Among the changes in cerebellar synaptic efficacy that are thought to underlie motor learning, the best-studied is long-term depression (LTD). More than 20 years of intensive research attest to the complexity of this phenomenon, which is modulated by at least 30 interacting molecules. New data on one of these modulators — protein phosphatase-2A (PP-2A) — published in the Proceedings of the National Academy of Sciences bring us closer to understanding how LTD is regulated.

A team led by Masao Ito applied the PP-2A inhibitors fostriecin, cytostatin or microcystin-LR to mature Purkinje cells from dissociated cerebellar cultures. Whole-cell voltage-clamp recordings showed that the amplitude of excitatory postsynaptic currents evoked by stimulation of a single granule cell was gradually depressed under these conditions. This depression was abolished by the calcium chelator BAPTA and by the kinase inhibitor staurosporin. By contrast, a metabotropic glutamate receptor antagonist had no effect.

Application of both cytostatin and AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid) in the presence of tetrodotoxin to prevent granule cell firing, reduced the amplitude of miniature excitatory postsynaptic currents and the number of AMPA receptor clusters per dendrite. But declustering did not occur in the presence of cytostatin alone. So, PP-2A might have a prominent role in LTD by regulating AMPA-receptor density at synapses through a use-dependent mechanism involving protein kinase activity, calcium signalling and activation of AMPA receptors.

The reversibility of LTD is crucial to its putative function in information storage. Another recent study also published in the Proceedings of the National Academy of Sciences used spike-probability measurements to shed light on interactions between long-term potentiation (LTP) and LTD in the mammalian cerebellum.

Extracellular monitoring of Purkinje cell spikes in rat cerebellar slices allowed Roger Tsien and colleagues to stably record for several hours. During this extended period, LTD could be saturated, LTP could be subsequently induced by various stimuli, and reinitiation of LTD could be attempted. Two types of LTP — presynaptic LTP evoked by 120 stimulations of parallel fibres at 4–8 Hz (4 Hz LTP) and postsynaptic LTP induced by 300 stimuli at 1 Hz (1 Hz LTP) — had opposite effects on postsynaptic LTD induced by synchronous stimulation of parallel fibres and a climbing fibre. A single round of 1 Hz LTP reversed LTD and allowed it to be reinduced, showing that these two forms of synaptic plasticity act in opposite directions on the same target (presumably postsynaptic AMPA receptors). By contrast, 4 Hz LTP did not reverse LTD, reflecting its presynaptic site of expression.

Like 1 Hz LTP, pharmacological induction of LTP with nitric oxide reversed LTD, whereas treatment with forskolin to elevate cyclic AMP levels did not. These results lend support to the model in which nitric oxide participates in the induction of synaptic plasticity, which takes the form of LTD or LTP depending on coincident elevation of postsynaptic calcium.