CellPrism · Resolving brain cell-types with non-invasive imaging

The brain is plastic: it constantly adapts to its environment thanks to numerous cell-types with varied functions. Brain cell-types fall into two broad categories: neurons and glial cells. These are further subdivided into multiple subtypes with specific functions tightly linked to their morphology. For example, astrocytes and microglia, two types of glial cells, respond very differently to brain inflammation: astrocytes increase in size and their morphology becomes more complex, whereas microglia become rounder and less complex. Gold-standard methods for assessing brain cell morphology are highly invasive. They are rarely applied in humans and cannot be used for repeated assessments over time, which is essential to understanding evolving processes and diseases. Magnetic resonance imaging can non-invasively image brain structure and function, but it cannot currently differentiate cell-type-specific changes.I will build a non-invasive brain cell-type decoder, to extract cellular composition and morphometry for a range of cell subtypes in physiological and pathological conditions, akin to a non-invasive biopsy. To achieve this I propose a novel paradigm built upon diffusion-weighted magnetic resonance spectroscopy (dMRS), a technique that is sensitive to the displacement of intracellular metabolites. I aim to enhance dMRS to achieve cell-subtype specificity and go beyond the glial/neuronal dichotomy. CellPrism’s strategy integrates prior knowledge about cell-types and encompasses the wealth of information available in dMRS in an integrated manner. CellPrism will be validated using causal pharmacological and genetic manipulations in mice. Translation to humans will be conducted at high field, whilst establishing sensitive protocols at accessible clinical fields. Finally, CellPrism will be piloted in patient populations where cellular mechanisms are well described in animal models and post-mortem analyses.