Key differences in RNA editing found between postmortem and living brain

By Pooja Toshniwal PahariaReviewed by Danielle Ellis, B.Sc.Jul 3 2024

In a recent study in Nature Communications, researchers examined the adenosine-to-inosine (A-to-I) nucleoside editing of postmortem and live prefrontal cortical tissues.

Researchers found that RNA editing levels were significantly higher in postmortem brain tissue compared to living tissue. Alexander W. Charney, MD, PhD, co-senior author of the study and Associate Professor of Psychiatry, Genetic and Genomic Sciences, Neuroscience, and Neurosurgery at Icahn Mount Sinai and co-lead of the Living Brain Project stated: "Understanding these differences helps improve our knowledge of brain function and disease through the lens of RNA editing modifications, which can potentially lead to better diagnostic and therapeutic approaches.”

Background

Recent research on molecular alterations in response to ischemia exposures has helped us grasp adenosine-to-inosine editing within the mammalian brain. Fresh brain tissue from living human donors enables a more accurate examination by eliminating postmortem tissue analysis confounds.

Adenosine-to-inosine editing is critical for the central nervous system's function, and incorrect control can result in neurological diseases. Deoxyribonucleic acid (DNA) is stable over long postmortem periods, but ribonucleic acid (RNA) is more vulnerable. The distinction between the living and postmortem central nervous system (CNS) tissues is critical for understanding brain illness and aging.

About the study

The present study investigated adenosine-to-inosine editing changes in human living and postmortem dorsolateral prefrontal cortices (DLPFC).

The researchers proposed that molecular reactions to ischemia exposures and innate immunological responses might modify the adenosine-to-inosine editing landscape in the postmortem brain. Using live Brain Project (LBP) data, they investigated the impact of postmortem vs. live DLPFC tissues on Alu editing activity. They analyzed genetic data from 164 alive individuals and 233 partially-matched postmortem DLPFC samples. They calculated an Alu editing index (AEI) for each study sample.

The researchers conducted a transcriptome-wide comparison study to determine how much of the worldwide Alu editing variation is explained by biological and technological variables. They undertook two further studies to investigate the impact of PMI and RNA degradation changes on Alu editing in live and postmortem tissues.

Researchers investigated RNA editing in living and postmortem DLPFC samples, sequencing 206,568 single nuclei from 21 postmortem and 31 living tissues. They also created pseudo-bulk pools for each cell type per donor and examined adenosine deaminases acting on RNA (ADAR) enzyme expression in live-type vs. postmortem DLPFC. They cataloged high-confidence RNA sites using two complimentary site-calling approaches and extensive detection-based criteria to prevent false positives.

The researchers used gene set variation analysis (GSVA) to discover biological pathways that might explain postmortem biases in RNA editing. They calculated single-sample scores for 10,493 gene-ontology biological processes for each bulk RNA-seq sample and plotted them against the AEI to discover predicted biological processes.

The researchers next investigated RNA editing quantitative trait loci (edQTLs) by identifying single-nucleotide polymorphisms (SNPs) that potentially alter adenosine-to-inosine editing levels in 195 postmortem-type and 155 living DLPFC tissues. They conducted two cis-edQTL studies to match adenosine-to-inosine editing levels to SNPs and an interaction analysis to examine context-dependent effects in live and postmortem tissues.

Results

The study found considerable changes in adenosine-to-inosine editing patterns between live and postmortem brains, especially in non-neuronal cells. The team noted enhanced universal Alu editing in the prefrontal cortices of postmortem samples, with significantly elevated AEI compared to the live DLPFC. They found increased ADAR, adenosine deaminase RNA-specific B1 (ADARB1), and ADARB2 levels in the postmortem dorsolateral prefrontal cortex. The ADAR gene ranked 15^th among differently expressed genes in the postmortem samples and was robustly associated with AEI.

Differences between postmortem and living tissues accounted for the most variability in Alu editing (72%). At the same time, other established factors, such as medical diagnosis, brain banks, predicted neuronal cellular percentages, RNA integrity (RIN), and extended postmortem intervals (PMI), explained the least. The secondary postmortem investigations revealed modest relationships between PMI and AEI, indicating that prolonged PMI is not likely to cause increased Alu editing in postmortem tissue.

The study discovered 193,195 editing sites per sample in live DLPFC and 295,343 sites across postmortem tissues, indicating ADAR-mediated RNA editing. The sites were A-to-I, mapped to Alu elements, were well known, had modest editing levels, and were frequently mapped to introns and 3′ UTRs.

The team also found considerable overrepresentation of LIV-PM sites, which comprised 15–31% of all A-to-I sites and have high editing levels. In total, 1,688 biological activities were positive predictors of global Alu editing, with genes associated with these processes distinguishing alive from postmortem samples and strongly predicting changes in the AEI.

Conclusion

The findings indicate that early biological reactions to human mortality, such as IFN-γ signaling and hypoxia, boost the expression of ADAR and ADARB1, resulting in a coordinated rise in transcriptome-wide adenosine-to-inosine editing. Postmortem brain tissues exhibit increased ADAR and ADARB1 expression and extensive adenosine-to-inosine editing compared to living DLPFC.

The study presents a novel approach for prioritizing sites critical for brain function. It shows genetic variations with varying impacts on adenosine-to-inosine editing levels in postmortem and live DLPFC. Living-biased-type sites are abundant in A-to-I sites, which show strict spatiotemporal control during brain development and are associated with neurological diseases.