Skipping Pluripotency

In January 2010, researchers at Stanford University announced that they had found a cell reprogramming shortcut that allowed them to convert mouse fibroblasts directly into functional neurons, without going through an intermediate pluripotent stage, using a combination of three transcription factors. Though this was not the first study to convert one mature cell type to another, it was by far the most dramatic switch. “This really was the first demonstration that such a long distance conversion can work at all,” says Stanford stem cell biologist Marius Wernig, who led the study.

Next researchers will have to show that these neurons can survive and function in the brain, says stem cell biologist Malin Parmar of Lund University in Sweden, whose group also recently directly converted fibroblasts to dopaminergic neurons. “In culture we’ve come as far as we can,” she notes. “The next step is to challenge these cells in in vivo models.” But so far, the results have been promising, and bypassing the pluripotent stage seems to eliminate the most common downside of stem cell therapies—the cells’ tendency to produce tumors. “Unlike the iPS cell reprogramming, we are never dealing with oncogenes, and we don’t deal with fast proliferating cells,” Wernig says.

In addition, this direct conversion approach is just plain simpler than iPSC techniques. “When you go the iPS route, you have two steps,” says Wernig. “First you have to make the iPS cells, but then you have to differentiate those cells into neurons—and that, it turns out, is not so trivial.” The whole process can be slow, taking many months to complete, and is relatively inefficient, with only a small percentage of cells successfully regressing to pluripotency. Those cells are then expanded in culture, and differentiated into the cell type of interest. Not only is the approach tedious, but it generates differentiated cells from only a handful of clonal pluripotent colonies, which eliminates the genetic variation found in the initial cell population, says molecular biologist Asa Abeliovich of Columbia University.

With direct conversion, on the other hand, it’s simply a matter of finding the right transcription factors. “The concept of just taking a skin biopsy, and in a few weeks having cells that you can potentially put in, and in a relatively robust and straightforward way, that’s much more realistic to me,” Abeliovich says.

Since early last year, at least half a dozen reports of such direct conversion have been published, and researchers are touting the new strategy as a safer, more efficient, and possibly more accurate alternative to iPSC technology. Multiple groups have shown that the technique can work for human cells, opening the doors for potential clinical applications. And at least two groups, including Broccoli’s, have demonstrated the ability to generate a specific neuron type—the dopaminergic neuron, which is lost from the brains of Parkinson’s patients. In both mouse and human cell cultures, researchers have shown that “these neurons really recapitulate all the functions that dopaminergic neurons directly taken from the brain have,” Broccoli says.

Mouse fibroblasts stained for lamin (green), tubulin (red), and DNA (blue).

FLICKR, GE HEALTHCARE

Furthermore, recent reports of fibroblasts directly converted to neurons with up to 50 percent efficiency suggest that direct conversion can be used to create disease models. Last month, for example, Abeliovich and his colleagues published a report detailing the successful generation of neurons from both healthy individuals and patients with a rare form of familial Alzheimer’s disease. Cells from unaffected individuals appeared normal and fully functional, even successfully integrating into mouse brains. But cells from Alzheimer’s patients exhibited abnormal processing, localization of amyloid precursor protein (APP), and increased production of amyloid beta, which forms the amyloid plaques characteristic of Alzheimer’s brains.

Having been focused on iPSC technology for the last few years, Abeliovich is excited to be perusing this new direction. “We had been really eager to set up disease models using iPS systems, but that turns out to be complicated, especially for a late onset disease,” he says. “Once we got [direct conversion] to work, we were much more optimistic that we could model a disease like Alzheimer’s.”

But direct conversion is not without weaknesses. For example, because researchers are only dealing with mature cells, which have limited expansion capabilities, it can be difficult to harvest or produce the large numbers of cells needed for research or therapy. “Fibroblasts go to senescence after 10 or 12 passages,” Broccoli says. “Then you have to get another biopsy.” With iPS cell technology, one the other hand, researchers can expand the pluripotent cells in culture until they have as many as they need.

As a work around, Sheng Ding of the Gladstone Institutes and the University of California, San Francisco, and his colleagues are taking a slightly different approach. Ding's team is inserting the four classic Yamanaka factors known to induce pluripotency, but only incubating the transfected fibroblasts for a brief period, inducing them to regress part way to iPSCs before culturing them in conditions that favor the formation of neural progenitor cells (NPCs). While this approach allows researchers to expand their cell populations in vitro, the technique’s efficiency is quite low, and some researchers are concerned that even partial regression towards pluripotency may trigger tumorigenic tendencies in the cells.

And most researchers are not worried about the numbers. Clinical trials of fetal neural transplants in Parkinson’s patients, for example, have demonstrated that only 100,000 functional dopaminergic neurons are needed for each hemisphere to get a therapeutic benefit, says stem cell biologist Malin Parmar of Lund University in Sweden, who has produced at least 70 to 142 million fibroblasts from one lung biopsy (Pfisterer et al, Cell Cycle, in press). “As you see, this is enough for cell therapy even with today’s efficiency,” Parmar says. “I don’t think it’s a major limitation.”

A more serious hurdle might be finding alternatives to the current lentivrial vectors used to deliver the transcription factors necessary for reprogramming in most direct conversion studies. Because this delivery strategy involves the integration of DNA into the host genome, it holds the potential to activate an oncogene, and thus “can never be used in the clinic,” Parmar says. But, she adds, one thing the iPSC research has taught us is that there are other ways to deliver the factors needed to reprogram cells. “If you follow the iPS field, they’ve progressed from lentivirals to non-integrated vectors to excisable plasmids, and now they’re at the ultimate goal, which is to use small molecules,” Parmar says. “I think that direct conversion has a lot to learn from how the iPS cells developed.”

While most studies to date have focused on making neurons, there are other scattered reports of the formation of other somatic cell types by direct conversion, including cardiomyocytes and hepatocytes. Though more work is required to show that these induced cells are truly functional, Wernig says, “it’s very comforting to see that things can go in many directions. All in all, these things together really point us to the conclusion that if you just find the right cocktail and conditions, you probably can convert skin cells into any other cell type you want.”

Editor's Note: This story has been updated from its original version to reflect both of Sheng Ding's affiliations: the Gladstone Institutes and the University of California, San Francisco.