What's Wrong With My Mouse?: Behavioral Phenotyping of Transgenic and Knockout Mice

array of choices available for assaying this behavioral construct, and seeing that there are somewhat murky borders between the constructs ‘anxiety’, ‘fear’ and ‘depression’, he or she will explore the question further with an array of tests, after careful consideration of the idiosyncratic parameters required to elicit interpretable data from each test. Such attention to detail matters in real life, even when highly experienced behavioral-testing laboratories are involved. For example, three groups recently independently produced null mutant mice for the corticotropinreleasing hormone 2 receptor. Each group was obviously interested in testing the effects of this gene deletion on anxiety. The results were published as three back-to-back-to-back papers in Nature Genetics 1‐3 . All three groups tested both male and female mice on a version of the elevated plus-maze anxiety test and reported their reluctance to enter open arms of the maze, the accepted index of anxiety. One group found no difference between knockout and wild type of either sex; one group found greater anxiety in knockouts of both sexes; and the third found greater anxiety in male knockouts only. All three also used an open field test, where reluctance to enter the center of the apparatus was the index of anxiety. Again, three different results were reported: one group found a tendency for greater anxiety in null mutants of both sexes; one found significantly greater anxiety in null mutants; and the third found that null mutant males (but not females) spent significantly more time in the center of the field, but interpreted the result to indicate greater anxiety! The different results could derive from many sources, including: differences in animal husbandry and handling before testing; relatively subtle differences in the test apparatus and test protocols; the specific genetic constructs inserted into embryonic stem cells; the stem-cell source strain (three different substrains of 129 inbred mice were used that are known to vary genetically 4