Intrinsic Macrolide Resistance of the Mycobacterium tuberculosis Complex Is Inducible

Abstract

Mycobacterium tuberculosis is intrinsically resistant to macrolides, a characteristic associated with expression of the erm(37) gene. This intrinsic resistance was found to be inducible with clarithromycin and the ketolide HMR3004. Furthermore, underlying the phenotypic induction was an increase in erm(37) mRNA levels.

The Mycobacterium tuberculosis complex (MTC) is intrinsically resistant to macrolides, such as clarithromycin (4, 5, 12, 14), and studies from this laboratory (8) demonstrated the inducibility of this phenotype in Mycobacterium microti (a member of the MTC). Subsequent reports associated the macrolide resistance of the MTC with the erm(37) gene (1, 6) and suggested that expression of this gene was inducible (6, 7) and may be regulated by the whiB7 gene (7). However, it was unclear whether regulation of erm(37) expression in M. tuberculosis conferred an inducible phenotype. Thus, the primary objectives of this study were to determine if the macrolide resistance of M. tuberculosis was induced by macrolides and ketolides and to analyze the underlying kinetics of erm(37) expression.

(This study was presented in part at the 104th General Meeting of the American Society for Microbiology, New Orleans, La., 23 to 27 May 2004 [abstr. U-019].)

For this study, the experimental organisms were M. tuberculosis strain H37Ra (ATCC 25177) and M. microti strain ATCC 19422; the macrolide susceptibilities of these organisms were considered equivalent to those of virulent M. tuberculosis and Mycobacterium bovis, respectively (1, 2). To assess phenotypic induction, M. tuberculosis and M. microti were cultured to mid-exponential growth phase in Middlebrook 7H9 broth (supplemented with 0.05% Tween 80, 1 g/liter digested casein, and 10% oleic acid-albumin-dextrose-catalase), and then the cultures were split and incubated in a range of clarithromycin concentrations (up to 32 μg/ml). The clarithromycin MIC was determined for each suspension by a broth microdilution assay based on CLSI (formerly NCCLS) guidelines (11).

In a preliminary experiment, preincubation of M. tuberculosis with clarithromycin at 2 and 8 μg/ml for 4, 8, 18, and 48 h indicated that ≤8 h of preincubation did not change the clarithromycin MIC (16 μg/ml), whereas the clarithromycin MIC was 128 μg/ml for organisms preincubated for 18 and 48 h. Thus, a preincubation of 18 to 24 h was used in subsequent experiments.

Table 1 shows the effect on the clarithromycin MIC of using an extended range of preincubation clarithromycin concentrations for M. tuberculosis and M. microti. Consistent with previous reports (1), the noninduced clarithromycin MIC for M. tuberculosis (16 μg/ml) was higher than the noninduced MIC for M. microti (2 μg/ml). However, preincubation of M. tuberculosis in ≤8 μg clarithromycin per ml increased the clarithromycin MIC to 64 to 128 μg/ml. For M. microti, preincubation in ≤0.5 μg clarithromycin per ml increased the MIC to 16 μg/ml. Thus, exposure to subinhibitory clarithromycin concentrations (i.e., below the noninduced MIC) caused four- to eightfold increases in clarithromycin MICs for both M. tuberculosis and M. microti.

TABLE 1.

Effect of preincubation with clarithromycin or the ketolide HMR3004 on susceptibilities of M. tuberculosis H37Ra and M. microti ATCC 19422

Expt and organism	Preincubation conditionsa		MIC (μg/ml)
	Drug	Amt (μg/ml)	Clarithromycin	HMR 3004	Azithromycin
Expt 1
M. tuberculosis	None		16	NDb	ND
	Clarithromycin	0.125	64	ND	ND
	Clarithromycin	0.5	128	ND	ND
	Clarithromycin	2	128	ND	ND
	Clarithromycin	8	64	ND	ND
	Clarithromycin	32	16	ND	ND
M. microti	None		2	ND	ND
	Clarithromycin	0.125	16	ND	ND
	Clarithromycin	0.5	16	ND	ND
	Clarithromycin	2	2	ND	ND
	Clarithromycin	8	1	ND	ND
	Clarithromycin	32	1	ND	ND
Expt 2
M. tuberculosis	None		16	16	128
	Clarithromycin	2	64	64	512
	HMR3004	2	64	64	512
M. microti	None		2	ND	32
	Clarithromycin	0.1	16	ND	128
	HMR3004	0.1	16	ND	128

^aOrganisms were incubated for 18 to 20 h in medium alone or in medium containing clarithromycin or the ketolide HMR3004.

^bND, not determined.

Further studies demonstrated that preincubation with either clarithromycin or the ketolide HMR3004 increased resistance to azithromycin and HMR3004, as well as to clarithromycin (Table 1); HMR3004 was used because it was considered one of the more active ketolides against mycobacteria (3, 12) and also only a weak inducer of erm genes in other bacteria (13). In contrast, preincubation of M. bovis BCG-Pasteur [erm(37) negative] with subinhibitory concentrations of clarithromycin or HMR3004 did not significantly change the clarithromycin MIC of 0.05 to 0.1 μg/ml. These findings indicated that the erm(37)-positive members of the MTC expressed inducible macrolide resistance, although the phenotype shift (four- to eightfold) was smaller than that observed for the rapidly growing mycobacteria (RGM) (8-10).

Ketolides were developed to overcome resistance conferred by erm genes, partly by not inducing these genes; therefore, it is interesting that the ketolide HMR3004 is able to induce resistance in the MTC and in the RGM (K. A. Nash, unpublished data). The inducing activity of ketolides for mycobacteria may reflect a novel regulatory mechanism, such as the involvement of the whiB7 gene as proposed by Morris et al. (7).

To investigate how clarithromycin affects erm(37) mRNA levels, real-time reverse transcription (RT)-PCR analysis was applied using methods described elsewhere (10), with the erm(37)-specific primers ERMMT-1 (TGTCCTCCGCGAGCGATTCC) and ERMMT-2 (AGGCCGACGGTCAGGGTGAA). Each sample was normalized to the level of 23S rRNA assessed by real-time RT-PCR (primers MS23-1 and MS23-3, detailed elsewhere [10]), using algorithms outlined by Vandesompele et al. (15). Initially, erm(37) RNA levels were analyzed in M. tuberculosis H37Ra organisms that had been incubated for 24 h in clarithromycin concentrations ranging from 1 to 32 μg/ml. This experiment demonstrated that erm(37) RNA levels were dependent on the drug concentration with a peak expression at 2 μg clarithromycin per ml (Fig. 1A). Furthermore, these results suggested that erm(37) induction occurred at clarithromycin concentrations at and above the noninduced MIC of 16 μg/ml.

FIG. 1.

A. The level of erm(37) mRNA in M. tuberculosis 24 h after the addition of clarithromycin (CLR) (1, 2, 4, 8, 16, and 32 μg/ml). Relative expression represents mRNA levels from each experimental condition compared to those of organisms incubated in medium alone (0 μg/ml). B. The change in erm(37) mRNA levels in M. tuberculosis following addition of clarithromycin (2 μg/ml) relative to the mRNA levels at time zero. Inset—change in erm(37) mRNA levels (relative to time zero) of M. microti (micr) and M. tuberculosis (tb) incubated for 18 h with clarithromycin at 0.1 and 2 μg/ml, respectively. Each experimental condition was tested in triplicate.

To analyze the kinetics of erm(37) induction, mRNA levels were assessed up to 30 h after the addition of 2 μg clarithromycin per ml (Fig. 1B). The initial increase in mRNA levels was linear; the correlation coefficient was 0.998 for time points between 1 and 9 h. After 9 h, the level of mRNA remained relatively constant at approximately 40-fold above that at time zero. Thus, phenotype induction (>8 h of preincubation) appeared to correlate with the plateau phase of erm(37) expression. A similar pattern was observed with the erm genes of the RGM, although the time course was faster, reaching the plateau phase within 1.5 to 2 h (10).

Like M. tuberculosis, erm(37) mRNA levels increased in M. microti following incubation in subinhibitory concentrations of clarithromycin (Fig. 1B). Interestingly, the erm(37) mRNA levels in noninduced M. microti were 80% of that for noninduced M. tuberculosis. Furthermore, the sequences of the erm(37) gene and surrounding DNA of the MTC (including M. microti) were found to be ≥99% identical (using the NCBI website http://www.ncbi.nlm.nih.gov/genomes/lproks.cgi). These results suggested that, in addition to erm(37), other factors must be involved in the macrolide resistance of the MTC in order to explain the distinct phenotypes of M. tuberculosis and M. microti.

Although the induced M. tuberculosis phenotype probably represents clinically significant resistance, the M. microti phenotype may be in the therapeutic range for clarithromycin. Thus, if M. microti is representative of M. bovis, then clarithromycin may have utility against tuberculosis caused by the latter. Furthermore, a study by Falzari et al. (3) indicated that several new macrolide derivatives have low MICs for M. tuberculosis, although the erm(37)-inducing ability of these agents is unknown. Thus, we believe that further study of macrolides and ketolides as potential agents for treatment of tuberculosis is warranted.