European Journal of Clinical Microbiology & Infectious Diseases

, Volume 32, Issue 6, pp 827–833

Combinations of cefoxitin plus other β-lactams are synergistic in vitro against community associated methicillin-resistant Staphylococcus aureus

Authors

    • Division of Pediatric Infectious DiseasesMayo Clinic
  • M. G. Fernandez
    • Paracelsus Medical Private University
  • N. Enthaler
    • Paracelsus Medical Private University
  • C. Graml
    • Paracelsus Medical Private University
  • K. E. Greenwood-Quaintance
    • Division of Clinical Microbiology, Department of Laboratory Medicine and PathologyMayo Clinic
  • R. Patel
    • Division of Clinical Microbiology, Department of Laboratory Medicine and PathologyMayo Clinic
Article

DOI: 10.1007/s10096-013-1817-9

Cite this article as:
Banerjee, R., Fernandez, M.G., Enthaler, N. et al. Eur J Clin Microbiol Infect Dis (2013) 32: 827. doi:10.1007/s10096-013-1817-9

Abstract

In vitro studies demonstrate that oxacillin minimal inhibitory concentrations (MICs) of methicillin-resistant S. aureus (MRSA) strains USA300 and 400 decrease in the presence of cefoxitin. The aim of this study was to characterize the activity of cefoxitin plus β-lactams against a collection of MRSA isolates. We assessed the in vitro antimicrobial activity of a selection of β-lactams alone and together with subinhibitory concentrations of cefoxitin against a collection of MRSA, methicillin-susceptible S. aureus (MSSA), and vancomycin-intermediate S. aureus (VISA) isolates using MICs and time kill assays. For community-associated (CA) MRSA strains USA300 and USA400, MICs of nafcillin, cefazolin, cephalexin, cefuroxime, ceftriaxone and cefotaxime decreased by 8- to 64-times in the presence of 10 μg/ml cefoxitin. In contrast, for hospital-associated (HA) strains COLn, N315, and Mu50, there was no change in any β-lactam MIC in the presence of cefoxitin. When combined with cefoxitin, the cephalexin MIC decreased for eight CA-MRSA and five MSSA sequence types but did not change for seven HA-MRSA sequence types. β-lactam/cefoxitin combinations were synergistic against CA- but not HA-MRSA strains in time kill assays. Cefoxitin combined with a variety of β-lactams enhances their activity against CA-MRSA strains in vitro. Further studies of combination β-lactam therapy may provide insight into β-lactam biology, penicillin binding protein cooperativity, and novel therapeutic strategies against MRSA.

Introduction

Methicillin-resistant S. aureus (MRSA) is a major cause of nosocomial and community-associated infection worldwide [1, 2]. Infection with MRSA causes clinical manifestations ranging from minor skin infections to life threatening bacteremia and pneumonia. MRSA prevalence has increased dramatically in recent years. MRSA is now the leading cause of skin and soft tissue infection in patients seen in US emergency departments [3] and causes more fatalities in the United States than any other infectious agent [4]. This has caused a shift in therapy for suspected staphylococcal infections such that non-β-lactam antibiotics with activity against MRSA are increasingly prescribed [5]. Consequently, MRSA strains which have typically been susceptible to non-β-lactam antibiotics (e.g., clindamycin, tetracyclines, trimethoprim-sulfamethoxazole) are now increasingly resistant to these antibiotics as well [6].

Vancomycin, the traditional drug of choice for treatment of serious MRSA infections, is associated with high rates of persistent bacteremia, treatment failure, and relapse [7, 8]. There are reports showing an increase in the number of clinical S. aureus isolates with elevated minimal inhibitory concentrations (MICs) to vancomycin, a phenomenon known as the vancomycin “MIC creep” [911]. Vancomycin-resistant (VRSA), vancomycin-intermediate (VISA) and vancomycin-heteroresistant (hVISA) S. aureus strains have emerged [1215]. There is an urgent need for new, more effective antibiotics to treat MRSA infections.

MRSA is resistant to β-lactam antibiotics because it produces penicillin binding protein 2a (PBP2a), which is encoded by mecA, and likely originated from a coagulase-negative Staphylococcus species [1619]. Unlike other staphylococcal PBPs which exhibit high affinity binding of β-lactams, PBP2a is a low affinity PBP and remains active at lethal concentrations of all β-lactams, except the newly approved cephalosporin, ceftaroline.

In recent years, an important role for penicillin-binding-protein 4 (PBP4) in S. aureus β-lactam resistance has emerged. PBP4, the only low molecular weight PBP in S. aureus, is believed to be a carboxypeptidase and transpeptidase that functions in cell wall cross-linking [20, 21]. The structure of PBP4 was published and the recombinantly expressed protein displayed β-lactamase activity [22].

It had been previously thought that PBP4 was not essential because its inactivation had minimal effects on methicillin resistance in COL, a prototypical HA strain [23]. However, Memmi et al. showed that disruption of PBP4 in CA-MRSA strains rendered these strains susceptible to β-lactam antibiotics, suggesting that PBP4 plays different roles in CA- versus HA-MRSA strains [24]. Furthermore, cefoxitin in combination with oxacillin caused reductions in oxacillin MICs and minimum bactericidal concentrations (MBC) in CA MRSA strains [24]. Cefoxitin is a second generation cephalosporin that binds irreversibly and tightly to PBP4, with a several hundred-fold greater affinity for this protein than most other β-lactams [25].

The findings of Memmi et al., as well as preliminary data from our group, led us to hypothesize that cefoxitin in combination with β-lactams may be a novel therapeutic strategy against MRSA. The aim of this study was to characterize the in vitro antimicrobial activity of cefoxitin and a variety of β-lactams alone and together against a collection of MRSA isolates using MICs and time kill assays.

Materials and methods

Reagents/strains

All strains were kindly provided and genotyped by Drs. Binh Diep and Henry Chambers, UCSF. CA-MRSA strains had SCCmec type IV, and HA-MRSA strains had SCCmec types I, II, or III. Strains were grown in trypticase soy broth (TSB) (Becton, Dickinson, and Co., Franklin Lakes, NJ). Nafcillin, cefoxitin, cefazolin, ceftriaxone, cefuroxime, cefotaxime, imipenem, clindamycin, vancomycin, cephalexin, linezolid, amoxicillin, and levofloxacin were obtained from Sigma Chemicals Co. (St. Louis, MO). Clavulanate and cefoxitin were obtained from The United States Pharmacopeia Co (Rockville, MD). Antibiotic dilutions were made fresh prior to each experiment.

Determination of MICs

Bacteria were grown for 4 h and diluted to 0.5 McFarland in TSB. Each strain was diluted (1:1000) in cation-adjusted Mueller-Hinton Broth (CAMHB) containing 25 μg/ml Ca2+ and 12.5 μg/ml Mg2+. A total of 0.5 ml of this dilution was added to tubes containing serial dilutions of antibiotic in CAMHB broth. After 16–20 h incubation at 37 °C, tubes were examined. The lowest concentration of drug that prevented appearance of turbidity was defined as the MIC. Experiments were done in triplicate.

Time kill assays

A 2-ml sample of an overnight broth culture was diluted to 0.5 McFarland and further diluted 1:2000 (low inoculum), or 1:100 (standard inoculum) into CAMHB media that contained antibiotics alone, or in combination. At 0, 6 h, and 24 h incubation at 37°, 0.1 ml was removed, diluted 10-fold in broth, and 0.1 ml plated onto blood agar plates. Plates were incubated for 24 h, after which the number of colonies was counted. For high inoculum time kill studies, overnight cultures were diluted 1:10,000 in CAMHB and incubated for 6 h with vigorous shaking before exposure to antibiotics, in an attempt to maintain bacteria in log rather than stationary phase. Synergy was defined as >2log10 reduction in CFU/ml by a combination of antibiotics compared to either antibiotic alone. Experiments were done in triplicate. Antibiotic concentrations chosen for time kill assays were equivalent to the MIC of the drug (cefazolin and nafcillin for USA 300), or below the MIC and estimated peak plasma levels when strains were highly resistant (cefuroxime for USA300 and USA400, nafcillin, cefazolin, and cefuroxime for all HA MRSA strains).

Results

CA- but not HA-MRSA strains are susceptible to a variety of β-lactams in combination with cefoxitin

Subinhibitory cefoxitin (10 μg/ml) combined with a variety of β-lactams reduced β lactam MICs of CA MRSA strains USA300 and USA400 (Table 1), by 8–64 times, for all agents except imipenem and amoxicillin/clavulanate. MICs for imipenem and amoxicillin/clavulanate alone were fairly low, and in the presence of cefoxitin decreased by 4–8 times for amoxicillin/clavulanate and by 0–4 times for imipenem (Table 1). The most dramatic MIC reduction occurred when cefoxitin was combined with cephalexin or cefuroxime. In contrast, there was no reduction in any β-lactam MIC for the three HA-MRSA strains, N315, COL, and Mu50. Only strain N315, a HA-MRSA strain which is heterogeneously resistant to β-lactams, displayed a minimal reduction in β-lactam MICs in the presence of cefoxitin. Both CA- and HA-MRSA strains displayed little or no reduction in MICs of non β-lactam antibiotics (vancomycin, clindamycin, trimethoprim-sulfamethoxazole, levofloxacin, or linezolid) when these agents were combined with cefoxitin.
Table 1

β-lactam MICs (micrograms/ml) alone and in combination with cefoxitin for select MRSA strains. Number before slash is MIC of β-lactam alone. Number after slash is MIC of β-lactam in combination with 10 mg/ml cefoxitin. USA 300 and 400 are CAMRSA; N315, Mu50, and COL are HA-MRSA

Strain

Cefoxitin

Nafcillin

Cefazolin

Cephalexin

Cefuroxime

Ceftriaxone

Imipenem

Vancomycin

Clindamycin

TMP-SMX

Levofloxacin

Linezolid

USA300

32

4/0.5

16/0.5

64/1

64/2

128/16

1/0.25

1/1

<0.125/<0.125

1/0.5

0.25/0.25

4/1

USA400

32

4/0.5

4/0.5

64/0.5

>128/2

128/8

0.1/0.1

1/1

<0.125/<0.125

0.5/0.5

0.25/0.06

2/1

N315

32

16/8

64/1

128/64

>128/128

>128/32

8/8

0.5/1

>128/>128

1/1

0.5/<0.125

2/1

Mu50

>128

>128/>128

>128/>128

128/64

>128/>128

>128/>128

32/32

4/4

>128/128

0.5/0.5

16/8

1/1

COL

>128

128/>128

>128/>128

128/64

>128/>128

>128/>128

32/32

0.5/1

<0.125/<0.125

4/2

0.25/0.25

1/1

TMP-SMX trimethoprim-sulfamethoxazole

In vitro susceptibility to cephalexin plus cefoxitin combinations distinguishes the major CA-MRSA strains from HA-MRSA strains

We next examined susceptibility to cephalexin/cefoxitin and cefuroxime/cefoxitin combinations among 20 S. aureus strains representing the major HA-MRSA, CA-MRSA and methicillin-susceptible S. aureus (MSSA) lineages from around the world. We evaluated susceptibility to the drug combinations by MIC testing (Table 2). We chose cephalexin and cefuroxime for these studies because CA MRSA strains displayed the most dramatic reduction in MIC for these agents when combined with cefoxitin (Table 1). None of the classic HA-MRSA strains containing SCCmec types I–III (COL, NY/Japan, UK, Portuguese/Brazilian, Iberian, USA100, and USA200) were susceptible to the antibiotic combinations with the exception of USA100, whose cefuroxime MIC decreased by 32 times in the presence of cefoxitin (Table 2).
Table 2

Cephalexin and cefuroxime MICs decrease in the presence of cefoxitin for CA-MRSA and MSSA strains. If MIC of drug/MIC drug plus cefoxitin = 1, no reduction was detected with combination therapy. Cefoxitin concentration was 10 μg/ml for MRSA strains and 1 μg/ml for MSSA strains. Range of values, tested in triplicate, are shown

PFGE

MLST

SCCmec type

MIC drug/MIC drug + cefoxitin

Cephalexin

Cefuroxime

USA100

ST5 HA

II

1

32

USA200

ST36 HA

II

1

1

USA300

ST8 CA

IV

32–64

32–64

USA400

ST1 CA

IV

64–128

64–128

USA500

ST8 CA

IV

64

32–64

USA600

ST45 CA

IV

128–256

8–64

USA700

ST72 CA

IV

2–8

16–32

USA800

ST5 CA

IV

16–64

8–128

USA1000

ST59 CA

IV

4–32

16–32

USA1100

ST30 CA

IV

8–128

32

COL

ST8 HA

I

2

1

NY/Japan MRSA

ST5 HA

II

2

1

UK Epidemic MRSA

ST36 HA

II

1

1

Portuguese/Brazilian

ST239 HA

III

1

1

Iberian Clone

ST247 HA

I

1

1

MSSA

ST1

8

MSSA

ST15

8

MSSA

ST22

8

MSSA

ST25

8

MSSA

ST20

8

CA community-associated, HA hospital-associated, PFGE pulse field gel electrophoresis type, MLST multilocus sequence type

All MRSA strains displaying susceptibility to the antibiotic combinations contained SCCmecIV. USA300, 400, 500, and 600 displayed a marked reduction (32–256 times) in cephalexin and cefuroxime MICs in the presence of cefoxitin. USA 700, 800, 1000, and 1100 displayed a more variable reduction in cephalexin and cefuroxime MICs (2–128 times) in the presence of cefoxitin (Table 2). In all five MSSA sequence types tested the cephalexin MIC decreased by eight times in the presence of 1 μg/ml cefoxitin (Table 2). In one VISA isolate the cephalexin MIC decreased in the presence of cefoxitin from 64 μg/ml to <0.125 μg/ml (data not shown in Table 2).

β-lactam/cefoxitin combinations are synergistic against CA but not HA MRSA strains in time kill assays

Time kill assays were next performed using nafcillin and cefazolin, narrow-spectrum anti-staphylococcal agents used commonly in the United States, and cefuroxime, which showed dramatic reductions in MIC when combined with cefoxitin. For CA-MRSA strains USA300 and 400, combinations of cefoxitin with nafcillin, cefazolin, or cefuroxime, were synergistic (defined as >2 log reduction at 24 h in CFU/ml between drug combination and the most active agent alone) in time kill assays at low (104.5–4.9 CFU/ml) and standard (105.5–6.5 CFU/ml) but not at high (107 CFU/ml) starting inocula (Table 3). Maximum synergy occurred when 10 μg/ml cefoxitin was combined with these β-lactams at their MIC concentrations, and did not increase when higher concentrations of cefoxitin or β-lactam were used (data not shown). No synergy was observed for the HA MRSA strains COL and N315.
Table 3

Change in log10CFU/ml at 24 h between β-lactam/cefoxitin combination vs. β-lactam alone in time kill experiments. Nafcillin 4 μg/ml; cefazolin 16 μg/ml, except for USA400 where 4 μg/ml was studied; cefuroxime 8 μg/ml

Strain

Inoculum size

10 4.5–4.9

10 5.5–6.5

107

Nafcillin

Cefazolin

Cefuroxime

Nafcillin

Cefazolin

Cefuroxime

Cefazolin

USA300

−4.9

−3.9

−4.9

−2.1

−2.8

−2.0

0

USA400

−4.7

−3.8

−4.3

−1.5

−4.7

−2.4

−0.9

COL

+0.3

−0.1

−0.1

0

0

+0.4

−0.4

N315

+0.2

+0.8

−0.01

−0.1

−0.02

−0.3

−0.02

At standard inocula (105.5–6.5 CFU/ml), activity of cefazolin plus cefoxitin, was superior to that of the nafcillin plus cefoxitin or cefuroxime plus cefoxitin for USA 300 and USA 400 (Fig. 1a and b). When comparing the log10 reduction in CFU/ml at 24 h from starting inoculum, the activity of the cefazolin plus cefoxitin combination was inferior to that of vancomycin, but superior to that of clindamycin (Fig. 1a and b). At standard and high inocula, bacterial killing occurred at 6 h, but regrowth was observed by 24 h for all cefoxitin-containing combinations vs. USA300 and for the cefuroxime/ cefoxitin combination vs. USA400. No drug combination showed activity against the HA MRSA strains COL and N315 in the time kill experiments (Fig. 1c and d). Only vancomycin displayed bactericidal activity against HA strains.
https://static-content.springer.com/image/art%3A10.1007%2Fs10096-013-1817-9/MediaObjects/10096_2013_1817_Fig1_HTML.gif
Fig. 1

Activity of cefoxitin and β-lactam combinations against MRSA strains in time kill experiments. a USA300. b USA400. c COL. d N315. Solid line, no drug; square, cefazolin (16 μg/ml for all strains except COL, 64 μg/ml for COL) plus cefoxitin (10 μg/ml); triangle, nafcillin (4 μg/ml for CA strains, 64 μg/ml for HA strains) plus cefoxitin (10 μg/ml); diamond, cefuroxime (8 μg/ml for CA strains, 32 μg/ml for HA strains) plus cefoxitin (10 μg/ml); circle, vancomycin (16 μg/ml); open triangle, clindamycin (8 μg/ml). Growth was similar to the no drug control for each β-lactam alone (not shown)

Discussion

We have demonstrated that combinations of cefoxitin plus a variety of other β-lactams act synergistically in vitro against CA-MRSA but not HA-MRSA. This study expands upon the study by Memmi et al., which demonstrated that oxacillin and subinhibitory cefoxitin are synergistic against CA-MRSA [24].

The clinical utility of cefoxitin/β-lactam combination therapy is unclear. Since the in vitro activity of the cefoxitin/β-lactam combinations is greatest at low inoculum, the drug combination may not be effective for treatment of high inoculum S. aureus infections like endocarditis or meningitis (where organism burden can be as high as 109–1011 bacteria per gram of tissue) [26, 27], but may be effective in patients with low organism burden infections, or removable foci of infection. In addition, cefoxitin/β-lactam combinations were most inhibitory early in time kill assays, suggesting they might become less effective over the course of antimicrobial therapy. While the in vitro activity of the drug combination is not superior to that of vancomycin, the current gold standard for treatment of severe MRSA infections, combination β-lactam therapy may be an option in treatment of infections caused by some VISA strains, for which few bactericidal agents exist, especially if daptomycin nonsusceptibility or intolerance occurs [28, 29].

The cephalexin MIC for one clinical VISA isolate decreased dramatically in the presence of cefoxitin. Because many VISA strains are USA100 [30], and the USA100 MRSA isolate tested in our experiments displayed a significant cefuroxime MIC reduction in the presence of cefoxitin (Table 2), the effect of combination β-lactam therapy on VISA strains deserves further evaluation. Also, an unexpected finding was that β-lactam MICs for all MSSA strains decreased by eight times in the presence of cefoxitin. MSSA strains cause significant morbidity and mortality, and improved killing of MSSA strains has therapeutic implications that should be examined further. Evaluation of the cefoxitin plus β-lactam combinations in animal models is needed before understanding the therapeutic potential of combination β-lactam therapy.

Combination β-lactam therapy has been used previously for treatment of other Gram positive infections. Ampicillin plus ceftriaxone or cefotaxime have been used clinically for treatment of enterococcal cardiac and orthopedic infections [3135]. The proposed mechanism in enterococci is sequential PBP inhibition, with ampicillin targeting PBPs 4 and 5, and ceftriaxone targeting PBPs 2 and 3 [33].

Sequential PBP inhibition is a therapeutic strategy that warrants further study in S. aureus. PBPs are considered part of a large multi-membered holoenzyme complex and interactions of PBPs with each other and with other proteins may regulate cell wall biosynthesis or degradation. There is evidence that PBP2, 2a, and 4 work cooperatively in S. aureus [36]. Clinically available β-lactams have differing affinities for the staphylococcal PBPs, which are presumed to have different cellular functions [25, 37]. This might explain why cefoxitin does not decrease the MICs of all β-lactams to the same extent in CA strains.

The basis for why cefoxitin plus β-lactam combinations are active against CA- but not HA-MRSA strains remains unclear, but may result from the tight binding of cefoxitin to PBP4, which plays an important role in β-lactam resistance of CA strains only. Memmi et al. demonstrated that disruption of pbp4 caused β-lactam susceptibility in CA-MRSA strains, but had no effect on β-lactam resistance of HA-MRSA strains. Furthermore, disruption of pbp4 reduced expression of pbp2 after challenge with oxacillin in CA but not HA MRSA strains [24]. It has been previously shown that meropenem and imipenem (carbapenems that bind tightly to PBP4, with affinity similar to that of cefoxitin) act synergistically with other β-lactams against MRSA [38, 39]. Why PBP4 is an important drug resistance determinant in CA- but not HA-MRSA strains may be related to the SCCmecIV element found in all CA-MRSA strains. There is evidence of differential gene expression among strains with SCCmec type II and SCCmec IVa elements [40].

A limitation of this study is that it investigates only in vitro antimicrobial activity, although we used physiologic antibiotic concentrations that are well below the peak plasma levels found in humans for cefazolin, nafcillin, cefuroxime, and cefoxitin [41]. In addition, we did not explore the properties of colonies that grew in the time kill experiments after exposure to the drug combinations. Some appeared morphologically abnormal and may represent persisters or drug-resistant subpopulations, as has been observed in earlier animal studies examining β-lactam therapy for MRSA [42, 43]. Despite these limitations, our in vitro observations are intriguing and suggest that further evaluation of cefoxitin plus β-lactam combinations may provide insight into β-lactam biology, PBP cooperativity, and novel therapeutic strategies against MRSA.

Acknowledgments

We thank Mark S. Rouse, Mayo Clinic, for helpful discussion and technical assistance, and are very grateful to Li Basuino, Binh Diep, and Henry Chambers of University of California, San Francisco, for providing strains and technical advice. M.G.F. was supported in part by funds from the German National Academic Foundation.

Funding

This work was supported by the Department of Pediatric and Adolescent Medicine, Mayo Clinic, Rochester, MN.

Transparency declarations

All authors: No conflicts of interest.

Copyright information

© Springer-Verlag Berlin Heidelberg 2013