IUPAC: 3,5-dimethyladamantan-1-amine · C12H21N · MW 179.31 g/mol · CAS 19982-08-2 (base) / 41100-52-1 (HCl) · ChEMBL CHEMBL807 · ATC N06DX01
Memantine. An aminoadamantane and the same kind of NMDA-receptor open-channel blocker as ketamine and PCP — yet it is a first-line Alzheimer's drug, not a dissociative. The difference is entirely kinetic: memantine is low-affinity, fast off-rate, and strongly voltage-dependent, so it blocks tonic, pathological Ca2+ flux while sparing the fast phasic signalling of normal synapses. Trade names: Namenda, Namenda XR, Ebixa, Axura; combined with donepezil as Namzaric.
Memantine plugs the identical site as ketamine, PCP and MK-801: the deep vestibule of the NMDA-type ionotropic glutamate receptor pore, a heterotetramer of two obligatory GluN1 subunits and two GluN2 subunits (GluN2B dominant in forebrain). The protonated amine on the adamantane cage sits near the GluN1/GluN2 asparagine (N-site) ring that also coordinates the physiological Mg2+ block, physically occluding Na+/Ca2+ flux. ChEMBL curates it as an NMDA-receptor negative allosteric modulator / open-channel blocker — it never touches the glutamate or glycine orthosteric sites. Structurally this is exactly what the memantine-bound cryo-EM map (PDB 7SAD) shows.
If the site is the same, why is memantine a well-tolerated Alzheimer's drug while ketamine is a dissociative anaesthetic and PCP is a psychotomimetic? The answer is kinetics, not location. Memantine is a low-affinity blocker (Ki ≈ 0.5–1 µM vs ketamine ~0.4 µM and MK-801 in the low-nM range) with a fast off-rate and only partial trapping, and its block is strongly voltage-dependent. Those three properties turn it into an activity- and voltage-filter that discriminates pathological tonic NMDA current from physiological phasic NMDA current.
Memantine's channel dwell time is short — it unbinds in the tens-to-hundreds of millisecond range, orders of magnitude faster than the near-irreversible trapping of MK-801. Low affinity means it never accumulates a deep, persistent block, so it silences NMDA current without abolishing it.
Block deepens at hyperpolarised potentials and is relieved by depolarisation. A brief synaptic EPSP transiently expels memantine (much like relieving the Mg2+ block), so the fast, strong depolarisation of a real synaptic event is spared.
Under chronic low-level (pathological) glutamate — the extrasynaptic, weakly depolarising drive of excitotoxicity — memantine stays lodged and clamps sustained Ca2+ entry. This is the therapeutic target: pathological tonic influx blocked, phasic synaptic transmission preserved (Chen & Lipton; Parsons; Danysz).
Where MK-801 is fully trapped (channel shuts and the blocker is locked in essentially forever), memantine is only partially trapped and can leave through both open and partly-closed states. This "foot-in-the-door" behaviour keeps unblock rapid and avoids cumulative, transmission-wrecking block.
Extrasynaptic GluN2B-containing receptors driving death signalling are a preferred functional substrate. GluN2 subtype also tunes memantine's voltage dependence and IC50 (Kotermanski & Johnson), giving physiological Mg2+ and memantine partly overlapping — but non-identical — filters.
By capping pathological Ca2+ load, memantine is neuroprotective/symptomatic in models of glutamatergic excitotoxicity. It does not clear amyloid or tau — it is a downstream brake on one final common pathway of neuronal injury, which is why the clinical effect is modest but real.
The explicit contrast. Ketamine and PCP are higher-affinity, slow-off, deeply-trapping blockers. Their block outlasts individual synaptic events and accumulates across normal, high-frequency transmission — so they disrupt phasic corticothalamic integration itself. That functional disconnection is the dissociation and psychotomimesis. MK-801 sits at the extreme (essentially permanent trap, powerful psychotogen, research-only). Memantine sits at the opposite pole: same pore, but the block clears fast enough to let real synapses through. Same drug class, opposite clinical identity — a textbook case of kinetics dictating pharmacology.
Memantine's PK is almost the mirror image of ketamine's. It is ~100% orally bioavailable, essentially unaffected by food, and — critically — largely not metabolised. Only a small fraction undergoes hepatic conversion to three inactive metabolites (N-3,5-dimethyl-gludantan / the N-glucuronide, 6-hydroxy-memantine, and 1-nitroso-deaminated memantine); the CYP system plays little role, so memantine has few CYP-mediated drug interactions. Roughly 48% is excreted unchanged in urine, partly via active tubular secretion through the organic-cation transport system.
The consequences: a long elimination half-life of 60–100 hours (functionally ~70 h), which mandates slow weekly titration and gives smooth once-daily (XR) or twice-daily steady-state exposure. Because clearance is renal and shares the cationic secretion pathway, urinary pH and renal function dominate exposure: alkaline urine (from carbonic-anhydrase inhibitors, sodium bicarbonate, or a vegetarian diet) sharply reduces clearance and raises plasma levels, and severe renal impairment requires dose reduction.
Metabolism cascade: the pharmacologically important point is what is absent — there is no active-metabolite chain to complicate the effect, unlike ketamine's norketamine/HNK cascade.
Because the parent molecule is the drug and it leaves the body slowly and renally, memantine exposure is stable and predictable. That pharmacokinetic calm — no first-pass metabolite pharmacology, no CYP tug-of-war, no route-dependent identity change — is part of why it is an easy chronic geriatric medication, and part of why it is a poor recreational dissociative (below).
Memantine (Merz, marketed in Germany from the 1980s; FDA approval as Namenda in 2003) is indicated for moderate-to-severe Alzheimer's disease, where cholinesterase inhibitors alone plateau. The effect size is modest but genuine — small improvements or slowed decline across cognition, activities of daily living, and behavioural/agitation measures — and it is frequently combined with donepezil (fixed-dose Namzaric) because the two mechanisms are complementary and non-overlapping.
By damping chronic excitotoxic Ca2+ load on vulnerable cortical and hippocampal neurons, memantine improves the signal-to-noise of glutamatergic transmission that AD pathology degrades. It is disease-modifying only in the loose sense of protecting downstream circuitry; it does not touch amyloid, tau, or the primary neurodegenerative cascade.
Memantine is unusually well tolerated. The most common adverse events are dizziness, headache, confusion, constipation and mild hypertension — and, crucially, it lacks the cholinergic GI burden (nausea, diarrhoea, bradycardia) that limits donepezil/rivastigmine. Discontinuation rates in trials approach placebo. Its therapeutic dose sits far below the concentration needed to substantially block phasic NMDA transmission, so cognition is not impaired at label doses.
Start 5 mg once daily, increase by 5 mg weekly to a target of 20 mg/day (10 mg twice daily, or 28 mg extended-release once daily). The gradual ramp is dictated by the 60–100 h half-life and minimises transient CNS effects during accumulation. Renal impairment (CrCl < 30 mL/min) caps the dose at ~5–10 mg/day.
Memantine is an NMDA channel blocker, so at doses well above the therapeutic 20 mg (into the hundreds of mg) the block becomes deep enough to produce PCP-like dissociation, confusion, ataxia, and hallucinations. But its low affinity, long half-life, and steep, delayed dose–response make it a clumsy and unrewarding recreational dissociative — effects come on slowly, linger for days, and offer poor control. This is the same molecular action as ketamine expressed at the wrong end of the dose curve, and it is a caution, not a use case.
Off-label, memantine has been trialled for neuropathic pain, OCD augmentation, and other conditions on the strength of its NMDA antagonism, with mixed and generally weak evidence. Its clean, well-mapped pharmacology makes it a favourite mechanistic tool for probing tonic-vs-phasic NMDA signalling in the lab.
Evidence-based, non-moralistic. Memantine has a wide safety margin and low abuse potential — the real issues are additive NMDA-antagonist stacking, renal/urinary-pH effects on exposure, and the CNS effects of overdose, not acute lethality.
| Target | Affinity | Rel. | Mechanism |
|---|---|---|---|
|
NMDA-R
Glutamate [NMDA] receptor · human
|
Ki = 540 nM
PCP/MK-801 site · pChEMBL 6.27
|
Open-channel block | |
|
NMDA PCP-site
[³H]MK-801 displ. · pig cortex
|
Ki = 429 nM
|
Channel blocker | |
|
GluN1/GluN2B
NMDA-R GRIN1/GRIN2B · human
|
Ki = 1.0 µM
HEK293 patch-clamp
|
Channel blocker | |
|
GluN2C
NMDA-R GRIN2C · rat
|
Ki = 700 nM
|
Channel blocker | |
|
GluN2B NTD
Ifenprodil site · rat cortex
|
Ki > 100 µM
no binding (pore-blocker, not NTD)
|
No effect | |
|
σ1R
Sigma-1 receptor (secondary)
|
Ki ~2.6 µM
|
Agonist | |
|
nAChR α7 / α4β2
Nicotinic ACh receptor (secondary)
|
IC50 ~1–8 µM
|
Antagonist | |
|
5-HT3
Serotonin-3 ion channel (secondary)
|
IC50 ~low µM
|
Antagonist |
Memantine binds the same pre-formed, water-filled vestibule as ketamine, and the rigid adamantane cage (zero rotatable bonds, per ChEMBL) surrenders essentially no internal conformational entropy on binding — even less than ketamine's two-rotor scaffold. The −TΔS budget is dominated almost entirely by desolvation of the symmetric hydrophobic cage and the release/reordering of pore waters around the N-site ring. In FlexAID∆S terms the Shannon-entropy collapse on binding lives in the solvent and side-chain microstates, not the ligand.
The interesting entropy story is kinetic, not thermodynamic. Memantine's low affinity and fast off-rate mean it never settles into the deep, fully-trapped, low-configurational-entropy bound state that locks in MK-801. Its bound ensemble stays "loose" — a shallow well with many escape microstates through open and partly-closed gates. A rigorous ΔG/ΔS decomposition should therefore weight the unbinding pathway entropy (a high-entropy, easily-accessed exit) as heavily as the binding desolvation term. That shallow, high-off-rate entropic profile is the molecular signature of a therapeutic channel blocker versus a dissociative one.