Long-term potentiation · Grey Matter

Long-term potentiation (LTP) is a lasting increase in the strength of a synapse produced by strong, coincident activity of the two cells it joins, and it is the best-studied cellular mechanism of learning and memory.

Long-term potentiation. Long-term potentiation (LTP) is a lasting increase in the strength of a synapse produced by strong, coincident activity of the two cells it joins, and it is the best-studied cellular mechanism of learning and memory.

It was first reported in 1973 by Bliss and Lømo, who found that a brief high-frequency train of stimuli to a pathway into the rabbit hippocampus made that pathway's synapses respond more strongly for hours. At most excitatory synapses in the hippocampus and cortex the trigger is the NMDA receptor: when glutamate release coincides with a depolarised postsynaptic membrane, the magnesium block is expelled and calcium flows into the dendritic spine. A large, fast calcium rise activates the enzyme CaMKII, which keeps itself switched on by phosphorylating its own subunits. CaMKII phosphorylates AMPA receptors so that they conduct more, and helps trap additional AMPA receptors in the synapse; the same glutamate now produces a larger current.

It has phases. Early LTP, lasting an hour or so, uses existing proteins; late LTP, lasting days, needs new gene expression and protein synthesis and comes with enlargement of the spine.

It is specific to the synapses that were active, and it is associative: a weak input active at the same time as a strong one is potentiated too, which is how two coincident signals become linked.

Its link to memory is strong but indirect. Blocking NMDA receptors or LTP-related molecules impairs some kinds of learning in animals, and learning produces LTP-like changes, but a single memory is spread over many synapses and changes.

LTP is Hebb's rule built into a synapse.

Calcium entering through the NMDA receptor at the moment of coincidence becomes more AMPA receptors, and the connection that helped fire the cell is heard more loudly afterwards.

Questions: How does calcium through NMDA receptors make a synapse stronger? Calcium entering a dendritic spine through NMDA receptors binds calmodulin, which activates the kinase CaMKII; CaMKII then phosphorylates its own neighbouring subunits and stays active after the calcium has gone. Active CaMKII phosphorylates AMPA receptors so that each one conducts more current, and helps capture additional AMPA receptors at the synapse, so the same release of glutamate produces a larger response. This early phase lasts an hour or so; keeping the change for days needs new proteins and growth of the spine. The direction depends on the calcium signal: a large, fast rise drives this potentiation, a modest, prolonged one drives depression instead. How can the same NMDA receptor make a synapse stronger or weaker? Both directions start with calcium entering a spine through NMDA receptors, and the pattern of that calcium decides which way the synapse goes. A large, fast rise, produced by strong coincident activity such as a high-frequency burst, activates kinases (above all CaMKII) that add and strengthen AMPA receptors, which is LTP. A modest, prolonged rise, produced by low-frequency activity such as 900 pulses at 1 Hz, favours phosphatases that dephosphorylate AMPA receptors and pull them out of the membrane, which is LTD. NMDA receptor blockers prevent both, which is how Dudek and Bear showed in 1992 that this form of LTD needs the same receptor as LTP. How stable are dendritic spines in the adult cortex, and does learning make new ones? Imaging the same dendrites over weeks in the mouse neocortex shows that in six-month-old mice about three quarters of spines persist for at least eight days, and spines that survive that long almost all last the rest of the observation; a minority are transient, appearing and vanishing within a few days. Turnover is much higher in young animals and falls with age. Learning a new motor or sensory task adds new spines on particular dendrites, some of which stabilise along with the skill, while potentiated spines enlarge. The adult cortex is therefore mostly stable, with a small turnover that learning can bias. Is long-term potentiation the same thing as Hebb's rule? Hebb's rule (1949) is a principle: a connection that repeatedly helps fire a cell is strengthened. LTP, found by Bliss and Lømo in 1973, is a measured phenomenon: synapses become stronger for hours after strong activity. NMDA-dependent LTP has the properties the rule asks for (it needs presynaptic release and postsynaptic depolarisation together, and it is specific to the active synapses), so it is the best candidate mechanism for Hebbian learning. They are still distinct: some forms of LTP are not Hebbian, and the rule as stated says nothing about weakening, which LTD and spike-timing rules add. What physically changes at a synapse that has been potentiated? Mostly the postsynaptic side gains AMPA receptors: more of them are inserted and held at the synapse, and existing ones are phosphorylated so each conducts more, so one vesicle of glutamate now produces a larger current. Within minutes to hours the spine head grows, its scaffold of proteins enlarges to hold the new receptors, and in late LTP new proteins are made to keep the change. Some forms also raise the probability of release on the presynaptic side. Synapses that had NMDA receptors but almost no AMPA receptors (silent synapses) can be switched on this way, which is common in development.