The diffusion of a dye (left to right) is slowed by spines, especially those with large head/neck diameters (inset, right).
AUGUSTINE/ELSEVIER
Spines are thought to act locally by preventing the chemical messengers made at a synapse from leaking into the rest of the dendrite. But Augustine turned this idea around: he wondered whether spines affect the long-distance journey of chemicals along dendrites. Using fluorescent imaging, the group now shows that dendritic diffusion of both an inert dye and calcium-mobilizing IP3 is slowed when spine density is high.
Slowing occurs because signals that enter a spine do not always escape quickly. How long they are delayed depends on spine geometry. Manipulating spine shape in vivo is difficult, so the group turned to computational modeling. They found that the spine property most critical to trapping is the ratio of the spine neck and spine head diameters. “With a constant head volume on a narrow neck, a cylindrical head is not a strong trap,” Augustine says. “If the head is more like a [sideways elongated] pancake, it's a huge trap.”
Other small dendritic travelers that are not actively transported might also be slowed by spines, including other second messengers, nucleotides, and even mRNAs. Calcium, however, was not slowed by spines because it did not travel far enough.
Augustine thinks that the experimental reliance on calcium has bolstered the prevailing view of spines as traps for synapse-generated signals. “But calcium is an exception,” he says, “because cells work really hard to buffer and pump calcium.” He points out that longer-lasting signals such as IP3 escape into dendrites just fine.
For now, the physiological outcome of slowed diffusion is anyone's guess. “Until we know more about what's getting trapped and under what circumstances,” says Augustine, “it's just speculation.” Still, he offers one tasty possibility. Spines swell during LTP, which might make them better able to trap more LTP-inducing proteins coming from the cell body. 
Reference:
