Erasing a Brain RNA Modifier Slashes Alzheimer's Amyloid Plaques in Mice
Scientists identify ALKBH3, an RNA demethylase, as a key driver of Alzheimer's pathology—and show that blocking it restores memory in mouse models.
Summary
Researchers at the Chinese Institute for Brain Research discovered that an RNA-modifying enzyme called ALKBH3 is significantly elevated in Alzheimer's disease brains—both in mouse models and human patients. ALKBH3 removes a chemical tag called m1A from RNA, and its overactivity depletes m1A from a critical gene (PINK1) that governs mitochondrial cleanup. When ALKBH3 was genetically reduced in 5xFAD Alzheimer's mice, amyloid plaques dropped substantially, mitochondrial health improved, and cognitive function was restored. Conversely, artificially boosting ALKBH3 worsened amyloid burden. A pharmacological inhibitor of ALKBH3 also reduced amyloid in cell models, suggesting a druggable target for Alzheimer's therapy.
Detailed Summary
Alzheimer's disease (AD) is defined by amyloid-beta (Aβ) plaque accumulation, mitochondrial failure, and cognitive decline, yet the upstream molecular triggers remain poorly understood. This study from Li et al. (2026) takes an epitranscriptomic approach—examining chemical modifications on RNA rather than DNA—to uncover a novel disease mechanism.
Using liquid chromatography-tandem mass spectrometry (LC-MS/MS) and immunofluorescence, the team profiled RNA modifications in the hippocampus of homozygous 6-month-old 5xFAD mice, a well-validated aggressive amyloid AD model. Among all modifications screened, N1-methyladenosine (m1A) showed the most dramatic change: a greater than 3-fold reduction compared to wild-type controls. Systematic profiling of m1A regulatory enzymes revealed that the m1A demethylase ALKBH3 was upregulated more than 2-fold, providing a mechanistic explanation for m1A loss. Importantly, ALKBH3 elevation was confirmed in human AD brain single-cell RNA-seq data across multiple cell types, establishing clinical relevance.
To establish causality, the researchers generated 5xFAD mice with heterozygous Alkbh3 deletion. These mice showed restored m1A levels, significantly reduced Aβ plaques (confirmed by immunofluorescence and Western blot), and improved cognitive performance. The opposite experiment—AAV-mediated hippocampal overexpression of ALKBH3 in 5xFAD mice—exacerbated Aβ pathology, confirming bidirectional control. In cell culture, a pharmacological ALKBH3 inhibitor (HUHS015) also reduced amyloid in APP-transfected SH-SY5Y neurons.
Mechanistically, the team identified PINK1 mRNA—encoding the master regulator of mitophagy—as a direct ALKBH3 target. ALKBH3-mediated removal of m1A from PINK1 mRNA destabilizes the transcript, reducing PINK1 protein and impairing mitophagy, the cellular process that clears dysfunctional mitochondria. This mitophagic failure leads to oxidative stress, elevated Aβ production, and impaired microglial phagocytic clearance of plaques, creating a self-reinforcing pathogenic loop.
The study positions the ALKBH3–m1A–PINK1 axis as a mechanistically validated, therapeutically tractable pathway. By linking epitranscriptomic dysregulation directly to both mitochondrial dysfunction and Aβ accumulation, it offers a unifying upstream explanation for two core AD hallmarks and a rationale for targeting ALKBH3 pharmacologically.
Key Findings
- m1A RNA modification is reduced >3-fold in 5xFAD AD mouse hippocampi, the largest change among all modifications profiled.
- ALKBH3 demethylase is upregulated >2-fold in 5xFAD mice and elevated in human AD brain single-cell RNA-seq data.
- Genetic reduction of Alkbh3 in 5xFAD mice significantly decreases Aβ plaques and restores cognitive function.
- ALKBH3 targets PINK1 mRNA, erasing m1A marks and impairing mitophagy—a key driver of neuronal dysfunction.
- Pharmacological inhibition of ALKBH3 with HUHS015 reduces amyloid burden in APP-expressing neuronal cells.
Methodology
The study used LC-MS/MS to quantify RNA modifications in mouse hippocampi, immunofluorescence, Western blotting, and single-cell RNA-seq analysis of human AD brain datasets. Causal experiments employed genetic Alkbh3 heterozygous deletion crossed into 5xFAD mice and AAV-mediated hippocampal overexpression, complemented by pharmacological inhibition in neuronal cell culture.
Study Limitations
All mouse experiments used the aggressive 5xFAD model, which may not fully recapitulate sporadic late-onset human AD. Alkbh3 reduction was heterozygous rather than complete knockout, and the full transcriptome-wide consequences of m1A loss in neurons require further characterization. The pharmacological inhibitor data is limited to cell culture.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: