Pinpointing a spatial address for RNA profiles in tissues

Monitoring messenger RNA in cells is a way to gather data that provides many biological insights. Improvements in the methods for detecting cellular mRNA are helping to provide more-detailed pictures of gene expression in the cells of a particular tissue. Writing in Science, Rodriques et al.¹ describe a new approach to assessing cellular RNA that they term Slide-seq. This versatile technique couples high-throughput RNA sequencing with a way to capture spatial information about the location of the analysed cells in tissues.

Scientists often try to tackle the complexity of a biological system by breaking it down to focus on cells, and cataloguing these individual units on the basis of their identity. Conventional microscopic imaging of cells in a tissue can provide much information, such as cell type and function, but imaging data of many cells can still lack a high level of detail about cellular features — the characteristics known as the cellular phenotype.

The development of techniques for rigorously analysing the molecular contents of individual cells offers an alternative way to capture thousands of cellular features and generate an unbiased picture of the cells in a given tissue. For example, single-cell RNA sequencing has been rapidly and widely adopted since its emergence ten years ago². The use of microfluidic-based technologies in this approach has brought huge efficiency gains and cost reductions^3–5. However, single-cell RNA sequencing requires tissue disruption and cell destruction, which causes loss of spatial information about cellular location that would be valuable for cell-type identification.

Efforts to retain spatial information are a current focus in the development of methods for single-cell analysis. For example, a method called multiplexed in situ hybridization, along with certain sequencing techniques, has enabled RNA to be monitored on a subcellular scale in intact tissues^6–8. However, these approaches require substantial technical expertise, which has limited their widespread adoption.

Rodriques and colleagues’ approach is conceptually rooted in an earlier³method for single-cell RNA analysis. In that technique, called Drop-seq, the tissue being analysed is disrupted to separate the cells, which are then loaded onto a microfluidic device. Individual cells are encapsulated in nanolitre-scale droplets together with a microbead coated with copies of a DNA sequence that provides a unique ‘barcode’ for the identification of material from each droplet (which is assumed to contain a single cell). The cells are broken open, which allows the cellular RNA to bind to the DNA barcodes. A step called reverse transcription generates DNA corresponding to the captured RNA sequences, and this DNA is tagged with the barcode DNA. DNA sequencing then enables all of the RNA sequences associated with a particular barcode sequence to be deduced. Drop-seq provided a huge increase in the numbers of individual cells that could have their RNA sequences determined in a standard experiment.

With Slide-seq, Rodriques and colleagues developed an innovative twist on the Drop-seq approach by assembling the DNA-barcoded microbeads in a layer on a glass slide (Fig. 1). DNA sequencing of this material on the slide allowed the authors to determine the position and sequence of the barcodes corresponding to each microbead. A frozen tissue sample was then placed on the microbead layer, and RNA from the cells was captured by the underlying beads. As with Drop-seq, in subsequent DNA sequencing steps, the unique barcodes allowed each RNA transcript to be traced back to its bead of origin, and hence allowed assembly of a bead-specific transcriptional profile. The ability to assign each RNA transcript to a specific microbead location enabled the RNA profiling information to be presented at a spatial resolution on the scale of individual cells in a tissue sample. Rodriques et al. confirmed that Slide-seq worked effectively for different tissue types, including mouse brain, liver and kidney, and for specimens of the human brain region called the cerebellum.

Slide-seq might not capture the full profile of RNA in single cells. For example, the RNA molecules bound to a microbead are not necessarily representative of an entire cell if only part of the cell is captured by a microbead. Alternatively, if a microbead is positioned at the boundary between two cells, RNA from both cells might be recorded. Therefore, to infer the single-cell identity corresponding to each microbead profile, the authors turned to more-comprehensive single-cell transcriptional data collected in previously published single-cell RNA sequencing work. Using these data sets, the authors constructed a cell-type ‘dictionary’ to infer cell identity from the Slide-seq data. This strategy required the development of a computational approach to reconstruct each Slide-seq bead gene-expression profile, based on a weighted combination of the cell-type transcriptional signatures detected in a corresponding single-cell RNA-sequencing experiment.

Rodriques et al. conducted a range of experiments that showcased Slide-seq’s power. They investigated the RNA profiles of cells in the hippocampal region of the mouse brain, using a strategy to capture the 3D structure of this region by analysing 66 sequential tissue sections. The authors generated RNA profiles of 1.5 million microbeads, demonstrating that Slide-seq can be scaled up as necessary. Their method identified gene-expression patterns that could be used to assign distinct cerebellar cell types together into spatially defined subpopulations in the mouse brain. Such expression of many genes is spatially dependent and is independent of cell type, but this valuable information is usually lost in conventional single-cell profiling approaches during the cell-dissociation step⁹. The authors also characterized cellular transcription dynamics in a mouse model of traumatic brain injury, revealing that an initial wave of cell proliferation after injury was followed by cellular differentiation.

One limitation of Slide-seq is that information about the size and shape of each cell and other key physical properties is not determined, unlike the case for RNA-profiling methods that use a microscopy technique called histology. A Nature paper¹⁰ published in March offers a way to capture such cellular features while also tracking the subcellular localization of mRNA — this technique to image RNA transcripts in cells in tissues is called seqFISH+. Previous approaches similar to seqFISH+ were restricted by the optical-resolution limits of microscopes and the high density of mRNA transcripts in a cell. SeqFISH+ overcame these problems through a method that sequentially uses probes that provide an imaging palette of 60 colours as a tool with which to monitor individual mRNA molecules. This considerable palette breadth allows each mRNA to be localized as a dot in an individual cell using a confocal microscope, enabling the expression of about 10,000 genes to be visualized.

Together, Slide-seq and seqFISH+ represent valuable and complementary tools to aid major advances in efforts to retain spatial information during transcriptional analyses of cells. A key feature of these tools lies in their relative ease of use. Before the development of SeqFISH+, similar types of approach required a complex imaging system called super-resolution microscopy. Slide-seq is perhaps even more accessible, with experimental costs of just a few hundred US dollars and only about 40 hours of work required to prepare the materials needed for the sequencing step. It is an exciting time in this field. A commitment to create access to these technologies will facilitate their adoption and continued development, placing them at the heart of current efforts to construct high-resolution cellular atlases.