Rapid Detection of Acquired and Inducible Clarithromycin Resistance in Mycobacterium abscessus Group by a Simple Real-Time PCR Assay

Abstract

By targeting the erm(41) and rrl genes in the Mycobacterium abscessus group, a multiplex real-time PCR assay for clarithromycin resistance showed 95% (38/40) concordance with nucleic acid testing and 95% (37/39) concordance with phenotypic testing. This assay provides a simple and rapid alternative to extended incubation or erm(41) sequencing.

TEXT

Mycobacterium abscessus group, which consists of M. abscessus, M. bolletii, and M. massiliense (1, 2), is a rapidly growing mycobacterium that causes significant and intractable infections, particularly in the skin, soft tissue, lungs, and blood (3). M. abscessus group infections are resistant to multiple antibiotics and require lengthy, sometimes lifelong, multidrug therapy (3, 4). Mutations in the peptidyltransferase-binding region of the 23S rRNA gene (rrl), in position 2058 or 2059, have been reported to confer acquired resistance to clarithromycin in the M. abscessus group (5). “Acquired” refers to spontaneous mutations that are selected under antibiotic pressure and do not involve horizontal transmission through genetic elements such as plasmids. Resistance via this mechanism can be detected by PCR, sequencing, or a 3-day conventional phenotypic antibiotic susceptibility testing method, such as broth dilution or Etest, although Etest is not currently a CLSI-recommended method (4, 5).

More recently, the presence of a functional erm(41) gene, where there is a thymine in position 28 (T28), was reported to confer inducible resistance to clarithromycin in the M. abscessus group through methylation of the adenine at position 2058 in the 23S rRNA (6, 7). Inducible resistance requires up to 14 days of incubation of isolates undergoing antibiotic susceptibility testing by phenotypic methods. These isolates may be misidentified as susceptible by a conventional 3-day incubation test (8, 9). Isolates with a nonfunctional erm(41) gene, where there is a cytosine in position 28 (C28), are susceptible to clarithromycin. Additionally, most isolates of M. massiliense were shown to have a truncated erm(41) gene, thus rendering it nonfunctional and the isolates susceptible to clarithromycin (6). Since species identification cannot predict the presence of a functional erm(41) gene, species identification by rapid molecular assays is insufficient for detecting inducible resistance (7).

An SYBR green-based real-time PCR assay, consisting of two duplex PCRs, with primers specific for mutated rrl, in position 2058 or 2059, or wild-type rrl and erm(41)-C28 or erm(41)-T28 was developed on a Rotorgene 6000 (Qiagen) for rapid identification of acquired and inducible resistance in the M. abscessus group. PCR primers and targets are described in Table 1. The first reaction is designed to detect mutated rrl and erm(41)-C28, while the second reaction detects the opposite, i.e., wild-type rrl and erm(41)-T28. A truncated erm(41) gene would show no amplification of erm(41) in either of the reactions. Examples of the results for isolates with mutated and wild-type rrl genes and functional and nonfunctional erm(41) genes are shown in Fig. 1. For DNA extraction, a sweep of multiple colonies on 5% sheep blood agar was suspended in sterile water to 0.5 McFarland density, boiled for 10 min, and then diluted 1:10.

TABLE 1.

PCR primers and targets

Primer mix	Primer namea	Target	Sequence	Product size (bp)	Product Tmb (°C)
Tube 1	erm(41)-C28-FWD-2	erm(41)-C28	GTCGCGACGCCAGC	197	86–88
	erm(41)-REV-4		GCCACCGGAAGGCGAG
	rrl-FWD-1	Mutated rrl	CAGTAAACGGCGGTGGTAAC	179	79–81
	rrl-2058mut-REV-3		AGGTCCCGGGGTCTTV
	rrl-2059mut-REV-3		AGGTCCCGGGGTCTV
Tube 2	erm(41)-T28-FWD-2	erm(41)-T28	GTCGCGACGCCAGT	197	87–89
	erm(41)-REV-4		GCCACCGGAAGGCGAG
	rrl FWD-1	Wild-type rrl	CAGTAAACGGCGGTGGTAAC	179	79–82
	rrl 2058-9wt-REV-1		AGGTCCCGGGGTCTTT

^aFWD, forward; REV, reverse.

^bT_m, melting temperature.

FIG 1.

Melting curve analysis of amplicons obtained by multiplex, real-time PCR allows the detection of acquired and inducible clarithromycin resistance in the M. abscessus group. Representative reaction 1 melting curves for isolates with wild-type rrl and nonfunctional erm(41)-C28 (A) and mutated rrl and truncated erm(41) (B) and representative reaction 2 melting curves for isolates with wild-type rrl and nonfunctional erm(41)-C28 (C) and wild-type rrl and functional erm(41)-T28 (D). A melting curve was generated by ramping from 60°C to 95°C and continuously measuring the fluorescence. Peaks corresponding to specific products are labeled. Fluorescence thresholds are shown with a dashed line.

The assay was first validated using 4 control isolates of the M. abscessus group with known erm(41) and rrl sequences obtained from ATCC (M. abscessus 19977) and ARUP Laboratories, along with 2 characterized M. chelonae isolates that lacked erm(41) and had wild-type rrl and 8 characterized M. avium isolates that lacked erm(41) and had either mutated or wild-type rrl obtained from the Stanford Health Care clinical microbiology laboratory strain collection. After initial validation of the control isolates, the assay was tested on 40 isolates of the M. abscessus group, consisting of 32 archived isolates from 26 patients seen at Stanford Health Care and 8 isolates from the University of Texas Health Science Center at Tyler Mycobacteria/Nocardia Laboratory. The results of the assay were compared to genotypic and phenotypic reference methods. The erm(41) sequencing was done on all isolates with a nontruncated erm(41) gene. PCR primers CTGACGGCACATCTGGTTG and ATCACCAGCACCGACGAAT were used to distinguish between full-length and truncated erm(41) genes and for sequencing. The rrl sequencing was done on isolates with mutated PCR results and to resolve discrepancies between PCR and phenotypic results. PCR primers rrl forward 1 and ACTACCCRCCAGGCACTGTC were used for rrl PCR and sequencing. Phenotypic testing consisted of a 3-day Etest (bioMérieux, Inc., Durham, NC) or a 14-day broth MIC test with a Sensititre Rapmyco panel (Thermo Fisher Scientific, Inc.) for all isolates that were still susceptible after the 3-day Etest. All discrepant isolates were sent to the University of Texas Health Science Center at Tyler Mycobacteria/Nocardia Laboratory for confirmatory phenotypic testing.

Of the 40 isolates tested, real-time PCR identified 18 (45%) with erm(41)-T28, 9 (22.5%) with truncated erm(41), 10 (25%) with erm(41)-C28, and 3 (7.5%) with mutated rrl. Of the 18 isolates with erm(41)-T28, 5 had a concurrent erm(41)-C28 and 1 had a concurrent rrl mutation. The real-time PCR results were 100% in agreement with sequencing results for 28 isolates with erm(41)-C28 and erm(41)-T28 and with full-length PCR results for 12 isolates with truncated erm(41). Sequencing confirmed that the 4 mutant rrl isolates had a mutation at position 2058 or 2059.

PCR demonstrated concordant results with phenotypic testing in 37 (92.5%) of 40 isolates. All three discordant isolates were susceptible to clarithromycin based on real-time PCR results but resistant by phenotypic testing. One isolate was found to be a mixture of two different strains and was therefore removed from the validation set. The remaining two discordant results belonged to a patient with cystic fibrosis and a long history of M. abscessus group pulmonary infection. In 2009 and 2010, clarithromycin-susceptible isolates were recovered from this patient. In February and May 2013, clarithromycin-resistant isolates, with MICs of 8 μg/ml and >256 μg/ml, respectively, were recovered. Sequence analysis of the rrl gene in these isolates identified a mutation at position 2057 in the latter isolate. As the PCR primers in our assay only detect mutations at positions 2058 and 2059, the mutation in the latter isolate was not detected. If we exclude the mixed isolate from the analysis, the PCR susceptibility results were in agreement with the phenotypic results in 37 (95%) of 39 isolates.

To our knowledge, this is the first published report of M. abscessus group isolates with a mixed erm(41)-C28 and erm(41)-T28 subpopulation. During the prevalidation phase of the study, 1 patient isolate demonstrated phenotypic clarithromycin resistance only after an extended 14-day incubation, suggesting the presence of a functional erm(41)-T28 gene. However, repeated real-time PCR detected a nonfunctional erm(41) gene with C28. Upon further investigation, it was found that PCR had been done on DNA extracted from a single colony of the mycobacterium. When DNA was extracted from a sweep of multiple colonies, a mixed population of erm(41)-T28 and erm(41)-C28 was identified. Single-plex PCR for erm(41)-T28 and erm(41)-C28 was done to confirm the presence of a mixed population. Although the presence of mixed subpopulations of mutated and wild-type organisms has been reported with rrl in M. abscessus (6), mixed subpopulations of functional and nonfunctional erm(41) genes have not been reported previously. During the validation phase, we found 5 (12.5%) isolates from 4 patients with mixed erm(41)-C28 and erm(41)-T28 populations. These findings highlight the importance of testing multiple colonies, whether performing genotypic or phenotypic testing.

Prior studies described the application of conventional and probe-based real-time PCR for detection of functional erm(41) genes (7, 10). The SYBR green-based real-time PCR assay described here provides a simpler alternative for detecting clarithromycin resistance in the M. abscessus group. Although doing phenotypic testing with an extended 14-day incubation can also detect inducible resistance, PCR is less labor-intensive and can provide results a few hours after a culture is isolated from a patient (7). PCR should even be able to detect the functional erm(41) gene directly in the clinical sample. Importantly, a representative portion of the sample (testing a sweep of several colonies if using solid agar) should be used to avoid missing an isolate with a mixed population of resistant and susceptible organisms. The rrl primers used in this study are also suited for rapid detection of acquired clarithromycin resistance in M. avium complex isolates. Although we showed that erm(41) and rrl accounted for the vast majority of clarithromycin resistance in the M. abscessus group isolates in our study, other mechanisms for resistance can also be present and would not be detected by this assay. Future research may help determine if patient outcomes can be improved with more rapid and accurate detection of clarithromycin resistance in this difficult-to-treat pathogen.