Activity of mecillinam alone and in combination with other beta-lactam antibiotics.

The in vitro activities of mecillinam, ticarcillin, cefamandole, and cefoxitin, singly and in all possible combinations, against 53 clinical isolates were studied by a checkerboard method of determining minimal inhibitory concentrations. For selected representative strains, bactericidal activity was determined by minimal bactericidal concentrations and killing curves. Mecillinam was the least active antibiotic against gram-positive cocci, Pseudomonas aeruginosa, and Bacteroides fragilis and the most active against Enterobacteriaceae. Reproducibility of mecillinam minimal inhibitory concentrations for susceptible Enterobacteriaceae was often poor, however, due to minor variations in inoculum size. When mecillinam resistance was observed with Enterobacteriaceae, partial inhibition could be demonstrated at concentrations below minimal inhibitory concentrations, and bacterial cells were consistently ovoid or round; under those conditions the addition of a second study antibiotic resulted in marked synergistic inhibition and killing which was independent of inoculum size and susceptibility to the second antibiotic. In contrast, synergy with mecillinam against mecillinam-susceptible strains or with other antibiotic combinations against any species was not consistently observed.

The in vitro activities of mecillinam, ticarcillin, cefamandole, and cefoxitin, singly and in all possible combinations, against 53 clinical isolates were studied by a checkerboard method of determining miniimal inhibitory concentrations. For selected representative strains, bactericidal activity was determined by minimal bactericidal concentrations and killing curves. Mecillinam was the least active antibiotic against gram-positive cocci, Pseudomonas aeruginosa, and Bacteroides fragilis and the most active against Enterobacteriaceae. Reproducibility of mecillinam minimal inhibitory concentrations for susceptible Enterobacteriaceae was often poor, however, due to minor variations in inoculum size. When mecillinam resistance was observed with Enterobacteriaceae, partial inhibition could be demonstrated at concentrations below miimal inhibitory concentrations, and bacterial cells were consistently ovoid or round; under those conditions the addition of a second study antibiotic resulted in marked synergistic inhibition and killing which was independent of inoculum size and susceptibility to the second antibiotic. In contrast, synergy with mecillinam against mecillinam-susceptible strains or with other antibiotic combinations against any species was not consistently observed.
Mecillinam (formerly called FL 1060) is the prototype of a relatively new class of fi-lactaxr antibiotics. Unlike the penicillins, cephalosporins, and cephamycins, it has a 6-amidino rathei than a 6-acylamino side chain. Its spectrum of activity is different, with marked activity against some Enterobacteriaceae but little activity against gram-positive organisms (6). Its effects on Escherichia coli, and presumably other Enterobacteriaceae, are also different than those of other ,8-lactam antibiotics. None of the known enzymes of cell wall synthesis is inhibited, and the formation of filaments or spheroplasts which rapidly lyse is not observed. Over a wide range of concentrations, ovoid, osmotically stable round cells are formed which divide slowly and eventually lyse (9). The unique biochemical and morphological effects of mecillinam on E. coli have been attributed to its affinity for interacting with penicillin-binding protein 2 and thus affecting cell shape. It does not interact with penicillin-binding proteins 1 or 3, the preferential targets of other fB-lactam antibiotics (11).
Because mecillinam has a unique mechanism of action among /.?-lactam antibiotics against bacterial cells, it was postulated and subsequently shown, both in vitro (15) and in animal models (5), that it may act synergistically with those drugs.
This paper reports on the in vitro activity of mecillinam alone and in combination with a penicillin (ticarcillin), a cephalosporin (cefamandole), and a cephamycin (cefoxitin) against a variety of clinical isolates. For comparison, the activities of ticarcillin, cefamandole, and cefoxitin, singly and in all possible combinations, were studied simultaneously.

MATERLALS AND METHODS
Organisms. Fifty-three bacterial strains which were isolated from patients hospitalized in The Ohio State University Hospitals were studied. They were selected because they represented a broad spectrum of common pathogens and demonstrated varied, but typical, patterns of susceptibility to the four study antibiotics. Included were strains of Staphylococcus aureus, Streptococcus faecalis, Escherichia coli, Klebsiella pneumoniae, Enterobacter cloacae, Enterobacter aerogenes, Serratia marcescens, Proteus mirabilis, Pseudomonas aeruginosa and Bacteroides fragilis. All isolates, except the strains of B. fragilis, were blood isolates.
Antibiotics. The antibiotics studied included mecilhinam obtained from Hoffmann-LaRoche Inc., Nutley, N.J.; ticarcillin obtained from Beecham Products, Pittsburgh, Pa.; cefamandole obtained from Eli Lilly and Co., Indianapolis, Ind.; and cefoxitin obtained from Merck Sharp and Dohme, West Point, Pa. Laboratory standards were diluted as recommended by manufacturers to stock concentrations of 1,000 ,tg/ml and used immediately or frozen at -20°C, for up to 1 month, until used. Susceptibility tests. Minimal inhibitory concentrations (MICs) of the four study antibiotics were determined singly and in combinations by the checkerboard method with microdilution techniques previously described (4). The media used were Trypticase soy broth (BBL Microbiology Systems, Cockeysville, Md.) for S. aureus and S. faecalis, Mueller-Hinton broth (Difco Laboratories, Detroit, Mich.) supplemented with CaCl2 and MgCl2 to contain 50 mg of calcium per liter and 25 mg of magnesium per liter for Enterobacteriaceae and P. aeruginosa and Schaedler broth (Difco) for B. fragilis. Antibiotic concentrations consisted of log2 dilutions which ranged from 128 to 0.13 pg/ml or 16 to 0.01 pg/ml. Inoculum sizes of approximately 2 x 104 and 2 x 105 colony-forming units (CFU) per ml were tested. Incubation was for 18 to 20 h at 37°C in a room-air incubator, except for B. fragilis which was incubated in an anaerobic glove box containing 85% N2, 10% H2, and 5% CO2 for 48 h. MICs were read as the lowest concentration of each antibiotic which completely inhibited visible growth. Antibiotics were considered to be synergistic when there was a fourfold or greater reduction in MICs of both drugs in the combination as compared with that of each drug alone. Antibiotics were considered to be antagonistic when there was a fourfold or greater increase in MICs of both drugs in the combination or when the addition of an inactive antibiotic to an active antibiotic increased the MIC of the latter by fourfold or more. Antibiotics were considered to be indifferent when changes in MICs were intermediate.
With three selected strains of Enterobacteriaceae (E. coli 1924, K. pneumoniae 1864, and S. marcescens 1923) which showed typical patterns of inhibition by mecillinam (see Results), MICs of mecillinam and one or more of the other study antibiotics, singly and in combinations, were repeated with inocula of approximately 2 x 103, 2 x104, 2 x 105, and 2 x 106 CFU/ml prepared from the same culture. Bactericidal activity of the antibiotics was studied by determining minimal bactericidal concentrations (MBCs) and performing killing curves (1) with the four inoculum sizes. To determine MBCs, MIC plates which had been incubated for 20 h were subcultured to microdilution plates containing sterile broth by using disposable inoculators (Dynatech Laboratories, Alexandria, Va.). The subcultures were incubated overnight. The absence of growth in the subculture from a given well of an MIC plate indicated a reduction in the original number of CFU to <102 CFU/ml (>99.9% kill with an original inoculum of 2 x 105 CFU/ml) in that well. MBCs were read as the lowest concentrations of antibiotics which yielded no growth in subcultures after overnight incubation. To perform killing curves, the same cultures tested above were used. The appropriate number of organisms was incubated with clinically achievable concentrations of the test antibiotics, singly and in combinations, in 50 ml of broth for 20 h. Organisms inoculated into antibiotic-free broth served as growth controls. After 0, 4, and 20 h of incubation, pour plate colony counts on portions of cultures were performed.

RESULTS
The MICs of the four study antibiotics for the 53 bacterial strains tested (with an inoculum of 2 x 104 CFU/ml) are shown in Table 1. When there was more than one pattem of susceptibil-ACTIVITY OF MECILLINAM 907 ities for certain species (E. coli, K. pneumoniae, E. cloacae, E. aerogenes, S. marcescens, and B. fragilis), strains in those species were subdivided. Among the Enterobacteriaceae, only eight strains (two E. aerogenes and six S. marcescens) were resistant to 2 jig of mecillinam per ml. Partial inhibition of those strains occurred at concentrations below the MICs. It was manifested as growth which was obviously less than that observed in control wells. There was a clearcut endpoint between the wells with partial inhibition and those with no inhibition of growth.
The interactions of the four study antibiotics in the six possible pair combinations are also show in Table 1. Against S. aureus, S. faecalis, P. aeruginosa, and B. fragilis, all antibiotic combinations were either synergistic or indifferent; when synergy was observed, it was not marked. Against Enterobacteriaceae, the patterns of antibiotic interaction were similar, with two exceptions. The first was that mecillinam showed marked synergy with ticarcillin and cefamandole against all eight (two E. aerogenes and six S. marcescens) and with cefoxitin against five (two E. aerogenes and three S. marcescens) of the eight strains which had mecillinam MICs greater than 2 (8 to >128) ,Lg/ml. The synergy was independent of susceptibility to the second antibiotic. The second exception was that cefoxitin antagonized the activity of ticarcillin and cefamandole against E. aerogenes and E. cloacae. Strains which were susceptible to clinically achievable concentrations of ticarcillin and cefamandole were resistant to those antibiotics in the presence of cefoxitin.
When MICs were performed with an inoculum of approximately 2 x 105 CFU/ml (not shown), mecillinam MICs for S. aureus, S. faecalis, P. aeruginosa, and B. fragilis remained 64 ug or higher per ml and MICs of E. coli remained 2 jig or less per ml. With some strains of Enterobacteriaceae which were mecillinam-susceptible with an inoculum of 2 x 104 CFU/ml, however, mecilhinam MICs were >128 ,tg/ml when tested with the higher inoculum and partial inhibition was observed at concentrations between the two MICs. When combination MICs were performed, mecilhinam showed marked synergy with other study antibiotics against mecillinamresistant strains including those which had demonstrated high mecillinam MICs only with the larger inoculum. Significant changes in MICs of ticarcilhin, cefamandole, or cefoxitin were not observed by changing the inoculum size from 2 x 104 to 2 x 105 CFU/ml, and the phenomenon of partial inhibition was never observed.
The effects of varying inoculum size from 2 X 103 to 2 x 106 CFU/ml on the MICs and MBCs of mecillinam and comparative study   (Table 1).
antibiotics, singly and in combinations, for three selected strains of Enterobacteriaceae are shown in Table 2. With E. coli 1924 (which typified mecillinam-susceptible Enterobacteriaceae whose MICs were unaffected by inoculum size in the above experiments), inoculum size only affected mecillinam MBCs; combinations of mecillinam and cefamandole were synergistic when mecillinam MBCs were high. With K. pneumoniae 1864 (which typified Enterobacteriaceae whose MICs were markedly affected by inoculum size in the above experiments), inoculum size affected mecillinam MICs as well as MBCs; combinations of mecillinam and cefamandole or cefoxitin were synergistic when mecillinam MICs and MBCs were high. With S. marcescens 1923 (which typified mecillinam-resistant Enterobacteriaceae whose MICs were unaffected by inoculum size in the above experiments), combinations of mecillinam and ticarcillin or cefamandole were synergistic in both inhibitory and bactericidal activities.
After MIC plates for E. coli 1924, K. pneumoniae 1864, and S. marcescens 1923 were subcultured for determination of MBCs, the contents of mecillinam-containing wells which had visible growth were examined by phase-contrast microscopy and Gram stains. Organisms from wells below MICs were ovoid or round regardless of strain, inoculum size, or antibiotic concentration.
The killing curves for mecillinam and comparative study antibiotics, singly and in combinations, with the three selected strains of Enterobacteriaceae were performed simultaneously with the MICs and MBCs shown in Table 2 and are illustrated in Fig. 1

DISCUSSION
Previous studies have indicated that mecillinam has little in vitro activity against grampositive bacteria, P. aeruginosa, and other nonfermentative gram-negative bacilli or anaerobic bacteria (2, 3, 6, 7). In one study (8), combinations with other fl-lactam antibiotics against gram-positive organisms and anaerobes were indifferent, but synergy was observed with some strains of P. aeruginosa. In another study (13), mecillinam antagonized the activity of ampicillin but was synergistic with carbenicillin against some strains of Bacteroides. In the present study, mecillinam had little activity against those organisms, and combinations with other B8-lactam antibiotics were usually indifferent. The occasional synergy which was observed occurred no more frequently than with nonmecillinam fi-lactam antibiotic combinations.
Ever since the original report of mecillinam (6), it has been clear that it has a broad spectrum of in vitro activity against Enterobacteriaceae (2,7,14) and that it is often synergistic with other ,B-lactam antibiotics against those organisms (8,15). The interpretation of susceptibility tests has been difficult, however, because of variations in media and inoculum size which markedly affected results. To solve the problem of inconsistency of results, the use of NIH medium (7) was recommended for dilution susceptibility tests because of its "proper" osmolality and conductivity, but acceptance of that recommendation is not practical because it is unknown whether MICs determined in NIH broth are better predictors of in vivo efficacy than MICs determined in other media and because cationsupplemented Mueller-Hinton broth has become the standard medium for the in vitro sus-ceptibility testing of Enterobacteriaceae (12). In the present study, which used cation-supplemented Mueller-Hinton broth, Enterobacteriaceae fit into three categories of susceptibility to mecillinam. Some, typified by E. coli 1924, were highly susceptible regardless of inoculum size, and marked synergy with other f8-lactam antibiotics was not observed. Others, typified by K. pneumoniae 1864, were variably susceptible, depending on inoculum size, and some, typified by S. marcescens 1923, were resistant regardless of inoculum size. Under conditions when mecillinam MICs and MBCs were high, there was partial inhibition of growth, morphologically abnormal cells and marked synergy with other f3lactam antibiotics at concentrations below MICs, indicating that mecillinam had some effect on all strains of Enterobacteriaceae studied.
The only antagonism among ,8-lactam antibiotics observed in the present study was the antagonism of cefamandole and ticarcillin by cefoxitin against cefoxitin-resistant strains of Enterobacter. Presumably, cefoxitin induced,t/lactamase production by those strains (10) which interfered with the activity of cefamandole and ticarcillin.