Cover Story

Flip through the pages of today's most prestigious scientific journals and you're likely to find study after study that used one of several powerful new genetic techniques to link a gene to a behavior or personality trait, from parenting to promiscuity to drug use.

These transgenic and knockout models allow researchers to add or delete entire genes from animals, most often the mouse. The resulting studies are grabbing newspaper headlines, announcing the discovery of an "anxiety gene" or the "genetic root for aggression."

And as molecular biologists continue to refine the techniques, the methods will become even more potent, allowing researchers to turn genes on and off at will in specified tissues of the body. As a result, the use of genetically mutated animals may begin to dominate some areas of behavioral research. Already, funding agencies are putting large amounts of money into the area, presenting a great opportunity for psychologists--as well as some challenges.

By most accounts, there's a severe shortage of behavioral experts working with the molecular and cellular biologists who create the mutant mice. The result is sometimes overstated conclusions from behavioral studies conducted by molecular biologists without strong behavioral training.

"To date, I have been dissatisfied with much of what's been published," says Steven Hyman, MD, director of the National Institute of Mental Health (NIMH).

Many studies use sophisticated, cutting-edge molecular genetic techniques, but simple, unsophisticated behavioral models, and then the researchers make claims that they've found genes for behavioral traits such as aggression and anxiety.

"That's absolutely inappropriate," says Hyman.

Part of the problem is that some molecular biologists think studying behavior is easy so they try to measure it without knowing the literature, says behavioral researcher Jacqueline Crawley, PhD, who runs an intramural lab at NIMH. But the bigger problem is a dearth of behavior experts working in the field, which forces the more molecular folks to go it on their own.

The solution, says Hyman, is to encourage more psychologists and behavioral neuroscientists to collaborate with molecular geneticists on gene mutation studies. But moving into the world of molecular genetics will require most psychologists to change the way they conduct their research. For one, most of these studies are conducted on mice. So psychologists used to working with rats will need to switch to a new, and less well studied, animal model.

In addition, psychologists will need to brush up on biology and learn more about genetics to build productive collaborations with molecular and cellular biologists.

"We need to train both biologists and behavioral scientists to collaborate with each other," says Alan Leshner, PhD, director of the National Institute on Drug Abuse (NIDA). "And we need to give them enough understanding of each other's domain so that it's a meaningful collaboration and not a pasting together of biology and behavior."

What's wrong with my mouse?

Many excellent biologist-psychologist collaborations are already up and running, including labs at New York University, Johns Hopkins University, the National Institutes of Health and Scripps Research Institute. Still, say Hyman and others in the field, studies continue to get published that take a precisely engineered strain of mouse, then conduct a superficial investigation of behavior.

There are several reasons for that. For one, some molecular and cellular biologists, often untrained in studying behavior, invent naïve and poorly controlled behavioral tests.

"The molecular biologists are developing all of these gene manipulations but often are doing crude behavioral assays and coming to incorrect conclusions," says Johns Hopkins psychologist and neuroscientist Randy Nelson, PhD, one of a small cadre of behavior experts working with genetically mutated mice.

In one case, a group of researchers claimed to have found a gene related to aggression, but it turned out the mice were aggressive because the mutated gene made them blind so they reacted aggressively when tested.

"They didn't bother to do a simple sensory battery before testing the mice for aggression," says Nelson.

In fact, the first line of testing for any new mutant strain of mouse is a battery of sensory and motor tests. They tease out artifacts that might cloud the findings from more precise behavioral tests, says NIMH's Crawley. Before you can blame a behavioral change on a gene mutation, she says, you have to ensure that some other deficit isn't causing the change.

One drawback of genetic manipulation techniques is their lack of specificity. Although they delete a single gene, they do so in every tissue of the body and the gene is absent throughout development.

"You're talking about a big whack," says Leshner. "To have an animal grow up with no dopamine reuptake transporter, for example, is serious stuff. You don't know all the systems it might affect."

Molecular biologists are developing techniques to bypass these limitations. One technique will produce what are called "inducible" knockout mice, which will be created with a gene that researchers can turn off and on. To ensure that an animal develops normally and at a specified time, researchers switch off the gene and watch what happens to behavior.

Another technique will produce "conditional" knockout mice, which will be missing a gene in one specific tissue--for example in only a certain area of the brain. This technique would allow researchers to examine how a gene in one area of the brain affects behavior without interference from other systems.

Combining these techniques would allow researchers to examine the effects of specific brain products on specific brain areas at precise times during development. For example, they might evaluate an Alzheimer's model in aged mice. Even then, however, molecular geneticists will need the help of behavioral experts to properly examine any changes in behavior that might arise.

"Behavior is a very important phenotype, and it's very complex," says Nelson. "Molecular biologists often don't know about gender differences, they don't know about developmental differences, they don't know about motivation. But we do."

Crawley is writing a book for John Wiley and Sons that will describe the complexity of studying behavior in mice--what's called behavioral phenotyping. The book, "What's Wrong with My Mouse? Behavioral Phenotyping of Transgenic and Knockout Mice," is written especially for molecular biologists.

But, says Crawley, she's added caveats "not to try this at home" without help from a good behavioral expert. Of course, that means biologists need to find psychologists and other behavioral experts willing and able to help them. But by most accounts there's a severe shortage.

"The genetics is becoming the easy step," says NIMH's Hyman. "The hard step is relating the genetics to brain and behavior."

Reflecting on rodents

Some psychologists resist moving into the area because, for them, the idea of abandoning the rat for the mouse is unthinkable.

The rat has proven an extremely good model for many human behaviors, it's relatively easy to train and it has a large enough brain to do elegant neuroanatomical work, they say. In addition, most of what psychologists know about behavior and its link to certain areas of the brain comes from models perfected in the rat. It may or may not be the case that what they've learned about learning and memory, motivation and mating behavior in the rat will translate to the mouse, says Nelson.

"It might not be the case that the hippocampus will do exactly the same in the mouse as it does in rats," he says. "The solution is to either make knockout rats or redo the past 15 years of rat behavioral work in mice."

The job will most likely fall to behaviorists to redo the behavior, admits Nelson, who has been working on models of aggression and mating behavior in the mouse for about five years. Too much money has been spent on deciphering the mouse genome, he says, to suddenly switch to rats.

Besides, he adds, doing the genetics in rats would be extremely difficult because of the large numbers of animals needed to create knockout and transgenic animals. Rats are simply bigger and far more costly to house to make them a practical alternative, say researchers.

"You can think of an analogy with molecular geneticists building a railroad track from the East Coast to Utah, using rails of a certain width," says Nelson. "Meanwhile the psychologists are laying tracks from California to Utah using a different rail dimension. The people who spent the most money are going to have their way. That means that for most situations we will have mice to evaluate the effects of genetic manipulations upon behavior."

The rat will not become obsolete, predict agencies that fund basic behavioral research.

"NIH won't abandon the rat," says Hyman. "But researchers should recognize that there are things that can be done in the mouse on the genetic level that can't be done in the rat. My advice would be, don't give up the rat but begin to think about mouse genetic models that can supplement your work. Not because that's where the money is, but because it's such a powerful tool."

But it is also where a lot of new money is. Although NIMH still funds far more behavioral research on the rat than it does on the mouse, the institute along with six other institutes including NIDA, the National Institute on Alcohol Abuse and Alcoholism and the National Institute on Aging have put nearly $40 million over the next five years into special grant solicitations designed to bolster work in the area of mouse behavioral genetics, and has hosted several large conferences on the future of the field.

"These animals are another tool," says psychopharmacologist Lisa Gold, PhD, of the Scripps Research Institute. "But I wouldn't dispense with pharmacology or brain lesion studies."

In fact, Nelson discourages his graduate students from focusing solely on studying animals with genetic mutations. He fears they will become, essentially, technicians for molecular geneticists who want someone to take their mutant mouse, test it for behavioral changes and then give it back.

"It's often very atheoretical work" to simply test the behavior of a mutant mouse, says Nelson. "But if we do it in conjunction with hypothesis-driven questions and with other tests-- confirming our findings through drug studies or brain manipulations--then it's an interesting enterprise."

Homing in on behavior

Behavioral researchers also have a critical role to play in designing specific, reliable behavioral tests for the mouse that can be replicated across laboratories.

"We need to have well-characterized, robust behavioral paradigms in specific strains of mice that can be reproduced from lab to lab," says Hyman.

The need for standardizing such techniques was emphasized by a much-cited article published in Science (Vol. 284, p. 1670-1672) last year. Psychologists John Crabbe, PhD, at Oregon Health Sciences University; Douglas Wahlsten, PhD, at the University of Alberta, in Canada; and Bruce Dudek, PhD, at the State University of New York, Albany, conducted the exact same behavioral tests on the exact same strains of mice, controlling for as many environmental variables as possible to see whether their techniques would reproduce across different labs.

Some held up beautifully. For example, across all labs, the same strains of mice preferred alcohol to water and vice versa. But other behaviors varied, depending on which lab the animals were tested in. For instance, with a test of activity that used an open-armed plus maze, a certain strain of knockout mouse showed increased activity, compared with normal mice in Oregon, decreased activity in New York and no difference in Canada.

The results sent a shock wave through the research community, with many people questioning whether behavioral work was worth doing in these animals. For Crabbe and his colleagues, it emphasized the importance of not only replicating a finding in one's own lab, but perhaps asking colleagues in other locations to attempt the same study.

The Crabbe study emphasizes the critical importance of having psychologists and other behavioral experts involved in the design and implementation of genetic mutation studies, says Donald Pfaff, PhD, of Rockefeller University.

"If behavioral assays are not done correctly, the variability of your results can be enormous," he says. "Psycho- logists are critical to this work's success because they understand the complexity of behavior and they know how to develop methods to deal with that variability."

Looking for a job in genetics?

Slews of prime university research positions are just waiting for properly trained psychologists.

Graduate students and postdocs at a handful of labs are getting trained to study behavior in mice with one of several types of genetic mutations. Some of the mice have an extra gene added from another species; others have a specific gene "knocked out" so that it's nonfunctional; and others have general mutations to a certain area of their DNA.

And because there's a shortage of behavioral expertise in this area, many are writing their own tickets into some of the best research jobs in the country.

The job of psychologists in these positions, among other things, is to analyze the mice for behavioral changes that might be caused by the gene mutation. If they find a gene that appears to affect some aspect of behavior, the knowledge they glean could help them and their collaborators begin to understand behaviors and traits, such as aggression, mating and anxiety, from molecules on up.

These techniques are proving so powerful, and are receiving so much funding and attention, that research universities and medical schools are ramping up efforts to build mouse- specific behavior laboratories. But they're finding a shortage of behavioral experts to work with the molecular and cellular biologists who create the genetic mouse mutants.

"There are not as many behavioral experts in the field as are needed," says Jacqueline Crawley, PhD, chief of the National Institute of Mental Health's (NIMH) behavioral neuropharmacology intramural lab, which conducts these types of studies.

"There needs to be more training at the graduate student and postdoctoral level in mouse behavioral tasks," she says. "My postdocs are in high demand right now. Not only is it a fascinating set of skills to acquire, but it will increase your job prospects."

Where do people go to get the training they need to work with these genetic mouse models?

Anyone trained in a psychology or behavioral neuroscience program to do experimental work with rats is most of the way there, says Crawley.

"Once you're trained in conducting rat behavioral tasks, it doesn't take too much more training," she says. "The mouse is not that different from the rat, but it is advisable to spend time in a real expert lab because there are several critical differences."

There are a handful of such "expert" labs that study behavior in genetically mutated mice, including Crawley's, Randy Nelson's at Johns Hopkins University, Donald Pfaff's at Rockefeller University and Lisa Gold's at Scripps Research Institute.

The National Institutes of Health has training grants available for students who want experience studying mouse behavior, says NIMH Director Steven Hyman, MD. "People should contact NIMH to find out about training grants," he says, adding that researchers should expect to see more targeted program announcements this fiscal year.

To get the best experience, it's important to find a lab with experts in both behavior and genetics, says Pfaff, who studies instinctive behaviors, such as mating. The best researchers are the ones who can speak the language of brain, behavior and genes, he says.

"Labs are really going after recent grads in behavioral neuroscience--what we once called physiological psychology," says John Crabbe, PhD, who runs a behavioral genetics lab at Oregon Health Sciences University. "Molecularly oriented groups are creating genetically engineered animals right and left and their next question is 'What have I done?' So they're looking to add well-trained postdocs with behavioral expertise. There are a lot of jobs out there."