With a new arsenal of robust models of ALS, drug development may move to the fast track.

HHMI investigator Nancy Bonini's sprawling laboratory at the University of Pennsylvania is filled with fruit flies. But they're not swarming. They're sequestered in glass tubes stoppered with cotton. Endless boxes of the white-tufted tubes crowd the benches and stack high on the shelves, overflowing onto the floor.

Postdoctoral fellow Hyung-Jun Kim taps the bottoms of two tubes on a lab counter, agitating the roughly two dozen flies in each. In one tube, flies zip around and scale the glass, some nearly reaching the top. In the other, the flies are sluggish. Most crawl on the bottom; a few make feeble attempts to climb but don't get far. "They don't look as happy," remarks Bonini.

The source of the stark contrast in energy and climbing ability is surprising: a human gene that is involved in amyotrophic lateral sclerosis (ALS), known as Lou Gehrig's disease. The sluggish flies carry the gene and an abundance of the protein it produces, while the sprightly flies do not.

Fruit flies have been a favorite of developmental biologists since the 1970s, with their rapid reproduction and easily manipulated genome. In 1998, however, Bonini authored an idea that radically extended the scientific reach of the humble insect. She mused that inserting genes related to human brain diseases might yield critical insights into poorly understood neurodegenerative conditions, including Huntington's disease, Parkinson's disease, and ALS. "I saw it as, 'Hey, there are all these terrible diseases and nobody is really studying them in model organisms,'" Bonini says. "I knew it was a high-risk thing."

That risk is paying off. In August, Bonini and colleagues announced in Nature a genetic factor that contributes to ALS in some patients. Bonini collaborated with Aaron Gitler, a Penn colleague who studies ALS genetics in yeast. Both are keen to define additional genes involved in ALS. Says Gitler: "My hope is that in the next three to five years we find all of the genetic contributors to ALS."

Bonini and Gitler's powerful fly and yeast tools are proving themselves just as other advanced animal models to study the disease are coming online. At Yale Medical School, HHMI investigator Arthur Horwich has developed Caenorhabditis elegans roundworms as a model organism for studying nerve degeneration, while both Horwich and HHMI investigator Hugo Bellen at Baylor College of Medicine have bred colonies of mutant mice that reliably develop ALS. Meanwhile, at the Harvard Stem Cell Institute, HHMI early career scientist Kevin Eggan is incubating dishes of human motor neurons grown from the skin cells of patients with the disease. Never before have researchers had access to human motor neurons in the laboratory, let alone nerve cells harboring the disease.

During the past 5 to 10 years, there has "almost been a renaissance in model organism genetics, applying them to human brain diseases," says Bonini. "It's been really exciting."

Sorely Needed Methods

Collectively, these new weapons in the fight against ALS provide a huge leap in scientists' ability to divine the genetic and molecular origins of the disease, says Amelie Gubitz, program director for neurodegeneration at the National Institute of Neurological Disorders and Stroke (NINDS). The menagerie of ALS animal models should also speed the development of drugs, she says.

In the past 15 years or so, patients and their families have experienced disappointment after disappointment as two dozen initially promising ALS drugs ultimately failed to help patients. Most recently, in 2009, a NINDS safety board halted a large trial of the brain drug lithium after determining it provided no benefit to patients. Though there could be many reasons why the trials failed, says Gubitz, one possibility is that the drugs were tested in animal models that did not faithfully reproduce the disease. That's why there is now "a big push to integrate the new animal models into the drug discovery pipeline."

Page 2 of 4

From left: Nancy Bonini, Arthur Horwich, Hugo Bellen, and Kevin Eggan. These HHMI investigators are contributing new models for studying ALS and testing new treatments. Photographs: Bonini: Nick Antony, Horwich: Chris Jones, Bellen: Misty Keasler, Eggan: New York Stem Cell Foundation

Gubitz is also pushing ALS researchers to collaborate more closely and to identify the strengths and weaknesses of each animal model. To that end, at the Society for Neuroscience meeting in November, Gubitz and Lucie Bruijn, of the ALS Association, gathered 125 scientists working with ALS animal models for an idea exchange. "We're trying to figure out which models should be made more broadly available," she says.

New methods to study ALS are sorely needed, as the disease remains a frustrating, deadly mystery. It was first described in the mid-19th century, when French physicians linked deaths from muscle spasticity and wastage to shriveled nerve fibers in the spinal column. In the time since, progress on understanding the causes of ALS and related motor neuron diseases—such as spinal muscular atrophy, which, in its severest form, is a fatal disease in children—has been maddeningly slow.

On a gross level, motor neurons, the long nerve cells that control movement, degenerate and die. Patients are typically diagnosed between the ages of 40 and 70, and they rarely survive more than five years. (About 10 percent of patients, including cosmologist Stephen Hawking, have a slowly progressing form of the disease and can survive for decades.) Often, control of the legs goes first, and then paralysis marches upward, eventually shutting down the lungs. In other cases, the face is affected first. The disease leaves the intellect and emotions of patients intact as their bodies wither. "It's a devastating disease," says Bonini, and it affects about 2 in 100,000 people worldwide.

The field is making progress," Hugo Bellen says. "Every year we make strides toward better understanding. But it's been a very tough nut to crack.

In 1991, an international team funded by the ALS Association and others, including HHMI, identified a form of ALS that runs in families. By 1993, the team had pinpointed mutations in a gene called SOD1 as responsible for some of these inherited cases. But only a small proportion of ALS cases—perhaps 2 percent—appear to be caused by the inherited SOD1 mutations, leaving researchers scratching their heads as to the cause of the vast majority of cases.

The search for treatments has been nearly fruitless, as well. Just one drug, riluzole (marketed under the brand name Rilutek), is approved to treat ALS, and it extends life by only a few months. A handful of other drugs are in clinical trials, but many more lie abandoned after failing in large studies. In one new treatment approach launched this year in two phase 1 trials, researchers implant nerve or bone marrow stem cells into the spinal columns of ALS patients. The hope is that the stem cells will pump out protective growth factors that rescue or rebuild dying motor neurons, says the leader of one of the trials, Clive Svendsen, director of the Regenerative Medicine Institute at Cedars-Sinai Medical Center in Los Angeles. But it will be years before researchers know whether stem cell therapy helps ALS patients.

Fingering Protein Clumps

"Oh, that's disturbing," Bonini proclaims. She's watching a big-screen monitor that displays microscope images of motor neurons autopsied from an ALS patient. Gitler, operating the microscope, points to the ugly brown smear that caused Bonini's reaction. It's a large clump of protein and it's not where it belongs.

In healthy motor neurons, this protein, TDP-43, accumulates only in the nucleus, where it's required for normal processing of RNA. But in the motor neurons of ALS patients, TDP-43 somehow escapes the nucleus and clumps in the cell body. "It's like a skein of yarn coming out of the nucleus," Gitler says. "It's just really striking. It's a dense aggregate."

In 2006, Penn Medicine neuroscientist Virginia Lee and colleagues reported finding clumps of TDP-43 in the motor neurons of ALS patients and in other nerve cells in patients with frontotemporal dementia, a condition that can cause sudden and baffling deviant behavior. The telltale clumps have also been found in athletes—including former professional football players—who later in life developed ALS-like disease and died. Subsequently, various research groups reported mutations in the TDP-43 gene (also called TARDBP) in families in which multiple members had ALS.

The discovery attracted Bonini's attention. "TDP-43 is probably a big player in different types of neurodegeneration," she says. "Everyone is trying to figure out how it hurts neurons."

Page 3 of 4

Around the same time as the TDP-43 discovery, Gitler arrived at Penn and set up a system to rapidly screen the effects of thousands of genes on yeast growth. When he spliced human TDP-43 into yeast, sure enough, clumps formed and the cells died early. Likewise, Bonini added human TDP-43 to her fruit flies, and the insects weakened, unable to move and climb normally. Those findings proved that yeast and fruit flies could serve as models of the deleterious effects of TDP-43 aggregation, an important confirmation. Bonini and Gitler hoped they could use the combined input of yeast and flies to address genes important for TDP-43 toxicity.

Gitler amped up a high-throughput genetic screening system, testing in TDP-43 yeast thousands of genes to see which, if any, accelerated or slowed cell death. They soon got an intriguing hit—the yeast homologue of a gene previously implicated in several human neurodegenerative conditions, ataxin-2.

When Bonini added ataxin-2 to her TDP-43 flies, they became even sicker; the gene clearly enhanced the toxic effects of TDP-43. "That was a 'Whoa' moment," says Bonini. "So we just went after ataxin-2."

Fruit flies on the left carry a human gene that’s linked to ALS, while those on the right are normal.

The pair enlisted the help of clinicians who treat ALS, testing samples from 900 patients and 900 controls for variations in ataxin-2. What they found was remarkable: Some 5 percent of patients carried an altered version, whereas just 1.4 percent of controls did, as they reported in Nature. "It's a risk factor for ALS," says Bonini.

The discovery also hints at a common mechanism of neuronal damage caused by interactions between ataxin-2 and TDP-43 proteins. Previously, a long form of ataxin-2 had been implicated in a nerve disease called spinocerebellar ataxia type 2 (SCA-2). That disease, like ALS, also features clumps of TDP-43 protein in certain neurons. The 5 percent of ALS patients with abnormal ataxin-2, in contrast, carry a medium-long version of the gene. "It may be that interactions between these proteins are underlying different presentations of neurodegenerative disease, but the same pathway is involved," says Bonini. "There might be a continuum of damage caused by different versions of ataxin-2."

In a relatively short time, the pair has inched closer to understanding the causes of ALS. And even if they never fully understand what triggers the disease—a fear that Bonini expressed—their yeast and fly systems can help identify potential treatments. Already, they have identified several genes that appear to slow the toxic effects of TDP-43 and ataxin-2. If confirmed, these protective genes could inspire drug development, Bonini says.

In the laboratory of Arthur Horwich, mice engineered to carry the human SOD1 gene linked to hereditary ALS serve as subjects for drug testing. To generate ALS-like symptoms quickly in the animals, Horwich revved up expression of the SOD1 protein. "In mice you might need 50 to 100 times the concentration of [SOD1]," he says. "Then you get roughly the same level of toxicity you see accreted in humans over many decades." He adds, "Of all the neurological diseases, ALS is really the most accurately recapitulated in mice. We're beginning to see what the toxic effects are at the level of the neuron. We're trying to connect the dots."

One connection has become clear: in all types of ALS, whether linked to SOD1 or TDP-43, clumps of protein aggregate where they shouldn't in motor neurons. In 2009, Horwich published two papers showing that worms or mice carrying mutant human SOD1 produce aggregates in neurons similar to what ALS patients produce. Intriguingly, SOD1 mice show that motor neurons can, for a time, successfully clear the aggregates. But eventually, the cells' ability to disperse the clusters wears down and the nerves become dysfunctional. Horwich hopes some of the wide range of drugs he's testing on mice and worms rev up the cells' ability to clear the clumps. "If we can do that, the downstream toxicity is prevented and the animals walk away free," he says.

Bellen at Baylor College of Medicine is also trying to connect the molecular dots in mice bred with ALS-causing SOD1 mutations. Over the past decade, he's focused on the junction between motor neurons and muscle cells, studying a protein that may one day serve as an early warning sign of ALS. The protein, called VAPB, directs proper development of the neuromuscular junction, and it appears to be depleted in ALS patients and SOD1 mice, several other research groups have found. A critical segment of this protein also circulates in the blood and functions as a hormone, which makes it an ideal candidate for an early detection blood test. Bellen is now supplying his ALS mice with extra VAPB to see if the protein slows or reverses symptoms—a small, early step toward developing the protein as a drug.

Though progress identifying the causes of ALS has been slow and piecemeal, researchers are inching closer to a unified theory of the molecular mechanisms of the disease. "The field is making progress," Bellen says. "Every year we make strides toward better understanding. But it's been a very tough nut to crack."

Page 4 of 4

Making Motor Neurons

Perhaps the newest means to study ALS will finally provide the scientific nutcracker researchers have long been searching for: dishes of human motor neurons grown from ALS patients. Until now, "never have we ever, ever been able to isolate a single motor neuron from a [nonfetal] source that could survive," says Kevin Eggan. He has perfected the molecular recipe for making the cells, a recipe that promises a nearly unlimited supply of motor neurons that can be used for basic physiological studies of the cells and for screens of potential new drugs.

The recipe builds on two decades of work centered in the laboratory of developmental biologist and HHMI investigator Tom Jessell of Columbia University. Jessell became fascinated with motor neurons early in his career, as they're fundamentally different from the 10,000 other types of neurons. Motor neurons send long axons—up to three feet long—outside the spinal sheathing to every muscle in the body. "They're the sole means of communication between the central nervous system and the body," he says. "So in this feature they are distinct from all other CNS neurons."

Working with chicken and mouse embryos, Jessell figured out the sequence of genetic switches that turn on and off in embryonic nerve tissue to produce motor neurons. Building on that work, a postdoctoral fellow in Jessell's lab, Hynek Wichterle, discovered in 2002 that adding two small molecules to embryonic stem cells coaxes them to generate motor neurons with high efficiency. Wichterle and Jessell told Eggan about the advance over coffee, and the young Eggan—who was then deliberating a career path—decided to work on motor neurons and ALS. "Hynek's work distilled 20 years of developmental biology into a simple and reproducible molecular recipe for making a motor neuron," Eggan says. "These cells were not like motor neurons, they really were motor neurons. They were electrophysiologically active; they made synapses with other cells."

After Eggan landed a faculty position at Harvard, he and his colleagues had some early successes in making human motor neurons, but the difficulty in obtaining a sufficient supply of human embryos slowed the work. Then in 2006, while attending a stem cell conference in Whistler, British Columbia, Eggan heard about a second advance that would prove crucial to pushing the work forward. Shinya Yamanaka of Kyoto University told attendees that he had reprogrammed ordinary skin cells to act like embryonic stem cells. Yamanaka called the new cells induced pluripotent stem (iPS) cells. "I was sitting in the back row and this wave of realization washed over me," Eggan says. He visualized a path to an unlimited supply of motor neurons carrying all the genetic mutations found in ALS patients. Step 1: Obtain skin cells from patients. Step 2: Reprogram those cells into iPS cells. Step 3: Apply Wichterle's recipe to grow those embryonic-like cells into motor neurons.

Now, after collecting skin cells from several dozen ALS patients and transforming them, Eggan is confident that the resulting cells "are real, functional motor neurons. We had to go to great lengths to show that," he says. The Harvard Stem Cell Institute is preparing to provide the cells to qualified researchers.

Valerie Estess, director of research for Project ALS, which helps fund Eggan's work, calls the advance "ALS in a dish." She says that, already, researchers at Harvard and at Columbia University are using the cells to screen for potential new treatments. "We can now model ALS more accurately," she says. "We hope that the drugs that emerge from the screens are contenders."

Of course, that's the ultimate goal for ALS researchers—to provide treatments for patients who now have few options. Yeast, worms, flies, mice, and, now, human neurons in a dish all offer platforms to quickly test drug candidates against the range of defects that lead to ALS. "It's great to have an arsenal of models to study disease," says NINDS's Gubitz. "Every model mirrors a specific aspect of disease, and they're much more powerful when combined."

And so with this solid foundation for discovery, ALS researchers expect an avalanche of advances in the coming years. Says Bonini: "You can't predict where the big breakthroughs are going to come from, which is why it's important to take many different approaches. It's really important to cast a wide net." Because, in the end, "the whole reason we're doing this is to make an impact for patients."

  • 1
  • 2
  • 3
  • 4

Scientist Profile

University of Pennsylvania
Yale University
Cell Biology, Neuroscience
Early Career Scientist
Harvard University
Baylor College of Medicine
Genetics, Neuroscience
Show More