Oral Presentation Australian Society for Microbiology Annual Scientific Meeting 2023

Repeated evolution of bedaquiline resistance in South African Mycobacterium tuberculosis is driven by truncation of the regulatory gene mmpR5 (93883)

Leah W Roberts 1 2 , Kerri M Malone 2 , Martin Hunt 2 , Lavania Joseph 3 , Penelope Wintringer 2 , Jeff Knaggs 2 , Derrick Crook 4 , Maha Farhat 5 , Zamin Iqbal 2 , Shaheed V Omar 3
  1. Centre for Immunology and Infection Control, Queensland University of Technology, Brisbane, QLD, Australia
  2. European Molecular Biology Laboratory's European Bioinformatics Institute, Hinxton, United Kingdom
  3. Centre for Tuberculosis, National Institute for Communicable Diseases, Johannesburg, South Africa
  4. Nuffield Department of Medicine, University of Oxford, Oxford, United Kingdom
  5. Department of Biomedical Informatics, Harvard Medical School, Boston, United States of America

Tuberculosis (TB) is a communicable disease caused by the bacteria Mycobacterium tuberculosis. Despite being curable, TB remains a global challenge. Drug-resistance (DR) in M. tuberculosis can be reliably predicted using in silico methods if the mechanism has been characterised. The antibiotic Bedaquiline (BDQ) is a key component of new WHO-recommended treatment  regimens for drug resistant TB but predicting BDQ resistance (BDQ-R) from genotypes remains challenging. South Africa has one of the highest usage of BDQ to date (>50% of global usage between 2015-2020), and represents a key region for the early detection of emerging BDQ-R. 

 

We analysed a collection (n=505) of whole genome sequenced M. tuberculosis from two high prevalence areas in South Africa (Cape Town and Johannesburg, 2019-2020). The dataset consisted mainly of lineage 2 and lineage 4 isolates, with BDQ-R isolates (n=32) spread throughout, indicating independent evolution of resistance rather than clonal spread. Many of the BDQ-R isolates had mutations in the efflux pump regulator gene mmpR5. We identified 53 independent acquisitions of 31 different mutations, with a particular enrichment of truncated MmpR5 by either frameshift, nonsense mutation, or introduction of an insertion element. Truncations impacted 66% of BDQ-R isolates and mainly occurred before or within the DNA binding domain of the protein. 

 

Extending our analysis to 5,253 publicly available South African M. tuberculosis from the European Nucleotide Archive (ENA), including 1,961 isolates from the CRyPTIC consortium with minimum inhibitory concentrations (MICs), revealed that mmpR5-disrupted isolates had a median BDQ MIC of 0.25 mg/L, compared to the “wild-type” (non-mmpR5-truncated) median of 0.06 mg/L. By matching mmpR5-disrupted isolates with phylogenetically close control isolates without the disruption, we were able to estimate the impact on MIC of individual mutations, including an average 2-fold increase in MIC associated with frameshifts in the DNA-binding domain of MmpR5. 

 

In conclusion, as the MIC increase we identified borders the ECOFF threshold for BDQ-R, we recommend the continued use of MICs and detection of MmpR5 truncations to identify shifts in BDQ-R in South Africa. The impact of elevated MIC on patient outcome is also of great interest, and warrants future study.