Four decades of transmission of a multidrug-resistant Mycobacterium tuberculosis outbreak strain

Abstract

The rise of drug-resistant strains is a major challenge to containing the tuberculosis (TB) pandemic. Yet, little is known about the extent of resistance in early years of chemotherapy and when transmission of resistant strains on a larger scale became a major public health issue. Here we reconstruct the timeline of the acquisition of antimicrobial resistance during a major ongoing outbreak of multidrug-resistant TB in Argentina. We estimate that the progenitor of the outbreak strain acquired resistance to isoniazid, streptomycin and rifampicin by around 1973, indicating continuous circulation of a multidrug-resistant TB strain for four decades. By around 1979 the strain had acquired additional resistance to three more drugs. Our results indicate that Mycobacterium tuberculosis (Mtb) with extensive resistance profiles circulated 15 years before the outbreak was detected, and about one decade before the earliest documented transmission of Mtb strains with such extensive resistance profiles globally.

Figure 1. Phylogenetic placement of the M outbreak strains.

(a) Maximum-likelihood whole-genome SNP phylogeny of sublineage 4.1.2.1 isolates from a global Mtb collection. Black dots indicate the acquisition of the katG S315T mutation conferring INH resistance. The grey dot indicates the gidB V110 frameshift mutation conferring STR resistance. (b) Histogram of pairwise SNP distances between M outbreak isolates.

Figure 2. Dated phylogeny of the M outbreak.

Resistance mutations are indicated directly on the tree. For resistance mutations that evolved early in the outbreak, the date of acquisition is indicated (95% CI). The timescale at the bottom of the figure illustrates the year of introduction of relevant antibiotics, whereas the vertical bar at the far right illustrates phenotypic drug resistance level as determined by drug susceptibility testing. *PZA was introduced for retreatment cases in 1961 and included in the standard first-line treatment scheme from 1979.

Figure 3. Correlation between ethionamide resistance and mutations in ethA, ndh, mshA and inhA promoter.

(a) The horizontal bar illustrates the ethA gene. Mutations are colour-coded by phenotypic MIC levels observed in the strains harbouring the mutations. Mutations indicated above the bar are nonsynonymous mutations that do not disrupt the reading frame, whereas mutations indicated below the bar result in transcriptional frameshift or the introduction of a premature stop codon. ETH MIC levels in relation to identified mutations in ndh, mshA and the inhA promoter are illustrated separately. Circle size indicates the number of isolates harbouring the individual mutations. (b) Summary of MIC levels in wild-type and mutant isolates.

Figure 4. Mutations in RNA polymerase subunits rpoB, rpoC and rpoA.

The rpoB S450L mutation common to 241 (light-blue-shaded background) of the 252 isolates is indicated by the dark green dot close to the root of the phylogeny. Other mutations in RNA polymerase subunits are indicated with coloured dots (rpoB: blue; rpoC: magenta; rpoA:pink). Mutations common to two or more closely related isolates, indicative of transmission, are indicated by shaded background following the same colour scheme. The amino-acid change caused by each mutation is annotated along the periphery of the phylogeny.