Mycobacterium abscessus infection in cystic fibrosis: molecular typing and clinical outcomes

Mycobacteriumabscessus infection in cystic fibrosis (CF)

Rapidly growing mycobacteria (RGM) are non-tuberculous mycobacteria (NTM) that infrequently cause disease in healthy people and are commonly environmental micro-organisms found in soil, bioaerosols and natural and chlorinated water. They grow in the laboratory on solid media, such as blood or chocolate agar, in 3–7 days. Colonies can have either a rough or smooth morphotype. The most notable RGM species are Mycobacterium abscessus, Mycobacterium chelonae, Mycobacterium fortuitum and Mycobacterium mucogenicum, all of which cause respiratory and soft tissue infections in humans. The source of these infections is still uncertain, but tap water and person-to-person transmission have both been suggested (Saiman & Siegel, 2004; Feazel et al., 2009; Bryant et al., 2013; Thomson etal., 2013).

M. abscessus is a rapidly growing, multidrug-resistant organism that has emerged as a significant pathogen in CF patients. Estimated to have an overall prevalence rate in this patient group of up to 13 % in Europe and 16 % in the USA, M. abscessus accounts for over half of all NTM isolated (CFF, 2012; Seddon etal., 2013). The rising prevalence of infection with M. abscessus in CF patients could be partly attributable to a number of factors, including improved laboratory practice, improved patient survival, increased environmental exposure (Feazel et al., 2009; Thomson etal., 2013), inhaled antibiotic usage (Saiman & Siegel, 2004) and person-to-person spread (Bryant et al., 2013). M. abscessus has been linked to poor clinical outcomes in the CF population, especially following lung transplantation, with many centres not listing infected patients for transplant assessment (Gilljam et al., 2010; Luong et al., 2010; Watkins & Lemonovich, 2012).

M. abscessus was first described as a species in 1992, having been previously considered a subspecies of M. chelonae (Kusunoki & Ezaki, et al. 1992). Mycobacterium bolletii and Mycobacterium massiliense were recognized as further new NTM species in 2006 (Adékambi & Drancourt, 2004; Adékambi etal., 2006; Euzéby, 2006). Collectively these three species are sometimes referred to as the M. abscessus complex. However, there has been some debate in the literature regarding the taxonomy of these organisms, which cannot be differentiated by phenotypic methods. Genotypic studies have led to the proposal that the three taxa should be combined into a single species, and that M. massiliense and M. bolletii should be further combined as a single subspecies (M. abscessus subsp. bolletii), while M. abscessus remains as a separate subspecies (Leao et al., 2009, 2011). However, recent higher-resolution typing, including whole-genome sequencing, appears to show that M. abscessus, M. massiliense and M. bolletii exist as three separate entities and supports the designation of three subspecies (Harris et al., 2012; Bryant et al., 2013). In the context of this review, we shall refer to M. abscessus subsp. abscessus, M. abscessus subsp. massiliense and M. abscessus subsp. bolletii for the sake of clarity, as this is consistent with the majority of the published studies that will be discussed here.

Identification of M. abscessus isolates

Identification of mycobacterial species is usually performed using genotypic methods. Two commercial line probe assays are employed by a number of laboratories, INNO-LiPA MYCOBACTERIA v2 (Innogenetics) and the Genotype Mycobacterium CM kit (Hain Lifescience). Both assays can accurately identify the majority of commonly encountered mycobacterial species, but the Genotype Mycobacterium CM kit may sometimes have difficulty differentiating M. chelonae from M. abscessus. Neither assay can differentiate between the three M. abscessus subspecies (Scarparo et al., 2001). This lack of accuracy in identification is concerning because it may impact on patient management, in particular the decision to list patients for lung transplantation, because M. abscessus has been linked with poorer clinical outcomes than those observed in patients infected with M. chelonae (Chalermskulrat et al., 2006; Broda et al., 2013). More accurate identification can be achieved by performing PCR and sequencing of housekeeping gene targets such as hsp65, rpoB, sodA and secA (Ringuet et al., 1999; Adékambi & Drancourt, 2004; Arnold et al., 2012; Blauwendraat et al., 2012). While most isolates can be identified with two targets, multilocus sequence analysis (MLSA) is required to differentiate the three M. abscessus subspecies reliably, particularly as some isolates give ambiguous results with single or dual gene targets (Macheras et al., 2009). Fig. 1 illustrates this for five example isolates. A number of reports in the literature describe ambiguous sequencing results from some isolates that have chimeric housekeeping gene sequences, presumably as a result of horizontal gene transfer and recombination (Zelazny et al., 2009; Macheras et al., 2011), which further highlights the challenges of accurate identification and differentiation of these subspecies from one another.

Molecular characterization of M. abscessus complex isolates

Further differentiation of M. abscessus isolates has been achieved using a number of different molecular methods, which are summarized below.

Impact of identification and typing on clinical outcomes

M. abscessus has emerged as a significant pathogen in CF patients, linked to disease progression and poor outcomes. Many centres consider M. abscessus infection to be a contraindication for lung transplantation. Conversely, patients infected with M. chelonae are still considered for transplantation (Gilljam et al., 2010; Luong et al., 2010; Watkins & Lemonovich, 2012). Reports in the literature that relate transplant outcome with the infecting mycobacterial species are likely to be confounded by the inherent inaccuracy in identification methods, with many M. abscessus isolates having been incorrectly identified as M. chelonae. Furthermore, unless PCR and sequencing techniques have been used, isolates will have been identified only as ‘M. abscessus’ and not differentiated further. This may turn out to be a very important requirement if the three M. abscessus subspecies are shown in the future to be associated with differing clinical outcomes, particularly following lung transplantation, as recent reports have suggested (Koh et al., 2011; Harada et al., 2012; Robinson etal., 2013). Predicting macrolide susceptibility by sequencing erm(41) and rrl, alongside accurate identification of the three M. abscessus subspecies, would help to inform antimicrobial treatment regimens if performed routinely on M. abscessus isolated from CF patients (Bastian et al., 2011; Choi et al., 2012; Nessar et al., 2012; Leung & Olivier, 2013). M. abscessus subsp. abscessus appears to be linked with worse clinical outcomes than the other subspecies, perhaps due in part to differing macrolide susceptibility (Koh et al., 2011; Harada et al., 2012; Robinson et al., 2013). However, it is important to consider the impact of other factors that may also contribute to clinical outcomes, such as smear positivity and host factors such as CF mutation and immunological status, as these things are also likely to play an important role. Further investigation into how all these factors combine to effect clinical outcomes is now required, and accurate identification of M. abscessus isolates will be critical.

In non-CF patient populations there have been several reports of outbreaks of M. abscessus subsp. massiliense in post-surgical, cutaneous infections (Duarte et al., 2009; Leao et al., 2009; Shang et al., 2011). These outbreaks were all caused by a single strain as determined by PFGE, which was thought to be particularly invasive and pathogenic. Recent studies have suggested that some strains of M. abscessus could be associated with chronic infection in a cohort of CF patients (Harris et al., 2012) and that certain strains may also be more transmissible in this patient group (Bryant et al., 2013). It has also been observed that the same strain or genotype is found in a number of CF patients with no apparent epidemiological links (Harris et al., 2012) and that common strains are seen in patients in remote geographical locations (Davidson et al., 2013). These early studies suggest that particular genotypes or strains of M. abscessus have differing clinical outcomes or are particularly well adapted to humans. The epidemiology of M. abscessus strains in CF and other patient populations warrants further investigation using the molecular methods we have described. However, whole-genome sequencing is the ultimate method for M. abscessus typing and, due to the high level of similarity seen between many M. abscessus strains, this may be required for differentiating closely related strains, particularly when investigating potential cross-transmission events.