Thread by @ZdenekVrozina

COVID doesn’t have to kill neurons to change brain function. It can retune how neurons regulate their messages.🧵

A new PLOS ONE paper shows that SARS-CoV-2 leaves a measurable molecular footprint in the brain - specifically in how neurons process their mRNA

A lot of COVID brain talk is inflammation / clots / imaging. This paper zooms in on something quieter - but powerful. How brain cells finish their mRNAs!

The key concept is alternative polyadenylation (APA) - the same gene can end its mRNA at different poly(A) sites - making shorter or longer 3’UTRs.

Why care? The 3’UTR is the control panel for where an mRNA goes, how long it survives, and how strongly it’s translated (RNA-binding proteins + miRNAs). Neurons lean heavily on this fine-tuning.

They reanalyzed single nucleus RNA-seq from frontal cortex, comparing controls vs COVID cases, and called poly(A) sites with a dedicated pipeline (SCAPTURE).

Their main readout is basically - How often does a cell choose a proximal poly(A) site? More proximal choice usually means shorter 3’UTRs.

And - wow. Neurons shift hard toward proximal poly(A) sites after infection. This shows up strongly in both glutamatergic and GABAergic neurons (very significant by KS test). Big result.

Meanwhile, non-neuronal cells show smaller / mixed shifts with the loudest signal in neurons.

Then they connect APA changes with gene-expression changes and ask - What biological programs sit under these shifted transcripts? Enrichment points to neuronal development/function and RNA processing/splicing themes

They also flag that some gene sets map to neurological pathways - one example in the paper’s enrichments is Parkinson disease (as a pathway term) among certain APA/expression quadrants.

Next step - they overlay Which of these altered genes overlap with known risk-gene sets? They selected 15 common neuro + psych disorders for joint analysis.
AD, PD, epilepsy, ASD, depressive disorder, bipolar disorder, ADHD, MS, ischemic stroke, schizophrenia, ALS, Huntington disease, migraine, brain aging, anxiety.

Across cell types they compile 408 genes with notable APA+expression patterns. 267 are classified as risk genes from DisGeNET/BioKA/GWAS resources.

And they say those 267 show enrichment especially for AD, PD, and schizophrenia-related traits. (Risk-gene overlap/enrichment, not COVID causes X yet)

They give concrete - this is how it could matter - examples
CALM1 downregulated + shorter 3’UTR in glutamatergic neurons.
APP increased expression. 3’UTR shifts differ by cell type (oligos vs astrocytes). They discuss how APP 3’UTR length can influence translation and AD risk context-dependently.

They also model a huge miRNA layer (thousands of predicted miRNAs targeting these genes) - but they explicitly note this part is predictive and needs wet-lab validation.

Sum:
This paper is a mechanistic regulation fingerprint in post-mortem cortex - showing that after SARS-CoV-2, neurons can flip transcript endings in a way that plausibly reshapes synaptic biology.

Chen at al., Single-cell alternative polyadenylation analysis reveals mechanistic insights of COVID-19-associated neurological and psychiatric effects. journals.plos.org/plosone/articl…

What we can’t tell from this dataset? Duration. It’s post-mortem and not longitudinal. The authors note limitations and that some findings (esp miRNA layer) are predictive. To prove persistence, you’d need time-stratified samples after recovery + disease controls.

There isn’t a single Undo button here. What papers like this suggest is a shift in mRNA regulation. That’s a network-level retuning - many small changes that can add up to noticeable brain function changes.