ABSTRACT
Mycobacterium abscessus is an emerging pathogen capable of causing invasive pulmonary infections in patients with chronic lung diseases. These infections are difficult to treat, necessitating prolonged multidrug therapy, which is further complicated by extensive intrinsic and acquired resistance exhibited by clinical M. abscessus isolates. Therefore, development of novel treatment regimens effective against drug-resistant strains is crucial. Prior studies have demonstrated synergistic efficacy of several β-lactams against M. abscessus in vitro; however, these combinations have never been tested in an animal model of M. abscessus pulmonary disease. We utilized a recently developed murine system of sustained M. abscessus lung infection delivered via an aerosol route to test the bactericidal efficacy of four novel dual β-lactam combinations and one β-lactam/β-lactamase inhibitor combination. All five of the novel combinations exhibited synergy and resulted in at least 6-log10 reductions in bacterial burden in the lungs of mice at 4 weeks compared to untreated controls (P = 0.038).
INTRODUCTION
Patients with chronic lung diseases, such as cystic fibrosis (CF) and bronchiectasis, are at increased risk for development of recurrent pulmonary infections, largely due to poor clearance of respiratory secretions, resulting in persistent colonization with pathogenic bacteria. Nontuberculous mycobacteria are ubiquitous in the environment and often colonize the airways of these patients. However, some species such as Mycobacterium abscessus can cause invasive pulmonary infections, especially in the CF population. M. abscessus infections in these patients lead to rapid decline in lung function and can significantly impair quality of life (1–3). Treatment of these infections in CF patients is a growing challenge, since M. abscessus is intrinsically resistant to several classes of antibiotics (4) and the incidence of multidrug-resistant strains is steadily rising (5–8).
Treatment of M. abscessus lung disease involves prolonged therapy with multiple antibiotics, several of which require intravenous administration and can cause substantial cytotoxicity (9, 10). Therefore, new drugs and novel regimens are urgently needed to effectively treat infections with M. abscessus strains that are resistant to drugs/regimens in current clinical use. One strategy that would allow for rapid implementation of effective treatments in the clinical setting involves repurposing of commercially available antibiotics in novel combinations. Several studies have demonstrated in vitro synergy between various antibiotics against M. abscessus, which, when used in combination, have the potential to overcome drug resistance (11–26).
One antibiotic class of particular interest along these lines is β-lactams, as they are the most widely used class of antibiotics for treatment of bacterial infections (27) and are generally well tolerated with minimal comparative cytotoxicity. There are currently only two β-lactams included in the M. abscessus treatment recommendations established by the U.S. Cystic Fibrosis Foundation and the European Cystic Fibrosis Society: cefoxitin (a cephalosporin) and imipenem (a carbapenem) (10, 28). The mechanisms of action of β-lactams against mycobacteria suggest that combinations of β-lactam subclasses should be effective against M. abscessus, but studies assessing synergy between dual β-lactams against this pathogen have only recently been undertaken and were limited to in vitro and ex vivo conditions (25, 26, 29).
β-Lactams exhibit their function via inhibition of synthesis of peptidoglycan, the exoskeleton of the bacterial cell wall that is essential for cellular growth and survival (30). The final step of peptidoglycan synthesis involves covalent bonding of monomeric peptide side chains to generate a single macromolecule that encapsulates the plasma membrane (31). Unlike most bacteria, whose peptidoglycans are predominantly cross-linked with 4→3 linkages generated by d,d-transpeptidases (DDT; also known as penicillin-binding proteins), the peptidoglycan of M. abscessus is cross-linked with both 4→3 and 3→3 linkages. These 3→3 linkages, generated by l,d-transpeptidases (LDT), predominate the peptidoglycan of M. abscessus (32).
LDTs and DDTs are differentially inhibited by β-lactam subclasses, with penicillins and cephalosporins harboring the greatest activity against DDTs and carbapenems being most active against LDTs (29, 33–35). This observation has given rise to the hypothesis that combinations of β-lactam subclasses may exhibit synergy and result in optimal killing of M. abscessus (25, 29). While the molecular mechanisms of dual β-lactam synergy have yet to be settled, it is noteworthy that these combinations have demonstrated synergistic efficacies against E. faecalis endocarditis (36) and tuberculosis (37).
M. abscessus also harbors a robust β-lactamase, BlaMab, against which many β-lactamase inhibitors are ineffective, including clavulanate, sulbactam, and tazobactam (38, 39). However, BlaMab is inactivated by the non-β-lactam β-lactamase inhibitor avibactam, resulting in the reduction of the MICs of several β-lactams against M. abscessus upon the addition of avibactam (25, 38, 40, 41). Avibactam may also directly inhibit LDTs (42), which may potentiate the synergy observed between dual β-lactams. In addition, relebactam and vaborbactam, which are also non-β-lactam β-lactamase inhibitors, reduce MICs of various β-lactams against M. abscessus when combined (43).
We recently tested these hypotheses and identified several combinations of β-lactam antibiotics that exhibited synergy against M. abscessus in vitro (25). Of these synergistic combinations, 13 were pairs of two β-lactams, and 5 consisted of a β-lactam paired with avibactam. However, since the in vitro efficacy of antibiotics against M. abscessus does not always correlate in vivo (1, 7, 44), further preclinical testing was necessary to elucidate combinations capable of treating M. abscessus pulmonary disease in an animal model.
Several attempts to establish an animal model of M. abscessus infection have been undertaken in recent years, including studies in zebrafish (45–50), Drosophila (15, 51), and Galleria mellonella larvae (52). However, these organisms are unable to mimic the complexity of mammalian physiology with regard to the treatment of infections.
A validated mouse model of chronic M. abscessus pulmonary disease resulting from an aerosol infection would be relevant in assessing efficacies of experimental treatment regimens. Prior studies of M. abscessus infection in immunocompetent mice have demonstrated gradual clearance of bacteria, while immunocompromised mice are capable of developing a sustained infection (53–56). However, in these studies, mice were infected with M. abscessus via tail vein injection, resulting in a systemic infection that is distinct from the aerosol route of infection that produces primary lung disease in humans. Attempts to establish pulmonary infection utilizing an intratracheal approach (13, 57, 58) have demonstrated sustained infections in immunocompromised mice; however, this surgical method is invasive and subject to high morbidity related to the procedure itself. Several laboratories have evaluated an aerosol route of M. abscessus infection, since it is noninvasive and mimics the route of acquisition in humans. While chronic infections were achieved in gamma interferon (59) and granulocyte-macrophage colony-stimulating factor (60) knockout mice, immunocompetent BALB/c (61) and ΔF508 “cystic fibrosis” (62) mice did not develop sustained pulmonary infections.
In order to more accurately assess pathophysiology of chronic M. abscessus lung disease in mice, a recent study leveraged the results of these prior studies, including pharmacologic immune suppression via corticosteroid injection to promote establishment of an invasive pulmonary infection in otherwise immunocompetent mice (63). C3HeB/FeJ mice were used, as they are capable of developing caseating granulomas in the setting of M. tuberculosis infection (64–66) and therefore likely also mimic the pathology caused by M. abscessus lung disease in humans. The use of injectable corticosteroids allowed for adjustment of the level of immunosuppression and chronic steroid use is considered a risk factor for invasive infections with rapidly growing mycobacteria (67–69), as well as reactivation of latent tuberculosis (70, 71), making this a clinically relevant modality. This study demonstrated that the use of dexamethasone immunosuppression in C3HeB/FeJ mice infected with M. abscessus via aerosol route allowed for establishment of an invasive M. abscessus pulmonary infection, with development of extensive alveolar inflammation with granuloma formation on histopathology following withdrawal of steroids and immune reconstitution (63). Although additional studies are needed to institute this as a standardized model of M. abscessus lung disease, this system appears to be a viable option for evaluation of in vivo treatment regimens. We therefore utilized this model to assess the in vivo efficacy of five β-lactam combinations found to be effective against M. abscessus in vitro (25).
RESULTS
Dual β-lactam combinations exhibit synergy against M. abscessus in mice.We evaluated efficacies of four novel dual β-lactam combinations and one β-lactam/β-lactamase inhibitor combination in C3HeB/FeJ mice harboring pulmonary infections with M. abscessus ATCC 19977 delivered via aerosol route. These combinations were selected based on evidence of synergy observed in vitro (25) and are as follows: imipenem and cefoxitin, imipenem and cefdinir, imipenem and doripenem, imipenem and biapenem, and biapenem and avibactam. We concurrently assessed efficacy of single drug treatments with imipenem, doripenem, biapenem, and cefdinir for comparison. Although therapeutic regimens containing imipenem or cefoxitin are used in the clinical setting to treat M. abscessus lung disease, none of these combinations have been systematically tested in an animal model against M. abscessus.
An important aspect of the dosing regimen in this study is that in dual β-lactam combinations, each β-lactam was used at only half the dose that was used in each single drug group. For instance, mice received 100 mg/kg imipenem and 20 mg/kg cefdinir per dose in this combination group, whereas single-drug-treated comparator groups received 200 mg/kg imipenem or 40 mg/kg cefdinir per dose. Our hypothesis for this dosing regimen was that, if combinations are synergistic, treatment with half dosages of each drug in a combination will produce lower bacterial burdens than the full dose of either drug when used alone.
At the onset of infection (designated week −1), the mean M. abscessus burden in the lungs of mice was 3.95 (standard deviations [SD] = 0.23) log10 CFU. After 1 week (week 0), the mean M. abscessus burden in the lungs was essentially unchanged at 3.93 (SD = 0.41) log10 CFU, and no colonies were isolated from splenic tissue; however, two colonies grew from hepatic tissue in one mouse, denoting some degree of dissemination. As demonstrated in a study describing this model (63), we observed a persistent uptrend in M. abscessus CFU in the lungs of mice that were immunosuppressed, but not receiving antibiotics throughout the experiment (Table 1). These mice were designated as “untreated” and were used as a positive control.
Mean CFU in the lungs of mice infected with M. abscessus ATCC 19977 and treated with antibiotics across 4 weeks, with associated reduction in lung CFU compared to untreated controls
At 1 week after antibiotic initiation (week 1), the average lung CFU in untreated mice was 4.61 log10 (SD = 0.73), whereas the CFU in the lungs of mice treated with antibiotic combinations all demonstrated >1-log10 reduction. By the week 2 time point, the mean lung CFU in untreated mice had increased to 6.77 log10 (SD = 1.15), whereas all antibiotic combination groups exhibited a >4-log10 reduction in mean lung CFU. At 4 weeks, the lung CFU in untreated mice was 7.55 log10 (SD = 1.63), compared to ≥6-log10 CFU reductions across all combination treatment groups. Single-drug treatments, which were administered at double the respective dosages of those in dual combinations, were also compared at this time point, with imipenem, doripenem, and biapenem conferring a >5-log10 pulmonary CFU reduction, whereas cefdinir only resulted in a 0.3-log10 decrease (Fig. 1).
M. abscessus CFU in the lungs of mice treated with antibiotic combinations versus single antibiotics versus untreated controls. All mice were immunosuppressed with dexamethasone. Data points represent individual mice. (A) Imipenem and cefdinir; (B) imipenem and doripenem; (C) imipenem and biapenem; (D) biapenem and avibactam.
The mean M. abscessus CFU in the liver and spleen of untreated mice at 4 weeks were 4.99 log10 (SD = 1.58) and 4.20 log10 (SD = 1.45), respectively, whereas no colonies were observed among the combination treatment groups throughout the experiment. In addition, no organ dissemination was noted among imipenem- or biapenem-treated mice; however, both liver and splenic disseminations were observed in one mouse in the doripenem group (organ cultures for the other four mice in this group were negative) and four mice in the cefdinir group. The doripenem-treated mouse with dissemination had 4.38 log10 CFU in the liver and 3.46 log10 CFU in the spleen, which averages to 0.88 log10 (SD = 1.96) CFU in the liver and 0.69 log10 (SD = 1.55) CFU in the spleen for the entire doripenem treatment group. For cefdinir-treated mice, the mean CFU counts were 3.94 log10 (SD = 2.72) and 2.97 log10 (SD = 2.33) CFU in the liver and spleen, respectively.
At 4 weeks, the difference in mean lung CFU in untreated mice versus all treatment groups was statistically significant via Kruskal-Wallis test (P = 0.001). Additional pairwise comparisons between treatment groups using the Wilcoxon rank sum test, with corrections for multiple comparisons, determined statistically significant differences in lung CFU between untreated mice and each of the combination treatment groups (P = 0.038), as well as those treated with biapenem (P = 0.038) or doripenem alone (P = 0.048), but imipenem alone did not confer a significant difference compared to untreated mice (P = 0.06). Cefdinir alone exhibited only a slight decrease in lung CFU compared to untreated controls and therefore was significantly inefficacious compared to all other treatment groups (P = 0.038).
In vitro activity versus in vivo efficacy.Although all of the antibiotic combinations tested in this study were effective, prior in vitro data (25) predicted that some combinations would perform better than others. For example, combinations of imipenem and doripenem, as well as imipenem and cefdinir, exhibited greater synergy and subsequent reduction in MICs compared to imipenem and biapenem in vitro; however, the opposite was observed in vivo. Although within the margin of error, imipenem and biapenem demonstrated a higher efficacy compared to both imipenem and doripenem and imipenem and cefdinir in mice. While antibiotic synergy was observed both in vitro and in vivo, the degree of synergy differed between testing modalities. Therefore, evaluation in an animal model enhances the rigor of results and allows for a more comprehensive evaluation of the therapeutic potential of novel combinations. That being said, dose ranging and pharmacokinetic/pharmacodynamic (PK/PD) studies are needed to optimize dosing and further characterize therapeutic levels relevant to humans.
DISCUSSION
The rising prevalence of multidrug-resistant M. abscessus strains observed among CF patients and the resulting morbidity from suboptimal treatments with cytotoxic antibiotics necessitate the urgent development of novel treatment regimens that are both effective and well tolerated. By leveraging our current understanding of the mechanisms of action of β-lactams against M. abscessus, this antibiotic class may be a largely untapped resource from which to develop these novel regimens, and their commercial availability would allow for rapid implementation in the clinical setting.
The potential for synergistic activity of dual β-lactams against mycobacteria was initially proposed based on enhanced susceptibility to amoxicillin of M. tuberculosis mutants lacking a dominant LDT (72). Subsequent studies that demonstrated similarities in molecular mechanisms of susceptibility of M. tuberculosis and M. abscessus to β-lactams (29, 32) provided the premise for potential dual β-lactam synergy against M. abscessus. Several combinations of β-lactams have demonstrated synergy against this bacterium in vitro (25, 26, 29). However, given the general lack of correlation between in vitro susceptibility and clinical efficacy (1, 7, 44), we tested five of these synergistic combinations in vivo in a recently developed murine model of M. abscessus pulmonary infection (63). This study not only provides proof of concept that this murine model is a viable option for assessment of treatment regimens against M. abscessus lung disease but also demonstrates bactericidal efficacy of these novel combinations, which have not previously been tested against this pathogen in mice.
Although most of the antibiotics tested in this study are dosed more frequently in the clinical setting (i.e., every 6 to 8 h), antibiotic administration in this experiment was limited to twice-daily dosing due to laboratory constraints. However, mice were treated 7 days per week for the duration of the experiment. With the exception of cefoxitin (54), none of the drugs tested in this study have previously been tested against M. abscessus infection in mice. Therefore, dosages were extrapolated from murine studies of other pathogens, since dose optimization studies were beyond the scope of this proof-of-concept study (73–78). As imipenem is currently used against M. abscessus in the clinical setting, the daily individual drug dose used in this study was similar to the calculated human equivalent dosage (HED) based on allometric scaling, which accounts for differences in body surface area and metabolic rate between mice and humans (79). Therefore, combination regimens included approximately half the HED of imipenem. However, because biapenem and doripenem have not previously been assessed against M. abscessus in mice, dosing concentrations tested were higher than the calculated HED but were extrapolated from prior murine studies in M. tuberculosis and K. pneumoniae, respectively (76, 78). Further studies are needed to assess PK/PD profiles and optimize antibiotic dosing; however, the efficacy demonstrated in this initial study provides a premise for future treatment studies using this murine model.
Dual β-lactam therapies are often avoided in clinical practice due to concern for potential toxicity. However, a recent meta-analysis of 13 randomized controlled trials assessed the efficacy and safety profiles of dual β-lactams compared to β-lactam plus aminoglycoside combinations against Gram-negative infections (80). This study found that dual β-lactams achieved slightly better clinical and microbiological responses, while conferring significant decreases in nephrotoxicity and ototoxicity compared to aminoglycoside-containing regimens, with the rates of other adverse events being largely comparable. Given relatively the high rate of adverse events related to current M. abscessus therapies, which often necessitate long-term use of an aminoglycoside (10), alternate treatment options associated with even a marginal reduction in cytotoxicity would be beneficial.
We found all of the novel antibiotic combinations tested to be bactericidal against M. abscessus ATCC 19977. After 4 weeks of treatment, all five combinations reduced the average lung CFU by at least 6 log10 compared to untreated controls, which was statistically significant. This not only illustrates the ability of the murine model to produce a sustained pulmonary M. abscessus infection via pharmacologic immunosuppression with dexamethasone but also provides direct examples of the bactericidal efficacy of these novel antibiotic combinations. A linear trend was observed for all treatment regimens throughout the 4-week duration of the experiment. We performed regression analysis on the data and generated trend equations that would permit prediction of time to sterilization of the infection (Table 2). This analysis predicted a linear trend, with R2 values ranging from 0.950 to 0.999, which represent correlation coefficients of the predicted trends to the actual data. Based on this analysis, we extrapolated the potential treatment duration necessary to completely clear M. abscessus infections in the lungs of mice. The combination of imipenem and biapenem was predicted to require 5.1 weeks to achieve sterilization compared to 5.3 weeks for imipenem and cefoxitin, the two β-lactams currently used to treat M. abscessus pulmonary infections in patients. It is possible that bactericidal activities of the combinations beyond the 4-week time point assessed in this study are nonlinear and require longer durations to achieve complete sterilization of M. abscessus from murine lungs. Since the aim of this study was to assess whether synergy between dual β-lactam combinations observed in vitro was preserved in vivo, determining sterilizing potential of the combinations was beyond the scope of this study, and further studies evaluating longer treatment durations are needed.
Regression analysis of observed M. abscessus burden in the lungs of mice
We also hypothesized that treatment with single drugs, at doses twice those in combination groups, would be less effective than synergistic combinations. Indeed, single-drug treatments were less effective than combinations; however, with the exception of cefdinir, all of the single-drug treatment groups exhibited a high level of bactericidal activity, with >5-log10 reductions in lung CFU at 4 weeks compared to controls. The most likely reasons for this were the relatively low initial MICs of imipenem (8 μg/ml), doripenem (16 μg/ml), and biapenem (16 μg/ml) against M. abscessus ATCC 19977, in addition to the higher than human equivalent dosages of doripenem and biapenem, as previously mentioned. Cefdinir, however, had an initial MIC of 64 μg/ml, which may explain its lack of efficacy when used alone, which was potentially further compromised by inadequate drug dosing. In addition, all of the combination groups resulted in greater CFU reductions compared to single drugs, which suggests that these combinations are synergistic in vivo, given equivalent dosing between the treatment groups.
Compared to BALB/c mice, C3HeB/FeJ mice tend to be more heterogeneous with regard to their response to both infection and treatment (66), a trait that was also observed in this study. Standard deviations in CFU among both untreated mice and those within treatment groups were initially quite small, but progressively increased throughout the course of the experiment (Table 1). This trend was particularly prevalent among untreated and single antibiotic treatment groups, resulting in higher standard deviations. In each single drug group, there was one mouse that had a much higher bacterial burden in the lungs compared to the others. For example, in the imipenem-treated group, four of five mice had lung CFU ranging from 0.0 to 1.6 log10. However, the fifth mouse had a CFU count of 5.4 log10. Similarly, for biapenem-treated mice, four of five had lung CFU counts ranging from 0.0 to 1.3 log10, whereas the CFU content in the fifth mouse was 4.7 log10. In mice treated with cefdinir, lung CFU ranges were 5.6 to 7.6 log10 for four mice, with 10.1 log10 CFU for the fifth mouse. This difference was slightly less pronounced in doripenem-treated mice, with lung CFU counts in four mice ranging from 1.0 to 3.1 log10 compared to 4.7 log10 CFU in the fifth mouse. Dissemination to the spleen or liver was only observed in one mouse in the doripenem group and four mice in the cefdinir group and did appear to correlate with higher pulmonary bacterial burdens; however, no gross pathology was observed in these mice. While heterogeneity between patients is observed in the clinical setting and the C3HeB/FeJ model may be a better representation of the varied pathophysiology expected from treatment of human infections compared to some other murine strains, additional preclinical studies with larger numbers of mice are warranted to help control for this heterogeneity.
This study illustrates proof of principle that the murine model of sustained M. abscessus pulmonary infection (63) may be a viable option for future treatment experiments but will require further optimization to improve our understanding of disease pathology. In addition, while the use of dexamethasone as an immunosuppressive agent would be relevant to patients with innate immune deficiencies or those on immunosuppressive regimens, this model is unable to mimic the complex pathophysiology of the cystic fibrosis lung with regard to impaired mucociliary clearance and biofilm formation and therefore may overemphasize the effect of antibiotic treatments. Further treatment studies in murine strains with lung pathophysiology similar to that seen in CF patients (i.e., βENaC-transgenic mice [81] that exhibit depletion of airway surface liquid due to increased sodium absorption) may provide a more accurate approximation of potential treatment efficacy in this population.
While this study assessed efficacy of novel regimens over the course of a month, we did not evaluate these regimens against a more chronic infection, nor did we assess for relapse potential after treatment cessation. Additional studies assessing rate of emergence of drug-resistant mutations after antibiotic exposure will also be important to identify combinations that are less prone to the selection of drug resistance and potentially denote greater clinical utility.
Prior in vitro work revealed several additional novel β-lactam combinations with potential efficacy against M. abscessus (25, 26, 29), which warrant further testing in vivo, in addition to studies of novel combinations against highly resistant clinical M. abscessus isolates to better assess synergistic efficacy. In addition, since the majority of antibiotics tested in this study have not previously been tested against M. abscessus in mice, the dosages were likely not optimal and therefore would necessitate dose-ranging studies to determine optimal dosages, as well as PK/PD assessments to allow for more comprehensive comparison of treatment efficacies. However, because the antibiotics tested are already commercially available and recommended dosages in humans have been established, this may not be as relevant in planning for future clinical trials.
Given our general lack of therapeutic options against M. abscessus pulmonary disease, the ability to rapidly implement novel regimens in the clinical setting will be crucial to improving patient outcomes. Use of dual β-lactams remains a novel approach to treatment of resistant infections and preliminary studies against M. abscessus both in vitro and in vivo are highly promising, warranting additional research.
MATERIALS AND METHODS
Ethics.Animal procedures used in the following studies were performed in adherence to the Johns Hopkins University Animal Care and Use Committee and to national guidelines.
Bacterial strains and growth conditions.M. abscessus reference strain ATCC 19977 was used in all studies. M. abscessus was cultured in Middlebrook 7H9 broth (Becton Dickinson) supplemented with 10% ADS (albumin, dextrose, NaCl) and 0.05% Tween 80 with constant shaking at 220 rpm in an orbital shaker at 37°C. Organ homogenates were cultured on Middlebrook 7H11 selective agar (Becton Dickinson) supplemented with 10% ADS, 50 μg/ml cycloheximide (Sigma-Aldrich), 25 μg/ml polymyxin B (Sigma-Aldrich), 50 μg/ml carbenicillin (Sigma-Aldrich), and 20 μg/ml trimethoprim (Sigma-Aldrich). Quantitative and CFU counts were enumerated after 5 days of incubation at 37°C on Middlebrook 7H11 selective agar.
Mice, infection, and corticosteroid administration.Female C3HeB/FeJ mice, 4 to 5 weeks old, were procured from Jackson Laboratories (Bar Harbor, ME). All mice were infected concurrently with 10 ml M. abscessus culture at an A600 of 0.1 via aerosolization using a Glas-Col inhalation exposure system (Glas-Col, Terre Haute, Indiana) according to the manufacturer’s instructions. The infection cycle protocol included 15 min of preheating, 30 min of nebulization, 30 min of cloud decay, and 15 min of surface decontamination.
To achieve adequate immunosuppression and ensure that any reduction in bacterial load observed in treated mice was due to antibiotic treatment rather than immune clearance, mice were treated with dexamethasone. A prior murine study demonstrated that dexamethasone is capable of adequately suppressing the immune response to achieve a sustained M. abscessus infection with an increase in pulmonary bacterial burden over time (63).
Dexamethasone (D1756; Sigma-Aldrich) was dissolved in sterile 1× phosphate-buffered saline (PBS; pH 7.4; Quality Biological) and administered via daily subcutaneous injection at 5 mg/kg/day, as described previously (63, 71, 82, 83). Daily dexamethasone treatment began 2 weeks prior to infection and was continued throughout the duration of the experiment. Mice were sacrificed at 24 h postinfection to determine the initial bacterial burden in the lungs as described above.
Antibiotic regimens.All antibiotics (imipenem, doripenem, biapenem, cefdinir, cefoxitin, avibactam, and cilastatin) were obtained through Sigma-Aldrich. Powdered drug stocks were dissolved in sterile 1× PBS to achieve the following concentrations: imipenem, 100 mg/kg; doripenem, 200 mg/kg; biapenem, 200 mg/kg; cefdinir, 20 mg/kg; cefoxitin, 200 mg/kg; and avibactam, 64 mg/kg. In addition, cilastatin was added to any regimen containing a carbapenem in a 1:1 dose ratio in an attempt to compensate for the higher activity of murine renal dihydropeptidase-1 compared to humans (84). Antibiotic regimens were administered via subcutaneous injection twice daily, 7 days a week, for a total of 4 weeks. Combination treatment groups were as follows: imipenem and cefoxitin, imipenem and cefdinir, imipenem and doripenem, imipenem and biapenem, biapenem and avibactam. Single-drug treatment groups included imipenem alone, doripenem alone, biapenem alone, and cefdinir alone. Antibiotic dosages for single-drug groups were twice those listed above to compensate for lack of a second agent. Each treatment group consisted of five mice per time point.
CFU determination.Mice treated with two drugs were sacrificed at 1, 2, and 4 weeks after treatment initiation. Mice treated with single drugs were sacrificed at 4 weeks only. Five mice per treatment group were sacrificed per time point, and the CFU burdens in the lungs, spleen, and liver were enumerated via culture as described above. Of note, homogenates of complete organs were cultured from the lungs and spleen, but only approximately 30% of hepatic tissue was homogenized and cultured; therefore, whole-liver CFU counts were extrapolated accordingly.
Data analysis.Data were analyzed via the Kruskal-Wallis rank sum test, with pairwise comparisons using a Wilcoxon rank sum test and with Bonferroni corrections for multiple comparisons.
ACKNOWLEDGMENTS
This study was supported by the Cystic Fibrosis Foundation award LAMICH17GO and National Institutes of Health award R21 AI137720 to G.L. E.S.-R. was supported by NIH T32 AI007291.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the Cystic Fibrosis Foundation or the National Institutes of Health.
FOOTNOTES
- Received 22 March 2019.
- Returned for modification 1 May 2019.
- Accepted 11 May 2019.
- Accepted manuscript posted online 20 May 2019.
- Copyright © 2019 American Society for Microbiology.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵