Effective cell replacement therapy for Parkinson’s disease has been an elusive target. Fetal grafts have met with mixed success. Induced pluripotent stem cells produced from a patient’s own cells hold promise as a source of dopaminergic neurons, but generating and differentiating these cells is a lengthy process, and residual stem cells in the graft often form tumors in animal models, making the method unsafe for humans. Now, two new papers report that dopaminergic neurons can be cooked up directly from adult fibroblasts, avoiding the generation of proliferative cells altogether. In a Nature paper published online July 3, researchers led by Vania Broccoli at San Raffaele Scientific Institute, Milan, Italy, describe how they used a cocktail of three transcription factors to coax mouse and human fibroblasts to transform themselves into dopaminergic neurons within a few days. Separately, scientists led by Malin Parmar at Lund University, Sweden, whipped up a different recipe of five transcription factors to achieve the same result, as reported in the June 21 Proceedings of the National Academy of Sciences. The findings not only hold promise for PD treatment, but also hint that the direct conversion method may have broad applicability for many types of cell replacement therapy.

“This proves that we can dissect the pathways to get a specific subtype of neuron, starting from a completely heterologous cell,” Broccoli told ARF. He predicted that scientists may soon be able to direct differentiation into specialized cell types with even more precision.

Last year Marius Wernig and colleagues at Stanford University, Palo Alto, California, identified a three-factor cocktail that directly converts mouse and human fibroblasts into mixed populations of glutamatergic and GABAergic neurons, proving that a one-step process can work (see ARF related news story on Vierbuchen et al., 2010 and ARF related news story on Pang et al., 2011). For cell replacement therapy, however, scientists need to be able to generate specific neuronal subtypes.

To achieve this, first author Massimiliano Caiazzo in the Italian group examined 14 transcription factors, narrowing them down to a set of three—Mash1, Nurr1, and Lmx1a—that converted mouse fibroblasts directly to dopaminergic neurons with about 18 percent efficiency. Mash1 plays an essential role in specifying neuronal fate in vivo, and Nurr1 and Lmx1a are known dopaminergic markers. The reprogramming took six days, contrasting with the approximately three months needed when going through a stem cell route.

Caiazzo and colleagues extensively characterized the induced neurons, finding that the cells expressed classic dopaminergic genes, and their global gene profile closely matched that of adult mesencephalic dopaminergic neurons. Importantly, the induced neurons also behaved like normal dopamine-producing ones, with similar membrane potential, spiking frequency, spontaneous activity, and dopamine release. When the authors grafted the induced neurons into newborn mouse brains, the cells integrated into existing circuitry. Six weeks after grafting, the neurons had mature shapes, expressed dopaminergic markers, and showed normal excitability, suggesting that they could be functional in vivo. In addition, the authors used several methods to show that converted cells do not pass through a proliferative stage, and therefore should not form tumors.

Caiazzo and colleagues then extended the results to human cells, taking fibroblasts from two healthy middle-aged donors and two people with genetic forms of PD. Both types of cells converted to dopaminergic neurons equally well, but efficiency was only about 3 percent, much lower than in mouse cells. Induced human neurons had normal dopaminergic gene expression and electrophysiology.

The Swedish group took a different approach to making dopaminergic neurons. They started with the cocktail of three factors—Mash1, Brn2, and Myt1l—developed by Wernig’s group for converting fibroblasts directly into mixed neurons. First author Ulrich Pfisterer added two dopaminergic transcription factors, Lmx1a and FoxA2, to the mix. The five-factor cocktail converted mouse fibroblasts into dopamine-producing neurons with about 10 percent efficiency, the authors found. The induced neurons expressed several classic dopaminergic markers and showed spontaneous and rebound action potentials consistent with dopamine neurons.

It is not yet clear if differences exist between neurons produced by the two methods. Parmar noted that the dopaminergic transcription factors each group chose act in the same pathway, with the ones her group used being upstream of the ones in the Italian cocktail. “It shows how plastic these cells are that you can use slightly different combinations and still get the same end result,” Parmar observed. Broccoli suggested that the neurons could nonetheless have slightly different functionality, and should be compared side by side in PD animal models.

Curt Freed at the University of Colorado, Aurora, agreed. “The final test for cells of this sort is to put them into animals, and see which recipe works best for curing Parkinson-like conditions,” he said.

Both groups are currently doing just that, using rodent models of Parkinson’s to test their cells’ ability to restore motor function. If the induced neurons pass this test, they may have potential for human therapy, suggested scientists contacted for this article. One question is whether the cells will survive long-term in vivo, Broccoli said. “The dopaminergic machinery leads to strong oxidative stress, which might impair cell survival,” he told ARF. “We are closely analyzing at different time points to see how many of these neurons can survive after months, or a year.”

In addition, the methods need refinement before the induced cells would be useful for human therapy, according to the researchers. For example, both studies used lentiviral vectors to deliver the transcription factors. These viruses integrate into the DNA and therefore could activate oncogenes. “We need to find a way to induce these neurons without integrating vectors,” Parmar said. Scientists working with induced pluripotent stem cells have developed non-integrating approaches that should translate well to this system, Broccoli said (see, e.g., ARF coverage of adenoviral approaches, plasmids, transgene removal, and episomal vectors).

Broccoli pointed out another problem, namely, that the 3 percent conversion efficiency from human fibroblasts would not provide enough neurons for transplants. As few as 30,000 surviving dopaminergic neurons in a graft are enough to restore movement and allow a patient to get off L-dopa, Freed said, but because many neurons do not survive transplantation, one or two million cells might be needed in the initial graft. Broccoli told ARF he hopes to improve conversion efficiency to around 15 percent by tweaking the process in various ways, for example, by putting all three transcription factors on one vector. These technical issues should be solvable with small modifications to the protocol, Broccoli said, and should not stand in the way of therapeutic applications.

The researchers see wide-ranging potential for this direct conversion technology. Parmar and colleagues are developing cocktails for converting fibroblasts into other neuronal subtypes, such as striatal and cortical neurons that would be useful for treating Huntington’s disease and stroke. Broccoli’s group has found that other mature cell types, for example, keratinocytes, can serve as the starting material for making neurons. He is also working to produce even more specific neuronal subtypes, in particular, the A9 midbrain dopaminergic neurons lost in PD. Freed told ARF, “We are in an exciting era, where it may be possible to create any cell type, depending on the condition you are trying to treat.”—Madolyn Bowman Rogers


  1. It's nice to see other groups reproduce our data about the conversion of human neurons directly from fibroblasts, and extend it to a linage-specific stage. Morphologically, the tyrosine hydroxylase-positive-induced neuronal cells converted by BAM+LF appear to be immature, and it is not clear if they can integrate into neuronal network, i.e., form synapses, and functionally release dopamine. More characterizations are needed, such as those performed by Wernig et al. (Wernig et al., 2008).

    Wernig and colleagues' neurons derived from reprogrammed fibroblasts functionally integrate into the fetal brain and improve symptoms of rats with Parkinson's disease.


    . Neurons derived from reprogrammed fibroblasts functionally integrate into the fetal brain and improve symptoms of rats with Parkinson's disease. Proc Natl Acad Sci U S A. 2008 Apr 15;105(15):5856-61. PubMed.

    View all comments by Zhiping Pang

Make a Comment

To make a comment you must login or register.


News Citations

  1. Research Brief: From Fibroblast to Neuron in One Easy Step
  2. Turning Human Fibroblasts Into Neurons; Making Safer Stem Cells
  3. Stem Cell Advance—A Safer, Inducible Pluripotent Cell?
  4. Newest Stem Cell Approaches Abandon Viruses, Tap Testes
  5. Without a Trace: iPS Cell Techniques Leave No Footprints
  6. Circuit Menders? Neurogenesis, Stem Cells Show Potential

Paper Citations

  1. . Direct conversion of fibroblasts to functional neurons by defined factors. Nature. 2010 Feb 25;463(7284):1035-41. PubMed.
  2. . Induction of human neuronal cells by defined transcription factors. Nature. 2011 Aug 11;476(7359):220-3. PubMed.

Further Reading

Primary Papers

  1. . Direct conversion of human fibroblasts to dopaminergic neurons. Proc Natl Acad Sci U S A. 2011 Jun 21;108(25):10343-8. PubMed.
  2. . Direct generation of functional dopaminergic neurons from mouse and human fibroblasts. Nature. 2011 Aug 11;476(7359):224-7. PubMed.