J-147

curcumin-derived neuroprotective

J-147 (curcumin-derived neuroprotective) is a cognitive & mood research compound. Derived from curcumin but a distinct molecule with far better brain penetration. Its target was identified as mitochondrial ATP synthase — modestly inhibiting it triggers an adaptive stress response that raises mitochondrial calcium buffering and downstream NGF and BDNF. It was selected in ageing-brain models rather than amyloid models, which is why its profile reads differently from the Alzheimer's drugs it sits beside.

How J-147 works

Derived from curcumin but a distinct molecule with far better brain penetration. Its target was identified as mitochondrial ATP synthase — modestly inhibiting it triggers an adaptive stress response that raises mitochondrial calcium buffering and downstream NGF and BDNF. It was selected in ageing-brain models rather than amyloid models, which is why its profile reads differently from the Alzheimer's drugs it sits beside.

Proposed benefits

Researched for focus, memory, neuroprotection, mood and stress resilience.

Where to get J-147

The evidence for J-147

Graded by what exists behind each claim.

✅ Clinically validated

• A phase 1 single-ascending-dose study in healthy volunteers (NCT03838185) completed in 2020 and never published its results. A human dose was therefore measured and not reported. There is nothing to cite. The 10-20mg on this page is the vendor's capsule size — it lands below the ~57mg you get by scaling the rodent dose, which is the conservative direction, but neither figure was established in a person.

📊 Correlative data

• Curcumin itself has a long record of poor oral bioavailability and poor brain penetration — the problem J-147 was built to solve. Read the parent compound's disappointing trial history as context for why a derivative was needed, not as evidence the derivative works.

🧪 Theoretical / extrapolated

• Its molecular target was identified as mitochondrial ATP synthase. Modestly inhibiting it triggers an adaptive stress response that raises mitochondrial calcium buffering and downstream NGF and BDNF — a hormetic mechanism rather than a blocking one.
• It was selected in ageing-brain models rather than amyloid models, which is why its preclinical profile reads differently from the Alzheimer's drugs it sits beside. Extensive rodent data; almost no human data.

How to read these tiers: they say how much human evidence exists, not how well something works — and ✗ flags harm, never a disappointing trial. How the evidence tiers work →

What J-147 actually does

J-147 is interesting for a reason that has nothing to do with its structure: it was not designed against amyloid. Twenty years of Alzheimer's drug development aimed at a single protein and mostly failed. J-147 came out of a phenotypic screen — compounds were tested for their ability to protect neurons against the toxicities associated with aging rather than for binding any chosen target — and the target was worked out afterward.

The chemistry. J-147 is a synthetic derivative of curcumin, retaining the pharmacophore that gave curcumin its activity in neuroprotection assays while removing the properties that made curcumin useless as a drug: negligible oral bioavailability, rapid conjugation and poor brain penetration. The compound is curcumin's activity without curcumin's pharmacokinetics, which is a much more useful description than calling it a curcumin analog Qiu 2023.

The target, and it is a bioenergetic one. The binding partner identified for J-147 is a component of the mitochondrial ATP synthase — complex V, the rotary enzyme that makes ATP from the proton gradient Qiu 2023. That enzyme is independently implicated in Alzheimer's disease pathology Patro 2021, which is why an ATP synthase ligand is a plausible neuroprotective agent rather than a curiosity.

Follow the consequence, because it inverts what a reader expects. J-147 partially inhibits ATP synthase. Less ATP production raises the AMP:ATP ratio. That activates AMP-activated protein kinase and inhibits mTOR. Raising AMPK and lowering mTOR is the canonical signature of caloric restriction, which is the most reproducible lifespan-extending intervention known. So a compound that makes mitochondria slightly less efficient is a caloric-restriction mimetic — and mild energetic stress that provokes an adaptive response is the same logic that makes exercise beneficial Austad 2022.

Downstream of that, the readouts reported for the compound — raised BDNF, reduced markers of oxidative stress and inflammation, improved mitochondrial homeostasis — are the expected consequences of an AMPK-directed adaptive program rather than five separate mechanisms Qiu 2023 Rahman 2024. A separate line of work reports J-147 attenuating neuroinflammation through the TLR4/NF-kappaB pathway in a sepsis model Qiu 2023, which is a distinct anti-inflammatory route and is worth keeping separate from the bioenergetic account rather than merging them.

Cell, rodent, human — and where it stops

Step one, the screen and the target: characterized. Phenotypic discovery, the ATP synthase binding partner, and the AMPK/mTOR consequence are set out in the current review of the compound Qiu 2023, and the wider case for targeting whole body metabolism and mitochondrial bioenergetics in Alzheimer's drug development is made independently Austad 2022.

Step two, the disease rationale. ATP synthase and mitochondrial bioenergetic dysfunction are documented features of Alzheimer's disease Patro 2021, and mitochondrial targets are reviewed as a therapeutic class for it Rahman 2024. That is the argument for why a bioenergetic compound belongs in this disease at all, and it is independent of J-147.

Step three, animals — and this is where the whole evidence base sits. J-147 has been studied in transgenic Alzheimer's models and in rapidly aging mouse strains, with cognitive and molecular endpoints, and in a sepsis-induced depressive-behavior model with a defined inflammatory pathway Qiu 2023. The current evidence for the compound across nervous system disease is collected in a narrative review Qiu 2023. A narrative review is the right venue for this compound, and that is itself the informative point: there is not enough randomized human evidence to systematically review.

Step four, humans, and this is the honest paragraph. No results from any human trial of J-147 have been published in the indexed literature. Not a phase 1 report, not a safety cohort, not a pharmacokinetic study. Every claim made for the compound — including every claim made by anybody selling it — is an extrapolation from rodents.

Where the chain breaks. (1) Transgenic Alzheimer's mouse models have an unbroken record of predicting human failures; dozens of compounds have rescued them and none has worked in a person. (2) The mechanism is a partial inhibition of a central bioenergetic enzyme, and the therapeutic window between beneficial mild stress and harmful energetic deficit is unmeasured in any species Austad 2022. (3) No human pharmacokinetics means no basis for a dose. (4) The anti-inflammatory route Qiu 2023 and the bioenergetic route are separate mechanisms studied in separate models, and nobody has shown which one carries the cognitive effect.

J-147 pharmacokinetics — how much of it actually gets in

The Vault card says J-147's half-life is not well characterized in public data. That is true of the number, and the reasoning that bounds it is content.

Start with what the molecule was built for. J-147 was optimized from curcumin explicitly to fix curcumin's pharmacokinetics Qiu 2023. Curcumin's failure has three named causes: it is poorly soluble, so little dissolves in the gastrointestinal tract; what is absorbed is almost entirely conjugated by glucuronidation and sulfation in the gut wall and liver on first-pass metabolism; and the conjugates do not enter the brain. An oral, brain-penetrant derivative had to defeat all three, and the fact that every rodent study dosed it orally and measured a central effect is itself the evidence that it did.

What clears it. A small synthetic molecule of this class is eliminated by hepatic cytochrome P450 oxidation followed by phase II conjugation, not by any peptidase or esterase route. That matters practically: the enzymes that would clear J-147 are the ones altered by common inducers and inhibitors, so exposure is modifiable by other drugs in a way that is completely uncharacterized for this compound.

The arithmetic that bounds the interval. The rodent studies on which every claim rests use once-daily oral dosing over weeks to months Qiu 2023. A once-daily schedule that produces a sustained central effect implies a half-life that is a meaningful fraction of a day — hours, not minutes, or the effect would track a spike and disappear between doses. That is not a measured half-life and it is a real constraint, and it is more than the blank on the card. There is no injectable comparator: the compound was developed as an oral drug and no parenteral human formulation exists.

And the limit of all of it: these are rodent numbers. Species differ several-fold in the conjugating enzymes that dominate this compound's clearance. No human pharmacokinetic data for J-147 has been published, so the interval above is an inference from animal dosing schedules and should be read as one.

What would have to be true, and how you would know it was not

Three predictions. The first follows from the mechanism and is measurable on a standard panel; the third is the one that should stop somebody.

1. If J-147 is a caloric-restriction mimetic, the metabolic markers move and they move in a specific direction. AMPK activation with mTOR inhibition improves insulin sensitivity. So: fasting insulin and HbA1c at baseline and 12 weeks, with a CMP for fasting glucose. A fall in fasting insulin is the cheapest available evidence that an AMPK-directed mechanism is engaged in a person Austad 2022, and if nothing moves there, the bioenergetic claim has no support in that individual.

2. The cognitive claim needs an instrument and two baselines. MoCA and a trail making test, administered twice before the first dose and again at 12 and 24 weeks. Two baselines are not fussiness: both instruments have large practice effects, and a single baseline guarantees that learning will be recorded as improvement. hs-CRP at the same points addresses the separate anti-inflammatory route Qiu 2023.

3. The falsification test that cuts against the compound, and it follows directly from the mechanism. J-147 partially inhibits ATP synthase Qiu 2023. The tissues that would notice first are the ones with the highest oxidative demand: heart, skeletal muscle, and the exercising body as a whole. So the specific adverse prediction is reduced exercise capacity — a fall in VO2 on a repeatable submaximal test, or rising heart rate at a fixed workload. If exercise capacity falls, the compound is inhibiting the enzyme more than mildly, and that is the difference between a hormetic stress and a toxic one. Nobody has established where that line is, which is exactly why it should be measured rather than assumed.

What nobody has tested yet

Four things nobody has published.

Nobody has published a human trial result of any kind. No phase 1, no pharmacokinetics, no safety cohort. For a compound with a named molecular target and a coherent mechanism, that gap is the whole story, and no amount of rodent data closes it.

Nobody has defined the therapeutic window on the enzyme. The mechanism is partial inhibition of ATP synthase Qiu 2023 Patro 2021. Mild inhibition is a hormetic signal; more is an energetic deficit. A dose-ranging study measuring both the AMPK response and mitochondrial respiratory capacity in the same animals would map the window, and until somebody does, every dose is a guess about where the curve turns.

Nobody has separated the two mechanisms. The bioenergetic account Qiu 2023 and the TLR4/NF-kappaB anti-inflammatory account Qiu 2023 are studied in different models. A study using an AMPK inhibitor or an AMPK-null animal to ask whether the cognitive benefit survives would say which mechanism carries the effect, and that determines what the compound should be developed for.

Nobody has compared it against the interventions it mimics. If J-147 works by raising AMPK and lowering mTOR, the comparators are metformin, rapamycin and exercise Austad 2022. A head-to-head in the same aging model with the same cognitive endpoints would say whether this compound adds anything to interventions that already exist — and it is the study most likely to make J-147 unnecessary or to justify it.

J-147 — its own safety story, not its class's

J-147 has no published human safety data. None. The class block above assumes a human record that does not exist, and the specific concerns below are derived from the mechanism and labeled as such.

The mechanism-derived concern is the target itself. ATP synthase is present in every mitochondrion in the body Patro 2021, not only in neurons. A systemically absorbed compound that partially inhibits complex V is inhibiting it in heart, skeletal muscle, liver and kidney at the same time, and the benefit depends entirely on the inhibition remaining partial. That is a dose question in an animal and an unknown in a person.

The second concern is mTOR inhibition, which is not free. Chronic mTOR suppression is the mechanism behind rapamycin's geroprotective effects and behind its immunosuppression, impaired wound healing and metabolic effects. A compound producing the same signature should be expected to produce some of the same consequences Austad 2022, and nobody has measured whether it does.

Interactions are entirely uncharacterized. Clearance is hepatic oxidation and conjugation, so inducers and inhibitors of those enzymes will move exposure. For a compound with no published human pharmacokinetics, that means the direction and size of any interaction is unknown — which is a different and worse situation than a known interaction.

Product identity is a real question. J-147 is a synthetic small molecule sold as a research chemical, and nothing about its supply chain is regulated. Purity, identity and the presence of synthetic intermediates are not established by any certificate a buyer is likely to receive.

What this page will not do. Recommend a dose, or convert a rodent dose into a human one. J-147 is one of the more intellectually satisfying compounds in this Vault — a phenotypic screen, a named bioenergetic target, and a caloric-restriction signature — and it has no human evidence at all, and holding both of those in view is the whole job of this page.

Sources read for this page

• Qiu F, et al. Current evidence for J147 as a potential therapeutic agent in nervous system disease: a narrative review. BMC Neurology 2023 · PMID 37674139
• Qiu F, et al. J147 ameliorates sepsis-induced depressive-like behaviors in mice by attenuating neuroinflammation through regulating the TLR4/NF-kappaB signaling pathway. Journal of Molecular Histology 2023 · PMID 37676534
• Austad SN, et al. Targeting whole body metabolism and mitochondrial bioenergetics in the drug development for Alzheimer's disease. Acta Pharmaceutica Sinica B 2022 · PMID 35256932
• Patro S, et al. ATP Synthase and Mitochondrial Bioenergetics Dysfunction in Alzheimer's Disease. International Journal of Molecular Sciences 2021 · PMID 34681851
• Rahman MA, et al. Drug Target to Alleviate Mitochondrial Dysfunctions in Alzheimer's Disease: Recent Advances and Therapeutic Implications. Current Neuropharmacology 2024 · PMID 39234772

J-147 — safety, predicted from mechanism

J-147 — interference & stacking

Bloodwork to run alongside J-147

Baseline first, then again at 8–12 weeks.

The Brain Fog & Cognition panel covers these in one order — 12 markers, $233.06 with the discount applied.

Check results you already have → · All 103 markers A–Z

J-147 — frequently asked questions

What J-147 is used for

J-147 appears under 1 goal in the goal router.

Related Cognitive & Mood compounds

← Explore the full Protocol Vault

↑ Back to on this page