Amyotrophic lateral sclerosis (ALS, also called Lou Gehrig's disease) causes motor neurons to degenerate and die; people with the disease gradually lose the ability to move. Although 10 percent of patients (including scientist Stephen Hawking) have a slowly progressing form of the disease, the vast majority of people diagnosed with ALS die within five years of diagnosis.
Progress toward effective treatments has been agonizingly slow. Many potential drugs failed in large clinical trials; now, only a handful of new drugs are in trials. And while researchers have found more than a dozen genetic mutations associated with ALS, no one's sure how these play into the 90–95 percent of cases that are sporadic (cases where patients are the first in their families to have the disease). Even for the small percentage of patients with known gene mutations, there's no real disease-modifying treatment available. Only two drugs are approved to treat ALS: one can extend patients' lives by a few months (riluzole); the other treats emotional instability that accompanies the disease (Nuedexta). Neither is the "cure" patients and their families want.
The fact that most cases of ALS develop in the absence of a known gene mutation isn’t helping research move more quickly. The current models—whether stem cell, animal, insect, or yeast—are all based on these genetic mutations. Sporadic cases may not develop by the same mechanism as genetic cases. Another reason that research has been so slow to yield results is that there isn't a coherent or consistent hypothesis among scientists about the underlying cause of ALS.
"The field is a complete mess," says Tom Jessell, a developmental biologist at Columbia University and a member of the Dana Alliance for Brain Initiatives. "Everyone's at liberty to pursue their own pet ideas, which our lab is doing as much as anybody else. At some point there has to be something so compelling that the entire field consolidates around a central idea." Some recent technical progress, though, could help find that compelling “something” faster.
A Disease in a Dish
Jessell thinks that the ability to quickly test ideas—in months, not years—about the underlying pathology of ALS is key to finding that central idea. For decades, he has studied the normal development of motor neurons, identifying the progression of genetic switches that transform embryonic tissues into motor neurons. In 2002, Hynek Wicheterle, working as a postdoc in Jessell's lab, used this knowledge to add two small molecules to embryonic stem cells in culture, causing them to grow into highly specific subtypes of motor neurons. "Suddenly, for the first time, you could make motor neurons in basically unlimited numbers," says Jessell.
However, using stem cells from human embryos is controversial, so they are in short supply. In 2006, Shinya Yamanaka of Kyoto University manipulated genes to revert normal, adult skin cells to an embryonic state. These induced pluripotent stem (iPS) cells act much like stem cells from a human embryo. Researchers have turned the skin cells of adults with ALS into iPS cells in the lab, studying and manipulating them to see if they grow into exact replicas of motor neurons with ALS. If it's proved that they do, scientists have what is essentially a "disease in a dish," a way to work directly with human motor neurons in the lab.
"I'd say that so far, nothing dramatic has been seen," says Tom Maniatis, a researcher at Columbia University who studies ALS and works with iPS cells. "The overall effort has been much more difficult than we anticipated." One of the problems, he says, has been that once iPS cells are generated, there is a lot of variation between cell lines. "The genetic background is such that if you isolated iPS cells from two patients carrying the same mutation, there would be a difference between those cell lines," says Maniatis. "It's an issue we're all struggling with: Just how much variability is introduced by the reprogramming process?"
This August, Yamanaka, with lead author Harushisa Inoue and colleagues, published a paper in Science Translational Medicine that suggests that iPS cells could be valid models for studying ALS as well as for screening potential treatments. The researchers made iPS cells from ALS patients with mutations in Tar DNA binding protein-43 (TDP-43). The mutation is associated with a small number of familial ALS cases, but clumps of TDP-43 also are found in many other people with ALS. These iPS cells did show some of the ALS phenotype, forming clumps of TDP-43, increasing the expression of genes involved in RNA metabolism (known to be affected by TDP-43), decreasing expression of genes that code for filaments that help structure neurons (decreased in postmortem samples from patients with ALS), and responding to a drug, anacardic acid, that improves ALS symptoms in mice. The effects were small but significant.
“If you look carefully at the data, the differences are exceedingly small, both in the patterns of expression and the response to these different drugs,” says Maniatis, who was not involved with the study. The results are encouraging as far as the viability of iPS cells for studying ALS, but there's more work to be done.
Mice, Yeast and Flies
In addition to pursuing a "disease in a dish" model, researchers are using animal and other models to push the field forward. Already, scientists have made discoveries by working with mouse-derived embryonic stem cells, for example. Looking at an inherited form of ALS that involves a mutation in a protein called SOD1, Maniatis, Eggan and colleagues found one reason motor neurons die in ALS: changes in glia, the supporting cells that surround all motor neurons. The team generated motor neurons and glia from embryonic stem cells with the SOD1 mutation, then isolated glial cells from the tissue culture and added them to a culture of typical (non-ALS) motor neurons, as well as to motor neurons with the SOD1 mutation. The SOD1 glia killed both types of motor neurons, although the effect was more pronounced on the SOD1 motor neurons. It seems, then, that the expression of SOD1 in mutated glial cells causes the glia to secrete a toxic factor that kills the motor neurons. The discovery has led to a hunt for the nature of the toxic factor: block it and perhaps you rescue the motor neurons.
Aaron Gitler, a geneticist at Stanford University who studies ALS in yeast, and Nancy Bonini, a biologist at the University of Pennsylvania who studies it in flies, have had a fruitful collaboration yielding data about the mechanisms that underlie ALS. After the genetic mutation in TDP-43 was identified, Gitler decided to study it in yeast. In healthy cells, TDP-43 stays in the nucleus, but in ALS, it's found in the cytoplasm, and forms clumps that lodge in the motor neurons.
Gitler expressed the TDP-43 gene mutations found in humans in his yeast model. "We found that it aggregates and is toxic in yeast as well," he says. While the "wild" form of the protein—the one that is not associated with disease—formed clumps as well, the mutated forms associated with ALS clumped and aggregated even more rapidly.
The next step was to screen the yeast genome—easy to do because it's well-characterized—to see if they could find any gene mutation that would either rescue the cell from TDP-43 toxicity or make it worse. "This is the power of yeast genetics," Gitler says. "We can overexpress or knock out every gene, one at a time."
The results yielded forty such genes; the researchers are still sifting through the data, but have found one mutation in particular that made cells significantly sicker when overexpressed. That gene codes for a protein called ataxin 2. That's where the collaboration with Nancy Bonini began.
"It turned out that Nancy had already been studying ataxin-2 for many years in fruit flies," says Gitler. And ataxin-2 was already implicated in another neurodegenerative disease, spinocerebellar ataxia 2, or SCA2. When the two scientists crossed flies with the TDP-43 mutation with flies with a mutation in ataxin-2, they found the flies' symptoms became worse, just at Gitler's team had seen in yeast, further validating ataxin-2's role in ALS.
SCA2 is caused when the gene that codes for ataxin-2 is mutated so that it has a segment that repeats the same series of nucleic acids many times, a so-called polyglutamine or polyQ segment. Abnormally long polyQ repeats have been associated with several neurodegenerative disorders, including Huntington's disease. In the normal population, the ataxin-2 protein has 22 such repeats. In people with SCA2, that number is 34. Gitler and Bonini hypothesized that maybe something in between those numbers, "say, in the 24 to 33 range," says Gitler, put people at risk for developing ALS.
To test their hypothesis, they analyzed the DNA of 1,000 people with ALS and 1,000 healthy controls, looking specifically at the size of the polyQ repeat in the ataxin-2 gene. The results, published in 2010 in Nature, confirmed it: expansions of an intermediate length (27 to 33 repeats) were much more common in people with ALS than in healthy controls. Although the study—confirmed by other labs who are testing their ALS patients as well—seem to point to these repeats as a key factor in the development of ALS, Gitler urges caution. "We have to be careful, because we did find some intermediate-length repeats in controls. We don't want someone to think they will get ALS because they have this length of repeat. It seems to be a risk factor, but it isn't a guarantee."
TDP-43 is an RNA binding protein, meaning it is involved in the transport, stability, and translation of messenger RNA, which carries DNA’s message to the ribosomes where the protein the DNA codes for is created. FUS is another recently-identified RNA binding protein that also forms clumps in the cytoplasm in ALS. Recently, RNA binding proteins have been implicated in other neurodegenerative disorders, including Alzheimer’s disease and a type of dementia called frontotemporal lobar degeneration with ubiquitin-positive inclusions, or FTLD-U. Mutations in TDP-43 and FUS have been linked to this type of dementia, and some patients with ALS show symptoms very similar to FTLD-U.
One Disease, or Many?
There may be a reason that research into ALS has failed, thus far, to yield a deep and coherent understanding of the disorder’s underlying pathology. It seems that the deeper scientists dig, the more genetic mutations and aberrant pathways surface. Probing into the genetic mutations has yielded multiple complex pathways that contribute to the death of neurons. For example, clumps of TDP-43 form in the cytoplasm, as Gitler and Bonini found. Other research has found that the stability of proteins, which must fold and maintain their shape to function properly, is affected. The way RNA is made, transported, and packaged is abnormal. Transport of cellular substances such as proteins, neurotransmitters, and enzymes along the axon, the long slender thread of a neuron that stretches to reach skeletal muscles, is disturbed. "This is an extremely complicated disease," says Jessell. "It may not even be only one disease, it may be many different diseases, all of which manifest in the same phenotype."
Scientists hope that by creating models of disease in different animals—from flies to reprogrammed human stem cells—research will converge on an answer, or multiple answers. "I think you have to use all these models, with different mutations, in the hopes that there will be one that really does provide the breakthrough," says Maniatis. For patients and families, that breakthrough can't come soon enough.