Blood Metabolome Study Links Ergothioneine, Lifestyle Factors to Midlife Brain Health and Alzheimer's Risk
核心洞察
A large-scale metabolomics study of 1,068 middle-aged participants identified 14 blood metabolites significantly associated with general cognition, with ergothioneine showing the strongest protective link.
All 14 cognition-associated metabolites replicated in at least one independent cohort, and the metabolite signature correlated significantly with incident Alzheimer's disease (搜索) risk in older adults.
Lifestyle factors, particularly smoking, explained the largest portion of variance in cognition-associated metabolites, while genetics played a relatively minor role.
A comprehensive analysis of the blood metabolome in middle-aged adults has uncovered distinct metabolic signatures tied to cognitive function and brain structure, while revealing that modifiable lifestyle factors—not genetics—exert the strongest influence on these brain health-associated metabolites.
The study, published in Nature Aging, analyzed 1,387 metabolites in plasma samples from 1,082 participants of the Rotterdam Study's RSIII-2 cohort using a nontargeted metabolomics platform. After quality control, 991 metabolites were tested for cross-sectional associations with general cognition and MRI markers of brain health in 1,068 dementia-free participants.
Fourteen Metabolites Tied to Cognition, Led by Ergothioneine
Researchers identified 14 metabolites significantly associated with general cognition (FDR < 0.05) after adjusting for age, sex, BMI, and lipid-lowering medication use. Ergothioneine, a naturally occurring sulfur-containing amino acid, demonstrated the strongest association with better cognition (adjusted mean difference = 0.122, P = 4.65 × 10⁻⁷). Uridine (0.093, P = 1.0 × 10⁻⁴) and 2-deoxyuridine (0.083, P = 5.48 × 10⁻⁴) also showed positive associations.
Conversely, lower levels of seven sulfated xenobiotic metabolites—including 4-vinylguaiacol sulfate, o-cresol sulfate, and 3-acetylphenol sulfate—were associated with better cognitive performance. These metabolites clustered into three correlated groups: sulfated xenobiotics, nucleotides plus ergothioneine, and uncharacterized metabolites.
MRI Markers Reveal Distinct but Concordant Metabolic Patterns
For brain imaging phenotypes, 21 metabolites were significantly associated with total brain volume, while S-adenosylhomocysteine (SAH) alone was linked to white matter lesion volume (adjusted mean difference = 0.123, P = 1.73 × 10⁻⁴). Higher levels of three sphingomyelins, glycerophosphorylcholine (GPC), 6-bromotryptophan, and argininate correlated with larger brain volume, whereas lower levels of caffeine-pathway metabolites and hydroxylated acylcarnitines showed the same direction of association.
Although the cognition-associated and MRI-associated metabolite sets did not directly overlap, their effect patterns were significantly concordant. Regression coefficients for cognition correlated with those for total brain volume (r = 0.33, P = 4.66 × 10⁻¹¹), hippocampal volume (r = 0.27, P = 1.15 × 10⁻⁷), and white matter lesion volume (r = −0.49, P = 5.39 × 10⁻²⁴).
Replication Across Independent Cohorts and Alzheimer's Disease (搜索) Link
All 14 cognition-associated metabolites replicated in at least one of four independent tests. Nine metabolites showed significant association (P < 0.05) with concordant effect directions in the RSI-4 cohort, an independent sample of 874 older participants recruited 15 years earlier. In a US-based cohort of 512 participants from seven Alzheimer's Disease (搜索) Research Centers, seven of ten available metabolites showed significant association with at least one cognitive score, including ergothioneine, 2-naphthol sulfate, and uridine.
Critically, the cross-sectional metabolite signature of cognition correlated strongly with prospective Alzheimer's disease (搜索) risk (r = −0.73, P = 1.03 × 10⁻¹³). β-cryptoxanthin and imidazole propionate showed the largest effects on incident AD diagnosis (β-cryptoxanthin: −0.199, P = 3.45 × 10⁻⁴; imidazole propionate: 0.200, P = 3.63 × 10⁻⁴).
Lifestyle Dominates Over Genetics in Shaping Brain Health Metabolites
Using gradient boosting decision tree models, the team estimated the proportion of metabolite variance explained by genetics, gut microbiome, lifestyle, clinical factors, and medication use. Among cognition-associated metabolites, lifestyle features explained considerable variance for nine metabolites, including 2-naphthol sulfate (28.6%), o-cresol sulfate (21.0%), and ergothioneine (4.7%). Genetics explained only 1.2% of variance for 2′-deoxyuridine and 0.9% for one uncharacterized metabolite.
Smoking emerged as the dominant lifestyle factor, with higher levels of sulfated xenobiotics and lower levels of uridine and 2′-deoxyuridine associated with smoking—matching the pattern of worse cognition. Ergothioneine was associated with all tested lifestyle factors except smoking: higher levels linked to better cognition were associated with higher alcohol intake, higher education, and lower BMI.
For MRI-associated metabolites, lifestyle (11 metabolites; explained variance range 0.2–8.8%), medication use (15 metabolites; 0.2–17.7%), and clinical factors (13 metabolites; 0.6–16.2%) were the major influencing factors. Gut microbiota explained variance for four metabolites (0.7–5.1%), while genetics contributed for three (1.5–2.9%).
Antacid Medication and Ergothioneine: A Mediation Pathway
A striking finding emerged from the analysis of medication effects: antacid use was strongly associated with lower blood ergothioneine levels (effect size = −0.45, P = 4.27 × 10⁻⁹). This association was confirmed in the US replication cohort, where lower ergothioneine levels were linked to proton pump inhibitor (PPI) intake (effect size = −0.159, P < 0.002).
Mediation analysis suggested that 31.5% (95% CI: 15.5–71%) of the total negative effect of antacid medication on cognition (−0.235, P < 2 × 10⁻¹⁶) was mediated through reduced ergothioneine levels. However, the authors caution that this finding should be interpreted carefully, as the temporal sequence of exposure, mediator, and outcome cannot be ascertained in the cross-sectional dataset.
Sex-Stratified and Gut Microbiome Insights
Sex-stratified analyses revealed suggestive evidence that lower levels of bile acids, including taurocholate and glycocholate, were associated with better cognition in females, while higher N-acetyl-aspartyl-glutamate (NAAG) was linked to better cognition in males. However, these sex-specific findings did not replicate in the RSI-4 cohort.
Of the 36 brain health-associated metabolites, 22 showed significant associations with specific gut microbiota. Ergothioneine was associated with 12 microbial genera, with higher levels linked to greater abundance of Lachnospiraceae ND3007, Fusicatenibacter, and Romboutsia, among others.
Implications for Prevention and Future Research
The study provides a systematic framework for understanding how modifiable exposures shape metabolites relevant to brain health. "Our findings highlight ergothioneine as a metabolite strongly linked to cognition and incident AD risk and reveal that antacid medication may be a relevant determinant of its circulating levels," the authors write. They emphasize that while the mediation analysis does not prove causality, it "offers a potentially actionable route through which medication use could affect brain health, if confirmable in longitudinal studies."
The researchers note several limitations, including the cross-sectional nature of mediation analyses, lack of dietary data at the time of blood collection, and the need for longitudinal cognitive assessments with repeated metabolomics measurements to confirm these findings over time.
