Spatial proteomics allows scientists to map the proteome of cells and tissues in a multidimensional fashion.1 While the concept of spatial proteomics has been around for some time, recent advances in mass spectrometry and microscopy techniques have allowed for a greater depth of proteome coverage than ever before. In particular, the development of deep visual proteomics has recently added to the growing interest in this field. Researchers can now use spatial proteomics and deep visual proteomics to elucidate complex cell biology, study diseases like cancer, and even make recommendations for personalized treatments.
What Is Spatial Proteomics?
Spatial proteomics is the study of the spatial distribution of the proteins within cells and tissues. The subcellular localization of proteins is intrinsic to cellular function, making spatial proteomics an important tool in cell biology research and the study of disease.2 Unlike bulk proteomics or single-cell proteomics, researchers can use spatial proteomics to study how different proteins interact in situ in their native environment.
Matthias Mann, a biochemist, is one of the key developers of deep visual proteomics, a spatial proteomics method for exploring the spatial distribution of the proteome within individual cells.
Max Planck Institute of Biochemistry, Munich
“Spatial proteomics is the marriage between the proteomics field and the imaging field,” said Matthias Mann, a biochemist at the Max Planck Institute of Biochemistry, Munich. “That allows you to see what all the proteins do in the body directly, instead of in cell cultures or another in vitro system.”
Andreas Mund, an engineer at the University of Copenhagen who has worked extensively in high-content microscopy, recently collaborated with Mann to develop deep visual proteomics. A key breakthrough in the spatial proteomics field, deep visual proteomics allows researchers to resolve the differential expression of the proteome within individual cells. According to Mund, the strength of this untargeted technique is that it is hypothesis-free.
“We look at whatever we can detect and quantify, and then the misregulated or deregulated proteins will tell us what's wrong… We often see very unexpected changes, which you would have not guessed or come up with,” Mund commented. “That's the beauty of it: we [capture] the whole complexity of biology.”
Spatial Proteomics Methods
Spatial proteomics spans several key methods: imaging-based approaches using antibodies, mass spectrometry, and more recent techniques such as deep visual proteomics.
Antibodies and imaging-based spatial proteomics
The earliest methods for visualizing the spatial distribution of proteins are based on immunofluorescence. Using fluorophore-conjugated antibodies, researchers can label a couple of proteins in a section of tissue and capture an image.3 After stripping off the first round of antibodies, they can apply a new set of antibodies targeting different proteins.
By repeating the process several times and layering the images, researchers can visualize the distribution of the different proteins in the tissue section. However, the spectral overlap of fluorophores and issues with signal-to-noise ratio limits this labor- and time-intensive method to the detection of only a handful of proteins, and it cannot detect protein isoforms.4
More recent technologies, including codetection by indexing (CODEX), have enabled multiplexed protein detection.4 CODEX uses antibody-conjugated DNA barcodes in a one-step incubation process, after which fluorescently-labeled DNA imaging probes are added and removed in cycles for barcode readout.5 Despite the significantly reduced experimental timeframes of the CODEX method, it still carries many of the other disadvantages of using fluorescently-labeled antibodies.5
Rong Fan, a chemist at Yale University, whose research focuses on developing technologies for spatial omics, cautioned that purists do not consider these methods to be true spatial proteomics because they target just a few known proteins and cannot analyze the complexity of the broader proteome. “I just call it spatial protein profiling,” Fan added. Researchers also use DNA-barcoded probes for the simultaneous detection of many more proteins and integration with transcriptomics approaches.4 For example, cellular indexing of transcriptomes and epitopes by sequencing (CITE-Seq) employs oligonucleotide-tagged antibodies to analyze cell surface proteins and transcriptomes simultaneously at the single-cell level.6
However, all these methods are used for targeted proteomics, in which there are already proteins of interest. “To perform unbiased, targeted discovery of proteins, you have to use mass spectrometry,” said Fan.
Mass spectrometry-based spatial proteomics
High-plex protein imaging of a human lymph node tissue section. On the left panel, cell types are identified by their protein profiles. On the right, antibodies for select proteins have been used to create a fluorescence image.
Rong Fan
Mass spectrometry (MS) involves extracting proteins and digesting them so they can be identified by their unique peptide sequences.1 Researchers combine MS with imaging techniques to analyze the proteins and peptides in tissue sections without the need for labelling, a method known as mass spectrometry imaging (MSI).4 Yet this method has limitations in its spatial resolution and the number of proteins that can be analyzed.
In terms of single-cell proteomics, there has been significant progress in MS-based methods over the last few years, enabling the quantification of over 1,000 proteins in cultured single cells.7 However, this approach too lacks the depth of the full proteome, and cannot be performed in a significantly high-throughput manner.
Moreover, it doesn’t provide the crucial spatial context that Fan believes is necessary for translational research. “Tissue is very heterogeneous... If you don't resolve [the proteome] spatially, you get bulk proteomics that doesn't make sense…it cannot give you any granular information about tissue biology,” Fan said.
Deep visual proteomics
Deep visual proteomics combines artificial intelligence (AI)-based analysis of high-resolution microscopy images and microdissection with ultra-high-sensitivity MS analysis of the proteome of individual cells or nuclei.8 Using this method, researchers can analyze cellular phenotypes, dissect out individual cells from a tissue section, then profile their proteomes using MS. Finally, they can map the resulting information back to the original tissue section to create a comprehensive view of the complex and heterogeneous spatial distribution of proteins within and among cells in that tissue.7 No longer limited by the use of antibodies, and with much greater depth of coverage of the proteome, researchers use deep visual proteomics to create maps of proteomics data at the subcellular level, all within the context of native tissues.8
Rong Fan’s research is focused on developing new technologies for spatial multi-omics.
Yale University
“Deep visual proteomics basically combines the best of the imaging world with the best of the mass spectrometry-based proteomics world,” explained Mund. “We can now identify and quantify thousands of proteins from single cells or a population of cells.”
In a deep visual proteomics workflow, sections of tissue are mounted on slides and stained to mark areas of interest. The entire tissue section is scanned, followed by AI-based segmentation to outline the cells. The precise locations of the cells are exported to a laser microdissection microscope much like GPS coordinates, after which individual cells are cut out.
“We can collect these cells, and then we can start to digest the proteins to peptides, and we analyze them using mass spectrometry-based proteomics,” elucidated Mund. “At the end, when we have quantified the thousands of proteins, we can then map them back to their original position.”
Fan remarked that spatial proteomics must be combined with other omics, such as epigenomics and transcriptomics, to provide a comprehensive view of biology. Using the analogy of a computer, he said that “…if you play a PC game, proteins are the game you’re playing, but what controls it is the underlying software.”
Applications of Spatial Proteomics
The depth and resolution of spatial proteomics, and in particular deep visual proteomics, makes these technologies incredibly valuable in biomedical research. “We look at the whole picture, the whole proteome in the cells, and this tells us what is wrong in these cells: What are the proteins which are misregulated and eventually are the cause for disease, or the cause for the progression of the disease. Then we can find new or better biomarkers, or eventually, trackable proteins,” said Mund.
One of the key areas where deep visual proteomics really shines is cancer research. “Cancer cells change their morphology during disease progression or development. We can identify them, we can study them and what they look like, and then we can measure their proteins with unprecedented sensitivity,” said Mund.
Andreas Mund, of the University of Copenhagen, used his expertise in high-content microscopy to develop deep visual proteomics alongside Matthias Mann.
University of Copenhagen
In the 2022 paper that introduced deep visual proteomics to the field, Mund, Mann, and their team explored the distinct changes in the proteome of melanocytes as they transitioned into invasive melanoma cells.8 “[You have] melanoma cells that are on the very surface of your of your skin, and then about half a millimeter in, they are busy degrading the extracellular matrix, which then makes them dangerous, because they can escape and metastasize,” Mann remarked.
Along with one of their students, Mann and Mund recently applied deep visual proteomics to explore the complexities of the tumor microenvironment. They found that an immunosuppressive barrier of macrophages was preventing the infiltration of T cells in a colorectal tumor.9
“When the T cells were on the one side of [the macrophage barrier], we could tell from the proteome that they are fully functional and aggressive,” explained Mann. “But the ones that passed it were sort of neutered, I would almost say. They were not aggressive, and they couldn't do anything to the cancer cells anymore.”
Mann and Mund are excited about the applications of deep visual proteomics in personalized medicine. By identifying which proteins are dysregulated in tumors, researchers can work with clinicians to tailor treatments in the future. “You can say ‘this chemotherapy that you're planning, don't even try it, because these pathways are not upregulated, but these other ones would be good’,” said Mann.
Their teams are also using deep visual proteomics in the study of cellular senescence and aging and neurological research. In one of their projects, Mann and his team are examining the protein aggregates that are a hallmark of Parkinson’s disease. “We can see how [the cells] are trying to cope with the aggregates and trying to remove them, or it can be the aggregates between the cells,” Mann explained. “You can also cut them out and see, what the immune system is trying to do with these aggregates.”
- Wu M, et al. Spatial proteomics: Unveiling the multidimensional landscape of protein localization in human diseases. Proteome Sci. 2024;22:7.
- Lundberg E, Borner GHH. Spatial proteomics: A powerful discovery tool for cell biology. Nat Rev Mol Cell Biol. 2019;20(5):285-302.
- Schueder F, et al. DNA-barcoded fluorescence microscopy for spatial omics. PROTEOMICS. 2020;20(23):1900368.
- Fan R. Integrative spatial protein profiling with multi-omics. Nat Methods. 2024;21(12):2223-2225.
- Bollhagen A, Bodenmiller B. Highly multiplexed tissue imaging in precision oncology and translational cancer research. Cancer Discov. 2024;14(11):2071-2088.
- Nettersheim FS, et al. Titration of 124 antibodies using CITE-Seq on human PBMCs.Sci Rep. 2022;12(1):20817.
- Rosenberger FA, et al. Spatial single-cell mass spectrometry defines zonation of the hepatocyte proteome. Nat Methods. 2023;20(10):1530-1536.
- Mund A, et al. Deep Visual Proteomics defines single-cell identity and heterogeneity.Nat Biotechnol. 2022;40(8):1231-1240.
- Zheng X, et al. Deciphering functional tumor-immune crosstalk through highly multiplexed imaging and deep visual proteomics. Mol Cell. 2025.
