Refining Spatial Proteomics by Mass Spectrometry: An Efficient Workflow Tailored for Archival Tissue

BackgroundFormalin-fixed, paraffin-embedded (FFPE) tissue remains the gold standard for extensively archiving biological specimens, providing biobanks with large repositories of retrospective potential. However, while formalin crosslinking is effective at preserving tissue, it poses significant challenges for extracting molecular information, including the proteome. Traditionally, this process required high levels of input material, which, in turn, limited the ability to preserve cell-type heterogeneity and spatial information. To address these limitations, we developed an easily adaptable and highly efficient workflow for extracting deep proteomes from low-input materials, such as biopsies used in routine histopathological diagnostics. MethodsWe compared the extraction efficiency of pancreatic acinar cells identified in FFPE tissue samples stained with conventional hematoxylin-eosin (H&E) against that of cells isolated from tissue samples immunostained for the epithelial cell adhesion molecule (EpCAM) across material inputs ranging from 1,166 to 800,000 {micro}m2 (estimated to 2 to 1,310 cells in volume). Cells were isolated using laser capture microdissection and subsequently analyzed using Liquid Chromatography-Tandem Mass Spectrometry. ResultsSimilar yields for both methods were observed, with EpCAM-positive cells yielding slightly higher results--approximately 1,200 unique protein groups at the lowest input and up to 5,900 at the highest. In cells isolated from H&E-stained tissue, [~]900 to [~]5,200 protein groups were identified. We decided that the optimal balance for our workflow, ensuring maximum protein identification while minimizing input material, lies within the range of approximately 50,000 to 100,000 {micro}m2. With these results, we tested spatial capabilities and biological relevance by isolating cancer cells from biopsies of pancreatic cancer, lung cancer, or glioblastoma, with the first two being stained with EpCAM and the latter being stained against the tumor-suppressor protein p53. We successfully identified tissue-specific protein expressions and observed prominent clustering of all cell populations. DiscussionOur results highlight the feasibility of performing spatial proteomics on FFPE tissue using minimal input material. This adaptable methodology opens up possibilities for investigating cell-type-specific biology while preserving spatial and histological information.