JAC Advance Access published online on June 13, 2007
Journal of Antimicrobial Chemotherapy, doi:10.1093/jac/dkm190
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Propensity to release endotoxin after two repeated doses of cefuroxime in an in vitro kinetic model: higher release after the second dose
1 Section of Infectious Diseases, Department of Medical Sciences, Uppsala University, Uppsala, Sweden 2 Section of Clinical Microbiology, Department of Medical Sciences, Uppsala University, Uppsala, Sweden
* Corresponding author. Tel: +46-18-6115663; Fax: +46-18-6115650; E-mail: gunilla.hjerdt.goscinski{at}akademiska.se
Received 28 December 2006; returned 28 January 2007; revised 27 March 2007; accepted 8 May 2007
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionTransparency declarationsReferences |
|---|
Objectives: To study endotoxin release from two strains of Escherichia coli after exposure to two repeated doses of cefuroxime in an in vitro kinetic model.
Methods: Cefuroxime in concentrations simulating human pharmacokinetics was added to the bacterial solution with a repeated dose after 12 h. In another experiment, tobramycin was given concomitantly with the second dose of cefuroxime. Samples for viable counts and endotoxin analyses were drawn before the addition of antibiotics and at 2 and 4 h after each dose.
Results: The propensity to release endotoxin, expressed as log10 endotoxin release (EU)/log10 killed bacteria, was higher after the second than after the first dose, 0.80 ± 0.04 and 0.65 ± 0.01, respectively, in the ATCC strain and 0.80 ± 0.04 and 0.65 ± 0.02, respectively, in the clinical strain (P < 0.001). Endotoxin was released earlier after the second dose (P < 0.001). Addition of tobramycin at the second dose reduced the endotoxin release in comparison with that of cefuroxime alone (P < 0.001).
Conclusions: The propensity to liberate endotoxin is higher after the second dose of cefuroxime than after the first, resulting in a higher release of endotoxin than expected from bacterial count. The release after the second dose can be reduced by the addition of tobramycin.
Key Words: aminoglycosides , Escherichia coli , morphology
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsReferences |
|---|
Numerous in vitro and animal experiments as well as some clinical studies have shown that endotoxin concentrations increase after antibiotic treatment of Gram-negative infections.1–4 Antibiotics differ in their capacity to cause endotoxin release depending on their mode of action. ß-Lactam antibiotics, acting on the cell wall, lead to a higher endotoxin release than aminoglycosides and other groups of antibiotics affecting bacterial protein synthesis.4,5 Among the ß-lactam antibiotics, there are also differences in capacity to liberate endotoxin depending on their affinities for the various penicillin binding proteins (PBPs) that are located in the cell wall.1 Furthermore, affinities for PBPs have been shown to be dose dependent.6,7 Ceftazidime has been demonstrated to bind to PBP 3 at low doses, leading to the formation of long filamentous structures with an increased endotoxin production before lysis. With increasing doses, ceftazidime also has a high affinity for PBP 1, leading to rapid lysis without elongation and with less endotoxin release.7 Similar mechanisms have been suggested for other cephalosporins, such as cefuroxime, for which higher doses have been shown to free less endotoxin per killed bacterium in vitro.8 It has also been demonstrated that the relation of the affinities affects the length of the rods and subsequent endotoxin liberation.9 In addition, combination of antibiotics may affect the endotoxin release since tobramycin has been shown to significantly reduce the cefuroxime-induced PBP 3 release.8
An important factor for the magnitude of endotoxin release is the number of killed bacteria, as demonstrated in in vitro experiments exposing the bacteria to constant concentrations of antitbiotics.8,10 This would imply that clinically the largest amount is liberated after the first dose when the bacterial count is at its highest. However, in a patient reported by Dofferhoff et al.,1 in whom the endotoxin concentrations were repeatedly measured after administration of two doses of a PBP 3-binding antibiotic, it was shown that the free plasma levels increased more following the second dose. From this it may be hypothesized that the release of endotoxin per killed bacterium might be higher at the second dose. The main purpose of the present investigation was therefore to further explore this and to study the endotoxin release from two Escherichia coli strains after exposure to two repeated doses of cefuroxime using an in vitro kinetic model simulating the human concentration–time profile. In addition, a secondary aim was to study whether the combination with tobramycin at the second dose resulted in a reduction similar to that previously shown when combined with the first dose.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsReferences |
|---|
Cultures and media
Two E. coli strains were used in the experiments, a clinical strain isolated from a patient with septicaemia (B049-3036) and a reference strain (ATCC 25922).
Brain heart infusion (BHI) made with pyrogen-free water was used as medium. Prior to each experiment, the test strains were inoculated in a pyrogen-free glass tube containing BHI and incubated for 4.5 h at 35°C, resulting in a logarithmic-phase culture of around 108 cfu/mL.
The antibiotics were obtained as reference powders with known potencies. Cefuroxime was purchased from Glaxo Wellcome AB (Gothenburg, Sweden) and tobramycin from Eli Lilly Sweden (Stockholm, Sweden).
Minimum inhibitory concentrations
MICs were determined in duplicate by 2-fold macrodilution in broth with an inoculum of 
105 cfu of the test strain per mL.
The in vitro kinetic model has been described in detail elsewhere.11 It consists of a spinner flask with a total volume of 110 mL with a filter membrane (0.45 µm) 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 tube 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 37°C during the experiments.
The bacteria were added to the culture vessel at time zero, resulting in a starting inoculum of 5 x 106 cfu/mL. Cefuroxime was diluted in PBS to a concentration of 10 000 mg/L, whereupon 0.66 mL of the solution was added to the spinner flask BHI medium. The initial cefuroxime concentration in the vessel was 60 mg/L, corresponding to the free fraction seen after an intravenous dose of 1.5 g which is an often recommended dose in adults with severe sepsis.12 The half-life of 1.5 h for cefuroxime was chosen in order to simulate human pharmacokinetics.13 The same dose was added to the vessel through the thin-plastic tube after 12 h when the inoculum had reached a level approximating that of the starting inoculum. Six experiments were performed with each strain.
In additional experiments, the bacteria were exposed to the same concentrations of cefuroxime but at the time of the second dose it was combined with tobramycin. Tobramycin was diluted in PBS to a concentration of 40 000 mg/L and 0.055 mL was added to the culture compartment resulting in a final concentration of 20 mg/L. The tobramycin concentration was chosen to correspond to the peak value seen after a single dose of 6 mg/kg often given in human sepsis.14 Two experiments with each strain were performed.
Samples for viable counts were drawn through the silicone membrane before addition of antibiotics and at 2, 4, 14 and 16 h. After dilution in PBS, the samples were seeded onto Columbia agar plates and incubated at 35°C. The samples were serially diluted 10-fold in PBS. At least three samples (10 or 100 µL) from the original bacterial suspension and/or dilutions were subsequently spread on agar plates, incubated at 35°C and counted after 48 h. To avoid antibacterial carryover, all samples were placed on the same spot on the agar plates and the samples allowed to diffuse into the agar for 3–5 min before spreading. If there was, after growth, a clear zone where the sample had been placed, the agar plate was divided into sections and the clear section was excluded from counting. The limit of detection of the viable counts was 1 x 101 cfu/mL. Growth controls were performed assuring normal growth during the experimental conditions up to 4 h when the high number of bacteria interfered with the filter.
Concomitantly with the samples for viable counts, and immediately after addition of the second antibiotic dose, samples were drawn for analyses of endotoxin. Since it has been shown that most of the endotoxin is released within 4 h after antibiotic administration, samples for analyses were not drawn after this time point.8 In two of the control experiments, samples for endotoxin analysis were obtained at 0, 2 and 4 h.
Endotoxin-free glass tubes were used for all endotoxin assays, the culture vessel was pre-heated at 180°C and the stirrer was washed in pyrogen-free water. Sterile, pyrogen-free plastic syringes and pipettes were used, and non-pyrogenic filters of 0.45 µm were employed to analyse the free endotoxin. The samples were immediately filtered and kept frozen at –70°C pending analysis. Analyses of endotoxin were performed in duplicate with the limulus amebocyte lysate assay (Endochrome-KTM; Charles River Endosafe, Charleston, SC 29407, USA).
The morphology of E. coli was examined using scanning electron microscopy. For this purpose, a separate experiment was performed with the ATCC strain given the same antibiotic concentrations as those used in the previous experiments and with samples drawn before the antibiotic doses and at 1 and 2 h after each dose. Samples of 10 mL of broth culture were removed and centrifuged at 1400 g for 10 min whereupon the bacteria were resuspended in glutaraldehyde 2.5% in 0.2 M sodium cacodylate buffer. The samples were then dehydrated for 10 min periods in increasing concentrations of acetone. Finally, the bacteria were critical point dried and gold-coated in a sputter coater at 10 mA and 1200 V for 2 min. A LEO Gemini 1530 scanning electron microscope at an accelerating voltage of 2 kV was used for examinations.
Reduction in cfu at various time points after the first dose was calculated by subtraction of the remaining number of cfu at 2 and 4 h from that at time zero. Similarly, reduction in cfu after the second dose was calculated by subtraction of the remaining number of cfu at 14 and 16 h from that at 12 h. In the further calculations, the reduction in cfu was considered to be equivalent to the number of killed bacteria.
The endotoxin release was similarly calculated by subtracting the endotoxin values at time zero and at 12 h from those obtained 2 and 4 h later.
The amount of endotoxin that was eliminated by the pump was calculated using the formula of first order kinetics and assuming a linear increase in concentration during the first two hours and between 2 and 4 h after each dose. It has previously been shown that the logarithmic release of endotoxin is proportional to the logarithmic number of killed bacteria.10 In order to reduce the effect of the variation in viable counts and to relate the endotoxin release to the bacterial killing, the propensity to release endotoxin, expressed as log10 endotoxin release per log10 killed bacteria, was calculated. Values obtained after the first dose were compared in the primary analysis with those obtained after the second dose. This ratio, as well as the logarithmic values of the endotoxin release and the number of killed bacteria, approximated normal distribution. Therefore, in the primary analysis a repeated measures ANOVA was performed. This statistical method was also used in the comparison between the release of endotoxin after the first and second doses and to test the effect of the addition of tobramycin at the second dose. A P value of <0.05 was considered significant. All values are expressed as mean ± SE, except the endotoxin concentrations which were the result of bacterial release, pre-exposure values caused by the preceding growth and elimination by the pump. Endotoxin concentration values are expressed as median and range. The software STATISTICA (StatSoft, Inc., Tulsa, OK, USA) was used in the statistical calculations.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsReferences |
|---|
The MIC values of cefuroxime were 2.0 mg/L for E. coli B049-3036 and 4.0 mg/L for E. coli ATCC 25922, whereas the MICs of tobramycin were 1.0 and 2.0 mg/L, respectively. In the control experiments, the mean log10 bacterial count after 4 h had increased from 6.9 at time zero to 7.9 for the B049 strain and from 6.2 to 7.5 for the ATCC strain. Corresponding endotoxin values were 3600 and 54 000 endotoxin units (EU)/mL, and 330 and 15 000 EU/mL, respectively.
The bacterial killing rates of E. coli B049 and E. coli ATCC 25922 after exposure to cefuroxime are shown in Figure 1, and the log10 number of killed bacteria at 2 and 4 h after the first and the second dose of the ATCC and B049 strains is demonstrated in Table 1. Irrespective of the bacterial killing rate, more bacteria were killed after the first than after the second dose, due to the somewhat higher number of bacteria at time zero than after 12 h. As expected, most bacteria were killed during the first two hours after administration of the cefuroxime doses. There was a larger variation in viable counts at 12 h than at time zero, and consequently there was also a larger variation in the number of killed bacteria after the second dose.
|
|
The endotoxin concentration at various time points is demonstrated in Table 2. There was a large variation, mainly due to the variation in the number of killed bacteria. The reduction in endotoxin concentration that occurred between 2 and 4 h after the second dose was due to a larger elimination than release during this period.
|
Geometric mean of the endotoxin release after correction for elimination via the outflow of the pump was at 2 and 4 h after the first dose 1700 and 12 000 EU/mL, respectively, for the ATCC strain and 9600 and 22 000 EU/mL, respectively, for the B049 strain (Figure 1). Corresponding values after the second dose were 45 000 and 67 000 EU/mL for the ATCC strain (P < 0.01) and 58 000 and 63 000 EU/mL (NS) for the B049 strain, respectively.
In Table 3, the propensity to liberate endotoxin is demonstrated. When the endotoxin release was related to the number of killed bacteria, there was a marked reduction in the variation and 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 time course was also significantly different, with an earlier release after the second dose (P < 0.001 for both strains).
|
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 for both the ATCC and the B049 strain (P < 0.001 for both strains) (Table 4).
|
Morphology
Morphological changes are demonstrated in Figure 2. After addition of cefuroxime into the bacterial suspension, an extensive elongation of the bacteria into filamentous forms occurred after 2 h, whereas only minor changes were seen at 1 h. At 12 h, before the second dose, elongated forms were noted among bacteria of normal size that were not seen before the first dose. After the second dose, filamentous forms together with spheroplasts were observed already at 1 h. When tobramycin was added, the filamentation was less pronounced.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsReferences |
|---|
In this in vitro kinetic model, in which the concentration–time profile of human serum levels of cefuroxime was simulated, it was demonstrated that the propensity to release endotoxin was higher after the second than after the first dose of cefuroxime for the two E. coli strains studied. 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 in effective treatment this will covariate with the bacterial concentration at the time of the dose. With this magnitude of the difference, and provided that 106 bacteria are killed, the log10 reduction of viable count has to surpass
1.1–2.0 log10 cfu/mL if the endotoxin release caused by the second dose should not exceed that of the first one. Alternatively expressed, the endotoxin release after the second dose would theoretically be 8–36-fold higher than that after the first dose provided that 106 bacteria are killed on both occasions. The starting inocula of 106–107 cfu/mL in our experiments were within the range of the variations found at sites of clinical infections, namely 104–109 cfu/mL in pus and peritonitis15 and 103–109 cfu/mL in meningitis,16 respectively. Thus, increased propensity to release endotoxin might explain the increased plasma endotoxin concentration observed after the second dose in the report by Dofferhoff et al.1
A possible mechanism behind the higher endotoxin release after the second dose of cefuroxime might be a continuing release from remaining filaments caused by the first dose. This hypothesis is supported by Jackson and Kropp6 who found an increased endotoxin release at sub-MIC concentrations of several ß-lactam antibiotics that offers an explanation of a sustained release. Our electron microscopy findings just before the second dose 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 after the second dose does not favour this hypothesis. Thus, the mechanism is not clear, but nevertheless 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. Since most of the elongated forms at 2 h after the first dose were eliminated at 4 h, the remaining filamentous structures seen 10 h later may be a sign of disturbed bacterial growth, but this needs further investigation.
In several experimental in vitro studies when the bacteria have been exposed to one dose resulting in constant antibiotic concentrations, it has been shown that addition of aminoglycosides lowers the endotoxin release induced by ß-lactam antibiotics if given concomitantly.8,10,17 However, if administration of the aminoglycoside is delayed 2 h this effect is no longer seen.8 In the present study, however, it was shown that addition of tobramycin reduced the cefuroxime-induced release also at the second dose even if this release was increased and other mechanisms of action may be present. The effect of a repeated combined aminoglycoside ß-lactam antibiotic treatment could not be studied because the clinical concentrations used resulted in all bacteria being killed already after the first dose at the inoculum sizes possible in our in vitro kinetic model.
The clinical relevance of antibiotic-induced endotoxin release is still under debate.18 Endotoxin that is released has been shown to be biologically active,17 but because of the variable levels of endotoxin binding serum proteins, the host response can be markedly altered.19 Thus, even if it has not been shown to affect mortality in prospective controlled clinical studies, antibiotic-induced endotoxin release may be of importance in the most seriously ill patients with septic shock, in whom further deterioration due to an increased endotoxin load may affect mortality. In order to rapidly reduce the initial bacterial load and reduce endotoxin release, addition of a single large dose of an aminoglycoside has been proposed.10 The present data suggest that if there is a clinical deterioration after the first ß-lactam dose, tobramycin may exert its effects despite preceding high ß-lactam concentrations and subsequent sub-MIC concentrations. However, the theoretical advantages of an initial aminoglycoside combination must be weighed against possible side effects and this should be further evaluated in experimental and clinical studies.
|
Transparency declarations |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsReferences |
|---|
None to declare.
|
Acknowledgements |
|---|
This work was supported by grants from the Olinder-Nielsen foundation.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsReferences |
|---|
1 . Dofferhoff ASM, Nijland JH, De Vries-Hospers HG, et al. Effects of different types and combinations of antimicrobial agents on endotoxin release from Gram-negative bacteria: an in-vitro and in-vivo study. Scand J Infect Dis (1991) 23:745–54.[Web of Science][Medline]
2 . Evans ME, Pollack M. Effect of antibiotic class and concentration on the release of lipopolysaccharide from Escherichia coli. J Infect Dis (1993) 167:1336–43.[Web of Science][Medline]
3 . Prins JM, Van Agtmael MA, Kuijper EJ, et al. Antibiotic induced endotoxin release in patients with Gram-negative urosepsis: a double-blind study comparing imipenem and ceftazidime. J Infect Dis (1995) 172:886–91.[Web of Science][Medline]
4 . Shenep JL, Barton RP, Mogan KA. Role of antibiotic class in the rate of liberation of endotoxin during therapy for experimental Gram-negative bacterial sepsis. J Infect Dis (1985) 151:1012–8.[Web of Science][Medline]
5
.
Crosby HA, Bion JF, Penn CW, et al. Antibiotic-induced release of endotoxin from bacteria in vitro. J Med Microbiol (1994) 40:23–30.
6 . Jackson JJ, Kropp H. Differences in mode of action of ß-lactam antibiotics influence morphology, LPS release and in vivo antibiotic efficiency. J Endotoxin Res (1996) 3:201–18.
7 . Neu HC. Relation of structural properties of ß-lactam antibiotics to antibacterial activity. Am J Med (1985) 79(Suppl 2A):2–13.[Medline]
8 . Sjölin J, Goscinski G, Lundholm M, et al. Endotoxin release from Escherichia coli after exposure to tobramycin: dose-dependency and reduction in cefuroxime-induced endotoxin release. Clin Microbiol Inf Dis (2000) 6:74–81.[CrossRef]
9 . Wientjes FB, Nanninga N. On the role of the high molecular weight penicillin-binding proteins in the cell cycle of Escherichia coli. Res Microbiol (1991) 142:333–44.[Medline]
10 . Goscinski G, Lundholm M, Odenholt I, et al. Variation in the propensity to release endotoxin after cefuroxime exposure in different Gram-negative bacteria: uniform and dose-dependent reduction by the addition of tobramycin. Scand J Infect Dis (2003) 35:40–6.[CrossRef][Web of Science][Medline]
11 . Löwdin E, Odenholt I, Bengtsson S, et al. Pharmacodynamic effects of sub-MICs of benzylpenicillin against Streptococcus pyogenes in a newly developed in vitro kinetic model. Antimicro Agents Chemother (1996) 40:2478–82.[Abstract]
12 . Amsden GW. Tables of antimicrobial agent pharmacology. In: Principles and Practice of Infectious Diseases, Sixth Edition—Mandell GL, Bennet JE, Dolin R, eds. (2005) Philadelphia, PA, USA: Churchill Livingstone. 656.
13
.
Goodwin CS, Dash CH, Hill JP, et al. Cefuroxime: pharmacokinetics after a short infusion, and in vitro activity against hospital pathogens. J Antimicrob Chemother (1977) 3:253.
14 . Inclan G, Suarez E, Calvo R, et al. Biocompartmental kinetics of tobramycin analysed with a wide range of covariates. Int J Antimicrob Agents (2005) 26:304–11.[CrossRef][Web of Science][Medline]
15
.
König C, Simmen HP, Blaser J. Bacterial concentrations in pus and infected peritoneal fluid—implications for bacterial activity of antibiotics. J Antimicrob Chemother (1998) 42:227–32.
16 . Arditi M, Ables L, Yogev R. Cerebrospinal fluid endotoxin levels in children with H. influenzae meningitis before and after administration of intravenous ceftriaxone. J Infect Dis (1989) 160:1005–11.[Web of Science][Medline]
17
.
Dofferhoff ASM, Esselink MT, De Vries-Hospers HG, et al. The release of endotoxin from antibiotic-treated Escherichia coli and the production of tumour necrosis factor by human monocytes. J Antimicrob Chemother (1993) 31:373–8.
18 . Lepper PM, Held TK, Schneider EM, et al. Clinical implications of antibiotic-induced endotoxin release in septic shock. Intensive Care Med (2002) 28:824–33.[CrossRef][Web of Science][Medline]
19 . Opal SM. The clinical relevance of endotoxin in human sepsis: a critical analysis. J Endotoxin Res (2002) 8:473–6.[CrossRef][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||