Inside the Minds of Mice And

TECHNOLOGY FEATURE INSIDE THE MINDS OF MICE AND MEN Monitoring technologies and genetic engineering are producing a growing array of animal models for psychiatric disorders, but researchers are still learning how best to use them. SAGE LABS SAGE

There is no such thing as an autistic rat, but researchers can study human psychiatric conditions by analysing the behaviour of rodents.

BY MONYA BAKER disorder known as fragile X syndrome arises factors from traumatic life experiences to in when humans lack a working copy of the gene utero conditions also contribute. Even when nyone familiar with Rett syndrome FMR1; mice without the gene show learn- researchers have decided which genes or factors will recognize the symptoms. Mouse ing deficits and hyperactivity similar to the to study, it is not always clear how to assess ani- models — animals that carry genetic symptoms of the human disorder. mal models: how can a researcher use a mouse Amutations similar to those that cause the con- But those are conditions with discrete, recog- to study diseases diagnosed by hallucinations or dition in humans — wring their paws, walk nized causes. Other neurocognitive disorders, an inability to understand figurative language? awkwardly and learn poorly. Other human such as autism, depression and schizophrenia, brain disorders have animal models, too. Mice have multiple and often mysterious causes, so FROM BEHAVIOUR TO BIOLOGY with extra copies of the genome regions dupli- mimicking them is more complicated. Studies Craig Powell, a neuroscientist at the Univer- cated in Down’s syndrome show motor prob- have implicated dozens of genetic variants in sity of Texas Southwestern Medical Center in lems and learning deficits. A developmental producing the disorders, and environmental Dallas, says that the goal of tweaking genes is

usually to uncover disease mechanisms. “So you make a mouse with a mutation that you know causes autism in humans, and you see that it has behaviours that resemble autism,” he says. Powell’s next step is to look at slices of the mouse’s brain, to see how it differs from normal mice. At that point, behaviour can help researchers C57BL/6j DBA/2 A/j PHONOMICS B. SPRUIJT/DELTA to home in on what really matters. “We might find a hundred things wrong with the brain function, but only one or two cause changes in behaviour, so we want to fix things and see what changes the behaviour,” says Powell. In work that has become a touchstone for the field, Mark Bear, a neuroscientist at the Massachusetts Institute of Technology (MIT) in Cambridge, and his colleagues showed that reducing expres- Balb/c 129S1/Sv C3H sion of a particular receptor in mice ameliorated Video tracking of different mouse strains (labelled) reveals that some explore a new cage more than others. the effects of a mutation that causes fragile X syndrome1. Several physiological abnormalities autism and developmental disorders. Tests wrong with their skin: open sores appeared were reversed, and engineered mice had fewer on the offspring showed that the sympto- on their faces. After tests showed nothing seizures and better memories. matic behaviours recurred in only some of the abnormal, video surveillance revealed that In another example, Adrian Bird, a geneticist genetic backgrounds. In another study5, mice the mice groomed themselves excessively, at the Wellcome Trust Centre for Cell Biology from a strain displaying a range of autism- literally rubbing through their fur. Further in Edinburgh, UK, and his colleagues reversed relevant symptoms were reared by and with work showed abnormalities in a region of the symptoms resembling Rett syndrome in mice2. mice from a strain known for high sociability. brain linked to obsessive–compulsive disorder They crafted a version of the defective gene The fostered mice showed no social deficits as (OCD). Although SAPAP3 had not previously that could be restored to normal activity with a adults, but other relevant symptoms, such as been implicated in the condition, drugs that supplement to a mouse’s diet, but administered repetitive grooming, were not reduced. And eased OCD symptoms in humans reduced the the supplement only after mice carrying the when a mutant version of DISC1, the first gene grooming in mice. gene began to exhibit symptoms. Activating to be implicated in schizophrenia, is present Studies8 of proteins the gene at this point was expected to have little in mice whose mother’s immune system has that interact with effect, but the mice showed marked improve- been stressed during pregnancy, the offspring SAPAP3 revealed that ment, raising hopes that recovery might also exhibit symptoms of affective disorders and they, too, had links to be possible in humans. A similar study3 into autism6. Without the environmental stressors, OCD and autism. spinal muscular atrophy found that restor- they show symptoms of schizophrenia. As human genetic ing a defective gene’s function four days after studies reveal gene birth essentially eliminated signs of the disease; HIDDEN SYMPTOMS variants with increas- doing so ten days after birth had little effect. Even the cleverest assays cannot capture some ingly smaller impacts Such studies could help researchers to predict important aspects of human disease, such as the on disease, there is which patients are most likely to benefit in paranoid delusions common in schizophrenia. “When in increasing demand clinical trials. (In fact, because the models will always be doubt, the for new behavioural Research into behaviour can also probe imperfect, most behavioural researchers object experimenter tests to assess them. how genetic variants interact with each other to phrases such as ‘schizophrenic mice’.) Mikhail needs to look at Results of assays are and with environmental factors. In one study Pletnikov, a neurobehaviourologist at Johns what the animal highly variable, and published this year4, wild-type males from five Hopkins University in Baltimore, Maryland, is is doing.” they may not measure mouse strains were bred with females carrying one of many scientists hoping to complement Douglas Wahlsten the most meaning-

C. TOUMA/MAX PLANCK INST. PSYCH. PLANCK INST. C. TOUMA/MAX a mutation that causes symptoms resembling behavioural tests with more readily measured ful symptoms, says biomarkers. People with schizophrenia exhibit Jeffrey Mogil, a neuroscientist at McGill Uni- a wide range of behavioural symptoms, but their versity in Montreal, Canada. “We’ve made a lot lateral ventricles — fluid-filled cavities on either of advances in making ever-fancier mice, but at side of the brain — tend to be larger than aver- the end of the day the question is, what’s your age, so Pletnikov is using brain scans to measure assay and what’s your measure and are they these structures in mice. It is not always nec- relevant?” he says. “The slow link in the chain, essary to see a behavioural change to probe a the messy link in the chain, has always been the diease’s biology, he says. “With humility, you can behavioural assays.” use mice or rats or even worms.” Current tests of animal behaviour are Unexpected behaviours have revealed unan- blunt tools. The Morris water-navigation ticipated biology. Several years ago, Guoping task evaluates an animal’s cognitive ability by Feng, a neuroscientist now at MIT, was trying assessing how it learns to use spatial cues to to work out the function of various proteins swim to an underwater platform that it can’t found on either side of the synapses that con- see. Changes in the animal’s performance can nect neurons. Knocking out one such protein, be used to measure learning and memory. SAPAP3, had no apparent effect on brain func- Tests for anxiety include the open-field test, tion: the mice walked and learned normally7. which measures the time a mouse spends in Less-anxious mice spend longer in open spaces. They did, however, seem to have something enclosed spaces or along the edges of its cage;

7 JULY 2011 | VOL 475 | NATURE | 125 © 2011 Macmillan Publishers Limited. All rights reserved TECHNOLOGY ANIMAL MODELS nervous animals avoid exposed areas. For both, of children without autism playing together psychiatric drugs effective in humans change and the ones with autism being off to the side, the outcomes. playing with a train or a computer,” she recalls. Such tests can be effective at screening new So Crawley designed a task that would assess drugs that act by the same mechanisms as whether a mouse chose to spend time with existing ones. But they are less useful for con- a social partner or an inanimate object. The ditions for which no effective drugs exist, or for assay is now used in many laboratories. gaining insight into pathology. So researchers Clinical tests have inspired other animal are trying to develop tests that capture more- counterparts. Mogil and his colleagues pro- specific components of human disorders. duced the mouse-grimace scale for pain “We talk more to the clinical researchers,” says assessment9, based on a scale that used facial Jacqueline Crawley, chief of behavioural neuro- expressions to determine pain in infants and science at the US National Institute of Mental other humans incapable of speaking. Mogil NORTH CAROLINA GREENSBORO D. WAHLSTEN/UNIV. Health in Bethesda, believes his scale will prove a more reliable The colours of mice and their bedding can confuse Maryland. “There are measure of chronic pain than commonly used video-tracking systems. opportunities for us assays such as the tail-flick test, which meas- to sit down and say, ures how quickly a mouse moves its tail out of mouse models of autism. ‘what do these dis- a beam of light. It should also help researchers Tim Bussey and Lisa Saksida, neuroscientists eases look like, what to design more-humane experiments. Other at the University of Cambridge, UK, have devel- is their variability in scientists have developed a mouse version10 of oped a mouse version of a touchscreen interface the real world, and the Wisconsin card-sorting test, in which par- originally developed for humans and primates. what do you consider ticipants are presented with cards displaying, It uses high-contrast images, tailored to mouse the fundamental core say, three red circles or four blue squares. Once eyesight; and instead of pressing a lever or pok- symptoms?’” humans recognize that rewards come for, say, ing its nose into a hole in the wall as in most Crawley’s own “An animal matching cards by colour, the reward criteria mouse-testing systems, the animal touches a studies of children model may are changed to matching by shape or number. screen with its nose or a paw. The technology, with autism inspired not be 100% The test is used to study disorders including commercialized by Campden Instruments in her to develop a translatable, but autism and schizophrenia. The mouse version Loughborough, UK, last year, could assess many mouse test for anal- maybe 80% is relies on scents such as cinnamon and garlic cognitive abilities in rodents; one battery of tests ogous behaviour. good enough.” alongside textures such as gravel and cotton assesses functions typically impaired in patients “You’ll see a group Jacqueline Crawley balls. Feng is currently evaluating the assay on with schizophrenia, including visual perception,

working memory and pattern-learning11. Bus- sey estimates that the system is now being used in more than 30 labs. Assessing the assays IRONING OUT THE KINKS Confounding variables are the bane of behav- Researchers evaluating animal models complicated. BTBR mice, a strain used to ioural testing, says Douglas Wahlsten, a neuro consider three kinds of validity. study autism, avoid interacting with other scientist at the University of North Carolina at mice and groom themselves excessively. Greensboro and the author of a handbook on Construct validity means that a test When BTBR males are exposed to female the topic. For example, if a mouse seizes up measures what it claims to. In animal urine, they do not vocalize and scent-mark with fear and stands still in the centre of an models, that means that whatever causes as males from other strains do. These traits open chamber, the time it spends there could symptoms in the animal is also what and others map well onto the diagnostic be misinterpreted as demonstrating reduced contributes to disease in humans. Such criteria for autism in humans, which include anxiety. And the walls of water-maze tanks are validity is relatively easy to achieve when deficits in interaction and communication, often so high above the water that mice cannot a condition is caused by a single gene, along with repetitive behaviour. see much of the room, which makes it hard for but most are more complicated. Mikhail them to use spatial cues to find the platform. Pletnikov, a neurobiologist at Johns Hopkins Predictive validity is the extent to which Genetic manipulation increases the scope University in Baltimore, Maryland, models a test predicts a future outcome. In an for artefacts in the data, because tinkering with schizophrenia by combining genes and animal model, the animal should respond to genes could alter how mice perform at tasks for environmental stressors. For complex drugs in a way that corresponds to human reasons that have little to do with the parameters disorders, he says, “we’ve passed that reactions. For example, antidepressants being tested (see ‘Assessing the assays’). If learn- period where we manipulate one gene to try are sometime evaluated by their effects on ing assessments are based on an animals’ ability to understand the whole disease”. the forced-swim test, which measures how to associate a sound with a mild electric shock, long a mouse will try to climb out of a tank for example, researchers should make sure that Face validity means that a test seems to of water before giving up. For disorders the animals have normal hearing and sensitivity measure what it needs to, for example such as autism, however, there are no to pain. The biggest confounding variable may that the symptoms in an animal model effective drugs to serve as positive controls. simply be moving the mouse from its cage to mirror those in a human. For heart rate or Even when drugs do exist, the symptoms the area where tests are performed. “It’s really tumour growth, such measures may be or mechanisms captured by a single rare that a small rodent would be lifted up by straightforward, but for diseases assessed behavioural assay are unlikely to capture another animal and survive,” says Laurence by behaviour, it is considerably more everything that is important. M.B. Tecott, a neurobiologist at the University of California, San Francisco. “We scare the hell out of the animal, then ask it if it’s anxious and into ‘bouts’ of activity and can keep track of a normal weight, but spend only about one-fifth how it can learn,” he says. “We do that routinely.” mouse as it eats, defecates and moves its bed- as much time walking around their cages. The To reduce distress, researchers are working ding. As part of the Mouse Phenome Project, htr2c mutants have normal activity and feed- on ways not just to observe behaviour, but to an international collaboration to collect pheno- ing most of the time, but leave their burrows in do so without physically transporting animals typic data on mouse strains used in labs, Tecott the middle of their resting periods for a series first. In IntelliCage, an observation system from is developing a lifestyle database for 16 strains. of snacks. Without automated analysis, such NewBehavior in Zurich, Switzerland, each Even in preliminary results, strains can be dis- insights into the behavioural components of animal is radiochipped. The system monitors tinguished by their distinct patterns of activity. obesity would be hard to detect. when each animal drinks and eats, and how it Tecott has also used his monitoring system More-expensive video-tracking systems, performs at various stations in its enclosure. on two lines of mice genetically engineered already widely used for many behavioural tests, Tecott, working with colleagues Evan Gould- for obesity: ob/ob mice, which lack the gene can also be used to monitor animals in their ing and Katrin Schenk, has developed an to make one of the hormones that regulates home cages, automatically detecting and catego- inexpensive system that can monitor animals appetite; and htr2c mutant mice, which lack a rizing behaviours. As more molecular biologists around the clock12. A cage sits on a weight- receptor for the neurotransmitter serotonin12. want to monitor genetically engineered mice, detecting platform that measures an animal’s The mice act like couch potatoes and midnight demand for and applications of automated sys- location 50 times a second. Self-correcting snackers respectively, says Tecott. The ob/ob tems are increasing, says Lucas Noldus, chief informatics organize the animal’s movements animals eat just slightly more than mice of executive of Noldus Information Technology in SAGE LABS SAGE

Mice (left) are commonly used as disease models, but rats have more complex and sociable behaviour, so are better suited to modelling neurocognitive disorders.

Video-tracking systems analyse the behaviour of mice in their home cages. Software can capture specific activities, such as grooming and sniffing.

Wageningen, the Netherlands. Noldus describes of converting existing equipment. But even something that has proved particularly diffi- one of his company’s latest systems as “an instru- the best systems require considerable effort to cult with mice. Rats’ larger size also makes it mented home cage that can be configured in a avoid false readings, warns Powell. “It’s hard to easier to take electrophysiological recordings, variety of ways, from a very bare cage to very sum up even in a book what the pitfalls are,” he brain images and tissue samples. rich stimuli. Depending on the cage conforma- says. “Don’t just take the Excel spreadsheet at Paylor predicts that labs working with mice tion you can perform all sorts of tests: an anxi- the end and analyse it. Watch what’s going on will find it difficult to switch to rats, which are ety test with a light spot, or a memory test with during the experiment.” expensive to buy and maintain, needing more automated pellet dispenser.” Even when behaviours can be observed accu- space and different equipment. “We will be sort The monitoring component doesn’t inter- rately, a mouse’s activity may not be robust or of a test case for labs that may want to study fere with the animal, says Vikrant Kobla, vice- subtle enough to reflect the effects of tweaking both mice and rats,” says Shannon Hamilton, president of business development at Clever a gene or the environment. Rats show more a postdoc in Paylor’s laboratory. Sys in Reston, Virginia. “You’re taking the same complex behaviours; for example, littermates But whether testing rats or mice, says cage and putting a camera in front of it.” His wrestle with each other, a behaviour that is Crawley, researchers must remember that the company’s systems recognize more than two considered social play. The most established goal of an animal model is not perfection but dozen behaviours, including head bobbing, behavioural tests were designed for rats, so utility. “An animal model may not be 100% grooming and standing on hind legs, and the many researchers are interested in modelling translatable, but maybe 80% is good enough repertoire is expanding. “We have so many disease using genetically modified rats. Rats to test for possible treatments.” ■ modules we’ve developed that we can adapt it lacking genes implicated in schizophrenia, Par- to deal with new behaviours,” says Kobla. kinson’s disease and autism are among the very Monya Baker is technology editor for Nature first strains being produced by Sigma-Aldrich and Nature Methods. CONSTANT VIGILANCE in Saint Louis, Missouri, which began offering 1. Krueger, D. D. & Bear, M. F. Annu. Rev. Med. 62, “If you want a rich detail of measure from many a suite of ready-made knockout rats earlier this 411–429 (2011). kinds of tests, you really need to use video year. Sigma and other companies also take on 2. Guy, J., Gan, J., Selfridge, J., Cobb, S. & Bird, A. tracking,” says Wahlsten. Nonetheless, he says, custom projects to genetically engineer rats. Science 315, 1143–1147 (2007). 3. Lutz, C. M. et al. J. Clin. Invest. (in the press). accuracy cannot be taken for granted, even for Richard Paylor, an autism researcher at Bay- 4. Spencer, C. M. et al. Autism Res. 4, 40–56 (2011). standard tests of animal behaviour. Tracking lor College of Medicine in Houston, Texas, has 5. Yang, M., Perry, K., Weber, M. D., Katz, A. M. & two animals at once is particularly difficult. It just begun testing on knockout rats, investigat- Crawley, J. N. Autism Res. 4, 17–27 (2011). 6. Abazyan, B. et al. Biol. Psychiatry 68, 1172–1181 is not uncommon for software to confuse an ing behaviours such as how the rats interact (2010). animal’s nose with its tail, for example. Bad and vocalize in social situations, and how they 7. Welch, J. M. et al. Nature 448, 894–900 (2007). lighting wrecks many experiments, particularly respond to social odours. He hopes to report 8. Peça, J. et al. Nature 472, 437–442 (2011). 9. Langford, D. J. et al. Nature Meth. 7, 447–449 if a mouse is moving from a lighter area to a results by the end of the year. It is too early to (2010). darker one, and different artefacts occur with say anything definitive, he says, but rats should 10. Bissonette, G. B. et al. J. Neurosci. 28, 11124– white, brown, black and patchy mice. Infrared allow finer behavioural assessments than mice. 11130 (2008). 11. Bussey, T. J. et al. Neuropharmacology backlighting vastly reduces these problems and In particular, it may be possible to use them doi:10.1016/j.neuropharm.2011.04.011 (2011). is commercially available, but is rarely used, to quantify the effects of potential treatments 12. Goulding, E. H. et al. Proc. Natl Acad. Sci. USA 105, owing to both lack of awareness and the cost for social and communication disorders, 20575–20582 (2008).