Replicability of Drug and Genetic Effects on Behavior

P.I.:  Douglas Wahlsten, Ph.D.
Co-I.:  John Crabbe, Ph.D.

Lab:  Mouse Psychometrics Centre, University of Alberta
         Portland Alcohol Research Center, Oregon Health Sciences University

Location:

Department of Psychology
University of Alberta
Edmonton, Alberta, Canada N2L 3G1

Description:
       One of the greatest challenges in compiling standard data on mouse behavior is the difficulty of obtaining similar results in different laboratories. Failures to replicate results of genetic experiments can arise from several sources. One is the well known genetic background problem (Gerlai, 1996; Crusio and Gerlai, 1999, Crawley, 2000), where effects of a targeted mutation sometimes depend on the genetic background or context in which it occurs. Two others pertain more to the broad domain of psychology - the behavioral test situation and the laboratory environment. The test situation includes the physical test apparatus and the protocol of procedures used to collect the data, whereas the lab environment comprises all those things that impinge on a mouse prior to the start of the behavioral test. In most research, it is not possible to separate these two psychological sources of differences between labs because most labs use apparatus and protocols that are specific or idiosyncratic to each lab. The experiment of Crabbe et al. (1999) addressed this matter by rigorously equating the test apparatus and protocols in three labs and running behavioral tests simultaneously. Despite this exceptional degree of cross-lab control, significant and substantial strain by lab interactions were observed for several behavioral domains. These results indicate that interactions between genetic strain and the lab environment are important sources of discrepancies among labs.
       The present phenotyping project extends our previous works in three directions. (1) We are assessing a larger number of behavioral tests in two specific domains - learning/memory and motor coordination - and conducting a psychometric analysis of the test results in order to devise an optimal array of tests with well documented reliabilities. (2) We are examining a wider spectrum of inbred strains, increasing the sample to 20 strains in the second phase of the project. (3) We are seeking to determine which of the tests give most consistent results across labs when effects of ethanol on behavior are examined. Whereas the previous study sought to equate many aspects of the lab environment, we are relaxing this requirement in the present study and instead hoping to identify those behavioral tests that give the most robust, consistent results despite variations among lab environments, as well as those that are most labile in different conditions.
        In the first phase of the project, 10 tests of learning and memory are being assessed in Edmonton, while 10 tests of motor coordination are assessed in Portland. Once we have done a thorough psychometric analysis of these tests, the best of the lot will be adopted for further work. Apparatus and protocols will be replicated and exchanged between the two sites, and then both sites will run all tests on a set of 20 inbred strains. Genetic covariance structure will be studied in order to refine further the array of tests, and this final array will then be used in experiments on strain-specific effects of alcohol.

Inbred strains:  For the first phase when reliability is being evaluated, the short list of strains recommended by the Strain Database Summit held at the Jackson Labs in May of 1999 will be employed, plus the F1 hybrid B6D2F1/J. These strains are A/J, BALB/cByJ, BTBR (as it becomes available), C3H/HeJ, C57BL/6J, DBA/2J, FVB/NJ, and 129/SvImJ. The wild-derived strain CAST/Ei will also be assessed, although we have reservations because of the notorious difficulties in handling these animals during behavioral tests.

Test protocols:  Apparatus designs and test protocols will be posted on a website as the refined versions become available. We do not intend to post the initial versions that are subject to a variety of changes, although we would be glad to discuss details with interested colleagues.

Selected References:

1. Crabbe, J.C., Wahlsten, D., and Dudek, B.C. (1999). Genetics of mouse behavior: Interactions with lab environment. Science, 284: 1670-1672.

2. Crawley, J. N. (2000). What's Wrong with My Mouse? Behavioral Phenotyping of Transgenic and Knockout Mice. NY: Wiley-Liss.

3. Crusio, W.E., and Gerlai, R. T. (eds.) (1999). Molecular-genetic Techniques for Behavioral Neuroscience. Amsterdam: Elsevier,

4. Gerlai, R. (1996). Gene-targeting studies of mammalian behavior: is it the mutation or the background genotype? Trends in Neuroscienc,. 19:177-181.

5. Wahlsten, D., Crabbe, J, and Dudek, B. (1999) Testing the genetics of behavior in mice. Response. Science, 285: 2069-2070.

6. Wahlsten, D. (1999). Single-gene influences on brain and behavior. Annual Review of Psychology, 50: 599-624.

7. Wahlsten, D. (1999). Experimental design and statistical inference. In W.E. Crusio and R. T. Gerlai (eds.), Molecular-genetic Techniques for Behavioral Neuroscience. Amsterdam: Elsevier, pp. 40-57.

8. Wahlsten, D. (1978). Behavioral genetics and animal learning. In H. Anisman and G. Bignami (Eds.), Psychopharmacology of Aversively Motivated Behaviors. New York: Plenum, pp. 63-118.