In nature, cell differentiation is a one-way process.  Cells in an embryo are pluripotent: each is capable of turning into any cell type.  As an embryo grows and divides, the cells specialize, until finally in an adult most of them are terminally differentiated – skin is skin, neurons are neurons, one cell type doesn’t just up and change into another.  An exception to this is hematopoeitic (blood-forming) stem cells in bone marrow, which differentiate into all manner of different blood and immune cell types.  But most adult cells are done differentiating.

But it would be incredibly useful to be able to grow all different types of human cells in culture – both for research and for transplantation.  So throughout the 1980s and 1990s, people worked on figuring out how to isolate pluripotent cells from embryos and grow them in culture in the lab.  The major breakthrough came with Thomson 1998 who was able to take cells from human blastocysts and grow them for months in culture while still maintaining their ability to differentiate into a wide range of different cell types.  These cells were called embryonic stem cells (ESC).

Human ESC have now generated over a decade of good research, political battles, a now-abandoned clinical trial for treating spinal cord injury with ESC-derived oligodendrocytes, and a new clinical trial for macular degeneration.  ESCs are promising and will probably continue to bring new medical advances.  However, odds are you didn’t have the foresight to donate some of your cells back when you were still an embryo (that’s tongue-in-cheek humor, but there actually are people working on extracting embryonic cells without killing the embryo – see Klimanskaya 2006).  Therefore if you ever get an ESC transplant it will be from an embryo with a different genome than yours, and that means you’ll be on immunosupressants to keep you from rejecting the transplant (or it from rejecting you).

For this and other reasons, the holy grail of stem cell research would be the ability to create any cell type from an adult’s cells.  And while we’re not there yet, incredible advances have been made in the past seven years.

The breakthrough came with Takahashi & Yamanaka 2006, who were able to reprogram mouse fibroblasts (skin cells) to a state of pluripotency by adding four transcription factors: Oct3/4, Sox2, Klf4, and c-Myc – now collectively known as Yamanaka factors. These reprogrammed cells were dubbed induced pluripotent stem cells (iPSCs).  By the way, don’t read the original paper; read Konrad Hochedlinger’s 2012 annotated version, which is the original paper plus comments telling you what we now know.  (You’ll need to download it and open it in Acrobat – the mouseover comments don’t work in browsers).

What does it mean to ‘add’ four transcription factors?  In Yamanaka’s case it meant integrating those four factors into the skin cell’s genome via a retrovirus.  The ‘retro’ in retrovirus refers to the flagrant violation of the DNA > RNA > protein central dogma that these viruses commit by reverse-transcribing their RNA into DNA and then integrating that DNA into the host cell’s genome.  Takahashi and Yamanaka constructed one retrovirus for each of the 24 transcription factors they wanted to test, carrying RNA coding for both the transcription factor and a drug resistance gene to enable selecting for successfully infected cells by killing off the uninfected cells with a toxic drug.  So the genomes of the transformed iPS cells were permanently altered.

You might worry that permanently integrating these transcription factor transgenes would mean that these cells would be permanently pluripotent, which is to say not pluripotent at all – that they might forever have embryonic stem cell-like characteristics rather than being capable of differentiating.  But embryonic cells are usually able to silence retroviral genes (with a specific zinc finger protein that binds to them – Wolf 2009), so basically the retroviral genes were expressed just long enough to induce pluripotency, but once pluripotency was acheived, the cell shut those transgenes off again.  That made it possible for the iPSC to differentiate into a variety of cell types, which Yamanaka demonstrated by creating tumors of diverse cell types in mice subcutaneously injected with the iPSC.  Yamanaka also showed integration of the iPS cells into other mouse embryos with some amount of differentiation, but this only went so far – by the time the mice were born, the iPS-derived cells were gone.  Yamanaka suspected this was due to incomplete inactivation of the Oct3/4 transgene.

Another question you might ask is exactly what this reprogramming consists of – after all, the original mouse cells had their own endogenous copies of Oct3/4, Sox2, Klf4 and c-Myc, yet weren’t pluripotent.  Yamanaka was able to account for much of the reprogramming of the cells in terms of epigenetic modifications: methylation and histone modifications of the four factors as well as other embryonic transcription factors including Nanog.

Yamanaka co-won a Nobel Prize for this work in 2012.  To win a Nobel Prize after just 6 years is quite a tribute to the immediate and enormous impact of Yamanaka’s work.  For comparison, the other half of the prize went to John Gurdon for work he did in 1962 (the year Yamanaka was born), pioneering somatic cell nuclear transfer in frogs – the technology that later created Dolly.   Most awesome inventions take decades to realize their full impact.  By comparison, iPSC revolutionized science overnight.

Here’s a sampling of the flood of discoveries that followed Yamanaka’s work.  Wernig 2007 studied the epigenetics and gene expression of iPSC in greater detail than Yamanaka had, and found them to be indistinguishable from ESCs; he also finally was able to get chimeric mice born with a mix of embryonic and iPS cells, which Yamanaka had not achieved.  Stadtfeld 2008 [Pubmed] acheived the same reprogramming feat as Yamanaka but with adenovirus - which reproduces using the host cell’s machinery but does not integrate DNA into the host genome – instead of retrovirus.  That meant that iPS could be created without disrupting the genome.  Park 2008 replicated Yamanaka’s success in human cells, using exactly the human homologues of the four Yamanaka factors: OCT4, SOX2, KLF4 and MYC. And Yamanaka himself led the work of showing that iPSC could be created without c-Myc transgenes, although it was slower and less efficient [Nakagawa 2008].  The creation of iPSC without c-Myc transgenes is a good thing because c-Myc is an oncogene, and the mice that were eventually born from iPS chimeras had a high incidence of tumors.

Today, about six and a half years after Yamanaka’s discovery, iPSC are everyday business.  Our lab at MGH has dozens of human iPSC lines we use to study Huntington’s Disease, about half of which are ‘integrated’ (i.e. containing retroviral transgenes of Yamanaka factors) and the other half of which are non-integrated.  Of course, for studying neurodegenerative diseases, what you want are often not the pluripotent iPSC themselves but brain-specific cell types which you can derive from iPSC.

Towards that end, the past few years have witnessed a proliferation of protocols for deriving a variety of neural cell types from iPSC: motor neurons [Karumbayaram 2009], cerebral cortex neurons [Shi 2012], neural crest cells [Menendez 2013], and so on.  Many approaches are be borrowed directly from the ESC literature from before iPSC existed – Conti 2005‘s description of the growth factors needed to induce neurons directly from pluripotent stem cells is in use with iPSCs today.   Some of the approaches (though not Conti’s) take advantage of the ability of pluripotent cells to form 3-dimensional agglomerations (neurospheres, embryoid bodies, rosettes, EZ spheres) where cells signal to each other to cause differentiation.  But although a lot of progress has been made, we still don’t know how to differentiate iPSC into many of the very specific neuronal subtypes affected in neurodegenerative disease – say, medium spiny neurons in Huntington’s Disease or thalamic neurons in fatal familial insomnia.

But amidst all the reprogramming of iPSC into differentiated cell types, a new frontier has emerged as well: transdifferentiation – the direct conversion of cells from one differentiated type (say, skin) to another (say, neurons) without a pluripotent step in between.  For a review of this field see Vierbuchen & Wernig 2011.  A number of different groups are working on conversion of skin cells (since, along with blood, those are easiest to obtain from living patients) to heart, liver, or other cell types.  Wernig’s group at Stanford has been a particular leader in the quest to convert skin cells to neurons.  Vierbuchen 2011 found 3 transcription factors that sufficed to convert mouse fibroblasts to neurons capable of firing action potentials and forming synapses.  And it’s fast: mature neurons formed in about a week, much quicker than if you had to convert to iPSC and then convert back to neurons.  Pang 2012 was able to achieve the same with human fibroblasts.

For more accurate human disease modeling, it’s clear that we need to be able to differentiate more specific neuronal subtypes.  Hopefully that will allow us to see and study in cell culture the disease phenotypes that, today, exist only in patients’ brains.  Whether we’ll get there through iPSC or transdifferentiation remains to be seen, but one way or another, this field of research is surging ahead at an incredible pace.