Biophysics

Theme 1: Modelling cell structures in tissues (J. Hague, H. Dickinson with C. McMullen (LHCS), K. Rietdorf (LHCS))

There is an unmet need for tools to design moulds and scaffolds for the growth of cultured tissues with bespoke cell organisations for applications in e.g. regenerative medicine, drug screening and cultured meat, and explain how large numbers of cells organise in polarised tissues. We have developed a microscopic biophysical model to explain the self-organisation of cells in polarised tissue. In the model, order is co-driven by short-range active forces between cells and the extracellular matrix (ECM) and macroscopic forces that develop within the ECM. We find close agreement with experiments on cell-laden hydrogels. 

Furthermore, we’ve demonstrated that machine learning tools can predict the role of this mechanobiology. The speed of the machine learning approach opens the possibility of real-time rational and automated design of moulds, scaffolds and 3D bioprinting strategies that account for the organisation of cells in the resulting cultured tissue. Ongoing work shows that the model applies to a broad range of cell types. Our ambition is to improve the quality and size of cultured tissues for a range of applications. Our impact development is ongoing through the creation of a cloud-based application to assist design in industry.

Theme 2: Understanding vascular structure from first principles (J. Hague)

Do the complex processes of angiogenesis during organism growth ultimately lead to near-optimal vasculatures in adult organs? We have examined this hypothesis by developing a powerful model driven by simulated annealing. The model finds the optimal structure of arterial trees that minimises power consumption. It provides insight into the properties of the whole arterial structure. Vasculatures generated using this novel method closely match coronary (figure, right) and cerebral morphologies. We are pursuing an important application for this work. Preformed vasculatures are essential to supply oxygen and nutrients to cultured tissues, meats, organoids and organs, since natural angiogenesis happens too slowly to prevent necrosis in large cultured tissues. Our latest work extends our optimisation algorithm to design vasculatures for cultured tissues. We have made key extensions that allow for the simultaneous optimisation of both arterial and venous vasculatures to avoid vessel intersections that would bypass the capillary network. Our method could be combined with 3D printing to generate vasculatures for arbitrary shapes of cultured tissue, be used to understand blood supply to organs, and form the core of simulations related to cardiovascular diseases, e.g. stroke.

Theme 3: Virtual patient / digital twin models of stroke (J. Hague, H. Dickinson with P. Wong (C&C))

Strokes have a major impact on the economy and wellbeing. Embolic strokes make up the majority of all strokes and are often the most severe. Computational stroke forecasting in digital twins could provide the next generation of advanced clinical monitoring tools for intra-operative stroke prevention. It could also be used to complement and replace animal models and to run in-silico trials. We have developed novel Monte Carlo models to predict how embolisation properties affect the severity of strokes. Our model simulates embolus motion and blood flow through the cerebral vascular. By extending the model with cerebral vasculatures “grown” in-silico using our simulated annealing approach, we created digital twin stroke simulations. The digital twin made it possible for us to run virtual trials to explore how embolisation origin, rate and composition (e.g. gaseous or solid) affect patient outcomes. We are investigating the application of machine learning to this problem. We aim to develop tools for personalised monitoring (both intra-operative and community) and use machine learning techniques to determine stroke origins.