Integrating Spatial Transcriptomics and Proteomics in Traumatic Brain Injuries: A Multimodal Image Analysis Pipeline

Poster Session C - Sunday, April 14, 2024, 5:00 – 7:00 pm EDT, Sheraton Hall ABC

Cai L. McCann¹ (cmccann@broadinstitute.org), James W. Aspden^2,3,4,5, Ruxandra F. Sîrbulescu^2,3,4, Samouil L. Farhi¹; ¹Broad Institute of MIT and Harvard, Cambridge, MA, United States, ²Vaccine and Immunotherapy Center, Charlestown, MA, United States, ³Massachusetts General Hospital, ⁴Harvard Medical School, ⁵Edinburgh Medical School, Edinburgh, UK

Spatial transcriptomics methods are increasingly utilized to investigate gene expression profiles with sub-cellular resolution in intricate brain cytoarchitecture while avoiding the loss of local and global spatial relationships. While gene expression analysis through RNA sequencing can simultaneously evaluate hundreds to thousands of genes, assessment of protein expression is still required for validation and can aid in characterization of cellular identities and behaviors. We have developed a methodological pipeline to register and analyze data across multiple modalities ensuring high-integrity spatial registration without disruption of biological information, using both spatial transcriptomics and immunofluorescence imaging on tissue sections. We demonstrate the utility of this method in the context of traumatic brain injury (TBI), a complex pathology in a highly complex tissue. TBI is characterized by a series of inflammatory and neurodegenerative events, as well as reactive gliosis. The latter represents activation and hypertrophy of astrocytes and is associated with a significant increase in glial fibrillary acidic protein (GFAP). Studies were conducted using the 10X Genomics Xenium instrumentation and protocol on tissue sections collected from male C57Bl6/J mice (N=6) at 48 hours after controlled cortical impact (CCI) TBI. A total of 349 genes, including GFAP, were evaluated. Subsequent immunolabeling and confocal imaging of GFAP was performed on the same tissue sections. Spatial transcriptomic and proteomic images were processed in a custom MATLAB script. Our study presents a novel and scalable approach for high-throughput multimodal data measurement in TBI pathology and biomarker research, and is further applicable for other neurological disorders.

Topic Area: METHODS: Neuroimaging