09 Jan 2006

The word "plasticity" means different things to different people. One of its common uses refers to the inherent ability of neuronal circuits to change and adapt to their environment. But scratch neuroscientists deep enough and they may admit to not knowing exactly how this plasticity occurs in humans. It could be a subtle molecular rearrangement that changes the flux through specific circuits, or it could be as dramatic as completely rewiring the axons and dendrites of the neuronal circuitry. Scientists think of these plasticities as functional versus structural, and while there is evidence that both can occur in the mammalian neocortex, the latter has only been consistently demonstrated after injury or some kind of external manipulation. But in the December 27 PLoS Biology, Elly Nedivi and colleagues at MIT reported that dendrites in the mouse neocortex spontaneously expand and contract. This finding might change the way scientists view neuronal plasticity and even raises the hope that future therapies designed to stimulate dendritic growth might be of benefit to those with certain neurodevelopmental or neurodegenerative disorders.

First author Wei-Chung Allen Lee and colleagues detected the structural plasticity when they used the powerful resolution of the two-photon microscope to visualize individual neurons inside the brains of 4- to 6-week-old living mice. Focusing their objectives on layer 2/3 of the visual cortex, they were surprised to see that all was not static. Images taken from exactly the same place, but at different times, revealed that dendrites, those arborlike branches that receive input from other neurons, appeared and disappeared with regularity.

Lee and colleagues monitored six pyramidal cells (named because of their shape) and eight nonpyramidal cells for three to 10 weeks. They found that each of the nonpyramidal cells had at least one, and as many as seven dynamic dendrite tips. In one neuron, four of 28 dendrite tips examined changed length. One of these branches extended by 16 μM over a 4-week period, while another extended by about 10 μM. In another nonpyramidal neuron, a single dendrite extended out of the field of view of the microscope. From age 11 to 13 weeks, that dendrite grew more than 90 μM. Overall, the authors found that about 14 percent (35 out of 259) of monitored dendrites grew (3 percent), retracted (2 percent), or did both (9 percent). By contrast, none of the 124 monitored pyramidal dendrites were dynamic, however, leaving the authors to conclude that “while dendritic branches of pyramidal cells remain stable, nonpyramidal interneurons in these layers are dynamic, exhibiting a range of structural changes on a week-to-week basis.” The authors found that the interneurons in question were γ-aminobutyric acid-, or GABAergic inhibitory neurons.

While this data indicates that the dendritic arbor can undergo spontaneous remodeling, the overall changes are small—only 1 to 5 percent of the total length of the dendritic arbor. This may explain why previous studies on remodeling have been inconsistent (for a review, see Chklovskii et al., 2004), but more importantly, it also raises questions of functional significance. “The functional test is a very hard one to pass,” said Karel Svoboda at the Howard Hughes Medical Institute at Cold Spring Harbor Laboratory in New York. He and other researchers, including Wen-Biao Gan at New York University School of Medicine, work to address this issue by looking at the dynamics of neuronal spines, which harbor the synapses that many neurons use to communicate with each other. Using this method, both groups have shown that spines turn over in pyramidal neurons of the cortex in response to stimuli. Using electron microscopy, Svoboda’s lab also showed that these spines do have synapses (see ARF related news story). Both groups also reported early last year that, though there are both persistent and transient spines in the pyramidal neurons of adult mice, the spines become more stable as the animals age (see Holtmaat et al., 2005 and Zuo et al., 2005). Many nonpyramidal dendrites do not contain spines. Despite this, Lee and colleagues were able to monitor the spines on one neuron, finding them to be motile. However, it is not clear if any new spines formed, or if any existing spines disappeared during the growth and retraction of the dendrites.—Tom Fagan