Rare tau gene variant reshapes brain cell transcriptomes in Alzheimer’s and PSP

A rare point mutation in the tau gene, known as A152T, leaves a distinctive molecular fingerprint in the brains of people who carry it, according to a new single-nucleus study of postmortem brain tissue. Researchers found that carriers of the variant who died with Alzheimer’s disease or progressive supranuclear palsy (PSP) showed gene-expression patterns that differed sharply from those of pathology-matched patients without the mutation, even though the cellular makeup of the tissue was essentially identical across groups. The findings, published in Acta Neuropathologica, suggest that pooling carriers with sporadic cases can mask important transcriptional signals, and they call attention to a variant that is rarely screened for in human genomic studies.

The A152T variant sits in the projection domain of the tau protein, which is encoded by the MAPT gene. Unlike many pathogenic MAPT mutations that cause familial tauopathies outright, A152T acts as a risk factor: it raises the odds of dementia with Lewy bodies roughly sixfold, frontotemporal dementia spectrum disorders including PSP and corticobasal degeneration about threefold, and Alzheimer’s disease 2.3-fold, making it the only known MAPT point mutation associated with Alzheimer’s risk. Earlier work in worms, transgenic mice and human induced pluripotent stem cells has linked the variant to neuronal hyperexcitability and impaired retrograde axonal transport, and previous biochemical studies showed that carriers accumulate soluble tau hyper-phosphorylated at a neighboring residue, a signature that distinguishes them from non-carriers.

To explore how the variant shapes disease at the level of individual cells, a team led by investigators at Mayo Clinic performed single-nucleus RNA sequencing on frozen tissue from the middle frontal gyrus, a cortical region affected relatively late in both diseases. The cohort comprised 28 donors: seven age-matched controls, six sporadic Alzheimer’s cases, six Alzheimer’s cases with A152T, four sporadic PSP cases and five PSP cases with the variant. Each sporadic case was matched as closely as possible to a carrier for pathological burden, Braak stage, comorbid alpha-synuclein pathology, APOE genotype, MAPT haplotype, age and sex. After stringent quality control, the analysis encompassed 73,441 nuclei, which clustered into excitatory and inhibitory neurons, oligodendrocytes, astrocytes, microglia, oligodendrocyte progenitor cells and endothelial cells.

A first and important observation was what the researchers did not find: using centered log-ratio transformed nuclei counts, they detected no significant differences in cell type or cluster composition between the disease groups and controls. Because the middle frontal gyrus is spared until late disease stages, this region offered a way to detect disease- and variant-associated transcriptome changes largely independent of substantial neurodegeneration. The transcriptional differences that did emerge were therefore changes in gene expression within cells, not changes in which cells were present.

Differential expression analysis, performed separately within each cell type using the MAST framework with age, sex and batch as covariates, identified excitatory neurons, oligodendrocytes and astrocytes as the most affected populations in both diseases. In Alzheimer’s tissue, excitatory neurons carried the heaviest burden: 2,787 differentially expressed genes relative to controls in sporadic cases compared with 1,898 in A152T carriers. Although 80 percent of the genes altered in carrier neurons were shared with sporadic Alzheimer’s, only 54.5 percent of the sporadic changes appeared in carriers, indicating that the variant narrows and redirects the neuronal response rather than simply amplifying it. Pathway analysis showed that downregulated genes shared by both groups were enriched for synaptic vesicle exocytosis and synaptic signaling, while genes unique to carrier neurons pointed to altered regulation of synapse organization and synaptic plasticity.

The glial results were more striking. In oligodendrocytes, sporadic Alzheimer’s cases yielded 2,534 differentially expressed genes versus 675 in carriers, with only 18.2 percent of the sporadic gene set shared with carriers. In astrocytes, the numbers were 1,503 versus 874, with just 14.1 percent overlap. Critically, when the researchers merged sporadic and carrier cases into a single disease group, the number of detectable differentially expressed genes dropped in several cell types, an effect seen again when excitatory neurons were subdivided into clusters: sporadic Alzheimer’s changes were concentrated in one cluster, carrier changes in another, and merging the groups obscured both. The authors argue that because A152T is not routinely genotyped in most cohorts, carriers are likely being folded into sporadic groups, potentially diluting signals in single-nucleus studies of Alzheimer’s disease.

In PSP, oligodendrocytes again emerged as a focal point. Sporadic PSP oligodendrocytes showed 2,121 differentially expressed genes, including strong downregulation of cytoplasmic translation and oxidative phosphorylation pathways, whereas carrier oligodendrocytes showed only 440, with heat-shock proteins such as HSPA1A and HSPA1B among the top upregulated genes, a pattern consistent with cellular stress. Only 17 percent of the sporadic gene set was shared with carriers. Upregulated genes in carriers were enriched for cholesterol and terpenoid biosynthesis, and network analysis highlighted enzymes of the mevalonate pathway, which supports myelination and myelin maintenance. Protein interactome analysis found 394 tau interactors among differentially expressed genes in sporadic PSP oligodendrocytes but only 46 in carriers, reinforcing the impression that the two groups experience distinct molecular insults.

Perhaps the most compelling cross-disease signal came from astrocytes. Comparing carriers with their respective sporadic groups, the researchers identified 20 genes differentially expressed in both Alzheimer’s and PSP carriers, including four members of the metallothionein family, MT1E, MT1F, MT1G and MT2A, all downregulated. Metallothioneins are cysteine-rich proteins that buffer zinc and copper and neutralize reactive oxygen species, and pathway analysis confirmed enrichment of copper detoxification and metal-ion homeostasis terms among downregulated genes. Donor-level pseudo-bulk analysis, an orthogonal validation approach, supported the metallothionein pattern in PSP, where all four genes were nominally significant, though none reached significance in the Alzheimer’s comparison. The authors suggest that reduced astrocytic metallothionein expression may reflect heightened susceptibility to heavy metal toxicity and oxidative stress in carriers across both diseases.

The study has important limitations that the authors address directly. The primary analysis treated individual nuclei as independent observations, which is susceptible to pseudo-replication and inflated false positives; the proportion of genes supported by donor-level pseudo-bulk analysis varied widely, from roughly 5 to 89 percent depending on cell type and disease group, and few genes survived false-discovery-rate correction, likely reflecting the small number of donors, an unavoidable consequence of the variant’s rarity. The tissue examined represents a single late-affected region and a single time point, so the sequence of neuronal and glial changes cannot be determined, and it remains unclear whether the carrier-specific signatures precede tau accumulation or arise as adaptations to it. Even so, as the first transcriptome study to compare A152T carriers directly with pathology-matched sporadic Alzheimer’s and PSP cases at single-nucleus resolution, the work delivers a clear methodological message: rare genetic risk variants can redistribute transcriptional vulnerability across neuronal and glial populations, and failing to screen for them risks averaging away biology that could ultimately inform therapeutic strategies for tauopathies.

Subject of Research: Cell type-specific transcriptional effects of the MAPT A152T tau risk variant in Alzheimer's disease and progressive supranuclear palsy

Article Title: Tau A152T risk variant drives distinct cell type transcriptional signatures in AD and PSP

Article References: Lilley, M. T., Ren, Y., Soni, D., Sanborn, E. C., Marks, J. D., Chen, Y., Yan, Y., DeTure, M., Rawlinson, B., Dickson, D. W., & Cook, C. N. (2026). Tau A152T risk variant drives distinct cell type transcriptional signatures in AD and PSP. Acta Neuropathologica, 152(1), Article 60. https://doi.org/10.1007/s00401-026-03110-6

Image Credits: AI Generated

DOI: 10.1007/s00401-026-03110-6

Keywords: MAPT, A152T, tauopathy, Alzheimer's disease, progressive supranuclear palsy, single-nucleus RNA sequencing, oligodendrocytes, astrocytes, metallothioneins, transcriptomics, postmortem brain, genetic risk variant