In Vitro Pharmacodynamics of Fosfomycin against Carbapenem-Resistant Enterobacter cloacae and Klebsiella aerogenes

The increase of carbapenem-resistant Enterobacterales (CRE) and lack of therapeutic options due to the scarcity of new antibiotics has sparked interest toward the use of intravenous fosfomycin against systemic CRE infections. We aimed to investigate the in vitro pharmacodynamics of fosfomycin against carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes. Time-kill studies and population analysis profiles were performed with eight clinical CRE isolates, which were exposed to fosfomycin concentrations ranging from 0.

ABSTRACT

The increase of carbapenem-resistant Enterobacterales (CRE) and lack of therapeutic options due to the scarcity of new antibiotics has sparked interest toward the use of intravenous fosfomycin against systemic CRE infections. We aimed to investigate the in vitro pharmacodynamics of fosfomycin against carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes. Time-kill studies and population analysis profiles were performed with eight clinical CRE isolates, which were exposed to fosfomycin concentrations ranging from 0.25 to 2,048 mg/liter. The 24-h mean killing effect was characterized by an inhibitory sigmoid maximum effect (E_max) model. Whole-genome sequencing was performed to elucidate known fosfomycin resistance mechanisms. Fosfomycin MICs ranged from 0.5 to 64 mg/liter. The isolates harbored a variety of carbapenemase genes including bla_IMP, bla_KPC, and bla_NDM. Five out of eight isolates harbored the fosA gene, while none harbored the recently discovered fosL-like gene. Heteroresistant subpopulations were detected in all isolates, with two out of eight isolates harboring heteroresistant subpopulations at up to 2,048 mg/liter. In time-kill studies, fosfomycin exhibited bactericidal activity at 2 to 4 h at several fosfomycin concentrations (one isolate at ≥16 mg/liter, two at ≥32 mg/liter, two at ≥64 mg/liter, two at ≥128 mg/liter, and one at ≥512 mg/liter). At 24 h, bactericidal activity was only observed in two isolates (MICs, 0.5 and 4 mg/liter) at 2,048 mg/liter. From the E_max model, no significant bacterial killing was observed beyond 500 mg/liter. Our findings suggest that the use of fosfomycin monotherapy may be limited against CRE due to heteroresistance and rapid bacterial regrowth. Further optimization of intravenous fosfomycin dosing regimens is required to increase efficacy against such infections.

INTRODUCTION

Carbapenem-resistant Enterobacterales (CRE) infections are associated with increased mortality rates, longer hospitalizations, and higher health care costs (1–3). Carbapenems, conventionally known as the last-resort antibiotics against multidrug-resistant (MDR) Gram-negative bacteria, are rendered useless in infections involving these contemporary pathogens (2). Infections caused by CRE are particularly problematic, as they are usually resistant to all available antibiotics except those with limited efficacy or safety (4). In addition, Enterobacter cloacae and Klebsiella aerogenes (formerly known as Enterobacter aerogenes) harbor chromosomally encoded AmpC-type β-lactamases, which confer resistance to most β-lactamase inhibitors. They can develop additional antibiotic resistance through enzyme induction and acquire numerous mobile genetic elements (5). With the declining development of new antimicrobials effective against CRE, recent focus has shifted toward reevaluating underutilized historical agents, such as the polymyxins and fosfomycin, against this growing global threat (6).

Fosfomycin is a phosphonic acid derivative discovered more than 40 years ago. It exerts bactericidal activity against a wide spectrum of Gram-positive and Gram-negative bacteria through cell wall synthesis inhibition. Unlike β-lactams and glycopeptides, fosfomycin blocks an earlier stage of peptidoglycan synthesis by inhibiting the enzyme UDP-N-acetylglucosamine enolpyruvyl transferase (MurA). The unique mechanism of action and chemical structure makes cross-resistance with other antibiotics unlikely. Hence, fosfomycin remains effective against several MDR pathogens (7). Currently, only the oral formulation, fosfomycin tromethamine, has been approved by the United States Food and Drug Administration (FDA) and used widely in Europe as a single-dose treatment against acute uncomplicated urinary tract infections caused by Escherichia coli and Enterococcus faecalis (8, 9). Hence, there has been renewed interest in the therapeutic potential of parenterally administered fosfomycin disodium against systemic infections caused by CRE.

Currently, a significant knowledge gap on the pharmacokinetics (PK) and pharmacodynamics (PD) of parenteral fosfomycin against CRE still exists. Knowledge on the PK/PD of fosfomycin is crucial in the development of safe and effective dosing regimens, especially for systemic CRE infections (10). With regard to PD, it is essential to know the fosfomycin potency against the microorganism, expressed as the MIC. Beyond the MICs, it is also important to evaluate the PD of fosfomycin through time-kill studies (TKS), as previous studies have demonstrated the detection of fosfomycin heteroresistance and emergence of fosfomycin resistance in Pseudomonas aeruginosa and Enterobacterales (11–13). In addition, dose fractionation studies that investigated the relationship between PK/PD indices and bacterial killing for fosfomycin against multidrug-resistant P. aeruginosa using a dynamic in vitro PK/PD model had shown that regimens that exceed clinical exposures in plasma after intravenous administration were unable to suppress the emergence of resistant subpopulations (14). In our study, we investigated the in vitro PD of fosfomycin against clinical isolates of carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes (formerly known as Enterobacter aerogenes) via time-kill studies and population analysis profiles (PAPs).

(This study was presented in part at the ASM Microbe 2017, New Orleans, Louisiana, 1 to 4 June 2017.)

RESULTS

Susceptibility and whole-genome sequencing.

The fosfomycin MICs together with the resistant genotypes are shown in Table 1. Seven out of eight isolates were susceptible (MIC ranges from 0.5 to 32 mg/liter) to fosfomycin. One isolate was resistant (MIC, 64 mg/liter) to fosfomycin. Five out of eight isolates were found to carry fosA genes, while none were found to carry fosL-like genes after analysis with the recently discovered fosL-like gene sequence (15). ENT 702 (Klebsiella aerogenes) carries the kpc-2 gene as well as the mcr-1 gene. Four out of eight isolates had sequence types (STs) assigned while we were unable to assign STs for the remaining isolates due to unknown schemes.

TABLE 1.

Fosfomycin susceptibilities and genotypic characteristics of carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes

Species	Isolate	Site of culture	ST	fos gene		Carbapenemase	Fosfomycin MIC [range] (mg/liter)a
				fosA gene	fosL-like gene
Enterobacter cloacae	ATCC 13047		1	Present	Absent		32 [32–64] (S)
Enterobacter cloacae	ENT 276	Tissue	94	Absent	Absent	NDM-7	0.5 [0.25–0.5] (S)
Enterobacter cloacae	ENT 427	Urine	New	Absent	Absent	IMP-1	1 [0.5–1] (S)
Enterobacter cloacae	ENT 197	Wound	New	Present	Absent	IMP-1	2 [1–2] (S)
Klebsiella aerogenes	ENT 265	Bile	116	Present	Absent	KPC-2	4 [2–4] (S)
Klebsiella aerogenes	ENT 702	Urine	New	Absent	Absent	KPC-2	8 [8–16] (S)
Enterobacter cloacae	ENT 1268	Thigh	182	Present	Absent	NDM-1	16 [16–32] (S)
Enterobacter cloacae	ENT 376	Blood	634	Present	Absent	KPC-2	32 [32–64] (S)
Enterobacter cloacae	ENT 263	Urine	New	Present	Absent	NDM-1	64 [64–128] (R)

^aBrackets denote susceptibility, which is defined according to EUCAST clinical breakpoints for fosfomycin against Enterobacteriaceae (MIC ≤ 32 mg/liter, susceptible [S], and MIC > 32 mg/liter, resistant [R]).

Population analysis profile studies.

The PAPs of ATCC 13047 and the eight isolates are shown in Fig. 1. Heteroresistant subpopulations were present in all isolates, with tested concentrations ranging from 0.25 to 2,048 mg/liter fosfomycin. Heteroresistant subpopulations that were able to grow in the presence of >32 mg/liter fosfomycin were present in 4/8 isolates that had MICs ranging from 2 to 64 mg/liter. No heteroresistant subpopulations (in the presence of >32 mg/liter fosfomycin) were detected in the remaining four isolates with an MIC range of 0.5 to 8 mg/liter. At the highest test concentration of 2,048 mg/liter, significant bacterial growth (bacterial proportion from 1.45 × 10⁻⁶ to 2.75 × 10⁻⁶) was still observed in 2 out of 8 isolates. We observed that the heteroresistant subpopulation bacteria required similar incubation duration when compared to the original isolate as well as to the reference strain before fosfomycin exposure during quantitative plate counting. This may suggest no significant difference in growth rates among the heteroresistant subpopulations.

FIG 1.

Population analysis profiles of carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes (*, Klebsiella aerogenes). Vertical dotted line represents the fosfomycin breakpoint of 32 mg/liter. FOS, fosfomycin.

Time-kill studies.

The time-kill profiles of ATCC 13047 and the eight isolates are shown in Fig. 2. Overall, bactericidal effect was observed between the first 8 h before regrowth occurred at almost all fosfomycin concentrations and was observed for six out of eight isolates and the ATCC strain. Fosfomycin was bactericidal against all isolates within 2 to 4 h. Bactericidal activity occurred at 32× MIC for isolates with fosfomycin MICs of 0.5 to 4 mg/liter. For isolates with fosfomycin MICs of 8 to 64 mg/liter, bactericidal activity occurred at 4× MIC. However, at 8 h, sustained bactericidal activity was only observed with fosfomycin concentrations at 32× to 1,024× MIC against 7 out of 8 isolates. The remaining CRE (ENT 263; fosfomycin MIC, 64 mg/liter) exhibited regrowth by 8 h even with fosfomycin at 2,048 mg/liter. At 24 h, significant regrowth was observed across all isolates at all fosfomycin concentrations, except for ENT 276 and ENT 265 (Klebsiella aerogenes) at 2,048 mg/liter. In summary, fosfomycin demonstrated early bactericidal activity that increased with higher fosfomycin exposures until regrowth set in at 24 h.

FIG 2.

Time-kill profiles (average colony counts) of carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes (*, Klebsiella aerogenes). FOS, fosfomycin.

Pharmacodynamic profiles of fosfomycin against Enterobacter cloacae and Klebsiella aerogenes.

The model fits of the bacterial killing at 24 h and fosfomycin concentrations for each strain are shown in Fig. 3. In general, greater bacterial killing was observed as fosfomycin exposures increased from ∼10 to 1,000 mg/liter in all isolates. On average, no significant bacterial killing was observed beyond ∼500 mg/liter.

FIG 3.

Model fits of the 24-h bacteria burden for the carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes isolates after exposure to various fosfomycin concentrations (*, Klebsiella aerogenes). FOS, fosfomycin.

Pharmacological estimates, maximum effect (E_max), Hill slope, and 50% effective concentration (EC₅₀)/MIC are displayed in Table 2. Overall, the E_max model fit well to the time-kill data with an R² of >0.95 for all isolates. From the E_max model, mean maximum bacterial killing (E_max) (mean percent standard error [%MSE]) (ranging from ∼6 log₁₀ CFU/ml to 1.22 log₁₀ CFU/ml; %MSE, 10.2) decreased with increasing fosfomycin MICs. The EC₅₀ (%MSE, 35.3) was normalized by MICs for optimal comparison of results, as we included study isolates that had a wide variation of fosfomycin MIC values. The effective fosfomycin concentrations that drove half of the maximal killing effect ranged from ∼3× to 175× MIC. Specifically, the EC₅₀/MIC ranged from 3.045× to 9.305× MIC for isolates with fosfomycin MICs of 4 to 64 mg/liter. In contrast, the EC₅₀/MIC was 22.38×, 57.76×, and 175× MIC for isolates with fosfomycin MICs of 0.5, 1, and 2 mg/liter, respectively. Of note, ENT 197 (IMP-1-producing Enterobacter cloacae; fosfomycin MIC, 2 mg/liter) had substantially higher EC₅₀/MIC values, which meant that higher fosfomycin concentrations were needed to drive bacterial killing of ENT 197 compared to other isolates with comparable fosfomycin MICs.

TABLE 2.

Pharmacodynamic analysis of fosfomycin exposure against different isolates of carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes

Isolate	Fosfomycin concn (mg/liter) (× MIC) that exhibited bactericidal activity in TKS			Pharmacodynamic parameter estimates after 24 h of fosfomycin exposure
	2–4 h	8 h	24 h	Emax (log10 CFU/ml)	Hill slope	EC50/MIC	R2
ATCC 13047	128 (4×)	2,048 (64×)		6.288	1.002	21.353	0.990
ENT 276	16 (32×)	512 (1,024×)	2,048 (4,096×)	5.942	0.910	22.380	0.971
ENT 427	32 (32×)	128 (128×)		5.159	2.362	57.760	0.964
ENT 197	64 (32×)	2,048 (1,024×)		5.127	3.734	175.000	0.999
ENT 265a	128 (32×)	128 (32×)	2,048 (512×)	5.315	7.756	4.785	0.999
ENT 702a	32 (4×)	512 (64×)		5.133	1.793	9.305	0.999
ENT 1268	64 (4×)	512 (32×)		3.635	6.830	5.410	0.999
ENT 376	128 (4×)	2,048 (64×)		1.221	4.068	3.045	0.982
ENT 263	512 (8×)			2.630	7.903	7.581	0.999

^aKlebsiella aerogenes.

We observed a broad range of Hill slope values among the eight isolates (0.910 to 7.903 [%MSE, 38.8]). This suggested that the rate of fosfomycin killing against carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes can be concentration dependent or independent and is independent of fosfomycin MICs.

DISCUSSION

This study presents information on the PD of fosfomycin used to treat infections caused by carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes. Early bactericidal activity was observed for all tested isolates. However, rapid bacterial regrowth occurred.

With the number of CRE infections rising globally, fosfomycin has been identified as a potential replacement for traditional CRE therapy. Moreover, fosfomycin is generally well-tolerated and penetrates well into various body tissues to achieve sufficiently high concentrations for inhibition of bacterial growth during systemic infections (7, 16). The maximum clinically achievable concentrations of fosfomycin following intravenous administration of an 8-g dose for infected patients are in the range of 300 to 600 mg/liter with a half-life of 2 to 2.5 h (17).

From the PAPs in our study, heteroresistance to fosfomycin appears to be prevalent among the carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes isolates in our setting, even for those that are considered susceptible based on EUCAST breakpoints. This can be a cause for concern, as current routine susceptibility testing methods are unable to detect the presence of these heteroresistant subpopulations. We also observed fosfomycin regrowth at 24 h across most tested carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes isolates even at fosfomycin concentrations of 32× to 1,024× MIC. This suggests that treatment failure can occur upon fosfomycin monotherapy even at high doses due to the killing of susceptible subpopulations and amplification of the resistant subpopulations. Our findings are consistent with data published in previous studies (12, 18). The clinical implications of our findings should be further explored in future studies that can examine novel susceptibility methods that can detect fosfomycin heteroresistance and/or examine the impact across clinically important Gram-negative bacteria species. Moving forward, the utility of fosfomycin-based antibiotic combinations should be further evaluated against carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes.

In the pharmacodynamic profiling analysis, we found that fosfomycin killing did not appear to be concentration dependent or concentration independent but appeared to differ for each strain. Our findings are in concordance with current published literature, where concentration- and time-dependent activity and area under the concentration-time curve (AUC)/MIC have been suggested depending on bacterial species (10–12, 19). Interestingly, a greater magnitude of bacterial killing was observed as fosfomycin exposures increased from ∼10 to 500 mg/liter in the tested isolates, and no considerable increase in bacterial killing was observed beyond 500 mg/liter. Our study findings also suggest that different dosing strategies for fosfomycin may be necessary for different strains of carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes. This will require in vitro hollow fiber infection model and in vivo murine model studies for further validation.

To date, there are limited studies investigating the in vitro pharmacodynamics of fosfomycin against carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes. Fransen and colleagues had investigated the pharmacodynamics of fosfomycin against extended-spectrum β-lactamase (ESBL) and/or carbapenemase-producing Enterobacterales via time-kill assays and emergence of resistance studies (12). Our results were similar, as they reported that fosfomycin exhibited bactericidal activity for a short period before they observed regrowth across all strains with different fosfomycin MICs. In addition, they demonstrated the quick emergence of resistant subpopulations with fosfomycin exposure. In our study, near-bactericidal activity was only achieved with 2,048 mg/liter of fosfomycin exposure at 24 h against ENT 276 (fosfomycin MIC, 0.5 mg/liter; −2.7 log₁₀ CFU/ml from baseline), and complete bactericidal activity was achieved against ENT 265 (fosfomycin MIC, 4 mg/liter; approximately −5 log₁₀ CFU/ml from baseline). Together with our observation of fosfomycin heteroresistance occurring in carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes from our study, the utility of existing fosfomycin monotherapy may be severely limited based on routine susceptibility testing and detection of fosA genes. Although fosA-like genes were identified in 5/8 isolates, we did not find a relationship between the various genotypes and our time-kill assays. Fosfomycin in combination with a synergistically plausible antibiotic should be considered instead.

Our study has a few limitations. The PD of fosfomycin against carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes were performed via in vitro static conditions. Thus, we were unable to further investigate the effect of fosfomycin fluctuating concentrations on its activity against these isolates. In addition, we did not study the emergence of resistance to fosfomycin, although we clearly demonstrated the presence of heteroresistance. In vitro pharmacodynamics systems, such as hollow-fiber infection models, could be used to investigate dosing regimens simulating higher exposures especially within the first 8 h to achieve sustained bactericidal activity and suppress emergence of resistance with repeated dosing. In addition, the impact of mutation rates of fosfomycin-resistant mutants and their fitness costs in vivo need to be evaluated when designing new dosing regimens. Lastly, other species of CRE, such as Escherichia coli and Klebsiella pneumoniae, may be included to broaden the investigation on the use of fosfomycin against them.

In conclusion, fosfomycin was bactericidal against all carbapenem-resistant Enterobacter cloacae and Klebsiella aerogenes tested within 8 h. However, regrowth was observed at 24 h when exposed to maximum clinically achievable concentrations regardless of their fosfomycin MICs. Fosfomycin-based combination therapies should be explored as a potential solution to overcome problems experienced with monotherapy.

MATERIALS AND METHODS

Microorganisms.

Eight clinical CRE isolates and reference strain ATCC 13047 were included in the study. Six Enterobacter cloacae and two Klebsiella aerogenes (formerly known as Enterobacter aerogenes) isolates were included in the study. The clinical isolates were from nonstool origin, had known diverse carbapenem resistance mechanisms, and had varied fosfomycin susceptibility.

Antibiotics.

Fosfomycin disodium and glucose-6-phosphate (G6P) were obtained from Toronto Research Chemicals Inc. (Toronto, ON, Canada) and Sigma-Aldrich (St. Louis, MO, USA), respectively. Stock solutions were prepared by reconstituting drug powder with sterile water, aliquoted, and stored at −80°C.

Susceptibility testing.

Fosfomycin MICs were determined in triplicates on different days via the agar dilution method according to CLSI guidelines (20). In brief, Mueller-Hinton II agar (MHA) (BBL, Sparks, MD, USA) supplemented with 25 mg/liter G6P with fosfomycin concentrations ranging from 0.25 to 1,024 mg/liter in serial 2-fold dilutions was prepared and inoculated on the surface with 2 μl of 7 log₁₀ CFU/ml and incubated for 16 to 20 h at 37°C. E. coli ATCC 25922 was used as the quality control strain. Categorical susceptibility was determined according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoints for Enterobacterales. Isolates with MICs of ≤32 mg/liter were considered susceptible to fosfomycin while those with MICs of >32 mg/liter were considered resistant to fosfomycin (21).

Whole-genome sequencing.

Genomic DNA was prepared from overnight bacterial cultures, extracted, and purified with the Qiagen DNeasy blood and tissue kit (Qiagen Inc., Valencia, CA, USA). Paired-end sequencing was conducted on the MiSeq or HiSeq system (Illumina Inc., San Diego, CA). Sequence reads were adaptor removed and quality trimmed using Trimmomatic software (v.0.36) (22). The trimmed reads were used for de novo genome assembly using SPAdes (v.3.11.1) (23). Sequence types (STs) were determined by performing a Basic Local Alignment Search Tool (BLAST) search of the assembled contigs against a multilocus sequence typing (MLST) database (https://pubmlst.org/databases/). Antibiotic resistance genes of interest were identified using ResFinder v.3.2 (24). Additionally, the presence of the newly discovered fosL gene was identified from assembled contigs using BLAST (15).

Population analysis profiles.

Population analysis profile (PAP) assays were conducted on all isolates to detect fosfomycin heteroresistance. Heteroresistance was defined as the ability of subpopulations of an isolate to grow in the presence of fosfomycin concentrations higher than that predicted to be effective based on the measured MIC (25). We prepared fosfomycin-containing MHA plates supplemented with 25 mg/liter G6P-containing fosfomycin ranging from 0.25 to 2,048 mg/liter using the same testing strategy with time-kill studies. We had constrained the lowest fosfomycin concentration to be impregnated into MHA plates for each isolate at 0.5× MIC. A bacteria suspension of approximately 8 log₁₀ CFU/ml was prepared prior to each experiment (26). Serial dilutions of the bacterial suspension were performed, and 50-μl aliquots were spread onto fosfomycin-containing MHA plates and drug-free MHA plates for bacterial enumeration.

Time-kill studies.

Time-kill studies (TKS) were performed by exposing each study isolate in duplicate to static concentrations of fosfomycin ranging from 0.25 to 2,048 mg/liter. As the isolates included in this study had varied fosfomycin susceptibility, representative fosfomycin concentrations were tested for each isolate from 0.5× MIC to 2,048 mg/liter to cover the wide concentration range. Bacterial cultures were prepared using the UV absorbance method as described elsewhere (26). Fosfomycin was then added to cation-adjusted Mueller-Hinton II broth (Ca-MHB) (BBL, Sparks, MD) supplemented with 25 mg/liter G6P in conical flasks containing a total of ∼5 log₁₀ CFU/ml of bacteria suspension and incubated at 35°C (final volume, 16 ml). A growth control flask and a sterility control flask were included. Aliquots were obtained in duplicates from each flask after 0 h (baseline), 2 h, 4 h, and 8 h, and in triplicates at 24 h. Serial dilutions of the aliquots were then plated on MHA plates to perform bacterial enumeration (26). The final limit of quantification was 2.6 log₁₀ CFU/ml.

The frequency of heteroresistant subpopulations was calculated by dividing the number of colonies on the plate containing a fosfomycin concentration greater than the fosfomycin MIC of the isolate by the number of colonies on the plate without fosfomycin (10). PAPs were generated by plotting the bacterial counts (log₁₀ CFU/ml) against their respective fosfomycin concentrations.

Pharmacodynamic analysis.

Bactericidal activity (primary endpoint) was defined as a ≥3 log₁₀ CFU/ml decrease in the colony count compared to that of the initial inoculum (27). The mean killing effect at 24 h was characterized by an inhibitory sigmoid E_max model using the ADAPT II program; fitting was weighted by the inverse of the observation variances (28). For comparison between the isolates, the maximal bacterial kill (E_max), 50% effective concentration normalized by the MIC (EC₅₀/MIC), Hill slope, and R² were calculated for each time-kill study.