INTRODUCTION

Top Abstract Introduction Materials and Methods References

Infection with gram-positive bacteria is increasingly important for two main reasons. First, the incidence of infection with gram-positive bacteria has increased over the last decade (4, 16, 17, 20). Second, bacteria such as staphylococci, enterococci, and Streptococcus pneumoniae have developed resistance to conventional antibiotics that have been used for decades. New antimicrobial agents with broader ranges of activity and/or greater potency against these organisms are needed. HMR 3647 (RU-66647) and HMR 3004 (RU-64004) are new antibiotics belonging to a novel class of antibiotics, the ketolides. The parent compound of the ketolides is erythromycin A, which is a 14-ring macrolide. The ketolides possess a 3-keto group in the macrolactone ring instead of the L-cladinose moiety of erythromycin A (1). HMR 3647 has a carbamate group linked to an imidazolium and pyridinium nucleus at C_11-12 (5). HMR 3004 (RU-64004) has a quinoline side chain linked to the 11-12 position of the 3-keto-6-methoxy erythromycin A skeleton (RU-006) (2). The mechanism of action of HMR 3647 and HMR 3004 appears to be similar to that of erythromycin A in that the drugs bind to the 50S ribosomal subunit leading to inhibition of protein synthesis in bacteria (6, 7). HMR 3647 has been shown to have antimicrobial activity similar to clindamycin (7). Previously published data indicate that HMR 3647 is more potent than other macrolide-lincosamide-streptogramin B (MLS_B) compounds against vancomycin-resistant enterococci and other gram-positive bacteria, but no direct comparison of activities of the two agents against enterococci was made (5, 7). Information regarding kill kinetics of HMR 3647 against enterococci and staphylococci is also limited (3).

In this study, we compared the in vitro activities of HMR 3647 and HMR 3004 with those of erythromycin, clindamycin, and quinupristin-dalfopristin against gram-positive bacteria with different susceptibility profiles, for instance, gentamicin-resistant enterococci, beta-lactamase-producing enterococci, and vancomycin-resistant enterococci. Time-kill studies were performed to determine if these compounds had bactericidal activities against methicillin-susceptible Staphylococcus aureus (MSSA), methicillin-resistant S. aureus (MRSA), and vancomycin-susceptible and vancomycin-resistant enterococci.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods References

Antibiotics. HMR 3647 and HMR 3004 were provided by Roussel UCLAF, Romainville, France; clindamycin was obtained from Upjohn Laboratories, Kalamazoo, Mich.; erythromycin was obtained from Sigma, St. Louis, Mo.; and quinupristin-dalfopristin was obtained from Rhône-Poulenc Rorer, Antony Cedex, France. Reconstitution, storage, and dilution of these antibiotics were performed according to the manufacturer's recommendations.

Bacterial strains. Two hundred eighty isolates of gram-positive bacteria, including MSSA, MRSA, Streptococcus species, Enterococcus species, Rhodococcus species, Pediococcus species, and Leuconostoc species, were tested. These bacteria either were collected from clinical specimens of various sources or were community fecal isolates. They were collected from the United States, Chile, Argentina, Belgium, and Thailand between 1980 and 1997; duplicate isolates from single patients were excluded, and many of the enterococci had been previously classified as distinct strains by pulsed-field gel electrophoresis or PCR (10, 12, 19). The streptococcal isolates included group A streptococci, group B streptococci, alpha-hemolytic streptococci, and S. pneumoniae. Some of the group A streptococci and S. pneumoniae were provided by Kenneth V. I. Rolston, Anderson Cancer Center, Houston, Tex. These bacteria were chosen based on varied antibiotic susceptibility profiles. The enterococcal isolates consisted of beta-lactamase-producing strains, isolates with high-level resistance to gentamicin, and vancomycin-resistant (Vm^r) strains. The genotypes of 31 Enterococcus faecium isolates in the Vm^r group were vanA, and 13 were vanB. Among Vm^r Enterococcus faecalis isolates, four were vanA genotype and nine were vanB genotype.

Susceptibility testing. MICs were determined by the agar dilution method according to National Committee for Clinical Laboratory Standards (NCCLS) guidelines for antimicrobial susceptibility testing (15). The agar was prepared with Mueller-Hinton (MH) agar II (Becton Dickinson and Company, Cockeysville, Md.). Escherichia coli ATCC 25922, E. faecalis ATCC 29212, Pseudomonas aeruginosa ATCC 27853, S. aureus ATCC 29213, and S. pneumoniae ATCC 49619 were used as control strains. The susceptibility breakpoints of erythromycin were defined according to NCCLS guidelines; erythromycin MICs of 8 µg/ml were considered resistant, and MICs of 0.5 µg/ml were considered susceptible for nonstreptococcal species. For Streptococcus spp., erythromycin MICs of 1 µg/ml and 0.25 µg/ml were considered resistant and susceptible, respectively (14). Isolates for which MICs were between these two values were grouped as resistant isolates for analysis of ketolide MICs according to erythromycin susceptibility. Time-kill studies were performed according to the methods described by Moellering et al. and NCCLS guidelines (11, 13). The experiments were performed with inocula of 10⁵ CFU/ml and a concentration of two to four times the MIC of an antibiotic to determine if there was any bactericidal activity of HMR 3647 and HMR 3004 compared to other antibiotics. Twenty-five microliters of bacterium-antibiotic mixture was taken from the flask for culture on a brain heart infusion (BHI) agar plate at 0, 4, and 24 h in duplicate. Antibiotic carryover effect was removed by centrifuging 1-ml aliquots of cells from the flask at 24 h and washing once with 0.9% NaCl. Serial 10-fold dilutions were made, and 25 µl of each dilution was put on a BHI agar plate in duplicate. A decrease in CFU by 3 log₁₀ CFU/ml was defined as the cutoff point for bactericidal effects.

	RESULTS AND DISCUSSION

The MICs at which 50% of isolates are inhibited (MIC₅₀s), MIC₉₀s, ranges of MICs, and percentages of isolates inhibited by each antibiotic at a concentration of 0.5 µg/ml are shown in Table 1. Among the 19 MSSA, two isolates were resistant to erythromycin, with MICs of >512 µg/ml. One of these showed an inducible MLS_B (iMLS_B) phenotype, as determined by placement of an erythromycin disk adjacent to a clindamycin disk on an agar plate (8). Another isolate was resistant to erythromycin and clindamycin, with an erythromycin MIC of >512 µg/ml and a clindamycin MIC of >256 µg/ml (constitutive MLS_B phenotype). The iMLS_B isolate had HMR 3647 and HMR 3004 MICs of 0.03 µg/ml; in contrast, the isolate which was highly resistant to both erythromycin and clindamycin had HMR 3647 and HMR 3004 MICs of >32 and 8 µg/ml, respectively. All erythromycin-susceptible MSSA were inhibited by HMR 3647 at a concentration of 0.03 to 0.06 µg/ml and by HMR 3004 at a concentration of 0.03 to 0.12 µg/ml, which were lower than the MICs of erythromycin by 8- to 16-fold. Both ketolides were also more potent than clindamycin and quinupristin-dalfopristin against MSSA, showing MICs lower than those of clindamycin and quinupristin-dalfopristin by 2- to 4-fold and 8- to 32-fold, respectively. MICs of HMR 3647 for MSSA were in the same range as MICs of HMR 3004. Time-kill studies were performed against three isolates of MSSA for HMR 3647 and four isolates for HMR 3004. One of these isolates was the iMLS_B strain, and three were erythromycin susceptible. Although the macrolides are considered bacteriostatic agents, one isolate was very susceptible to HMR 3647, showing a decrease of CFU by 3.83 log₁₀ CFU/ml at 24 h. This isolate was also killed by HMR 3004, having a decrease in CFU of 3.37 log₁₀ CFU/ml at 24 h. We obtained the same result in a repeat experiment. Another isolate showed a decrease of 3 log₁₀ CFU/ml when tested against HMR 3004, while a decrease of 2.54 log₁₀ CFU/ml with HMR 3647 was seen. Both ketolides showed bacteriostatic effects against the remaining isolates (Table 2 and Fig. 1). In general, the ketolides had similar activities against MSSA and were superior to the compared drugs. All MRSA in our study were highly resistant to erythromycin and clindamycin and are likely constitutive MLS_B producers, as MICs for each isolate were >512 µg/ml and >256 µg/ml, respectively. They were not inhibited by either ketolide at a concentration of 64 µg/ml.

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Time-kill curves of HMR 3647 against S. aureus isolates. Concentrations of HMR 3647 for the time-kill of MSSA JT9, MSSA JT22, and MSSA JT10 (4× MICs) were 0.25, 0.12, and 0.25 µg/ml, respectively.

Because we did not see a bactericidal effect of nafcillin, a potent antistaphylococcal antibiotic, against any of the MSSA tested, the MICs of nafcillin for those isolates were determined by broth microdilution in MH broth according to NCCLS guidelines (15). The MICs in BHI broth were determined simultaneously to simulate the conditions in time-kill studies. For MSSA JT9, the MIC of nafcillin in BHI broth was 0.12 µg/ml. For the remaining isolates, the MIC was 0.25 µg/ml. In MH broth, the MIC for all isolates was 0.25 µg/ml. Then, another time-kill experiment was performed in duplicate, with nafcillin (1 µg/ml) in BHI broth. The changes in colony counts from baseline determined at 24 h of incubation were 0.5 to 1.0 CFU/ml, which were in the same range as the original experiment.

Of 82 vancomycin-susceptible (Vm^s) E. faecalis isolates, 17 were susceptible to erythromycin, 14 had erythromycin MICs in the intermediate range (1 to 4 µg/ml), and 51 were resistant to erythromycin. All erythromycin-susceptible isolates were inhibited by the ketolides at concentrations of 0.007 to 0.03 µg/ml, which were 4- to 64-fold lower than the MICs of erythromycin. All 14 isolates with intermediate susceptibility to erythromycin had HMR 3647 MICs of 0.015 µg/ml and HMR 3004 MICs of 0.007 to 0.06 µg/ml, except for one isolate which showed an HMR 3004 MIC of 2 µg/ml. For erythromycin-resistant (Em^r) isolates, erythromycin MICs were in the range of 128 to >512 µg/ml, while HMR 3647 MICs ranged from 0.03 to 8 µg/ml and HMR 3004 MICs ranged from 0.03 to 64 µg/ml. The MIC₅₀ and MIC₉₀ of HMR 3647 for Em^r isolates were 2 and 8 µg/ml, respectively; the MIC₅₀ and MIC₉₀ of HMR 3004 were 4 and 64 µg/ml, respectively. MICs of the two ketolides for beta-lactamase-producing and gentamicin-resistant E. faecalis showed the same trend as for isolates that do not have these properties.

One of 13 Vm^r E. faecalis isolates was susceptible to erythromycin, which showed a MIC of 0.25 µg/ml; this isolate had an HMR 3647 MIC of 0.015 µg/ml and an HMR 3004 MIC of 0.03 µg/ml. Another isolate had an erythromycin MIC of 1 µg/ml; for this isolate, MICs of both ketolides were 0.015 µg/ml. The remaining 11 isolates were erythromycin resistant, with MICs of erythromycin ranging from 256 to >512 µg/ml; MICs of both ketolides ranged from 2 to 16 µg/ml. The relationship of ketolide and erythromycin MICs for all E. faecalis isolates is shown in Fig. 2.

View larger version (10K): [in this window] [in a new window]   FIG. 2.   MICs of HMR 3647 for E. faecalis in relation to erythromycin (Em) susceptibility. The MICs for erythromycin-susceptible isolates were 0.5 µg/ml, those for erythromycin-intermediate isolates were 1 to 4 µg/ml, and those for erythromycin-resistant isolates were 128 to >512 µg/ml. For no isolates were erythromycin MICs 8 to 64 µg/ml.

A similar trend was found when the MICs of ketolides against E. faecium were compared to those of erythromycin. Overall, 69 isolates of E. faecium (25 Vm^s, 44 Vm^r) were tested. None of the Vm^r E. faecium isolates were susceptible to erythromycin. Three Vm^s isolates and six Vm^r isolates had MICs of erythromycin in the intermediate range (1 to 4 µg/ml). MICs of HMR 3647 against these isolates were 0.015 to 0.12 µg/ml, with only one isolate showing a MIC of 0.12 µg/ml. MICs of HMR 3004 were 0.007 to 0.5 µg/ml, with most isolates showing MICs of 0.015 µg/ml. Three isolates of Vm^s and one isolate of Vm^r E. faecium with moderate-level resistance to erythromycin (erythromycin MICs, 8 to 16 µg/ml) were inhibited by both ketolides at the concentrations 0.015 to 0.03 µg/ml. Two erythromycin-resistant (MICs, 16 µg/ml) Vm^s isolates were susceptible to clindamycin, and five Vm^r isolates showed intermediate erythromycin MICs but were susceptible to clindamycin, suggestive of the iMLS_B phenotype. These isolates were all inhibited by ketolides at concentrations of 0.015 to 0.06 µg/ml. Among the 25 Vm^s E. faecium isolates, 40% were inhibited by both ketolides at a concentration of 0.03 µg/ml. MICs of HMR 3647 for the two isolates of Vm^s E. faecium that were susceptible to erythromycin were 0.015 and 0.03 µg/ml, lower than the erythromycin MIC by 4- to 32-fold; MICs of HMR 3004 for these two isolates were 0.015 µg/ml, lower than the erythromycin MICs by 8- to 32-fold. Subgroup analysis of both Vm^r and Vm^s E. faecium isolates revealed a pattern similar to that seen with E. faecalis. That is, isolates which showed erythromycin MICs of 512 µg/ml had ketolide MICs ranging from 1 to 64 µg/ml, with a MIC₉₀ of 16 µg/ml for both ketolides. On the other hand, isolates which had erythromycin MICs of 16 µg/ml had ketolide MICs ranging from 0.007 to 0.5 µg/ml.

Among the 13 Vm^r E. faecalis isolates, 4 vanA and 9 vanB isolates showed MICs of HMR 3647 in the range of 0.015 to 8 and 0.015 to 16 µg/ml, respectively. MICs of HMR 3004 for vanA and vanB E. faecalis isolates were the same as those of HMR 3647. Among the 44 Vm^r E. faecium isolates, 31 vanA and 13 vanB isolates showed a MIC₅₀ and MIC₉₀ of HMR 3647 of 8 and 16 µg/ml (range, 0.015 to 64 µg/ml, with only one isolate of the vanB genotype showing an HMR 3647 MIC of 64 µg/ml). The MICs of HMR 3004 for vanA and vanB E. faecium isolates were approximately the same as those of HMR 3647.

Time-kill studies of HMR 3647 against one erythromycin-susceptible and one erythromycin-resistant E. faecalis isolate revealed a minimal to modest decrease in CFU at 24 h (1.99 and 0.2 log₁₀ CFU/ml [Table 2]). Similar results were observed with HMR 3004, with changes in CFU of 0.05 and 0.30 log₁₀ at 24 h. Against one Vm^r E. faecium isolate, both compounds also showed a bacteriostatic effect. There was a decrease of 3.3 log₁₀ CFU/ml for this isolate when tested with quinupristin-dalfopristin. No time-kill study of quinupristin-dalfopristin against E. faecalis was performed, because this species is usually resistant to quinupristin-dalfopristin (9).

One of the 16 group A streptococci was resistant to ketolides, erythromycin, and clindamycin but susceptible to quinupristin-dalfopristin, with a MIC of quinupristin-dalfopristin of 0.25 µg/ml. MICs of HMR 3647, HMR 3004, erythromycin, and clindamycin for this isolate were 64, >64, >512, and >256 µg/ml, respectively, indicative of the constitutive MLS_B phenotype. Almost all (15 of 16, 94%) were susceptible to ketolides, erythromycin, clindamycin, and quinupristin-dalfopristin. For the susceptible isolates, MICs of HMR 3647 were 0.007 to 0.12 µg/ml, lower than those of erythromycin by 2- to 8-fold; the MICs of HMR 3004 were 0.003 to 0.25 µg/ml, lower than those of erythromycin by 2- to 16-fold.

All group B streptococci were inhibited by 0.03 µg/ml of ketolides, erythromycin, and clindamycin, but none were inhibited by this concentration of quinupristin-dalfopristin (Table 1). Ketolides were approximately four- to eightfold more potent than erythromycin against group B streptococci.

All S. pneumoniae isolates had ketolide MICs lower than those of erythromycin by 4- to 16-fold. The majority of S. pneumoniae isolates in our study (9 of 14) were susceptible to erythromycin, with MICs of HMR 3647 and HMR 3004 ranging from 0.003 to 0.03 µg/ml. The remaining five isolates had erythromycin MICs of 2 to 8 µg/ml. These five isolates were susceptible to clindamycin (MICs, 0.015 to 0.25 µg/ml) and likely express either an efflux system (15) or the inducible MLS_B phenotype. The MICs of HMR 3647 for these five isolates ranged from 0.003 to 0.5 µg/ml.

HMR 3647 was the most active agent among the antibiotics tested against alpha-hemolytic streptococci. All 10 isolates were inhibited by HMR 3647 at a concentration of 0.5 µg/ml, with a MIC₅₀ of 0.03 µg/ml. With the breakpoint of 0.5 µg/ml, 60, 50, and 80% of isolates were susceptible to HMR 3004, erythromycin, and clindamycin, respectively. The Em^s isolates were all inhibited by HMR 3647 and HMR 3004 at concentrations of 0.015 to 0.03 µg/ml. Five of 10 isolates were resistant to quinupristin-dalfopristin and showed MICs of 4 µg/ml, four showed MICs of 2 µg/ml, and one showed a MIC of 1 µg/ml.

Twenty-one percent of Rhodococcus species in this study were susceptible to erythromycin. These isolates had HMR 3647 MICs lower than those of erythromycin by two- to fourfold. MICs of HMR 3004 were lower than those of erythromycin by four- to eightfold. Erythromycin-resistant isolates had ketolide MICs in the range of 2 to 16 µg/ml. All Leuconostoc spp. were inhibited by HMR 3647 at a concentration of 0.03 µg/ml, while four of seven had HMR 3004 MICs of 2 µg/ml and three of seven had HMR 3004 MICs of 0.015 to 0.03 µg/ml. MICs of HMR 3647 were lower than those of erythromycin by four- to eightfold. HMR 3647 and HMR 3004 showed similar activities against pediococci, diphtheroids, and lactobacilli. Both compounds were more active than erythromycin, showing lower MICs by 4- to 64-fold.

Conclusions. In summary, HMR 3647 and HMR 3004 have similar activities against gram-positive bacteria. Both compounds were more potent than erythromycin against almost all species, except rhodococci, for which these three agents were similar. The MICs of ketolides tended to vary with the MIC of erythromycin. Isolates which showed high erythromycin MICs were more likely to have increased ketolide MICs as well. Strains with the inducible MLS_B phenotype of S. aureus were susceptible to ketolides, while constitutive MLS_B strains were resistant to ketolides. Some erythromycin-resistant enterococci were inhibited by low concentrations of ketolides. The five S. pneumoniae strains with reduced susceptibility to erythromycin were susceptible to ketolides at a concentration of 0.5 µg/ml. Ketolides showed bactericidal activities against some S. aureus isolates, as determined by time-kill studies. Therefore, ketolides might be useful agents in the treatment of gram-positive bacterial infections when the infecting organisms are susceptible to these compounds. Further in vivo studies of these compounds appear warranted.