Bedaquiline and clofazimine resistance in Mycobacterium tuberculosis: an in-vitro and in-silico data analysis

Lindsay Sonnenkalb, Joshua James Carter, Andrea Spitaleri, Zamin Iqbal, Martin Hunt, Kerri M. Malone, Christian Utpatel, Daniela María Cirillo, Camilla Rodrigues, Kayzad Nilgiriwala, Philip W. Fowler, Matthias Merker, Stefan Niemann, Ivan Barilar, Simone Battaglia, Emanuele Borroni, Angela Pires Brandão, Alice Brankin, Andrea Maurizio Cabibbe, Pauline Claxton, David A. Clifton, Ted Cohen, Jorge Coronel, Derrick W. Crook, Viola Dreyer, Sarah G. Earle, Vincent Escuyer, Lucilaine Ferrazoli, George F. Gao, Jennifer L. Gardy, Saheer E. Gharbia, Kelen Teixeira Ghisi, Arash Ghodousi, Ana Lúıza Gibertoni Cruz, Louis Grandjean, Clara Grazian, Ramona Groenheit, Jennifer L. Guthrie, Wencong He, Harald Hoffmann, Sarah Hoosdally, Nazir Ismail, Lisa Jarrett, Lavania Joseph, Ruwen Jou, Priti Kambli, Rukhsar Khot, Jeff Knaggs, Anastasia Koch, Donna Kohlerschmidt, Samaneh Kouchaki, Alexander S. Lachapelle, Ajit Lalvani, Simon Grandjean Lapierre, Ian F. Laurenson, Brice Letcher, Wan-Hsuan Lin, Chunfa Liu, Dongxin Liu, Ayan Mandal, Mikael Mansjö, Daniela Matias, Graeme Meintjes, Flávia de Freitas Mendes, Marina Mihalic, James Millard, Paolo Miotto, Nerges Mistry, David Moore, Kimberlee A. Musser, Dumisani Ngcamu, Ngoc Nhung Hoang, Camus Nimmo, Nana Okozi, Rosângela Siqueira de Oliveira, Shaheed Vally Omar, Nicholas I. Paton, Tim Peto, Juliana Maíra Watanabe Pinhata, Sara Plesnik, Zully M. Puyén, Marie Sylvianne Rabodoarivelo, Niaina Rakotosamimanana, Paola M. V. Rancoita, Priti Rathod, Gillian Rodger, Timothy C. Rodwell, Eaysha Roohi, David Santos-Lázaro, Sanchi Shah, Thomas A. KohlThe Lancet Microbe 2023 · DOI

TL;DR

This study tackles the growing problem of drug resistance in Mycobacterium tuberculosis (TB) by creating a comprehensive catalog of genetic variants associated with resistance to bedaquiline and clofazimine. By combining novel in-vitro evolution experiments with extensive in-silico analysis and a large clinical dataset, they've uncovered new resistance mechanisms and provided a crucial resource for improving TB diagnostics and treatment strategies.

What Problem Does This Paper Address?

Multidrug-resistant tuberculosis (MDR-TB) is a major global health threat, and while new drugs like bedaquiline have become cornerstones of treatment, the emergence of resistance is a significant concern. Understanding the genetic basis of this resistance is crucial for developing rapid diagnostic tools and tailoring effective treatment regimens. However, our knowledge of how Mycobacterium tuberculosis develops resistance to bedaquiline and clofazimine, particularly the underlying genetic mechanisms and their phenotypic consequences, has been incomplete. This gap hinders the development of accurate molecular diagnostics and can lead to treatment failures.

Methods

The researchers employed a multi-pronged approach:

  • Novel In-Vitro Evolution Model: They developed a unique method using sub-inhibitory concentrations of bedaquiline and clofazimine to select for resistant M. tuberculosis mutants over multiple bacterial passages. This mimics conditions where drug exposure might not be fully effective, potentially reflecting real-world scenarios.
  • Genome Sequencing: Both Illumina and PacBio sequencing were used to characterize the genetic makeup of the selected in-vitro mutants, allowing for the identification of new mutations and even large-scale genomic rearrangements.
  • Phenotypic Characterization: Minimum Inhibitory Concentrations (MICs) were determined for the selected mutants using resazurin microtitre assays and the BACTEC MGIT 960 system to quantify their resistance levels.
  • In-Silico Analysis:
    • Protein Modelling: Structural models of the Rv0678 transcriptional repressor were used to investigate how identified mutations might affect protein stability, DNA binding, protein dimerization, and ligand affinity.
  • Molecular Dynamics Simulations: These simulations were used to explore how specific mutations in the Rv0678 hinge region could disrupt conformational changes necessary for DNA binding, potentially leading to resistance.
  • Clinical Data Integration: They integrated their findings with a large dataset of over 14,000 clinical M. tuberculosis isolates from the Comprehensive Resistance Prediction for Tuberculosis: an International Consortium (CRYPTIC) project, as well as publicly available data, to build a comprehensive mutation catalogue.

Datasets

  • In-vitro evolved mutants: A novel dataset generated through the study's experimental evolution model.
  • Clinical M. tuberculosis isolates: Over 14,000 isolates from the CRYPTIC Consortium, providing a large-scale, real-world dataset of genotypic and phenotypic resistance information.
  • Publicly available data: Existing literature and databases of M. tuberculosis resistance variants.

The study effectively created an expanded and validated mutation catalogue for bedaquiline and clofazimine resistance, building upon existing knowledge by incorporating novel experimental and computational findings.

Key Insights and Ideas

  • Sub-inhibitory concentrations are potent drivers of resistance: The study demonstrated that even low drug concentrations, significantly below the critical MIC, can effectively select for resistant mutants. This has critical implications for drug dosing and treatment strategies.
  • Rv0678 is a central player, but not the only one: While the majority of bedaquiline resistance variants identified were in the Rv0678 gene (encoding a transcriptional repressor of an efflux pump), the study also identified resistance-associated variants in other genes like atpE and pepQ, and importantly, a novel resistance mechanism involving large-scale genomic rearrangements.
  • Multiple mechanisms of Rv0678 dysfunction: Protein modelling revealed that resistance can arise through diverse mechanisms affecting Rv0678, including impaired DNA binding, reduced protein stability, disrupted dimerization, and altered affinity for its fatty acid ligand.
  • Genotypic testing can refine phenotypic interpretation: The study highlights that genotypic data can be invaluable for clarifying borderline resistance phenotypes, which is essential for accurate diagnosis and treatment selection.
  • In-vitro evolution as a predictive tool: The in-vitro model proved effective in identifying novel resistance variants and mechanisms that may not yet be prevalent in clinical settings, offering a prospective approach to understanding resistance evolution.

Key Results

  • 265 genomic variants identified: The study catalogued 265 genomic variants implicated in bedaquiline resistance, with 250 (94%) affecting the Rv0678 gene.
  • 40 new variants discovered in vitro: The in-vitro evolution model identified 40 novel variants associated with bedaquiline resistance.
  • New resistance mechanism: Genomic rearrangement: A significant finding was the identification of a large-scale genomic inversion disrupting the Rv0678 gene, representing a novel mechanism of bedaquiline resistance.
  • Cross-resistance confirmed: The study confirmed that many bedaquiline resistance variants also confer cross-resistance to clofazimine.
  • Validation of clinical findings: 15% of mutations found in clinical bedaquiline-resistant isolates were also identified in their in-vitro work.
  • Structural insights: Protein modelling elucidated four major mechanisms by which Rv0678 variants confer resistance.

Limitations and Future Work

  • Limitations of external MIC data: The authors acknowledge that external MIC data from previous studies were collected using different phenotypic assays, which may introduce variability.
  • Undetermined phenotypes: Some variants had a wide range of phenotypes or discrepancies between methods, leading to their classification as "undetermined."
  • Epistatic interactions: The role of secondary variants in genes like mmpS5 or mmpL5 in potentially reverting resistance phenotypes needs further experimental exploration.

Future Directions:

  • Biochemical investigation: The proposed disruption of ligand binding in Rv0678 needs biochemical validation.
  • Structure-based machine learning: The study suggests that structure-based machine learning approaches, similar to those used for other TB drugs, could be developed for predicting bedaquiline resistance.
  • Clinical relevance of novel variants: Further investigation is needed to understand the prevalence and clinical impact of the newly identified resistance variants and mechanisms.
  • Refining diagnostic tools: The comprehensive mutation catalogue can be directly implemented into diagnostic platforms (e.g., Deeplex Myc-TB assay) and bioinformatic pipelines to improve the interpretation of molecular TB testing.

Summary updated

BibTeX
@inproceedings{sonnenkalb2023bedaquiline,
  author = {Lindsay Sonnenkalb and Joshua James Carter and Andrea Spitaleri and Zamin Iqbal and Martin Hunt and Kerri M. Malone and Christian Utpatel and Daniela María Cirillo and Camilla Rodrigues and Kayzad Nilgiriwala and Philip W. Fowler and Matthias Merker and Stefan Niemann and Ivan Barilar and Simone Battaglia and Emanuele Borroni and Angela Pires Brandão and Alice Brankin and Andrea Maurizio Cabibbe and Joshua A. Carter and Daniela María Cirillo and Pauline Claxton and David A. Clifton and Ted Cohen and Jorge Coronel and Derrick W. Crook and Viola Dreyer and Sarah G. Earle and Vincent Escuyer and Lucilaine Ferrazoli and George F. Gao and Jennifer L. Gardy and Saheer E. Gharbia and Kelen Teixeira Ghisi and Arash Ghodousi and Ana Lúıza Gibertoni Cruz and Louis Grandjean and Clara Grazian and Ramona Groenheit and Jennifer L. Guthrie and Wencong He and Harald Hoffmann and Sarah Hoosdally and Martin Hunt and Zamin Iqbal and Nazir Ismail and Lisa Jarrett and Lavania Joseph and Ruwen Jou and Priti Kambli and Rukhsar Khot and Jeff Knaggs and Anastasia Koch and Donna Kohlerschmidt and Samaneh Kouchaki and Alexander S. Lachapelle and Ajit Lalvani and Simon Grandjean Lapierre and Ian F. Laurenson and Brice Letcher and Wan-Hsuan Lin and Chunfa Liu and Dongxin Liu and Kerri M. Malone and Ayan Mandal and Mikael Mansjö and Daniela Matias and Graeme Meintjes and Flávia de Freitas Mendes and Matthias Merker and Marina Mihalic and James Millard and Paolo Miotto and Nerges Mistry and David Moore and Kimberlee A. Musser and Dumisani Ngcamu and Ngoc Nhung Hoang and Stefan Niemann and Kayzad Nilgiriwala and Camus Nimmo and Nana Okozi and Rosângela Siqueira de Oliveira and Shaheed Vally Omar and Nicholas I. Paton and Tim Peto and Juliana Maíra Watanabe Pinhata and Sara Plesnik and Zully M. Puyén and Marie Sylvianne Rabodoarivelo and Niaina Rakotosamimanana and Paola M. V. Rancoita and Priti Rathod and Gillian Rodger and Camilla Rodrigues and Timothy C. Rodwell and Eaysha Roohi and David Santos-Lázaro and Sanchi Shah and Thomas A. Kohl},
  title = {{Bedaquiline and clofazimine resistance in Mycobacterium tuberculosis: an in-vitro and in-silico data analysis}},
  booktitle = {The Lancet Microbe},
  year = {2023},
  doi = {10.1016/s2666-5247(23)00002-2},
}
Figure