Skip to main content

Long-term fixation impact on archived human nervous tissues for sequencing-based transcriptomics.

Howe S, McCann H, Xiong Z, Chang J, Steyn FJ, Ngo ST, Nguyen Q

Publication: Brain communications (2025) • Vol 7 • Page fcaf428PMID: 41262414View on PubMed

Purpose

This paper compared RNA integrity, single-cell transcriptomics (Chromium) and spatial transcriptomics profiles among autopsy hypothalamic specimens that were fixed for 2 days or >6 years  (>2,520 days) and autopsy lumbar spinal cord specimens fixed for 14 days or >6 years (>2,352 days). Specimens were obtained after a postmortem interval of 9-57 h and, in most cases, both specimen types or fixation durations were not available.

Conclusion

RNA isolated from spinal cord specimens that were fixed for 14 days had a lower RIN than hypothalamic specimens fixed for 2 days (P=0.005), but the DV200 was only insignificantly lower and DV200 and RIN were not affected by fixation duration with a linear mixed effects model. Spinal cord specimens fixed for 14 days had higher median reads per nuclei (P=0.003), fewer median genes per nuclei (P=0.003), a lower median unique molecular identifiers (UMI) count per nuclei (P=0.003), and a lower percentage of reads that were confidently mapped (P=0.003) than hypothalamic specimens fixed for 2 days. In the linear mixed effects model (all specimens were included), there was a significant negative effect of fixation duration on the median number of genes per nucleus/spot (P=0.039) and the percentage of reads that were confidently mapped (P< 0.001), but not on the number of reads or UMIs per nuclei/spot.  Additionally, in the linear mixed effects model, transcriptomics with Visium (spinal cord only) versus Chromium Next Gem Flex led to higher numbers of median genes per spot (P=0.027), UMI per spot (P=0.039), and percentages of mapped reads (P=0.039), and analysis of spinal cord versus hypothalamus specimens led to a higher number of reads per nuclei/spot (P=0.049) but lower RIN (P<0.001).

The average number of nuclei or spots was lower in hypothalamic specimens fixed for >6 years than 2 days (69.25 versus 2348 nuclei/spots, respectively) and in spinal cord specimens fixed for >6 years than 14 days (34.50 versus 366.8, respectively).  Single-cell transcriptomics analysis (Chromium) of hypothalamic specimens fixed for 2 days found 7-9 clusters that represented different cell types (neurons, oligodendrocytes, oligodendrocyte precursor cells, astrocytes, microglia, ependymal motile ciliated cells and vascular endothelial cells), but only 1-2 clusters were identified in lumbar spinal cord specimens that were fixed for 14 days. In contrast, spatial transcriptomics (Visium) of spinal cord specimens fixed for 14 days had 8 clusters (Schwann cells, vascular smooth muscle cells, astrocytes, oligodendrocytes, ependymal ciliated motile cells and neurons), while those fixed for >6 years only had 3-4 clusters and no clusters were identified in spinal cord and hypothalamic specimens, respectively. There was a significant effect of fixation duration, specimen type (spinal cord versus hypothalamus), and technology (Visium versus Chromium) on gene expression variance by principal component analysis (P<0.001, P=0.005 and P=0.005, respectively). Specimens fixed for 14 days (spinal cord only) had lower expression in 7 gene markers genes relative to hypothalamic specimens fixed for 2 days. Analysis of all specimens found that longer fixation durations led to significant decreases in 8 genes, Visium rather than Chromium resulted in higher expression of 5 genes, and higher GFAP expression occurred in spinal cord than in hypothalamic specimens.

Studies (1)

Purpose

This study compared RNA integrity, single-cell transcriptomics (Chromium) and spatial transcriptomics profiles among autopsy hypothalamic specimens that were fixed for 2 days or >6 years (>2520 days) and autopsy lumbar spinal cord specimens that were fixed for 14 days or >6 years (>2352 days).  Brain (including hypothalamus) and/or lumbar spinal cord specimens were collected from 4 patients with motor neuron disease (MND), 5 patients with behavioral variant frontotemporal dementia (bvFTD), and 4 patients with FTD-MND after a postmortem delay of 9-57 hours (PMI temperature was not reported). Eight of the eleven brains collected were divided in half and  stored at −80°C and fixed for 2 days before analysis (2 days fixation). The other half of three brain specimens and all of the lumbar spinal cord specimens were fixed for 2 weeks in 15% neutral buffered formalin. Coronal sections (3 mm) of fixed brain regions were then returned 10% formalin for long-term storage (>6 years) at room temperature before paraffin embedding. Transverse sections of seven fixed spinal cord specimens were then embedded in paraffin and stored at room temperature (2week fixation), while transverse sections of the remaining four fixed spinal cord specimens were returned to 10% formalin for long-term storage (>6 years) at room temperature before paraffin embedding. RNA was extracted using the Qiagen RNeasy FFPE Kit and quality was assessed using the Agilent Bioanalyzer or RNA ScreenTape assay. For single nucleotide transcriptomics, tissue sections of 12 hypothalamic and 8 lumbar spinal cord sections were deparaffinized, libraries were constructed using a Chromium fixed RNA profiling kit and were paired-end sequenced on a NovaSeq 6000 instrument. For spatial sequencing, 5 µm sections of 10 lumbar spinal cord specimens were shipped with desiccant, processed using the Visium CytAssist spatial gene expression method and sequenced on NovaSeq 6000 instrument.

 

Summary of Findings

RNA isolated from spinal cord specimens that were fixed for 14 days had a lower RIN than hypothalamic specimens fixed for 2 days (P=0.005), but the DV200 was only insignificantly lower and DV200 and RIN were not affected by fixation duration with a linear mixed effects model. Spinal cord specimens fixed for 14 days had higher median reads per nuclei (P=0.003), fewer median genes per nuclei (P=0.003), a lower median unique molecular identifiers (UMI) count per nuclei (P=0.003), and a lower percentage of reads that were confidently mapped (P=0.003) than hypothalamic specimens fixed for 2 days. In the linear mixed effects model (all specimens were included), there was a significant negative effect of fixation duration on the median number of genes per nucleus/spot (P=0.039) and the percentage of reads that were confidently mapped (P< 0.001), but not on the number of reads or UMIs per nuclei/spot. Additionally, in the linear mixed effects model, transcriptomics with Visium (spinal cord only) versus Chromium Next Gem Flex led to higher numbers of median genes per spot (P=0.027), UMI per spot (P=0.039), and percentages of mapped reads (P=0.039), and analysis of spinal cord versus hypothalamus specimens led to a higher number of reads per nuclei/spot (P=0.049) but lower RIN (P<0.001).

The average number of nuclei or spots was lower in hypothalamic specimens fixed for >6 years than 2 days (69.25 versus 2348 nuclei/spots, respectively) and in spinal cord specimens fixed for >6 years than 14 days (34.50 versus 366.8, respectively).  Single-cell transcriptomics analysis (Chromium) of hypothalamic specimens fixed for 2 days found 7-9 clusters that represented different cell types (neurons, oligodendrocytes, oligodendrocyte precursor cells, astrocytes, microglia, ependymal motile ciliated cells and vascular endothelial cells), but only 1-2 clusters were identified in lumbar spinal cord specimens that were fixed for 14 days. In contrast, spatial transcriptomics (Visium) of spinal cord specimens fixed for 14 days had 8 clusters (Schwann cells, vascular smooth muscle cells, astrocytes, oligodendrocytes, ependymal ciliated motile cells and neurons), while those fixed for >6 years only had 3-4 clusters and no clusters were identified in spinal cord and hypothalamic specimens, respectively. There was a significant effect of fixation duration, specimen type (spinal cord versus hypothalamus), and technology (Visium versus Chromium) on  gene expression variance by principal component analysis (P<0.001, P=0.005 and P=0.005, respectively). Specimens fixed for 14 days (spinal cord only) had lower expression in 7 gene markers genes relative to  hypothalamic specimens fixed for 2 days. Analysis of all specimens found that longer fixation durations led to significant decreases in 8 genes, Visium rather than Chromium resulted in higher expression of 5 genes, and higher GFAP expression occurred in spinal cord than in hypothalamic specimens.

  • Biospecimens

    TissueBrain
    TissueSpinal Cord
  • Preservation Types

    Formalin
  • Diagnoses

    Other diagnoses
    Autopsy
  • Platforms

    RNAAutomated electrophoresis/Bioanalyzer
    RNANext generation sequencing
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    Biospecimen AcquisitionBiospecimen location
    Hypothalamus
    Lumbar spinal cord
    Biospecimen PreservationTime in fixative
    14 days (spinal cord)
    2 days (hypothalamus)
    >6 years (spinal cord and hypothalamus)
    Next generation sequencing SpecificTechnology platform
    Chromium (single cell)
    Visium (spatial)