Mapping functional changes to retinal output neurons in photoreceptor degeneration using two-photon imaging and spatial transcriptomics

• Applicants: Professor Dr. Philipp Berens; Professor Dr. Thomas Euler
• Subject Area: Experimental and Theoretical Network Neuroscience ; Molecular and Cellular Neurology and Neuropathology
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 552589001

Classification of neurons into types – the “building blocks” of complex neural networks – is an important approach towards understanding brain function. Cell types in the brain have been defined based on functional responses, morphologies, or transcriptomes. However, to date, such cell type classifications have rarely integrated data across modalities. In addition, cell types are not static: Neurons can change or lose their function during aging and disease. This is a problem when clustering relies on a single modality, such as neural responses: Is a “new” functional type in a disease model indeed “new” or a known type with altered function? How are cell types changing when they age? To address both issues, a dynamic, multimodal cell type concept is needed: Dynamic, to capture changes in cell type function, and multimodal, to use information across modalities for robust cell type identification. Here, we propose to classify mouse retinal ganglion cells (RGCs) based on aligned functional and transcriptomic data from the same tissue, generated with two-photon calcium imaging and spatial transcriptomics. We hypothesize that RGC types can vary in their function and transcriptome over time, e.g., during normal aging or a progressive disease. To test this hypothesis, we will develop novel computational methods for multimodal cell type classification. We will use this classification to follow RGC types over the disease progression in rd10 mice, a model for hereditary retinal degeneration, such as Retinitis Pigmentosa. In these mutant mice, photoreceptors degenerate over the course of 6 months, while the inner retina remains structurally intact and is partially remodeled. We will start by collecting a multimodal dataset in young adult wildtype mice. We will record RGC light responses in the explanted retina and perform spatial transcriptomics in the same tissue to determine the mRNA fingerprints for the recorded cells. We will develop a new multimodal clustering approach and seed it by established cell atlases to integrate functional and transcriptomic fingerprints of the RGC types. Next, we will collect wildtype data for different ages. Accordingly, we will extend our clustering model to gain insights into how RGC types change their function and transcriptome as they age. Finally, we will acquire a dataset in the rd10 mouse retina at different time points during degeneration. Our set of target genes will contain RGC type-specific and disease-associated markers, allowing us to correlate functional and transcriptomic changes in distinct RGC types over disease progression. Our work will result in the first RGC type clustering that is equally informed by multiple modalities and tracks cell types changing over time. From this, we expect new insights into changes (and stability) of RGC types in healthy vs. degenerating retina. These insights will help to understand the brain’s “building blocks”, and their susceptibility and resilience to disease and aging.

• DFG Programme: Research Grants
• International Connection: Israel
• International Co-Applicant: Professor Shahar Alon, Ph.D.