Like Scarlett O'Hara, every cell must ask itself: "Where shall I go? And what shall I do?" But DNA, even if it did give a damn, couldn't contain enough information to order all the body's cells into position.
Fortunately, it doesn't have to. For example, Lucia Galli-Resta, of the Institute of Neurophysiology, Pisa, has discovered that all cells need to assemble the complex structures of the eye's retina is a little self-knowledge and a few simple rules.
The retina contains many different types of cell in several layers. Within a layer, different cell types are interspersed, creating a 'retinal mosaic'. Within the retina, some neurons make horizontal connections, which help control the eye–brain pathway. One class of these neurons is the 'amacrine' cells.
Now Galli-Resta has worked out how new amacrine cells know what they are destined to become, and also how, once they get to the retina, they arrange themselves into the orderly patterns of a retinal mosaic.
In the developing embryo, retinal neurons are made in a rapidly dividing region of the central nervous system. They then migrate into their final positions (how they know where to go is not yet known).
Amacrine cells are the only neurons in the retina that communicate using a molecule called 'acetylcholine'. This makes them easy to track, as the chemical compositions of most cells overlap so much that they can't be identified before they have taken up their final shape and position. "The difficulty lies in the lack of markers to identify cells of a single type," says Galli-Resta.
For amacrine cells, he found two markers: two large protein molecules that make and transport acetylcholine. Some of the cells migrating to the retina contain these, he reports in
Next, Galli-Resta looked into what controls the assembly of the mosaic. Using computer modelling he found that a regular patterning of cells could result from a simple 'minimal distance' rule: if a cell excluded others of the same type from an area around it.
Galli-Resta tested whether amacrine cells were doing this, by investigating the interactions between them and another type of neuron called 'ganglion' cells, which surround, and connect to, amacrine cells in the retina.
He severed the developing optic nerve of newborn rats so that the animals would have a normal number of amacrine cells, but far fewer ganglion cells. He also studied a mutant mouse that produces about twice the normal number of ganglion cells.
The minimal distance rule seems to work well for amacrine cells, he found -- they adopted the same spacing, regardless of their surroundings. Apparently they recognize each other, but ignore other types of cell. But "there might be more than a single biological process going on," he adds.
Spacing patterns emerge early in development -- cells arrange themselves regularly even when there are only a few, and shuffle sideways to accommodate newcomers. Jeremy Cook, of University College London, and Leo Chalupa, of the University of California, Davis, described them as being "like passengers on a crowded railway platform -- [moving] enough to maintain a small 'personal space', but no more".
The difficulty of distinguishing different types of cell makes it hard to say how general a method this is of assembling complex structures, but the use of simple rules on a local scale, without any top-down masterplan, seems a good way to keep development economical and flexible. "The minimal-distance rule works for other cell types in the retina," says Galli-Resta, and he suspects it operates in other nerve centres.