Does BDNF Have Pre- or Postsynaptic Targets?

See allHide authors and affiliations

Science  01 Mar 2002:
Vol. 295, Issue 5560, pp. 1651-1653
DOI: 10.1126/science.1070163

Learning and memory depend on the persistent modification of synapses between brain neurons. New memories are formed in the hippocampus, which is situated beneath the cerebral cortex, and damage to the hippocampus impairs certain kinds of learning in humans and animal models. In this brain region, excitatory synaptic transmission depends on glutamate, the most common excitatory amino acid neurotransmitter in the central nervous system (CNS). This neurotransmitter is released into the synapse by the presynaptic neuron and binds to glutamate receptors on the postsynaptic neuronal membrane. There are two sorts of glutamate receptor: AMPA receptors, which regulate basal synaptic transmission, and NMDA receptors, which modulate a specific type of synaptic plasticity called long-term potentiation (LTP) that may form the cellular basis of learning and memory (1). NMDA receptor activation increases the Ca2+ concentration in the postsynaptic neuron, activating intracellular biochemical pathways that result in persistent modification of postsynaptic AMPA receptor numbers and activity. In addition to AMPA and NMDA receptors, many other molecules localized at synapses regulate LTP induction. For example, a neurotrophin called brain-derived neurotrophic factor (BDNF) is thought to act presynaptically to induce LTP. On page 1729 of this issue, Kovalchuk et al. (2) reveal that BDNF in concert with glutamate induces LTP by binding to receptors on the postsynaptic neuronal membrane.

BDNF and other neurotrophins such as nerve growth factor (NGF) are trophic substances that promote the survival or differentiation of neurons. They are also crucial for synaptic transmission and plasticity (3). In contrast to their slow effect on neuronal survival or differentiation (hours or days), their modulation of synaptic transmission is much faster (seconds or minutes). A pioneering study by Lehof et al. (4) showed that treating cultured neurons and muscle cells with exogenous BDNF potentiates synaptic transmission at the developing neuromuscular junction within minutes by enhancing neurotransmitter release from presynaptic terminals. Subsequently, similar acute effects of exogenous BDNF on synaptic transmission in the CNS were reported. BDNF potentiates excitatory synaptic transmission in the CA1 region of the hippocampus (5) and has acute effects on both excitatory and inhibitory synapses in the CNS, although BDNF's effects on excitatory synapses are still controversial (3).

Several lines of evidence indicate that these actions of BDNF are mediated by the TrkB receptor tyrosine kinase. It has been shown pharmacologically and genetically that binding of BDNF to TrkB receptors is essential for LTP induction in the hippocampus. Mice engineered to lack BDNF have impaired hippocampal LTP (6), and conditional knockout mice that lack TrkB receptors in the forebrain during the postnatal period show reduced hippocampal LTP and impaired learning behavior (7). Furthermore, LTP is blocked when endogenous BDNF is reduced by a TrkB-immunoglobulin G fusion protein that scavenges BDNF (8). These findings strongly suggest that when endogenous BDNF binds to TrkB receptors, permissive and/or instructive signals are generated that induce LTP. It has been presumed that these actions of BDNF are mediated presynaptically, leading to modulation of neurotransmitter release (although LTP in the CA1 hippocampal region and dentate gyrus is known to be induced postsynaptically). However, because the manipulations in these studies should have affected both pre- and postsynaptic BDNF-TrkB pathways equally, it is unclear whether the action of BDNF is solely due to presynaptic modification.

Enter Kovalchuk et al. (2) with their study showing that the postsynaptic BDNF-TrkB pathway is crucial for regulation of excitatory synaptic transmission and LTP induction. Using electrophysiology in mouse hippocampal slices, the authors show that exogenously applied BDNF induces depolarization of postsynaptic dentate granule cells, presumably through activation of Na+ channels (9) and Ca2+ ion influx through Ca2+ channels. In contrast, BDNF has no effect on presynaptic perforant path fibers originating from cortical neurons, which form synapses with the dentate granule cells. An increase in Ca2+ concentration in dentate granule cells induced by BDNF, but not their depolarization, is blocked by a Ca2+ channel inhibitor D890. Postsynaptic voltage clamping, which prevents depolarization of dentate granule cells, also blocks an increase in Ca2+ concentration in response to BDNF. These findings suggest that the primary action of BDNF is to induce postsynaptic depolarization of dentate granule cells leading to voltage-dependent opening of Ca2+ channels (see the figure). BDNF application alone has no effect on synaptic transmission. However, when it is combined with a weak burst of presynaptic activity that itself has little effect on synaptic transmission, persistent enhancement of synaptic transmission is induced, similar to that seen with tetanic stimulation. Furthermore, this persistent enhancement is blocked by the D890 inhibitor or by an NMDA receptor antagonist, demonstrating that BDNF-mediated LTP is induced postsynaptically.

Postsynaptic modulation by BDNF.

The neurotrophin BDNF postsynaptically modulates hippocampal synaptic transmission and plasticity (2). Activation of postsynaptic TrkB receptors by BDNF causes depolarization of the postsynaptic neuron, presumably through opening of Na+ channels (9) and concomitant activation of Ca2+ channels, resulting in an increase in the Ca2+ concentration. During synaptic transmission, postsynaptic cell depolarization caused by the activation of TrkB receptors enhances NMDA receptor opening by removing a Mg2+ block from the receptor channel, which then facilitates the induction of LTP in the postsynaptic cell.

The Kovalchuk et al. study clearly shows that TrkB receptors are present on the postsynaptic neuron and that exogenous BDNF has some effect on postsynaptic excitability and Ca2+ signaling in the dentate granule cell. Furthermore, the authors point out that the BDNF-TrkB pathway can regulate the induction of NMDA receptor- dependent LTP. Their work provides concrete evidence for the importance of BDNF in postsynaptic regulation of LTP. However, because the investigators only examined TrkB receptor activity by applying BDNF exogenously, it will be important to show in future experiments that the same phenomenon is induced by endogenous BDNF, and to identify the site of BDNF release. Although this study strongly suggests that the target of BDNF is TrkB receptors in the postsynaptic neuronal membrane, it remains possible that presynaptic TrkB receptors are also involved in synaptic transmission and plasticity. This puzzle could be solved by engineering conditional knockout mice that do not express BDNF or TrkB receptors in specific brain regions, such as the dentate gyrus or the CA1 and CA3 regions of the hippocampus.


View Abstract

Navigate This Article