Spatially resolved 3D metabolomic profiling in tissues

Abstract

Spatially resolved RNA and protein molecular analyses have revealed unexpected heterogeneity of cells. Metabolic analysis of individual cells complements these single-cell studies. Here, we present a three-dimensional spatially resolved metabolomic profiling framework (3D-SMF) to map out the spatial organization of metabolic fragments and protein signatures in immune cells of human tonsils. In this method, 3D metabolic profiles were acquired by time-of-flight secondary ion mass spectrometry to profile up to 189 compounds. Ion beams were used to measure sub–5-nanometer layers of tissue across 150 sections of a tonsil. To incorporate cell specificity, tonsil tissues were labeled by an isotope-tagged antibody library. To explore relations of metabolic and cellular features, we carried out data reduction, 3D spatial correlations and classifications, unsupervised K-means clustering, and network analyses. Immune cells exhibited spatially distinct lipidomic fragment distributions in lymphatic tissue. The 3D-SMF pipeline affects studying the immune cells in health and disease.

Principles of Spatial Transcriptomics Analysis: A Practical Walk-Through in Kidney Tissue

“…In this manuscript, we have reviewed the biological and computational basis of spatial transcriptomics analysis, with an example of cell type mapping and downstream applications in kidney tissue. Spatial transcriptomics technologies are evolving at a rapid pace and extending beyond transcriptomics into metabolomics and proteomics ( Lundberg and Borner, 2019 ; Ganesh et al, 2021 ; Yuan et al, 2021 ). The integration of unbiased spatial omics technologies will provide a powerful set of tools to characterize disease processes in intact tissue ( Dries et al, 2021a ).…”

Metabolomics-Guided Elucidation of Plant Abiotic Stress Responses in the 4IR Era: An Overview

“…For instance, the development of novel mass spectrometry imaging software, in addition to several existing examples of software such as MSiReader [ 110 ], with ML capabilities to predict mass accuracy of collected data, will enable reduction in molecular ambiguity, enhanced data quality and interpretability. Machine learning dimension reduction methods, such as t -distributed stochastic neighbor embedding (t-SNE) and uniform manifold approximation and projection (UMAP), enable visualization of the large complex image spectral data that will aid in identifying similar pixel or image clusters and data interpretation [ 102 , 111 ]. Cloud computing platforms such as METASPACE and OpenMSI, have enabled the construction of imaging MS libraries that contribute to rapid and accurate metabolite identification by resolving ionization pathways and integrating all signals corresponding to a particular metabolite, and data analysis, thus increasing the widespread use of MSI [ 102 ].…”

Single‐Cell Metabolomics by Mass Spectrometry: Opportunities and Challenges

“…Recent advances in the metabolomics of tissue embedded single cells have opened the door to explore these spatial arrangements (Opportunity 6) . These methods grew out of the many modalities of MS imaging (MSI), including SIMS, MALDI, and f‐LAESI [35–39] . These methods represent different tradeoffs between spatial resolution (∼0.1 μm, ∼5 μm, and ∼30 μm, for SIMS, MALDI, and f‐LAESI, respectively), degree of ion fragmentation (SIMS > MALDI ≈ f‐LAESI), and the complexity of sample preparation (MALDI > SIMS > f‐LAESI).…”

“…Ion sources must be able to efficiently produce ions from the metabolites and lipids in the cell. Several options have demonstrated single cell capabilities, including SIMS, ESI, MALDI, f‐LAESI, and LDI from silicon nanopost arrays (NAPA) [6,15,23,35–37,49–50] . High ionization efficiency goes a long way to identify metabolites even in microbial cells.…”