In host mutational adaptation of Mycobacterium tuberculosis complex strains
BackgroundTuberculosis (TB) is the second leading cause of death from an infectious disease worldwide. Mycobacterium tuberculosis complex (MTBC), the causative bacteria of TB, is becoming increasingly resistant to anti-TB drugs, resulting in poorer treatment outcomes. The mutations arising in MTBC strains during infection provide a record of bacterial diversification and adaptation, and indirect evidence on the selective pressures and conditions that MTBC encountered in vivo, such as those exerted by the immune system and antibiotic therapies. MethodsWe conducted a meta-analysis of MTBC sequenced strains with multiple clinical isolates sequenced from the same patient, from published studies and TB Portals. We applied stringent genomic QC criteria to keep only clonal isolates and applied robust genomic pipelines to identify mutations arising de novo during infection. A convergent evolution approach was applied to identify heavily mutated genes, operons, and candidate promoter across strains of multiple patients. We estimated the frequency of drug resistance acquisition during treatment for the subset of patients with available drug treatment data. ResultUsing 5,899 high-quality genomes from 1,056 TB patients after ensuring clonality of isolate genomes we identified limited within-host diversity was identified including 3,296 fixed mutations across 501 patients. A total of 21 genes, 25 operons and 27 promoter regions were statistically enriched by mutations compared to the rest of the genome, and additional loci with established or plausible adaptive roles approaching statistical significance. Significantly, this included multiple loci known to be involved in resistance to first-, second-, and last-line anti-TB drugs. Fluoroquinolone resistance was acquired more frequently during treatment than resistance to any other anti-TB drug. Previously reported candidate drug-resistance and -tolerance genes (prpR, Rv2571c, fadD11, helY, ndhA, Rv0139, fadE5, and mce1 operon) were also identified. Genes encoding regulators (phoR, whiB6 and mycP1) and effectors (espK and eccE1) of the virulence ESX-1 locus were frequently mutated in host. ConclusionsHere we analysed a large dataset of MTBC within-host genetic diversity. We show that frequently mutated genes in MTBC during infection reveal known and biologically plausible in host adaptations, predominantly associated drug resistance, but also in genes involved in pathogenesis. The higher resistance acquisition rate observed for fluoroquinolones may have important clinical relevance. We reveal a list of candidate loci which will require mechanistic characterisation and whose impact on disease progression will need to be investigated.