Stem cells can transform into whatever cell the body tells them to. Unfortunately, scientists have yet to master that particular gift of gab. But investigators at Stanford University may soon crack the language with tiny "chat rooms" for stem cells.

In their natural milieu, stem cells have a variety of neighbors that pass on chemical messages at exact spots at particular times in specific amounts to guide the cells' development into a given cell type. In today's laboratory, however, researchers often bathe the whole cell with chemicals--kind of like out-of-control beer keggers compared with the sophisticated cocktail parties the body normally throws for stem cells.

To uncover the mostly unknown placement, timing and identity of the cues, Stanford materials scientist Nicholas A. Melosh and his colleagues are re-creating the niche where stem cells normally dwell. They are developing a microscopic lab on a silicon chip that surrounds a stem cell with as many as 1,000 cavities, each 500 nanometers wide. The nano reservoirs each hold roughly an attoliter (1018 liter) of liquid--comparable to the size of cellular secretions--and are sealed with the same type of lipid bilayer that makes up cell membranes. Tenths of a volt open pores in these layers, so that "when researchers want to deliver a specific chemical to the cell at a particular stage in its development, they will merely have to press a button," Melosh remarks. The team is now working to grow stem cells derived from adult fat.

In addition to growth factors, scientists could try alternative means to direct stem cells, adds Richmond Wolf, director of technology transfer at the California Institute of Technology. He points to suppressing gene expression using RNA interference.

Melosh also hopes to use their invention to grow tissues from stem cells, layer by layer. This ability could permit the growth of compound tissues that are, for instance, bone on one side and cartilage on another. "Right now if you tear cartilage, you have to screw it back on. There is no way yet to reproduce that interface between bone and cartilage," Melosh states. The hope is to build composite tissues where the artificially grown cartilage naturally bonds with the body's.

A concern is that the chemicals in the nano reservoirs could react with the lipids. So the researchers hope to replace the lipid seals with nonreactive gold ones, which they can dissolve with electric current, if necessary. The voltages used to open seals could also affect the stem cells, but Melosh explains they could solve this problem by recessing the pores so that the electric fields are more distant from the cells.

Standard electronics industry manufac-turing can fabricate the device so that it could make it to market in five to eight years, Melosh predicts. But he and colleagues will use it for experiments well before then. "I could see research-level products being used by the beginning of next year if all goes well," Wolf adds.