Researchers use rodent models to unlock the mystery of the autistic mind and bring autistic children back into the social world.
Humans are social creatures. Any psychiatrist or social scientist would subscribe to this credo. However, some people have less interest in social interaction than others and for them, it is hard-wired into their brains, i.e., introversion; but introversion is not a disease. For others, there is also hard-wiring for nonsocial behavior where it is a disease: autism.
“When you look at autism in humans, what you see are deficits in language, deficits in social interaction, bizarre or aberrant motor patterns, and repetitive behavior, sometimes to the point of self-injury,” says George Wagner, PhD, professor of psychology, Rutgers University, Busch Campus, New Brunswick, N.J. “The child might fixate on a particular behavior and just repeat it non-stop until you can distract or bring in some other behavior, probably the behavior that is most likely to be incredibly frustrating for the parent and may even lead to institutionalization [of the child].”
According to Wagner, between 30% and 40% of children with autism start off with an apparently normal developmental trajectory, whereby they learn language skills, motor skills, etc. Wagner classifies behaviors in human autism as retardations; autistic children are often mentally-retarded or retarded in the development of many skill sets. At some point early in childhood, he says, some event triggers a regression, after which “the child either loses skills that they had already mastered or their trajectory of development is severely altered that they start falling behind their normal trajectory.”
Regression can be devastating to parents of autistic children. But luckily, dedicated autism researchers like Wagner are trying to develop animal models that allow for a better understanding of autism and may some day lead to better drugs to the disease. As discussed below, there are rodent models for human autism. Like every animal model, rodent models of autism have their limitations, which make finding good models so challenging. So how does one find a good model for autism?
The good model
“The endeavor to model in animals, particularly rodents, is challenging because the clinical manifestation of the disorder tends to involve behaviors for which there is not an obvious direct counterpart in animals,” says Mark Stanton, PhD, professor of psychology, University of Delaware, Newark, Del. “So that makes the issue of animal models controversial.”
According to Stanton, there are four dimensions to consider in designing the rodent model of autism. The first, the behavioral dimension, defines how the disorder under study manifests itself in the behavior of the animal model. In the second or neurobiological dimension, this behavior may be due to neuroanatomical or neurochemical alterations in both the human and the animal model. “In many cases, it is easier to argue with parallelism between the human disorder in the animal model [and the human] at the level of these neural alterations than it is at the level of behavior,” says Stanton, who explains that of all of the brain regions that have been reported to be altered in human autism, cerebellar alterations are the most common.
Since autism is a developmental disorder, there must be a developmental dimension in its animal models. Stanton tests his rodent model for eyelid conditioning, a simple form of reflex conditioning that is homologous throughout many animal species. This reflex has a profile of development that’s understood in rodents and in humans and involves development of the brain stem and cerebellum, areas that are affected in autism. Changes in brain development implicated in autism occur early, hence the reason for the manifestation of autistic symptoms during early childhood. So, if the animal model only exhibits the autism-like behavior as an adult, it is not likely to be a good model for autism. Therefore, the animal model must exhibit the autism-like behavior at an age equivalent to that of an autistic child.
The final dimension has to do with the etiology or cause of the disorder. So in looking for the cause in the human disease, researchers would look to see if there is a similar cause in the animal model. One way to identify etiology of disease is to pharmacologically interfere with the suspected cause of the disease. And if the disease is undetectable after the drug is administered, then the cause is identified.
Of mice and rats
In humans, autism is a broad spectrum of behaviors that appear in different degrees across children, researchers say that a good animal model should test at least some of these behaviors. “We want to ask some of the questions you can’t ask in the clinical population and basically what we do is look at the clinical data and we try to see if there are some behavioral features that we can model in a particular animal model,” says Benjamin Walker, PhD, assistant professor of psychology, Georgetown University, Washington D.C.
Walker uses the Sprague-Dolly rat because there is a lot of information on its neural circuitry. His initial experiments with the rat model of autism tried to answer the question: why do a large number of the autistic children get seizures? The seizures were thought of as a by-product, not a cause of autism, says Walker, but that they occur before the other behavioral abnormalities and some are sub-clinical. So his hypothesis was that seizures seen in children prior to onset of autism change the neurobiology of specific neural circuits, they might lead to some of the symptoms associated with autism. Walker is trying to determine which circuits are responsible for the social behavioral response in rats with autism-like behavior. By perturbing in the model those candidate systems that have been reported to play a role in development of human autism, Walker is getting closer to determining the neurobiological basis of the disease. For example, he destroyed the cholinergic basal forebrain in adult rats and then tested the rat on a social interaction task.
“A lot of people were doing social interactions but they were putting two rats together in an open field to see how they attract each other or go towards each other or interact in some way. And we thought that that wasn’t necessarily a good model for what was going on in autism,” says Walker. “We wanted to see, if given a preference, would the rat (and we think this is also the case in humans) prefer to be in a social interaction versus not.”
The social interaction task he uses is a three-chamber task where a “test” rat is placed in a center chamber. Then, on either the left or right side of that chamber, there is a control rat in one chamber and the net from the “test” rat’s home cage in the other. The “test” rat is “asked” to decide how it likes to spend its time across a 15-minute test. Photo beams record the time the “test” rat spends in each chamber. Walker has found that the control rats are overwhelmingly social. However, rats with lesions in the acetylcholine or serotonin systems in their forebrains did not engage in social interaction and spent more of their time in their home cage.
In terms of weaknesses of the Sprague-Dolly model of autism, Walker had this to say. “It’s an animal model of a very complex human syndrome and when we change the acetylcholine pathway, we see similar symptoms to autism. However, these are symptoms common to a lot of underlying pathologies.”
Beyond transgenics
Finding a good animal model of autism is so difficult that autism researchers must often feel fortunate to find a model that works. Researchers like Valerie Bolivar, PhD, research scientist and director of the Mouse Behavioral Phenotype Analysis Core, Wadsworth Center, New York State Department of Health, Albany, N.Y., know the feeling. A few years ago, she discovered a normal inbred mouse strain that shows very low levels of social interactive behavior. This mouse, called BTBR, comes from Jackson Laboratories, Bar Harbor, Me., and has enabled Bolivar and her colleagues to study autism. With this model, she has been able to establish that these social abnormalities occur at a young age.
“The advantage of this mouse over the transgenic and knock-outs is that it’s an inbred strain, so you don’t have any of these nasty problems of what happens when a gene is totally gone or something is added and where it’s inserted,” says Bolivar. It’s more of a natural type of model system.
In a three-compartment test for social behavior similar to the one Walker is using, the normal mouse prefers the company of a stimulus rat, whereas the BTBR does not. “It is just as likely to spend its time with the empty container as with a live mouse. They seem to pass like ships in the night,” Bolivar says.
There are also suspected genetic factors that greatly increase the risk of autism. And although only some of the genes have been identified, what is clear is that autism is polygenic, requiring alterations in many genes to cause disease. Bolivar is also interested in finding genes that are involved in this abnormal social behavior and then to hopefully extrapolate some of the genes to behaviors characteristic of human autism.
Although Bolivar knows that she is not going to find a mouse with autism, she does see the possibility of finding some of the underlying pathways involved in social behavior. And by understanding “normal” social behavior, she is hopeful that researchers will one day understand social behavior in autistic children.
This article was published in Drug Discovery & Development magazine: Vol. 10, No. 11, November, 2007, pp. 18-22.
Filed Under: Drug Discovery
