How much human evidence is there for J147?

One Phase 1 study, NCT03838185, in 64 healthy young and elderly volunteers, sponsored by Abrexa Pharmaceuticals, running from January 2019 to February 2020 and recorded as completed. No results are posted to the registry and no peer-reviewed publication of the outcome was found. No Phase 2 or later trial of J147 is registered. That is the entire human record.

Is there Chinese research on J147?

Yes, and it is the most active recent literature on the compound. Chinese groups have published on J147 in sepsis-induced neuroinflammation via TLR4 and NF-kB, in depression via 5-HT1A and cAMP signalling, in diabetic peripheral neuropathy via AMPK and TRPA1, and in tPA-induced brain haemorrhage after experimental stroke. A 2023 narrative review from Shenzhen Second People's Hospital and Guangdong Medical University consolidates the field.

Is there Russian research on J147?

We searched CyberLeninka, eLibrary and the Russian neurological journals in Russian and found no primary Russian-language research on J147. The Russian-language material that exists is secondary and popular. We publish that as a stated result rather than leaving a silence.

Reference table · Primary sources · Stated negatives

Every claim on this site, and where it comes from

This is the table the rest of the site is built on. Each row names the claim, its primary source, the model it was shown in, the dose where one is stated, and the year. A claim with no row here does not appear anywhere on this site — and the rows recording where we looked and found nothing are included on purpose.

Read the model column carefully. Most of what is known about J147 is known in mice.

Origin and identity

The last two rows exist because secondary sources conflate these compounds, including one that printed “J147 (CAD-031)”. They are three different molecules.
Claim	Source	Model	Year
A synthetic derivative developed from curcumin — not curcumin	Chen Q, Prior M, Dargusch R, … Maher P, Schubert D, PLoS One 2011;6(12):e27865 · PMID 22194796	chemistry	2011
Found by phenotypic screening against old-age brain toxicities, not against amyloid	Salk Institute; Goldberg et al.	cell-based screens	2011–2018
Neurotrophic where curcumin is not; greater stability and bioavailability	Chen et al. 2011; Qiu et al. review	in vitro	2011 / 2023
CAD-031 is a separate compound derived from J147 — not another name for it	Salk Institute; geroneuroprotector literature	—	2018–2019
CMS121 is a fisetin derivative — a different lineage entirely	Salk Institute (Maher, Currais)	—	2018

Mechanism

Claim	Source	Model	Year
Target is the alpha subunit of mitochondrial F1-ATP synthase (ATP5A)	Goldberg J, Currais A, Prior M, Schubert D, Maher P et al., Aging Cell 2018;17(2):e12715 · PMID 29316249	cell + mouse	2018
Alters intracellular calcium handling; engages CAMKK2 / AMPK stress resilience	Goldberg et al.	cell	2018
ATP synthase described as a target shared by ageing and dementia	Goldberg et al. — the paper's title	—	2018
Neither an acetylcholinesterase nor a phosphodiesterase inhibitor	Prior et al.	—	2013

Animal efficacy and pharmacokinetics

Claim	Source	Model	Dose	Year
Reverses cognitive impairment in aged APP/PS1 mice, dosed late in disease	Prior et al., Alzheimer's Res Ther 5:25	aged transgenic mouse	oral	2013
Preserves synaptic proteins and cognition across ageing / neurodegeneration models	Qiu et al. review, aggregating primaries	mouse, rat	`1–200 mg/kg/day` oral; 200 ppm diet	2011–2022
Extends lifespan in flies	cited in Goldberg et al. / Salk	Drosophila	—	2018
Lowers plasma fatty acids via AMPK/ACC1 in liver	Pharmacological Research	mouse	—	2022
Oral bioavailability ≈ 28 %	Prior et al., PK section	mouse	`20 mg/kg single`	2013
Plasma half-life ≈ 1.5 h; brain half-life ≈ 2.5 h	Prior et al.	mouse	`20 mg/kg`	2013
Brain-to-blood ratio ≈ 0.5; high BBB penetration in MDCK-MDR1	Prior et al.	mouse; in vitro	`20 mg/kg`	2013

The human record, in full

Checked September 2026. This table is the entire human evidence base for J147.
Fact	Source	Detail
One Phase 1 trial, completed	ClinicalTrials.gov `NCT03838185`	“A Phase I, Randomized, Double-Blind, Placebo-Controlled Study to Assess the Safety, Tolerability and Pharmacokinetics of Single Ascending Oral Doses of J147 in Healthy Young Volunteers and Healthy Elderly Volunteers”
Sponsor and size	NCT03838185	Abrexa Pharmaceuticals, Inc. · 64 participants
Dates	NCT03838185	started 22 January 2019 · completed 1 February 2020
Comparator	NCT03838185 intervention record	a single oral dose of corn oil
Results	NCT03838185	`hasResults: false` — nothing posted, and no peer-reviewed publication of the outcome found
Later-phase trials	ClinicalTrials.gov search	none registered

A completed trial with unpublished results is not evidence of anything — in either direction. It is not a hidden success and it is not a concealed failure. It is an absence, and the honest thing is to record it as one.

The Chinese literature

The most active recent work on J147 is Chinese, and it is under-cited in Western summaries. The groups are named here because they did the work.

Finding	Source	Model	Dose	Year
Attenuates sepsis-induced depressive-like behaviour via TLR4/NF-κB in hippocampus and microglia	Qiu F et al., J Mol Histol 2023 · PMID 37676534 (PMC10635911)	mouse + microglia	not stated in abstract	2023
Narrative review consolidating J147 across nervous-system disease	Qiu F, Wang Y, Du Y, Zeng C, Liu Y, Pan H, Ke C — Shenzhen Second People's Hospital; Guangdong Medical University; Center for Human Tissues & Organs Degeneration. BMC Neurol 2023 · PMID 37674139	review	— (review)	2023
Antidepressant-like effects involving 5-HT1A	Lian L et al., Neuropharmacology 2018 · PMID 29626566	mouse	not stated in abstract	2018
Sub-acute dosing improves depression-like behaviour via 5-HT1A-mediated cAMP signalling	Li J et al., Front Neurosci 2020;14:701 · PMID 32733195	mouse	9 mg/kg, gavage	2020
Ameliorates diabetic peripheral neuropathy via AMPK negative regulation of TRPA1	Lv J et al., Acta Cir Bras 2018 · PMID 30020315	STZ rat	not stated in abstract	2018
Reduces tPA-induced brain haemorrhage in acute experimental stroke	Jin R et al., Front Neurol 2022;13:821082 · PMID 35309561	rat	not stated in abstract	2022
Drives microglia toward a resolving phenotype via CAMKK2/AMPK, reducing neuroinflammation — the first primary to test this site’s own mechanistic claim directly	He L, Ali T, Wei T, Yin H, Yang Y, Zhao Z, Liu H, Tan Z (Shenzhen; Longhua People’s Hospital), Biochem Biophys Res Commun 2025;778:152395 · PMID 40712391	microglia + mouse	not stated in abstract	2025
Protects against traumatic brain injury by inhibiting neuronal endoplasmic-reticulum stress, potentially via AMPK/SREBP-1	Jin R, Wang M, Shukla M, Lei Y, An D, Du J, Li G (Penn State Hershey), Transl Res 2024;274:21–34 · PMID 39245209	mouse	not stated in abstract	2024
Affects cognition and anxiety after surgery — a peri-operative model, not a neurodegeneration one	Oberman K, van Leeuwen BL, Nabben M, Villafranca JE, Schoemaker RG (Groningen), Physiol Behav 2024;273:114413 · PMID 37989448	Zucker rat	not stated in abstract	2024

Where we looked and found nothing

Question	What was searched	Result
J147 in autism	PubMed / PMC; the Chinese literature in Chinese; the Russian literature in Russian; the autism-model literature directly (BTBR, prenatal valproic acid, Shank3, Fmr1)	Nothing. No mouse model, no cell model, no case report, no trial. See the autism page, which is built on this absence rather than around it.
Russian-language primary research on J147	CyberLeninka, eLibrary.ru, Russian neurological journals; native-language queries for «J147», «куркумин производное», «нейропротекция», «АТФ-синтаза»	None found. Russian-language coverage of J147 is secondary and popular. Russian primary work on curcumin and on mitochondrial ATP synthase exists and is context — it is never cited on this site as J147 evidence.
Full patent family and territory-by-territory expiry	Google Patents / WIPO for the Salk filing	Not established. We have the filing; we do not have the complete family. The chemistry page says so rather than implying a clear field.