- Email: [email protected]
Motor skills and motor learning in Lurcher mutant mice during aging
Motor learning in Lurcher mice during aging
Pergamon
PII: S0306-4522(00)00509-1
Neuroscience Vol. 102, No. 3, pp. 615±623, 2001 615 q 2001 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306-4522/01 $20.00+0.00
www.elsevier.com/locate/neuroscience
MOTOR SKILLS AND MOTOR LEARNING IN LURCHER MUTANT MICE DURING AGING P. HILBER* and J. CASTON UPRES PSY.CO EA 1780, Laboratoire de Neurobiologie de l'Apprentissage, Universite de Rouen, Faculte des Sciences, 76821 Mont Saint Aignan Cedex, France
AbstractÐMotor learning abilities on the rotorod and motor skills (muscular strength, motor coordination, static and dynamic equilibrium) were investigated in three-, nine-, 15- and 21-month-old Lurcher and control mice. Animals were subjected to motor training on the rotorod before being subjected to motor skills tests. The results showed that control mice exhibited decrease of muscular strength and speci®c equilibrium impairments in static conditions with age, but were still able to learn the motor task on the rotorod even in old age. These results suggest that, in control mice, ef®ciency of the reactive mechanisms, which are sustained by the lower transcerebellar loop (cerebello-rubro-olivo-cerebellar loop), decreased with age, while the ef®ciency of the proactive adjustments, which are sustained by the upper transcerebellar loop (cerebello-thalamo-cortico-ponto-cerebellar loop), did not. In spite of their motor de®cits, Lurcher mutants were able to learn the motor task at three months, but exhibited severe motor learning de®cits as soon as nine months. Such a de®cit seems to be associated with dynamic equilibrium impairments, which also appeared at nine months in these mutants. By two months of age, degeneration of the cerebellar cortex and the olivocerebellar pathway in Lurcher mice has disrupted both lower and upper transcerebellar loops. Disruption of the lower loop could well explain precocious static equilibrium de®cits. However, in spite of disruption of the upper loop, motor learning and dynamic equilibrium were preserved in young mutant mice, suggesting that either deep cerebellar nuclei and/or other motor structures involved in proactive mechanisms needed to maintain dynamic equilibrium and to learn motor tasks, such as the striatopallidal system, are suf®cient. The fact that, in Lurcher mutant mice, motor learning decreased by the age of nine months suggests that the above-mentioned structures are less ef®cient, likely due to degeneration resulting from precocious and focused neurodegeneration of the cerebellar cortex. From this behavioral approach of motor skills and motor learning during aging in Lurcher mutant mice, we postulated the differential involvement of two transcerebellar systems in equilibrium maintenance and motor learning. Moreover, in these mutants, we showed that motor learning abilities decreased with age, suggesting that the precocious degeneration of the cerebellar Purkinje cells had long-term effects on motor structures which are not primarily affected. Thus, from these results, Lurcher mutant mice therefore appear to be a good model to study the pathological evolution of progressive neurodegeneration in the central nervous system during aging. q 2001 IBRO. Published by Elsevier Science Ltd. All rights reserved. Key words: precocious neurodegeneration, aging, cerebellum, motor skills, motor learning, Lurcher mutant mice.
The cerebellum is involved in motor control 56,57 and motor learning (see Ref. 12 for review). Indeed, the latency before falling of rats and mice placed on a rotorod, a test widely used to re¯ect equilibrium and motor learning capabilities, 1,13,31,35 is decreased after cerebellectomy 5,11,59 and after lesion of the olivocerebellar pathway by 3-acetylpyridine. 46 Motor skills and motor learning are also altered in mutant mice exhibiting cerebellar degeneration, such as weaver, 29 staggerer 18,28 and Lurcher. 13,30,32 Heterozygous Lurcher (1/Lc) mutant mice have been widely investigated. They are ataxic 14,43,55 and motor learning, as tested on the rotorod, is delayed and even abolished providing the rotation rate is high. 13 These problems are both due to a precocious degeneration of the Purkinje cells of the cerebellar cortex, which begins at the end of the ®rst postnatal
week and which is almost 100% by two months of age, 7±9,20 and a retrograde degeneration of most granule cells (90%) and of the inferior olivary neurons (60± 75%). 8,9,23,58 The Lurcher mutation is located in the glutamate receptor delta 2 subunit, leading to an increase of the constitutive inward Ca 21 and Na 1 currents, and to apoptosis of the Purkinje cells by excitotoxicity. 60 Although, as mentioned above, the behavior of Lurcher mutant mice has well been investigated, nothing is known about the evolution of motor behavior of these animals during aging. We have demonstrated previously in non-ataxic heterozygous staggerer (1/sg) mutant mice, in which the degeneration process begins late, between three and six months of age, leading to a 30± 40% decrease of Purkinje cells, granule cells and olivary neurons by the age of 12 months, 19,49 a precocious impairment of motor abilities that increased with age but which was not greater than those of control (1/1) mice from 18 months of age. 10 The 1/sg mice may therefore represent a model for genetic contribution of acceleration of the normal aging process of the cerebellum, as proposed previously by Shojaeian-Zanjani et al. 48 The
*Corresponding author. Tel.: 133-2-35-14-67-22; fax: 133-2-35-1463-49. E-mail address: [email protected] (P. Hilber). Abbreviations: 1 /Lc, heterozygous Lurcher mutant mice; 1 /sg, heterozygous staggerer mutant mice; 1 /1, control mice. 615
616
P. Hilber and J. Caston
problem of motor aging in 1/Lc mice is quite different, since degeneration of the cerebellar cortex is precocious and is completed by the age of two months. The aim of the present study was therefore to look for the long-term consequences of this focused and precocious degeneration on the evolution of motor skills and motor learning during aging. EXPERIMENTAL PROCEDURES
Animals Lurcher mutant mice (1/Lc) and normal controls (1/1) of the same strain (B6CBA) were obtained in our laboratory and reared in standard conditions: a 12-h light (08.00±20.00)/12-h dark (20.00±08.00) cycle, at 21±228C, with food and water available ad libitum. Eighty animals were tested, and each age group (three, nine, 15 and 21 months) contained 10 Lurcher mice and 10 controls (gender ratio 1:1). All research was performed under guidelines established by ªle Comite Consultatif National d'Ethique pour les Sciences de la Vie et de la Santeº. All efforts were made to minimize the number of animals used and their suffering.
walking. Grasping consisted of the animals hanging on the rotorod while being passively rotated. During asynchronous walking, the movements of the four limbs was not perfectly coordinated; therefore, the animals stumbled and avoided falling by hanging and being passively rotated, until, at the top of the rotorod, they walked again. During synchronous walking, the mice walked evenly, the four paws being precisely coordinated, and the walking speed was perfectly synchronized with the rotation rate so that the animals walked at the top of the rotorod. The relative proportions of these strategies were calculated for the ®rst three trials and the last three trials of the training session. Hanging This test was designed to evaluate the muscular strength of the animals. The mice were hung by their two forepaws in the middle of a thin string (3 mm diameter and 30 cm length), located 80 cm above the ¯oor, which was covered with a foam carpet to cushion the falls. Their motion was limited by plastic cardboard placed at both ends of the string and in front of it. The latency before falling was measured, the maximal time being ®xed to 600 s. During the time the animals kept hanging, the percentage of time they used one or two paws or three or four paws was also measured. Each animal was subjected to two trials separated by a 20-min interval, during which it was returned to its cage with its congeners.
Experimental design
Wooden beam
In order to evaluate the motor abilities which are involved in achievement of the rotorod task, all the animals were subjected to a battery of tests, measuring their muscular strength (hanging test), equilibrium capabilities in dynamic (wooden beam test) and static (elevated unsteady platform test) conditions, and motor coordination (hole board test). All these motor skills were tested after the training session on the rotorod to determine, in the case of unsuccessful learning, whether it was due to learning disability or to motor skill de®cits. However, because motor training could in¯uence these motor skills, they were also tested, in non-trained mice, in the same conditions (see Discussion). The motor skills tests are described below in chronological order and began, for each mouse, the day after the last day of training on the rotorod. They were not processed randomly but always in the same order for each mouse to allow comparisons between mutants and controls, and between the different age groups. The animals were subjected to one test per day and, after each experiment, each apparatus was cleaned with an alcohol solution (50% ethanol).
The wooden beam test was used to evaluate the equilibrium abilities of the mice when their motion was not limited. The apparatus consisted of a motionless wooden beam with a square section (3 cm £ 3 cm), 1 m in length, placed 80 cm above a foam carpet. At the onset of the single trial, each animal was placed at the middle of the beam, its body axis being perpendicular to the beam long axis. We recorded the latency before falling, ®xed to a maximum of 600 s, the distance covered and the time spent walking. Then, for each animal, the walking speed and the percentage of walking time were calculated.
Rotorod The device was similar to the one used in previous studies. 11,13 It consisted of a horizontal wooden mast 3 cm in diameter and 40 cm in length, covered with sticking plaster in order to increase roughness and situated 25 cm above a landing platform covered with a thick carpet to cushion the eventual fall of the animals. The rod rotated around its longitudinal axis. It was driven by a d.c. electric motor. The rotation speed was ®xed at 20 r.p.m., a speed chosen to allow comparisons with previous studies 10,13 and because preliminary experiments demonstrated that this procedure was perfectly adapted to reveal differences between both mutants and non-mutants, and the different age groups. The animals were subjected daily to two series of ®ve trials, with 30 min between the two consecutive series. The time interval was 5 min between two successive trials. Each trial was performed in the following way: the mouse was placed on the middle of the rotating rod, its body axis being perpendicular to the rotation axis and its head directed against the direction of rotation at the onset of the trial. Between trials, the animals were returned to their cage with their congeners. Quantitative and behavioral data were collected. The latency before falling was measured, a maximal value being ®xed to 300 s for each animal. Training was stopped when the mouse was able to stay on the apparatus during 300 s for three successive trials, or after 15 days of training. For each trial, we recorded the strategy used by the mice to keep balance: grasping, asynchronous walking and synchronous
Hole board This test permitted an evaluation of the motor coordination of the mice. The apparatus consisted of a wooden, square painted box (33 cm £ 33 cm £ 33 cm) containing a ¯oor with 36 holes (hole board). The holes were 1 cm deep, 2 cm in diameter and arranged in a 6 £ 6 array. At the beginning of the single trial, each mouse was placed at the middle of the hole board. We measured, during a 5-min period, the time spent walking and the number of fore- or hindpaw slips into the holes. We then calculated the slip frequency (number of slips per minute of walking). Unstable platform The aim of this test was to evaluate the capabilities of the mice to maintain balance when their displacements were limited. The apparatus consisted of a light circular platform (diameter 8.5 cm, weight 16 g), ®xed at its center on a vertical axis (1 m high) and which could tilt by 308 in either direction. 24 This platform was covered with sticking plaster in order to avoid sliding and grasping. Each animal was subjected to three trials and returned to its cage during the inter-trial intervals (10 min). At the beginning of the trial, the mouse was placed at the middle of the board (horizontal situation). Only motions of the animal could provoke tilting of the platform, and only adapted repartition of muscular strength in the limbs and the body could permit the mouse to restore equilibrium and to maintain balance. We determined the latency before falling, the maximal time being ®xed to 180 s, and the number of slips (when a paw was out of the circumference of the platform). Then, for each trial, we calculated the slip frequency (number of slips per minute). Statistical analysis ANOVA and post hoc comparisons with the Tukey Honestly Signi®cant Difference (HSD) test were used for the purpose of
Motor learning in Lurcher mice during aging
Fig. 1. Rotorod. (A) Evolution of the scores of the animals (in seconds, ^S.E.M.; ordinates), when trained on the rotorod for 15 consecutive days (abscissae). The age of mice (in months) is given on the right. (B) Scores of the mice (in seconds, ^S.E.M.) during the ®rst three trials (open bars) and the last three trials (®lled bars) on the rotorod. Lower histograms: scores of the 1/Lc mice; upper histograms: scores of the 1/1 mice. Abscissa: age of mice (in months). (C) Scores de®cit (in %) of the 1/Lc mice compared to the scores of 1/1 during the ®rst three trials (open bars) and the last three trials (®lled bars) on the rotorod. Abscissa: age of mice (in months). (D) Strategies used (in % of time, ^S.E.M.) on the rotorod during the ®rst three trials as a function of age on the abscissa (in months). Upper histograms: 1/1 mice; lower histograms: 1/Lc mice. G., grasping; S.W., synchronized walking; A.S.W., asynchronized walking. (E) Strategies used (in % of time, ^S.E.M.) on the rotorod during the last three trials as a function of age on the abscissa (in months). Upper histograms: 1/1 mice; lower histograms: 1/Lc mice. G., grasping; S.W., synchronized walking; A.S.W., asynchronized walking.
statistical analysis. In all cases, P , 0.05 was considered statistically signi®cant. RESULTS
Rotorod For each animal, the scores of the 10 daily trials were averaged to obtain the mean score per day. We then p calculated the mean daily score (^S.E.M.: s/ n) achieved by each group of mice. The mean scores of the ®rst three and the last three trials were also calculated in order to ®nd the score de®cits of the 1/Lc mice in comparison to 1/1 mice at the beginning and at the end of the training session. The mean scores are given ^S.E.M. For each group, we also calculated the relative proportion (^S.E.M.) of grasping and walking. Analysis of the scores performed during training (Fig. 1A) by a 2 £ 4 £ 15 ANOVA (two genotypes, four age groups, 15 days of training, with repeated measures on
617
the last factor) revealed signi®cant effects of training (F14,1008 43.076, P , 0.0001), genotype (F1,72 1737.59, P , 0.0001) and age (F3,72 2.163, P , 0.0001). There was a signi®cant age £ genotype £ repeated measures interaction (F42,1008 17, P , 0.0001), which indicates that the evolution of the performances during training differed for the two genotypes as a function of age. Indeed, whereas all the 1/1 mice were able to reach the learning criterion whatever their age, seven 1/Lc mice reached it at three months, only one at nine months, and none at 15 and 21 months. Scores achieved during the ®rst three trials on day 1 measured equilibrium abilities of the animals on the rotorod before training (Fig. 1B). Analysis of these scores revealed signi®cant effects of age (F3,72 43.426, P , 0.0001) and genotype (F1,72 508.823, P , 0.0001), with an age £ genotype interaction (F3,72 3.216, P , 0.0001). The scores achieved by the mutants were lower than those of controls at three, nine and 15 months (P , 0.0001), but not at 21 months (P . 0.05). This lack of difference between the two genotypes at 21 months is probably due to the signi®cant decrease of the scores of 1/1 mice at 15 months (P , 0.0001), whereas no signi®cant decrease of performance with age was observed in Lurcher mice (obviously due to a ¯oor effect). Concerning the scores achieved during the last three trials (Fig. 1B), i.e. at the end of the training session, the ANOVA also revealed effects of age (F3,72 18.23, P , 0.0001) and genotype (F1,72 355.065, P , 0.0001), with a signi®cant interaction between these two factors (F3,72 18.23, P , 0.0001). The scores of 1/Lc mice were always lower than those of controls, whatever the age (P , 0.05 at three months, P , 0.0001 at nine, 15 and 21 months). The performances did not decrease with age in 1/1 mice (they were maximal at all ages), but decreased signi®cantly in the 1/Lc mice as early as nine months of age (P , 0.0001). From Fig. 1C, it is clear that, during the ®rst three trials, the score de®cit of the 1/Lc mice, compared to the score of 1/1 mice, was about 95% at all ages. Since, during these ®rst three trials, the mice were naive (not trained), this de®cit revealed the poor equilibrium abilities of the mutants on the rotorod. It is also clear that, during the last three trials, the de®cit of the 1/Lc mice, compared to the score of the 1/1 mice, was very low at three months (26.2%), much greater at nine months (77.6%), and nearly 100% at 15 and 21 months (Fig. 1C). This indicates that, at three months, intensive training was ef®cient in restoring equilibrium abilities of 1/ Lc mice. This restoration by training was very poor at nine months and impossible thereafter. The strategies used by the 1/Lc and the 1/1 mice to maintain balance were quite different. During the ®rst three trials (Fig. 1D), all 1/Lc mice, whatever their age, maintained balance by grasping (100% of the time), whereas control mice adopted a walking strategy very early, although the frequency of this strategy decreased with age. During the last three trials (Fig. 1E), the percentage of walking in 1/1 mice was greater than during the ®rst three trials (compare Fig. 1E and D) and did not decrease with age. Whatever their age, the
618
P. Hilber and J. Caston
Fig. 2. Hanging. (A) Latency before falling (in seconds, ^S.E.M.). (B) Percentage (^S.E.M.) of use of one or two paws. Abscissae: age of the mice (in months).
1/1 mice were able to adopt synchronized walking at the end of training. The 1/Lc mice maintained balance by walking about 50% of the time at three months; this walking behavior decreased at nine months and completely disappeared thereafter. Therefore, it is obvious that the latencies before falling in 1/1 mice and 1/Lc mice are associated with their walking abilities. These abilities improved with training in the young 1/Lc mice, but not in the older ones. It can therefore be concluded that the motor learning abilities (walking) of the 1/1 mice were preserved during aging, while those of 1/Lc mice decreased very early (nine months) and were null in old animals (15 and 21 months). Hanging For each animal, the scores of the two trials (latency before falling and percentage of time the animals used one/two or three/four paws) were averaged. The mean scores (^S.E.M.) were then determined. Concerning the latencies before falling (Fig. 2A), ANOVA revealed effects of genotype (F1,70 73.917, P , 0.0001) and age (F3,70 18.205, P , 0.0001), with an interaction between these two factors (F3,70 13.942, P , 0.0001). Post hoc comparisons revealed that the scores of Lurcher mutants were signi®cantly lower than those of controls at three and nine months (P 0.0001), but not thereafter. The scores of control mice decreased signi®cantly between three and nine months (P , 0.01), and between nine and 15 months (P , 0.01), whereas there was no signi®cant difference in mutant mice, whatever their age. This ®nal observation could be due to the poor performances of the mutants as early as three months (¯oor effect). Concerning the number of paws used to keep hanging
Fig. 3. Wooden beam. (A) Latency before falling (in seconds, ^S.E.M.). (B) Walking time (in seconds, ^S.E.M.). (C) Proportion of time spent walking (in %, ^S.E.M.). (D) Distance covered (in cm, ^S.E.M.). Abscissae: age of the mice (in months).
(Fig. 2B), ANOVA revealed effects of age (F3,70 5.9897, P , 0.001) and genotype (F1,70 7.5488, P , 0.01), but no interaction between these two factors (F3,70 0.1357, P . 0.05). Wooden beam Analysis of the falling latencies (Fig. 3A) revealed effects of genotype (F1,70 85.461, P , 0.0001) and age (F1,70 16.064, P , 0.0001), with an age £ genotype interaction (F3,70 16.069, P , 0.0001). All the 1/1 mice, whatever their age, were able to stay on the beam for at least 600 s. Only the youngest (three months old) 1/Lc mice achieved this performance, which decreased thereafter. The performances of 1/Lc mice were signi®cantly lower than those of controls at 15 and 21 months (P , 0.0001), but not at three and nine months. Their performances decreased signi®cantly between nine and 15 months (P , 0.05). Concerning the time spent walking (Fig. 3B), ANOVA revealed signi®cant effects of age (F3,70 34.7197, P , 0.0001) and genotype (F1,70 13.5236, P , 0.001), with a signi®cant interaction between these two factors (F3,70 7.8319, P , 0.0001). The walking time was signi®cantly higher in Lurcher mutant mice than in controls only at three months (P , 0.0001). The values decreased signi®cantly in both groups between three and nine months (P , 0.0001 for 1/Lc and P , 0.05 for 1/1). Concerning the percentage of walking on the beam (Fig. 3C), statistical analysis revealed effects of genotype (F1,70 25.676, P , 0.0001) and age (F3,70 14.286, P , 0.0001), but no interaction between these two factors (F3,70 2.215, P . 0.05). The percentage of walking decreased between three and nine months, and was higher in Lurcher mutants, especially at three months (33.6% for the mutants and 15.3% for controls).
Motor learning in Lurcher mice during aging
Fig. 4. Hole board. (A) Time spent walking (in seconds, ^S.E.M.). (B) Number of slips (^S.E.M.). (C) Slip frequency (^S.E.M.). Abscissae: age of the mice (in months).
The distance covered on the beam is shown in Fig. 3D. ANOVA revealed no effect of genotype (F1,70 0.0017, P . 0.05), but a signi®cant effect of age (F3,70 12.6104, P , 0.0001) and an interaction between these two factors (F3,70 4.3562, P , 0.01). The distance covered by the mutants decreased signi®cantly between three and nine months (P , 0.001), whereas statistical analysis did not reveal any signi®cant difference between age groups in 1/1 mice. Hole board Statistical analysis of the time spent walking on the hole board (Fig. 4A) revealed no effect of genotype (F1,70 0.53, P . 0.05), but a signi®cant effect of age (F3,70 4.255, P 0.01). There was no interaction between these two factors (F3,70 0.408, P . 0.05), indicating that, in this test, the decrease with age of the time spent walking was similar in the two genotypes. Whatever the age, the number of slips (Fig. 4B) and the slip frequency (Fig. 4C) were always higher in 1/Lc mice than in controls. For these two variables, ANOVA revealed an effect of genotype (F1,70 103.2, P , 0.0001 for the number of slips and F1,70 269.781, P , 0.0001 for the slip frequency). There was no signi®cant effect of age (F3,70 0.481, P . 0.05 for the number of stumbles and F3,70 2.729, P . 0.05 for the slip frequency) and no signi®cant age £ genotype interaction (F3,70 0.392, P . 0.05 for the number of slips and F3,70 1.84, P . 0.05 for the slip frequency). These results indicated that Lurcher mice exhibited severe impairments in motor coordination and that these de®cits did not increase with age. The absence of a genotype effect concerning the
619
Fig. 5. Unstable board. (A) Latency before falling (in seconds, ^S.E.M.). (B) Slip frequency (^S.E.M.). Abscissae: age of the mice (in months).
time spent walking on the hole board showed that the activity of the mutants on the hole board was similar to that of controls. Unstable platform For each animal, the scores of the three trials (latency before falling and slip frequency) were averaged. The mean scores (^S.E.M.) per group were then calculated. At three months, all the 1/1, but not the 1/Lc, mice were able to stay on the platform for at least 180 s (Fig. 5A). ANOVA revealed signi®cant effects of genotype (F1,70 270.48, P , 0.0001) and age (F3,70 4.51, P , 0.01), but no interaction between these two factors (F3,70 1.06, P . 0.05). The mean scores of the mice decreased with age in the two genotypes. On the other hand, Fig. 5A clearly shows that the scores achieved by the mutants were always lower than those of controls, indicating that, whatever their age, the static equilibrium of 1/Lc mice was greatly impaired. These results are clearly different from those obtained with the wooden beam test, when the Lurcher mice could walk. On the elevated unsteady board, the mice could not walk and the strategies used to stay on the platform were limited: only an adapted repartition of the muscular strength in the limbs and in the body could permit them to stay on the apparatus. Their inability to quickly adapt their body posture could thus explain the static equilibrium impairments (when the motions were reduced). In the Lurcher mutant mice, these equilibrium impairments were obvious from three months of age under static conditions, while they were observed only from nine months of age in dynamic conditions (when motions were permitted). The slip frequency is shown in Fig. 5B. ANOVA
620
P. Hilber and J. Caston
revealed signi®cant effects of genotype (F1,70 196.724, P , 0.0001) and age (F3,70 33.88, P , 0.0001), with an interaction between these two factors (F3,70 22.648, P , 0.0001). The slip frequency was always higher in Lurcher mice than in controls (P , 0.05 at three and nine months, P , 0.0001 at 15 and 21 months). This frequency increased signi®cantly between nine and 15 months and between 15 and 21 months (P , 0.0001) in Lurcher mice, whereas statistical analysis revealed no signi®cant difference between age groups in control mice. DISCUSSION
Motor skills and motor learning in young mice In Lurcher mutant mice, massive degeneration of Purkinje cells in the cerebellar cortex and degeneration of the olivocerebellar pathway are associated with de®cits in muscular strength, motor coordination impairments and equilibrium disturbances. 14,32,33,43,54 However, in spite of these abnormalities due to lack of cerebellar cortex output, postural sensorimotor learning is not abolished. 13,30,31,34 Indeed, speci®c motor training allows the mutants to improve their performance and to reduce their motor de®cits by improving motor coordination. The ef®ciency of motor training can appear very quickly 30 and can be preserved for one week. 34 In the present study, we investigated motor learning on the rotorod with regard to speci®c motor abilities. In order to discover whether the scores of the animals on the rotorod at the end of the training session were due to either learning disabilities or motor skills de®cits, motor skills had to be tested after training. However, given that training could in¯uence motor skills scores, we have tested, in another group of mice, motor abilities before training on the rotorod (unpublished data). The results obtained showed that there was no signi®cant difference between motor skills scores of trained and untrained mice, whatever their age (data not shown). Results of the present study showed that Lurcher mutant mice were able to learn a sensorimotor task and to adopt an adapted walking behavior on the rotorod, which agrees with previous studies. 13,30,31,34 Training on the rotorod, which was ef®cient in improving motor coordination for a speci®c task (rotorod), did not permit the mutant mice to reduce their muscular strength, motor coordination and equilibrium de®cits in the other tests. Moreover, equilibrium de®cits in Lurcher mice were test speci®c. Indeed, when placed on the wooden beam, a test with low spatiotemporal constraints, the mutants did not exhibit any apparent de®cit in terms of latency before falling, but spent more time walking than 1/1 mice. These results have also been observed by Lalonde et al. 32 and Le Marec et al. 34 In our study, the long time spent walking by the mutants on the wooden beam was unlikely due to hyperactivity, because their locomotor activity was similar to that of controls on the hole board. It is more likely a strategy associated with regulation of equilibrium. 14 If so, walking behavior can be understood as a compensatory process which contributes to equilibrium maintenance,
by elaborating body adjustments preceding equilibrium disturbances in order to anticipate the fall (feedforward mechanisms). On the contrary, when Lurcher mice were subjected to an experimental task with higher spatiotemporal constraints (on the rotorod and the unstable platform), their equilibrium de®cits were obvious and they were unable to maintain balance for long. In these tests, maintenance of balance required a quick adaptation of muscular synergies, in response to modi®cations of the support, these reactive motor adjustments being sustained by feedback mechanisms. 24 Our results showed that, in spite of speci®c equilibrium impairments, resulting from alteration of reactive adjustments, Lurcher mutant mice were still able to elaborate and execute proactive adjustments in order to compensate their de®cits. These anticipatory body adjustments, corresponding to walking behavior, appeared when the spatiotemporal constraints were low (i.e. on the wooden beam). They were also observed when the constraints were higher, after a speci®c motor training, since on the rotorod, the mice learned to anticipate the rotation by walking. It therefore appears that the cerebellar cortex is not a crucial structure for the acquisition of a complex motor behavior. The role of the cerebellum in motor regulation can be understood from its implication in two transcerebellar loops, involved respectively in feedback and feedforward mechanisms, termed ªlowerº (cerebello-rubro-olivocerebellar) and ªupperº (cerebello-thalamo-cortico-pontocerebellar) transcerebellar loops (see Ref. 2). These two loops originate in the deep nuclei, but the cerebellar cortex constitutes a key in the relay system between sensory afferents and motor efferents. In Lurcher mutant mice, degeneration of Purkinje cells and the olivocerebellar pathway provokes a disruption of these two loops. Alteration of the reactive adjustments can be explained by disruption of the lower transcerebellar loop. In spite of the disruption of the upper transcerebellar loop, the feedforward mechanisms are still functional in Lurcher mice. This can be explained by the fact that they are relayed by other motor structures, such as the striatopallidal system, as suggested by Le Marec et al.; 34 indeed, its functional integrity in Lurcher mutant mice 44,50 leads us to think that it could be involved in motor learning. An alternative hypothesis is that the cerebellar cortex would not be mainly involved in the anticipatory body adjustments and that the deep cerebellar nuclei would be suf®cient. Indeed, it has been demonstrated that motor learning on the rotorod was abolished in cerebellectomized Lurcher mice, 13 suggesting that the deep cerebellar nuclei are involved in the elaboration and execution of anticipatory body adjustments. In this mutant, the loss of neurons in the cerebellar deep nuclei is moderate, 21,22 and a high cytochrome oxidase activity was observed in all three cerebellar deep nuclei. 50 Taken together, these data suggest that the deep cerebellar nuclei are mainly involved in the automatization of postural adjustments during motor training in Lurcher mutant mice. The role of the deep cerebellar nuclei in the cerebello-thalamocortical pathway, 4,40,41 the in¯uences they receive from the cerebral cortex 39,47 and their connections to the
621
Motor learning in Lurcher mice during aging
thalamus 42,45 suggest that they are involved in the organization of synergic limb movements and are, at least in part, independent of the in¯uence of Purkinje cell input. We can therefore postulate that, in young Lurcher mutant mice, the upper loop was not completely disrupted. It can therefore be stated that the cerebellar cortex is an essential link in the functioning of the lower transcerebellar loop, which is responsible for feedback mechanisms, and is not essential in the functioning of the upper transcerebellar loop, responsible for the elaboration and the maintenance of feedforward mechanisms. The functioning of this loop requires only the integrity of the deep cerebellar nuclei. The neurodegeneration appears very early in heterozygous Lurcher mice 8,20 and it is not excluded that neural plasticity in cerebellar circuitry, in particular between deep cerebellar nuclei and mossy ®bers, is involved in the organization of motor control in these mice. Such a plasticity could also be involved in the functioning of the upper loop and could be the basis of the automatization process expressed through motor learning.
Evolution of motor abilities and motor learning during aging Control mice. In rodents, aging is commonly associated with motor impairments, 10,15,38 which are test and strain dependent. 26 In spite of the progressive decrease of muscular strength with age, all the control mice were able to reach the learning criterion on the rotorod. Motor training was ef®cient whatever the age, and only a few days of training was required to compensate the age-related equilibrium de®cits. As demonstrated by Caston et al. 10 and Ingram et al., 26 the acquisition of motor tasks was not abolished in aged mice, showing that old control mice were still able to elaborate an internal motor program sustaining the organization and control of anticipatory body adjustments. Indeed, we did not detect any alteration of motor coordination in the hole board and equilibrium impairments on the wooden beam (i.e. when the spatiotemporal constraints were low) during aging, although we noticed a decline of equilibrium maintenance on the unstable board and on the rotorod before training (i.e. when the spatiotemporal constraints were high). We therefore postulate that the age-related decrease of equilibrium is associated with progressive dysfunction in structures involved in reactive body adjustments, structures which are included in the lower regulation loop. The neurochemical and structural modi®cations which appear in the cerebellar cortex 6,25,51±53 and in the vestibular system in old mice 51 could explain the progressive loss of the reactive postural adjustments in response to environmental changes. Given that the cerebellar cortex is not crucial for proactive adjustments, the mice were still able to anticipate equilibrium de®cits, whatever their age, by feedforward mechanisms associated with the functionality of the upper transcerebellar loop. Moreover, old control mice were still able to elaborate synchronized walking after training on the rotorod.
Lurcher mice. The aging pro®le of Lurcher mutant mice was different from that of controls. Indeed, we observed a severe alteration of motor learning abilities during aging, associated with impairments in equilibrium on the wooden beam. No further alteration was found in equilibrium on the rotorod before training and on the unstable board in old mice, probably due to the precocious and maximal alterations of these abilities at three months of age (¯oor effect). Disappearance of the walking behavior on the rotorod at the end of training, and decrease of the time spent walking on the wooden beam with age, would re¯ect the progressive inability to produce anticipatory body adjustments, and then motor training becomes progressively inef®cient. The pro®le of old mutant mice on the rotorod therefore resembles that of cerebellectomized Lurcher mice. 13 The inability of aged Lurcher mice to elaborate proactive adjustments can therefore be explained by dysfunctions of key structures implicated in the upper transcerebellar loop with age, such as the deep cerebellar nuclei, which, as suggested above, are involved in feedforward mechanisms. While a massive and precocious degeneration of Purkinje cells provokes only a mild neuronal loss in the deep cerebellar nuclei, a ªpossible evolution of the neurodegeneration process in these structuresº cannot be excluded, as suggested by Heckroth et al. 21 This degeneration or dysfunction in aged animals would result in the inability to elaborate and execute anticipatory body adjustments. However, the deep cerebellar nuclei are not the only structures which are involved in proactive adjustments; other structures integrated in the upper loop and directly or indirectly connected to the cerebellum could be affected (i.e. the thalamic motor nuclei and the striatopallidal system). Indeed, Strazielle et al. 50 found a high level of metabolic activity in several motor nervous structures in Lurcher mice (i.e. in the ventrolateral nuclei of the thalamus and the deep cerebellar nuclei). One can surmise that such high activity could be associated with a considerable production of oxygen free radicals, toxic for the cells. Thus, the accumulation of reactive oxidative components with age would provoke oxidative stress in neurons involved in motor regulation and highly activated in young Lurcher mice. Alteration and/or death of these structures resulting from this stress could explain the age-related motor impairments observed in old Lurcher mice. CONCLUSION 37
1
Marr and Albus have proposed a functional model of the cerebellum in which the cerebellar cortex plays a central role in motor learning. Structural and functional modi®cations in the cerebellar cortex after a learning session agree with this model. 3,16,17,27 Such a role seems to be devoted mainly to the olivocerebellar pathway, 46 which synchronizes the rhythmical activity of Purkinje cells. 36,46 However, studies in Lurcher mutant mice obviously show that lack of the cerebellar cortex and olivocerebellar pathway does not abolish the acquisition of motor tasks provided they are not too dif®cult. 13 Our results suggest that the cerebellar cortex, which is
622
P. Hilber and J. Caston
primarily involved in reactive compensatory adjustments conferring to this part of the cerebellum a facilitating role in motor skills, is not implicated in anticipatory motor regulations which require the integrity of the deep cerebellar nuclei. Therefore, alterations of the dynamic equilibrium and motor learning in aged Lurcher mice would re¯ect dysfunction and/or neuronal loss in nervous structures sustaining feedforward mechanisms, particularly in the deep cerebellar nuclei. The age-related motor skill impairments observed in control mice would result from alteration in nervous structures involved in reactive
adjustments (i.e. the cerebellar cortex). Thus, the evolution of motor status of Lurcher mutant mice is clearly different from that of controls. These behavioral data suggest that precocious and focused neurodegeneration of a speci®c nervous structure can provoke long-term disturbances or neuronal cell loss in nervous system structures usually connected to the cells which have disappeared. Lurcher mutant mice therefore appear to be a good model to study the pathological evolution of progressive neurodegeneration in the central nervous system during aging.
REFERENCES
1. Albus S. S. (1971) A theory of cerebellar function. Math. Biosci. 10, 25±69. 2. Altman J. and Bayer S. A. (1997) Development of the Cerebellar System: In Relation to its Evolution, Structure and Function. CRC, Boca Raton, FL. 3. Anderson B. J., Alcantara A. A. and Greenough W. T. (1996) Motor skills learning: changes in synaptic organization of the rat cerebellar cortex. Neurobiol. Learn. Mem. 66, 221±229. 4. Angault P. and Cicirata F. (1994) Organization anatomo-fonctionnelle de la voie d'in¯uence du neÂocervelet sur le cortex ceÂreÂbral moteur. Rev. Neurol. 150, 39±45. 5. Auvray N., Caston J., Reber A. and Stelz T. (1989) Role of the cerebellum in the ontogenesis of the equilibrium behavior in the young rat: a behavioral study. Brain Res. 505, 291±301. 6. Baurle J. and Grusser-Cornehls U. (1994) Axonal torpedoes in cerebellar Purkinje cells of two normal mouse strains during aging. Acta neuropath., Berlin 88, 237±245. 7. Caddy K. W. T. and Biscoe T. J. (1975) Preliminary observations on the cerebellum in the mutant mouse Lurcher. Brain Res. 91, 276±280. 8. Caddy K. W. T. and Biscoe T. J. (1976) The number of Purkinje cells and olivary neurons in normal and Lurcher mutant mouse. Brain Res. 111, 396±398. 9. Caddy K. W. T. and Biscoe T. J. (1979) Structure and quantitative studies on the normal C3H and Lurcher mutant mouse. Phil. Trans. R. Soc. Lond. 287, 167±201. 10. Caston J., Delhaye-Bouchaud N. and Mariani J. (1996) Motor behavior of heterozygous staggerer mutant (1/sg) versus normal (1/1) mice during aging. Behav. Brain Res. 72, 97±102. 11. Caston J., Jones N. and Stelz T. (1995) Role of preoperative and postoperative sensorimotor training on restoration of the equilibrium behavior in adult mice following cerebellectomy. Neurobiol. Learn. Mem. 64, 195±202. 12. Caston J., Lalonde R., Delhaye-Bouchaud N. and Mariani J. (1998) The cerebellum and postural sensorimotor learning in mice and rats. Behav. Brain Res. 95, 17±22. 13. Caston J., Vasseur F., Stelz T., Chianale C., Delhaye-Bouchaud N. and Mariani J. (1995) Differential roles of cerebellar cortex and deep cerebellar nuclei in the learning of the equilibrium behavior: studies in intact and cerebellectomized Lurcher mutant mice. Devl Brain Res. 86, 311±316. 14. Fortier P. A., Smith A. M. and Rossignol S. (1987) Locomotor de®cits in the mutant mouse, Lurcher. Expl Brain Res. 66, 271±286. 15. Gage F. H., Dunnett S. B. and BjoÈrklund A. (1984) Spatial learning and motor de®cits in aged rats. Neurobiol. Aging 5, 43±48. 16. Gilbert P. F. (1974) A theory of memory that explains the function and structure of the cerebellum. Brain Res. 70, 1±18. 17. Gilbert P. F. and Thach W. T. (1977) Purkinje cell activity during motor learning. Brain Res. 128, 309±328. 18. Guastavino J. M. (1988) EthogeÂneÁse de la souris mutante staggerer. Facteurs eÂpigeÂneÂtiques et reÂcupeÂrations fonctionnelles. TheÁse Doctorat d'Etat, Universite Paris-Nord, France. 19. Hadj-Sharaoui N., Frederic F., Zanjani H., Herrup K., Delhaye-Bouchaud N. and Mariani J. (1997) Purkinje cell loss in heterozygous staggerer mutant mice during aging. Devl Brain Res. 98, 1±8. 20. Heckroth J. A. (1992) Development of glutamic acid decarboxylase-immunoreactive elements in the cerebellar cortex of normal and Lurcher mutant mice. J. comp. Neurol. 315, 85±97. 21. Heckroth J. A. (1994) A quantitative morphological analysis of the cerebellar nuclei in normal and Lurcher mutant mice. II. Volumetric changes in cytological components. J. comp. Neurol. 343, 183±192. 22. Heckroth J. A. (1994) Quantitative morphological analysis of the cerebellar nuclei in normal and Lurcher mutant mice. I. Morphology and cell number. J. comp. Neurol. 343, 173±182. 23. Heckroth J. A. and Eisenmann L. M. (1991) Olivary morphology and olivocerebellar topography in adult Lurcher mutant mice. J. comp. Neurol. 312, 641±651. 24. Hilber P., Lalonde R. and Caston J. (1999) An unsteady platform test for measuring static equilibrium in mice. J. Neurosci. Meth. 88, 201±205. 25. Hussain S., Slikker W. Jr and Ali S. F. (1995) Age-related changes in antioxidant enzymes, superoxide dismutase, catalase, glutathione peroxidase and glutathione in different regions of mouse brain. Int. J. devl Neurosci. 13, 811±817. 26. Ingram D. K., London E. D., Reynolds M. A., Waller S. B. and Goodrick C. L. (1981) Differential effects of age on motor performance in two mouse strains. Neurobiol. Aging 2, 221±227. 27. Kleim J. A., Vij K., Ballard D. H. and Greenough W. T. (1997) Learning-dependent synaptic modi®cations in the cerebellar cortex of the adult rat persist for at least four weeks. J. Neurosci. 17, 717±721. 28. Lalonde R. (1987) Motor abnormalities in staggerer mutant mice. Expl Brain Res. 68, 417±420. 29. Lalonde R. (1987) Motor abnormalities in weaver mutant mice. Expl Brain Res. 65, 479±481. 30. Lalonde R. (1994) Motor learning in Lurcher mutant mice. Brain Res. 639, 351±353. 31. Lalonde R., Bensoula A. N. and Filali M. (1995) Rotorod sensorimotor learning in cerebellar mutant mice. Neurosci. Res. 22, 423±426. 32. Lalonde R., Botez M. I., Joyal C. C. and Caumartin M. (1992) Motor abnormalities in Lurcher mutant mice. Physiol. Behav. 51, 523±525. 33. Lalonde R., Joyal C. C., Guastavino J.-M. and Botez M. I. (1993) Hole poking and motor coordination in Lurcher mutant mice. Physiol. Behav. 54, 41±44. 34. Le Marec N., Caston J. and Lalonde R. (1997) Impaired motor skills on static and mobile beams in Lurcher mutant mice. Expl Brain Res. 768, 310±316.
Motor learning in Lurcher mice during aging
623
35. Le Marec N., Stelz T., Delhaye-Bouchaud N., Mariani J. and Caston J. (1997) Effects of cerebellar granule cell depletion on learning of the equilibrium behaviour: study in postnatally X-irradiated rats. Eur. J. Neurosci. 9, 2472±2478. 36. Llinas R. and Welsh J. (1993) On the cerebellum and motor learning. Curr. Opin. Neurobiol. 3, 958±965. 37. Marr D. A. (1969) A theory of cerebellar cortex. J. Physiol., Lond. 202, 437±470. 38. Marshall J. F. (1982) Sensorimotor disturbances in the aging rodent. J. Geront. 37, 548±554. 39. Massion J. (1973) Intervention des voies ceÂreÂbello-corticales et cortico-ceÂreÂbelleuses dans l'organisation et la reÂgulation du mouvement. J. Physiol., Paris 67, 117A±170A. 40. Massion J. (1976) The thalamus in the motor system. Appl. Neurophysiol. 39, 222±238. 41. Massion J. and Rispal-Padel L. (1972) Spatial organization of the cerebello-thalamo-cortical pathway. Brain Res. 40, 61±65. 42. McCormick D. A. (1995) The cerebellar symphony. Nature 374, 412±413. 43. Phillips R. J. S. (1960) Lurcher, a new gene in linkage group XI of house mouse. J. Genet. 57, 35±42. 44. Reader T. A., Strazielle C., Botez M. I. and Lalonde R. (1998) Brain dopamine and amino acid concentrations in Lurcher mutant mice. Brain Res. Bull. 45, 489±493. 45. Rispal-Padel L. (1993) Contribution des effeÂrences ceÂreÂbelleuses aÁ l'organisation des synergies motrices. Rev. Neurol., Paris 149, 716±727. 46. Rondi-Reig L., Delhaye-Bouchaud N., Mariani J. and Caston J. (1997) Role of the inferior olivary complex in motor skills and motor learning in the adult rat. Neuroscience 77, 955±963. 47. Shinoda Y., Sugiuchi Y. and Futami T. (1993) Organization of excitatory inputs from the cerebral cortex to the cerebellar dentate nucleus. Can. J. neurol. Sci. 20, Suppl. 3, S19±S28. 48. Shojaeian-Zanjani H., Mariani J., Delhaye-Bouchaud N. and Herrup K. (1992) Neuronal cell loss in heterozygous staggerer mutant mice: a model for genetic contributions to the aging process. Devl Brain Res. 67, 153±160. 49. Shojaeian-Zanjani H., Mariani J. and Herrup K. (1990) Cell loss in the inferior olive of the staggerer mutant mouse is an indirect effect of the gene. J. Neurogenet. 6, 229±241. 50. Strazielle C., Kremarik P., Ghersi-Egea J. F. and Lalonde R. (1998) Regional brain variations of cytochrome oxidase activity and motor coordination in Lurcher mutant mice. Expl Brain Res. 121, 35±45. 51. Sturrock R. R. (1989) Age-related changes in Purkinje cell number in the cerebellar nodulus of the mouse. J. Hirnforsch. 30, 757±760. 52. Sturrock R. R. (1989) Changes in neuron number in the cerebellar cortex of the ageing mouse. J. Hirnforsch. 30, 499±503. 53. Sturrock R. R. (1990) A comparison of quantitative histological changes in different regions of the ageing mouse cerebellum. J. Hirnforsch. 31, 481±486. 54. Thifault S., Girouard N. and Lalonde R. (1996) Climbing sensorimotor skills in lurcher mutant mice. Brain Res. Bull. 41, 385±390. 55. Thullier F., Lalonde R., Cousin X. and Lestienne F. (1997) Neurobehavioral evaluation of Lurcher mutant mice during ontogeny. Devl Brain Res. 100, 22±28. 56. Welsh J. P. and Harvey J. A. (1992) The role of the cerebellum in voluntary and re¯exive movements: history and current status. In The Cerebellum Revisited (eds Llinas R. and Sotelo C.), pp. 301±334. Springer, New York. 57. Welsh J. P., Lang E. J., Sugihara I. and Llinas R. (1995) Dynamic organization of motor control with the olivocerebellar system. Nature 374, 453±457. 58. Wilson D. B. (1976) Histological de®cits in the cerebellum of adult Lurcher (Lc) mice. J. Neuropath. exp. Neurol. 35, 40±45. 59. Zion C., Auvray N., Caston J., Reber A. and Stelz T. (1990) Effects of cerebellectomy at day 15 on the ontogenesis of the equilibrium behavior in the rat. Brain Res. 515, 104±110. 60. Zuo J., De Jager P. L., Takahashi K. A., Jiang W., Linden D. and Heintz N. (1997) Neurodegeneration in Lurcher mice caused by mutation in delta 2 glutamate receptor. Nature 388, 769±773. (Accepted 18 October 2000)
Pergamon
PII: S0306-4522(00)00509-1
Neuroscience Vol. 102, No. 3, pp. 615±623, 2001 615 q 2001 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306-4522/01 $20.00+0.00
www.elsevier.com/locate/neuroscience
MOTOR SKILLS AND MOTOR LEARNING IN LURCHER MUTANT MICE DURING AGING P. HILBER* and J. CASTON UPRES PSY.CO EA 1780, Laboratoire de Neurobiologie de l'Apprentissage, Universite de Rouen, Faculte des Sciences, 76821 Mont Saint Aignan Cedex, France
AbstractÐMotor learning abilities on the rotorod and motor skills (muscular strength, motor coordination, static and dynamic equilibrium) were investigated in three-, nine-, 15- and 21-month-old Lurcher and control mice. Animals were subjected to motor training on the rotorod before being subjected to motor skills tests. The results showed that control mice exhibited decrease of muscular strength and speci®c equilibrium impairments in static conditions with age, but were still able to learn the motor task on the rotorod even in old age. These results suggest that, in control mice, ef®ciency of the reactive mechanisms, which are sustained by the lower transcerebellar loop (cerebello-rubro-olivo-cerebellar loop), decreased with age, while the ef®ciency of the proactive adjustments, which are sustained by the upper transcerebellar loop (cerebello-thalamo-cortico-ponto-cerebellar loop), did not. In spite of their motor de®cits, Lurcher mutants were able to learn the motor task at three months, but exhibited severe motor learning de®cits as soon as nine months. Such a de®cit seems to be associated with dynamic equilibrium impairments, which also appeared at nine months in these mutants. By two months of age, degeneration of the cerebellar cortex and the olivocerebellar pathway in Lurcher mice has disrupted both lower and upper transcerebellar loops. Disruption of the lower loop could well explain precocious static equilibrium de®cits. However, in spite of disruption of the upper loop, motor learning and dynamic equilibrium were preserved in young mutant mice, suggesting that either deep cerebellar nuclei and/or other motor structures involved in proactive mechanisms needed to maintain dynamic equilibrium and to learn motor tasks, such as the striatopallidal system, are suf®cient. The fact that, in Lurcher mutant mice, motor learning decreased by the age of nine months suggests that the above-mentioned structures are less ef®cient, likely due to degeneration resulting from precocious and focused neurodegeneration of the cerebellar cortex. From this behavioral approach of motor skills and motor learning during aging in Lurcher mutant mice, we postulated the differential involvement of two transcerebellar systems in equilibrium maintenance and motor learning. Moreover, in these mutants, we showed that motor learning abilities decreased with age, suggesting that the precocious degeneration of the cerebellar Purkinje cells had long-term effects on motor structures which are not primarily affected. Thus, from these results, Lurcher mutant mice therefore appear to be a good model to study the pathological evolution of progressive neurodegeneration in the central nervous system during aging. q 2001 IBRO. Published by Elsevier Science Ltd. All rights reserved. Key words: precocious neurodegeneration, aging, cerebellum, motor skills, motor learning, Lurcher mutant mice.
The cerebellum is involved in motor control 56,57 and motor learning (see Ref. 12 for review). Indeed, the latency before falling of rats and mice placed on a rotorod, a test widely used to re¯ect equilibrium and motor learning capabilities, 1,13,31,35 is decreased after cerebellectomy 5,11,59 and after lesion of the olivocerebellar pathway by 3-acetylpyridine. 46 Motor skills and motor learning are also altered in mutant mice exhibiting cerebellar degeneration, such as weaver, 29 staggerer 18,28 and Lurcher. 13,30,32 Heterozygous Lurcher (1/Lc) mutant mice have been widely investigated. They are ataxic 14,43,55 and motor learning, as tested on the rotorod, is delayed and even abolished providing the rotation rate is high. 13 These problems are both due to a precocious degeneration of the Purkinje cells of the cerebellar cortex, which begins at the end of the ®rst postnatal
week and which is almost 100% by two months of age, 7±9,20 and a retrograde degeneration of most granule cells (90%) and of the inferior olivary neurons (60± 75%). 8,9,23,58 The Lurcher mutation is located in the glutamate receptor delta 2 subunit, leading to an increase of the constitutive inward Ca 21 and Na 1 currents, and to apoptosis of the Purkinje cells by excitotoxicity. 60 Although, as mentioned above, the behavior of Lurcher mutant mice has well been investigated, nothing is known about the evolution of motor behavior of these animals during aging. We have demonstrated previously in non-ataxic heterozygous staggerer (1/sg) mutant mice, in which the degeneration process begins late, between three and six months of age, leading to a 30± 40% decrease of Purkinje cells, granule cells and olivary neurons by the age of 12 months, 19,49 a precocious impairment of motor abilities that increased with age but which was not greater than those of control (1/1) mice from 18 months of age. 10 The 1/sg mice may therefore represent a model for genetic contribution of acceleration of the normal aging process of the cerebellum, as proposed previously by Shojaeian-Zanjani et al. 48 The
*Corresponding author. Tel.: 133-2-35-14-67-22; fax: 133-2-35-1463-49. E-mail address: [email protected] (P. Hilber). Abbreviations: 1 /Lc, heterozygous Lurcher mutant mice; 1 /sg, heterozygous staggerer mutant mice; 1 /1, control mice. 615
616
P. Hilber and J. Caston
problem of motor aging in 1/Lc mice is quite different, since degeneration of the cerebellar cortex is precocious and is completed by the age of two months. The aim of the present study was therefore to look for the long-term consequences of this focused and precocious degeneration on the evolution of motor skills and motor learning during aging. EXPERIMENTAL PROCEDURES
Animals Lurcher mutant mice (1/Lc) and normal controls (1/1) of the same strain (B6CBA) were obtained in our laboratory and reared in standard conditions: a 12-h light (08.00±20.00)/12-h dark (20.00±08.00) cycle, at 21±228C, with food and water available ad libitum. Eighty animals were tested, and each age group (three, nine, 15 and 21 months) contained 10 Lurcher mice and 10 controls (gender ratio 1:1). All research was performed under guidelines established by ªle Comite Consultatif National d'Ethique pour les Sciences de la Vie et de la Santeº. All efforts were made to minimize the number of animals used and their suffering.
walking. Grasping consisted of the animals hanging on the rotorod while being passively rotated. During asynchronous walking, the movements of the four limbs was not perfectly coordinated; therefore, the animals stumbled and avoided falling by hanging and being passively rotated, until, at the top of the rotorod, they walked again. During synchronous walking, the mice walked evenly, the four paws being precisely coordinated, and the walking speed was perfectly synchronized with the rotation rate so that the animals walked at the top of the rotorod. The relative proportions of these strategies were calculated for the ®rst three trials and the last three trials of the training session. Hanging This test was designed to evaluate the muscular strength of the animals. The mice were hung by their two forepaws in the middle of a thin string (3 mm diameter and 30 cm length), located 80 cm above the ¯oor, which was covered with a foam carpet to cushion the falls. Their motion was limited by plastic cardboard placed at both ends of the string and in front of it. The latency before falling was measured, the maximal time being ®xed to 600 s. During the time the animals kept hanging, the percentage of time they used one or two paws or three or four paws was also measured. Each animal was subjected to two trials separated by a 20-min interval, during which it was returned to its cage with its congeners.
Experimental design
Wooden beam
In order to evaluate the motor abilities which are involved in achievement of the rotorod task, all the animals were subjected to a battery of tests, measuring their muscular strength (hanging test), equilibrium capabilities in dynamic (wooden beam test) and static (elevated unsteady platform test) conditions, and motor coordination (hole board test). All these motor skills were tested after the training session on the rotorod to determine, in the case of unsuccessful learning, whether it was due to learning disability or to motor skill de®cits. However, because motor training could in¯uence these motor skills, they were also tested, in non-trained mice, in the same conditions (see Discussion). The motor skills tests are described below in chronological order and began, for each mouse, the day after the last day of training on the rotorod. They were not processed randomly but always in the same order for each mouse to allow comparisons between mutants and controls, and between the different age groups. The animals were subjected to one test per day and, after each experiment, each apparatus was cleaned with an alcohol solution (50% ethanol).
The wooden beam test was used to evaluate the equilibrium abilities of the mice when their motion was not limited. The apparatus consisted of a motionless wooden beam with a square section (3 cm £ 3 cm), 1 m in length, placed 80 cm above a foam carpet. At the onset of the single trial, each animal was placed at the middle of the beam, its body axis being perpendicular to the beam long axis. We recorded the latency before falling, ®xed to a maximum of 600 s, the distance covered and the time spent walking. Then, for each animal, the walking speed and the percentage of walking time were calculated.
Rotorod The device was similar to the one used in previous studies. 11,13 It consisted of a horizontal wooden mast 3 cm in diameter and 40 cm in length, covered with sticking plaster in order to increase roughness and situated 25 cm above a landing platform covered with a thick carpet to cushion the eventual fall of the animals. The rod rotated around its longitudinal axis. It was driven by a d.c. electric motor. The rotation speed was ®xed at 20 r.p.m., a speed chosen to allow comparisons with previous studies 10,13 and because preliminary experiments demonstrated that this procedure was perfectly adapted to reveal differences between both mutants and non-mutants, and the different age groups. The animals were subjected daily to two series of ®ve trials, with 30 min between the two consecutive series. The time interval was 5 min between two successive trials. Each trial was performed in the following way: the mouse was placed on the middle of the rotating rod, its body axis being perpendicular to the rotation axis and its head directed against the direction of rotation at the onset of the trial. Between trials, the animals were returned to their cage with their congeners. Quantitative and behavioral data were collected. The latency before falling was measured, a maximal value being ®xed to 300 s for each animal. Training was stopped when the mouse was able to stay on the apparatus during 300 s for three successive trials, or after 15 days of training. For each trial, we recorded the strategy used by the mice to keep balance: grasping, asynchronous walking and synchronous
Hole board This test permitted an evaluation of the motor coordination of the mice. The apparatus consisted of a wooden, square painted box (33 cm £ 33 cm £ 33 cm) containing a ¯oor with 36 holes (hole board). The holes were 1 cm deep, 2 cm in diameter and arranged in a 6 £ 6 array. At the beginning of the single trial, each mouse was placed at the middle of the hole board. We measured, during a 5-min period, the time spent walking and the number of fore- or hindpaw slips into the holes. We then calculated the slip frequency (number of slips per minute of walking). Unstable platform The aim of this test was to evaluate the capabilities of the mice to maintain balance when their displacements were limited. The apparatus consisted of a light circular platform (diameter 8.5 cm, weight 16 g), ®xed at its center on a vertical axis (1 m high) and which could tilt by 308 in either direction. 24 This platform was covered with sticking plaster in order to avoid sliding and grasping. Each animal was subjected to three trials and returned to its cage during the inter-trial intervals (10 min). At the beginning of the trial, the mouse was placed at the middle of the board (horizontal situation). Only motions of the animal could provoke tilting of the platform, and only adapted repartition of muscular strength in the limbs and the body could permit the mouse to restore equilibrium and to maintain balance. We determined the latency before falling, the maximal time being ®xed to 180 s, and the number of slips (when a paw was out of the circumference of the platform). Then, for each trial, we calculated the slip frequency (number of slips per minute). Statistical analysis ANOVA and post hoc comparisons with the Tukey Honestly Signi®cant Difference (HSD) test were used for the purpose of
Motor learning in Lurcher mice during aging
Fig. 1. Rotorod. (A) Evolution of the scores of the animals (in seconds, ^S.E.M.; ordinates), when trained on the rotorod for 15 consecutive days (abscissae). The age of mice (in months) is given on the right. (B) Scores of the mice (in seconds, ^S.E.M.) during the ®rst three trials (open bars) and the last three trials (®lled bars) on the rotorod. Lower histograms: scores of the 1/Lc mice; upper histograms: scores of the 1/1 mice. Abscissa: age of mice (in months). (C) Scores de®cit (in %) of the 1/Lc mice compared to the scores of 1/1 during the ®rst three trials (open bars) and the last three trials (®lled bars) on the rotorod. Abscissa: age of mice (in months). (D) Strategies used (in % of time, ^S.E.M.) on the rotorod during the ®rst three trials as a function of age on the abscissa (in months). Upper histograms: 1/1 mice; lower histograms: 1/Lc mice. G., grasping; S.W., synchronized walking; A.S.W., asynchronized walking. (E) Strategies used (in % of time, ^S.E.M.) on the rotorod during the last three trials as a function of age on the abscissa (in months). Upper histograms: 1/1 mice; lower histograms: 1/Lc mice. G., grasping; S.W., synchronized walking; A.S.W., asynchronized walking.
statistical analysis. In all cases, P , 0.05 was considered statistically signi®cant. RESULTS
Rotorod For each animal, the scores of the 10 daily trials were averaged to obtain the mean score per day. We then p calculated the mean daily score (^S.E.M.: s/ n) achieved by each group of mice. The mean scores of the ®rst three and the last three trials were also calculated in order to ®nd the score de®cits of the 1/Lc mice in comparison to 1/1 mice at the beginning and at the end of the training session. The mean scores are given ^S.E.M. For each group, we also calculated the relative proportion (^S.E.M.) of grasping and walking. Analysis of the scores performed during training (Fig. 1A) by a 2 £ 4 £ 15 ANOVA (two genotypes, four age groups, 15 days of training, with repeated measures on
617
the last factor) revealed signi®cant effects of training (F14,1008 43.076, P , 0.0001), genotype (F1,72 1737.59, P , 0.0001) and age (F3,72 2.163, P , 0.0001). There was a signi®cant age £ genotype £ repeated measures interaction (F42,1008 17, P , 0.0001), which indicates that the evolution of the performances during training differed for the two genotypes as a function of age. Indeed, whereas all the 1/1 mice were able to reach the learning criterion whatever their age, seven 1/Lc mice reached it at three months, only one at nine months, and none at 15 and 21 months. Scores achieved during the ®rst three trials on day 1 measured equilibrium abilities of the animals on the rotorod before training (Fig. 1B). Analysis of these scores revealed signi®cant effects of age (F3,72 43.426, P , 0.0001) and genotype (F1,72 508.823, P , 0.0001), with an age £ genotype interaction (F3,72 3.216, P , 0.0001). The scores achieved by the mutants were lower than those of controls at three, nine and 15 months (P , 0.0001), but not at 21 months (P . 0.05). This lack of difference between the two genotypes at 21 months is probably due to the signi®cant decrease of the scores of 1/1 mice at 15 months (P , 0.0001), whereas no signi®cant decrease of performance with age was observed in Lurcher mice (obviously due to a ¯oor effect). Concerning the scores achieved during the last three trials (Fig. 1B), i.e. at the end of the training session, the ANOVA also revealed effects of age (F3,72 18.23, P , 0.0001) and genotype (F1,72 355.065, P , 0.0001), with a signi®cant interaction between these two factors (F3,72 18.23, P , 0.0001). The scores of 1/Lc mice were always lower than those of controls, whatever the age (P , 0.05 at three months, P , 0.0001 at nine, 15 and 21 months). The performances did not decrease with age in 1/1 mice (they were maximal at all ages), but decreased signi®cantly in the 1/Lc mice as early as nine months of age (P , 0.0001). From Fig. 1C, it is clear that, during the ®rst three trials, the score de®cit of the 1/Lc mice, compared to the score of 1/1 mice, was about 95% at all ages. Since, during these ®rst three trials, the mice were naive (not trained), this de®cit revealed the poor equilibrium abilities of the mutants on the rotorod. It is also clear that, during the last three trials, the de®cit of the 1/Lc mice, compared to the score of the 1/1 mice, was very low at three months (26.2%), much greater at nine months (77.6%), and nearly 100% at 15 and 21 months (Fig. 1C). This indicates that, at three months, intensive training was ef®cient in restoring equilibrium abilities of 1/ Lc mice. This restoration by training was very poor at nine months and impossible thereafter. The strategies used by the 1/Lc and the 1/1 mice to maintain balance were quite different. During the ®rst three trials (Fig. 1D), all 1/Lc mice, whatever their age, maintained balance by grasping (100% of the time), whereas control mice adopted a walking strategy very early, although the frequency of this strategy decreased with age. During the last three trials (Fig. 1E), the percentage of walking in 1/1 mice was greater than during the ®rst three trials (compare Fig. 1E and D) and did not decrease with age. Whatever their age, the
618
P. Hilber and J. Caston
Fig. 2. Hanging. (A) Latency before falling (in seconds, ^S.E.M.). (B) Percentage (^S.E.M.) of use of one or two paws. Abscissae: age of the mice (in months).
1/1 mice were able to adopt synchronized walking at the end of training. The 1/Lc mice maintained balance by walking about 50% of the time at three months; this walking behavior decreased at nine months and completely disappeared thereafter. Therefore, it is obvious that the latencies before falling in 1/1 mice and 1/Lc mice are associated with their walking abilities. These abilities improved with training in the young 1/Lc mice, but not in the older ones. It can therefore be concluded that the motor learning abilities (walking) of the 1/1 mice were preserved during aging, while those of 1/Lc mice decreased very early (nine months) and were null in old animals (15 and 21 months). Hanging For each animal, the scores of the two trials (latency before falling and percentage of time the animals used one/two or three/four paws) were averaged. The mean scores (^S.E.M.) were then determined. Concerning the latencies before falling (Fig. 2A), ANOVA revealed effects of genotype (F1,70 73.917, P , 0.0001) and age (F3,70 18.205, P , 0.0001), with an interaction between these two factors (F3,70 13.942, P , 0.0001). Post hoc comparisons revealed that the scores of Lurcher mutants were signi®cantly lower than those of controls at three and nine months (P 0.0001), but not thereafter. The scores of control mice decreased signi®cantly between three and nine months (P , 0.01), and between nine and 15 months (P , 0.01), whereas there was no signi®cant difference in mutant mice, whatever their age. This ®nal observation could be due to the poor performances of the mutants as early as three months (¯oor effect). Concerning the number of paws used to keep hanging
Fig. 3. Wooden beam. (A) Latency before falling (in seconds, ^S.E.M.). (B) Walking time (in seconds, ^S.E.M.). (C) Proportion of time spent walking (in %, ^S.E.M.). (D) Distance covered (in cm, ^S.E.M.). Abscissae: age of the mice (in months).
(Fig. 2B), ANOVA revealed effects of age (F3,70 5.9897, P , 0.001) and genotype (F1,70 7.5488, P , 0.01), but no interaction between these two factors (F3,70 0.1357, P . 0.05). Wooden beam Analysis of the falling latencies (Fig. 3A) revealed effects of genotype (F1,70 85.461, P , 0.0001) and age (F1,70 16.064, P , 0.0001), with an age £ genotype interaction (F3,70 16.069, P , 0.0001). All the 1/1 mice, whatever their age, were able to stay on the beam for at least 600 s. Only the youngest (three months old) 1/Lc mice achieved this performance, which decreased thereafter. The performances of 1/Lc mice were signi®cantly lower than those of controls at 15 and 21 months (P , 0.0001), but not at three and nine months. Their performances decreased signi®cantly between nine and 15 months (P , 0.05). Concerning the time spent walking (Fig. 3B), ANOVA revealed signi®cant effects of age (F3,70 34.7197, P , 0.0001) and genotype (F1,70 13.5236, P , 0.001), with a signi®cant interaction between these two factors (F3,70 7.8319, P , 0.0001). The walking time was signi®cantly higher in Lurcher mutant mice than in controls only at three months (P , 0.0001). The values decreased signi®cantly in both groups between three and nine months (P , 0.0001 for 1/Lc and P , 0.05 for 1/1). Concerning the percentage of walking on the beam (Fig. 3C), statistical analysis revealed effects of genotype (F1,70 25.676, P , 0.0001) and age (F3,70 14.286, P , 0.0001), but no interaction between these two factors (F3,70 2.215, P . 0.05). The percentage of walking decreased between three and nine months, and was higher in Lurcher mutants, especially at three months (33.6% for the mutants and 15.3% for controls).
Motor learning in Lurcher mice during aging
Fig. 4. Hole board. (A) Time spent walking (in seconds, ^S.E.M.). (B) Number of slips (^S.E.M.). (C) Slip frequency (^S.E.M.). Abscissae: age of the mice (in months).
The distance covered on the beam is shown in Fig. 3D. ANOVA revealed no effect of genotype (F1,70 0.0017, P . 0.05), but a signi®cant effect of age (F3,70 12.6104, P , 0.0001) and an interaction between these two factors (F3,70 4.3562, P , 0.01). The distance covered by the mutants decreased signi®cantly between three and nine months (P , 0.001), whereas statistical analysis did not reveal any signi®cant difference between age groups in 1/1 mice. Hole board Statistical analysis of the time spent walking on the hole board (Fig. 4A) revealed no effect of genotype (F1,70 0.53, P . 0.05), but a signi®cant effect of age (F3,70 4.255, P 0.01). There was no interaction between these two factors (F3,70 0.408, P . 0.05), indicating that, in this test, the decrease with age of the time spent walking was similar in the two genotypes. Whatever the age, the number of slips (Fig. 4B) and the slip frequency (Fig. 4C) were always higher in 1/Lc mice than in controls. For these two variables, ANOVA revealed an effect of genotype (F1,70 103.2, P , 0.0001 for the number of slips and F1,70 269.781, P , 0.0001 for the slip frequency). There was no signi®cant effect of age (F3,70 0.481, P . 0.05 for the number of stumbles and F3,70 2.729, P . 0.05 for the slip frequency) and no signi®cant age £ genotype interaction (F3,70 0.392, P . 0.05 for the number of slips and F3,70 1.84, P . 0.05 for the slip frequency). These results indicated that Lurcher mice exhibited severe impairments in motor coordination and that these de®cits did not increase with age. The absence of a genotype effect concerning the
619
Fig. 5. Unstable board. (A) Latency before falling (in seconds, ^S.E.M.). (B) Slip frequency (^S.E.M.). Abscissae: age of the mice (in months).
time spent walking on the hole board showed that the activity of the mutants on the hole board was similar to that of controls. Unstable platform For each animal, the scores of the three trials (latency before falling and slip frequency) were averaged. The mean scores (^S.E.M.) per group were then calculated. At three months, all the 1/1, but not the 1/Lc, mice were able to stay on the platform for at least 180 s (Fig. 5A). ANOVA revealed signi®cant effects of genotype (F1,70 270.48, P , 0.0001) and age (F3,70 4.51, P , 0.01), but no interaction between these two factors (F3,70 1.06, P . 0.05). The mean scores of the mice decreased with age in the two genotypes. On the other hand, Fig. 5A clearly shows that the scores achieved by the mutants were always lower than those of controls, indicating that, whatever their age, the static equilibrium of 1/Lc mice was greatly impaired. These results are clearly different from those obtained with the wooden beam test, when the Lurcher mice could walk. On the elevated unsteady board, the mice could not walk and the strategies used to stay on the platform were limited: only an adapted repartition of the muscular strength in the limbs and in the body could permit them to stay on the apparatus. Their inability to quickly adapt their body posture could thus explain the static equilibrium impairments (when the motions were reduced). In the Lurcher mutant mice, these equilibrium impairments were obvious from three months of age under static conditions, while they were observed only from nine months of age in dynamic conditions (when motions were permitted). The slip frequency is shown in Fig. 5B. ANOVA
620
P. Hilber and J. Caston
revealed signi®cant effects of genotype (F1,70 196.724, P , 0.0001) and age (F3,70 33.88, P , 0.0001), with an interaction between these two factors (F3,70 22.648, P , 0.0001). The slip frequency was always higher in Lurcher mice than in controls (P , 0.05 at three and nine months, P , 0.0001 at 15 and 21 months). This frequency increased signi®cantly between nine and 15 months and between 15 and 21 months (P , 0.0001) in Lurcher mice, whereas statistical analysis revealed no signi®cant difference between age groups in control mice. DISCUSSION
Motor skills and motor learning in young mice In Lurcher mutant mice, massive degeneration of Purkinje cells in the cerebellar cortex and degeneration of the olivocerebellar pathway are associated with de®cits in muscular strength, motor coordination impairments and equilibrium disturbances. 14,32,33,43,54 However, in spite of these abnormalities due to lack of cerebellar cortex output, postural sensorimotor learning is not abolished. 13,30,31,34 Indeed, speci®c motor training allows the mutants to improve their performance and to reduce their motor de®cits by improving motor coordination. The ef®ciency of motor training can appear very quickly 30 and can be preserved for one week. 34 In the present study, we investigated motor learning on the rotorod with regard to speci®c motor abilities. In order to discover whether the scores of the animals on the rotorod at the end of the training session were due to either learning disabilities or motor skills de®cits, motor skills had to be tested after training. However, given that training could in¯uence motor skills scores, we have tested, in another group of mice, motor abilities before training on the rotorod (unpublished data). The results obtained showed that there was no signi®cant difference between motor skills scores of trained and untrained mice, whatever their age (data not shown). Results of the present study showed that Lurcher mutant mice were able to learn a sensorimotor task and to adopt an adapted walking behavior on the rotorod, which agrees with previous studies. 13,30,31,34 Training on the rotorod, which was ef®cient in improving motor coordination for a speci®c task (rotorod), did not permit the mutant mice to reduce their muscular strength, motor coordination and equilibrium de®cits in the other tests. Moreover, equilibrium de®cits in Lurcher mice were test speci®c. Indeed, when placed on the wooden beam, a test with low spatiotemporal constraints, the mutants did not exhibit any apparent de®cit in terms of latency before falling, but spent more time walking than 1/1 mice. These results have also been observed by Lalonde et al. 32 and Le Marec et al. 34 In our study, the long time spent walking by the mutants on the wooden beam was unlikely due to hyperactivity, because their locomotor activity was similar to that of controls on the hole board. It is more likely a strategy associated with regulation of equilibrium. 14 If so, walking behavior can be understood as a compensatory process which contributes to equilibrium maintenance,
by elaborating body adjustments preceding equilibrium disturbances in order to anticipate the fall (feedforward mechanisms). On the contrary, when Lurcher mice were subjected to an experimental task with higher spatiotemporal constraints (on the rotorod and the unstable platform), their equilibrium de®cits were obvious and they were unable to maintain balance for long. In these tests, maintenance of balance required a quick adaptation of muscular synergies, in response to modi®cations of the support, these reactive motor adjustments being sustained by feedback mechanisms. 24 Our results showed that, in spite of speci®c equilibrium impairments, resulting from alteration of reactive adjustments, Lurcher mutant mice were still able to elaborate and execute proactive adjustments in order to compensate their de®cits. These anticipatory body adjustments, corresponding to walking behavior, appeared when the spatiotemporal constraints were low (i.e. on the wooden beam). They were also observed when the constraints were higher, after a speci®c motor training, since on the rotorod, the mice learned to anticipate the rotation by walking. It therefore appears that the cerebellar cortex is not a crucial structure for the acquisition of a complex motor behavior. The role of the cerebellum in motor regulation can be understood from its implication in two transcerebellar loops, involved respectively in feedback and feedforward mechanisms, termed ªlowerº (cerebello-rubro-olivocerebellar) and ªupperº (cerebello-thalamo-cortico-pontocerebellar) transcerebellar loops (see Ref. 2). These two loops originate in the deep nuclei, but the cerebellar cortex constitutes a key in the relay system between sensory afferents and motor efferents. In Lurcher mutant mice, degeneration of Purkinje cells and the olivocerebellar pathway provokes a disruption of these two loops. Alteration of the reactive adjustments can be explained by disruption of the lower transcerebellar loop. In spite of the disruption of the upper transcerebellar loop, the feedforward mechanisms are still functional in Lurcher mice. This can be explained by the fact that they are relayed by other motor structures, such as the striatopallidal system, as suggested by Le Marec et al.; 34 indeed, its functional integrity in Lurcher mutant mice 44,50 leads us to think that it could be involved in motor learning. An alternative hypothesis is that the cerebellar cortex would not be mainly involved in the anticipatory body adjustments and that the deep cerebellar nuclei would be suf®cient. Indeed, it has been demonstrated that motor learning on the rotorod was abolished in cerebellectomized Lurcher mice, 13 suggesting that the deep cerebellar nuclei are involved in the elaboration and execution of anticipatory body adjustments. In this mutant, the loss of neurons in the cerebellar deep nuclei is moderate, 21,22 and a high cytochrome oxidase activity was observed in all three cerebellar deep nuclei. 50 Taken together, these data suggest that the deep cerebellar nuclei are mainly involved in the automatization of postural adjustments during motor training in Lurcher mutant mice. The role of the deep cerebellar nuclei in the cerebello-thalamocortical pathway, 4,40,41 the in¯uences they receive from the cerebral cortex 39,47 and their connections to the
621
Motor learning in Lurcher mice during aging
thalamus 42,45 suggest that they are involved in the organization of synergic limb movements and are, at least in part, independent of the in¯uence of Purkinje cell input. We can therefore postulate that, in young Lurcher mutant mice, the upper loop was not completely disrupted. It can therefore be stated that the cerebellar cortex is an essential link in the functioning of the lower transcerebellar loop, which is responsible for feedback mechanisms, and is not essential in the functioning of the upper transcerebellar loop, responsible for the elaboration and the maintenance of feedforward mechanisms. The functioning of this loop requires only the integrity of the deep cerebellar nuclei. The neurodegeneration appears very early in heterozygous Lurcher mice 8,20 and it is not excluded that neural plasticity in cerebellar circuitry, in particular between deep cerebellar nuclei and mossy ®bers, is involved in the organization of motor control in these mice. Such a plasticity could also be involved in the functioning of the upper loop and could be the basis of the automatization process expressed through motor learning.
Evolution of motor abilities and motor learning during aging Control mice. In rodents, aging is commonly associated with motor impairments, 10,15,38 which are test and strain dependent. 26 In spite of the progressive decrease of muscular strength with age, all the control mice were able to reach the learning criterion on the rotorod. Motor training was ef®cient whatever the age, and only a few days of training was required to compensate the age-related equilibrium de®cits. As demonstrated by Caston et al. 10 and Ingram et al., 26 the acquisition of motor tasks was not abolished in aged mice, showing that old control mice were still able to elaborate an internal motor program sustaining the organization and control of anticipatory body adjustments. Indeed, we did not detect any alteration of motor coordination in the hole board and equilibrium impairments on the wooden beam (i.e. when the spatiotemporal constraints were low) during aging, although we noticed a decline of equilibrium maintenance on the unstable board and on the rotorod before training (i.e. when the spatiotemporal constraints were high). We therefore postulate that the age-related decrease of equilibrium is associated with progressive dysfunction in structures involved in reactive body adjustments, structures which are included in the lower regulation loop. The neurochemical and structural modi®cations which appear in the cerebellar cortex 6,25,51±53 and in the vestibular system in old mice 51 could explain the progressive loss of the reactive postural adjustments in response to environmental changes. Given that the cerebellar cortex is not crucial for proactive adjustments, the mice were still able to anticipate equilibrium de®cits, whatever their age, by feedforward mechanisms associated with the functionality of the upper transcerebellar loop. Moreover, old control mice were still able to elaborate synchronized walking after training on the rotorod.
Lurcher mice. The aging pro®le of Lurcher mutant mice was different from that of controls. Indeed, we observed a severe alteration of motor learning abilities during aging, associated with impairments in equilibrium on the wooden beam. No further alteration was found in equilibrium on the rotorod before training and on the unstable board in old mice, probably due to the precocious and maximal alterations of these abilities at three months of age (¯oor effect). Disappearance of the walking behavior on the rotorod at the end of training, and decrease of the time spent walking on the wooden beam with age, would re¯ect the progressive inability to produce anticipatory body adjustments, and then motor training becomes progressively inef®cient. The pro®le of old mutant mice on the rotorod therefore resembles that of cerebellectomized Lurcher mice. 13 The inability of aged Lurcher mice to elaborate proactive adjustments can therefore be explained by dysfunctions of key structures implicated in the upper transcerebellar loop with age, such as the deep cerebellar nuclei, which, as suggested above, are involved in feedforward mechanisms. While a massive and precocious degeneration of Purkinje cells provokes only a mild neuronal loss in the deep cerebellar nuclei, a ªpossible evolution of the neurodegeneration process in these structuresº cannot be excluded, as suggested by Heckroth et al. 21 This degeneration or dysfunction in aged animals would result in the inability to elaborate and execute anticipatory body adjustments. However, the deep cerebellar nuclei are not the only structures which are involved in proactive adjustments; other structures integrated in the upper loop and directly or indirectly connected to the cerebellum could be affected (i.e. the thalamic motor nuclei and the striatopallidal system). Indeed, Strazielle et al. 50 found a high level of metabolic activity in several motor nervous structures in Lurcher mice (i.e. in the ventrolateral nuclei of the thalamus and the deep cerebellar nuclei). One can surmise that such high activity could be associated with a considerable production of oxygen free radicals, toxic for the cells. Thus, the accumulation of reactive oxidative components with age would provoke oxidative stress in neurons involved in motor regulation and highly activated in young Lurcher mice. Alteration and/or death of these structures resulting from this stress could explain the age-related motor impairments observed in old Lurcher mice. CONCLUSION 37
1
Marr and Albus have proposed a functional model of the cerebellum in which the cerebellar cortex plays a central role in motor learning. Structural and functional modi®cations in the cerebellar cortex after a learning session agree with this model. 3,16,17,27 Such a role seems to be devoted mainly to the olivocerebellar pathway, 46 which synchronizes the rhythmical activity of Purkinje cells. 36,46 However, studies in Lurcher mutant mice obviously show that lack of the cerebellar cortex and olivocerebellar pathway does not abolish the acquisition of motor tasks provided they are not too dif®cult. 13 Our results suggest that the cerebellar cortex, which is
622
P. Hilber and J. Caston
primarily involved in reactive compensatory adjustments conferring to this part of the cerebellum a facilitating role in motor skills, is not implicated in anticipatory motor regulations which require the integrity of the deep cerebellar nuclei. Therefore, alterations of the dynamic equilibrium and motor learning in aged Lurcher mice would re¯ect dysfunction and/or neuronal loss in nervous structures sustaining feedforward mechanisms, particularly in the deep cerebellar nuclei. The age-related motor skill impairments observed in control mice would result from alteration in nervous structures involved in reactive
adjustments (i.e. the cerebellar cortex). Thus, the evolution of motor status of Lurcher mutant mice is clearly different from that of controls. These behavioral data suggest that precocious and focused neurodegeneration of a speci®c nervous structure can provoke long-term disturbances or neuronal cell loss in nervous system structures usually connected to the cells which have disappeared. Lurcher mutant mice therefore appear to be a good model to study the pathological evolution of progressive neurodegeneration in the central nervous system during aging.
REFERENCES
1. Albus S. S. (1971) A theory of cerebellar function. Math. Biosci. 10, 25±69. 2. Altman J. and Bayer S. A. (1997) Development of the Cerebellar System: In Relation to its Evolution, Structure and Function. CRC, Boca Raton, FL. 3. Anderson B. J., Alcantara A. A. and Greenough W. T. (1996) Motor skills learning: changes in synaptic organization of the rat cerebellar cortex. Neurobiol. Learn. Mem. 66, 221±229. 4. Angault P. and Cicirata F. (1994) Organization anatomo-fonctionnelle de la voie d'in¯uence du neÂocervelet sur le cortex ceÂreÂbral moteur. Rev. Neurol. 150, 39±45. 5. Auvray N., Caston J., Reber A. and Stelz T. (1989) Role of the cerebellum in the ontogenesis of the equilibrium behavior in the young rat: a behavioral study. Brain Res. 505, 291±301. 6. Baurle J. and Grusser-Cornehls U. (1994) Axonal torpedoes in cerebellar Purkinje cells of two normal mouse strains during aging. Acta neuropath., Berlin 88, 237±245. 7. Caddy K. W. T. and Biscoe T. J. (1975) Preliminary observations on the cerebellum in the mutant mouse Lurcher. Brain Res. 91, 276±280. 8. Caddy K. W. T. and Biscoe T. J. (1976) The number of Purkinje cells and olivary neurons in normal and Lurcher mutant mouse. Brain Res. 111, 396±398. 9. Caddy K. W. T. and Biscoe T. J. (1979) Structure and quantitative studies on the normal C3H and Lurcher mutant mouse. Phil. Trans. R. Soc. Lond. 287, 167±201. 10. Caston J., Delhaye-Bouchaud N. and Mariani J. (1996) Motor behavior of heterozygous staggerer mutant (1/sg) versus normal (1/1) mice during aging. Behav. Brain Res. 72, 97±102. 11. Caston J., Jones N. and Stelz T. (1995) Role of preoperative and postoperative sensorimotor training on restoration of the equilibrium behavior in adult mice following cerebellectomy. Neurobiol. Learn. Mem. 64, 195±202. 12. Caston J., Lalonde R., Delhaye-Bouchaud N. and Mariani J. (1998) The cerebellum and postural sensorimotor learning in mice and rats. Behav. Brain Res. 95, 17±22. 13. Caston J., Vasseur F., Stelz T., Chianale C., Delhaye-Bouchaud N. and Mariani J. (1995) Differential roles of cerebellar cortex and deep cerebellar nuclei in the learning of the equilibrium behavior: studies in intact and cerebellectomized Lurcher mutant mice. Devl Brain Res. 86, 311±316. 14. Fortier P. A., Smith A. M. and Rossignol S. (1987) Locomotor de®cits in the mutant mouse, Lurcher. Expl Brain Res. 66, 271±286. 15. Gage F. H., Dunnett S. B. and BjoÈrklund A. (1984) Spatial learning and motor de®cits in aged rats. Neurobiol. Aging 5, 43±48. 16. Gilbert P. F. (1974) A theory of memory that explains the function and structure of the cerebellum. Brain Res. 70, 1±18. 17. Gilbert P. F. and Thach W. T. (1977) Purkinje cell activity during motor learning. Brain Res. 128, 309±328. 18. Guastavino J. M. (1988) EthogeÂneÁse de la souris mutante staggerer. Facteurs eÂpigeÂneÂtiques et reÂcupeÂrations fonctionnelles. TheÁse Doctorat d'Etat, Universite Paris-Nord, France. 19. Hadj-Sharaoui N., Frederic F., Zanjani H., Herrup K., Delhaye-Bouchaud N. and Mariani J. (1997) Purkinje cell loss in heterozygous staggerer mutant mice during aging. Devl Brain Res. 98, 1±8. 20. Heckroth J. A. (1992) Development of glutamic acid decarboxylase-immunoreactive elements in the cerebellar cortex of normal and Lurcher mutant mice. J. comp. Neurol. 315, 85±97. 21. Heckroth J. A. (1994) A quantitative morphological analysis of the cerebellar nuclei in normal and Lurcher mutant mice. II. Volumetric changes in cytological components. J. comp. Neurol. 343, 183±192. 22. Heckroth J. A. (1994) Quantitative morphological analysis of the cerebellar nuclei in normal and Lurcher mutant mice. I. Morphology and cell number. J. comp. Neurol. 343, 173±182. 23. Heckroth J. A. and Eisenmann L. M. (1991) Olivary morphology and olivocerebellar topography in adult Lurcher mutant mice. J. comp. Neurol. 312, 641±651. 24. Hilber P., Lalonde R. and Caston J. (1999) An unsteady platform test for measuring static equilibrium in mice. J. Neurosci. Meth. 88, 201±205. 25. Hussain S., Slikker W. Jr and Ali S. F. (1995) Age-related changes in antioxidant enzymes, superoxide dismutase, catalase, glutathione peroxidase and glutathione in different regions of mouse brain. Int. J. devl Neurosci. 13, 811±817. 26. Ingram D. K., London E. D., Reynolds M. A., Waller S. B. and Goodrick C. L. (1981) Differential effects of age on motor performance in two mouse strains. Neurobiol. Aging 2, 221±227. 27. Kleim J. A., Vij K., Ballard D. H. and Greenough W. T. (1997) Learning-dependent synaptic modi®cations in the cerebellar cortex of the adult rat persist for at least four weeks. J. Neurosci. 17, 717±721. 28. Lalonde R. (1987) Motor abnormalities in staggerer mutant mice. Expl Brain Res. 68, 417±420. 29. Lalonde R. (1987) Motor abnormalities in weaver mutant mice. Expl Brain Res. 65, 479±481. 30. Lalonde R. (1994) Motor learning in Lurcher mutant mice. Brain Res. 639, 351±353. 31. Lalonde R., Bensoula A. N. and Filali M. (1995) Rotorod sensorimotor learning in cerebellar mutant mice. Neurosci. Res. 22, 423±426. 32. Lalonde R., Botez M. I., Joyal C. C. and Caumartin M. (1992) Motor abnormalities in Lurcher mutant mice. Physiol. Behav. 51, 523±525. 33. Lalonde R., Joyal C. C., Guastavino J.-M. and Botez M. I. (1993) Hole poking and motor coordination in Lurcher mutant mice. Physiol. Behav. 54, 41±44. 34. Le Marec N., Caston J. and Lalonde R. (1997) Impaired motor skills on static and mobile beams in Lurcher mutant mice. Expl Brain Res. 768, 310±316.
Motor learning in Lurcher mice during aging
623
35. Le Marec N., Stelz T., Delhaye-Bouchaud N., Mariani J. and Caston J. (1997) Effects of cerebellar granule cell depletion on learning of the equilibrium behaviour: study in postnatally X-irradiated rats. Eur. J. Neurosci. 9, 2472±2478. 36. Llinas R. and Welsh J. (1993) On the cerebellum and motor learning. Curr. Opin. Neurobiol. 3, 958±965. 37. Marr D. A. (1969) A theory of cerebellar cortex. J. Physiol., Lond. 202, 437±470. 38. Marshall J. F. (1982) Sensorimotor disturbances in the aging rodent. J. Geront. 37, 548±554. 39. Massion J. (1973) Intervention des voies ceÂreÂbello-corticales et cortico-ceÂreÂbelleuses dans l'organisation et la reÂgulation du mouvement. J. Physiol., Paris 67, 117A±170A. 40. Massion J. (1976) The thalamus in the motor system. Appl. Neurophysiol. 39, 222±238. 41. Massion J. and Rispal-Padel L. (1972) Spatial organization of the cerebello-thalamo-cortical pathway. Brain Res. 40, 61±65. 42. McCormick D. A. (1995) The cerebellar symphony. Nature 374, 412±413. 43. Phillips R. J. S. (1960) Lurcher, a new gene in linkage group XI of house mouse. J. Genet. 57, 35±42. 44. Reader T. A., Strazielle C., Botez M. I. and Lalonde R. (1998) Brain dopamine and amino acid concentrations in Lurcher mutant mice. Brain Res. Bull. 45, 489±493. 45. Rispal-Padel L. (1993) Contribution des effeÂrences ceÂreÂbelleuses aÁ l'organisation des synergies motrices. Rev. Neurol., Paris 149, 716±727. 46. Rondi-Reig L., Delhaye-Bouchaud N., Mariani J. and Caston J. (1997) Role of the inferior olivary complex in motor skills and motor learning in the adult rat. Neuroscience 77, 955±963. 47. Shinoda Y., Sugiuchi Y. and Futami T. (1993) Organization of excitatory inputs from the cerebral cortex to the cerebellar dentate nucleus. Can. J. neurol. Sci. 20, Suppl. 3, S19±S28. 48. Shojaeian-Zanjani H., Mariani J., Delhaye-Bouchaud N. and Herrup K. (1992) Neuronal cell loss in heterozygous staggerer mutant mice: a model for genetic contributions to the aging process. Devl Brain Res. 67, 153±160. 49. Shojaeian-Zanjani H., Mariani J. and Herrup K. (1990) Cell loss in the inferior olive of the staggerer mutant mouse is an indirect effect of the gene. J. Neurogenet. 6, 229±241. 50. Strazielle C., Kremarik P., Ghersi-Egea J. F. and Lalonde R. (1998) Regional brain variations of cytochrome oxidase activity and motor coordination in Lurcher mutant mice. Expl Brain Res. 121, 35±45. 51. Sturrock R. R. (1989) Age-related changes in Purkinje cell number in the cerebellar nodulus of the mouse. J. Hirnforsch. 30, 757±760. 52. Sturrock R. R. (1989) Changes in neuron number in the cerebellar cortex of the ageing mouse. J. Hirnforsch. 30, 499±503. 53. Sturrock R. R. (1990) A comparison of quantitative histological changes in different regions of the ageing mouse cerebellum. J. Hirnforsch. 31, 481±486. 54. Thifault S., Girouard N. and Lalonde R. (1996) Climbing sensorimotor skills in lurcher mutant mice. Brain Res. Bull. 41, 385±390. 55. Thullier F., Lalonde R., Cousin X. and Lestienne F. (1997) Neurobehavioral evaluation of Lurcher mutant mice during ontogeny. Devl Brain Res. 100, 22±28. 56. Welsh J. P. and Harvey J. A. (1992) The role of the cerebellum in voluntary and re¯exive movements: history and current status. In The Cerebellum Revisited (eds Llinas R. and Sotelo C.), pp. 301±334. Springer, New York. 57. Welsh J. P., Lang E. J., Sugihara I. and Llinas R. (1995) Dynamic organization of motor control with the olivocerebellar system. Nature 374, 453±457. 58. Wilson D. B. (1976) Histological de®cits in the cerebellum of adult Lurcher (Lc) mice. J. Neuropath. exp. Neurol. 35, 40±45. 59. Zion C., Auvray N., Caston J., Reber A. and Stelz T. (1990) Effects of cerebellectomy at day 15 on the ontogenesis of the equilibrium behavior in the rat. Brain Res. 515, 104±110. 60. Zuo J., De Jager P. L., Takahashi K. A., Jiang W., Linden D. and Heintz N. (1997) Neurodegeneration in Lurcher mice caused by mutation in delta 2 glutamate receptor. Nature 388, 769±773. (Accepted 18 October 2000)