Genomic epidemiology of Mycobacterium abscessus in a Canadian cystic fibrosis centre

Abstract

The Mycobacterium abscessus complex causes significant morbidity and mortality among patients with Cystic Fibrosis (CF). It has been hypothesized that these organisms are transmitted from patient to patient based on genomics. However, few studies incorporate epidemiologic data to confirm this hypothesis. We longitudinally sampled 27 CF and 7 non-CF patients attending a metropolitan hospital in Ontario, Canada from 2013 to 2018. Whole genome sequencing along with epidemiological data was used to evaluate the likelihood of transmission. Overall, the genetic diversity of M. abscessus was large, with a median pairwise distance (IQR) of 1,279 (143-134) SNVs between all Ontario M. abscessus isolates and 2,908 (21-3,204) single nucleotide variants (SNVs) between M. massiliense isolates. This reflects the global diversity of this pathogen, with Ontario isolates widely dispersed throughout global phylogenetic trees of each subspecies. Using a maximum distance of 25 SNVs as a threshold to identify possible transmission, we identified 23 (of 276 total) pairs of closely-related isolates. However, transmission was probable for only one pair based on both genomic and epidemiological data. This suggests that person-to-person transmission of M. abscessus among CF patients is indeed rare and reinforces the critical importance of epidemiological data for inferences of transmission.

Figure 1

Sampling date and subspecies assignment. Isolates collected from 2013 to 2018 from 30 patients with CF and 7 non-CF patients. (a) Isolates either could not be regrown or used for sequencing (open circles) or were sequenced and successfully assigned a subspecies (M. abscessus-squares, M. massiliense-triangles, M. bolletii-filled grey square). More isolates were available from patients with CF (1–30) than non-CF patients (31–37, grey background). (b) Fraction of available isolates sequenced by month.

Figure 2

Subspecies maximum likelihood phylogeny. These phylogenies were constructed from the recombination-masked SNV alignment, and drawn using the midpoint root. Bootstrap support > 80% is indicated on nodes with red circles. The seven locus (argH, cya, gnd, murC, pta, purH, and rpoB) PubMLST sequence types (ST) are indicated in coloured squares where an existing ST could be assigned (see Methods). Isolates from patients with CF are indicated in black, while those from non-CF patients are indicated in dark red. a M. abscessus. b M. massiliense.

Figure 3

Summary of pairwise SNV distances. Recombination-masked SNV distances for the indicated clades were computed for each subspecies and clade in Fig. 2. Lines and boxes indicate the median and interquartile range (IQR), while whiskers extend to 1.5 × the IQR. Colours match the indicated clades in the  Fig. 2. Where missing, median lines coincide with upper quartiles. (a) M. abscessus. (b) M. massiliense. c Within- and Between-patient pairwise SNV distances by subspecies.

Figure 4

Distribution of unadjusted pairwise SNV distances of global M. abscessus isolates. The 95th percentile of pairwise within-patient SNVs of the Ontario isolates is indicated by vertical lines, but could not be calculated for M. bolletii. (a) M. abscessus (b) M. massiliense (c) M. bolletii.