Gene expression is often described by measuring how much RNA is present in a cell at a given moment. But RNA levels are constantly changing as new molecules are produced and older ones are broken down. Measuring both processes can provide a more complete picture of how cells respond to their environment.

Researchers at the Perelman School of Medicine, University of Pennsylvania have developed a method called spatial NT-seq that measures RNA abundance and RNA turnover while preserving information about where cells are located in tissue.

Benchmark and validation of in vivo transgenesis-free metabolic RNA labeling for cell-type specific RNA turnover analysis in mouse brains

a, A schematic depiction of the integrated workflow of in vivo metabolic RNA labeling followed by scNT-seq2. b, UMAP visualization of cortical cells from four intraperitoneally injected UPRT transgenic mice colored by annotated cell types. DMSO injection serves as the control for the 4tU-labeled mouse, while saline injection is the control for the 4sU-labeled mouse. c, A box plot showing the 4sU- or 4tU-labeled new RNA fraction per cell for both Neurod6-positive and -negative cell-types. d, Scatter plots comparing transcriptome-wide gene-level new RNA fractions between 4sU- (y axis) and 4tU-labeled (x axis) UPRT mice (P12 CTX) in Neurod6/NEX-cre targeted (Neurod6-positive cortical Ex neurons on the left) or nontargeted (Neurod6-negative cortical cells on the right) cell types. e, Violin plot illustrating the proportion of 4sU-labeled new transcripts per cell from in vivo-labeled embryonic day 16.5 (E16.5) mouse cortical tissues compared to in vitro-labeled cultured E16.5 cortical cells.  f, A box plot (top) showing the 4sU-labeled new transcript fraction per cell for wild-type postnatal day 7 (P7) mouse cortical samples in labeling time evaluation experiments. g, Scatter plots showing NTRs (x axis, from RNA labeling) and USTRs (y axis, from RNA splicing) for gene groups with distinct gene lengths (short – blue, <10 kb; medium – yellow, 10-100 kb; long – red, >100 kb) in cortical Ex neurons. h, A scatter plot showing the NTR derived from RNA labeling and USTR from RNA splicing for cell-type-specific TFs (red, n = 124) and broadly expressed slow turnover genes (blue, n = 269) in cortical Ex neurons. 

Measuring newly made and existing RNA

Traditional RNA sequencing can show which genes are active, but it usually does not reveal how quickly RNA molecules are being produced or degraded. Spatial NT-seq combines metabolic RNA labeling with chemical changes made directly in tissue, allowing researchers to distinguish newly synthesized RNA from RNA that was already present.

The researchers used the approach to map RNA turnover across the mouse brain. They found substantial differences between brain regions, showing that RNA is produced and degraded at different rates depending on cellular location.

One region, the dentate gyrus of the hippocampus, showed particularly high RNA turnover even under normal conditions. The dentate gyrus is involved in processes such as learning, memory, and the formation of new neurons.

Tracking the brain’s response to stimulation

The team also examined how RNA regulation changed after electroconvulsive stimulation, a laboratory model related to electroconvulsive therapy used to treat severe depression.

Cells in the dentate gyrus showed especially strong changes in both RNA synthesis and degradation following stimulation. The researchers described a process called “kinetics scaling,” in which cells coordinate the production and breakdown of RNA so they can rapidly adjust their overall collection of transcripts.

This means that changes in gene activity may involve more than simply turning RNA production up or down. Cells can also alter how quickly RNA molecules are removed, providing another layer of control over gene expression.

Using machine learning to understand RNA stability

The researchers combined spatial NT-seq with computational analysis in a broader framework called in vivo Timescope. Machine learning was used to identify sequence features and regulatory factors associated with differences in RNA stability between brain regions and cell types.

The approach allows RNA abundance, synthesis, degradation, and spatial location to be examined together. This could help researchers better understand how RNA regulation changes during brain activity, development, and disease.

Because spatial NT-seq does not require genetically engineered animals, the researchers suggest that it could potentially be adapted to other biological systems, including human tissue cultures and organoid models.

By extending RNA sequencing beyond measurements of RNA abundance, spatial NT-seq provides a way to examine how quickly the transcriptome changes and how those dynamics differ across individual regions of complex tissues.

Availability – The analysis source code is available via GitHub at https://github.com/wulabupenn/spatialNT-seq.