Neboglamine (CR 2249)

Chemical identity, and a name that was replaced

Two names sit on one PubChem record. CID 3074827 carries Neboglamine and Nebostinel together, along with CAS 163000-63-3, molecular formula C13H24N2O3, molecular weight 256.34 g/mol, InChIKey VCRGLZYPNNAVRP-JTQLQIEISA-N and UNII 12EA34U5B8. The IUPAC name on that record is (4S)-4-amino-5-[(4,4-dimethylcyclohexyl)amino]-5-oxopentanoic acid, which is a glutamic acid derivative carrying a 4,4-dimethylcyclohexyl amide. It is a small synthetic molecule and not a peptide, despite the amino-acid backbone. Two development codes attach to the same structure: CR 2249, used by the Italian originator, and XY-2401, used by the American licensee. The FDA Global Substance Registration System files neboglamine as the official name and nebostinel as a common one, against INN number 7929. The public registry page built on that same record prints the designation the other way round, with nebostinel as the official name, so the two displays disagree.

Stereochemistry is load-bearing here. The published assays used the (S) enantiomer, written in the older papers as (S)CR 2249, and the opposite enantiomer, CR 2361, was reported inactive in the same preparation. One laboratory reported that inactivity, in two papers three years apart (PMID 9336311, PMID 10381808); no independent group has published a test of CR 2361, and it is not mentioned in the record of Lanza 1997 (PMID 9294970). Catalogue material more often lists a hydrochloride salt, which has its own record: CAS 2759182-59-5, PubChem CID 165412562, C13H25ClN2O3, 292.80 g/mol. The free acid and the salt are separate registry entries and should not be read as interchangeable numbers.

Nothing under any of these names is approved anywhere. A Drugs@FDA query on neboglamine as an active ingredient returns no match, and an openFDA label search returns no match. The Global Substance Registration System records the approval year as unknown. An International Nonproprietary Name is a naming decision taken by the World Health Organization and carries no regulatory weight; the registry records INN 7929 and links a WHO extranet page for it, but it does not name a proposed list, and the WHO proposed lists are published as scans with no text layer, so no list number could be established from a source.

What the in vitro work established about the glycine site

The kynurenate test, built by Pittaluga and colleagues in 1997, measures function: rat hippocampal slices, NMDA-evoked release of tritiated noradrenaline, blocked by kynurenic acid, with the question being which compounds reverse that block. No binding measurement is involved. CR 2249 reversed the block and its enantiomer did not. The same paper reported a result that complicates the usual description of this molecule. In synaptosomes, glycine and D-cycloserine enhanced NMDA-evoked release, while oxiracetam and CR 2249 did not, and the authors concluded that drugs active in their test may bind to sites different from the glycine site of the NMDA receptor.

Lanza and colleagues examined the interaction directly in the same year. Working in rat hippocampal slices, they reported that (S)CR 2249 at 3, 10 and 30 micromolar raised the efficacy of glycine's reversal of kynurenate antagonism in a concentration-dependent way while glycine's potency was not significantly changed (glycine EC50 21.5 plus or minus 4.2 micromolar). In binding experiments the molecule increased tritiated dizocilpine binding concentration-dependently, with positive cooperative interaction between glycine and the compound. Their stated conclusion was that it probably acts at a separate allosteric site to increase NMDA receptor functionality.

Human tissue enters the record once, and only in vitro. Pittaluga and colleagues repeated the assay in 1999 in slices of human cerebral cortex taken from patients undergoing neurosurgery, and reported that CR 2249 prevented the antagonism produced by 300 micromolar kynurenate with an EC50 under 10 micromolar, alongside aniracetam, oxiracetam and D-cycloserine. The enantiomer was again inactive. That is human neocortex in a superfusion chamber, not a person taking a compound, and the paper reports no patient count for the tissue donors in its abstract.

The rodent record, and the part of it that sits in patents

Garofalo and colleagues in 1996 reported the first behavioural results: CR 2249 reduced the memory-retention deficit produced by scopolamine in a step-through passive avoidance procedure in rats and by electroconvulsive shock in a step-down procedure in mice, and performance also changed in animals with no induced cholinergic deficit. Timing was specific: an effect was reported only when the compound was given 45 minutes before training, and none when it was given immediately after training or before the retention trial. Microdialysis in freely moving rats showed increased hippocampal noradrenaline release and reduced efflux of its metabolite. Doses, route and group sizes are absent from the retrieved abstract.

Chiusaroli and colleagues published the last primary paper on this molecule in 2010. Counting Fos-like immunoreactivity in rat forebrain, they reported increases of 3.2-fold in prefrontal cortex, 4.8-fold in nucleus accumbens and 4.5-fold in lateral septal nucleus, and no effect in dorsolateral striatum, where haloperidol produced a 390-fold increase. In the same 2010 rat study, oral neboglamine inhibited phencyclidine-induced hyperlocomotion and rearing dose-dependently while basal locomotor activity was unchanged, unlike haloperidol and clozapine in the same experiment (PMID 20045056); the retrieved abstract states neither dose nor animals per group.

Two patents carry experiments that were never published as papers. US 7,737,180, priority May 2004 and granted 2010, describes rats of roughly 250 grams given 1, 3 and 10 mg/kg intraperitoneally, ten per group, in a phencyclidine-disrupted prepulse inhibition procedure, with about 80 per cent inhibition of the phencyclidine effect at the top dose; mice at 3 and 10 mg/kg intraperitoneally, ten per group, in which immobility rose from 48 seconds in controls to 130 seconds after fourteen days' daily phencyclidine at 10 mg/kg subcutaneously and was 58 seconds at neboglamine 10 mg/kg, which the patent reports as 87.8 per cent inhibition of the phencyclidine effect by its own stated formula; and gerbils at 12.5 to 50 mg/kg in an ischaemia model, where the patent attributes the neuroprotective effect principally to the metabolite CR 2863, which was more potent than the parent in the same model. US 8,258,185, priority October 2005 and granted 2012, reports rats given 30 mg/kg intraperitoneally with cocaine-induced hyperactivity inhibited by 88 per cent and morphine-induced by 77.5 per cent, without stating group sizes. Patent examples are not peer reviewed and were filed by the party seeking the claim.

People were dosed, and nothing was published

A company announcement dated 8 January 2007 is the only public account of human exposure. It states that single-dose escalation and food-interaction Phase 1 studies had been completed at a clinical pharmacology unit in Rouffach, France, and that a multiple-dose escalation study was due to finish in the second quarter of 2007. The release stated that volunteers in the then-ongoing multiple-dose escalation study would be evaluated with 24-hour quantitative electroencephalography, polysomnography, 24-hour Holter electrocardiography and psychometric tests. It does not report that any of those assessments were completed, and no result from any of them has been published. The release asserts, in its own words, that the compound is "safe and well-tolerated within the dose range that is expected for efficacy". It reports no participant count, no dose, no adverse event, and no efficacy endpoint; nothing supporting that sentence has been published.

The subject of that announcement was a funding agreement. It reported a cooperative research and development agreement with the United States National Institute on Drug Abuse, valued at 40 million dollars, to test whether the compound could support abstinence from cocaine or prevent relapse. Implementation was described as conditional on completing the multiple-dose escalation study and on the institute's own evaluation of interactions between cocaine and the compound in animal and human models. Nineteen years later, no report of that evaluation, of the interaction study, or of any of the Phase 1 work appears in the indexed literature.

The published record is five papers. A PubMed search across neboglamine, nebostinel and CR 2249 returns five records, the newest from 2010, every one of them preclinical or in vitro. A Europe PMC search across the same terms on 19 August 2026 returns 23 records: twelve patent-family members and eleven literature records, with a search on nebostinel alone returning zero. Several of the literature records are reviews that mention the compound in passing. There is no pharmacokinetic paper, no Phase 1 report, no electroencephalography or polysomnography dataset, and no toxicology publication.

What the Phase 2 claim rests on

Aggregator pages describe this molecule as having reached Phase 2 for schizophrenia and for cocaine dependence, usually dated to June 2015. ChEMBL records a maximum phase of 2 for CHEMBL1255840 and attaches no trial identifier to it. Searches of ClinicalTrials.gov in August 2026, by intervention and by free text, across neboglamine, nebostinel and CR 2249, returned zero studies. A search of the EU Clinical Trials Register returned the message that the query did not match any clinical trials. Neither registry holds a record of a trial of this compound at any phase, for any indication, in any country.

The phase figure traces to a commercial pipeline database, record 800004810, which the National Center for Advancing Translational Sciences cites on its own page for this substance. That URL redirects to a subscription identity provider and the record cannot be read without access, so the assertion cannot be checked against anything a reader can open. The same page carries two further data points, a potency of 10 micromolar typed as a Ki and a single human oral dose figure, both sourced to a conference abstract hosted on the 2010 FENS Forum website. The dose figure is not reproduced here. That host no longer resolves, and the Wayback Machine holds no capture of the file.

Ownership moved more than once. The patents were filed by Rottapharm SpA and are now assigned to Rottapharm Biotech Srl; Xytis Pharmaceuticals held the molecule under the code XY-2401 during the period when the Phase 1 work was run. What can be established from open sources is that a compound was taken into healthy volunteers, that a large public funding agreement was announced for a cocaine indication, that nothing was registered in either major trials registry, and that the last primary publication appeared in 2010.

Where the molecule appears now

A market surveillance study published in 2025 by twelve official medicines control laboratories across Europe and Australia, coordinated through the General European Official Medicines Control Laboratory Network, documented 159 samples between January 2020 and September 2024, yielding 166 molecular identification entries across 34 distinct molecules. Sixty-nine per cent of samples came from the illegal market, 49 per cent were presented as dietary supplements and 32 per cent as medicines. Neboglamine appears in the appendix table with a detection frequency of one, found by one of the twelve laboratories, presented mainly as a substandard or falsified medicine, with legal status recorded as research chemical. That is the only record of the compound in a seized or purchased sample.

Its neighbours in that table are not uniform. Dihexa, NSI-189, J-147, IDRA-21 and 9-methyl-beta-carboline appear with the same one-or-two detection counts and the same bare research-chemical status. Bromantane and picamilon are recorded differently: research chemical in the EU and Australia, but in use as a prescription drug in Russia. The authors' framing is that a set of clinically uncharacterised molecules is reaching consumers through products presented as supplements or as medicines, and that consumers may not be aware of what they contain. That paper is, at present, the only peer-reviewed publication that records this compound existing outside a laboratory.

Claims attached to it in that market do not come from the five papers. A specific milligram human dose range, an approximate four-hour half-life and weak noradrenaline reuptake inhibition above a stated dose all circulate on aggregator library pages and catalogue copy, and none of the three appears in Garofalo 1996, Pittaluga 1997, Lanza 1997, Pittaluga 1999 or Chiusaroli 2010. No human pharmacokinetic measurement of any kind is retrievable. The characterisation as an antidepressant is a different case: it is absent from the five papers but present in the 2007 trade announcement, which claims anti-depressive activity in preclinical animal models, and in the forced swim example of US 7,737,180 B2. Each is recorded below with the search behind it.

What is not known

There is no published human data of any kind. Healthy volunteers were dosed: a January 2007 company announcement describes completed single-dose escalation and food-interaction Phase 1 studies at a unit in Rouffach, France, and a multiple-dose escalation study then in progress, in which it stated that volunteers would be evaluated with quantitative electroencephalography, polysomnography, Holter electrocardiography and psychometric tests. It does not report that any of those assessments were completed, and it gives no date for the dosing itself. None of that work has ever appeared in a journal, in a registry, or in any document a reader can open, so no exposure, no adverse event rate, no elimination half-life, no food effect and no participant count is publicly established. Neither ClinicalTrials.gov nor the EU Clinical Trials Register holds a record of any trial of this compound, which means the widely repeated Phase 2 designation cannot be checked against a protocol, an enrolment figure or an endpoint. The binding site remains unidentified: the two 1997 papers each concluded that the compound acts somewhere other than the glycine site itself, and no radioligand displacement study has been published in the twenty-nine years since. Nothing has been reported on chronic administration, reproductive or developmental effects, carcinogenicity, drug interactions, or behaviour in any species other than rat, mouse and gerbil. The most recent primary publication is from 2010, and the animal literature outside the two patents amounts to two behavioural papers and three in vitro studies from the originating groups.

Questions