‘Slide-tags’ Method Sharpens Spatial Transcriptomics

Spatial transcriptomics may have just achieved single-cell resolution. Researchers led by Evan Macosko, Fei Chen, and colleagues at the Broad Institute in Cambridge, Massachusetts, bound together spatial gene expression and single-cell transcriptomics to a closeness not achieved before.

In the December 13 Nature, they described their new technology, “Slide-tags,” whereby thousands of unique DNA barcodes coating a microscope slide tag cells in a thin tissue slice. The nuclei are then isolated, their transcriptomes determined, and each cell traced back to their location based on which barcode it had absorbed. Slide-tags identified receptor-ligand pairs in adjacent cells within tonsil tissue, found distinct microenvironments within a solid tumor, and located specific cell types in different cortical layers of the human brain.

“This work fills a substantial gap in the field: in-depth data about the molecular content of single cells can now be mapped at high resolution,” Patrik Ståhl of the Royal Institute of Technology in Sweden wrote in an accompanying editorial. Andrew Yang, University of California, San Francisco, found the technology exciting. “This … brings high-throughput workflows of single-cell analyses into spatial genomics,” he wrote (comment below).

Previous iterations of spatial transcriptomics have been used to map what happens around amyloid plaques, and to define which subtype of dopaminergic neurons are prone to perishing in people with Parkinson’s disease (Feb 2023 news; May 2022 news). But those methods had low resolution.

In those experiments, frozen tissue is placed on a microscope slide coated with microscopic polymer beads. Each bead is tagged with a unique oligonucleotide barcode. When the tissue thaws, cells degrade, and RNA slips out to bind beads. The tissue is removed, and the RNA quantified to map gene expression at 10-micron resolution, the size of each bead. “The trouble when you sample any 10 microns in the brain is that while one cell will predominate, many others in the neighborhood will also donate RNA,” Macosko explained. Brain cells can be as small as 2 microns in diameter.

To increase the resolution, co-first authors Andrew Russell, Jackson Weir, and Naeem Nadaf turned spatial transcriptomics on its head. Rather than having RNA bind barcodes on beads, they released the barcodes to seep into cells of a tissue slice. Ultraviolet light did the trick, since the nucleotide tags were attached with a photocleavable linker. The tissue was then removed, the nuclei isolated, the transcriptomes analyzed, and the cells mapped back to the bead array (image below). The advantage of this method is that the nuclei, not the beads, are tagged, enabling the researchers to be certain which cell contained the RNA at any location.

Tagging Cells. During Slide-tags, a thin tissue slice placed on a bead array absorbs ultraviolet light, freeing DNA barcodes to slip into its cells. The tissue is removed, nuclei sequenced, and gene expression mapped to the bead array based on the barcodes each nucleus had absorbed. [Courtesy of Ståhl, Nature, 2023.]

Russell and colleagues first tried Slide-tags on a 20-micron-thick slice of wild-type mouse hippocampus. From a 3 mm2 piece, they isolated and sequenced about 1,660 nuclei, then traced 840 back to their original locations in the tissue slice. Why only half? Macosko said they detected few to no barcodes from some of those nuclei, perhaps because they did not reach cells near the top of the tissue slice. He doesn’t foresee this being a problem. “Scientists can scale up analysis to gather more nuclei, and there is no bias in nuclei loss since we see all the cell types that we normally do with regular RNA-Seq,” Macosko said. They found 10 types of cells, including neurons, microglia, and endothelial cells.

What about human brain? The researchers analyzed prefrontal cortex tissue from a 78-year-old healthy woman. A 100 mm2 sagittal slice yielded almost 17,500 nuclei mapped to the six cortical layers. Gene expression delineated seven types of excitatory neuron and 14 inhibitory, as well as three populations of astrocyte: those in the white matter, gray matter, and the bushy protoplasmic astrocytes that wrap their endfeet around blood vessels to create the glial limitans of the blood-brain barrier (image below).

Mapping Single Cells. Within a slice of human prefrontal cortex (top left), 17,500 cells were traced back to their original locations (top right). They included seven subtypes of excitatory neuron in cortical layers 2 to 6 (bottom left), 14 subtypes of inhibitory neuron (bottom middle), protoplasmic astrocytes and astrocytes from gray and white matter (bottom right). [Courtesy of Russell et al., Nature, 2023.]

To see if Slide-tags can analyze dense immune tissue—a shortcoming of current spatial transcriptomic methods—the scientists assessed a human tonsil. From a 7 mm2 slice, they recovered a whopping 81,000 nuclei, sequencing and mapping about 5,800 of those. They spotted areas rich in B or T cells, laid out as expected for the tonsil.

The scientists wondered if Slide-tags would map receptor-ligand pairs based on the proximity of cells. Indeed, after searching for the expression of known pairs in tonsil tissue, they found 645 of them in cells adjacent to one another, many of which were involved in B cell maturation.

Macosko noted that RNA-Seq isn’t the only thing researchers can do with the barcoded cells. Rather, they are amenable to any type of single-cell assay. As a proof-of-concept for spatial multiomic analysis, the researchers Slide-tagged a section of a metastatic melanoma sample, then used both snRNA- and ATAC-Seq to correlate gene expression with chromatin accessibility. Both were heterogeneous throughout the melanoma. Different gene expression and T cell responses mapped to a second section indicated distinct microenvironments in adjacent areas within the solid tumor.

As for next steps, Macosko says Slide-tags can be combined with traditional spatial transcriptomics methods by mixing original and photocleavable beads on the slide. This would improve resolution even further.

Macosko will use Slide-tags to analyze brain biopsy samples obtained during shunt surgeries for hydrocephalus (Jul 2023 news; Jul 2023 news).

He is also working with Steve McCarroll at the Broad Institute to analyze 250 brain samples from healthy adults, using this method as part of the next iteration of the human brain atlas (Oct 2023 news). Last week, Macosko and colleagues across the U.S. published the most extensive mouse brain atlas, though they used older methods for spatial transcriptomics) (Dec 2023 news).—Chelsea Weidman Burke

Comments

  1. This exciting technology brings the depth and established workflows of single-cell analyses into spatial genomics. With the understanding that spatial context is key to understanding biology, genomics tools have begun mapping molecules in situ. Alas, these emerging spatial genomics technologies often lack the depth, breadth, and quantitative frameworks that more established single-cell techniques have.

    The authors' Slide-tags technology cleverly barcodes the location of nuclei in tissue before dissociating them for established single-cell analyses. This way, they can capture quantitative, genome-wide transcriptome data with location data, whereas most alternative spatial techniques would fall short.

    In short, the best of both worlds: the depth and established workflows of single-cell analyses while retaining spatial context.

    View all comments by Andrew C. Yang

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Higher-Resolution Spatial Transcriptomics Maps Mayhem Near Plaques
  2. Single-Cell Sleuthing Nabs Neurons Prone to Perish in Parkinson’s
  3. Fresh Brain Every Friday: Biopsies Transform Alzheimer's Science
  4. Cortical Biopsies Hint at Start of Alzheimer's 'Cellular Phase'
  5. Behold, the Human Brain Like Never Seen Before
  6. New Atlas Charts Mouse Brain in Exquisite Detail

Further Reading

Primary Papers

  1. Russell AJ, Weir JA, Nadaf NM, Shabet M, Kumar V, Kambhampati S, Raichur R, Marrero GJ, Liu S, Balderrama KS, Vanderburg CR, Shanmugam V, Tian L, Iorgulescu JB, Yoon CH, Wu CJ, Macosko EZ, Chen F. Publisher Correction: Slide-tags enables single-nucleus barcoding for multimodal spatial genomics. Nature. 2023 Dec 18; PubMed.
  2. Ståhl PL. Gene expression of single cells mapped in tissue sections. Nature. 2024 Jan;625(7993):38-39. PubMed.

Annotate

To make an annotation you must Login or Register.