ES cells can differentiate into any cell type when placed in a normal embryo. Can we control this amazing potential and harness it for experimental and therapeutic purposes? In principle it should be possible to culture ES cells in the presence of instructive molecules such as growth and differentiation factors and get them to form a desired cell type. Although people have appreciated this potential for about 15 years, only recently have realistic systems been developed. My laboratory and that of my colleague James Huettner have shown that we can "direct" ES cells to adopt a neural fate. To do so we start with normal ES cells. These are then cultured for 4 days as aggregates (which are termed embryoid bodies (EBs)). Next, retinoic acid (RA) is added to the cultures for an additional 4 days. Our studies have demonstrated the following main points:

1. the majority of cells become committed to the neural lineage and make neural progenitor cells.
2. Neural progenitor cells differentiate into neurons and glia.
3. The neurons have the cardinal properties of normal CNS neurons. They are post-mitotic and have axons and dendrites. Axons and dendrites form chemical synapses with one another. Neurons diversify in terms of neurotransmitter type. About 75% are glutaminergic, 20% GABAergic and 5% glycinergic. Synaptic transmission for each appears just like that in synapses in the brain.
4. The developmental pathway leading to neural cells mimics the pathway found in the early embryo. Thus most of the control switches utilized in this system are probably the same ones used in the normal developing nervous system.

Potential uses of this ES cell-based system are the following:

1. As an invesitgational tool for studying mechanisms of nerve cell growth, injury, repair and models of human diseases. Note that any engineered ES cell line can now be differentiated in a dish. This provides a powerful way of quickly viewing the effect of a mutation or other genetic alteration. The same cells can be used to make mice if desired.
2. As a source of cells for transplant research. The system provides large numbers of cells from all stages of the neural developmental lineage. Cells can be engineered to express markers useful for transplant experiments that allow cells to be easily detected against the background of host cells. Finally, the ES cells can be engineered to express growth factors and other important regulatory proteins with any desired spatial or temporal specificity.
3. As a potential source of therapeutic transplants. ES cells have already been obtained from the rhesus monkey ( work of J. Thompson, University of Wisconsin). The likelihood of getting human ES cells is high. Presumably such cells could be differentiated in ways similar to mouse cells.