Identification of genes required for host-cell adhesion and antibiotic persistence of uropathogenic Escherichia coli via high-content screening

Uropathogenic Escherichia coli (UPEC) are the most common bacterial pathogens causing urinary tract infections (UTIs). Driven by the development of antibiotic resistant UPEC strains, UTIs have become a major public health issue and generate substantial healthcare costs. Even in the absence of antibiotic resistance, UTIs are notoriously difficult to treat and have high rates of recurrence, mostly due to: 1. the ability of UPEC to adhere to and invade bladder epithelial cells, which protects them from the action of antimicrobial agents and the immune system 2. the presence in antibiotic-susceptible UPEC populations of a small fraction of so-called persister cells, which are killed at a slower rate than the bulk of the population and can therefore survive prolonged antibiotic exposureThese two features allow some bacteria to remain viable in a protected niche for the whole duration of the antibiotic treatment, potentially leading to infection relapse when the treatment is terminated.High-content screening (HCS), which uses microscopy as a screening tool, allows the visualization of bacterial phenotypes at the single-cell level and can therefore provide a powerful and unbiased tool to understand the genetic basis of these phenotypes. The present thesis describes two complementary high-content screening approaches aimed at identifying UPEC mutants with altered adhesion to bladder cells and with altered antibiotic persistence respectively.In a first chapter, we report preparatory steps that led to the construction of a transposon insertion library of UPEC mutants that is suitable for high-content imaging, using the CFT073 clinical UTI isolate as background strain.In the second chapter, we performed fluorescence microscopy-based high-content screening to identify mutants with defects in early adhesion to bladder epithelial cells. We recovered 82 mutants with decreased adhesion and 54 mutants with increased adhesion. Unexpectedly, nine low-adhesion ÂÂhitsÂÂ mapped to the two P pili operons encoded by CFT073, which are thought to mediate adhesion to kidney cells rather than bladder cells. Additionally, six high-adhesion hits mapped to the operon coding for F1C pili, and we show that this phenotype is linked to increased P pili synthesis. These results therefore reveal a critical role for P pili in UPEC adhesion to bladder epithelial cells, which may inform the development of novel anti-adhesion therapies to prevent UTI recurrence.In the third chapter, we combined microfluidics with time-lapse microscopy to identify mutants with altered persistence to fosfomycin (FOS) âÂÂ a cell-wall inhibitor used as first-line agent in the treatment of UTIs âÂÂ in synthetic human urine (SHU), which mimics the physiology of UPEC in real urine. We recovered four mutants with decreased persistence, three of which had defects in lipopolysaccharide (LPS) synthesis. These results identify the outer membrane of UPEC as a key component for survival to FOS treatment, thus opening up new avenues in the fight against antibiotic persistence.

Single-cell division and killing dynamics of bacteria exposed to isocratic and time-dependent concentrations of antibiotic

Katrin Esther Schneider

Antibiotic resistance, defined as the ability of a bacteria to survive an antibiotic drug to which it was initially sensitive, has become increasingly worrying. While the discovery of antibiotics is considered to be one of the most important of all biomedical advances of the twentieth century, bacterial infections are still problematic, and the medical and financial costs that they entail are a great burden to society. In the last 30 years the number of new antibiotic approvals has been decreasing dramatically, leading to a lack of replacements for antibiotics whose effectiveness has been compromised. The drug discovery pipeline is very long and expensive with high rates of attrition at each stage of the process. Massive economic investments have attempted to address this problem but they will bear little fruit if methods are not improved to make the process faster, less expensive, and more efficient. One of the fundamental steps in the evaluation of any new compound is the characterization of the relationship between its pharmacokinetic (PK) and pharmacodynamic (PD) parameters, which will ultimately determine the optimal dosing regimen to be used in clinical trials. A more detailed understanding of these parameters can make a critical difference in early go/no-go decisions, thereby increasing the effectiveness of the drug discovery process. The use of novel assays with single-cell resolution has demonstrated how conventional measurements, taken at the population level, can hide important underlying dynamics of cellular behavior. Since none of the PK/PD models available today allow single-cell measurements, this thesis is focused on the development and implementation of new tools that address this need by enabling the study of essential parameters of antibiotic effectiveness at the single-cell level. We first assessed the effect of different isocratic concentrations of antibiotic by exposing Mycobacterium smegmatis to various concentrations of isoniazid. We find that the higher number of surviving bacteria at lower concentrations of antibiotic is entirely due to increased division rates, as killing rates remain constant. Second, in order to study more realistic pharmacokinetic profiles, we construct a novel microfluidic platform that uses computer-controlled microvalves to enable the administration of time-dependent antibiotic concentration gradients. These gradients serve as input to an array of 960 microchambers in which the cells are imaged by time-lapse microscopy. By studying uropathogenic Escherichia coli (UPEC) treated with ampicillin, we evaluated dosing regimens of intermittently dosed patients. As a proof-of-principle, we accurately predict the pharmacokinetic driver for dose-optimization. Finally, we employ quantitative phase microscopy to complement our single-cell analyses with dry-mass measurements, which give us deeper insights into fundamental processes. Interestingly, we find that this technique can be used as a more rigorous tool to identify dead cells by microscopy.

EPFL2018

A Microfluidic Platform Enables Large-Scale Single-Cell Screening to Identify Genes Involved in Bacterial Persistence

Amanda Verpoorte

Persistent cells, representing a small fraction of a bacterial population, are refractory to antibiotic therapy. It is a phenotypic transient state. The underlying mechanisms of persistence remain largely unknown, but even divergent hypotheses agree on a fundamental component: persistence arises from phenotypic heterogeneity. Novel approaches are needed to elucidate which factors allow some bacteria to persist but not others. High-content screening (HCS), which uses microscopy as a screening tool, enables the visualization of single-cell phenotypes and is thus particularly suited for studying bacterial persistence. Traditionally performed with eukaryotic cells in micro-well plate format, executing HCS with prokaryotic cells required the development of a screening novel strategy due to their much smaller size. Microfluidics has proved to be a remarkable toolbox for bacterial studies. In combination with time-lapse microscopy, it permits the study of dynamic processes with single-cell resolution in a tightly controlled microenvironment. Though numerous microfluidic platforms for prokaryotes imaging have been developped, the current state of the art does not allow the imaging of thousands of genotypes in parallel, as is required for high-content screening. The focus of this thesis is to develop a high-throughput microfluidic screening platform enabling high-content imaging of prokaryotic cells. This platform allows 1152 bacterial strains to be imaged in parallel on both the single-cell and sub-cellular level with precise control over the growth environment. First, we demonstrate the ability of the platform to effectively screen for both altered cell division phenotypes through subcellular markers imaging in a static environment and for phenotypes emerging in complex medium profiles, such as bacterial persistence. Second, we employed our platform in a genetic screen, exposing a mutant library of Mycobacterium smegmatis to the antibiotic Isoniazid (INH), and identified a total of 31 mutants with altered persistence rates during antibiotic exposure as compared to the wild-type. Eight of these mutants had decreased persistence rates while 23 mutants displayed increased persistence. Third, using the same platform we performed in depth analysis of mutants with altered rates of persistence in order to identify the cell division and cell death dynamics underlying persistence. This single-cell analysis revealed unexpected dynamics where a mutant with decreased persistence displayed a lower killing rate compared to the wild-type. This high-throughput microfluidic screening platform is highly versatile and can be used in a wide range of bacterial studies, as it is capable of imaging not only single cells, but also subcellular markers, such as those based on the cell cycle or bacterial septa. This novel technology thus opens new doors for microbiologists.

EPFL2017

Microtissue models for studying the dynamics of host-pathogen interactions in the urinary bladder

Kunal Sharma

Urinary tract infections (UTIs) are amongst the most common bacterial infections and are the second-most frequent reason for antibiotic prescriptions. Moreover, in about 25% of all treated cases, recurrence of infection occurs. Uropathogenic Escherichia coli (UPEC) is the most common causative agent of UTIs. Much of our current understanding about UTIs has come from mouse models of UTI which have highlighted the intracellular lifestyle of this pathogen in the forms of intracellular bacterial communities (IBCs) as the prominent cause for bacterial persistence. Recurrent infections are thought to be caused by bacterial fluxing from persistent IBCs in the umbrella cell layer of the epithelia in the bladder. Although in vitro model systems of UTIs have been developed to study acute phases of infection including the formation of IBC and their dispersal, they suffer from certain shortcomings. For example, these models lack vasculature and the possibility to impose mechanical stresses, such as those experienced by the cells during bladder filling and voiding. Moreover, the design of these in vitro systems makes it difficult to study long-term UPEC persistence when exposed to antibiotic or neutrophil-mediated stresses. In the last two decades, two bioengineering approaches have emerged to generate functional and physiological tissues: organoid and organ-chip systems. Differentiating cells self-organize in a hydrogel and enable the formation of organoids, which mimic various aspects of organ physiology and function in a three-dimensional structure. Organ-chip systems rely upon co-culturing pre-differentiated cells across a synthetic membrane in a defined geometry. The organ-chip systems allow dynamic control over environmental (nutritive) state and over the mechanical stresses experienced by the emulated tissue. In this thesis, I showcase the development of model systems to study UPEC pathogenesis using both these approaches. In the first part of this thesis, we developed a bladder organoid model of UPEC infection that recapitulates the stratified bladder architecture within a volume suitable for high-resolution live-cell imaging. Bacteria injected into the organoid lumen rapidly enter umbrella cells and proliferate to form IBCs. Unexpectedly, we identified populations of individual âsolitaryâ bacteria that form independently of IBCs and which penetrate deeper layers of the organoid wall, and evade killing by antibiotics and neutrophils. Volumetric electron microscopy revealed that these solitary bacteria may be intracellular or pericellular (sandwiched between uroepithelial cells) and are rod-shaped and flagellated, unlike coccoid-shaped and unflagellated bacteria within IBCs. Bacterial fluxing from IBCs is therefore not an essential requirement for the establishment of persistent bacterial populations in the bladder wall. The dynamic responses of IBCs to host stresses and antibiotic therapy are difficult to assess in situ. In another study, we focussed on examining the lifecycle and persistence of IBCs during early stages of UTIs. We developed a human bladder-chip model wherein superficial umbrella cells and bladder microvascular endothelial cells are co-cultured under flow in urine and nutritive media respectively. Bladder filling and voiding was mimicked mechanically by application and release of linear strain. Using time-lapse microscopy, we show that rapid recruitment of neutrophils from the vascular channel

EPFL2021