Lineage-specific in-vitro susceptibility Mycobacterium tuberculosis to Pretomanid

Previous work reported unprecedented differences in the intrinsic in vitro susceptibility of the Mycobacterium tuberculosis complex (MTBC) to pretomanid (Pa) using the Mycobacteria Growth Indicator Tube (MGIT) system. The goal of this study was two-fold. First, we used MGIT to refine the current understanding of the impact of the MTBC diversity with a particular focus on the less frequent lineage 5 (L5), lineage 6 (L6), lineage 7 (L7), lineage 8 (L8), and lineage 9 (L9) that either did not feature at all in Bateson et al. or were represented by only a limited number of strain. Second, we used Middlebrook 7H11 (7H11) as an alternative medium to investigate whether the differences observed with MGIT were media-specific. We tested 125 L1-L9 strains that lacked known resistance mutations in the six canonical Pa-resistance genes from patients who had never received nitroimidazole treatment originating from 45 different countries in five different continents. All 125 strains were tested on 7H11 whereas a subset of 41 isolates were tested using MGIT. In addition, four strains with known resistance to nitroimidazoles were tested on both media. Our data confirmed that MTBC, unlike most bacteria-antibiotic combinations, displayed substantial differences in the susceptibility relative to the technical variation of Pa MIC testing. This was also the case for the 7H11 medium, demonstrating that these differences were not specific to MGIT. Notably, lineage 1 was confirmed to have intrinsically elevated MICs compared with lineages 2, 3, 4 and 7 (L2-4/7), underlining the urgent need for WHO to publish its decision of whether lineage 1 should be considered susceptible to BPaL(M), the now preferred all-oral regiment for treating rifampin-resistant tuberculosis. Lineage 5 and 6, which are most frequent in West Africa, responded differently to Pa, with the former being more similar to L2-4/7 and the latter being more susceptible. Of all 129 strains included in this study, 123 had whole genome sequencing data and for 21 of these, new Fastq data was generated and published under this study project.