The regulatory code of injury responsive enhancers enables precision cell state targeting in the CNS [multiome]

Enhancer elements in the genome direct cell type-specific gene expression programs. In response to injury, cells undergo dynamic changes in gene expression that drive adaptation and initiate tissue repair. Here, we investigate how injury-induced gene expression programs are encoded within enhancer elements in the mammalian CNS. Leveraging single-nucleus transcriptomics and chromatin accessibility profiling, we identify thousands of injury-induced, cell type-specific enhancers in the mouse spinal cord. These enhancers are more abundant in glial cells and retain cell type specificity, even when regulating shared wound response genes. By modeling glial injury-responsive enhancers using deep learning, we reveal that their architecture encodes cell type specificity by integrating generic stimulus response elements with cell identity programs. Finally, through in vivo enhancer screening, we demonstrate that injury-responsive enhancers can selectively target reactive astrocytes bordering lesion sites across the CNS using therapeutically relevant gene delivery vectors. Our decoding of the principles of injury-responsive enhancers enables the design of sequences that can be programmed to therapeutically target disease related cell states. Overall design: Single nuclei from the mouse spinal cord were processed with a multiomic approach, sampling transcriptomics and chromatin accessibility information. We collected tissue from uninjured mice, as well as mice subjected to a spinal cord injury (ie., contusion, at 1, 3, 7, and 28 dpi). Nuclei were either sampled unbiased or subjected to FAC sorting to enrich for specific populations (e.g., astrocytes, ependymal cells, immune cells and perivascular cells). Each experimental condition encompassed 1-2 animals per sex.