TrkB phosphorylation in serum extracellular vesicles correlates with cognitive function enhanced by ergothioneine in humans

Dementia affects over 50 million people worldwide and lacks satisfactory disease-modifying treatments, emphasizing the need for early, preventive interventions. Nutritional compounds can reduce dementia risk, but mechanisms are often unclear. Ergothioneine (ERGO), a dietary amino acid taken up via OCTN1, has lower circulating levels in patients with dementia and correlates negatively with neurodegeneration, while supplementation improves memory in animals and some cognitive domains in humans. TrkB signaling supports neuroplasticity and cognition, and small-molecule TrkB agonists ameliorate memory deficits in preclinical models. This study asks whether ERGO deficiency impairs cognition and hippocampal neurogenesis, whether ERGO supplementation rescues these functions via TrkB phosphorylation, and whether TrkB phosphorylation in serum extracellular vesicles (EVs) tracks ERGO exposure and cognitive enhancement in humans.

Prior work shows certain food-derived compounds (astaxanthin, epigallocatechin gallate, resveratrol) can improve cognitive outcomes, but mechanisms remain uncertain. The flavone 7,8-dihydroxyflavone acts as a TrkB agonist and improves cognition in dementia models. ERGO is diet-derived, efficiently absorbed via the OCTN1 transporter, and accumulates in brain despite hydrophilicity. Lower ERGO levels are reported in dementia and correlate with hippocampal atrophy; repeated ERGO supplementation improves learning and memory in healthy and dementia-model mice, and ERGO-containing food-extract tablets enhance verbal memory in humans (Cognitrax). ERGO promotes synaptogenesis in neurons and neuronal differentiation of neural stem cells (NSCs) in vitro, potentially via neurotrophin/TrkB signaling. TrkB polymorphisms associate with dementia risk, and TrkB activation is neuroprotective and promnesic in models. However, the involvement of neurotrophin/TrkB signaling in ERGO’s cognitive benefits and the impact of ERGO deficiency on cognition were previously unclear.

Animal studies: Male ICR mice were fed an ERGO-free diet (Basal Diet 5755) from 3 to 8 weeks to generate ERGO-deficient mice; controls received a normal diet (PicoLab Rodent Diet 5053). From 9 weeks, ERGO-deficient mice were orally administered ERGO (0, 2, or 20 mg/kg) in sterile water on days 0, 2, 4, 7, 9, and 11 (three times per week for two weeks). Cognitive function was assessed with the novel object recognition test (NORT) and spatial recognition test (SRT) 14 days after start of dosing; discrimination index (DI) was computed from exploration times. Plasma and hippocampal ERGO levels were quantified by LC–MS/MS. Immunohistochemistry was performed on hippocampal dentate gyrus sections stained for doublecortin (Dcx, newborn neurons) and NeuN (neuronal nuclei) to quantify neurogenesis (Dcx+ area normalized to NeuN+ area). Western blotting of dentate gyrus lysates measured p-TrkB (Tyr816) and total TrkB, normalized as p-TrkB/TrkB. Time-course experiments evaluated ERGO (50 mg/kg) or vehicle for 4, 7, or 14 days, with NORT/SRT, Dcx/NeuN immunostaining, and p-TrkB/TrkB at each timepoint. TrkB involvement was tested by co-administering the TrkB inhibitor ANA-12 (0.5 mg/kg i.p. daily or 30 min before ERGO) with ERGO (50 mg/kg) for two weeks; effects on DI, p-TrkB/TrkB, and neurogenesis were assessed. Human clinical study: A randomized, double-blind, placebo-controlled, parallel-group trial (UMIN000034386) was conducted for 12 weeks in healthy volunteers and subjects with mild cognitive impairment (MCI) (placebo n=23; ERGO n=25). Participants were assigned by age and MMSE scores; the ERGO group took food-extract tablets containing ERGO daily (four tablets/day; total ERGO content indicated as 5 mg), and the placebo group took matched tablets. Participants avoided ERGO-containing foods (e.g., mushrooms). Serum was collected and Cognitrax cognitive testing performed at weeks 0, 4, 8, and 12. ERGO and metabolites (S-methyl-ERGO, hercynine) were quantified in serum by LC–MS/MS using HILIC separation and ESI+ MRM; AAUC for serum ERGO (incremental AUC vs baseline) was calculated across weeks 0, 4, 8, 12. EV isolation and analysis: Serum was sequentially centrifuged (1,200 × g, then 17,000 × g), filtered (0.22 μm), and ultracentrifuged (100,000 × g, 90 min). EVs were also isolated via size-exclusion chromatography (Sepharose CL-2B), combined across fractions 6–12, concentrated, and subjected to immunoprecipitation with anti-SNAP25 to enrich neuron-derived EVs. Nanoparticle size/concentration was characterized by tunable resistive pulse sensing (qNano). Western blots probed p-TrkB (Tyr816), TrkB, NT-5, SNAP25, CD63; p-TrkB/TrkB and TrkB/CD63, NT-5/CD63, SNAP25/CD63 were quantified by densitometry. Correlations between EV protein ratios and serum ERGO, blood ERGO, AAUC, and Cognitrax domain scores were assessed by Pearson’s correlation. As an in vitro validation, Neuro2a cells were transfected with TrkB-3xFlag or tdTomato; EVs from the conditioned medium were probed for CD63 and TrkB/p-TrkB/Flag to confirm TrkB cargo in neuronal EVs. Statistical analyses used t tests or ANOVA with Tukey’s post hoc tests; normality via Kolmogorov–Smirnov test; correlations via Pearson r.

- ERGO deficiency reduced systemic and hippocampal ERGO and impaired memory in mice: • Plasma ERGO was below quantification (<0.03 μM) in ERGO-deficient mice vs 1.24 ± 0.42 μM in controls. • Hippocampal ERGO was 0.23 ± 0.03 nmol/g in ERGO-deficient mice vs 2.00 ± 0.11 nmol/g in controls. • ERGO-deficient mice showed reduced discrimination index (DI) in NORT and SRT vs controls. - ERGO supplementation restored ERGO levels and rescued cognition: • Oral ERGO (2, 20 mg/kg, 2 weeks) increased plasma and hippocampal ERGO dose-dependently; 20 mg/kg restored hippocampal ERGO to control levels. • NORT/SRT DI significantly improved with 2 and/or 20 mg/kg ERGO; ERGO-deficient mice administered 20 mg/kg ERGO showed significantly higher exploration of novel/moved objects. - Hippocampal neurogenesis and TrkB activation were ERGO-responsive: • ERGO-free diet reduced Dcx+/NeuN+ percentage and p-TrkB in dentate gyrus; ERGO supplementation increased Dcx+/NeuN+ and p-TrkB in a dose-dependent manner, with total TrkB unchanged. • Time-course: ERGO (50 mg/kg) increased Dcx+/NeuN+, p-TrkB/TrkB, and DI only after 14 days, not at 4 or 7 days. - TrkB inhibition blocked ERGO’s effects in mice: • Co-treatment with ANA-12 suppressed ERGO-induced increases in NORT and SRT DI and prevented the ERGO-driven rise in hippocampal p-TrkB/TrkB and Dcx+/NeuN+. - Human RCT findings (placebo n=23; ERGO n=25): • Serum ERGO increased at weeks 4, 8, and 12 in the ERGO group; S-methyl-ERGO and hercynine remained much lower than ERGO. • Blood ERGO at week 12 was higher in ERGO vs placebo (114 ± 54 μg/mL vs 72.1 ± 42.1 μg/mL); blood values were ~100-fold higher than serum. • EV characterization confirmed typical size (~104 nm) and neuronal enrichment via SNAP25 IP; p-TrkB and TrkB were detected in SNAP25-positive EVs. • The p-TrkB/TrkB ratio in serum EVs was significantly higher in ERGO vs placebo at week 12; within the ERGO group, p-TrkB/TrkB at weeks 8 and 12 exceeded week 0. • TrkB/CD63, NT-5/CD63, and SNAP25/CD63 did not change with ERGO supplementation. - Correlations supporting biomarker utility: • Serum ERGO concentration correlated with EV p-TrkB/TrkB (Pearson r ≈ 0.177, p < 0.05); no correlation for TrkB/CD63, NT-5/CD63, or SNAP25/CD63. • In the ERGO group, EV p-TrkB/TrkB correlated with multiple Cognitrax domains: composite memory (r = 0.286, p < 0.01), verbal memory (r = 0.215, p < 0.05), processing speed (r = 0.204, p < 0.05), visual memory (r = 0.235, p < 0.05), cognitive flexibility (r = 0.332, p < 0.01), executive function (r = 0.335, p < 0.01), working memory (r = 0.256, p < 0.05); negative correlations for reaction time (r = -0.284, p < 0.01) and complex attention (r = -0.265, p < 0.01). • ERGO AAUC correlated with composite memory (r = 0.244, p < 0.05), verbal memory (r = 0.240, p < 0.05), and processing speed (r = 0.240, p < 0.05). - Baseline cognition and EV markers: • p-TrkB/TrkB did not correlate with baseline MMSE and was similar between healthy controls and MCI pre-intervention. • NT-5/CD63 and SNAP25/CD63 at baseline correlated positively with MMSE and were higher in healthy controls than MCI.

This work demonstrates that ERGO is essential for maintaining hippocampal neurogenesis and cognitive function in mice, and that supplementation restores both by increasing TrkB phosphorylation in the dentate gyrus. The blockade of ERGO’s effects by ANA-12 provides causal evidence for TrkB involvement. The temporal requirement of ~2 weeks aligns with ERGO brain distribution and the slower dynamics of neurogenesis. In humans, ERGO intake elevated the p-TrkB/TrkB ratio in serum EVs and this ratio correlated with systemic ERGO exposure and improvements across multiple Cognitrax domains, indicating that TrkB phosphorylation is associated with ERGO-induced cognitive enhancement. Detection of p-TrkB and TrkB in SNAP25-enriched serum EVs and TrkB-bearing EVs from TrkB-transfected neuronal cells supports a CNS origin for at least a fraction of circulating TrkB-positive EVs. Together, these findings link dietary ERGO, TrkB activation, hippocampal plasticity, and cognition, and propose serum EV p-TrkB/TrkB as a quantitative pharmacodynamic marker of ERGO’s cognitive efficacy.

ERGO deficiency impairs hippocampal neurogenesis and recognition/spatial memory in mice; oral ERGO supplementation restores ERGO levels, increases hippocampal TrkB phosphorylation, rescues neurogenesis, and improves cognition, effects abrogated by TrkB inhibition. In humans, ERGO elevates the p-TrkB/TrkB ratio in serum EVs, which correlates with systemic ERGO exposure and cognitive domain improvements, supporting p-TrkB/TrkB in serum EVs as a quantitative biomarker of ERGO-induced cognitive enhancement. Future studies should validate this biomarker in larger and diverse cohorts, clarify the CNS contribution to circulating TrkB-positive EVs, define the physiological roles of TrkB-loaded EVs, and refine dosing/duration strategies to optimize cognitive outcomes.

p-TrkB/TrkB in mouse serum EVs could not be reliably detected, likely due to limited serum volume, leaving the cross-species EV readout unconfirmed. The physiological roles and tissue origins of TrkB-bearing circulating EVs remain unresolved. Correlation magnitudes between EV p-TrkB/TrkB and cognitive domains were modest, and while consistent with mechanistic involvement, do not establish causality in humans. Additional validation with larger samples and standardized EV isolation/quantification methods is needed.