Physiology& Behavior,Vol. 50, pp. 259-261. ©Pergamon Press plc, 1991. Printed in the U.S.A.
0031-9384/91 $3.00 + .00
BRIEF COMMUNICATION
Early Postnatal Hyperthyroidism Improves Both Working and Reference Memory in a Spatial Radial-Maze Task in Adult Mice WIM E. CRUSIO AND HERBERT SCHWEGLER* Gdndtique, Neurogdndtique et Comportement, URA 1294 CNRS, UFR Biomddicale, Universitd Paris V Rend Descartes, 45 rue des Saints-P~res, 75270 Paris Cedex 06, France *Anatomisches lnstitut, Universitiz't Freiburg, Albertstra~e 17, 7800 Freiburg, FRG
Received 29 May 1990 CRUSIO, W. E. AND H. SCHWEGLER. Early postnatal hyperthyroidism improves both working and reference memory in a spatial radial-maze task in adult mice. PHYSIOL BEHAV 50(1) 259-261, 1991.--Newborn male pups of the inbred mouse strain DBAJZI were injected with a buffered solution of L-thyroxine, known to induce a hyperplasia of the inlJra- and infrapyramidal mossy fiber terminal fields (iip-MF). Because previous studies have shown that the size of the iip-MF correlates positively with both reference memory and working memory in a spatial radial maz, task, we expected improved performance on such a task in the treated animals. Such an improvement was in fact found for both reference and working memory. We conclude that the iip-MF are functionally involved in the regulation of spatial learning and memory in mice. Thyroxine Working memory DBA/2J inbred mice
Reference memory
Spatial learning
NATURALLY occurring variations in hippocampal neuroanatomy covary with behavior in mice and rats (6,14). More specifically, very strong correlations have been found between the sizes of the intra- and infrahippocampal mossy fiber terminal fields (iip-MF) and learning behavior: a negative correlation was obtained with two-way active-avoidance learning (6) and a positive one with spatial radial-maze learning (3,10). Depending on the experimental setup, radial mazes may be used to dissociate different types of memory. First, by supplyhag animals exclusively with either extramaze or intramaze cues, spatial or nonspatial learning (8), respectively, may be tested. Second, following Olton et al. (9), we may define working memory (WM) as concerning trial-specific information and reference memory (RM) as concerning information that is relevant over all trials. Thus, in a radial maze in which only some arms are always reinforced and others never, information about which arms of the maze have already been visited during a specific trial will be stored in an animal's WM, whereas information identifying the arms of the maze that are reinforced will be stored in the RM (7). In a previous experiment, we tested mice from nine different inbred mouse strains in an 8-arm radial maze. All arms of the maze were reinforced, the animals were provided with clear extramaze cues, and a confmement procedure was employed. Thus only spatial WM capability was tested in that task. A strong, negative correlation was obtained between the number of errors
Radial maze
Hippocampus
and the extent of the iip-MF (10). A causal relationship between this terminal field and processes modulating spatial WM was rendered highly probable by the finding that an induced hyperplasia of the iip-MF (see below) was accompanied by increased learning performance in that simple test (11). In a follow-up study in which only a subset of 4 out of 8 arms was reinforced, we were able to show that the sizes of the iip-MF in nine different inbred mouse strains correlate positively with both WM and RM in a spatial arrangement of the radial maze, but not in a nonspatial one: strains with larger iip-MF terminal fields performed better on both the WM and the RM aspects of the spatial task than strains with smaller ones (3). This result suggests that variations in the size of the hippocampal iip-MF underlie the variations in these mnemonic capabilities. A strong test of this hypothesis would be to modify the iipMF distribution during its development, in order to test whether an experimentally induced alteration of this neuroanatomical trait would remain correlated with both WM and RM in adulthood. For such a test we selected the inbred strain DBAJZI (13), which has a small iip-MF terminal field and shows poor radialmaze learning (3,10). Animals from this strain were treated neonatally with thyroxine, because it is known that a transient hyperthyroidism produces an irreversible hyperplasia of the iipMF in mice and rats (5, 6, 11). If our hypothesis is correct, this treatment should also improve the genetically dependent poor WM and RM performance of DBA/ZI inbred mice.
259
260
CRUSIO AND SCHWEGLER METHOD
25]
WORKING MEMORY ERRORS
Animals and Treatment During the first 12 days after birth, a critical period in the development of the mouse hippocampus, male DBA/2J pups were daily injected subcutaneously with 0.05 ml of a buffered solution of L-thyroxine (2 p,g L-thyroxine in 0.15 M Soerensen, pH 9.0, total dose: 24 p,g; n = 6 ) . Controls ( n = 9 ) were injected with an identical volume of carrier solution only. Treated mice showed typical signs of hyperthyroidism such as accelerated development and lower body weight at adulthood (2).
! MEMORY ERROf~
+1
20.
E 5
15.
10.
$ Apparatus and Procedure At the age of 12 weeks, animals were tested for spatial WM and RM capacity using an eight-ann radial maze. The central part measured 22 cm in diameter. The arms (25 cm long, 6 cm high, 6 cm wide) were closed and made of transparent Plexiglas. At the end of each arm some food pellets were deposited behind a perforated wall in order to prevent the mice from smelling the presence or absence of a food reward. For each mouse a specific subset of 4 arms was reinforced by placing a small food pellet (approximately l0 mg) behind a low barrier hindering the animal from seeing whether a specific arm was baited or not. Although the maze was dally rotated by 45 ° , it always remained oriented in space in a similar way. This was done in order to impede animals from solving the maze problem by using possible intramaze cues, such as olfactory markings. Several extramaze cues were provided close to the arms. A confinement procedure was used utilizing transparent guillotine doors at the entrance of each arm. These doors were lowered and kept closed for 5 s after the animal had returned to the center box. This procedure is known to disrupt chaining responses and kinesthetic strategies in rats (1) and mice (10). Forty-eight hours prior to training, the mice were deprived of food, but not of water, and, 24 h apart, received two 10-rain habituation trials with free access to all arms (no rewards present). During training body weight was kept at 80-90% of the pretest body weight. Animals were trained on five consecutive days, one trial per day. Trials were terminated after the animal had eaten all 4 rewards or after a maximum of 15 min, whatever came first. From day 3 onwards, all animals always retrieved the 4 rewards within this time limit. An arm entrance was scored if an animal entered an arm with all 4 paws. Repeated entries of arms belonging to both subsets were counted as WM errors, an RM error was noted if an arm belonging to the subset that was never reinforced was entered. The variables chosen for analysis were the total numbers of WM and RM errors made during the last 3 trials.
Statistics Nonparametric statistics were used throughout. We used MannWhitney U-tests to test the significancy of differences between groups; one-tailed in the cases of the learning variables where improvements were predicted, two-tailed in the cases of other variables. Spearman's rank-correlation coefficient was used to estimate correlations between variables. All tests were corrected for ties (12). RESULTS
As expected, thyroxine-treated animals performed better than controls on both the WM and the RM aspects of the present spatial radial-maze task (Fig. 1). This predicted improvement was significant for both RM (p
0031-9384/91 $3.00 + .00
BRIEF COMMUNICATION
Early Postnatal Hyperthyroidism Improves Both Working and Reference Memory in a Spatial Radial-Maze Task in Adult Mice WIM E. CRUSIO AND HERBERT SCHWEGLER* Gdndtique, Neurogdndtique et Comportement, URA 1294 CNRS, UFR Biomddicale, Universitd Paris V Rend Descartes, 45 rue des Saints-P~res, 75270 Paris Cedex 06, France *Anatomisches lnstitut, Universitiz't Freiburg, Albertstra~e 17, 7800 Freiburg, FRG
Received 29 May 1990 CRUSIO, W. E. AND H. SCHWEGLER. Early postnatal hyperthyroidism improves both working and reference memory in a spatial radial-maze task in adult mice. PHYSIOL BEHAV 50(1) 259-261, 1991.--Newborn male pups of the inbred mouse strain DBAJZI were injected with a buffered solution of L-thyroxine, known to induce a hyperplasia of the inlJra- and infrapyramidal mossy fiber terminal fields (iip-MF). Because previous studies have shown that the size of the iip-MF correlates positively with both reference memory and working memory in a spatial radial maz, task, we expected improved performance on such a task in the treated animals. Such an improvement was in fact found for both reference and working memory. We conclude that the iip-MF are functionally involved in the regulation of spatial learning and memory in mice. Thyroxine Working memory DBA/2J inbred mice
Reference memory
Spatial learning
NATURALLY occurring variations in hippocampal neuroanatomy covary with behavior in mice and rats (6,14). More specifically, very strong correlations have been found between the sizes of the intra- and infrahippocampal mossy fiber terminal fields (iip-MF) and learning behavior: a negative correlation was obtained with two-way active-avoidance learning (6) and a positive one with spatial radial-maze learning (3,10). Depending on the experimental setup, radial mazes may be used to dissociate different types of memory. First, by supplyhag animals exclusively with either extramaze or intramaze cues, spatial or nonspatial learning (8), respectively, may be tested. Second, following Olton et al. (9), we may define working memory (WM) as concerning trial-specific information and reference memory (RM) as concerning information that is relevant over all trials. Thus, in a radial maze in which only some arms are always reinforced and others never, information about which arms of the maze have already been visited during a specific trial will be stored in an animal's WM, whereas information identifying the arms of the maze that are reinforced will be stored in the RM (7). In a previous experiment, we tested mice from nine different inbred mouse strains in an 8-arm radial maze. All arms of the maze were reinforced, the animals were provided with clear extramaze cues, and a confmement procedure was employed. Thus only spatial WM capability was tested in that task. A strong, negative correlation was obtained between the number of errors
Radial maze
Hippocampus
and the extent of the iip-MF (10). A causal relationship between this terminal field and processes modulating spatial WM was rendered highly probable by the finding that an induced hyperplasia of the iip-MF (see below) was accompanied by increased learning performance in that simple test (11). In a follow-up study in which only a subset of 4 out of 8 arms was reinforced, we were able to show that the sizes of the iip-MF in nine different inbred mouse strains correlate positively with both WM and RM in a spatial arrangement of the radial maze, but not in a nonspatial one: strains with larger iip-MF terminal fields performed better on both the WM and the RM aspects of the spatial task than strains with smaller ones (3). This result suggests that variations in the size of the hippocampal iip-MF underlie the variations in these mnemonic capabilities. A strong test of this hypothesis would be to modify the iipMF distribution during its development, in order to test whether an experimentally induced alteration of this neuroanatomical trait would remain correlated with both WM and RM in adulthood. For such a test we selected the inbred strain DBAJZI (13), which has a small iip-MF terminal field and shows poor radialmaze learning (3,10). Animals from this strain were treated neonatally with thyroxine, because it is known that a transient hyperthyroidism produces an irreversible hyperplasia of the iipMF in mice and rats (5, 6, 11). If our hypothesis is correct, this treatment should also improve the genetically dependent poor WM and RM performance of DBA/ZI inbred mice.
259
260
CRUSIO AND SCHWEGLER METHOD
25]
WORKING MEMORY ERRORS
Animals and Treatment During the first 12 days after birth, a critical period in the development of the mouse hippocampus, male DBA/2J pups were daily injected subcutaneously with 0.05 ml of a buffered solution of L-thyroxine (2 p,g L-thyroxine in 0.15 M Soerensen, pH 9.0, total dose: 24 p,g; n = 6 ) . Controls ( n = 9 ) were injected with an identical volume of carrier solution only. Treated mice showed typical signs of hyperthyroidism such as accelerated development and lower body weight at adulthood (2).
! MEMORY ERROf~
+1
20.
E 5
15.
10.
$ Apparatus and Procedure At the age of 12 weeks, animals were tested for spatial WM and RM capacity using an eight-ann radial maze. The central part measured 22 cm in diameter. The arms (25 cm long, 6 cm high, 6 cm wide) were closed and made of transparent Plexiglas. At the end of each arm some food pellets were deposited behind a perforated wall in order to prevent the mice from smelling the presence or absence of a food reward. For each mouse a specific subset of 4 arms was reinforced by placing a small food pellet (approximately l0 mg) behind a low barrier hindering the animal from seeing whether a specific arm was baited or not. Although the maze was dally rotated by 45 ° , it always remained oriented in space in a similar way. This was done in order to impede animals from solving the maze problem by using possible intramaze cues, such as olfactory markings. Several extramaze cues were provided close to the arms. A confinement procedure was used utilizing transparent guillotine doors at the entrance of each arm. These doors were lowered and kept closed for 5 s after the animal had returned to the center box. This procedure is known to disrupt chaining responses and kinesthetic strategies in rats (1) and mice (10). Forty-eight hours prior to training, the mice were deprived of food, but not of water, and, 24 h apart, received two 10-rain habituation trials with free access to all arms (no rewards present). During training body weight was kept at 80-90% of the pretest body weight. Animals were trained on five consecutive days, one trial per day. Trials were terminated after the animal had eaten all 4 rewards or after a maximum of 15 min, whatever came first. From day 3 onwards, all animals always retrieved the 4 rewards within this time limit. An arm entrance was scored if an animal entered an arm with all 4 paws. Repeated entries of arms belonging to both subsets were counted as WM errors, an RM error was noted if an arm belonging to the subset that was never reinforced was entered. The variables chosen for analysis were the total numbers of WM and RM errors made during the last 3 trials.
Statistics Nonparametric statistics were used throughout. We used MannWhitney U-tests to test the significancy of differences between groups; one-tailed in the cases of the learning variables where improvements were predicted, two-tailed in the cases of other variables. Spearman's rank-correlation coefficient was used to estimate correlations between variables. All tests were corrected for ties (12). RESULTS
As expected, thyroxine-treated animals performed better than controls on both the WM and the RM aspects of the present spatial radial-maze task (Fig. 1). This predicted improvement was significant for both RM (p
