Brain Lithium Deficiency Emerges as an Early Driver of Alzheimer's Disease
A Harvard study finds endogenous lithium drops in the brain before Alzheimer's symptoms appear, and restoring it prevents disease in mice.
Summary
Researchers at Harvard Medical School measured 27 metals in human brain tissue and found lithium was uniquely depleted in the prefrontal cortex of people with mild cognitive impairment (MCI) and Alzheimer's disease (AD). Amyloid plaques were shown to sequester lithium, reducing its bioavailability. In mouse models, cutting dietary lithium by ~50% in the brain accelerated amyloid deposition, phospho-tau accumulation, neuroinflammation, synapse loss, and cognitive decline. These effects were mediated partly through GSK3β activation. Single-nucleus RNA sequencing revealed that lithium deficiency produces transcriptomic changes overlapping with human AD. Crucially, replacing lithium using lithium orotate—a salt with low amyloid binding—prevented pathology and memory loss in both AD mouse models and aging wild-type mice.
Detailed Summary
Alzheimer's disease is defined by amyloid plaques and phospho-tau tangles, but the earliest molecular triggers remain poorly understood. This landmark Nature study from the Yankner lab at Harvard proposes that disruption of endogenous lithium homeostasis is one such early event, potentially preceding and promoting classic AD pathology.
Using inductively coupled plasma mass spectrometry (ICP-MS), the investigators measured 27 metals in the prefrontal cortex (PFC) and cerebellum of 133 cognitively normal, 58 MCI, and 94 AD participants from the ROSMAP cohort, plus an independent replication cohort. Of all metals analyzed, lithium was the only one significantly reduced in the PFC at the MCI stage—before full AD diagnosis—and remained depleted in AD. Cerebellar levels and mean serum lithium were not significantly altered, pointing to a brain-specific deficit. Strikingly, laser ablation ICP-MS showed that amyloid plaques concentrate lithium 3–4 fold relative to surrounding tissue, and subfractionation confirmed that the non-plaque cortical fraction is lithium-depleted in AD. This amyloid sequestration effect was reproduced in J20 APP-transgenic mice and was absent before plaque onset, establishing a mechanistic link between amyloid accumulation and reduced lithium bioavailability.
To establish causality, the team fed wild-type, 3xTg, and J20 mice a chemically defined lithium-deficient diet, achieving ~50% cortical lithium reduction matching the human deficit. In AD mouse models this produced a 3–4-fold increase in hippocampal phospho-tau (both early pSer202 and advanced pSer396/Ser404 epitopes), accelerated amyloid plaque burden, thioflavin S-positive neurofibrillary tangle-like structures, pro-inflammatory microglial activation, and loss of synapses, axons, and myelin. Ageing wild-type mice on the deficient diet also showed elevated Aβ42 and significant memory impairment in Morris water maze and novel-object recognition tests. These pathological changes were evident as early as five weeks into the diet. Mechanistic studies identified GSK3β activation as a key mediator, consistent with lithium's well-known role as a GSK3β inhibitor at physiological concentrations.
Single-nucleus RNA sequencing of lithium-deficient mouse brains revealed transcriptomic changes across neurons, astrocytes, microglia, and oligodendrocytes that significantly overlapped with gene-expression signatures from human AD brain tissue, suggesting that lithium deficiency recapitulates a broad molecular landscape of the disease.
Critically, the study demonstrated therapeutic potential. Lithium orotate—a lithium salt shown to bind amyloid less avidly than lithium chloride—restored cortical lithium to physiological levels and prevented amyloid deposition, phospho-tau accumulation, neuroinflammation, and memory loss in both AD mouse models and aging wild-type mice, without reaching the high serum concentrations associated with lithium toxicity in psychiatric use. This positions sub-therapeutic, physiological-range lithium supplementation as a plausible preventive or disease-modifying strategy.
Key Findings
- Lithium was the only metal significantly depleted in the prefrontal cortex at the MCI stage, preceding full AD diagnosis.
- Amyloid plaques sequester endogenous lithium 3–4 fold, reducing its bioavailability in affected brain regions.
- Dietary lithium depletion (~50% cortical reduction) accelerated amyloid, phospho-tau, neuroinflammation, and synapse loss in AD mice.
- Lithium deficiency produced AD-like transcriptomic changes across multiple brain cell types, overlapping with human AD signatures.
- Lithium orotate supplementation at physiological doses prevented AD pathology and memory loss in mouse models and aging wild-type mice.
Methodology
ICP-MS quantified 27 metals in human postmortem PFC and cerebellum from ROSMAP and a replication cohort (total >400 cases spanning NCI, MCI, and AD). Causal mechanisms were tested in 3xTg and J20 AD mouse models and aging wild-type mice fed a chemically defined lithium-deficient diet, with behavioral, histological, biochemical, and single-nucleus RNA-seq readouts. LA-ICP-MS and cortical subfractionation localized lithium to amyloid plaques.
Study Limitations
The study is primarily mouse-model-based for causal claims; human data are observational and cross-sectional, limiting causal inference in people. Dietary lithium depletion in mice achieves a ~92% reduction in intake to model a ~50% cortical deficit, which may not precisely reflect human lithium insufficiency patterns. Long-term safety and efficacy of lithium orotate supplementation in humans have not yet been established in randomized trials.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.