Assessing heterogeneity of Mycobacterium tuberculosis clinical isolates using ODELAM, a novel time-lapse microscopy platform
Magdalena Donczew1, Thurston Herricks2, Fred D. Mast2, John D. Aitchison2, David R. Sherman1
1 Department of Microbiology, University of Washington, Seattle, WA 98109
2 Center for Global Infectious Diseases Research, Seattle Children's Research Institute, Seattle, WA, 98109
We have developed a time-lapse microscopy-based platform, ODELAM (One-Cell-Doubling-Evaluation-of-Living-Arrays-of-Mycobacterium), for assaying the growth kinetics of Mycobacterium tuberculosis (Mtb) on solid medium in real time with single-colony resolution. Our image processing pipeline quantitates the growth kinetics of up to one hundred thousand individual colony forming units (CFUs) in up to five different conditions simultaneously. ODELAM provides a multivariate analysis of cell growth kinetics, individually measuring the lag-, log- and stationary phase of each CFU. By focusing on the physiology of single cells growing into microcolonies we are able to assess bacterial heterogeneity, a property highly relevant to bacterial physiology, pathogenesis and drug tolerance.
We used ODELAM to analyze the growth and heterogeneity of Mtb strains belonging to different phylogenetic lineages that collectively represent the majority of TB globally. We investigated the response of these strains to a range of concentrations of commonly used antitubercular drugs. To assess the level of heterogeneity, we quantified the inter-quartile range (IQR) for lag- and doubling time for each strain in each condition. At sub-MIC concentrations of drugs, all strains show increased variability in measured growth parameters, compared to no-drug condition. We also noticed that strains with the highest MIC to certain drugs show especially high heterogeneity upon exposure to sub-MIC concentrations of these drugs. Exposure of Mtb isolates to bedaquiline (BDQ) resulted in unique bacterial behavior. At low concentrations of BDQ, we observed that each strain shows two phenotypes with significantly different growth kinetics. Both cell types are alive but only one subpopulation is actively growing and dividing. Interestingly, dividing cells appear to grow better in sub-MIC drug concentrations than on the drug-free media, as showed by the median doubling times.
We hypothesize that stress-induced heterogeneity, both within and across isolates, facilitates adaptation to unfavorable conditions and contributes to drug tolerance. We also propose that sub-MIC levels of BDQ may have a beneficial effect on mycobacterial growth and is an example of the biological phenomenon known as hormesis.