Reprogramming adult cells into embryonic stem cells

Blogging on Peer-Reviewed Research

As promised, I'm going through the three papers from last week about the re-programming of adult cells into an embryonic-like phenotype. Since it is three papers I'll go through first what's common to all three, and then what each group did special.

First of all, let's summarize the method one more time.


All of these papers are based on the "rational identification" of 4 critical transcription factors by Yamanaka in 2006. What they did was take 20 proteins that drive the expression of other genes that were known to be in embryonic stem cells, and added them to adult cells to see which were critical for the embryonic stem cell phenotype. They screened which cells had been reprogrammed by using a control region upstream of the Fbx15 gene to drive a drug-resistance marker in the cells. Only cells which turned on the Fbx15 gene - a gene expressed in embryonic stem cells - would survive subsequent selection with the drug. Since many cells were infected and only a few transformed into the embryonic-like cells, the authors needed a way to rapidly determine which ones had done this.

They then subtracted factors one-by-one until they identified just 4 factors which seemed critical for this transformation of adult cells into ES cells. These four were Oct4, Sox2, c-Myc, and Klf4, and it was a surprise because only four transcription factors were needed to reprogram the cells back into being pluripotent, that is, capable of making many lineages of cells.

The main problems were that: (1) gene-array analysis (comparing the levels of expression of tens of thousands of genes) showed that while they were similar to embryonic stem cells, there were significant differences (2) there were problems in reprogramming the chromatin - the scaffolding molecule that binds DNA that also is critical to epigenetic regulation of gene expression - and other post-translational modifications of DNA suggesting the cells were somewhere in-between adult fibroblasts and embryonic stem cells, (3) the method used to deliver the genes - a retrovirus - creates problems for translation into therapy as it causes these genes to insert randomly into the genome, (4) the promoters used to express the genes are always on - or constitutive - and these genes can have negative consequences (cancer) when expressed inappropriately.

These new papers have replicated and expanded upon Yamanaka's original work.

Maherali et. al. Cell Stem Cell [1]
The first paper I'll discuss - from Kathrin Plath and Konrad Hochedlinger (from UCLA and Harvard respectively) was published simultaneously in Cell Stem Cell in the journal's inaugural issue. The first major difference is these researchers, instead of using the Fbx15 gene to identify the cells, used a gene called Nanog - which is also required for the ES-cell phenotype - to drive the selection gene and the green fluorescent protein (GFP) to make it easier to visualize which cells successfully converted.

The cells from this group additionally were able to grow without a feeder-layer - a good sign for those who don't want their custom ES cells grown in the presence of xenobiotics - and also were capable of reprogramming the ES-like phenotype onto other somatic cells - another good control to show they had the correct phenotype. They also did more testing to show that the transformation isn't immediate. After adding the factors you had to wait a good while - about a week, for the changes to occur. This is a very long time for a cell culture experiment if you ask me, and just goes to show that the process of re-arranging the chromatin in the cells must be a gradual process.

These authors also did the proof-of-principle experiment of controlling one of the genes - Oct4 - with a tetracycline-inducible promoter which would in the future allow researchers to turn these genes off when they are no longer needed (potentially solving problem 4). These are promoters which are only active in the presence of a drug like tetracycline or doxycycline. They also can be made to work with other chemicals like tamoxifen.

Also, for those who want to know how to test pluripotency - there are a few ways. You can remove the anti-differentiation factor, LIF, and then just watch them differentiate in the dish. Then there are in vivo methods. The easiest is to inject the cells into an immune-deficient or "nude" mouse and wait for the cells to make a characteristic tumor - the teratoma - a benign tumor containing cells from the 3 primordial germ layers. The second, which is more rigorous, is to inject the cells into a mouse blastocyst and then implant the blastocyst into a female. The resulting mice will be "chimeras" or mixtures of the two cell types in the blastocyst. The resulting mice look like these ones from their last figure:
The chimera is the top one, and the lighter-colored hair is from the injected ES-cells which have genes for a brown or agouti coat. There's a third method we'll get into with the last paper.

Finally, for some reason not fully clear to me, their methods were a slight improvement on the original as their cells had a DNA methylation pattern closer to that of real embryonic stem cells - as well as proper demethylation of both X-chromosomes (ES cells should not have an inactivated X chromosome - if the cells are female both X chromosomes should be active until differentiation) - possibly bypassing problem 2. This might have been because the Nanog locus was better for selection than Fbx15 - the following paper's results support this - or it might have just been something simple like how they waited 7 days rather than 3 for the cells to transform. Time will tell.

Okita et al. Nature [2]
This second paper, is from Yamanaka's group at Kyoto university - the one that got this method up-and-going, uses similar methods to his original paper - but this time they demonstrate "germ-line transmission". That is, they are able to use their technique to make embryonic stem cells from adult cells, they then make mice out of them using the chimera technique described above, and in a certain portion of these mice the cells contributed to the germ-line - sperm or egg cells - allowing them to create viable offspring from breeding. It should also be noted that this paper is an excellent example of just beautiful data-presentation. Everything is very clear and the data is very crisp and pretty - it's more important than one would think.

Now, the first thing take away from this paper are that by switching to selection for Nanog from Fbx15 they were able to make their cells fully pluripotent and capable of making real live mice - which they failed to do in their first paper. However, they also waited 7 days after retroviral induction - so it's hard to be sure it still isn't a timing issue.

Here's a picture of their mice after 2 generations - and an example of great presentation of data:
The offspring of the chimeras are ligher-colored - the light coat shows the trait has been inherited from the chimeras and passed to subsequent generations. This time the coat is uniform, because rather than being a mixture of engineered and wild-type cells, these animals were generated from a cross of the chimeric animals with normal mice - showing the cells can result in fertile offspring.

The second main thing to take from this paper is that the mice generated from this process developed tumors about 20% of the time, and the authors traced the origin to the myc oncogene (one of the critical 4 factors is a notorious oncogene). This is a major limitation of therapeutic potential. For whatever reason the other factors are silenced during development despite being transgenes - while the myc gene was refractory to silencing and this resulted in a variety of cancers in the animals. It should also be noted that they showed the cells generated with Nanog rather than Fbx15 selection were capable of being passaged longer in culture in the undifferentiated state - another excellent outcome.

Wernig et al., Nature [3]
The second paper from Nature is from Rudolph Jaenisch's group at Whitehead/MIT. The main thing to take away from this paper - which also selected with Nanog and created germline chimeras much the same as the first two - is that they did an even more rigorous test of the cells pluripotency, and actually demonstrated what should be called totipotency - the ability to make a complete viable embryo and every type of cell in it.

The previous chimeric assays relied on injected the reprogrammed ES cells into a mouse blastocyst - which also contains wild-type ES cells in the inner cell mass. So the resulting embryos are a mixture of wild-type and engineered cells. It's possible that the wild-type cells can compensate for defects in the engineered cells, and it would difficult to detect. So Jaenisch's group went to the next level and used a technique called tetraploid rescue - it's complicated - that allows them to generate embryos that are 100% made from the engineered cells. This shows, with a high degree of certainty, that the re-programmed cells can be totipotent.

Hopefully this hasn't been too technical, if there is something that can be clarified, please comment away. The main things you should take away though are:

  1. In mice scientists have been able to reprogram adult/differentiated cells back into embryonic stem cell-like cells.
  2. These cells can make any cell type - they are totipotent.
  3. These cells can be passaged (probably) indefinitely
  4. There are still major problems - the use of retroviruses (random insertion into the genome), the potential for cancer from failure of one of the oncogenes to be silenced, and the need to adapt this to human cells.

Still this is a major success and improvement from last year's results using Fbx15 which seemed to create a adult/embryonic hybrid that wasn't fully totipotent. With time the system should be able to be adapted to human cells, and if we're fortunate we'll be able to eliminate the other technical difficulties that the proof-of-principle experiments were subject to.

1. Nimet Maherali, Rupa Sridharan, Wei Xie, Jochen Utikal, Sarah Eminli, Katrin Arnold, Matthias Stadtfeld, Robin Yachechko, Jason Tchieu, Rudolf Jaenisch, Kathrin Plath, and Konrad Hochedlinger. Directly Reprogrammed Fibroblasts Show Global Epigenetic Remodeling and Widespread Tissue Contribution. Cell Stem Cell, Vol 1, 55-70, 07 June 2007

2. Okita,Keisuke; Ichisaka,Tomoko; Yamanaka,Shinya. Generation of germline-competent induced pluripotent stem cells Nature. Published online 6 June 2007

3. Wernig,Marius; Meissner,Alexander; ; Foreman,Ruth; Brambrink,Tobias; Ku,Manching; Hochedlinger,Konrad; Bernstein,Bradley E.; Jaenisch,Rudolf In vitro reprogramming of fibroblasts into a pluripotent ES-cell-like state Nature. Published online 6 June 2007.

More like this

Today three papers came out, two in Nature, and one in the inaugural edition of Cell Stem Cell, that basically confirm the results from last year's landmark manuscript by Kazutoshi Takahashi and Shinya Yamanaka (for details on this paper, see this post). Just to remind you, in that original…
I hope this time I'm finally right about this. I've been hopeful that some strategy of developing stem cells would allow us to bypass the absurd ethical restrictions from those who think one type of destruction of an embryo is worse than another. Particularly promising were spermatogonial stem…
Since the discovery of IPS Cells, the stem cell field has exploded. Here's a few links on the latest developements. First, two cool papers came out recently. In the first from the Jaenisch lab, mouse IPS Cells were differentiated into erythrocytes and used to cure sickle celled anemia in a mouse…
You guys might have noticed I've been quiet lately, that's because I've scheduled a thesis defense and am under deadlines. However, I couldn't let these two (1) papers(2) on reprogramming of human adult cells into stem cells slip by without some comment (NYT piece here) These reports are a follow-…

Excellent summary! Now I can consider myself well informed without reading the papers... Specially since sometimes you can't read the papers without paying at least 30 USD for each (that's the case with Nature and the majority of the most important magazines).

There has been a lot of hype about "obtaining embryonic stem cells without harming embryos", which is of course nonsense. They are an equivalent, but the most obvious thing is that they're not embryonic; more care should be exercised with the terms used. When I read the approach to transform adult cells into an ES cell phenotype by using stably transfected/transduced/integrated transgenes, it became quite obvious to me that there would be problems for a therapeutic use. It's not unusual for transgenes to survive into the adult organism, or transgenic mice and other organisms would be a rarity. And having extraneous regulatory genes being expresed inside one's organism, in places they shouldn't be expressed, isn't something that would make me feel very safe about. What is great is that these works are a proof of concept. Perhaps the next step could be the use of a Cre-Lox type of system to definitively elliminate the transgenes once the desired phenotype has been achieved. Using an inducible promoter is, in my opinion, not a solution, as mutations or chromosomal rearrangements that substitute a constitutive overexpression for an inducible one would be selected as advantageous for the cells that carry them, leading eventually to malignancy.

When I read the approach to transform adult cells into an ES cell phenotype by using stably transfected/transduced/integrated transgenes, it became quite obvious to me that there would be problems for a therapeutic use. It's not unusual for transgenes to survive into the adult organism, or transgenic mice and other organisms would be a rarity.

Absolutely this is true. What was interesting was that they authors observed that 3/4 genes were regularly silenced during differentiation despite being driven by the transgenic promoter - yet the myc would penetrate about 20% of the time.

A very curious result. It sounds like epigenetic reprogramming during differentiation is targeting these transgenic loci fairly efficiently, but myc for some reason isn't being silenced by the reprogramming. If anything I'm surprised that the genes are getting turned off by the cells at all since they are being driven transgenically by constitutive promoters. You would expect some to be silent just by luck, usually about 1/3 of random insertions into a trangenic line will be PCR positive but inserted into a silent loci. What is strange is that this was happening to 3 of the genes frequently, but not myc. It's a mystery.

Either way, switching to an inducible promoter should hopefully fix that.

Being as how I have no biological training whatever, the description was Greek to me. Perhaps a question will get to the bottom line.

Is there a reasonable possibility that these procedures will lead to the development of embryonic stem cells and negate the current necessity of destroying an embryo to get them? Does it follow, therefore, that experimentation on embryonic stem cells is no longer required?

Well SLC.

1. They are not embryonic. They act like it, but they are not. But, in theory, yes. The problem being that its not certain what the long term result *could* be to having such cells in a body, especially if they opted to use the technique to "grow" entire organs for transplant, which is one of the major tricks they have been considering to replace the current iffy, and rejection prone, method of transplant.

2. Umm. Hard to say. First off, there might actually be some more complex, but *better* system hidden in the cells that doesn't have the 20% cancer risk. That would require a more complex analysis though, which isn't possible unless you have a cell to *compare* to. I.e., if you want to know why an embryonic cell does X instead of Y, you have to compare a normal cell to an embryonic cell. This is something the, "eliminate research on embryonic cells!", crowd is severely self deluded about. Second, it could be easily argued that we couldn't have gotten here without it, nor is it certain we could stop at this point, without a more complete understanding of how the differentiation works. Again, controlling the "mechanism" that turns something off/on is usually better than simply inserting stuff that *forces* them on. We have identified the switches, but we don't necessarily understand the circuit board under the panel, if you get my meaning. Suggesting that the answer to this is "yes" is imho premature, by a long shot.

What Kagehi said. And also this is still in mouse. This has not worked in humans yet, and that may be a really significant barrier (or not - impossible to be sure). But keep in mind it took 17 years after the discovery of embryonic stem cells in mice to generate them from humans. Although that was more of a raw materials problem, instant translation is not always a guarantee.

Most the other details I feel are mostly technical. A little bit of creative molecular biology should allow scientists to turn the genes off after they've done their job.

Regarding Kagehi's speculation about a better, although more complicated, approach that obviates the cancer problem caused by the transgenes, I must insist that a Cre-Lox site specific deletion system (if you're not familiar with it, here is a very good, brief, description of how it works) would be great to at least delete the c-Myc transgene.

I haven't been able to find a decent map of the transgene constructions used in the different works, but from what little I found, it seems the retroviral LTRs are used as the transgene's promoters. Anyway, such strong constitutive promoters don't usually reflect physiological levels of expression of the genes linked to them, and cells usually tend to silence that kind of transgene (it has happened to me). That explains the disappearence of the expression of 3 of the transgenes. With regard to c-Myc (and I don't know if Okita et al. used the same version as Maherali et al., which is even more active than the wild type), it's expression could be expected to give a survival advantage to the cells, so it wouldn't be so likely to be suppressed by epigenetic changes.

Nevertheless, if his work eventually leads to significant therapeutical advances, Yamanaka might well be bound for the Nobel Prize.

An impressive explanation.

I think the implications for denialism are interesting, because it's potentially allowing 'wild' (love the terminology!) embryos to be given sacrosanct status while undercutting the vitalist arguments that are used to justify that status.

Crazy atheist scientist: Give me your babies and I will cure your diseases!

Nervous conservative public: Uh-oh. No way can you have our babies.

Crazy atheist scientist: OK, fair enough, I will make the babies myself from your dandruff in my Faustian laboratory. Mwahahahaha...

Nervous conservative public: Righhhht... we're supposed to think that's a good thing. Are we?

Re: Will this translate into a therapy.

We could use the cre-lox system as suggested by ribozyme, or we could use an inducible promoter. But you still have the problem that the foreign DNA will be introduced randomly within the genome. This has caused many problems in the past for other "gene therapy" procedures.

In the past inserting genes in trans (i.e. gene therapy) has proved to be inefficient untrustworthy, and induce many complications. From this history, and from the fact that the procedure involves the addition of two oncognes (c-Myc and klf4), one might reasonably guess that this "reprogramming" of cells might have many unintended consequences (i.e. teratoma formation). But there is a difference. This procedure can be done in cells outside of the animal. We have many tools to get DNA, mRNA and proteins into cells when they are grown in isolation. And those technologies are rapidly expanding. Once we figure out how long the four transgenes have to be active, I'm 99% sure that we can figure out a way to reprogram cells without having c-myc remaining on.

Very nice summary.

Don't forget about another paper that was lost in the mix last week. Kevin Eggan's group at Harvard discovered a new way to reprogram cells. Its a report in Nature.

If you excuse my shameless self-link, I've put a little summary and some more links at my page.

Re-programming adult Stem cells to function as embryonic stems could reduce most of the controversa on embryonic stem cell research.

If some researcher could devise a procedure for extracting a few cells from the embryos without damaging the embryos, then most of the controversa and ethics issues could be greatly reduced.

Is anyone in the stem cell research community devising methods of extracting a few cells from an embryo, without damaging the embryo and using those cells for all kind of treatment? I am reading this novel titled " Ageless- A Stem Cell Mystery" . This is a great novel because it stimulated my interest in this subject. There is a website devoted to this book and the address for the site is:

You can find a lot of information about this book on this website