Comparative In Vivo Antibacterial Activity of Human-Simulated Exposures of Cefiderocol and Ceftazidime against Stenotrophomonas maltophilia in the Murine Thigh Model

Cefiderocol is a novel siderophore cephalosporin that utilizes bacterial ferric iron transports to cross the outer membrane. Cefiderocol shows high stability against all classes of β-lactamases, rendering it extremely potent against carbapenem- and multidrug-resistant Gram-negative organisms.

ABSTRACT

Cefiderocol is a novel siderophore cephalosporin that utilizes bacterial ferric iron transports to cross the outer membrane. Cefiderocol shows high stability against all classes of β-lactamases, rendering it extremely potent against carbapenem- and multidrug-resistant Gram-negative organisms. Using a neutropenic murine thigh model, we compared the efficacies of human-simulated exposures of cefiderocol (20-g, 3-h infusion every 8 h [Q8H]) and ceftazidime (2-g, 2-h infusion Q8H) against Stenotrophomonas maltophilia, an emerging opportunistic Gram-negative organism associated with serious and often fatal nosocomial infections. Twenty-four S. maltophilia isolates were studied, including isolates resistant to ceftazidime, trimethoprim-sulfate, and/or levofloxacin. The thighs were inoculated with bacterial suspensions of 10⁸ CFU/ml, and the human-simulated regimens were administered over 24 h. Efficacy was measured as the change in log₁₀ CFU/thigh at 24 h compared to 0-h controls. Cefiderocol human-simulated exposure demonstrated potent bacterial killing; the mean bacterial reduction at 24 h was −2.67 ± 0.68 log₁₀ CFU/thigh with ≥2-log reduction achieved in 21 isolates (87.5%) and a ≥1-log reduction achieved in the remaining 3 isolates (12.5%). In comparison, ceftazidime human-simulated exposure produced a mean bacterial reduction of −1.38 ± 1.49 log₁₀ CFU/thigh among 10 ceftazidime-susceptible isolates and a mean bacterial growth of 0.64 ± 0.79 log₁₀ CFU/thigh among 14 ceftazidime-nonsusceptible isolates. Although ceftazidime showed modest efficacy against most susceptible isolates, humanized cefiderocol exposures resulted in remarkable in vivo activity against all S. maltophilia isolates examined, inclusive of ceftazidime-nonsusceptible isolates. The potent in vitro and in vivo activity of cefiderocol supports the development of this novel compound for managing S. maltophilia infections.

INTRODUCTION

Stenotrophomonas maltophilia is a ubiquitous Gram-negative organism that causes infections with high mortality rates—up to 69% in bacteremia cases (1). Capable of adhering to and forming biofilms on medical devices, this pathogen has become increasingly notorious for causing health care-associated infections, especially in immunocompromised patients (1, 2). The most common types of infection are pneumonia and bacteremia (3–6). Endocarditis, mastoiditis, peritonitis, meningitis, soft tissue infections, wound infections, and urinary tract infections due to S. maltophilia have also been reported (7–18).

Owing to the relatively low resistance rates as currently defined (19), trimethoprim-sulfamethoxazole administered alone or as part of a combination regimen is currently considered the cornerstone of S. maltophilia treatment. While frequently utilized, this agent is associated with a considerable incidence of hypersensitivity—including serious cutaneous reactions such as Stevens Johnsons syndrome—and hyperkalemia, particularly among HIV-infected individuals and the elderly (20). Myelosuppression is also a concern, especially in bone marrow transplant recipients and neutropenic patients (21, 22). The risk of adverse side effects is further increased due to the high dosages required to address S. maltophilia infections (at least 15 mg/kg/day of trimethoprim) (2, 23–25). Consequently, the fluoroquinolones and ceftazidime are frequently used as alternative therapies (2). However, due to the substantial repertoire of resistance mechanisms (i.e., efflux and β-lactamases) encountered in S. maltophilia, resistance rates of 48 and 15% have been reported for ceftazidime and levofloxacin, respectively (19).

Even more concerning are the clinical failure rates that appear discordant to susceptibility as defined for trimethoprim-sulfamethoxazole (1, 2, 26). Available alternatives do not fare any better. In their retrospective study of 98 patients, Wang et al. found that the overall clinical success rates of trimethoprim-sulfamethoxazole and fluoroquinolone monotherapy were only 61 and 52%, respectively (27). Another retrospective study compared trimethoprim-sulfamethoxazole and minocycline treatment among 45 patients and reported clinical failure rates of 41 and 30%, respectively, despite 100% susceptibility rates against the selected monotherapies (28). Given the shortcomings of the currently available therapies, there is a clear need for new agents with activity against this organism.

Cefiderocol is a novel siderophore cephalosporin with a molecular structure comprising a ceftazidime molecule conjugated with a catechol moiety (29). It exerts potent chelating activity with ferric iron and can thus take advantage of bacterial ferric iron uptake transporters that recognize the ferric-bound catechol moiety and permit cefiderocol entry into the periplasmic space, the site of action (29, 30). It is also highly stable in the presence of all classes of β-lactamases, including metallo- and serine-based carbapenemases (31). Thus, cefiderocol is an attractive potential option against carbapenem- and multidrug-resistant Gram-negative organisms inclusive of S. maltophilia. However, limited data pertaining to the in vivo efficacy of cefiderocol against S. maltophilia are currently available.

The objective of this study was to assess the efficacy of cefiderocol against a variety of S. maltophilia isolates, including those resistant to ceftazidime, trimethoprim-sulfamethoxazole, and/or levofloxacin, in a neutropenic murine thigh infection model and to compare its in vivo bactericidal activity to that of ceftazidime. In order to improve the translational application of the outcomes of this comparative assessment to clinic, the murine cefiderocol doses utilized in the study were selected to attain plasma exposures in mice that mimicked those achieved in humans following the administration of the dose currently used in phase III clinical trials: 2 g every 8 h (Q8H) as a 3-h infusion (32). For ceftazidime, a murine regimen that mimicked the plasma exposure achieved in humans following a dose of 2 g Q8H as a 2-h infusion was utilized (33).

(This study was presented in part at ASM Microbe 2019, San Francisco, California, USA, 20 to 25 June 2019 [poster presentation AAR09-2213].)

RESULTS

In vitro susceptibility.

Twenty-four S. maltophilia isolates were examined in this investigation. The following susceptibility categorizations are based on breakpoints published by the Clinical and Laboratory Standards Institute (CLSI) (34). All isolates were determined to be susceptible (≤4 mg/liter) to cefiderocol with MICs ranging from 0.015 to 0.5 mg/liter. Ceftazidime MICs ranged from 1 to >64 mg/liter. Ten isolates were ceftazidime susceptible (≤8 mg/liter), and fourteen isolates were ceftazidime resistant (≥32 mg/liter). Two isolates were resistant to trimethoprim-sulfamethoxazole (≥4/76 mg/liter). Fourteen isolates were levofloxacin susceptible (≤2 mg/liter), six bore intermediate resistance (4 mg/liter), and four were resistant (>4 mg/liter). All isolates were susceptible to minocycline (≤4 mg/liter). Ceftolozane-tazobactam MICs ranged from 0.5 to >64 mg/liter. The ceftazidime-avibactam MIC range was 1 to >64 mg/liter. Aztreonam MICs were largely ≥64 mg/liter, with the exception of two isolates that had MICs of 8 and 32 mg/liter. The MIC results are summarized in Table 1 .

TABLE 1.

Modal MICs for studied S. maltophilia isolates

Isolate	Antibiotic MIC (mg/liter)
	Cefiderocol	Ceftazidime	Trimethoprim-sulfamethoxazole	Ceftolozane-tazobactam	Ceftazidime-avibactam	Aztreonam	Levofloxacin	Minocycline
STM C43-32	0.015	2	0.5/9.5	2	1	64	4	0.5
STM C49-90	0.015	2	0.25/4.75	1	2	32	1	0.25
STM C49-56	0.03	4	4/76	16	2	>64	4	2
STM C49-92	0.03	8	0.25/4.75	4	2	>64	2	0.5
STM-4	0.06	1	0.125/2.38	0.5	1	64	2	2
STM C41-71	0.06	2	0.5/9.5	1	2	8	2	0.5
STM-3	0.06	2	0.125/2.38	1	1	64	2	2
STM-29	0.06	2	0.25/4.75	1	2	64	4	1
STM C41-40	0.06	4	1/19	4	4	>64	2	0.25
STM-28	0.06	4	0.25/4.75	2	2	64	4	1
STM C43-37	0.06	32	1/19	8	16	>64	0.5	0.25
STM C41-97	0.06	64	0.125/2.38	16	16	>64	4	0.25
STM C42-17	0.06	64	0.25/4.75	64	32	>64	8	1
STM C43-10	0.12	32	0.25/4.75	16	32	>64	4	1
STM C41-44	0.12	>64	0.5/9.5	>64	64	>64	32	4
STM C42-4	0.12	>64	0.25/4.75	32	4	>64	1	0.25
STM C42-15	0.12	>64	4/76	>64	32	>64	32	1
STM C43-11	0.12	>64	1/19	64	64	64	8	0.25
STM C45-46	0.12	>64	0.125/2.38	>64	>64	>64	1	0.25
STM C49-22	0.12	>64	0.25/4.75	64	32	>64	0.5	0.5
STM C42-23	0.25	>64	0.25/4.75	64	64	>64	2	0.25
STM C45-76	0.5	64	0.5/9.5	64	64	>64	1	0.25
STM C41-45	0.5	>64	0.25/4.75	>64	>64	>64	1	0.5
STM C42-70	0.5	>64	2/38	>64	>64	>64	2	0.5

Confirmatory pharmacokinetic studies.

Previously determined cefiderocol and ceftazidime regimens in mice which simulate human free plasma pharmacokinetic profiles after cefiderocol doses of 2 g Q8H (3-h infusion) (32) and ceftazidime doses of 2 g Q8H (2-h infusion) (33) were reconfirmed to validate these exposures in S. maltophilia-infected mice. The observed cefiderocol and ceftazidime concentrations are displayed in Fig. 1, and the percentages of the dosing interval during which the free unbound drug in plasma remained above MIC (%fT>MIC) for all profiles are detailed in Tables 2 and 3.

FIG 1.

Free concentration-time profiles of cefiderocol (A) and ceftazidime (B) human-simulated exposures. Observed murine concentrations are displayed as means ± the standard deviations.

TABLE 2.

%fT>MIC values of the humanized cefiderocol regimen at each cefiderocol MIC in humans and in mice^a

Species	%fT>MIC values for the following MICs (mg/liter):
	0.015	0.03	0.06	0.12	0.25	0.5	1
Mouse	100	100	100	100	100	100	100
Human	100	100	100	100	100	100	100

^aData were adapted from Ghazi et al. (32). %fT>MIC, percentage of the dosing interval during which the free unbound drug in plasma remained above MIC.

TABLE 3.

%fT>MIC values of the humanized ceftazidime regimen at each ceftazidime MIC in humans and in mice^a

Species	%fT>MIC values for the following MICs (mg/liter):
	1	2	4	8	16	32	64
Mouse	100	100	100	93	62	24	0
Human	100	100	100	87	58	28	0

^aData were adapted from Crandon et al. (33). %fT>MIC, percentage of the dosing interval during which the free unbound drug in plasma remained above MIC.

Efficacy of human-simulated exposure studies.

The results of the efficacy studies are presented in Fig. 2. Across the 24 isolates, the bacterial burden was 6.78 ± 0.43 log₁₀ CFU/thigh at 0 h, and the in vivo growth was robust, as signified by a 1.75 ± 0.62 increase over 24 h in the untreated control mice. Ceftazidime treatment resulted in a mean bacterial density decrease of –1.38 ± 1.49 log₁₀ CFU/thigh among the ten ceftazidime-susceptible isolates. While seven isolates achieved substantive kill (≥2-log reduction, n = 3; ≥1-log reduction, n = 4), net stasis or growth was observed in three susceptible isolates. Among the 14 ceftazidime-resistant isolates, a net bacterial growth of 0.64 ± 0.79 log₁₀ CFU/thigh was observed after humanized ceftazidime exposures. None of the ceftazidime-resistant isolates achieved ≥1-log reduction. Cefiderocol human simulated exposures produced a mean bacterial reduction of −2.67 ± 0.68 log₁₀ CFU/thigh at 24 h among the S. maltophilia test isolates. Cefiderocol treatment resulted in a ≥2-log reduction in 21 isolates (87.5%) and a ≥1-log reduction in the three remaining isolates (12.5%). The activity of cefiderocol was unencumbered by the multidrug-resistant phenotypic profile of the S. maltophilia studied. Compared to ceftazidime, cefiderocol displayed significantly more antibacterial activity in seven of the ten ceftazidime-susceptible isolates and all fourteen ceftazidime-resistant isolates (P ≤ 0.05).

FIG 2.

Mean change in log₁₀ CFU/thigh ± the standard deviation at 24 h by cefiderocol and ceftazidime human-simulated exposures against ceftazidime-susceptible (A) and ceftazidime-resistant (B) S. maltophilia. Asterisks (*) indicate statistically significant differences (P ≤ 0.05) between cefiderocol and ceftazidime.

DISCUSSION

Cefiderocol, which retains activity against multidrug-resistant Gram-negative bacteria, provides a potential valuable therapeutic option against S. maltophilia infections. The SIDERO-WT-2015 study demonstrated the in vitro activity of cefiderocol and comparators against 8,954 clinical Gram-negative bacilli isolates from patients in North America and Europe, including S. maltophilia (35). This international surveillance data set showed that 99.4% of the examined S. maltophilia isolates (n = 165), including those not susceptible to carbapenems, were susceptible to cefiderocol (MIC₉₀ = 0.5 mg/liter). On the other hand, the susceptibility of the S. maltophilia isolates to all comparator agents were much lower, with MIC₉₀s ranging from 8 to >64 mg/liter. This remarkable in vitro activity provided evidence that cefiderocol could be a potent therapeutic option against S. maltophilia infections and lent support to further in-depth in vivo investigations. While several previously published studies have demonstrated the in vivo efficacy of cefiderocol against a variety of Gram-negative pathogens such as Pseudomonas aeruginosa, Acinetobacter baumannii, and Enterobacteriaceae, inclusive of carbapenem-resistant strains (32, 36), limited data exist in literature on its in vivo activity against S. maltophilia. Ito-Horiyama and coworkers reported that in a neutropenic murine systemic model, cefiderocol was effective against S. maltophilia (ED₅₀, 1.17 mg/kg/dose), while its comparators (cefepime [>100 mg/kg/dose], ceftazidime-avibactam [>100 mg/kg/dose], meropenem-cilastatin [>100 mg/kg/dose], and colistin [>10.0 mg/kg/dose]) were not (37). The same study also demonstrated that cefiderocol exhibited dose-dependent bactericidal activity against S. maltophilia (37). These studies provided evidence of cefiderocol potent in vivo activity against S. maltophilia infections; nonetheless, assessment of efficacy using clinically relevant exposures of cefiderocol was not conducted.

Our current investigation utilizing the neutropenic murine thigh infection model was carried out to assess the efficacy of cefiderocol human-simulated regimen against a diverse collection of S. maltophilia isolates, including those resistant to current standard therapies, as well as to compare the cefiderocol efficacy profile to that of ceftazidime human-simulated regimen. The study isolates were selected to also provide a wide MIC range for cefiderocol among S. maltophilia isolates, including those at the upper end of the MIC distribution. The isolates were also chosen to encompass the MIC distribution of ceftazidime (MIC₉₀ > 64 mg/liter [38]). A high starting inoculum was utilized to establish infection in this study (10⁸ CFU/ml). This inoculum, which was higher than what is typically utilized in in vivo investigations (10⁷ CFU/ml) (36), and the relatively high cefiderocol MICs of the examined isolates were selected for this study to allow us to examine the outcome of cefiderocol therapy in the context of the most challenging clinical scenarios.

Ceftazidime in vitro susceptibility was generally predictive of the in vivo activity of the ceftazidime human-simulated exposure, with susceptible isolates achieving net bacterial killing at 24 h, with the exception of isolates STM C49-90, STM C49-56, and STM C49-92. The lack of activity against the latter isolates could have been attributed to the high starting inoculum utilized in this investigation or the induction of chromosomal β-lactamases, which has been previously reported upon exposure to cephalosporins (39). Likewise, the majority of the ceftazidime-resistant isolates demonstrated a lack of in vivo efficacy upon the administration of ceftazidime human-simulated regimen. The outcome with ceftazidime provided validation for the utility of this infection model in the assessment of the cefiderocol efficacy against S. maltophilia.

Against all isolates examined, cefiderocol demonstrated potent in vivo efficacy, including those resistant to trimethoprim-sulfamethoxazole, ceftazidime, and/or levofloxacin, as well as isolates with cefiderocol MICs at the upper end of the distribution (0.5 mg/liter). This efficacy is also notable because it was observed with a higher than standard inoculum, and an inoculum effect has been observed for β-lactams with S. maltophilia (40). Previous cefiderocol pharmacokinetic/pharmacodynamic studies showed that a target exposure of approximately 75 to 90% %fT>MIC was required for in vivo activity (36). Although the observed concentrations with the human-simulated regimen were somewhat higher than the simulated profile in the current murine model, the exposure achieved is predicted to reasonably approximate the target exposure; the cefiderocol regimen utilized provided a %fT>MIC value of 100% against the isolates examined, and thus the bacterial killing observed in our murine model is consistent with the previously established pharmacokinetic/pharmacodynamic predictions. Against the majority of the isolates, cefiderocol human-simulated regimen resulted in bacterial killing of a magnitude of >2 log, an outcome that is typically considered to be predictive of robust clinical efficacy. Among ceftazidime-susceptible isolates, cefiderocol demonstrated enhanced antibacterial activity compared to ceftazidime; cefiderocol produced more than an additional 1 log of kill. This direct comparison of efficacies indicates that cefiderocol could offer a better therapeutic alterative to ceftazidime even among isolates that are deemed ceftazidime susceptible.

This study utilizing a preclinical murine model has some limitations. Although trimethoprim-sulfamethoxazole is a mainstay of treatment for S. maltophilia infections, it was not used as a comparator because the high thymidine content found in mouse serum would have interfered with trimethoprim in vivo antibacterial activity (41, 42). However, the attainment of >1- to 2-log₁₀ killing with cefiderocol against all studied isolates, including the two isolates that were resistant to trimethoprim-sulfamethoxazole, suggests that cefiderocol may have at least comparable clinical efficacy. Another limitation of this study is that biofilm formation, which has been linked to relapsing S. maltophilia infection (2), was not assessed. The impact of cefiderocol on biofilm formation and its efficacy in the setting of preformed biofilms are opportunities for future research. In addition, experiments utilizing study periods longer than 24 h are also warranted to assess the likelihood of resistance emergence.

In conclusion, cefiderocol exposures simulating those achieved in humans after a 2-g dose Q8H as a 3-h infusion demonstrated potent in vivo activity against all tested S. maltophilia isolates with various degrees of susceptibilities to the current standard of care treatments. In light of the need for alternative therapies, these results support the development of cefiderocol for patients infected with S. maltophilia.

MATERIALS AND METHODS

Antibiotics.

Cefiderocol vials (lot 12M01) were supplied by Shionogi & Co., Ltd. (Osaka, Japan) and stored at –80°C. Commercially available ceftazidime vials (lot 107209C; Sagent Pharmaceuticals) were purchased through Cardinal Health (Dublin, OH) and stored at room temperature. All drug vials were reconstituted according to the manufacturer’s recommendations and then further diluted with 0.9% saline (Hospira, Inc., Lake Forest, IL) to reach target dosing solution concentrations. Mice received doses of cefiderocol or ceftazidime previously shown to simulate steady-state plasma exposures observed in humans receiving cefiderocol at 2 g Q8H (3-h infusion) (32, 43) or ceftazidime at 2 g Q8H (2-h infusion) (33). All doses were administered subcutaneously in a volume of 0.2 ml.

Bacteria.

Twenty-four clinical S. maltophilia respiratory isolates from teaching hospitals in the United States were selected for this study. Bacterial isolates were frozen on skim milk (BD BioSciences, Sparks, MD) at –80°C until needed. The isolates were transferred twice on Trypticase soy agar plates with 5% sheep blood (TSA II; Becton Dickinson, Sparks, MD) and incubated for approximately 24 h at 37°C. MICs were measured in triplicate according to CLSI broth microdilution methodology for the following drugs: cefiderocol, ceftazidime, trimethoprim-sulfamethoxazole, ceftolozane-tazobactam, ceftazidime-avibactam, aztreonam, levofloxacin, and minocycline (34). Cefiderocol MICs were measured in iron-depleted CA-MHB per CLSI recommendations (34). Modal MICs were reported.

Neutropenic murine thigh infection model.

Female CD-1 mice weighing 22.1 ± 1.45 g were purchased from Charles River Laboratories, Inc. (Wilmington, MA). In order to induce neutropenia, the mice received intraperitoneal (i.p.) cyclophosphamide injections 4 days prior to (150 mg/kg) and 1 day prior to (100 mg/kg) inoculation. Uranyl nitrate at 5 mg/kg i.p. was administered 3 days before inoculation to decrease renal clearance and facilitate achieving human-simulated exposures of study agents in mouse plasma. Protocols were reviewed and approved by Hartford Hospital’s Institutional Animal Care and Use Committee. For all experiments, both thighs of each mouse were inoculated with 0.1-ml bacterial suspensions of 10⁸ CFU/ml 2 h before the start of antibiotic treatment.

Confirmatory pharmacokinetic studies.

The human-simulated regimens of cefiderocol and ceftazidime used in this study were previously established in the murine thigh infection model (32, 33). Confirmatory pharmacokinetic studies were completed to verify these treatments resulted in exposures equivalent to humans receiving cefiderocol at 2 g Q8H (3-h infusion) and ceftazidime at 2 g Q8H (2-h infusion) in the current investigation (32, 33, 43). The target %fT>MIC values are summarized in Tables 1 and 2. To account for murine protein binding of cefiderocol (31.6%) and ceftazidime (26.0%), total drug concentrations were multiplied by 68.4 and 74.0%, respectively, to obtain free drug concentrations (32, 33, 43). At each time point, six mice were euthanized via CO₂ asphyxiation, and blood samples were collected via cardiac puncture. Blood was collected in sodium heparin Vacutainer tubes (BD, Franklin Lakes, NJ) and centrifuged to separate plasma, which was frozen at –80°C until assayed. Cefiderocol plasma concentrations were assayed by Shionogi & Co. using validated liquid chromatography-tandem mass spectrometry, as previously reported (36). Ceftazidime samples were analyzed via high-performance liquid chromatography (HPLC) by the Center for Anti-Infective Research and Development (Hartford, CT). The HPLC assay was composed of a Waters 515 Pump (Waters Associates, Milford, MA), a Waters WISP 717 autoinjector (flow rate 1.2 ml/min), a Hitachi L -2400 UV detector (ESA, Inc., Chelmsford, MA; wavelength, 257 nm), and a C₁₈ Resolve column (5 μm, 3.9 by 150 mm). Peak heights were determined by an EZChrom chromatography data system (Scientific Software, Inc., Pleasanton, CA). The mobile phase was composed of 0.05 M phosphate buffer (pH 3.5 to 4.0) and acetonitrile, and the internal standard used was 0.1 M phosphate buffer (pH 6). The range of the standard curve was 0.5 to 50 μg/ml. The interday coefficients of variation were −1.70 and −0.40%, and the intraday coefficients of variation were 2.73 and −3.19% for the low (1 μg/ml) and high (40 μg/ml) check samples, respectively.

Efficacy of human-simulated exposure studies.

Groups of three mice were randomly assigned to no treatment (0-h controls), 0.9% saline placebo (24-h controls), or treatment with cefiderocol or ceftazidime human-simulated regimens for 24 h. Placebo injections were administered vehicle on the same schedule as the most frequent treatment. At 24 h, the mice were sacrificed by CO₂ asphyxiation and cervical dislocation. Thighs were harvested and homogenized in 0.9% saline prior to serial dilution and plating on Trypticase soy agar plates with 5% sheep blood for CFU determination. Efficacy was defined as the change in log₁₀ CFU/thigh at 24 h compared to 0-h controls. Comparisons between the changes in log₁₀ CFU/thigh at 24 h with the cefiderocol and ceftazidime treatment groups were carried out using a Student t test (SigmaPlot 14.0; Systat Software, Inc., San Jose, CA). A P value of ≤0.05 was considered statistically significant.