The last session I attended at the meeting we the Tuesday morning stem cell symposium. Now that stem cells have been coopted as a political football, everyone has their own idea about the nature of the stem cell debate, and their own expectations about what activists in the field would be talking about. Most people would be wrong. There was absolutely no discussion about adult vs. embryonic stem cells: that's been long settled in the scientific community, and there is near-universal agreement that we ought to be researching both. No, the question here wasn't about whether we should do stem cell research, but what we have learned from the current work, and where to go next.

John Gurdon was first, on "nuclear reprogramming by Xenopus oocytes", although his talk covered more ground than was implied by the title. (I've got to say that John Gurdon is just a wonderful fellow to listen to: he's got that rich British accent, and he's slender and dapper and topped with those 'long flowing locks'...it's like getting a lecture from an elder David Bowie.) Reprogramming is an important issue. It has become increasingly clear that the egg nucleus isn't just a blank slate, a pile of DNA stripped of the restrictive modifications of the adult differentiated cell. It has it's own special pattern of modifications.

The functioning genome is more than just its sequence of DNA. The DNA is looped and folded in specific ways by the arrangement of proteins that ensheath it, and carries a suite of specific transcription factors that enable and disable transcription of different genes. In addition, one property that Gurdon discussed is methylation: methyl groups are covalently attached to stretches of the DNA backbone that silence those regions. The pattern of methylation is different in different cell types, including the oocyte, and there are also sex-specific methylations.

The reason we need to continue to study embryonic stem cells is that they already possess the appropriate pattern of extra-genetic modifications for a totipotent cell, and we currently have no idea how to impose that pattern on a nucleus. We don't even know what that pattern is. I think that someday we'll have the information or the recipe to switch the naked genome to a totipotent state, but we'll only get that by working with embryonic cells that already have that desired state; I can imagine a future in which we don't need embryonic stem cells at all, but the current political situation is, ironically, blocking progress towards that condition.

Gurdon gave a specific example. In frogs, we know that the egg cytoplasm has some reprogramming activity. If you extract the nucleus from a fully differentiated adult cell type, such as from an intestinal cell, and inject it into an enucleated egg, you will sometimes get a seemingly normal adult frog. We've been doing this for almost 50 years, but the efficiency hasn't improved much: it works about 15% of the time. One trick is to do serial transfers. Transplant the intestinal nucleus to an egg, let it incubate for a while, transfer it to another, then another, and the extended exposure to the egg cytoplasm boosts efficiency to 22%. Furthermore, one common result is that the developing embryo is abnormal, but may have patches of tissue that look healthy; extracting a nucleus from the healthy bits of your ugly embryo and resetting it by transplanting to another oocyte brings you up to a 30% success rate.

What is the mechanism of nuclear reprogramming? It's obvious that the egg is doing something to the DNA to prepare it for development. One system Gurdon is working with is a method of inducing the reprogramming in a large number of nuclei, or even better, plasmids, by injecting numerous quantities into the germinal vesicle of the frog egg. He can squirt 100 nuclei into the egg, which is not going to produce something that can develop to adulthood, but all he wants to see is how the egg biochemically modifies the injected DNA in the first few steps. He is assaying one particular gene, OCT4, a mammalian transcription factor that is not expressed in somatic cells, but is a reliable marker for stem cells. The promoter region of OCT4 is heavily methylated in cells of the thymus, which blocks its expression there. Upon transfer to the oocyte, those sites are demethylated by some activity there, and OCT4 is expressed. He can see exactly the same thing by putting a minimal promoter for OCT4 on a plasmid, and that also gets demethylated. One advantage there is that he can tweak the DNA and identify the bits that signal the occyte to demethylate, and he has identified several sites in the DNA that must be present for demethylation to occur.

The bottom line is that he has identified a strong, selective demethylating activity in frog oocytes, and that selective promoter demethylation is a necessary step in reprogramming somatic cell nuclei.

Another speaker, R. Jaenisch, hit on the same theme of nuclear cloning, stem cells, and reprogramming of the genome. The central problem of cloning of any kind, exemplified by Dolly the sheep, is that embryonic genes must be activated and somatic genes must be deactivated, and the current procedures all involve the hope that something in the oocyte will do that job for us.

Jaenisch is doing a kind of hopeless cloning—he's using cells that he knows aren't going to result in a healthy adult, or that are extremely difficult to reprogram, because, like Gurdon, he is interested in the process, not just the end result. For instance, he is cloning mice from lymphoid cells. Cells of the immune system rearrange their DNA during differentiation, stripping out chunks of immunologically significant sequence that they aren't going to use, which means a successful clone would be immuno-compromised...but that rearrangement is an unambiguous marker that one is working with what was a differentiated cell, and that one isn't working with a clone fortuitously derived from a somatic stem cell. He has also been working with nuclei from neurons, which are post-mitotic and about as differentiated as you can get, if one could put differentiation on a scale.

Most of his clones die after implantation. There is a decline in potency with increasing age of the donor—and the state of differentiation matters. The frequency of survival to adulthood in clones derived from B or T cells, or neurons, is less than 0.001%, which is why working with mice is an advantage for this kind of research.

Jaenisch has found that there are many genes like OCT4, and all of them need to be reactivated for success, but OCT4 is an instructive place to start. He has made a genetically modified mouse carrying an inducible OCT4—that is, a copy of the gene that he can switch on at will, at any time in development, including adulthood. Remember, OCT4 is always off in somatic tissues; turning it on in an adult mouse does horrible things. It triggers rapid proliferation of epithelial cells while suppressing fibroblast cells, producing multiple, massive, invasive cancers and death in less than a week. (One side effect of stem cell research that isn't mentioned enough is that it's good for more than just cloning. This reprogramming of cells into a proliferative, multipotent state is also a cellular property at the heart of cancer; knowing how to reprogram cell states is a step on the road to a cure.) The next step in his work is to use his inducible OCT4 cells to see if they can be used to improve the efficiency of cloning.

He's optimistic. There are no barriers in principle to therapeutic cloning, there are only technical problems.