Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, November 2004, p. 4144-4147, Vol. 48, No. 11
0066-4804/04/$08.00+0 DOI: 10.1128/AAC.48.11.4144-4147.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Centro Nacional de Microbiología, Instituto de Salud Carlos III,1 Medical Department, GlaxoSmithKline, Madrid, Spain2
Received 4 December 2003/ Returned for modification 20 April 2004/ Accepted 5 July 2004
| ABSTRACT |
|---|
|
|
|---|
20% when the hyperimmune serum was used, thus showing cross-protection and synergism (defined as the situation in which there is no response to the single agents [no differences versus placebo] while the combination exhibits significant activity) with subinhibitory concentrations of the antibiotic. The presence of antipneumococcal antibodies allowed antibiotic efficacy with negligible values of pharmacodynamic parameters (Cmax/MIC ratio of <1 and thus a null value for the time that serum levels exceed the MIC). This in vivo synergism offers a potential therapeutic strategy against resistant strains. | INTRODUCTION |
|---|
|
|
|---|
T>MIC, where T is time): low rates of survival for
T>MICs of <20% and high rates for
T>MICs of >40% (7). With respect to the immune system, immunogenicity depends on the pneumococcal serotype (9), and passive immunization against S. pneumoniae has shown, in animal models, that an anti-capsular antibody protective threshold exits (18). Nevertheless a number of other pneumococcal cell surface antigens, such as choline binding protein A (CbpA) and pneumococcal surface protein A (PspA), induce opsonophagocytic antibodies that may provide cross-immunity regardless of serotype (4), offering some degree of protection in murine models (1). This fact is important if, as in previous studies (2, 21, 22), animals have been passively immunized with hyperimmune serum obtained after serial immunizations with a whole-cell pneumococcal heat-inactivated inoculum, leading to the probability of participation of antibodies other than those directed to the serotype-specific polysaccharide. In these studies (2, 21, 22), the protection obtained by passive immunization with hyperimmune serum was elevated, and, with the addition of antibiotic treatment, the combined effect was similar to the addition of effects obtained by the administration of hyperimmune serum and the antibiotic alone, thus suggesting an additive effect. The present study investigates the activity of amoxicillin concentrations subinhibitory over the treatment period against an amoxicillin-resistant strain causing infection in animals immunized with homologous hyperimmune serum (obtained with the infecting strain) or heterologous hyperimmune serum (obtained with a strain belonging to a serotype other than that of the infecting strain). In addition, based on the definition of in vivo synergism, i.e., "the protective dose of the combination is one-fourth of the antibiotic or no response to the single agentsantibiotic or antibodiesis obtained while the combination exhibits significant activity" (5), this study tries to elucidate if the combined effect of antibodies and antibiotic is synergistic rather than additive by using a dilution of hyperimmune serum that had been demonstrated to cause no reduction in mortality.
| MATERIALS AND METHODS |
|---|
|
|
|---|
Infecting strain. A serotype 6B S. pneumoniae strain (MIC and minimum bactericidal concentration [MBC] of penicillin, 4 µg/ml) was selected for the study based on its resistance to ß-lactams and virulence in mice. MICs of penicillin, amoxicillin, erythromycin, and levofloxacin were 4, 8, >128, and 32 µg/ml, respectively.
After serial passages in mice, the microorganism was grown three times in Todd-Hewitt broth supplemented with 0.5% yeast extract (THYB; Difco, Detroit, Mich.) and enriched with 5% fetal bovine serum until an absorbance of 0.3 at 580 nm (UV-visible spectrophotometer, UV-1203; Shimadzu Scientific Instruments, Inc., Columbia, Md.). This procedure assures a highly encapsulated strain (10). The final bacterial suspension was then aliquoted and stored at 70°C in 15% glycerol until its use.
In vitro studies. MICs and MBCs of amoxicillin against the infecting strain were determined by a broth dilution method following NCCLS procedures (14). Modal values from five separate determinations were taken as the working values.
Animals. Eight- to 12-week-old female BALB/c mice weighing 19 to 22 g were used.
Determination of minimal lethal dose and challenge dose. Groups of 10 mice per dilution were intraperitoneally (i.p.) inoculated with different inocula ranging from 105 to 108 CFU/ml (spectrometrically measured) to determine the minimal dose that produced a 100% mortality rate over a 7-day follow-up period. Bacteria in the logarithmic phase of growth in enriched THYB were centrifuged, and the pellet was washed three times and resuspended in THYB to reach 108 CFU/ml (spectrometrically measured). The inoculum was confirmed by culture of serial dilutions onto blood Mueller-Hinton agar incubated at 37°C in 5% CO2 air. Dead mice were recorded daily. The minimal lethal dose was determined from the results obtained in three independent experiments. Twice the minimal lethal dose was used as the infective inoculum (challenge dose).
Hyperimmune serum. A serotype 23F S. pneumoniae strain was chosen to obtain the heterologous hyperimmune serum, whereas the infecting strain was used for the homologous serum. In both cases, bacteria in a logarithmic phase of growth were inactivated at 60°C for 1 h. Groups of animals were inoculated weekly with 200 µl of the 6B or 23F inactivated-bacterium suspensions (109 CFU/ml in phosphate-buffered saline [PBS]) by the i.p. route for 5 weeks. Animals were exsanguinated by cardiac puncture to obtain the serum samples, which were subsequently pooled, aliquoted, and frozen until use.
Determination of protection by hyperimmune sera. To determine the degrees of protection of the immune sera, groups of 10 mice per dilution and serum were inoculated i.p. with 200-µl serial double dilutions (in PBS) of immune serum up to the dilutions that demonstrated no difference in protection versus controls. Mice in the control groups received injections of the placebo (PBS) or nonimmune serum. After 1 h, mice received one challenge dose of bacteria by the i.p. route. Animals were observed for 7 days.
Antibiotic regimen. Amoxicillin doses were selected based on the results of a previous study (2). The dose of 3.12 mg/kg of body weight was chosen to assure subinhibitory concentrations in serum over the entire treatment period, and the 25-mg/kg dose was chosen as a positive therapeutic control of antibiotic efficacy.
Determination of antibiotic efficacy in the presence of antibodies. To investigate the in vivo combined effect of both doses of amoxicillin and both hyperimmune sera, groups of 10 animals were passively immunized with hyperimmune serum (at the dilutions that had demonstrated null protection) and dosed with the antibiotic. On each experimental day of inoculation, two untreated control groups were included: 10 animals received i.p. 200 µl of THYB as a placebo, and 10 animals received i.p. 200 µl of nonimmune mouse serum as a control to rule out a biological effect of nonspecific immunity. Antibiotic treatment was initiated 1 h after the pneumococcal challenge and continued every 8 h, with a total of six subcutaneous doses being administered. Animals were observed and deaths were recorded for 7 days.
Determination of antibiotic concentrations in serum. Amoxicillin concentrations in serum in six healthy animals after a single subcutaneous dose of antibiotic were determined as previously described (2) to confirm the concentrations determined in previous experiments (2).
Statistical analysis. Survival curves were obtained by the Kaplan-Meier method. An ordinal log rank test was used to compare different study groups.
| RESULTS |
|---|
|
|
|---|
Mortality rates were 100% with placebo or nonimmune serum, with all animals dead by day 2.
Survival rates in animals immunized with 6B or 23F nondiluted hyperimmune serum were 100 and 80%, respectively, and, when serum was diluted 1/4, survival rates were 70 (significantly [P < 0.05] different from controls) and 0%, respectively. The minimal dilutions of hyperimmune sera producing 0% survival rates (all animals dead by day 3) were 1/6 for the homologous serum (anti-6B) and 1/4 for the heterologous one (anti-23F). No statistical differences between the effects on mortality of these dilutions of hyperimmune serum and placebo or nonimmune serum were found (Table 1).
|
90%. In contrast, the amoxicillin subtherapeutic dose (3.12 mg/kg), achieving subinhibitory concentrations in serum (Cmax of 6.1 µg/ml; Cmax/MIC ratio of 0.7) produced a final survival rate of 10%, with significant (P < 0.05) delay in mortality with respect to that for groups receiving the placebo, the nonimmune serum, or the noneffective dilutions of hyperimmune sera (1/6 for 6B and 1/4 for 23F). This antibiotic dose produced significantly (P < 0.05) lower survival rates than the 25-mg/kg amoxicillin dose.
As shown in Table 1, the combined effect of homologous or heterologous hyperimmune serum and subinhibitory concentrations of amoxicillin was synergistic: survival rates from day 4 onwards were more than four times higher with the combination (mortality
20%) than with the antibiotic or the immunization alone (mortality
90%). Moreover, there were no differences between the group receiving the combination of the subtherapeutic amoxicillin dose plus the minimal dilutions of hyperimmune serum producing 0% survival rates and control groups for therapeutic efficacy, whereas there were significant differences (P < 0.05) between the group receiving the combination and control groups for nontherapeutic efficacy.
There were no differences in the synergistic effect achieved by amoxicillin subtherapeutic doses between animals passively immunized with the hyperimmune serum directed to the infecting strain and animals passively immunized with the hyperimmune serum directed to a different serotype (100 versus 80% survival).
| DISCUSSION |
|---|
|
|
|---|
Cross-seroprotection. The high protection obtained in animals immunized with nondiluted anti-23F hyperimmune serum, close to the protection obtained with the anti-6B hyperimmune serum, suggests the participation of non-serotype-specific antibodies. In this study, the hyperimmune serum was obtained through serial immunizations with a heat-inactivated whole-cell inoculum (whether 6B or 23F pneumococci). Probably due to the use of the whole cell, antibodies other than polysaccharide antibodies, directed to common surface pneumococcal antigens, are involved in protection. Previous studies have demonstrated that anti-PspA antibodies induce opsonophagocytosis, offering some degree of protection in murine models (1, 20). In humans, pneumococcal carriage induces an immunoglobulin G response with strain-to-strain cross-reactivity with respect to 23F and 6B serotypes (12). Thus, the postcolonization immune response involves non-polysaccharide-specific serotype antibodies. The in vivo synergy between antipneumococcal antibodies and amoxicillin subinhibitory concentrations in this animal model may be non-serotype specific because other antigens and antibodies (not directed to polysaccharides) are involved: The same effect is obtained when using passive heteroimmunization from the perspective of polysaccharide antibodies (23F versus 6B) and probably homoimmunization from the perspective of other antibodies (i.e., PspA homology among different serotypes).
Synergism. In previous studies (2, 21, 22), a synergistic effect between the 3.12-mg/kg subtherapeutic dose and the 1/4 dilution of hyperimmune serum specific to the infecting strain, a 6B non-amoxicillin-resistant (amoxicillin MIC of 4 µg/ml) S. pneumoniae strain (different from the serotype 6B strain used in the present study), with respect to mortality, was not clearly demonstrated due to the high protection provided by immunization. With respect to synergism, the more interesting interactions are those that result in enhanced rates or absolute magnitudes of bacterial killing by the combination compared with those by either drug alone (8), and it has been proposed that this effect be measured in vivo (17). Despite a significantly higher in vivo bactericidal activity (bacterial clearance from blood) of the combination than of amoxicillin alone, at supra- and infrainhibitory concentrations (21, 22), no translation into a synergistic effect with respect to survival was demonstrated (only a positive additive effect) (2).
In the present experiment, hyperimmune sera were diluted until nonstatistical differences between them and the placebo or nonimmune serum were reached, showing 0% protection (100% mortality). Considering that (i) the subtherapeutic dose of amoxicillin in nonimmunized animals produced 10% survival from day 4 onwards, (ii) the 1/6 dilution of anti-6B hyperimmune serum produced 0% survival from day 4 onwards, and (iii) the subtherapeutic dose of amoxicillin in previously immunized animals produced 100% survival over the entire follow-up period, it can be concluded that, from the survival rate perspective, a synergistic effect is present in this case since the combined effect of the subtherapeutic dose of amoxicillin and passive immunization is 10 times higher than the effect of the antibiotic or passive immunization alone. The importance of this synergistic effect is augmented if we consider that there are numerous examples of in vitro synergistic combinations of antimicrobial agents but that thus far synergistic antimicrobial combinations have proved more effective than single agents in only a limited number of clinical settings (13). In our case, the synergism is demonstrated in an in vivo therapeutic model, and we speculate that, with penicillins, this could also occur in the clinical setting since, in contrast to what has been found for other antimicrobial agents, no clear report has been published showing documented microbiological failures in pneumococcal respiratory infections with most penicillins at high doses despite a large number of published studies warning of penicillin resistance. Previous classical reports have shown that bacteremic pneumonia due to pneumococci for which the penicillin MIC was
2 µg/ml responds to high penicillin doses but that cases involving higher resistance may require an alternative antibiotic (15). Increase of the penicillin dose to cover resistant strains is a strategy to obtain response to therapy in cases of penicillin resistance. But the facts that both antibodies (whether capsular or noncapsular) and ß-lactams act on the cell surface and that synergism has now been demonstrated in mice suggest that a possible combined effect of preexisting natural antibodies and penicillins in the clinical setting cannot be ruled out in cases of very high penicillin resistance, since the level of antibodies in healthy middle-age adults can be as high as several hundred micrograms per milliliter of blood for some serotypes (and cross-serotype inhibition of antibody binding is also observed) (6). In this model we have used one strain with penicillin and amoxicillin MICs of 4 and 8 µg/ml, respectively.
Pharmacodynamics of the antibody-antibiotic combination.
For several years, the relationship between the pharmacological properties of an antimicrobial agent and the susceptibility of the infecting organism, as indicated by pharmacodynamic parameters, have been used as the most accurate predictor of antimicrobial efficacy both in animal models and in the clinical setting. The additive effect demonstrated, with respect to survival rates, in the previous experiment translated pharmacodynamically into a decrease of
T>MIC from 25.6% with the antibiotic alone to 2.8% in the presence of specific antibodies (2). In that experiment, an intermediately resistant strain (amoxicillin MIC of 4 µg/ml) was used as the infecting strain. In the present experiment, a fully resistant strain was used (amoxicillin MIC of 8 µg/ml) to assure subinhibitory concentrations (Cmax = 6.1 µg/ml) over the entire treatment period. In this case, pharmacodynamic parameters had negligible values (
T>MIC of 0% and Cmax/MIC ratio of <1), predicting failure of therapy, as occurred with the antibiotic therapy in nonimmunized animals (90% mortality). Nevertheless, in the presence of pneumococcal antibodies, therapeutic efficacy was obtained with negligible values of pharmacodynamic parameters predicting efficacy.
From the pharmacodynamic point of view, and considering the definition of in vivo synergism based on the
T>MICs for the individual drugs and the combination (17), the effect of the combination can also be considered synergistic since
T>MIC decreased to 0 from 25.6% with amoxicillin (3.12 mg/kg) alone. The significant decrease in pharmacodynamic parameters indicative of efficacy in the presence of pneumococcal antibodies until negative values are reached makes antipneumococcal vaccination a possible strategy (not only because it is directed to serotypes carrying resistance determinants) to overcome pneumococcal resistance (3). This is due to the facts that the highest amoxicillin MIC generally found is 8 µg/ml (prevalence of strains for which the MICs are
8 µg/ml in Spain is 5.1% [16]) and that fully resistant strains are isolated mainly in the pediatric and elderly populations, where antipneumococcal vaccines are recommended (annually for the elderly). Strategies to overcome resistance based on pharmacodynamics led to the development of such new formulations as the sustained-release amoxicillin-clavulanic acid formulation (11), which produces
T>MICs of at least 50 and 35% against previously non-amoxicillin-susceptible pneumococcal strains for which the MICs are 4 and 8 µg/ml, respectively. On the other hand, if the results of this model could be extrapolated to humans, this area has the potential to impact antibiotic dosing to decrease pneumococcal resistance (19). The combination of both strategies would probably provide the coverage for 100% of the pneumococcal population.
| ACKNOWLEDGMENTS |
|---|
This work was supported by a grant from GlaxoSmithKline S.A., Madrid, Spain.
| FOOTNOTES |
|---|
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Clin. Vaccine Immunol. | Clin. Microbiol. Rev. |
|---|---|
| J. Clin. Microbiol. | ALL ASM JOURNALS |