Using stem cells for clinical therapies is an idea still bathed in a futuristic glow, but one such treatment already has a history of success going back almost 40 years. Tens of thousands of patients treated with bone marrow transplants have shown that an infusion of healthy stem cells can regenerate a failing body part. In most of these cases, the patients suffered from congenital blood or immune disorders, or their bone marrow had been damaged by cancer treatment. As a result, the haematopoietic stem cells in their marrow, which normally produce billions of blood and immune cells daily, needed replacing.

Since 1968, these transplants have triumphantly repaired patients' capacity to manufacture healthy blood and immune cells. Over the past decade, as scientists discovered additional stem cell types throughout the human body, enthusiasm has grown for the possibility that other failing body parts might also be regenerated with a transplant of stem cells.

Yet the more researchers learn about the characteristics and behaviour of adult stem cells, the less they seem to agree on answers to some fairly fundamental questions, such as what these cells really are, where they originate, what they are capable of doing, and how they do it. Consequently, although adult stem cells may not provoke much political rancour today, they have become more scientifically controversial than their embryonic counterparts.

Fortunately, the majority of scientists can at least agree on a basic definition: a stem cell (whether adult or embryonic) must renew itself indefinitely through cell division, while remaining in its generic state and retaining its potential to give rise to daughter cells of more specialised types. These progeny often start out only partially differentiated themselves, with some flexibility to serve as progenitors of several cell varieties within a particular organ or system. For example, descendants of mesenchymal stem cells found in bone marrow can become bone, as well as cartilage, fat cells, various kinds of muscle and the cells that line blood vessels.

Although the tissues that sprout from these bone marrow stem cells are seemingly diverse, they have one thing in common: when the human body is first forming, they all originate in the middle layer, or mesoderm, of the developing embryo. This fact is at the heart of one of the most important questions debated by stem cell scientists: whether adult stem cells can transdifferentiate, that is, produce functional new tissues outside the lineage of their embryonic layer. The answer could be crucial to some of the more ambitious regenerative therapies based on adult stem cells.

Traditionally, adult stem cells have been considered limited in their potential, able only to manufacture cell varieties within their own lineage. Hence, they are usually described as multipotent, rather than pluripotent like embryonic stem cells. In recent years, however, many research groups have claimed to have made adult stem cells cross lineage lines---for example, by turning haematopoietic stem cells into liver, neural stem cells into blood vessels and mesenchymal stem cells into neurones.

In 2002 Catherine Verfaillie of the University of Minnesota first described a new adult stem cell from bone marrow that could produce cell types of all three embryonic lineages. Dubbing it a multipotent adult progenitor cell (MAPC), Verfaillie speculated that its flexibility might equal that of embryonic stem cells. Indeed, she thought MAPCs might be left over from embryonic development to serve as a universal repair mechanism for the adult body.

Such a one-size-fits-all adult stem cell would certainly solve the problem of regenerating tissues where no local progenitors have been discovered, such as in the adult heart, or where local stem cells are extremely rare and difficult to obtain, as in the brain. Unfortunately, other investigators have had difficulty reproducing some of the original MAPC results, so the jury is still out on their real potential. Further scrutiny has also thrown cold water on many of the transdifferentiation claims for other types of adult stem cells.

Even in tissues that share a lineage, transplanted stem cells do not always work enthusiastically. In particular, attempts to make stem cells taken from blood or bone marrow generate new tissue in the heart have produced conflicting results.

In clinical trials involving patients whose hearts were scarred by heart attacks, modest tissue regeneration has sometimes been observed. This improvement can occur even when the studies find no evidence that the stem cells contributed new heart cells to the healing organ. The key to this seeming paradox may be that stem cells can secrete helpful growth-signalling chemicals and contribute to the formation of new blood vessels. In other words, the transplanted bone marrow stem cells may not be producing new heart cells themselves, but they could be laying essential groundwork for the heart's own as yet undiscovered progenitor cells to do so.

Opponents of further human testing have argued that performing these transplants before the regenerative mechanisms at work are fully understood puts patients unnecessarily at risk for tumourlike growths or abnormal heartbeats. Given the lack of effective alternatives for people with failing hearts, however, the trials are very likely to continue, making heart repair potentially the first widespread therapeutic application of adult stem cell therapy beyond traditional bone marrow transplants.

Treatments for less life-threatening conditions may not be far behind. An ongoing clinical trial is already testing the safety of breast reconstruction material created from the stem cells found in fat. In the past two years, both skin and hair stem cells have also been discovered, each of which might be marshalled for cosmetic work. Dental researchers hope to make stem cells discovered in and around teeth regenerate enamel or crowns, although growing an entirely new tooth from scratch might be more than adult stem cells could muster anytime soon.

So far the cells seem to do best when applied within their own lineage to produce small amounts of new tissue or to boost natural regeneration. Last December, for example, German doctors reported having repaired a large gap in a young girl's skull using a combination of bone graft and stem cells derived from her own fatty tissue.

Injections of fat-derived stem cells are already gaining popularity as a means to speed healing of bone and cartilage injuries in horses. For certain uses in humans, too, these cells could be easier to harvest than mesenchymal stem cells from bone marrow. Researchers are finding, however, that like all other adult stem cells studied to date, this type shows a definite decline in vigour as their owners age. Late in life when repairs are most likely to be needed, one's own stem cells might therefore not be the best bet. Where, then, might patients turn?

One potential source of fresh therapeutic stem cells is the donated tissue of miscarried and aborted foetuses. These stem cells are classified as "adult" because they are found in differentiated tissues. Their extreme youth, however, gives scientists hope that when transplanted they will adapt easily to new surroundings and energetically produce new cells.

A major test for both foetal stem cells and the prospects of cell-based brain therapies in general could come in the next year if California-based StemCells, Inc., receives US government approval for its proposed clinical trial. The company, co-founded by the Salk Institute's Fred Gage, who first discovered neural stem cells, plans to transplant foetal neural stem cells into the brains of children with Batten disease. That lethal illness arises from the failure of brain cells to produce an enzyme that clears away cellular wastes. If the stem cells manufacture healthy new brain cells that produce the missing enzyme, the treatment could alleviate the disease, with exciting implications for other related brain disorders.

The Batten trial would be Western scientists' first transplant of neural stem cells into the human brain, an environment that some fear could be difficult for stem cell therapy. Unlike skin, liver and other tissues that naturally repair themselves after an injury, the brain, spinal cord and other nervous tissues do not, and no one is quite sure why. The very existence of adult neural stem cells suggests that they should be able to replace damaged neural tissue. Their failure to do so has prompted speculation that something inhibits them.

Researchers at the Schepens Eye Research Institute in Boston, Mass., reported a breakthrough on this problem earlier this year. Just by manipulating genes responsible for sending "blocking" signals to stem cells, they were able to regrow damaged optic nerves in mice. The experiment highlights a new and promising approach to stem cell therapy. The idea is to learn the language of signals that normally direct stem cells' behaviour well enough to be able to recruit a patient's own stem cells to make repairs on demand.

Studying the cues that stem cells send and receive in their natural environment is also improving scientists' basic understanding of what gives a stem cell its potential. If the secret to "stemness" were as simple as having particular genes active at specific times, then any cell of the body might conceivably be turned into a stem cell as needed.

Ongoing investigations of both adult and embryonic stem cells will likely reveal whether such a feat is feasible. The adult versions so far appear to lack the versatility of the embryonic kind, and even within their own tissue families they show diminishing vigour. Still, certain types of adult stem cells have already proved themselves extremely useful for modest regeneration and repairs. The diverse research currently focused on these cells worldwide promises to unlock further the power of the body's own repair system.