TAK-653 (osavampator)

TAK-653 (osavampator)

TAK-653 (also called osavampator or NBI-1065845) is an experimental drug originally developed by Takeda and now licensed to Neurocrine Biosciences for treatment-resistant depression. It works by gently amplifying the brain's main excitatory signal, glutamate, through positive allosteric modulation of the AMPA receptor. In plain terms, it makes existing glutamate signalling more efficient without firing the receptor itself, which is why it doesn't cause the seizures or dissociation that have killed off most other drugs in this class.
Most people taking it outside of a clinical trial are using it as a nootropic for sustained cognitive enhancement, working memory, focus, mental clarity, and as a non-stimulant lever for mood and motivation, especially in people who haven't responded well to SSRIs. The half-life is long (around 33 to 48 hours), so it's a once-daily compound that builds to steady state over a week. It's not an acute-feel drug. Most users describe a slow lift in cognitive bandwidth and motivation that becomes noticeable after 5-14 days, not a same-day kick.

Deep-dive

AMPA receptors are the workhorse glutamate receptors in the brain. They handle the fast excitatory signalling that underpins learning, memory, and mood circuit function. Most antidepressants act indirectly on this system through serotonin or noradrenaline, which is part of why SSRIs take 4-6 weeks to work and miss roughly a third of patients entirely. The newer model, supported by ketamine's rapid antidepressant effect, is that boosting AMPA-driven synaptic plasticity directly is what actually drives mood improvement, and the serotonin pathway is just one slow way to get there.
TAK-653 is a positive allosteric modulator (PAM), not an agonist. The distinction matters. An AMPA agonist activates the receptor on its own, which causes runaway excitation and seizures (this is why earlier AMPA drugs like CX516 and Org26576 stalled out). A PAM only works when glutamate is already there, slowing receptor desensitisation and amplifying the normal signal. Preclinical work showed TAK-653 has minimal intrinsic agonist activity, which is what gives it a wide safety margin against seizures while still producing robust pro-cognitive and antidepressant effects.
Pharmacokinetics. Oral, well-absorbed, with peak plasma levels at 1.25 to 5 hours and a terminal half-life of 33 to 48 hours. It crosses the blood-brain barrier readily. Linear pharmacokinetics across the studied dose range (0.3 to 18 mg single, up to 9 mg multiple). Metabolism is mostly hepatic via CYP3A4 and CYP2D6, with renal excretion under 10%.
Cognitive evidence in humans. A randomised crossover study in 24 healthy volunteers tested 0.5 mg and 6 mg against placebo using transcranial magnetic stimulation and the NeuroCart battery. TAK-653 increased cortical excitability on TMS in a dose-dependent way, consistent with target engagement, and produced a stimulant-like cognitive profile without the dissociative or sedative side effects of ketamine. Stroop performance shifted in a way that suggests effects on executive function, though the signal was inconsistent across doses. There are no large healthy-volunteer trials on long-duration cognitive enhancement, so most of the nootropic claims rest on this single mechanistic study plus the broader AMPA PAM literature (piracetam and similar compounds have decades of cognitive enhancement data, and TAK-653 is more selective and potent).
Antidepressant evidence. The Phase 2 SAVITRI trial randomised 183 adults with MDD who had failed at least one antidepressant to placebo, 1 mg osavampator, or 3 mg osavampator added on top of their existing antidepressant for 8 weeks. The 1 mg dose hit the primary endpoint with a 4.3-point placebo-adjusted MADRS reduction at day 28 and a 7.5-point reduction by day 56 (effect size around 0.73, which is roughly double what most approved antidepressants achieve). Remission rate was 48.7% at day 56 versus around 23% on placebo. The 3 mg dose underperformed and didn't reach statistical significance, suggesting a U-shaped or downward dose response above 1 mg. Phase 3 trials are now enrolling.
Why the 3 mg dose did worse. This is one of the most interesting findings and is consistent with what's known about glutamate signalling. AMPA receptors operate within a narrow optimal window. Push too hard and you get receptor desensitisation, internalisation, and disrupted excitation-inhibition balance. The same inverted-U shows up with most cognitive enhancers (dopamine, noradrenaline, choline). For TAK-653 specifically, the ceiling seems to sit around 1 mg for mood, with cognitive effects potentially extending higher but with diminishing or reversing returns. More is not better.
Women. The drug interaction work explicitly tested women on combined oral contraceptives. A dedicated phase 1 study showed TAK-653 doesn't induce CYP3A4 and doesn't reduce ethinyl estradiol or levonorgestrel exposure, meaning it's compatible with hormonal birth control without dose adjustment. The Phase 2 SAVITRI trial included men and women across the dose groups and didn't report sex differences in response, though it wasn't powered to detect them. Population pharmacokinetic models found no clinically significant sex differences in exposure. There's no evidence that women need a different dose or schedule. Caveat: like all of psychiatry, MDD trials historically underrecruit women in proportion to actual disease burden, and there's no published menstrual cycle, perimenopause, or postmenopause subgroup analysis. Avoid in pregnancy and breastfeeding (no data, and animal reproductive toxicology hasn't been published).
Limitations of the evidence. All human data comes from healthy volunteer phase 1 studies and one phase 2 trial in MDD. There are no published long-term human safety data beyond a few months, no head-to-head trials against other antidepressants or nootropics, no studies in older adults specifically, and the cognitive claims in healthy users are largely extrapolated from mechanism and the small TMS/NeuroCart study. The compound is sold by research chemical vendors with variable quality control, which is a real risk separate from the drug itself.

Dosage:

  • Standard dose: 1 mg once daily, taken at the same time each day. This is the dose that worked in the SAVITRI MDD trial and the dose the Phase 3 program is being built around. For nootropic use, most people land in the 0.5-2 mg range
  • Start low: 0.5 mg for the first 5-7 days to assess tolerance, then move to 1 mg if no issues. Given the long half-life, you don't reach steady state until day 7-10 anyway, so there's no rush
  • Don't go higher chasing more effect. The 3 mg arm of SAVITRI underperformed the 1 mg arm. The dose-response is not linear, it's an inverted U. Going to 3-6 mg is more likely to blunt the benefit than amplify it
  • Timing: Morning is typical to align with cognitive demands of the day. The half-life is long enough that timing doesn't matter much for effect, but morning dosing reduces the chance of any subtle activation interfering with sleep
  • With or without food: Doesn't meaningfully affect absorption. Take it however you'll remember consistently
  • Onset: This is not an acute-feel compound. Plasma peaks at 1-5 hours, but clinical effects build over days. Expect 1-2 weeks before you can fairly judge whether it's working. The MDD response peaked at day 56, not day 7
  • Cycling: No established protocol. If using as a nootropic, taking 4-8 week courses with breaks is more conservative than continuous use, given that long-term human safety data is limited
  • Women on combined oral contraceptives: No dose adjustment needed. TAK-653 doesn't reduce ethinyl estradiol or levonorgestrel exposure
  • Stacking: Not well studied. Combining with other AMPA PAMs (piracetam, aniracetam) is theoretically additive and increases risk of overshooting the optimal window. Combining with stimulants (caffeine, modafinil) is fine in principle but compounds the cardiovascular and overstimulation risk. Avoid stacking with any compound that lowers the seizure threshold

Here's what you can expect:

First week is usually nothing dramatic. The half-life means it takes 7-10 days to reach steady state, and the antidepressant data shows the response builds over weeks, not hours.
By week 2-4, most people who respond report a steady lift in cognitive bandwidth, easier sustained attention, sharper working memory, more fluent verbal recall, and a baseline mood improvement that feels like a low background hum rather than a stimulant high. Motivation tends to come back before mood does. Tasks that felt heavy feel manageable. People often describe it as the absence of a fog they didn't realise was there.
In the MDD trials, the bigger response showed up at day 56 than at day 28, suggesting the mechanism is genuinely cumulative, not acute. If you're using it for mood, give it 6-8 weeks before deciding it's not working.
Non-responders exist. Roughly half of MDD patients in SAVITRI didn't reach remission. The healthy-volunteer cognitive data is thinner, so the response rate for nootropic use is genuinely unknown. Some people will feel nothing meaningful.
There is no euphoria, no buzz, no obvious subjective drug effect. If you're chasing a feeling, this is the wrong compound.

Side effects & risks:

  • Headache is the most common reported side effect across both Phase 1 and Phase 2 trials. Usually mild, often fades within the first week as you reach steady state
  • Nasopharyngitis-like symptoms (mild upper respiratory irritation) showed up at slightly elevated rates versus placebo in SAVITRI. Mechanism unclear, generally self-resolving
  • Sleep disruption. Some users report mild activation or shallower sleep, especially in the first week or at higher doses. Morning dosing helps. If it persists, dose is probably too high
  • No dissociation, no psychotomimetic effects. Unlike ketamine, TAK-653 does not cause derealisation, hallucinations, or out-of-body experiences. This was specifically tested in healthy volunteers and confirmed across dose ranges up to 18 mg
  • No seizure signal in clinical trials so far. This was the major risk for earlier AMPA drugs and was the central design goal for TAK-653. Phase 1 and Phase 2 data have shown no seizures and no convulsions across hundreds of participants. That said, anyone with a history of epilepsy or seizure was excluded from the trials, and the compound has not been tested in that population. If you have a seizure history, do not take this
  • Inverted-U dose response. Going above 1-2 mg increases the chance of cognitive impairment rather than enhancement, plus more side effects without benefit. Higher is worse, not better
  • Long-term safety data is genuinely limited. The longest published trial is 8 weeks. There is no human data on what daily use over 1+ years looks like
  • CYP3A4 and CYP2D6 metabolism. TAK-653 is metabolised through these pathways. Strong inhibitors (ketoconazole, ritonavir, some SSRIs like fluoxetine and paroxetine for CYP2D6) could elevate exposure. Strong inducers (rifampin, St John's wort) could lower it. TAK-653 itself is not a CYP3A4 inducer, so it doesn't reduce the effectiveness of birth control or other CYP3A substrates
  • Drug interactions with antidepressants. TAK-653 was tested as add-on therapy in SAVITRI, meaning patients stayed on their existing SSRI/SNRI. No serious interaction signals emerged, but the combination pharmacology is not fully characterised, and you're stacking glutamatergic action on top of monoaminergic action
  • Source quality. TAK-653 is not approved or sold as a supplement. It's available only through research chemical vendors, with all the quality control concerns that implies. Vendor variability, contamination, mislabelled doses, and counterfeit products are real risks. If you go this route, source matters more than dose
  • Pregnancy and breastfeeding: Skip it. No human data, no published reproductive toxicology
🩸

Blood markers

Liver enzymes (ALT, AST, GGT), baseline before starting and recheck at 8-12 weeks if using continuously. The compound is heavily hepatically metabolised and long-term human liver safety data is limited.
Comprehensive metabolic panel including kidney function (creatinine, eGFR), baseline. Renal clearance is minor but worth a reference point.
Full blood count, baseline, especially if using research-grade material where contamination is possible.
Thyroid panel (TSH, free T4) is not specifically indicated for TAK-653 itself but is worth having if you're using it for low mood, since untreated hypothyroidism mimics depression and is the most commonly missed diagnosis in this space.
Most people using TAK-653 situationally for short courses don't need bespoke bloodwork beyond a baseline comprehensive panel. The people who actually need monitoring are those running it continuously for months, anyone with a history of liver issues, and anyone combining it with multiple other compounds.
Not approved by any regulatory authority. Currently in Phase 3 trials for major depressive disorder.