and 5.5
kPa. The respiratory rate was 25 per min and the inspiratory-expiratory ratio
was 1:3. Ventilator settings were then kept constant throughout the experi20
ment. Atelectasis was prevented by placing the piglets into prone position
after preparatory procedures and by maintaining a positive end expiratory
pressure of 5 cm H2O.A central venous line as well as a 7 F Swan-Ganz
catheter equipped with thermistor were inserted through the internal jugular
vein into the superior caval vein and the pulmonary artery, respectively. In
addition, a cervical artery was catheterised with an arterial line used for
blood sampling and continuous measurement of the arterial blood pressure.
A minor vesicotomy was performed for the urinary catheter and a heating
pad (Operatherm 200WTM, KanMed, Bromma, Sweden) was used to keep
the animals at a constant body temperature.
Administration of endotoxin and antibiotics (III)
As the preparation was completed, at least 20 minutes of stabilisation time
passed before baseline values were registered and baseline blood samples
were taken prior to administration of antibiotics and endotoxin. All animals
were subjected to endotoxin infusion (Escherichia Coli O111:B4; Sigma
Chemical, St. Louis, MO, USA) with an initial infusion rate of 4 Pg/kg per
hour. After 30 minutes the infusion rate was reduced to 1 Pg/kg per hour
which was continued during the rest of the experiment. At least a two-fold
increase in MPAP (mm Hg) was taken as a sign of severe endotoxemic
shock.
If the mean arterial pressure (MAP) decreased to the level of the MPAP during the first hour of the experiment, a single dose of 0.2 mg adrenaline was
given intravenously. Values were registered hourly after baseline for six
hours after which all surviving piglets were killed by an intravenously injected overdose of potassium chloride.
The animals were randomised by the sealed envelope method to receive
either tobramycin, ceftazidime or saline solution. Tobramycin was given as
an iv infusion of 140 mg in 100 ml saline solution for 20 minutes, starting
10 minutes before initiation of endotoxin infusion. Ceftazidime was injected
as an iv bolus dose of 1 g in 10 ml saline 5 minutes before the start of endotoxin infusion and followed by 90 ml of saline for a total infusion time of
10 minutes. In the control group, an infusion of 100 ml saline was initiated 5
minutes before the start of endotoxin administration and given for 10 minutes.
21
Measurements (III)
MAP and MPAP were continuously monitored through the arterial and the
Swan-Ganz catheter, respectively. The central line was used for injection of
10 mL cold saline for determining the cardiac output (CO). CO was expressed as cardiac index (CI) and body surface area was calculated as
weight0.67 x 0.112. CI was assessed hourly by the thermodilution method
using the thermistor in the Swan Ganz catheter. The average value of at least
three serial measurements was registered. Heart rate (HR) was continuously
monitored. Blood samples for analysis of endotoxin, TNF-D and IL-6 were
drawn immediately before addition of antibiotics and hourly until the end of
the experiment. At the same time points, arterial blood gas and base deficit
were analyzed (ABLTM 300, Radiometer, Brønhøj, Denmark).
Determination of plasma endotoxin, TNF-D and IL-6
concentrations (III)
Blood samples for analysis of plasma endotoxin, TNF-D, and IL-6 concentrations were drawn immediately before addition of antibiotics and hourly
until the end of the experiment at 6 h. Endotoxin-free, heparinized tubes
were used for the endotoxin samples and EDTA tubes for the TNF-D and IL6 samples. After centrifugation the supernatants were transferred to endotoxin-free tubes and subsequently kept at –70oC until analysis.
Analyses of endotoxin were performed in duplicate with the limulus amoebocyte lysate assay (CoatestTM Plasma ChromoLAL; Charles River Endosafe, Charleston, SC, USA). Commercial sandwich enzyme-linked immunosorbent assays were used for detection of IL-6 (QuantikineTM porcine IL6, P6000; R&D Systems, Minneapolis, MN, USA) and TNF-D (KSC3012;
Biosource International, Nivelles, Belgium).
Calculations and statistical methods
Reduction in cfu at various time points was calculated by subtraction of the
remaining number of cfu from that at time zero (I,II,IV,V) and from that at
the time for the second antibiotic dose (IV). The number of cfu at time zero
was calculated as the mean of the values obtained immediately before and
after addition of antibiotics (I, II, V). The endotoxin and SpeA release was
22
calculated similarly (I, II, IV,V). The bacterial killing rate in papers I and II
was calculated as the time needed for the antibiotic to kill 2 log10 cfu and in
the same studies analysis of variance (ANOVA) with repeated measurements
was used to compare propensities to release endotoxin and bacterial killing
rates. Regression analysis was used to determine the relationship between
the number of killed bacteria and the increase in endotoxin (I,II). In papers I
and II, the propensity to release endotoxin was expressed as endotoxin release per killed bacterium, whereas in paper IV, in which the variation in the
viable count was greater, this was expressed as the log10 endotoxin release
per log10 killed bacteria.
To compare the differences in killing rates and SpeA release between antibiotic regimens, a repeated measures ANOVA was performed (V). A Spearman rank correlation test was employed to test relationships between SpeA
release, inoculum size and killing rate.
In the in vivo study (III), the TNF-D response at 1 h and the IL-6 response at
2 h were compared in the primary analyses of the immediate initial cytokine
response. Since there was a variation in the immediate cytokine response, the
reduction in the concentrations of the TNF-D and IL-6 from 1 h and 2 h,
respectively, to 6 h, as well as the concentrations at 6 h, were considered as
late cytokine responses. In order to be able to detect a difference in the initial
log concentration of the cytokines of at least 20 % with an D-error of 0.05, a
E-error of 0.2, a power of 0.8, and a calculated interindividual variation of 14
%, eight evaluable animals should be included in each treatment group. In
this exploratory trial, an interim analysis was planned when there were at
least four animals in each group that had survived for 6 h. The distribution of
the logarithm of the initial cytokine response approximated normal distribution in this model and was therefore analysed by an unpaired t-test. Other
differences between the treatment groups in this study were calculated by the
non-parametric Mann-Whitney U-test.
P-values <0.01 were considered significant in study I due to multiple statistical analyses, in the remaining studies p-values <0.05 were significant. Results were expressed as mean + SE or median ± range.
23
Results and discussion
MIC values (I II, IV,V)
The MICs for the gram-negative strains are shown in Table 1. The difference
in MIC-values for the B049 strain against cefuroxime in papers I (MIC=8
mg/L) and V (MIC=2 mg/L) may be explained by differences in methods
and a method error of plus one step.
Table 1. MICs against cefuroxime and tobramycin for the gram-negative strains.
Bacterial strain
E. coli B 049-3036 (I)
E. coli B 049-3036 (IV)
E.coli CCUG 17620 (II)
E. coli ATCC 25922 (IV)
E. coli B 940-3012 (II)
E. coli B 942-1035 (II)
Klebsiella pneumoniae B 541-1003 (II)
Salmonella enteritidis CCUG 35852 (II)
Neisseria meningitidis CCUG 15117 (II)
Cefuroxime
Tobramycin
8,0
2,0
2,0
4,0
2,0
1,5
2,0
8,0
0,05
0,25
0,5
1,5
1,0
1,0
1,0
1,0
1,0
2,0
The group A streptococcus strain used was highly sensitive to penicillin and
clindamycin, with MIC values of 0.008 mg/L and 0.125 mg/L against penicillin G and clindamycin, respectively.
Killing rates (I, II)
The killing rate was expressed as the time needed for 2 log10 killing. In all
the tested bacteria, including the four E. coli strains, the Salmonella enteriti24
dis, Klebsiella pneumoniae and N. meningitidis strains, there was an increased killing rate after exposure to tobramycin and cefuroxime in combination with tobramycin as compared with cefuroxime alone (p<0.001). At
higher concentrations within the same treatment, the killing rate was significantly greater for all three regimens. The concentration-dependency of the
killing rate was greater for tobramycin and tobramycin plus cefuroxime than
for cefuroxime alone, which was confirmed by using interaction terms in the
multivariate analysis (p<0.01).
Antibiotic-induced endotoxin release (I, II)
The PBP-3 specific cefuroxime led to a considerably greater endotoxin release than tobramycin or the combination of the two at all concentration
levels, even if there was a large species variation (Table 2).
25
Table 2. Increase in endotoxin release after 2 and 4 hours. EU/ml + SE.
Cefuroxime
2xMIC
10xMIC
50xMIC
E.coli CCUG
2h
4h
141+80 3006+1484
414+83 4186+2470
292+50 5027+3602
E.coli B940
2h
4h
843+272 2699+534
452+181 3727+654
520+217 2656+587
E.coli B942
2h
4h
791+296 6563+769
1038+283 7060+882
1312+478 5876+858
Tobramycin
2xMIC
10xMIC
50xMIC
140+81
197+58
134+26
250+13
197+34
139+65
242+31
142+15
144+22
199+61
124+48
168+78
200+51
122+45
156+10
300+69
182+34
160+37
20+2
21+10
19+7
41+13
41+16
29+8
645+231
348+88
296+68
851+220
590+160
383+103
2+1
6+3
4+2
14+8
6+3
5+3
Cefuroxime +
Tobramycin
2xMIC
10xMIC
50xMIC
160+11
130+29
97+18
202+9
193+17
142+11
216+32
184+47
194+49
238+43
218+49
238+81
199+21
157+25
205+17
263+28
204+30
215+6
108+26
81+21
89+21
190+59
119+32
112+29
823+115
467+63
420+47
1165+288
589+88
459+88
10+4
11+3
8+1
35+8
20+5
21+5
26
Salmonella
2h
4h
97+50
1163+79
1064+62 1421+145
426+61
477+26
Klebsiella
2h
4h
1848+1288 3997+1724
2754+1953 4653+2011
1957+871
3126+1275
Neisseria
2h
4h
2+2
83+51
6+4
76+52
24+17
49+36
Most of the endotoxin was liberated within 2 h after exposure to tobramycin
or the combination, whereas an important additional increase was seen between 2 and 4 h from most of the strains exposed to cefuroxime alone (I,II).
The prolonged and high endotoxin liberation by cefuroxime can be explained by the PBP 3-associated filamentous elongation of the bacteria with
continuing endotoxin production and liberation (24,78).
Endotoxin release in relation to the number of killed
bacteria (I, II)
For the E. coli B049-3036, investigated in paper I, there was a linear relationship between the endotoxin release and the number of killed bacteria
during the first 4 h for each antibiotic regimen at each dosage level, with r2
values 0.61-0.96. This indicated a proportionality, and that the propensity to
release endotoxin could be expressed as endotoxin per killed bacterium.
At 4 h, there was a significant difference between the antibiotics, with the
highest propensity to release endotoxin for cefuroxime, lower for tobramycin
and the lowest for the combination of the two (p<0.001) (Fig.3).
Figure 3. Endotoxin release per killed bacterium 4h after exposure to antibiotics at
2xMIC, 10xMIC, and 50xMIC.
27
With increasing doses, there was a significant reduction (p<0.001) in endotoxin release per killed bacterium at 10xMIC and 50xMIC for tobramycin
and the combination and at 50xMIC also for cefuroxime alone. It seems
reasonable to assume that the reduction in the propensity to release endotoxin by cefuroxime at the highest doses is due to changes in PBP-affinity,
as has earlier been shown for ceftazidime (78). Thus, the affinity for PBP 3
decreases in favour of PBP 1, resulting in bacterial lysis without filament
formation and with less endotoxin release. When the bacterial killing rate
was correlated to the propensity to release endotoxin in bacteria exposed to
tobramycin or the combination of tobramycin and cefuroxime, a significant
negative correlation was found.
The intention with paper II was to see if the demonstrated reduction in the
propensity to release endotoxin by tobramycin in one strain was applicable
to other E. coli-strains and to other gram-negative species. Moreover, the
proportionality between endotoxin release and number of killed bacteria,
observed in our previous study, was studied over the wider range of bacteria
often seen in clinical practice.
After 4 h of antimicrobial exposure, large species variations in the propensity to release endotoxin were demonstrated. After exposure to cefuroxime,
the Klebsiella strain liberated 6.3 – 9.9 x 10-3 EU/bacterium, whereas the N.
meningitidis strain released considerably less, 0.1 - 0.4 x 10-3 EU/bacterium.
Even among the E. coli strains, there was a difference in cefuroxime-induced
liberation, the B 940 strain releasing only half of that of the B 942 strain.
Corresponding variations were seen after exposure to tobramycin or the
combination of cefuroxime and tobramycin, but these were of smaller magnitude. The propensity to release endotoxin upon antimicrobial exposure
seems to be a strain-specific characteristic, since there was a significant correlation between the release caused by tobramycin and that by cefuroxime
(p<0.05 at each concentration level). The endotoxin released by the N. meningitidis strain was remarkably low in comparison with that of the other
strains, especially when the exposure to cefuroxime was compared. The
findings are in agreement with the results from an ex vivo study by Prins et
al. (79). It may also be the result of the large intraspecies variation that has
earlier been described for meningococci (5). In addition, PBP 1, 2 and 3
have all been found in the meningococcal cell wall and varying affinities to
E-lactam antibiotics have been described for all of them (67,75,82).
In the endotoxin release per killed bacterium after 4 h, there was a significant treatment effect in the E.coli CCUG strain (p<0.001), the B940-3012
strain (p<0.001), the B942-1035 strain (p<0.001), the Klebsiella strain
(p<0.001), and the Salmonella strain (p<0.01). Paired comparisons demon28
strated highly significant differences between cefuroxime and the combination of tobramycin and cefuroxime. The reduction was 96%, 93%, 97%,
86%, and 85%, respectively, from that caused by cefuroxime alone. Similarly, the difference between cefuroxime and tobramycin alone was large and
strongly significant. For the N. meningitidis strain, these differences did not
reach statistical significance but a similar trend was seen with a reduction of
52% when tobramycin was added to cefuroxime. In none of these strains
there was a statistical difference between the combination and tobramycin
alone.
After 4 h, there was a significant dose-dependent reduction in endotoxin
release per killed bacterium in the Salmonella, the E. coli CCUG, the E.coli
B942, and the Klebsiella strains (p<0.05 for each strain). The reductions in
tobramycin-induced endotoxin release from 2xMIC to 50xMIC were 67%,
48%, 46% and 29%, respectively. Corresponding values for the combination
were 44%, 33%, 16%, and 41%, respectively. In the N. meningitidis strain, a
similar trend was demonstrated (p=0.05), with a mean reduction of 71% and
29% by tobramycin and the combination, respectively. No dose-dependency
at all was seen in the E. coli B940 strain.
There was no significant dose-dependency in the cefuroxime-induced release
in any of the strains.
Despite large variations in antibiotic-induced endotoxin release, there was in
all strains a considerable reduction in the cefuroxime-induced endotoxin
release per killed bacterium when tobramycin was given concomitantly. This
indicates that the effect observed in the previously investigated E. coli strain
seems to be applicable to many other gram-negative bacteria. Similarly,
dose-dependency with a reduction in endotoxin release at higher tobramycin
concentrations was seen in all strains with the exception of the E. coli B940
strain. This response constitutes another advantage besides the increase in
bacterial killing rate and the reduction in toxicity associated with the once
daily high-dose regimen of aminoglycosides.
In a separate experiment, the endotoxin release of K. pneumoniae with different starting inoculae (103, 105 and 107 cfu/ml) after exposure to 2x MIC of
cefuroxime, tobramycin or a combination of cefuroxime and tobramycin,
was studied in relation to the number of killed bacteria (Fig 4).
29
Figure 4. Klebsiella pneumoniae. Correlation between endotoxin release and number of killed bacteria after 4h at starting inoculae of 103, 105 and 107 cfu/ml, respectively.
The correlations between logarithmic values of endotoxin release and number of killed bacteria were highly significant for cefuroxime (r=0.99;
p<0.001), tobramycin (r=0.93; p<0.001), and the combination (r=0.99;
p<0.001). The slopes of the regression lines did not differ significantly, the
mean value being 0.75+0.04. This implies that there is a considerably higher
release at higher bacterial concentrations but per killed bacterium the endotoxin release is somewhat lower. The reason for this is not known but it
may be speculated that a high environmental endotoxin concentration may
lead to formation of endotoxin aggregates and a reduction in the biologial
activity. The magnitude of the slope demonstrates that within 1 log10 the
deviation from the straight line leads to only a small systematic error when
expressing the endotoxin release in terms of release per killed bacterium and
in comparison with the effect on endotoxin release due to variations in inoculum size, this error is negligible.
Possible mechanisms for reduction in endotoxin release
by tobramycin (I,III)
The initial hypothesis to explain the reduction in endotoxin release by the
aminoglycosides was by an unspecific binding similar to that demonstrated
30
for polymyxin B (47,78,85). However, in paper I it was shown that tobramycin at 2.5, 12.5 and 25mg/L added to BHI spiked with endotoxin did not
reduce the endotoxin activity in the solution as compared with the control
In a separate experiment with cefuroxime at a concentration of 10xMIC in a
bacterial solution, it was demonstrated that if the addition of tobramycin was
delayed for 2 h, no reduction in the cefuroxime-induced endotoxin release
took place, neither immediately nor after 4 h. Thus, neutralisation can be
excluded as the principal explanation of the reduction in cefuroxime-induced
endotoxin release.
The absence of in vitro neutralisation indicates that there is no aminoglycoside binding that affects the biological activity in the limulus amoebocyte
lysate assay but, on the other hand, a tobramycin binding that in vivo neutralises the endotoxin effect cannot be excluded. Inhibition of the binding of
endotoxin to LBP, the most important carrier protein in the body which is
crucial for the attachment to the CD14 receptor on the macrophages, represents one possible mode of action. Another might be that binding with LBP
is unaffected, whereas the tobramycin-endotoxin-LBP complex has a reduced affinity for the CD14 receptor or a lower potential to elicit a signal via
TLR 4 for a subsequent TNF-D- and IL-6- production.
In the porcine in vivo model (III), the primary endpoints were therefore to
analyse the immediate TNF-D and IL-6 responses after 1 h and 2 h, respectively, in relation to the endotoxin concentration. These time points were
chosen since cytokine values obtained later might be influenced by an aminoglycoside-induced protection against oxidative injury, which has been
demonstrated in vitro (15). The plasma TNF-D levels reached peak values 1
h after the start of endotoxin infusion with a median value of 7265 ng/L
(range 2492 – 10625 ng/L). There were no significant differences in peak
TNF-D values between the treatment groups. TNF-D concentration decreased to baseline levels within 3 h in most animals.
Plasma IL-6 levels peaked after 2 h or 3h with a median maximum IL-6
value of 2372 ng/L (range 1544–7138 ng/L). There were no significant differences in IL-6 concentration at 2 h between treatment groups (Fig. 5).
31
Figure 5. Plasma IL-6 concentrations 2-6 hours after start of endotoxin infusion
(n=4 in each group).
These results indicate that there is no in vivo neutralisation of endotoxininduced cytokine production and consequently that there is no interaction
with the binding to LBP or the CD14 receptor.
In order to maximise a possible neutralising effect, the tobramycin dose chosen was in the highest range of that used clinically. Nonetheless, neither
tobramycin nor ceftazidime had any endotoxin neutralising effect in the LAL
assay in comparison with that caused by saline. These results are thus in
agreement with the in vitro results described in paper I.
An alternative hypothesis to endotoxin neutralisation may be that the mode
of action of tobramycin is responsible for the reduced release of endotoxin.
The significant negative correlation between the bacterial killing rates of
tobramycin and the combination of tobramycin and cefuroxime, on the one
hand, and the endotoxin release on the other, supports this hypothesis. Failure of tobramycin to reduce cefuroxime-induced endotoxin release if added
after a 2 h pre-exposure to cefuroxime, indicates that initial mechanisms
probably are the most important. Previously, Kusser et al. have shown that
aminoglycosides are able to inhibit the synthesis of endotoxin (53). Our results strongly support that such a mechanism is quantitatively more important in comparison with neutralization for the reduction in endotoxin release.
32
Possible anti-inflammatory effects by antibiotics (III)
The inflammatory damage to endothelial cells and other tissue cells mediated by oxygen radicals induces cytokine production (55,76). In several in
vitro experiments, it has been demonstrated that ceftazidime and some of the
other E-lactam antibiotics inactivate hypochlorous acid (HOCl) (21) and, in
the case of ceftazidime, also singlet oxygen (66). In clinically relevant concentrations, ceftazidime has also been shown to partially protect endothelial
cells from the oxidative stress of activated neutrophils in vitro (55). The
documentation for aminoglycosides is less extensive but both tobramycin
and gentamicin have been shown to protect lung epithelial cells against myeloperoxidase-dependant oxidant injury by binding to anionic cell surfaces
and neutralising HOCl (15). These findings led to the hypothesis that the
anti-inflammatory effects might affect the cytokine response during the last
phase of our porcine model. At the end of the experiment, at 6 h, there was a
treatment effect expressed as a significantly greater reduction in IL-6 by
ceftazidime as well as tobramycin (p<0.05) and a lower absolute IL-6 level
in the ceftazidime group as compared with placebo (p<0.05). Since there are
in vitro data that ceftazidime neither directly affects IL-6 production (66) nor
neutrophil function (34). this result may have been caused by its antioxidative properties. For tobramycin, in vitro data indicate that neutrophil function
and oxidative metabolism might be influenced but these results have not
been reproduced in vivo (34). Whether our findings are due to antioxidative
properties of the antibiotics or to other mechanisms cannot be concluded
from data presented, but the result indicates that in vitro reported antiinflammatory properties may be demonstrable in an animal model.
Effect of repeated dosing on endotoxin release and
morphology (IV)
In an in vitro kinetic model, simulating the human concentration timeprofile, the endotoxin release from two E. coli strains after exposure to two
repeated doses of cefuroxime was investigated. In addition, it was studied
whether the combination with tobramycin at the second dose resulted in a
reduction similar to that previously shown when combined with the first
dose.
33
Figure 6. Bacterial killling rate (mean + SE) and endotoxin release (geometric mean
+ antilog of the SE of the logaritmic values) after the first and second dose of cefuroxime in the ATCC strain (a) and the B049 strain (b). Note that a considerably
lower number of bacteria were killed after the second dose.
34
There was a large variation in endotoxin concentrations, mainly due to the
variation in the number of killed bacteria. Median endotoxin release after
correction for elimination via the outflow of the pump was 2 h and 4 h after
the first dose 1720 and 12600 EU/ml, respectively, for the ATCC strain and
8570 and 21100 EU/ml, respectively, for the B049 strain. Corresponding
values after the second dose were 102000 and 124000 EU/ml for the ATCC
strain (p<0.01) and 128000 and 149000 EU/ml (NS) for the B049 strain,
respectively. The geometric mean of the endotoxin release is shown in
Fig. 6.
When the endotoxin release was related to the number of killed bacteria,
there was a marked reduction in the variation in both strains. The increase in
the tendency to release endotoxin after the second dose in comparison with
that after the first was highly significant (p<0.001 for both strains). The difference between the doses ranged from 0.15 to 0.26 log10 EU/log10 number of
killed bacteria. This implies that the difference between the doses will be
greater at higher numbers of killed bacteria and this will, in effective treatment, covariate with the bacterial concentration at the time of the dose. The
time course was also significantly different, with an earlier release after the
second dose (p<0.001 for both strains).
One possible mechanism behind the higher endotoxin release after the second dose of cefuroxime might be the continued release from remaining filaments caused by the first dose. This hypothesis is supported by Jackson and
Kropp, who found an increased endotoxin release at sub-MIC concentrations
of several E-lactam antibiotics (48). Our electron microscopy findings just
before the second dose with a few elongated forms may also be in agreement
with this. However, at that time, the endotoxin concentration was relatively
low, indicating that even if there may be some sustained release, the contribution of this to the total release after the second must be limited. A change
in PBP affinity with a higher binding to PBP 3 than to PBP 1 represents another possibility, but the presence of spheroplasts together with filaments
after the second dose does not favour this hypothesis. Thus, the mechanism
is not clear but it might be speculated that there is an enduring antibiotic
effect on cell wall synthesis after the first dose resulting in a quicker and
more extensive bacterial elongation after the second dose.
After addition of tobramycin to the second dose of cefuroxime, the propensity to release endotoxin was significantly reduced in comparison with that
after cefuroxime alone (p<0.001), even if the bacteria had previously been
exposed to cefuroxime. In addition, the filamentation was less pronounced.
The effect of a repeated combined aminoglycoside E-lactam antibiotic treatment could not be studied, because, at the clinical concentrations used all
35
bacteria were killed already after the first dose at the inoculum sizes possible
to use in our in vitro kinetic model.
Exotoxin release (V)
The purpose was to investigate the amount and time course of SpeA liberation from a GAS strain, at different starting inoculae, after exposure to different doses of penicillin, clindamycin or a combination of the two, and to
relate the release to the bacterial concentration and the killing rate.
The starting inoculum sizes in the low and high inoculum experiments were
7.0 + 0.1 and 8.6+ 0.2 log10 cfu/ml, respectively. At the low, as well as the
high initial inoculum, a considerably higher killing rate was seen for penicillin than for the combination of penicillin and clindamycin (p<0.01 and
p<0.001, respectively), but there was no difference in bacterial killing rate
between low and high antibiotic doses within any of the treatment regimens.
A similar result was seen in the high inoculum experiment when penicillin
was compared with clindamycin alone (p<0.01), whereas in the low inoculum experiment, only a trend in the similar direction was observed (p=0.09).
In contrast to antibiotic-induced endotoxin liberation from gram-negative
bacteria which, after exposure to E-lactam antibiotics, initially leads to
higher endotoxin concentration than in the control (24), all treatment regimens in the present study resulted in a SpeA release, which was many times
lower than that of the control. In the high inoculum experiments (Fig.7),
there was a large variation in the SpeA production, from 39+9 ng/ml for low
dose penicillin to -2+3 ng/ml for the high dose of the combination. The toxin
release was significantly higher at low than at high concentrations of penicillin (p<0.05). This dose-dependency was also seen for the combination
(p<0.05) but not for clindamycin alone. Since there was no difference in
killing rate between the two doses this cannot explain the result. For certain
E-lactam antibiotics, such as ceftazidime or cefuroxime, binding to different
E.coli PBPs has been demonstrated at varying concentrations, which results
in a higher endotoxin release after low than after high doses (71).It may be
speculated that one possible mechanism is that penicillin has a similar dosedependent varying affinity to different PBPs. This would imply a different
binding pattern at low and high concentrations, which then may lead to a
divergent effect on cellular activity prior to lysis. Varying clindamycin concentrations did not influence the SpeA release but addition of clindamycin
could not abolish the dose-effect of penicillin in the combination.
36
Figure 7. SpeA production in the high inoculum experiment, at low (a) and high (b)
antibiotic doses, Pg/L + SE.
37
In the low inoculum experiments, the highest release of SpeA was seen at
low penicillin concentration. However, the release was >1 log10 lower than
that seen in the high inoculum experiment and no statistically significant
differences between treatments could be demonstrated.
Clindamycin alone released less SpeA than penicillin, which, however, was
significant only at low concentrations (p<0.01). When clindamycin was
added to penicillin, there was a significant reduction in SpeA release both at
low and high concentrations (p<0.01 for both comparisons)which is in accordance with an in vitro study by Coyle et al. (17). The most probable explanation of this is that clindamycin, as a protein synthesis inhibitor, reduces
the production of toxins or proteins responsible for their release to the environment (52,89).
Most of the speA was released within the first 3 h. The SpeA production
during this period correlated significantly to the number of killed bacteria
during the same period (n=18, r=0.50; p<0.05). When analysed separately
for each treatment regimen, a positive correlation could be demonstrated
after treatment with penicillin and the combination, the latter being significant (p<0.01), whereas no similar correlation was seen for clindamycin.
Within the limited range of the inoculum size of 0.8 log10 cfu/ml, there was a
significant positive correlation between this and the speA concentration at
time zero (r=0.54; p<0.05). In contrast, there was a trend towards a negative
correlation to the increase during the following hours (r=-0.42; NS), being
negative for all treatment regimens. Therefore, variation in the initial inoculum size does not explain the correlation between toxin release and number
of killed bacteria. The reason for this is not clear, but the killing rate of Elactam antibiotics has been associated with growth rate and hence the metabolic activity of the bacteria (90), and it may therefore be speculated that an
initially higher metabolism also results in an increased toxin production.
When a protein synthesis inhibitor, such as clindamycin, shuts this off, the
relationship is no longer seen.
In several experimental infections of animals with GAS, clindamycin has
turned out to be more effective than penicillin (12,91). Therefore, the combination of penicillin and clindamycin has often been recommended for the
treatment of severe GAS infections (41,87). Even if there is a limited reduction in the killing rate, this seems to be outweighed by the reduced inoculum
effect (90,91), longer post-antibiotic effect (89), suppressed synthesis of
virulence factors, such as M-protein and capsule (12,32), enhanced opsonisation and phagocytosis (32) and decreased toxin release by clindamycin.
38
Clinical implications
In animal models, there is strong evidence that soluble endotoxin released
from gram-negative bacteria by antibiotics contributes to pathogenesis and
mortality of the sepsis (70). However, the clinical significance of antibioticinduced endotoxin release is still not known (78). Anecdotal clinical data
from the studies by Dofferhoff et al. and Arditi et al. demonstrate that increases in endotoxin concentration may be associated with clinical deterioration (7,24) and the results from a prospective clinical study in patients with
gram-negative urosepsis by Prins et al. suggest that the difference between
imipenem and ceftazidime in endotoxin release may also be reflected by
differences in the systemic inflammatory response (77). Furthermore, a retrospective study by Mock et al. has demonstrated that mortality was higher
in trauma patients treated with endotoxin-releasing PBP 3-specific antibiotics than in those given other antibiotics, which lends some support to a possible clinical significance (69).
On the other hand, there are clinical studies that do not find any differences
in the concentrations of endotoxin and cytokines or in clinical outcome between PBP 2-specific imipenem and PBP 3-specific ceftazidime treatment of
various gram-negative infections (14,60,63). In a study by Maury et al., in
which patients had both a verified gram-negative bacteremia and a severe
infection in the form of severe sepsis or septic shock, endotoxin concentration was significantly decreased after one and four hours of antibiotic treatment, suggesting that antibiotic-induced endotoxin release is not a clinically
relevant phenomenon (65). However, in this non-comparative study, the total
endotoxin concentration was measured, also including bacterial cell-bound
endotoxin. As has been shown by Arditi et al. in studies on children with
gram-negative meningitis, there is a fall in total endotoxin concentration that
parallels the reduction in bacterial count, whereas the free endotoxin level
concomitantly increases (7). In another study, Hurley et al. found that free
endotoxin concentrations increased when chronically bacteriuric patients
were treated with different antibiotics (44) Thus, if measured as free endotoxin, which is the most biologically active form, antibiotic-induced endotoxin liberation has been demonstrated to occur also in patients.
39
In clinical trials, there are difficulties to prove an effect between treatments,
even if there is one, due to the large variations in the endotoxin release. As
shown in our study, endotoxin liberation is highly dependent on the number
of bacteria which may clinically vary considerably. If the release of endotoxin is related to the number of bacteria, it may be possible to reduce the
interindividual variation and consequently increase the chance in clinical
trials to detect a difference between treatments. However, even if this variation is reduced, the strain specific propensity to release endotoxin still represents an important confounding factor.
Taken together with the similarity between the responses of human volunteers and animals to administration of endotoxin, the resemblance of this
response to the symptoms seen in sepsis and septic shock (96), and the reduction in systemic inflammatory response demonstrated in at least one prospective and controlled study (77), available data suggest that reduction in
endotoxin release may be of clinical importance. In patients with uncomplicated gram-negative infections, antibiotic-induced endotoxin release is most
probably of limited value. On the other hand, in patients with progressing
severe sepsis or who have developed septic shock, an additional dose of
endotoxin may be deleterious (58).
Despite a great inter- and intraspecies variation among the gram-negative
bacteria, it was possible in our in vitro endotoxin studies to demonstrate
significant reductions in endotoxin release when the combination of the PBP
3-specific E-lactam antibiotic, cefuroxime, and the protein synthesis inhibitor, tobramycin, was compared with cefuroxime alone. Furthermore, increasing doses of tobramycin led to a lower endotoxin release and a higher bacterial killing rate. If tobramycin was added as late as 2 h after cefuroxime administration, no inhibitory effect on the endotoxin release was registered.
These findings indicate that it is probably an advantage to add an aminoglycoside to the E-lactam antibiotic for the treatment of severe infections, such
as progressing severe sepsis and septic shock, a combination that has been
recommended for a long time (101). Besides a broad spectrum and a high
killing rate, also the release of endotoxin is diminished. If the treatment is
initiated with a PBP 3-specific antibiotic but the clinical condition of the
patient deteriorates, an aminoglycoside dose may be given concomitantly or
before the second dose. Our data also indicate that it is an advantage to give
a high dose of the aminoglycoside. High doses have also been recommended
in critically ill patients due to the increase in the volume of distribution
(13,40). However, the beneficial effects should be weighed against the
nephro- and neurotoxic effects of the aminoglycosides. In a study in critical
care patients with a creatinine clearance of >70 ml/min, very high doses
40
were given for six days without impairment of renal function, provided that
the aminoglycoside concentrations were followed (54). In these patients,
kidney function was more or less preserved but, a recent study in patients
with poor renal function and peritoneal dialysis has demonstrated that aminoglycoside treatment did not affect residual renal function (9). The use of
aminglycosides in a high dose regimen in oliguric or anuric patients or patients who present with a rapidly decreasing renal function needs further
investigation. However, taken together, these data indicate that one or a few
doses of an aminoglycoside in the initial phase of the treatment of severe
sepsis or septic shock may have beneficial effects and that the risk of renal
injury, with the exception of the patients with the poorest function, seems to
be very low.
In some similarity with the discussion above, there seems to be an advantage
to combine penicillin with the protein synthesis inhibitor, clindamycin, for
the treatment of severe group A streptococcal infections.
In animal studies clindamycin alone, or in combination with penicillin, has
shown beneficial effects in the treatment of invasive group A streptococcal
infections (56,91). This is in part due to inhibited expression of different
streptococcal virulence factors, such as SpeA (62,90).
The findings in our in vitro study that clindamycin released less SpeA from a
GAS strain than penicillin and that the addition of clindamycin to penicillin
resulted in less toxin production than treatment with penicillin alone, are in
agreement with earlier in vitro and animal experiments. The study also demonstrated that the SpeA concentration was dependent on the number of bacteria prior to antibiotic exposure and the dose of penicillin. In the clinical
setting, early treatment seems to be very important and, in addition, combination therapy from the start with high doses of penicillin may offer an advantage.
41
Conclusions
The in vitro release of endotoxin is associated with the number of killed
bacteria, the logarithm of the release being proportional to the logarithm of
the number of killed bacteria.
The protein synthesis inhibitor, tobramycin, reduces endotoxin release in
vitro from various gram-negative strains induced by the PBP 3-specific Elactam antibiotic, cefuroxime. Increasing doses of tobramycin lead to less
endotoxin liberation.
There is no unspecific binding between endotoxin and tobramycin in vitro.
Nor does tobramycin affect the binding of the LBP-endotoxin complex to the
macrophages in vivo.
The reduction in endotoxin release is associated with the killing rate of tobramycin.
The propensity to liberate endotoxin in vitro is higher after the second dose
of cefuroxime than after the first, which results in a higher release of endotoxin than expected from the bacterial count. The release of the second
dose can be reduced by the addition of tobramycin
There is a possible anti-inflammatory effect given by ceftazidime and tobramycin expressed as a late cytokine inhibition in porcine endotoxin shock.
The SpeA release from a GAS strain in vitro is correlated to the initial number of bacteria and the killing rate. Addition of a protein synthesis inhibitor,
clindamycin, reduces the toxin release in comparison with that of the Elactam antibiotic, penicillin G, alone. The SpeA release is also dependent on
the dose of penicillin, with high doses leading to lower SpeA levels as compared with low doses.
42
Acknowledgements
I would like to express my sincere gratitude to all the people who made this
thesis possible. In particular I would like to thank:
Jan Sjölin, my tutor, who persuaded me to undertake this project and since
then, has always been available for questions and discussions, despite an
immense work load. I am most grateful for your never ending support and
enthusiasm, your enormous knowledge and your efforts to find new solutions when everything looked hopeless, especially the statistics.
Monica Lundholm, my co-tutor, unfortunately no longer with us. Monica
introduced me to the endotoxin studies in the Laboratory of Microbiology
and were most helpful and supporting during my first not so skilful activities
in the lab. R.I.P.
Inga Odenholt, my co-tutor, always optimistic and encouraging. Your
pharmacokinetic knowledge, advices and critical comments were most valuable, especially before you left us for Malmö.
Eva Tano, co-author, my support in the laboratory during the last years, as
well as my talking partner when problems arose. Without your skill in planning and performing the laboratory work, day or night, there would have
been nothing. You are a real guarantee of quality.
Caroline Stenbom-Johansson and Christina Nylander, skilful laboratory
technicians performing the analyses of the first two studies. I will remember
our sports discussions and other fun during work.
Anders Nordgren, intensive care nurse, for giving anaesthesia to the pigs,
supervising them as if they were human patients and taking care of everything.
All my co-authors in the papers. Especially I want to thank Miklos Lipcsey for doing the invasive operations on the pigs, Anna Norrby-Teglund
for introducing me into the “superantigenic” world and Pontus Thulin for
43
the time-consuming work with the SpeA-analyses.
Professor Göran Friman for allocating research time during the years.
Elisabeth Löwdin, friend and loyal room-mate since 1986, lately also coauthor and always willing to take on my duties in the clinic when necessary.
In our office, where visitors have to find space between desks and bikes
there has been time for laughter as well as gossip talk.
Anders Lannergård, Cecilia Ryden and Erik Torell, my colleagues, for
doing my clinical work during the last crucial weeks of the thesis.
Birgitta Sembrant, you never stop impressing me with your technical skills
with figures and tables in any document form. A life-line during the final
stress.
Osvaldo, my husband and life time supporter. Thank you for your patience,
consideration and for serving coffee, porridge and dinners during the breaks.
Tomas and Pia, Jens, Hannes, Jan and Nea, all of you remind me of other
qualities in life like the travels and outdoor activities we have enjoyed together and hopefully will continue to do.
For financial support I would like to thank: The Olinder-Nielsen Foundation,
Eli Lilly Sweden AB, Stockholm, Sweden, Glaxo Wellcome AB, Gothenburg, Sweden, Laerdal Foundation for Acute Medicine, Swedish Foundation
for Strategic Research and the Swedish Research Council.
44
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Abraham E, Laterre PF, Garbino J et al. (2001). Lenercept (p55 tumor necrosis factor receptor fusion protein) in severe sepsis and early septic shock: a
randomized, double-blind, placebo-controlled, multicenter phase III trial with
1,342 patients. Crit Care Med 29(3), 503-10.
Abraham E, Reinhart K, Opal S et al. (2003). Efficacy and safety of tifacogin
(recombinant tissue factor pathway inhibitor) in severe sepsis: a randomized
controlled trial. JAMA 290(2), 256-8.
Abraham E, Wunderink R, Silverman H et al. (1995). Efficacy and safety of
monoclonal antibody to human tumor necrosis factor alpha in patients with
sepsis syndrome. A randomized, controlled, double-blind, multicenter clinical
trial. TNF-alpha Sepsis Study Group. JAMA 273(12), 934-41.
The ACCP/SCCM Consensus Conference Committee. (1992). Definitions for
sepsis and organ failure and guidelines for the use of innovative therapies in
sepsis. Chest 101, 1644-55.
Andersen BM, Solberg O. (1984). Endotoxin liberation and invasivity of Neisseria meningitidis. Scand J Infect Dis 16, 247-54.
Annane D, Sebille V, Charpentier C et al. (2002). Effect of treatment with low
doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA 288(7), 862-71.
Arditi M, Ables L, Yogev R. (1989). Cerebrospinal fluid endotoxin levels in
children with H. influenzae meningitis before and after administration of intravenous ceftriaxone. J Infect Dis 160, 1005-11.
Artenstein AW, Cross AS. (1989). Inhibition of endotoxin reactivity by aminoglycosides. J Antimicrob Chemother 24, 826-8.
Baker RJ, Senior H, Clemenger M et al. (2003). Empirical aminoglycosides
for peritonitis do not affect residual renal function. Am J Kidney Dis 41, 670-5.
Bernard GR, Margolis BD, Shanies HM et al. (2004). Extended evaluation of
recombinant human activated protein C in the United States Trial (ENHANCE
US): a single-arm, phase 3B, multicenter study of drotrecogin alpha (activated) in severe sepsis. Chest 125, 2206-16.
Bingen E, Goury V, Bennani H et al. (1992). Bactericidal activity of ß-lactams
and amikacin against Haemophilus influenzae: effect on endotoxin release. J
Antimicrob Chemother 30, 165-72.
Brook I., Gober A.E, Leyva F. (1995). In vitro and in vivo effects of penicillin
and clindamycin on expression of group A beta-hemolytic streptococcal capsule. Antimicrob Agents Chemother 39:7, 1565-8.
Buijk SE, Mouton JW, Gyssens IC et al. (2002). Experience with a once-daily
dosing program of aminoglycosides in critically ill patients. Intensive Care
Med 28, 936-42.
Byl B, Clevenbergh P, Kentos A et al. (2001). Ceftazidime- and imipeneminduced endotoxin release during treatment of gram-negative infections. Eur J
Microbiol Infect Dis 20, 804-7.
45
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
13. Cantin A, Woods DE: (1993). Protection by antibiotics against myeloperoxidase-dependent cytotoxicity to lung epithelial cells in vitro. J Clin Invest
91, 38-45.
14. Cohen J. (1996). INTERSEPT: an international, multicenter, placebocontrolled study of monoclonal antibody to human necrosis factor-alpha in patients with sepsis. International Sepsis Trial Study Group. Crit Care Med
24(9), 1431-40.
Coyle EA, Cha R, Rybak MJ. (2003). Influences of linezolid, penicillin, and
clindamycin, alone and in combination, on streptococcal pyrogenic exotoxin A
release. Antimicrob Agents and Chemother 47, 1752-5.
Crosby HA, Bion JF, Penn CW et al. (1994). Antibiotic-induced release of
endotoxin from bacteria in vitro. J Med Microbiol 40, 23-30.
Cunningham MW. (2000). Pathogenesis of group A streptococcal infections.
Clin Microbiol Rev 13, 470-511.
Darenberg J, Ihendyane N, Sjolin J et al. StreptIg Study Group. (2003). Intravenous immunoglobulin G therapy in streptococcal toxic shock syndrome: a
European randomized, double blind, placebo-controlled trial. Clin Infect Dis
37(3), 341-3.
Deby-Dupont G, Deby C, Mouithys-Mickalad A et al. (1998). The antibiotic
ceftazidime is a singlet oxygen quencher as demonstrated by ultra-weak
chemiluminescence and by inhibition of AAP consumption. Biochim Biophys
Acta 8:61-8.
Dhainault JF, Tenaillon A, Le Tulzo Y et al. (1994). Platelet activating factor
receptor antagonist BN 52021 in the treatment of severe sepsis: A randomized,
double-blind, placebo-controlled multi-center clinical trial. Crit Care Med 22,
1720-8.
Dofferhoff ASM, Esselink MT, de Vries-Hospers HG et al. (1993). The release of endotoxin from antibiotic-treated Escherichia coli and the production
of tumor necrosis factor by human monocytes. J Antimicrob Chemother 31,
373-84.
Dofferhoff ASM, Nijland JH, de Vries-Hospers HG et al. (1991). Effects of
different types of combinations of antimicrobial agents on endotoxin release
from Gram-negative bacteria: an in vitro and in vivo study. Scand J Inf Dis 23,
745-54.
Eagle H. (1952). Experimental approach to the problem of treatment failure
with penicillin. 1. Group A streptococcal infection in mice. Am J Med 13, 38999.
Flegel WA, Baumstark MW, Weinstock C et al. (1993). Prevention of endotoxin-induced monokine release by human low- and high-density lipoproteins and by apolipoprotein A-1. Infect Immun. 61, 5140-6.
Foca A, Matera G, Iannello D et al. (1991). Aminoglycosides modify the in
vitro metachromatic reaction and murine generalized Shwartzman phenomenon induced by Salmonella minnesota R595 lipopolysaccharide. Antimicrob
Agents Chemother 35, 2161-4.
Freeman R, Gould EK. Prevention of fever and gramnegative infection after
open heart surgery by antiendotoxin. (1985). Thorax 40, 846-8.
Frey EA, Miller DS, Gullstein Jahr T et al. (1992). Soluble CD14 participates
in the response to endotoxin. J Exp Med 176, 1665-71.
Gazzano-Santoro H, Parent JB, Grinna L et al. (1992). High-affinity binding
of the bactericidal/permeability increasing protein and a recombinant amino-
46
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
terminal fragment to the lipid A region of lipopolysacharide. Infect Immun
1992, 4754-61.
Gemmel CG. (1993). Antibiotics and neutrophil function – potential immunomodulating activities. J Antimicrob Chemother 31 Suppl B, 23-33.
Gemmel CG., Peterson PK, Schmeling D et al. (1981). Potentiation of opsonization and phagocytosis of Streptococcus pyogenes following growth in the
presence of clindamycin. J Clin Invest 67, 1249-56.
Gilbert DN. (2000). Aminoglycosides. In: Mandell GL, Bennet JE, Dolin R,
eds. Principles and practice of infectious diseases, 5th edn. New York: Churchill Livingstone, 690-705.
Gnarpe H, Belsheim J, Blomqvist C et al. (1984). The in-vitro influence of
ceftazidime on host defence mechanisms. J Antimicrob Chemother 13:369-75.
Gould EK, Harvey JA, Dytrych JK. (1989). Antibody to endotoxin is associated with decreased frequency of postoperative infection. Am J Obstet Gynecol 160, 317-9.
Greenman RL, Schein RMH, Martin MA et al. (1991). A controlled clinical
trial of E5 murine monoclonal IgM antibody to endotoxin in the treatment of
gramnegative sepsis. JAMA 266, 1097-102.
Hacket SP, Stevens DL. (1993). Superantigens associated with staphylococcal
and streptococcal toxic shock syndrome are potent inducers of tumor necrosis
factor-beta synthesis. J Infect Dis 168(1), 232-5.
Hamilton-Miller JM. (2001). Immunopharmacology of antibiotics: direct and
indirect immunomodulation of defence mechanisms. J Chemother 13, 107-11.
Hanberger H, Nilsson LE, Kihlström E et al. (1989). Postantibiotic effect of Elactam antibiotics on Escherichia coli evaluated by bioluminescence assay of
bacterial ATP. Antimicr Agents Chemother 34, 102-6.
Hansen M, Christrup LL, Jarlov JO et al. (2001). Gentamicin dosing in critically ill patients. Acta Anaesthesiol Scand 45, 734-40.
Holm SE. (1989). Invasive group A streptococcal infections. N Engl J Med
335, 590-1
Hooper DC. (2000). Quinolones. In: Mandell GL, Bennet JE, Dolin R, eds.
Principles and practice of infectious diseases, 5th edn. New York: Churchill
Livingstone, 404-23.
Hurley JC. Antibiotic-Induced Release of Endotoxin: A Reappraisal. (1992).
Clin Infect Dis 15, 840-54.
Hurley JC, Louis WJ, Tosolini FA et al. (1991) Antibiotic-induced release of
endotoxin in chronically bacteriuric patients. Antimicrob Agents Chemother
35, 2388-94.
Ingalls RR, Heine H, Lien E et al. (1999). Lipopolysacharide recognition,
CD14, and lipopolysacharide receptors. Infect Dis Clin North Am 13, 341-53.
Ishiguro EE, Vanderwel E, Kusser W. Control of lipopolysacharide biosynthesis and release by Escherichia coli and Salmonella typhimurium. (1986). J
Bacteriol 168, 328-33.
Issekutz AC. (1983). Removal of Gram negative endotoxin from solutions by
affinity chromatography. J Immunol Methods 61, 275-81.
Jackson JJ, Kropp H. Beta-lactam antibiotic-induced release of free endotoxin:
in vitro comparison of penicillin-binding protein (PBP) 2-specific imipenem
and PBP 3-specific ceftazidime. (1992). J Infect Dis 165, 1033-41.
47
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
Jackson JJ, Kropp H. (1996). Differences in mode of action of E-lactam antibiotics influence morphology, LPS release and in vivo antibiotic efficiency. J
Endotoxin Res 3, 201-18.
Kaul R, McGeer A, Norrby-Teglund A. (1999). Intravenous immunoglobulin
therapy for streptococcal toxic shock syndrome—a comparative observational
study. The Canadian Streptococcal Study Group. Clin Infect Dis 28(4), 800-7.
Korzeniowski OM. (1989). Effects of antibiotics on the mammalian immune
system. Infect Dis Clin North Am 3, 469-78.
Kotb M. (1995). Bacterial pyrogenic exotoxins as superantigens. Clin Microbiol Rev 8:3, 411-26.
Kusser WC, Ishiguro EE. (1988). Effects of aminoglycosides and spectinomycin on the synthesis and release of lipopolysaccharide by Escherichia coli. Antimicrob Agents Chemother 32, 1247-50.
Laczika K, Staudinger T, Hollenstein U et al. (1997). Renal tolerability of four
different once-daily dose regimens of netilmicin in critical care patients. Wien
Klin Wochenschr 109, 840-4.
Lamy M, Mathy-Hartert M, Deby-Dupont G. (1996). Neutrophil-induced
Oxidative Stress. In: Vincent JL ed. Yearbook of Intensive Care and Emergency Medicine. Berlin, Heidelberg: Springer-Verlag. 83-95
Lee PK, Schlievert PM. (1989). Quantification and toxicity of group A streptococcal pyrogenic exotoxins in an animal model of toxic shock syndrome-like
illness. J Clin Microbiol 27, 1890-2.
Leeson MC, Morrison DC. (1994). Induction of proinflammatory responses in
human monocytes by particulate and soluble forms of lipopolysacharide.
Shock 2, 235-45.
Lepper PM, Held TK, Schneider EM et al. (2002). Clinical implications of
antibiotic-induced endotoxin release in septic shock. Intensive Care Med 28,
824-33.
Levin M, Quint PA, Goldstein B. (2000). Recombinant bactericidal/permeability increasing protein (rBPI21) as adjunctive treatment for
children with severe meningococcal sepsis: a randomized trial. Lancet
356(9234), 961-7.
Luchi M, Morrison DC, Opal S et al. (2000). A comparative trial of imipenem
versus ceftazidime in the release of endotoxin and cytokine generation in patients with gram-negative urosepsis. J Endotox Res 6, 25-31.
Löwdin, E., I. Odenholt, S. Bengtsson, and O. Cars. (1996). Pharmacodynamic
effects of sub-MICs of Benzylpenicillin against Streptococcus pyogenes in a
newly developed in vitro kinetic model. Antimicrob. Agents. Chemother
40,2478-82.
Mascini, EM, Jansze M, Shouls LM et al. (2001). Penicillin and clindamycin
differentially inhibit the production of pyrogenic exotoxins A and B by group
A streptococci. Int J Antimicr Agents 18, 395-8.
Maskin B, Fontan PA, Spinedi EG. (2002). Evaluation of endotoxin release
and cytokine production induced by antibiotics in patients with gram-negative
nosocomial pneumonia. Crit Care Med 30, 349-54.
Mathison JC, Tobias PS, Wolfson E et al. (1992). Plasma lipopolysacharide
binding protein. A key component in macrophage recognition of gramnegative LPS. J Immunol 149, 200-6.
48
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
Maury E, Barakett V, Blanchard H et al. (1998). Circulating endotoxin during
initial antibiotic treatment of severe gram-negative bacteraemic infections. J
Infect Dis 178, 270-3.
Meloni F, Ballabio P, Bianchi L et al. (1995). Cefodizime modulates in vitro
tumor necrosis factor-alpha, interleukin-6 and interleukin-8 release from human peripheral monocytes. Chemotherapy 41, 289-95.
Mendelman PM, Campos J, Chaffin DO et al. (1988). Relative penicillin G
resistance in Neisseria meningitidis and reduced affinity of penicillin-binding
protein 3. Antimicrob Agents Chemother 32, 706-9.
Michie HR, Manogue KR, Spriggs DR et al. (1988). Detection of circulating
tumor necrosis factor after endotoxin administration. N Engl J Med 318, 14816.
Mock CN, Jurkovich GJ, Dries DJ et al.(1995). Clinical significance of antibiotic endotoxin-releasing properties in trauma patients. Arch Surg 130, 123441.
Morrison DC, Bucklin SE, Leeson MC et al. (1996). Contribution of soluble
endotoxin released from gramnegative bacteria by antibiotics to the pathogenesis of experimental sepsis in mice. J Endotox Res 3, 237-43.
Neu, HC. (1985). Relation of structural properties of beta-lactam antibiotics to
antibacterial activity. Am J Med 79, 2-13.
Norrby-Teglund A, Thulin P, Gan BS. (2001). Evidence for superantigen
involvement in severe group A streptococcal tissue infections. J Inf Dis 184,
853-60.
Norrby-Teglund A, Kaul R, Low DE et al. (1996). Evidence for the presence
of streptococcal-superantigen-neutralizing antibodies in normal polyspecific
immunoglobulin G. Infect Immun 64(12), 5395-8.
Opal SM, Cross AS, Jhung JW et al. (1996). Potential hazards of combination
immunotherapy in the treatment of experimental septic shock. J Infect Dis
173(6), 1415-21.
Orus P, Vinas M. (2001). Mechanisms other than penicillin-binding protein-2
alterations may contribute to moderate penicillin resistance in Neisseria meningitidis. Int J Antimicrob Agents 18, 113-9.
Ottonello L, Dallegri F, Dapino P et al. (1991). Cytoprotection against neutrophil-delivered oxidant attack by antibiotics. Biochem Pharmacology 42,
2317-21.
Prins JM, van Agtmael MA, Kuijper EJ et al. (1995). Antibiotic-induced endotoxin release in patients with Gram-negative urosepsis: a double blind study
comparing imipenem and ceftazidime. J Infect Dis 172, 886-91.
Prins JM, van Deventer SJH, Kuijper EJ et al. (1994). Clinical relevance of
antibiotic-induced endotoxin release. Antimicrob Agents Chemother 38, 12118.
Prins JM, Speelman P, Kuijper EJ et al. (1997). No increase in endotoxin
exposure during antibiotic killing of meningococci. J Antimicrob Chemother
39(1), 13-8.
Rangel-Frausto MS, Pittet D, Costigan M et al. (1995). The natural history of
the systemic inflammatory response syndrome (SIRS): A prospective study
JAMA 273, 117-23.
Reinhart K, Wiegand-Lohnert C, Grimminger F et al. (1996). Assessment of
the safety and efficacy of the monoclonal anti-tumor necrosis factor antibodyfragment, MAK 195F, in patients with sepsis and septic shock : a multicenter,
49
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
randomized, placebo-controlled, dose-ranging study. Crit Care Med 24(5),
733-42.
Ropp PA, Nicholas RA. (1997). Cloning and characterization of the ponA
gene encoding penicillin-binding protein 1 from Neisseria gonorrhoeae and
Neisseria meningitidis. J Bacteriol 179, 2783-7.
Salvo I, de Cian W, Musicco M, et al. The Italian sepsis study: preliminary
results on the incidence and evolution of SIRS, sepsis, severe sepsis and septic
shock. (1995) Intensive Care Med 21, 244-9.
Schlievert PM, Assimacoupoulos AP, Cleary PP. (1996). Severe invasive
group A streptococcal disease: Clinical description and mechanisms of pathogenesis. J Lab Clin Med, 127, 13-22.
Shenep JL, Barton RP, Mogan KA. (1985). Role of antibiotic class in the rate
of liberation of endotoxin during therapy for experimental gram-negative bacterial sepsis. J Infect Dis 151, 1012-18.
Sjölin, J. (1996). Pathophysiology in septic shock. In: Risberg B, ed. Trauma
care- an update. Pathophysiology of trauma. Mölndal, Sweden. PR-Offset.
88-103.
Stevens DL, Madaras-Kelly KJ, Richards DM. (1998). In vitro antimicrobial
effects of various combinations of penicillin and clindamycin against four
strains of Streptococcus pyogenes. Antimicrob Agents Chemother 42, 1266-8.
Stevens DL, Tanner MH, Winship J et al. (1989). Severe group A streptococcal infections associated with a toxic shock-like syndrome and scarlet fever
toxin A. N Engl J Med 321, 1-8.
Stevens DL, Mayer KA, Mitten JE. (1987). Effects of antibiotics on toxin
production and viability of Clostridium perfringens. Antimicrob Agents Chemother 31, 213-8.
Stevens DL, Yan S, Bryant AE. (1993). Penicillin-binding protein expression
at different growth stages determines penicillin efficacy in vitro and in vivo:
an explanation for the inoculum effect. J Infect Dis 167, 1401-5.
Stevens DL, Gibbons AE, Bergstrom R et al. (1988). The Eagle effect revisited: Efficacy of clindamycin, erythromycin, and penicillin in the treatment of
streptococcal myositis. J Infect Dis 158, 23-8.
Stevens DL, Bryant AE, Hackett SP. (1992). Grampositive shock. Curr Op Inf
Dis 5, 355-63.
Tobias PS, Mathison J, Mintz D et al. (1992). Participation of lipopolysacharide-binding protein in lipopolysacharide-dependent macrophage activation.
Am J Resp Cell Mol Biol 7, 239-45.
Tomasz A.(1986). Penicillin-binding proteins and the antibacterial effectiveness of E-lactam antibiotics. Rev Infect Dis 8, 260-78.
Van den Berg C, de Neeling AJ, Schot CS et al. (1992). Delayed antibioticinduced lysis of Escherichia coli in vitro is correlated with enhancement of
LPS release. Scand J Infect Dis 24, 619-27.
van Deventer SJH, Buller HR, ten Cate JW et al. (1990). Experimental endotoxemia in humans: analysis of cytokine release and coagulation, fibrinolytic and complement pathways. Blood 76,2520-6.
Warren BL, Eid A, Singer P et al. (2001). Caring for the critically ill patient.
High dose antithrombin III in severe sepsis: a randomized controlled trial.
JAMA 286(15), 1869-78.
50
98.
99.
100.
101.
102.
103.
Weiss J, Elsbach P, Olsson I et al. (1978). Purification and characterization of
a potent bactericidal and membrane active protein from the granules of human
polymorphonuclear leukocytes. J Biol Chem 253, 2664-72.
Wright SD, Tobias PS, Ulevitch RJ et al. (1989). Lipopolysacharide (LPS)
binding protein opsonizes LPS bearing particles for recognition by a novel receptor on macrophages.J Exp Med 170, 1231-41.
Xu S, Collins CM. (1996). Temperature regulation of the streptococcal pyrogenic exotoxin A-encoding gene (speA). Infect Immun 64, 5399-402.
Young LS. Sepsis syndrome. (1995). In: Mandell GL, Bennet JE, Dolin R,
eds. Principles and practice of infectious diseases, 4th edn. New York: Churchill Livingstone, 690-705.
Ziegler EJ, Mc Cutchan JA, Fierer J et al. (1982). Treatment of gramnegative
bacteremia and shock with human antiserum to a mutant E. coli. N Engl J Med
307, 1225-30.
Ziegler EJ, Fischer CJ, Sprung CL et al. (1991). Treatment of gramnegative
bacteremia and septic shock with HA-1A human monoclonal antibody against
endotoxin. N Engl J Med 324, 429-36.
51
Acta Universitatis Upsaliensis
Comprehensive Summaries of Uppsala Dissertations
from the Faculty of Medicine
Editor: The Dean of the Faculty of Medicine
A doctoral dissertation from the Faculty of Medicine, Uppsala University,
is usually a summary of a number of papers. A few copies of the complete
dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine.
(Prior to October, 1985, the series was published under the title “Abstracts of
Uppsala Dissertations from the Faculty of Medicine”.)
Distribution:
Uppsala University Library
Box 510, SE-751 20 Uppsala, Sweden
www.uu.se, [email protected]
ISSN 0282-7476
ISBN 91-554-6109-3