As the resident biologist, I've been asked by friends and family about the significance of the recent research paper describing the generation of apparent stem cells from, well, non-stem cells. Here goes.
The paper in question is "Induced Pluripotent Stem Cell Lines Derived from Human Somatic Stem Cells," headlined by lead author Junying Yu, and coming out of the stem-cell powerhouse state of Wisconsin (not being facetious here, incidentally -- they really have led the way on stem cell work).
You can click here to go the article's PubMed citation. You'll need a Science subscription or institutional access to be able to read the actual article.
They started from the known fact that fusing an embryonic stem cell with a myeloid precursor (a "pre white blood cell", basically) effectively "reprograms" the myeloid cell into a stem cell. With that in mind, they made a list of genes that are highly expressed (lots of message output) in embryonic cells but not in myeloid cells. These are good candidates for the embryonic "on switch(es)" that lead to having stem-cell-like behavior. They organized this list in terms of which genes were already known to be involved in pluripotency -- that is, the ability to become any cell type. This is what we're looking for in stem cells, after all.
A set of these genes were then placed in a lentiviral vector. Lentiviruses are slow-growing viruses that can incorporate themselves -- and any passenger DNA loaded into them -- into the genome of a target organism. HIV is the best-known lentivirus. Based on their useful DNA-incorporating traits, lentiviruses are often used for work of this type.
The loaded lentivirus was then used to incorporate these genes into a line of hematopoetic (blood-forming) cells that were themselves originally derived from embryonic stem cells. They saw an increased expression of a gene that's normally seen only in stem cells, as well as stem-cell-like morphology, and the ability to make tumor-like growths in immune-compromised mice. So far, so good.
The group then winnowed down the genes until they were left with a set of four -- or possibly three. Three genes they found to be absolutely required for this transformation from somatic (normal body) cell to stem cell. The fourth gene significantly increases the efficiency of the process, and so they decided to keep using it as well.
With the selection narrowed to four genes, they tried again with a less stem-cell-like starting material -- fetal fibroblasts (basically, skin-forming cells). Once again, the loaded lentivirus was used to successfully convert these non-stem-cells into stem-cell-like cells. These product cells once again looked like embryonic stem cells, seem to have the right disposition of genetic material, and have been humming along for weeks and weeks now without problems (normally, the fibroblast cultures peter out as the cells can only divide twenty or so times before running out of steam).
Finally, they moved on to post-natal cells, and tried again with fibroblasts derived from foreskin (a ready supply of skin-forming cells, as it happens). Once again, the four genes were able to convert these to apparent stem cells. Notably, unlike the first two types of cells, these post-natal-derived cells did not all have the same behavior -- some of them kept differentiating into neural cells, whereas others kept differentiating into epithelial cells. This little quirk of behavior will be discussed in the punchline.
The overall upshot of this work is that Junying Ju and collaborators have found a remarkably small set of genes that appear to be able to "revert" certain kinds of body cells back into stem cells. There are some important caveats:
- The lentivirus approach can't be used as-is in the clinic. The weird variations they saw in the last experimental group probably come down to the lentivirus picking different areas to stick itself in the target genome. This can cause problems because it can disrupt normal expression of genes that are already there. This has actually caused cancer in at least one earlier gene therapy trial.
- It's not clear how to intentionally make the cells be not stem cells. Having these four genes cranked all the way on means that once you've converted a cell into a stem cell, it's still receiving a strong and continuous "be a stem cell" message. Part of the natural cell development process involves turning these signals off so that you can be an actual body cell. This is doable, but it's one more way that this is not an immediate replacement for natural stem cells (that will just do this themselves, with appropriate prompting).
- Finally, these cells just haven't been around that long. As the researchers themselves note, "...further work is needed to determine if human iPS [induced pluripotent stem] cells differ in clinically significant ways from ES cells."
One of the exciting upshots of this work that has not received much attention is what a great opportunity this is to learn about the natural process of cells being pluripotent or focusing on a specific function. It is fascinating, if perhaps not strictly surprising, that the process mostly comes down to a handful of genes. Knowing this kind of thing can help in our understanding of normal development and in helping to understand and perhaps prevent developmental (birth) defects.