Excitotoxicity

Excitotoxicity begins with too much glutamate acting for too long at ionotropic receptors, principally the NMDA receptor, whose channel admits calcium once the magnesium block is relieved by depolarisation. The calcium load activates calpains, neuronal nitric oxide synthase and phospholipases, and overwhelms mitochondrial buffering until the permeability transition opens, collapsing membrane potential and generating reactive oxygen species. The process self-amplifies under energy failure, because the sodium gradient driving glutamate uptake collapses and astrocytic transporters run in reverse, releasing glutamate instead of clearing it.

It is the dominant mechanistic account of the ischaemic border zone and of part of traumatic brain injury, and it has been the most heavily targeted and least successful mechanism in stroke drug development. A long series of NMDA antagonists and free radical scavengers reduced infarct volume in rodent occlusion models and then failed in human trials, the free radical trapping agent NXY-059 being the best documented. Two drugs on this axis are approved elsewhere: memantine, a low-affinity uncompetitive NMDA antagonist, in moderate to severe Alzheimer's disease, and riluzole in amyotrophic lateral sclerosis, both with modest effects.

The interpretive lesson is that excitotoxicity runs on a very short clock, over minutes to a few hours. Any agent aimed at it must be present inside that window, so a model that dosed before or at the moment of occlusion tested something no clinical pathway can deliver.

The specific error to watch for is a neuroprotection claim built from cell culture. Glutamate challenge in a dish is usually applied at concentrations and durations far beyond anything an intact brain sustains, and protection at micromolar drug concentrations means little when unbound brain exposure would be orders of magnitude lower.