Mouse embryo grown from reprogrammed skin cells. Credit: Marius Wernig and Alex Meissner (c) Whitehead Institute for Biomedical Research

Genetically modified mouse skin cells that act like embryonic stem cells (ES cells) could show researchers how to make versatile human stem cells from a simple skin biopsy without having to negotiate the ethical minefield of working with human eggs and early embryos.

Research published this week shows, for the first time, that differentiated mouse cells can be reprogrammed to create any tissue in the body. This work advances a breakthrough published last year, when Shinya Yamanaka and Kazutoshi Takahashi at Kyoto University identified a quartet of genes, Oct3/4, Sox2, c-Myc and Klf4, that caused cultured mouse skin cells to behave remarkably like ES cells1. Cells engineered to express these genes could make teratomas (a type of tumour composed of many different types of differentiated cells) and most tissue types. But these so-called induced pluripotent stem cells (iPS cells) could still not do everything ES cells can do. For instance, when iPS cells were injected into blastocysts that were then put into surrogate mothers, they did not form any of the tissues in those embryos that grew into live-born mice. They did contribute to tissues in embryos that grew to mid-gestation, but even then, the iPS cells lacked one crucial trait: they did not definitively make the germ line, the reproductive cells that produce sperm or eggs. This meant that, despite their name, these iPS cells were not truly pluripotent, or capable of forming every type of cell in the body, and so lacked the full potential of ES cells.

Now, new work by Yamanaka shows that male iPS cells injected into blastocysts and grown into adult mice can contribute to all tissue types, including sperm2. Konrad Hochedlinger of Harvard Medical School and his colleagues did the same with female iPS cells, and showed that iPS cells gave rise to oocytes3. And a group led by Rudolf Jaenisch at the Whitehead Institute in Cambridge, MA produced late-stage mouse embryos derived entirely from iPS cells4.

Click here to read an interview with Shinya Yamanaka.

The trick is a better way to screen for cells that are successfully reprogrammed. As before, scientists use retroviruses to insert the four genes into cultured skin cells (fibroblasts), but this time the cultured cells are different: they have first been modified to signal when one of their own pluripotency genes becomes active. For example, Yamanaka created mice in which the gene for Nanog was fused with the gene for green fluorescent protein. Then they used these mice to make cultures of skin cells. These cells turn green when Nanog is expressed. The details of the screens vary between researchers, but in each case it allows researchers to pick out the right cells.

The research also shows that iPS cells closely resemble ES cells in their genetic activity. Hochedlinger found that engineered female cells first reactivated their silent X chromosome, and later silenced one of the two X chromosomes at random, mimicking what happens early in embryonic development. In fact, gene silencing was reset across the genome to resemble ES cells. Perhaps most important, that activity appears to be maintained by the cells' own genes, not the ones inserted by the retrovirus. All three groups found that the virally introduced transgenes had been silenced in iPS cells. Such results support the notion that differentiated cells can be made to behave like, and perhaps substitute for, ES cells.

This could be the first step on the long road to growing neurons from skin cells taken from a patient with Parkinson's disease. The nerve cells could then be used to test therapies — or even as therapy themselves.

If converting cultured skin cells to iPS cells proves successful in humans, it could be the first step on a very long road to, say, growing nerve cells from skin cells taken from a patient with Parkinson's disease. The nerve cells could then be used to test therapies — or even as therapies themselves.

“If what they've got works in human cells, the simplicity of taking a skin biopsy from human patients opens up a door for a lot of work to be done,” says Fred Gage, who studies neurodegenerative disease at the University of California, San Diego. In many diseases known to have a genetic cause, the cause itself is poorly understood. Researchers can engineer mice to express the suspected genes, but such models frequently fail to predict what happens in humans. Far better would be to have an unlimited source of the patient's own cells to study. Very few cell types isolated directly from a person can be cultured for long, but iPS cells derived from a patient's skin biopsy could, in principle, be cultured indefinitely, and be made to differentiate into whatever cells were required.

But even if iPS cells created from a patient could be made into neurons, those neurons would not be helpful unless they showed some sort of defect that could be measured readily. Neurodegenerative diseases tend to afflict brain tissues that are several decades old, for example, but no one is likely to keep neurons in a dish that long.

Such discussions are, perhaps, premature since mouse iPS cells have not yet been used to create well-characterized populations of differentiated cells in the laboratory. However, researchers don't seem to think that doing so will be a problem. “If they make all the tissues in the mouse, they should be able to make all the tissues in the lab,” says Alan Trounson of Monash University in Clayton, Australia.

The far bigger barrier is moving from mice to humans. Hochedlinger is trying to reproduce his experiments in human cells, starting with the four genes Yamanaka used in the mouse studies. Rapid success is unlikely, he acknowledges: “It remains unclear whether reprogramming is possible in humans or whether the same factors would do the job.” Human ES cells behave quite differently from their murine counterparts. For example, they need a different protein recipe to maintain themselves, and while mouse ES cells can regrow again if separated into individual cells, human ES cells do not.

Even if the technique does work for human cells, there are significant hurdles to using it practically. Retroviruses insert into the genome randomly, so differences between cell lines from individual patients might be due to varying numbers and locations of inserted genes rather than to real differences between patients. From a therapeutic standpoint, the nature of the pluripotency genes that need to be inserted could cause problems. One of these genes (c-Myc) is often activated in cancer, and progeny of mice grown from iPS-containing embryos cells developed carcinomas at an unusually high rate. Even if cMyc could be avoided, the presence of any transgenes in iPS cells could limit their therapeutic use.

Fortunately, the work published this week implies that the virally introduced cells are needed merely to initiate, not sustain, skin cells' transformation into iPS cells. That could mean that permanent genetic modifications might not be necessary, which would make the cells more useful for both research and, potentially, cell therapies.

Bruce Conklin, who studies cardiomyocytes differentiated from both mouse and human ES cells at the University of California, San Francisco, is confident researchers can find alternate ways to induce pluripotency. “Once it can be done with Myc, it is only a matter of time before it can be done with a cocktail of drugs, or other agents.”

The search for such factors is already under way. Jaenisch is collaborating with Peter Schultz at the Scripps Research Institute in La Jolla, California, to screen over a million small molecules to find those that activate pluripotency genes in mouse embryonic fibroblasts. Other approaches expose cells to proteins or introduce genes into cells temporarily. But again, moving from mice to humans can stymie efforts. Last year, Schultz and colleagues identified a small molecule that could help maintain mouse ES cells in culture. It did not, however, translate to human cells.

Despite the promise, most researchers believe the potential of iPS cells for drug screens or therapies is no reason to abandon work on ES cells. “This is a very first step, a very first discovery,” says Hochedlinger. “It's too early to tell if this will replace ESC [embryonic stem cells] or SCNT [somatic cell nuclear transfer]. At this point, the answer is no.” Still, Hochedlinger and other researchers are planning to spend years figuring out exactly what iPS cells are capable of. Researchers working on specific diseases are eager for their success.

“It's not going to be easy,” says Gage, “but it could be great.”