In Vitro Activity of the New Ketolide HMR3647 in Comparison with Those of Macrolides and Pristinamycins against Enterococcus spp.

Abstract

Ninety-four erythromycin-susceptible and 107 erythromycin-resistant enterococcal strains (MIC of ≥512 μg/ml) were inhibited by the ketolide HMR3647 at MICs of ≤0.007 to 0.06 and 0.03 to 8 μg/ml, respectively. Eighteen vanA-positive isolates and 29 high-level-penicillin-resistant isolates, all of them erythromycin resistant, were inhibited by HMR3647 at an MIC range of 0.015 to 4 μg/ml. The new ketolide has excellent activity against Enterococcus species.

Ketolides are novel semisynthetic 14-membered-ring macrolides. They are characterized by a 3-keto function on the aglycone A instead of an l-cladinose. The lack of l-cladinose allows a greater stability in acidic environments (4). Ketolides have the same spectrum of action as macrolides, but they have better in vitro activity against gram-positive microorganisms (1, 7, 8, 11).

The genus Enterococcus is increasingly being recognized as a nosocomial pathogen; indeed, it is considered the third most common cause of nosocomial infections (5). The treatment of enterococcal infections has always been of concern, due to the intrinsic resistance of the genus to several antibiotics. Moreover, this genus frequently acquires mechanisms of antibiotic resistance to aminoglycosides, β-lactams, glycopeptides, and macrolides. Therefore, any new therapeutic alternative for the treatment of enterococcal infections should be urgently evaluated.

In the present study, the in vitro activity of the ketolide HMR3647 was compared with that of macrolides and pristinamycins on 202 Enterococcus strains with different antibiotic resistance phenotypes.

Bacterial strains.

A total of 202 Enterococcus strains of different species were studied (163 Enterococcus faecalis strains, 32 E. faecium strains, 3 E. gallinarum strains, 2 E. durans strains, and 2 E. hirae strains). Forty-seven of them were selected according to their pattern of resistance to vancomycin and penicillin, including 18 vanA-positive Enterococcus strains (13 E. faecium, 2 E. hirae, 2 E. faecalis, and 1 E. durans strain), characterized by PCR (13), and 29 non-β-lactamase-producing, penicillin-resistant E. faecium strains (MIC of ≥16 μg/ml). The other 155 strains were randomly selected. Therefore, MIC distributions in this work are not intended to reflect the behavior of natural enterococcal strains and are used to compare the relative activities of the different antibiotics.

Susceptibility testing.

MICs were determined by agar dilution in Mueller-Hinton agar medium according to the National Committee for Clinical Laboratory Standards method (10). The antibiotics included were the ketolide HMR3647, erythromycin A, clarithromycin, azithromycin, and roxithromycin, all of them supplied by Hoechst Marion Roussel (Romainville, France); spiramycin (Sigma); and pristinamycin I and II (supplied by Rhône-Poulenc Rorer).

Table 1 shows the MIC data for the erythromycin A susceptibility pattern. A total of 108 of 202 strains studied were erythromycin A resistant, and 94 were erythromycin A susceptible. Erythromycin-susceptible strains (probably lacking specific mechanisms of resistance) were those for which the MIC was ≤4 μg/ml, due to the clear bimodal distribution observed (Fig. 1), even though some of these strains (for which MICs were 1 to 4 μg/ml) are classified “intermediate” by the National Committee for Clinical Laboratory Standards general criteria (10).

TABLE 1.

Resistance of 202 Enterococcus strains to HMR3647, macrolides, and pristinamycins, according to phenotype

Antibiotic and phenotype of isolatesa	MIC			% of MICs
	50%	90%	Range	≥0.5	≥2	≥4	≥8
HMR3647
Eryr	2	4	0.03–8	88.8	56.4
Erys	0.03	0.03	≤0.007–0.06	0	0
VanA	0.5	4	0.125–4	88.8	22.2
Penr	1	4	0.015–4	86.2	41.3
Pens	0.03	4	0.007–8	41.0	28.3
Erythromycin A
Eryr	≥512	≥512	64–≥512				100
Erys	2	4	≤0.06–4				0
VanA	≥512	≥512	≥512				100
Penr	≥512	≥512	0.5–≥512				89.6
Pens	4	≥512	≤0.006–≥512				47.3
Clarithromycin
Eryr	≥512	≥512	32–≥512				100
Erys	1	2	≤0.06–4				0
VanA	≥512	≥512	≥512				100
Penr	≥512	≥512	0.125–≥512				89.6
Pens	2	≥512	≤0.006–≥512				47.3
Roxithromycin
Eryr	≥512	≥512	256–≥512				100
Erys	2	8	0.125–16				20.2
VanA	≥512	≥512	≥512				100
Penr	≥512	≥512	1–≥512				89.6
Pens	8	≥512	0.125–≥512				58.3
Azithromycin
Eryr	≥512	≥512	64–≥512				100
Erys	4	16	0.06–64				39.3
VanA	≥512	≥512	≥512				100
Penr	≥512	≥512	2–≥512				89.6
Pens	16	≥512	0.06–≥512				68.7
Spiramycin
Eryr	≥512	≥512	4–≥512		100
Erys	2	2	0.25–2		72.3
VanA	≥512	≥512	≥512		100
Penr	≥512	≥512	1–≥512		96.5
Pens	2	≥512	0.25–≥512		85.5
Pristinamycin I
Eryr	128	≥256	2–≥256			99.1
Erys	2	4	1–32			26.6
VanA	128	≥256	8–≥256			100
Penr	128	≥256	2–≥256			93.1
Pens	4	≥256	1–≥256			60.7
Pristinamycin II
Eryr	≥256	≥256	≤1–≥256			93.5
Erys	≥256	≥256	1–≥256			98.9
VanA	128	≥256	2–≥256			94.4
Penr	128	≥256	≤1–≥256			75.8
Pens	≥256	≥256	1–≥256			99.4

^aEry^r, erythromycin A resistant (n = 108); Ery^s, erythromycin A susceptible (n = 94); Pen^r, penicillin resistant (n = 29); Pen^s, penicillin susceptible (n = 173). n = 18 for VanA isolates. 

FIG. 1.

MICs of HMR3647 and erythromycin A for 202 Enterococcus strains. Black bars, erythromycin A; stippled bars, HMR3647 against erythromycin A-susceptible strains; white bars, HMR3647 against erythromycin A-resistant strains.

For the 94 erythromycin-susceptible Enterococcus strains, the MICs at which 50 and 90% of the isolates were inhibited (MIC₅₀ and MIC₉₀) of the tested macrolide antibiotics, in micrograms per milliliter, were 2 and 4 for erythromycin, 1 and 2 for clarithromycin, 2 and 8 for roxithromycin, 4 and 16 for azithromycin, and 2 and 2 for spiramycin, respectively. For these strains, the MIC₅₀ and MIC₉₀ of the ketolide HMR3647 were 0.03 and 0.03 μg/ml, respectively. The MIC₅₀ of HMR3647 was 6 to 8 dilutions lower than that of erythromycin A for erythromycin-susceptible and -resistant strains. This illustrates the better intrinsic activity of HMR3647 against enterococcal isolates compared with that of macrolides. Excellent activity of HMR3647 against a few erythromycin-susceptible enterococcal strains was previously observed, with exactly the same MIC₅₀ and MIC₉₀ as in our study (8).

The erythromycin MIC for all but one of the 108 erythromycin-resistant strains tested was ≥512 μg/ml. The MIC₅₀s and MIC₉₀s for C₁₄, C₁₅, and C₁₆ macrolides (erythromycin A, clarithromycin, roxithromycin, azithromycin, and spiramycin) were ≥512 and ≥512 μg/ml in all cases. Interestingly, the MIC₅₀ and MIC₉₀ of the ketolide HMR3647 were only 2 and 4 μg/ml, respectively (MIC range of 0.03 to 8 μg/ml), significantly lower than those of macrolide antibiotics. Similar results were obtained with Enterococcus strains studied by other authors (6, 9); in one study, the MICs obtained were 1 dilution higher than those in our study (12).

Figure 1 shows the MIC distribution for erythromycin A and HMR3647 against 202 Enterococcus strains. For erythromycin, a bimodal distribution was observed. The first distribution peak corresponded to strains for which MICs were in the range of ≤0.06 to 4 μg/ml. The second peak included the erythromycin-resistant strains for which MICs were usually ≥512 μg/ml. The MIC distribution for the ketolide HMR3647 also showed a bimodal shape. The first peak corresponded to the erythromycin-susceptible strains, for which MICs of the ketolide were in the range of ≤0.007 to 0.06 μg/ml, while the second peak corresponded to the erythromycin-resistant strains, for which MICs were 0.06 to 8 μg/ml. Thus, the erythromycin resistance mechanism slightly influences the MIC of the ketolide. The better intrinsic activity of the ketolide against Enterococcus, compared with that of macrolides, does not account for the low MICs observed for erythromycin-resistant isolates, which indicates that this compound is less affected by the major erm-encoded macrolide-lincosamide-streptogramin B mechanism of resistance in this genus. Ketolides do not induce macrolide-lincosamide-streptogramin B resistance against erythromycin A in inducibly resistant bacteria (2, 3).

The MIC₅₀ and MIC₉₀ of pristinamycin I for erythromycin-resistant and erythromycin-susceptible strains were 128 and ≥256 μg/ml and 2 and 4 μg/ml, respectively; the MIC₅₀ and MIC₉₀ of pristinamycin II were both ≥256 μg/ml, irrespective of the erythromycin A susceptibility pattern (Table 1). Most E. faecalis strains (162 of 163) were highly resistant to pristinamycin II (MIC of ≥16 μg/ml); for the only susceptible strain the MIC was 4 μg/ml. The MIC of pristinamycin I was related to the erythromycin A susceptibility or resistance (for 86 of 89 erythromycin-susceptible strains the MIC of pristinamycin I was ≤4 μg/ml, and for 73 of 74 erythromycin-resistant strains it was ≥8 μg/ml). Most E. faecium strains studied were erythromycin resistant (30 of 32), and for 27 and 20 of these 30 strains the MICs of pristinamycin I and pristinamycin II were ≥8 μg/ml.

Table 1 shows the activities of HMR3647 and macrolides against Enterococcus, according to their vancomycin or penicillin pattern of susceptibility. All vancomycin-resistant strains were resistant to macrolide-streptogramin antibiotics. The MIC₅₀ and MIC₉₀ of erythromycin A, clarithromycin, roxithromycin, azithromycin, and spiramycin were ≥512 and ≥512 μg/ml, respectively, and those of both pristinamycin I and pristinamycin II were 128 and ≥256 μg/ml, respectively. The activity of the new ketolide appeared to be better than those of the other drugs against these strains. All vanA strains were inhibited by HMR3647 at concentrations of ≤4 μg/ml, and the MIC₅₀ and MIC₉₀ were 0.5 and 4 μg/ml, respectively, a range identical to that observed with vanA-negative, erythromycin-resistant strains. Jones and Biedenbach (8) reported an MIC₅₀ and MIC₉₀ of the ketolide HMR3647 of 1 and 8 μg/ml, respectively, for vanA strains. Therefore, the susceptibility of vancomycin-resistant enterococci to HMR3647 depends on the phenotype of resistance to erythromycin A, as previously suggested (1). The MIC₅₀s of erythromycin A, clarithromycin, roxithromycin, azithromycin, and spiramycin for the penicillin-susceptible strains were 4, 2, 8, 16, and 2 μg/ml, and the MIC₅₀ of the ketolide HMR3647 was 0.03 μg/ml. For penicillin-resistant strains, the MIC₅₀s of erythromycin A, clarithromycin, roxithromycin, azithromycin, and spiramycin were ≥512 μg/ml and that of the ketolide was 1 μg/ml. As in the case of vancomycin-resistant isolates, this difference was attributable not to the penicillin-resistant phenotype but to the higher frequency of erythromycin resistance among penicillin-resistant strains.

All vanA enterococcal strains and all high-level-penicillin-resistant strains were resistant to erythromycin A (MIC of ≥512 μg/ml); nevertheless, the MIC of the ketolide HMR3647 was always ≤4 μg/ml, with MIC₅₀s of 0.5 and 1 μg/ml in the two groups of strains, respectively. This observation may suggest that this compound may be eventually useful, alone or in combination, in some difficult-to-treat enterococcal infections. The results in animal models should provide some evidence to support such a possibility.