Mycobacterium abscessus and Antibiotic Resistance: Same As It Ever Was

(See the Major article by Shaw et al on pages 1678–86.) It seems that no matter how much we learn about Mycobacterium abscessus, we inevitably encounter more impediments to successful treatment, while improved therapeutic outcomes elude us. Mycobacterium abscessus subsp abscessus (hereafter referred to as M. abscessus) confronts us with formidable, innate antibiotic resistance mechanisms that defeat most antibiotics [1, 2]. Further, for the few antibiotics with some activity against M. abscessus, such as the macrolides and amikacin, even if they can overcome innate resistance mechanisms, they are vulnerable to the development of acquired mutational resistance by M. abscessus, usually as the consequence of weak and ineffective companion antibiotics. The Mycobacterium abscessus blues linger, despite our best efforts [3]. In this issue of Clinical Infectious Diseases, Shaw et al [4] shed more light on the complicated processes contributing to our therapeutic frustrations with M. abscessus. The authors report the results of whole-genome sequencing on 32 M. abscessus isolates from multiple body sites in 2 patients with cystic fibrosis who were undergoing anti-mycobacterial treatments. A clustering analysis of single nucleotide variants revealed that each patient harbored multiple M. abscessus subpopulations that were differentially abundant between sputum, lung samples, chest wounds, and pleural fluid. Sputum isolates did not reflect the overall within-patient diversity, and failed to detect subclones with mutations associated with macrolide resistance (rrl 2058/2059). Both patients died as a result of this infection. The authors concluded that M. abscessus isolates from sputum do not always reflect the diversity present within the patient, which can include subclones with differing antimicrobial resistance profiles. Dheda et al [5] recently evaluated patients with cavitary, multidrug-resistant tuberculosis who failed antibiotic therapy with second-line anti-tuberculosis drugs and underwent lung resections. Antibiotic concentrations were measured at multiple biopsy sites within each cavity. The Mycobacterium tuberculosis minimum inhibitory concentrations (MICs) and drug resistance–associated single nucleotide variants differed between the various locations within each cavity and with pretreatment and serial sputum isolates. The pretreatment sputum MICs had an accuracy of only approximately 50% in predicting cavitary MICs. The MIC heterogeneity within cavities was such that at some positions, MICs indicated resistance, not present before treatment, and at some positions within the same cavity, they indicated susceptible isolates. The authors concluded that a drug penetration gradient across the tuberculous cavity wall in multidrug-resistant tuberculosis created variable antibiotic exposures that amplified the acquired drug resistance. Presumably, the relatively weak anti-mycobacterial activity of the second-line anti-tuberculosis drugs contributed to the emergence of acquired resistance as well. Because the first-line anti-tuberculosis drugs, such as isoniazid and rifampin, have very low MICs for M. tuberculosis and have good penetration of most spaces, the risk for acquired drug resistance by this mechanism with drug-susceptible M. tuberculosis is small. These 2 studies share important and instructive features. The tuberculosis patients in the study by Dheda et al [5] were described as medically untreatable, which is a clinical scenario much closer to that of a M. abscessus lung infection than to drug-susceptible tuberculosis. Both studies involved extremely drug-resistant mycobacterial isolates; limited treatment options, with weak anti-mycobacterial agents; and the presence of infecting organisms in niches with poor antibiotic penetration. This combination of factors results in low drug concentrations and creates an environment where the anti-mycobacterial drug combinations are not adequate to prevent or protect against the emergence of resistant mycobacterial isolates. An unfortunate conclusion is that there appears to be a vicious cycle, in which the more difficult the mycobacterial infection is to treat with available antibiotics, the more likely it is that worsening drug resistance will develop during therapy, resulting in an even more treatment-refractory mycobacterial infection. In both studies, mycobacterial isolates cultured from sputum were an unreliable indicator of the emerging organism’s MIC variability. Shaw et al [4] suggest that awareness of the M. abscessus phenotypic variability, with sampling of multiple body sites in conjunction with whole-genome sequencing, may be necessary to achieve an optimal treatment outcome. This suggestion is reasonable for an organism where there are multiple therapeutic options, such as with drug-susceptible tuberculosis. Unfortunately, most patients with serious M. abscessus infections receive as aggressive an antibiotic approach as possible from the onset of therapy, with few available alternative agents for an intensification of the regimen based on worsening in vitro susceptibility profiles and/or clinical deterioration. For M. abscessus, the most important potential casualties of the progressive emergence of drug resistance are the macrolides and amikacin. It is well known that most M. abscessus subsp abscessus isolates are macrolide resistant, due to an inducible macrolide resistance gene, erm(41). However, approximately 20% of M. abscessus subsp abscessus isolates have an inactive erm gene and are, therefore, macrolide susceptible, which significantly improves the chances of a successful treatment outcome with macrolide-based therapy [6–9]. It cannot be overstressed how profoundly disappointing it is for a patient with an inactive erm gene and macrolide-susceptible M. abscessus to subsequently develop acquired mutational macrolide resistance [10]. In the study by Shaw et al [4], the most important factor related to mutational macrolide resistance was macrolide exposure. The 2 reported patients had severe and multifocal diseases, where exposure to the macrolide by the organism was likely inconsistent and suboptimal. More important, however, was the unavoidable use of weak companion anti-microbials, greatly magnifying the risk of selecting M. abscessus isolates with mutational macrolide resistance. This study exposes another perspective on the inadequacy of weak companion medications for protecting against the emergence of M. abscessus isolates with mutational macrolide resistance. Unfortunately, it also appears that amikacin, an equally important drug for treating M. abscessus, was a casualty of the same phenomenon in these patients. An interesting corollary to the above consideration is reflected in the statement, “macrolide use still drives development of high-level macrolide resistance even in the presence of a functional erm (41) gene.” This observation suggests that, even with the functional erm (41) gene, macrolide must still have enough of a therapeutic impact to result in the selection of M. abscessus isolates with mutational macrolide resistance. Although unusual, some patients with macrolide-resistant M. abscessus pulmonary infection due to an active erm gene will respond favorably to macrolide-containing regimens [9]. The finding by Shaw et al [4] suggests that the inclusion of macrolide in M. abscessus treatment regimens for patients whose organism has an active erm gene may legitimately be of some, albeit limited, benefit. For patients with M. abscessus disease, there are few available options on the immediate horizon to remedy these treatment obstacles. Inhaled liposomal amikacin was recently shown to significantly improve sputum conversion rates for patients with refractory Mycobacterium avium complex lung disease [11]. Limited preliminary data suggest similar activity for M. abscessus pulmonary disease, although confirmation requires a larger and more rigorous trial [12, 13]. Even if inhaled liposomal amikacin is effective for pulmonary M. abscessus disease, this form of amikacin would not be of benefit for non-pulmonary disease. The findings in this study reinforce the urgent need for better antimicrobial agents against M. abscessus. Macrolides and amikacin are not adequate to eradicate M. abscessus infections in desperately ill patients, such as those described by Shaw et al [4]. The motley collection of other available antibiotics is simply not adequate to successfully treat patients with severe M. abscessus disease, protect macrolides and amikacin against the emergence of mutational resistance, or, ultimately, to prevent a progression of the infection and clinical decline of the patient. All of us in the non-tuberculous mycobacterial disease community can only hope that anti-mycobacterial drug development will improve with the identification of new agents, as well as with support for testing currently available agents, if even such testing is only remotely promising. Surely, we must be approaching a tipping point where we can utilize our expanding knowledge of the myriad ways that M. abscessus frustrates our therapeutic efforts and use that knowledge to design better treatment strategies [2]. Potential conflicts of interest. Author certifies no potential conflicts of interest. The author has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.