Hidden RNA Switches in the Human Brain Reveal New Genetic Routes to Alzheimer’s and Parkinson’s Disease

For decades, scientists hunting the genetic roots of brain disease have focused almost entirely on one question: how much of a given gene is switched on or off. A sweeping new study published in Nature Genetics argues that this focus has missed an entire regulatory dimension. An international team led by researchers at Zhejiang University, Harvard Medical School, the Broad Institute and MIT has built the first cell-type-resolved atlas of alternative polyadenylation in the human brain, revealing that the tail end of messenger RNA molecules is governed by its own genetic rules and that these rules help explain risk variants for Alzheimer’s disease, Parkinson’s disease, schizophrenia and bipolar disorder that gene-expression studies could not account for.

The molecular process at the heart of the study is called alternative polyadenylation, or APA. When a gene is transcribed into messenger RNA, the transcript’s 3-prime untranslated region, or 3-prime UTR, is trimmed and capped with a polyadenylated tail at one of several possible positions. Which cleavage site the cell chooses matters enormously: the sequence, length and structure of the 3-prime UTR determine where the mRNA travels inside the cell, how stable it is, and how efficiently it is translated into protein. In neurons, which must deliver specific mRNAs to distant dendrites and synapses, this choice is especially consequential. Yet until now, APA in the human brain had never been mapped at single-cell resolution, and its genetic control remained largely unknown.

The team mined an extraordinary resource: single-nucleus RNA sequencing data from 379 postmortem human brains, part of the Religious Orders Study and Rush Memory and Aging Project, spanning roughly two million cells across seven major brain cell types, including excitatory and inhibitory neurons, astrocytes, oligodendrocytes, oligodendrocyte progenitor cells, microglia and vascular cells. Because 3-prime-biased sequencing directly captures the polyadenylation sites of detected transcripts, the researchers could identify more than 30,000 unique cleavage sites and quantify, for each gene in each cell type in each individual, the relative usage of proximal versus distal polyadenylation sites, an index known as PDUI.

The atlas immediately revealed striking cell-type signatures. Neurons, on average, carry significantly longer 3-prime UTRs than other brain cells, while microglia carry shorter ones, and these differences persisted even among genes expressed broadly across all cell types, suggesting they arise from the global regulatory environment of each cell rather than from which genes are expressed there. The team also predicted 374 RNA-binding proteins and transcription factors whose expression tracks 3-prime UTR length across individuals, including known polyadenylation regulators such as PABPN1 and CPSF7 as well as candidates like ZNF460 and EHMT2. Knockdown experiments in SH-SY5Y-derived neurons validated several of these regulators, with U2AF2, PCBP2 and PABPN1 each controlling hundreds of reproducible APA targets.

When the researchers compared individuals across the spectrum of Alzheimer’s disease, they found that APA changes were widespread but largely invisible to conventional expression analysis. More than 3,000 genes showed differential polyadenylation in late-stage Alzheimer’s, concentrated in neurons, yet the vast majority of these genes showed no change in expression at all. Genes with altered 3-prime UTR usage in neurons were enriched for pathways central to Alzheimer’s biology, including microglial activation, microtubule transport, axodendritic transport and mitochondrial function, and included established risk genes such as MAPT, ABCA7 and MEF2C. In other words, APA acts as an independent, largely orthogonal layer of dysregulation that converges on the same disease pathways through a different set of genes.

The study then connected APA to genetics by integrating the atlas with whole-genome sequencing, identifying 3-prime UTR quantitative trait loci, or 3-prime aQTLs, for 4,288 genes. These are genetic variants that bias which polyadenylation site a gene uses. The team validated two such variants, at the DHX34 and DNAJA1 loci, using CRISPR-based prime editing in HEK293T cells, finding strong correlations between editing frequency and polyadenylation usage. Notably, the aQTLs colocalized far more often with protein quantitative trait loci than with expression QTLs, implying that many of these variants act on translation or protein abundance rather than on mRNA levels, a mechanism that expression-based studies structurally cannot detect.

The disease implications are substantial. Across 17 brain traits and diseases, the researchers linked 168 genome-wide association study loci to 3-prime aQTLs, and only about 17.5 percent of these were shared with eQTLs. In Alzheimer’s disease, the PLEKHA1 and APOC2 loci showed aQTL-dependent associations; in schizophrenia, PAK6 and AP3B2; in bipolar disorder, NDUFA13; and the MTCH2 locus, associated with multiple brain traits, appears to alter protein output through stop-codon readthrough controlled by 3-prime UTR choice. Particularly striking was the APOC2 finding: a microglial aQTL within the APOE locus remained genome-wide significant even after conditioning on eight independent variants, and had no effect on APOC2 expression, marking it as an APOE-independent, expression-independent contribution to Alzheimer’s risk.

Perhaps the most dramatic result concerns Parkinson’s disease. The top Parkinson’s risk locus, near the SNCA gene that encodes alpha-synuclein, colocalized strongly with a neuronal and oligodendrocyte aQTL. By dissecting SNCA’s 3-prime UTR into four distinct polyadenylation sites, the team showed that the risk variant rs356220 promotes usage of the most proximal site, with little to no effect on total SNCA expression. Imaging experiments told the biological story: the short pA1 isoform preferentially localized to neurites, while longer isoforms remained more abundant in the cell body, and individuals with Parkinson’s disease showed increased proximal SNCA 3-prime UTR usage in excitatory neurons. The risk allele’s effect direction matched the disease-associated shift, suggesting the locus acts by rerouting SNCA mRNA within neurons rather than by changing how much of it is made.

The authors are candid about limitations. Sparse single-nucleus data introduce technical noise, statistical power was limited for some non-neuronal cell types, and further experiments are needed to distinguish direct from secondary regulator effects. But the broader message is hard to overstate: a post-transcriptional regulatory layer that is essentially free to measure from existing 3-prime-biased single-cell datasets explains a meaningful slice of brain disease heritability that expression analyses miss entirely. The researchers propose that APA profiling should become a standard complementary modality in future single-cell studies, and that 3-prime UTR-targeted therapeutics, informed by cell-type-specific aQTL maps, represent a promising new frontier for treating Alzheimer’s, Parkinson’s and other devastating neurological disorders.

Subject of Research: Single-cell mapping of alternative polyadenylation and its genetic regulation in the human brain and its role in neurodegenerative and neuropsychiatric disease

Article Title: Single-cell profiling and genetic regulation of alternative polyadenylation in the human brain

Article References: Nan, J., Boix, C. A., Shi, S., Fan, X., Ni, J., Wang, K., Shuai, X., Ding, K., Wu, P., An, Y., Sun, N., Hou, L., Chen, K., Huang, X., Li, C., Akay, L., Louderback, K., Nawaid, H., Park, Y. P., … Xiong, X. (2026). Single-cell profiling and genetic regulation of alternative polyadenylation in the human brain. Nature Genetics. https://doi.org/10.1038/s41588-026-02758-w

Image Credits: AI Generated

DOI: 10.1038/s41588-026-02758-w

Keywords: alternative polyadenylation, 3' UTR, single-cell RNA sequencing, Alzheimer's disease, Parkinson's disease, SNCA, 3'aQTL, GWAS, gene regulation, neurons, microglia, post-transcriptional regulation