#	Pagina
attuale pagina	/open-h2020/projects/202598/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 2 - EvoStruc (The physics of antibiotic resistance evolution in spatially-structured multicellular assemblies)

Teaser

Summary

Antimicrobial resistance (AMR), meaning bacterial infections that have become resistant to antibiotic drugs, is a huge and growing challenge to modern healthcare worldwide. In 2016 the O\'Neill report for the UK government estimated the cost of AMR as $100 trillion in lost productivity by the year 2050, if action is not taken. Combating AMR is a challenge for physicists as well as for biologists and clinicians. This is because we need mathematical and computational models to understand how bacterial infections develop and grow, and to predict the effects of antibiotics on how bacteria grow and evolve. Improving our basic understanding of how antibiotics work and how resistance evolves is essential if we are to design better treatment regimes for bacterial infections, that can use less antibiotics, and prevent resistance from happening.

One of the reasons that we need the tools of physicists to understand bacterial infections is that they can be spatially structured: often, bacteria in an infection grow in densely packed communities, known as biofilms, which can organise themselves into complicated spatial patterns. Statistical and soft matter physics can help us understand how these spatial structures form. We expect the details of how an infection is spatially structured to affect how it responds to antibiotics, and how it evolves to become resistant to antibiotics. In turn, antibiotic exposure and evolution of resistance may alter how the infection is structured.

The EVOSTRUC project aims to uncover the two-way link between the emergence of spatial structure in bacterial populations, such as infections, and the evolution of antibiotic resistance. We will do this using a combination of laboratory experiments, computer simulations and mathematical models. In the first stage of the project, our objective is to understand how a spatial gradient of an antibiotic affects how AMR evolves in a bacterial population. Later, we will extend this understanding to investigate how real biofilms grow and form spatial structure, and how this affects evolution of resistance. Finally, in the last stage of the project, we will try to use the knowledge we have gained to help design evolution-resistant surface coatings for applications in medical devices.

Work performed

\"
In the first stage of our project, we aim to understand how a spatial gradient of an antibiotic affects the evolution of resistance. We started this work by running computer simulations of a population of Escherichia coli (E. coli) bacteria growing and evolving in a gradient of the antibiotic ciprofloxacin. In these simulations we used previously obtained measured data on how mutations in the E. coli DNA can make the bacteria resistant to ciprofloxacin, an antibiotic that targets bacterial DNA. We compared our simulation predictions to the results of experiments that we performed in-house, in which we tracked E. coli bacteria as they grew and evolved in a gradient of ciprofloxacin. We were surprised to find that our simulation predictions and our experimental results did not match. In our experiments, evolution of resistance happens less often than our simulations predicted. This unexpected discovery led us to look in more detail at how the antibiotic ciprofloxacin works, and how bacteria evolve resistance to it. This has led to several significant achievements. First, we have developed what we believe is the first mathematical model for how ciprofloxacin kills bacteria. Second, we have found that there is a delay between when a resistance mutation happens in the DNA of a bacterium, and when it actually starts to be resistant to ciprofloxacin. This delay is enough to explain why our simulation and experimental results do not match. We are also in the process of performing similar analyses for antibiotics that target bacterial cell walls.

In parallel with this work, we have also used computer simulations to understand better how a bacterial population colonises a drug gradient. These simulations reveal that antibiotics that work better on fast-growing bacteria are much more effective at inhibiting spread of the population than those that work better on slow-growing bacteria. However, if the antibiotic kills the bacteria (as opposed to just preventing them from growing), then there is a tradeoff: the antibiotic that works better on the slow-growing cells kills a higher percentage of the population even though the spatial spread of the population is greater.

We have also started work on the second stage of our project: investigating how bacterial biofilms grow on surfaces and the link between the spatial structure of the biofilm and how it evolves resistance to antibiotics. Using computer simulations in which individual bacteria are tracked as they proliferate and interact with each other, we have shown that biofilms can form into distinct kinds of spatial structures depending on the environmental conditions. We have also found that these spatial structures are largely controlled by the formation of non-growing \"\"pinning sites\"\" along the expanding edge of the biofilm (these are shown in pink in the attached image). If no pinning sites form (eg when nutrients are abundant), the biofilm is rather flat. When pinning sites form but rapidly disappear, the biofilm becomes rough in structure. When pinning sites form and do not disappear, the biofilm becomes \"\"fingered\"\". We are just beginning the next part of this investigation, on how this spatial structure impacts evolution. In the second attached image, we colour parts of the biofilm according to which of the original bacteria they are descended from. It is clear to see that the flat biofilm (left picture) has much more genetic diversity than the rough biofilm (right picture).

In parallel with this simulation work, we have been developing methods for growing and evolving biofilms in our laboratory. This work is still in progress but we have several promising lines of research underway.

The final part of our project aims to optimise biofilm spatial structure to prevent the evolution of resistance. Although we are not yet working extensively on this, we have started to investigate ways to topologically modify surfaces, as a way to control biofilm spatial structure.\"

Final results

Our project has already led to new understanding of how DNA-targeting antibiotics work and how the evolution of resistance to some antibiotics (including ciprofloxacin) is more complicated than standard models suggest. We have also gained new understanding of how the coupling between nutrient availability and a drug gradient controls how effective different antibiotics are at stopping a bacterial population from spreading. Our biofilm simulations have also revealed the importance of pinning sites along the growing biofilm edge in controlling what spatial structure the biofilm forms. We are in the process of writing research papers, or have already submitted them, about all of these topics.

In the process of this investigation, we have also developed several new techniques. We have found a way of measuring bacterial growth rates more accurately than the commonly-used log fitting method, and we have also constructed a turbidostat, which is a bacterial growth chamber in which the growth rate can be measured every few minutes without perturbing the bacterial population. We have also developed an efficient way of running very long computer simulations of biofilms, which has allowed us to go well beyond the state of the art in terms of our simulations. For our biofilm experiments, we have been exploring a range of experimental setups that will allow us to track the spatial structure of a biofilm as it grows while also being able to perform evolution experiments.

In the remaining part of the project, we plan to extend the models that we have made for ciprofloxacin to understand better another class of antibiotics, those that target bacterial cell walls. Our preliminary work suggests that simple mathematical models can explain much about how cell wall targeting antibiotics work differently on different growth media. We have also set up a collaboration with a manufacturer of surface coatings, in which we are using our models to understand how biofilms form in the gradients of biocide that are released from the surfaces of antifouling paints. We will extend our biofilm simulations to include the evolution of antibiotic resistance, and we will perform laboratory studies of how biofilm spatial structure emerges and how it affects the evolution of resistance. Finally, we will move on to controlling biofilm spatial structure by modifying the surfaces that biofilms grow on. We have already set up a collaboration with researchers at Heriot-Watt University in Edinburgh who can supply us with surfaces on which grooves have been etched; we expect these to impact on the types of biofilms that form and on the evolution of resistance.