The standard view of how a neuron functions is described in hundreds of textbooks: dendrites receive inputs, in the form of action potentials, from other cells. These signals travel to the cell body, where they are integrated. Action potentials are then generated in the proximal region of the axon, and, from there, propagated to the nerve terminal. Here, synaptic vesicles containing neurotrans- mitters release their contents into the synaptic cleft by a process called exocytosis, during which the vesicles fuse, either fully or partially, with the presynaptic membrane. The transmitter molecules diffuse across the synapse and bind to receptors on the dendrites of the postsynaptic cell, in which the same process is initiated.
This model of nerve cell function has served neuroscientists well for over a hundred years: anatomists and histologists determined the fine structure of the neuron in the latter half of the nineteenth century, and Sir Charles Sherrington coined the term “synapse” in 1897. The model is generally accepted because it provides an explanation of how a brain consisting hundreds of billions of cells, all of which produce action potentials in basically the same way, can work: information is encoded in the patterns of action potentials, and the connectivity of cells is what determines the specificity of neuronal signalling. Thus, action potenitals generated in one cell population will encode different information from signals produced by other populations of cells, because each population projects to a different target region, and therefore release transmitters only onto those target cells.
Recently, researchers have obtained evidence that neurotransmitters can be released from sites other than the presynaptic membrane. For example, at the terminals of climbing fibres, which originate in the inferior olivary nucleus and project to the cerebellar cortex, glutamate is released at sites that do not face the synapse. However, this and similar findings are still consistent with the view that neurotransmitter release occurs at nerve terminals. But now, two papers, published back-to-back in the March issue of Nature Neuroscience, provide evidence that neurotransmitters can be released along the length of axons, suggesting that the classical model of how neurons function may be inadequate.
Both papers describe experiments performed on rodents, and both show that axons in the corpus callosum are capable of releasing the neurotransmitter glutamate, to which adjacent glial cells are responsive. The corpus callosum is the bundle of approximately 100 million axons which connects the two hemispheres of the brain and enables them to communicate with each other. It consists almost entirely of white matter – it contains axons of neurons projecting from one side of the brain to the other, but is devoid of cell bodies or dendrites, and therefore is not a part of the brain where one would expect neurotransmission to take place.
Dirk Dietrich and his colleagues, from the Experimental Neurophysiology Laboratory at the University of Bonn’s Department of Neurosurgery, prepared slices of tissue containing the corpus callosum from juvenile rats. At 8-16 days of age, the rat corpus callosum contains large numbers of oligodendrocyte precursor cells (OPCs). Mature oligodendrocytes are the cells which ensheath axons with a fatty protein called meylin; this insulates the fibres and therefore increases the velocity at which they conduct nervous impulses.
Because OPCs are known to express a variety of voltage-gated ion channels, Deitrich’s team used the patch clamp technique to record the electrical activity of OPCs. When adjacent corpus callosum axons were stimulated, inward currents were recorded from the OPCs. These currents were recorded from every OPC into which the microelectrodes were inserted, and were always dependent upon stimulation of the adjacent axon. The researchers then used glutamate receptor antagonists to show that the currents are mediated by a type of glutamate receptor called the AMPA receptor; these receptors mediate fast synaptic transmission in the central nervous system. It was also found that OPCs are sensitive to lower concentrations of glutamate than are postsynaptic neurons, and that axons of the rat optic nerve are also capable of releasing glutamate. As in nerve terminals, glutamate is released from both corpus callosum and optic nerve axons in response to local increases in the concentration of calcium ions (calcium “microdomains”) in the areas around calcium channels in the axonal membrane.
The molecules involved in synaptic transmission – neurotansmitters and the constituents of synaptic vesicles – are synthesized in the cell body and transported along the axon to the nerve terminal, and, because there is a constant turnover of these proteins, the components of neurotransmission are always present along the length of axons. It is, therefore, possible that the currents recorded in the OPCs were generated by leakage of these components from the axonal transport apparatus. To rule out this possibility, OPCs were continuously stimulated for periods of several seconds. The cells continued to release glutamate, at an estimated 8 times per second, showing that the release was not due to the sporadic leakage of neurotransmitters being transported along the axon.
Electron microscopy was then used to examine the callosal axons. This revealed the presence of synaptic vesicle-like structures in the axons; it showed that at least some of these vesicles were docked at the axonal membrane; and that vesicles vesicles are recycled quickly after use. Thus it appears that the release of neurotransmitters from axons in the corpus callosum occurs in the same way as at classical synapses – by exocytosis. The vesicles in the axons cause small protrusions in the membrane, which form invaginations in the OPCs. However, although the structures in the axon and the OPCs at which transmission occurs are closely apposed to each other, the area between OPCs and axons, into which the neurotransmitter is released, is wider than the cleft of classical synapses. It has an irregular width, and, unlike well-characterized synapses, lacks an extracellular matrix.
Electron micrograph showing a biocytin-laballed oligodendrocyte precursor cell (Bio) and an adjacent callosal axon containing synaptic vesicle-like structures. The region of the axon at which the vesicles are docked causes an invagination of the oligodendrocyte precursor cell membrane. (Scale bar = 200nm; from Kukley, et al, 2007.)
Deitrich’s team examined this surprising phenomenon only in juvenile mice. The other paper, by Ziskin et al, describes similar experiments performed in mice, but also shows that axo-glial neurotransmission occurs in the corpus callosum of adult animals. Both papers therefore provide strong evidence that neurotransmitter release is not restricted to nerve terminals, but can also occur at discrete regions along the entire length of the axon. This neuron-glial signalling may provide a means of controlling oligodendrocyte differentitation during neural development. It remains to be seen whether or not this type of signalling takes place in other brain regions.
These findings show that our understanding of basic neuronal function is far from complete, and add yet another level of complexity to information processing in the brain. They also have potential medical applications, as excess release of glutamate can be damaging to nerve cells. This excitotoxicity has been implicated in a wide variety of conditions, including autism and epilepsy, as well as neurodegenerative diseases such as Alzheimer’s and Parkinson’s. The findings may therefore open up new avenues for the development of treatments for some of these conditions, based on drugs that target glutamate release in the white matter.
Kukley, M. et al. (2007). Vesicular glutamate release from axons in white matter. Nat. Neurosci. 10: 311 – 320. [Abstract]
Ziskin, J. L., et al. (2007). Vesicular release of glutamate from unmyelinated axons in white matter. Nat. Neurosci. 10: 321 – 330. [Abstract]