Science and Reason: Induced pluripotent stem cells with one transcription factor

Just under three years ago, in October 2006, some important stem cell research was announced by a Japanese scientific team led by Shinya Yamanaka. The team showed how ordinary mouse skin cells could be transformed into cells that turned out to be pluripotent, just like embryonic stem cells (ESCs). The new cells were called induced pluripotent stem cells (iPSCs). Although "ground-breaking" is an over-used term, this research genuinely deserved the description.

Aside from the fact that it could be done at all, the surprising thing was that the transformation could be effected by adding transcribable genes for just four transcription factors to the skin cell DNA. Those genes were Oct4, Sox2, c-Myc, and Klf4. And now very recent research shows that, under the right conditions, just the addition of Oct4 alone can accomplish the same feat.

We discussed some of the early research here, with additional reports here, here, and here.

In the three years since the original announcement, research has extended and improved the process in a number of ways. The ultimate goal is to be able to produce pluripotent human stem cells that are in all important respects equivalent to embryonic stem cells, by a process that meets several important criteria:

Cells to be reprogrammed into a pluripotent state should be readily obtainable from human subjects (unlike embryonic cells or rare types of adult stem cells).
No permanent changes to cellular DNA should be made, only changes to gene expression.
The process should be relatively quick and efficient, so that reasonable number of pluripotent cells can be obtained for routine therapeutic or experimental uses.

Reprogramming of other cell types into pluripotent cells is important not just as a technical feat to prove it can be done. There are two other important objectives. The first is to develop human cell lines that model many types of pathology (cancer, Parkinson's disease, or whatever) to facilitate research into therapeutics for these diseases. The best way to develop such lines is first to obtain pluripotent cells with the appropriate pathology, derived from human subjects with the disease, which can't generally be done from embryonic sources. From there, several techniques can be used to produce appropriate cell cultures with the desired model pathology.

The second objective is longer-range but even more important: to manufacture cells, for patients with certain diseases, that can be used as therapeutic replacements for the patient's own malfunctioning tissue. This would be accomplished by obtaining pluripotent cells derived from the patient, correcting genetic problems in those cells, and then inducing the cells to differentiate into the required tissue type. Diseases that should be treatable in this way include Parkinson's disease, Type 1 diabetes, and heart disease. Starting with cells from the actual patient eliminates the problem of tissue incompatibility.

The criteria listed above that are imposed on the process are important for meeting both of these objectives.

In the three years since the original work was announced, dozens of research groups have set about testing improvements to the original procedures in order to progress towards the ultimate objectives. The improvements that have been made include:

adapting the procedures to work in species other than mice – including pigs and fruit flies, as well as humans
reducing the number of transcription factors that need to be introduced, or finding other suitable transcription factors
finding other cell types besides skin cells to start with, generally various types of non-pluripotent stem cells – which makes other improvements in the process easier to accomplish
changing the way that the transcription factors are introduced into the target cells, in order to avoid alteration of the original DNA (since such alterations may introduce risks of cancer or other cellular malfunction)
finding other proteins or small molecule compounds that can be added to enhance the efficiency and speed of the process

Quite a few important improvements have been announced within the past several months, along with other related news. The most interesting related news is a demonstration that iPSCs really are not only equivalent to ESCs in terms of gene expression, but are in fact equally pluripotent. This latter fact was convincingly demonstrated by cloning several generations of live, healthy mice from iPSCs. (We'll discuss that in a separate article, but here's an overview.)

What I want to discuss here is how the list of transcription factors (or their genes) that need to be added to a non-pluripotent cell has been reduced to just one: Oct4. The work was done by a mostly German team led by Hans Schöler of the Max Planck Institute for Molecular Biomedicine.

So how was this accomplished? Well, the trick is, you have to start with the right kind of cells. In this case the researchers used human fetal neural stem cells (HFNSCs). While such cells aren't pluripotent, they are "multipotent", which means they can normally differentiate into various other cell types.

Back in February the researchers in this study reported that reprogramming with just Oct4 could be done in mouse neural stem cells (see here, here, or here). But would this also work with human cells?

Yes. The latest report shows that HFNSCs can be reprogrammed to a pluripotent state using only Oct4 and Klf4, and (generally) even with Oct4 alone. How is this possible? It is known that mouse neural stem cells already express Sox2, c-Myc, and Klf4. As for the human case, the paper says, cautiously, that "The feasibility to reprogram directly NSCs by OCT4 alone might reflect their higher similarity in transcriptional profiles to ES cells than to other stem cells like haematopoietic stem cells or than to their differentiated counterparts."

And the main indication of this is that the process works: "One-factor human NiPS cells resemble human embryonic stem cells in global gene expression profiles, epigenetic status, as well as pluripotency in vitro and in vivo. These findings demonstrate that the transcription factor OCT4 is sufficient to reprogram human neural stem cells to pluripotency."

What this is saying is that there are several criteria for similarity to embryonic stem cells that the reprogrammed HFNSCs meet. At a molecular level the reprogrammed cells express the same genes and have the same epigenetic markers as ESCs. In addition, they can differentiate into many adult cell types both in vitro and in vivo (in the latter case, by forming teratomas (mixed masses of cell types) when implanted in mice).

There are still several drawbacks to this method for practical purposes, even of research. For one thing, human fetal neural stem cells are not exactly easily obtainable. And in addition, retroviruses were used (as in the original Yamanaka work) to introduce Oct4 into the cells. For therapeutic applications it would be absolutely necessary to use one of the other methods that have been explored and that do not disrupt the existing cell DNA or leave exogenous DNA in derived cells – since either alternative means the derived cells might revert to a more undifferentiated state. On top of all that, the process is still inefficient and slow.

Reprogramming methods that have been explored in other research include the introduction of genetic material in forms other than retroviruses, as well as direct delivery of the transcription factor proteins. The researchers in this study intend to investigate such possibilities, as well as use of other initial cell types: "Future studies will show if direct reprogramming is possible with small molecules or OCT4 recombinant protein alone. ... It will be interesting to extend this study to human NSCs derived from other sources, such as dental pulp, as well as to other stem-cell types."

Kim, J., Greber, B., Araúzo-Bravo, M., Meyer, J., Park, K., Zaehres, H., & Schöler, H. (2009). Direct reprogramming of human neural stem cells by OCT4 Nature DOI: 10.1038/nature08436