Scientists have succeeded in remyelinating damaged areas of the central nervous system and restoring lost function by transplanting glial cells directly into the spinal cord of animals. This reconstruction feat, until recently thought impossible, raises hopes of finding treatments for diseases with persistent myelin loss, such as multiple sclerosis and the leucodystrophies. Bill Blakemore, professor of neuropathology at the University of Cambridge, told neuroscientists at a Novartis Foundation symposium in London last week, that his team had identified the cells with the greatest capacity to make myelin sheaths. Grafting cells into lesions to replace the lost function seems to be a safe approach. Clinical experience with transplants of 8-10 week old fetal tissue into the brains of Parkinson's disease patients has been growing over the past few years. The results are reassuring: the cell therapy is both successful in improving symptoms and carries no obvious hazards. But although cell transplantation is opening up the possibility of finding treatments for disorders of the brain and nervous system, including Parkinson's disease, stroke, Huntington's chorea, and motorneurone disease, the consensus reached at the meeting was that relying on fresh fetal neural tissue remains an obstacle. In addition to the ethical objections, there is the practical disadvantage that a single fetus does not yield enough cells for one transplant. Several novel approaches may circumvent this problem for brain cell transplants. One promising alternative is to use immortalised neural stem cells, which have been modified by genetic engineering to grow indefinitely in culture. The advantage of this strategy, currently at the laboratory stage, is that the immortalised cells can be expanded to large numbers and be kept in the fridge until the best time for transplantation. Whether these cells will be as effective as primary foetal grafts, however, has yet to be determined. A different approach is to transplant polymer-encapsulated cells genetically engineered to deliver nerve growth factors direct to the brain. This technique has been tested in patients with Huntington's chorea and motorneurone disease, and the first clinical trials are in progress. The situation is different for remyelination. "Growing oligodendrocytes is much easier [than growing nerve cells], but we still need buckets of cells," said Professor Blakemore, who admits that the large numbers of fetal cells needed for a successful transplant cannot be provided. Today, however, it is possible to keep human embryonic stem cells in culture. These pluripotent cells are capable of developing into virtually any cell in the body, and prior experience with embryonic stem cells in rodents indicates that it may be relatively straightforward to grow large quantities of human oligodendrocyte precursors for glial cell transplantation. Yet there is an important caveat. For widespread myelination, the cells need to migrate, and transplantation studies with cells from rodents and pigs tell researchers that the area they remyelinate is small, about 1-2 mm3. In humans, however, the lesions are huge, so if implanted cells are to restore function they will have to migrate considerably. "We can get myelination in our animal models, but that does not tell us with any degree of certainty how extensive this can be [in humans]. And we will not know how effective glial cell transplants will be until we try."