Bacterial reasons for treatment failure in Mycobacterium abscessus

Mycobacterium abscessus (MAB), a gram-positive, non-tuberculous mycobacterium is an emerging pathogen comprising three subspecies; M. a. abscessus, M. a. massiliense, and M. a. bolletii. MAB is predominantly associated with pulmonary infections and leads to accelerated lung damage, imposing a significant burden, particularly on patients with conditions like Cystic Fibrosis and other pre-existing lung conditions. The incidence of MAB infection is on the rise worldwide, often requiring extended periods of multidrug therapy spanning several months to years. Nonetheless, treatment of this pathogen leads to unfavourable clinical outcomes; with treatment failures frequently exceeding 50%.
MAB is intrinsically resistant to most antibiotics including all first line drugs for tuberculosis. Moreover, MAB frequently acquires new resistance, particularly against macrolides and aminoglycosides. However, there exists limited knowledge about the evolutionary trajectory driving mutations and de novo resistance in this pathogen. Classical drug resistance, measured by the minimum inhibitory concentration (MIC), is a poor predictor of treatment outcome in chronic respiratory infections. Therefore, other reasons for treatment failure including host determinants, pharmacokinetics as well as other bacterial mechanisms may contribute to the success of treatment.
This thesis focusses on highlighting and understanding bacterial mechanisms that may contribute to treatment failures in MAB infections. My goal was to identify factors driving the evolution of de novo resistance. An invitro assay was established to assess the trajectory of evolved mutations in MAB. Subsequently, I explored heteroresistance, a phenomenon which depicts a bacterial subpopulation displaying different antibiotic susceptibility levels compared to the dominant population. After evaluating heteroresistance using population analysis profile, the gold standard method, I particularly aimed to establish whether these phenotypes were stable within the population. Finally, a significant portion of this thesis was devoted to exploring drug tolerance in MAB. Drug tolerance is a crucial survival strategy that enables bacterial subpopulations to persist during drug treatment. Drug tolerance is evident in the kinetics of bacterial killing, yet it eludes detection through traditional MIC measures and may correlate with poor treatment outcomes.
I discovered that MAB readily acquired a high-level aminoglycoside resistance through a point mutation (A1408G) in the 16S rRNA gene. This resistance was identified across all three subspecies, with its emergence irrespective of the growth phase. Additionally, I identified heteroresistant phenotypes in certain antibiotics, including imipenem, linezolid, and clarithromycin. Subsequent evaluations showed that the imipenem heteroresistant phenotypes were transient and reverted to features of the dominant population.
In the area of drug tolerance, the traditional method of assessing bacterial killing using colony forming units (CFUs) posed significant challenges due to its highly time-consuming and labour-intensive nature, hindering scalability. Overcoming this hurdle marked a significant achievement: the establishment of the Antimicrobial Single Cell Testing (ASCT), a collaborative project in our laboratory where I established the wet lab part. ASCT, an imaging-based platform, enabled us to evaluate drug tolerance at the single-cell level across an international cohort comprising 191 MAB clinical isolates.
With this platform, we demonstrate a highly diverse killing kinetics in MAB across and within clinically relevant drugs, indicating that bacterial killing is a fundamental bacterial phenotype and a drug feature. We detected that cefoxitin and imipenem kill-kinetics, which were most rapid and presumably drive microbial sterilization in MAB, were predictive for clearing infections independent of MICs. In contrast, only macrolide MICs showed an association with treatment outcome. Lastly, we illustrate that time-kill kinetics of drugs sharing a similar mode of action are correlated, underscoring a shared underlying killing mechanism specific to certain drug targets.
Overall, in this doctoral thesis, I tackled critical aspects of MAB resistance and heteroresistance, and delved into drug tolerance in detail. I show that combining drug tolerance and classical resistance metrics could enhance the prediction of treatment outcomes. These findings underscore the significance of mechanisms involved in bacterial killing, offering valuable insights to guide the treatment of mycobacterial diseases.