Long-term potentiation (LTP) is an activity-dependent form of synaptic plasticity which provides a compelling model for the synaptic basis of learning. First described in the HIPPOCAMPUS (Bliss & Lømø, 1973; see Bliss & Collingridge, 1993 for review), LTP is also found in other brain regions, including NEOCORTEX and AMYGDALA. In hippocampal slices, or in anaesthetized animals, LTP induced by brief tetanic stimulation (see TETANUS) can last for several hours; in animals with chronically implanted electrodes (see ELECTRODE) it can persist for several days. The induction of LTP requires the tetanus to be above a threshold intensity (the property of cooperativity). LTP displays two other properties which enhance its attraction as a mnemonic device: input specificity which limits potentiation to activated synapses, and associativity, which allows LTP to be induced by a weak sub-threshold tetanus to one pathway, provided it is coincident with a strong tetanus to a second, converging pathway (see HEBBIAN SYNAPSE).
In most cases, the induction of LTP is controlled by postsynaptically located NMDA RECEPTORS. (An exception is hippocampal LTP at synapses made by granule cell axons onto CA3 pyramidal cells). Specific antagonists of the NMDA receptor, such as APV (D-aminophosphonovalerate), suppress the induction of LTP, without affecting the predominantly AMPA receptor-mediated synaptic responses evoked by single stimuli. Once LTP is induced, however, NMDA receptors appear to play no further role in its expression. At resting levels of membrane potential, the NMDA receptor channel is blocked by magnesium (Mg2+) ions, and the opening of the channel requires both the binding of transmitter and strong local DEPOLARIZATION; it is this voltage dependence of the NMDA receptor which leads directly to the properties of cooperativity, associativity and input specificity. Activation of the NMDA receptor is both necessary and sufficient for the induction of LTP. The block of LTP by APV establishes the former; the latter is shown by experiments in which LTP is induced by procedures which satisfy the requirements for activation of the NMDA receptor—for example, by repeatedly pairing single afferent stimuli with depolarizing pulses delivered to the target cell through an intracellular electrode. All this can be summarized in the following induction rule for LTP: a synapse will be potentiated if and only if it is active at a time when the dendrite on which it is located is strongly depolarized. Whether this rule is rigorously obeyed is not known; there are indications that input specificity, for example, is not absolute. Since the NMDA channel is permeant to calcium (Ca2+) ions, and since the induction of LTP is blocked by postsynaptic injection of Ca2+ chelators (see CHELATING AGENT), the immediate trigger for the induction of LTP is probably the entry of Ca2+ through activated NMDA channels.
Two temporal components of LTP have been distinguished: early-LTP, lasting 4–6 hours, and late-LTP. Late-LTP is blocked by protein synthesis inhibitors providing these are present around the time of induction. Early-LTP is sensitive to PROTEIN KINASE inhibitors, in the presence of which only SHORT-TERM POTENTIATION (STP), lasting about an hour, survives. It is likely that early-LTP is maintained in part by postsynaptic changes, including PHOSPHORYLATION of GLUTAMATE receptors, and in part by presynaptic modifications leading to a sustained increase in transmitter release. Since induction is a postsynaptic process, any presynaptic contribution to the expression of LTP implies the existence of a mechanism which informs the presynaptic side of the synapse about the postsynaptic inductive event. A number of candidates have been proposed for the role of ‘retrograde messenger’, including NITRIC OXIDE and ARACHIDONIC ACID. Late-LTP may be mediated by structural changes involving both pre- and postsynaptic elements. The requirement for GENE TRANSCRIPTION in late-LTP suggests that a signal passes from the activated synapses to the soma to initiate transcription, and that a gene product then travels back from the soma to act only on the activated synapses. There is evidence that the activated synapses are tagged in some way which permits them to selectively respond to the SOMATOFUGAL message(s).
How good is the evidence that LTP is exploited by the brain to store information? Early experiments were consistent with the hypothesis: when infused into the hippocampus, the NMDA receptor antagonist APV blocked both LTP and LEARNING in the WATER MAZE. However, subsequent experiments have shown that once rats have learnt the task, APV does not prevent relearning in a different context. More recently, studies of second-generation transgenic animals, in which LTP in the hippocampus can be modulated on demand (Mayford et al., 1997), and experiments showing that in fear conditioning, the auditory CONDITIONED STIMULUS evokes a response in the AMYGDALA which is potentiated by an LTP-like process as the result of training (Rogan et al., 1997) have provided powerful new evidence in favour of the hypothesis.
Bliss T.V.P. & Lømø T. (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. Journal of Physiology 232:331–356.
Bliss T.V.P. & Collingridge G.L. (1993) A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361:31–39.
Mayford M., Mansuy I.M. Muller R.U. & Kandel E.R. (1997) Memory and behavior: a second generation of genetically modified mice. Current Biology 7: R580-R618.
Rogan M.T., Staubli U.V. & LeDoux J.E. (1997) Fear conditioning induces associative long-term potentiation in the amygdala. Nature 390:604–607
T.V.P.BLISS
This is the complete article, containing 879 words
(approx. 3 pages at 300 words per page).
