Reprogrammed stem cells are loaded with errors

View Images

This is an updated version of another post, edited to meld together a brand-new discovery with an intimately related one from last year.

Imagine trying to rewind the clock and start your life anew, perhaps by moving to a new country or starting a new career. You would still be constrained by your past experiences and your existing biases, skills and knowledge. History is difficult to shake off, and lost potential is not easily regained. This is a lesson that applies not just to our life choices, but to stem cell research too.

Over the last four years, scientists have made great advances in reprogramming specialised adult cells into stem-like ones. By turning back the clock, they can once again imbue the potential to produce any of the various cells in the human body. It’s the equivalent of erasing a person’s past and having them start life again.

But two groups of scientists – one led by Kitai Kim, and the other by Ryan Lister and Mattia Pelizzola – have found a big catch. They showed that these reprogrammed cells, formally known as “induced pluripotent stem cells” or iPSCs, still carry a memory of their past specialities. A blood cell, for example, can be reverted back into a stem cell, but it keeps a record of its history that constrains its future. It would be easier to turn this “stem cell” back into a blood cell than, say, a brain cell.

The history of iPSCs is written in molecular marks that annotate its DNA. These ‘epigenetic’ changes can alter the way a gene behaves even though its underlying DNA sequence is still the same. They are like Post-It notes – you can stick them to a book to point out parts to read or ignore, without editing the underlying text. Epigenetic marks separate different types of cells from one another, influencing which genes are switched on and which are switched off. And according to Kim, they’re not easy to remove, even when the cell has apparently been reprogrammed into a stem-like state.

But reprogramming adult cells is just one of two ways of making stem cells tailored to a person’s genetic make-up. The other is known as nuclear transfer. It involves transplanting a nucleus (and the DNA inside it) from one person’s cell into an empty egg. The egg becomes an embryo, which yields stem cells containing the donor’s genome. Kim has found that these cells (known as nuclear transfer embryonic stem cells or ntESCs) are ‘stemmier’, for lack of a better word. They’re much more like genuine embryonic stem cells than the reprogrammed iPSCs.

Nuclear transfer certainly steers into trickier ethical territory than iPSCs, since the process of harvesting stem cells destroys embryos. And it is still trailing behind technically. So far, it has only worked in monkeys and other non-human mammals, and it has been mired in scientific scandal.

Meanwhile, work on iPSCs has raced ahead (see my interactive timeline for an overview). The starting pistol was fired in 2006, when Shinya Yamanaka first showed that it was possible to create these cells in mice. The race intensified in 2007, when three research groups independently managed to do the same for human cells. The cells have been used to cure at least two genetic diseases in mice. They’ve even been used to create live mice, passing the ultimate test of their stem-like status.  Various groups have made the technique more efficient, sped it up, found ways of sorting out the most promising cells, and changed the details so that it doesn’t use viruses (or uses only viruses).

But all along, scientists knew that there are subtle differences between iPSCs and genuine embryonic stem cells and, indeed, between iPSCs produced from different tissues. For a start, some types of cell are easier to reprogram than others. Skin, stomach or liver cells, for example, are easier to convert than cells from connective tissues. And the older or more specialised the cells are, the harder the task becomes.

Once the cells are converted, there are further issues. Kim could produce blood cells more easily from iPSCs that themselves came from blood cells than those that came from connective tissue or brain cells. However, if they wanted to make bone cells, iPSCs from connective tissue were the better choice.

Kim thinks that this is because the common reprogramming techniques fail to strip away a cell’s epigenetic markers. He focused on one such marker – the presence of methyl groups on DNA, which typically serve to switch off genes. They’re like Post-it notes that say “Ignore this”. Kim found that iPSCs have very different methylation patterns depending on the cells they came from. Those that come from brain or connective cells have methyl groups at genes that are necessary for making blood cells, and vice versa. The iPSCs even have distinctive methyl marks if they come from slightly different lineages of blood cells.

Now, Ryan Lister and Mattia Pelizzola from The Salk Institute have found the same reprogramming errors in human iPSCs, and to a much greater extent than even Kim had suspected. “You might compare all past studies to looking through a key hole at what is inside a room. [We opened] up the door to view the entire landscape,” says Joseph Ecker, who led the study.

Ecker’s team looked for methyl marks across the entire genomes of five lines of iPSCs, each produced by different laboratories around the world. They also compared these to the methyl marks of adult cells and genuine embryonic stem cells. For each cell line, Lister and Pelizzola looked for methyl marks at 1.17 billion sites across the genome, around 250 times more detailed than the search that Kim did.

At first, the iPSCs seemed to have a spread of methyl marks that looked superficially similar to those of embryonic cells. But when Lister and Pelizzola looked more closely, the cracks started to appear in this tidy picture. The duo found plenty of hotspots around the iPSC genomes that were unusually ridden with methyl marks. None of these marks existed in true embryonic stem cells, and some sat in places that could switch off important genes.

Many of these errors were common to all iPSC lines, and some were unique to individual ones. Around half of them were remnants from the iPSCs past lives, but the other half were fresh mistakes, found neither in the adult cells nor the embryonic ones. In either case, the iPSCs could pass these marks to their own daughters. Any cell that’s born from reprogrammed ones will inherit the same legacy of errors.

When Kim published his discovery last year, another group led by Jose Polo found the same epigenetic problem along with a seemingly simple solution. When scientists grow cells in culture, they frequently split among fresh containers so they don’t run out of room. This is called “passaging” and Polo found that he could solve the epigenetic problem by doing it constantly for a long time. But Lister and Pelizzola studied several iPSC lines that had been passaged repeatedly, and they still carried reprogramming errors.

Another solution might be to abandon iPSCs altogether. Last year, Kim showed that ntESCs (those produced by nuclear transfer) were far more similar to genuine embryonic stem cells than any of the iPSCs. Their methyl patterns are a closer match and they’re easier to convert into any type of adult cell. This certainly makes sense – when the nucleus is transferred into an empty shell, it its DNA is rapidly and actively stripped of its methyl groups. Its history is erased with far greater efficiency.

The nuclear transfer method might have the edge here, but that’s unlikely to stop scientists from trying to improve the reprogramming technique. There are two big challenges ahead. First: work out why the iPSCs haven’t been reprogrammed in quite the right way. Second: find ways of fixing that to give cells that are a closer match to true embryonic stem cells. This might not be easy.

Lister and Pelizzola found large swathes of the genome – a few million DNA letters across – that are rife with methyl marks. These areas, which clustered near the centre and tips of each chromosome, proved to be especially resistant to reprogramming. The duo thinks the DNA at these places might be packaged and folded in such a way that makes methyl marks harder to remove. Now, they’re testing chemicals that target or open up these areas, in the hope of producing a ‘stemmier’ breed of iPSCs.

References: Lister, Pelizzola, Kida, Hawkins, Nery, Hon, Antosiewicz-Bourget, O’Malley, Castanon, Klugman, Downes, Yu, Stewart, Ren, Thomson, Evans & Ecker. 2011. Hotspots of aberrant epigenomic reprogramming in human induced pluripotent stem cells. Nature.

Kim, Doi, Wen, Ng, Zhao, Cahan, Kim, Aryee, Ji, Ehrlich, Yabuuchi, Takeuchi, Cunniff, Hongguang, Mckinney-Freeman, Naveiras, Yoon, Irizarry, Jung, Seita, Hanna, Murakami, Jaenisch, Weissleder, Orkin, Weissman, Feinberg & Daley. Epigenetic memory in induced pluripotent stem cells. Nature

More on stem cells: