Spatial transcriptomics is useful for understanding the molecular organization of a tissue and providing insights into cellular function in a morphological context. In order to obtain reproducible results in spatial transcriptomics, we have to maintain tissue morphology and RNA molecule stability during the image acquisition and biomolecule collection processes. Here, researchers at Waseda University developed a tissue processing method for robust and reproducible RNA-seq from tissue microdissection samples. In this method, the researchers suppressed RNA degradation in fresh-frozen tissue specimens by dehydration fixation and effectively collected a small amount of RNA molecules from microdissection samples by magnetic beads. They demonstrated the spatial transcriptome analysis of the mouse liver and brain in serial microdissection samples (100 μm in a diameter and 10 μm in thickness) produced by a microdissection punching system. Using this method, the researchers could prevent RNA degradation at room temperature and effectively produce a sequencing library with Smart-seq2. This resulted in reproducible sequence read mapping in exon regions and the detection of more than 2000 genes compared to non-fixed samples in the RNA-seq analysis. This method would be applied to various transcriptome analyses, providing the information for region specific gene expression in tissue specimens.
Evaluation of tissue lysis and RNA purification effects on RNA-seq
from ethanol-fixed tissue microdissection samples
(a) Workflow of RNA-seq from tissue microdissection samples. The microdissection samples were collected from ethanol-fixed liver tissue using a punching needle. Then, the microdissection samples were lysed by Triton-X100 or Proteinase K, followed by poly(A) RNA purification by oligo (dT) magnetic beads. (TN: Triton-X100, no RNA purification and PP: Proteinase K and RNA purification) The tissue microdissection samples were serially collected from the same mouse liver slice. (b) Electropherograms of cDNA constructed under different tissue processing conditions. (c) The number of protein-coding genes estimated from RNA-seq results. Stars indicate p-value <0.005 determined by Welch’s t-test. (d) Sequencing read proportions assessed by mapping to a reference genome. (e) Comparisons of gene expression levels obtained between fresh tissue bulk RNA and tissue microdissection samples prepared under two different conditions. TPM values were averaged from four samples in the bulk sample pool and eight samples in the microdissection sample pool. (f) Pearson’s correlation coefficients across samples in the dataset including control samples and sample obtained by each method. Box plots show the within-sample range.