In This Section      

The Mouse Model: Less than Perfect, Still Invaluable

Researchers are devising new strategies to overcome limitations of the genetic mouse model, as they search for causes and cures.

October 2010-- Valina Dawson had high hopes several years ago when she set out to design a mouse model of Parkinson’s disease. The illness affects more than 1 percent of people over age 60, and no method exists to cure the disease or slow its progression. An animal model could allow scientists like Dawson, a professor of physiology and neurology, to understand the disease and test candidate therapies.

She succeeded—more or less. Genetically, the “parkin knockout mice” she engineered mimicked a genetic mutation that causes some cases of Parkinson’s disease. However, the mice did not display the disease’s hallmark features—no trembling limbs, rigid body movements or unsteady gait.

Dawson is not the first scientist to encounter a glitch in a genetic mouse model. Since the 1980s, when biomedical scientists began adding genes to mice to engineer transgenic animals and subtracting, or “knocking out,” genes to generate knockout mice, they’ve discovered unexpected quirks.

For example, certain genetic mouse models of Alzheimer’s disease looked promising until scientists examined the animals’ brains and found missing there the amyloid plaques or neurofibrillary tangles that characterize the disease. Other models would recapitulate the pathology of Alzheimer’s disease—plaques and tangles—but not significant memory deficits.

And scientists using a mouse model that they developed of amyotrophic lateral sclerosis (ALS) found that the drug creatine appeared to delay the onset of ALS-like symptoms. But in a clinical study spurred on by those results, patients attained no benefit from the drug.

While the drawbacks of the mouse model may be especially prevalent in neurodegenerative diseases, they aren’t limited to them. “There are many, many examples of drugs curing cancer in mice that do not work in humans,” says neuroscientist David Borchelt, director of the Santa Fe Health Alzheimer’s Disease Research Institute at the University of Florida College of Medicine. “Unless a drug shows very robust results in mice, the probability that it will translate into humans is very, very low.” The mice in the ALS studies received only “marginal benefit” from the drug, points out Borchelt, a former faculty member in the Johns Hopkins Department of Pathology.

Despite such pitfalls, Dawson and other scientists are not abandoning the mouse model. Instead, they are discovering ways to circumvent or overcome the hurdles. The mouse model, they say, remains an unsurpassed research tool, despite its imperfections.

Alike, Yet Different
The mouse and human genomes are about 85 percent the same, and those similarities have made the mouse a powerful model for studying human biology and disease. Biomedical research scientists use millions of mice each year. At the Johns Hopkins School of Medicine, the Transgenic Core Laboratory generates 100 to 120 new mouse models annually. But those small genetic differences between mouse and human translate into big distinctions.
For one thing, the mouse brain is 70 percent neurons and 30 percent glia, while the human brain has the opposite ratio, says Dawson. Initially believed to be simply the “glue” or structural support of the brain, glia appear to play a more significant role in various brain functions, including helping the brain to recover after an injury.
Another difference is topography. The mouse brain’s surface is smooth, while the human brain has wrinkles, ridges, and crevices, a design that may allow us to pack more neurons into the skull’s restricted space.
And finally, mice live only about two years, while people can live for 80 years or more. That’s significant, especially for scientists who study Parkinson’s and Alzheimer’s disease and other diseases of aging, says professor of physiology Roger Reeves, faculty director of the Transgenic Core Laboratory. Such diseases may only manifest themselves after certain defective proteins accumulate in the brain and reach a tipping point where they begin to cause symptoms. Alzheimer’s disease may be one such case: Amyloid proteins may be deposited over years or decades but only begin to have an effect when someone reaches their seventh or eighth decade.

Fine-Tuning The Models
So rather than abandon the mouse model, researchers are trying to understand why their models do not perform as expected.

In the case of Parkinson’s disease models, Dawson and her colleagues have zeroed in on a group of cells in the midbrain called dopaminergic neurons. In Parkinson’s disease patients, these dopamine-producing cells gradually die or degenerate, a loss that eventually leads to tremor, muscle rigidity and other motor problems. Researchers have struggled to replicate this key facet of the disease in their animal models. But in recent years, Dawson and other neuroscientists have proposed some explanations.

One hypothesis has to do with the plasticity of the dopamine-producing cells. “When we knock out or overexpress a protein from the beginning, the dopaminergic system compensates,” says Dawson. “The mouse has compensatory mechanisms that prevent the loss of these neurons.” 

The underlying problem, Dawson proposed, was that Parkinson’s disease does not commence until a person is well into adulthood, possibly because the human brain, like the mouse brain, compensates for the loss of dopamine-producing cells. If researchers could design a mouse model to bypass this compensation so that relevant genes did not activate or inactivate until the mouse matured, then that animal might begin to lose its dopamine-producing neurons, as occurs in the human disease.

And that is essentially what Dawson has done. Using special methods of crossbreeding or using viruses to introduce enzymes that turn particular genes on or off at designated times, she believes she has scored a victory: “a mouse model that shows neural degeneration reminiscent of early-stage Parkinson’s disease.”

Dawson does not want to say more about the model before her results are published. But she is hopeful that the mice can be used to screen and test new therapies for reducing the loss of neurons.

It’s important to note that a model represents only a part of a disease process, says Dawson. Within each model, “there are important clues to be uncovered that simply cannot be done in humans. “It is through the use of multiple models and technologies that the underlying biology is revealed.” 

And models, by their nature, will always have limitations, says Reeves. “Only humans get human diseases. A model of a disease is not the condition.”

–Melissa Hendricks