Harnessing Stem Cells to Solve ALS


by Brenda Patoine

May, 2009

/uploadedImages/News_and_Publications/Special_Publications/Articles/1_jessell_thomas.jpg Thomas M. Jessell, Ph.D.
Professor of Biochemistry & Molecular Biophysics
Columbia University
Investigator, Howard Hughes Medical Institute


 Q: Why has Amyotrophic Lateral Sclerosis (ALS) been such a tough disease to crack?

A: I think all neurodegenerative diseases are tough nuts in their own way; if they weren’t we might have better therapies or even cures for some of them. I think it’s a reflection of how difficult these problems are in general, whether it’s Parkinson’s or Huntington’s or spinal muscular atrophy (SMA). There is progress, but it is slow and tough.

Therapeutically there is nothing in my view that works very effectively for ALS. To some extent that doesn’t distinguish it from dozens of other neurodegenerative and neurological disorders. This is a problem facing the field of neurodegenerative research in general. Despite many advances in understanding neuronal development and function, we are still not yet at the point where we can intervene in an effective way.

ALS is particularly tough for several reasons. For one, compared to another motor neuron degenerative disorder like spinal muscular atrophy, for example, which affects children, ALS is a very diverse disease. It isn’t genetically uniform. In contrast, almost all kids with SMA have a mutation in the same gene, so there is at least a rational thought process about how to approach the disease. In ALS, probably 85 percent of individuals have a sporadic nonfamilial form, so there isn’t the uniformity you see in something like SMA or Rett syndrome. This makes attempts to find a common pathway or common hypothesis all the more difficult.

Even with familial cases of ALS, there are many different genes. Superoxide dismutase (SOD1), the first gene identified as a cause of ALS, has been known for 15 years, but no one yet understands why mutations in SOD1 lead to ALS. So the heterogeneity of the disease complicates matters. In reality ALS is probably many different diseases with a common cellular and behavioral phenotype.

It has also been difficult to think of rational hypotheses or even to test hypotheses effectively because the neuron that is affected in ALS, the motor neuron, is so inaccessible. It’s buried in the spinal cord. To some extent the basic research that is focusing on ALS at the moment is trying to overcome that problem through the use of stem cell biology.

The situation is particularly stark in ALS because of the rapid progression of the disease, which typically claims lives in three to five years. There is still very little that one can do. It is a challenge to the field. The encouraging thing is because of certain basic advances, people now feel guardedly optimistic that a focus on this disease will have an impact over the next five to ten years.

Q: What are the key scientific advances that are fueling this optimism?

A: It comes back to the point that 85 percent of people with ALS have no obvious genetic component. So the question is, why do motor neurons die? Is it an environmental toxin? Is it a disorder in glutamate clearance? Is it a protein defect? What is cause and  what is consequence? None of these questions have really been answered. The problem is, how do you  test hypotheses in an adult-onset disease that affects motor neurons?

One way forward is to find a way to study in vitro large numbers of human motor neurons that carry hallmarks of the disease. If you can do that, it means you can test some things rationally and apply the advances in genomics and biochemistry to the problem. Remarkably, that is beginning to become possible now.

Our interest in ALS has been primarily at the basic scientific level of trying to understand more about the normal program of motor neuron differentiation and connectivity. Our view is that, if we understood how motor neurons are generated normally, how they form connections with muscle, how central connections in the spinal cord form, and so on, then we would have a better chance of understanding what has gone wrong in a disease like ALS or SMA. So we’ve spent years trying to understand the normal program of motor neuron development.

A few years ago a fellow in my lab, Hynek Wichterle, decided that if one really knew enough about the normal program of motor neuron differentiation, then one should be able to take the same developmental signals and turn another cell type, such as an embryonic stem cell, into motor neurons. This would allow one essentially to generate unlimited numbers of motor neurons. Then one could think about applying ALS genetic insights onto that. That scenario has now become possible, not only in the mouse where we did it, but also in humans.

There was a very important paper published in Science1 last August by my Columbia colleagues Wichterle and Chris Henderson, together with Kevin Eggan at Harvard. They took advantage of the ability to generate motor neurons from stem cells, using this remarkable method for reprogramming skin fibroblasts to become stem cells that was first demonstrated by Satoshi Yamanaka. Wichterle and colleagues took fibroblasts from a patient with ALS and used the transcription factors Yamanaka has defined to reprogram those fibroblasts into stem cells. Then they used the differentiation protocol we had developed to turn those stem cells into human motor neurons. So it was the combination of the basic research that we’ve done on the motor neuron, the Yamanaka reprogramming, and Eggan’s expertise in human embryonic stem (ES) cells that all had to come together to produce this. Three different backgrounds had to converge to get this result, which was the first real example of putting all those individual pieces together.

Now one can generate patient-specific human motor neurons in the billions. You can do that in familial cases where you know the genetic lesion as well as in sporadic cases where you don’t know the nature of the insult. Eggan has done this: there are now at least two dozen patient-specific ALS embryonic stem cell lines that can be effectively converted into motor neurons. This gives one the ability to test any hypothesis in human motor neurons that are bearing some reflection of the disease. This is very much the state of the art at the moment.

Q: You and your colleagues have also found that astrocytes play a key role in the demise of motor neurons in ALS. What is the significance of this finding?

A: This work was published in 2007 in Nature Neuroscience2 with my Columbia colleague Serge Przedborski, who was really the leader on this; my group provided the basic scientific perspective to ask the right questions of embryonic stem cells. Serge made the remarkable discovery, which was also made by Tom Maniatis’s group at Harvard, that part of the reason motor neurons die in ALS is because when the mutant SOD1 gene is expressed in astrocytes, the astrocytes release a toxic factor or factors that act selectively on motor neurons to prompt their demise. I think these two reports together have really motivated the field to ask questions about whether this is true in humans.

There were two papers published in Stem Cell in December 2008, one from Eggan’s group and one from Rusty [Fred] Gage’s group at Salk Institute, that have shown that the same phenomenon is true in humans, perhaps even in a more robust way. Now we can start asking questions about what actually is being released by astrocytes expressing the mutant gene that damages motor neurons. At the moment, there are various candidates that might explain this.

To me, what’s most important is that, with the availability of induced pluripotent stem (iPS) cell-derived motor neurons from individual patients, you can now test hypotheses in a rational and a rigorous way, including hypotheses aimed at understanding what is going on with the astrocytes. That was just not possible before. You couldn’t think of doing high-throughput drug screens to find compounds that prevent the death of motor neurons, regardless of what the astrocyte toxic factor is. That is going on now. Suddenly, with this iPS programming, that limitation or bottleneck in actually getting cells to do research has disappeared. This all happened in the last three to six months of 2008.

Now you can try to identify what the astrocyte factor is, and you can try to identify what the mutant gene does in motor neurons or in astrocytes. Suddenly you can bring the world of contemporary biology to bear on this problem. One has to be optimistic–at least I’m optimistic–that this rational and rigorous approach is going to change the way we think about the disease in the next five years, to the point that there will be some consensus as to common cellular mechanisms that link the familial and the sporadic cases. This will in turn provide more objective, rational ways of finding therapies.

Q: How do you foresee the approach to therapeutic development changing as a result of the ability to generate motor neurons from ALS patients’ cells?

A: Most of the drugs that have been screened to date in clinical trials have been disappointing failures. I think that’s because, even though you can find drugs that have some effectiveness in the mouse model, if you don’t have a better cellular assay you have only a limited shot at success in humans. So at the very least, these ES cell-derived assays are going to allow you to come up with a better set of leads for drug development. Eventually that has got to result in better therapies.

You also want to get the pharmaceutical companies interested, and they don’t know how to deal with mouse model assays; such assays are just incompatible with the way that pharmaceutical or biotech companies generally think. But if you can create assays based on ES cell-derived motor neurons, it will allow you to screen chemical libraries of a million or more compounds to find lead drug targets. Then you can optimize those targets and test the most promising of them in the mouse models. You then have a sort of pipeline, or a way forward, to move from basic drug discovery to something that will work in animal models that should then allow you to identify more effective compounds for human clinical trials.

The great thing about these iPS cell lines is that you can derive these lines from patients with sporadic forms of the disease, where you don’t know the nature of the biochemical lesion. As a result, you now have a chance to ask if the changes in motor neurons expressing the mutant SOD1 gene are the same as the changes in motor neurons from a patient with sporadic disease. By doing that with two dozen lines, you should be able to get better characterization of disease phenotypes. It may turn out that ALS is really seven different diseases, so therapies that will affect ALS disease type one may be quite different from those that affect ALS disease type five or seven.

Q: The induced stem cell work as applied to ALS has been hailed as the “discovery of the year” by the science and lay media alike. Why so much attention on such a relatively rare condition as ALS?

A: ALS is the first disease to be studied using skin cells from patients to create disease-specific stem cells for research. But I think part of the reason this has received so much attention is that you could in principle do this for any neurodegenerative disease. It’s just that the methods for turning stem cells into motor neurons are much better worked out than for almost any other cell type, so ALS was a natural first target. It’s a fortunate coincidence of scientific developments over the last five years that has propelled ALS to the forefront of this kind of research.

There was also a paper by George Daley published in Cell shortly after the ALS work was published that did the same thing with a number of other diseases, not only neurological diseases. He showed that you could take these fibroblasts from different patients and make embryonic stem cell lines, although he didn’t actually differentiate them into various cell types. I think it’s still true that ALS is the only one so far where researchers have gone all the way through the program and generated well-characterized human motor neurons from an individual patient.

Now we can create assays to ask what biochemical or genetic changes there are in motor neurons from ALS patients that you don’t find in normal unaffected individuals. That work is going on in half a dozen places as we speak. Because this is all so recent, many researchers are just now taking the initial observations and running with them. It’s happening in biotech labs, in stem cell labs, in motor neuron labs, even in biochemistry labs, which have gotten interested in the disease because suddenly you can do science in a way that you couldn’t before.

This is what you want. These are such tough problems that it’s unclear where the next breakthrough is going to come from. If there are only two labs pursuing it, it is going to take longer to reach that breakthrough than if there are three dozen labs doing it, each imposing its own biases and prejudices and testing different theories. In principle, all of these theories can now be tested on the same cell line or the same set of motor neurons, so more cross-validation will emerge. This is what is needed to drive the field forward. If I were a patient or a relative of a patient, that’s the way I would want the field to move, as rapidly as possible.