Dissociative decline of spatial learning and recall with aging in male CD1 mice

On aging, spatial memory declines to some degree in both, rodents and humans. It is unknown, however, whether aging brings about a decline of encoding and retrieval of spatial information in parallel. The present study examined spatial encoding and retrieval in male CD-1 mice at 4, 9, and 12 months of age in a complex dry-land maze. The total time to reach the goal zone was age-dependent favoring mice at 4 months of age. We showed previously that moving time represents encoding of spatial information and resting time represents recall of previously learned spatial information. The average moving time decreased from 69.8 ± 5.3 s (mean ± SEM), 69.7 ± 8.0 s, and 78.9 ± 4.9 s to 17.0 ± 2.3 s (p < 0.001 ), 24.7 ± 2.7 s (p < 0.001), and 31.0 ± 3.5 s (p < 0.001) at 4, 9, and 12 months of age, respectively. The average resting time decreased from 34.9 ± 5.6 s, 22.2 ± 4.2 s, and 41.7 ± 5.3 s to 3.6 ± 1.2 s (p < 0.001), 5.3 ± 1.8 s (p = 0.009), and 22.7 ± 4.9 s (p = 0.007) at 4, 9, and 12 months of age, respectively. We conclude that age-related deficits of spatial memory in mice manifest with an encoding deficit prior to a retrieval deficit.


INTRODUCTION
Spatial learning in experimental animals is assessed using mazes since the beginning of the 19th century [1].While dry-land mazes have been used in early studies, the use of water mazes has become the mainstay of research on spatial memory in rodents for about 30 years [2][3][4][5].Recently, it has been argued that use of dry-land mazes may offer additional insight into the mechanisms of spatial learning.Thus, the number of animal studies using dry-land mazes has increased over the last couple of years.Spatial learning in water mazes is subject to the influence of non-cognitive processes such as fear which may blur the insight into the cognitive processes at work [6,7].Furthermore, use of dry-land mazes allows separation of different cognitive processes performed during resting and moving time in the maze [8].This is not possible in water mazes since animals need to move continuously.
On moving around in an unknown or familiar environment, the localization information is encoded [8][9][10] and animals build an internal representation of the space around them, and the cognitive map [11,12].During resting times, animals update and recall the localization information and compare currently experienced sequence information with stored sequences [8,13,14].
It was the goal of the present study to assess age-related spatial memory performance in a complex dryland maze in male CD-1 mice.

MATERIALS AND METHODS
All animal experiments were performed according to institutional guidelines and the "Principles of laboratory animal care" (NIH publication No. 86-23, revised 1985).

Animals
Three groups of male CD-1 mice (Charles River) from different age cohorts, i.e., young (4 months, n = 9), middle-aged (9 months, n = 4), and aged (12 months, n = 9) mice were used in this study.Previous reports in humans [47] and experimental animals [48] show a clear-cut gender dependence of spatial orientation with males relying on hippocampal integrity to a much greater extent than females.In addition, hippocampus-dependent spatial learning in females may be influenced by the stage of the estrus cycle [49].At the given age of the transgenic animals used in this study, hippocampal impairment is predominant.Therefore, the present study was restricted to male animals.Four or five animals were housed in a cage and were maintained on a 12 h light/dark cycle in a temperature (22˚C ± 2˚C) and humidity (55% ± 5%) controlled room, with water ad libitum available.

Maze and Behavioural Testing
To assess learning we used a complex maze that has been described previously [50][51][52] (Figure 1).The complex maze was constructed of gray rigid PVC and the overall size was 200 × 180 cm with 10 cm high, opaque walls and 10 cm wide, red alleys the mice traveled on.The maze has a starting point, crossings, t-junctions, blind alley and a goal zone and was surrounded by a curtain with optical geometric cues at four cardinal directions.Animals were trained five times a day for eight consecutive days.They had a maximum time of 300 s to find the goal zone, where they were rewarded with a food pellet.The behavioural observation was made between 10:00 AM and 4:00 PM.Experiments were recorded by a tracking system (Multitrack, Accuscan, USA).

Statistical Analysis
All statistical analyses were carried out using the statistics program SPSS (SPSS 17.0 for Windows, SPSS Inc. IL 60606, Chicago).Statistical significance was accepted at p < 0.05.

RESULTS
To assess spatial learning we analyzed total time to reach the goal zone.A two-way ANOVA with Fisher LSD multiple comparison testing and factors trial and age shows a significant effect of trial (F 39,1120 = 6.937, p < 0.001) and age (F 2,1120 = 45.448,p < 0.001) (Figure 2, Table 1).
Previously we argued, that separate analysis of moving and resting time allows distinction of encoding the spatial information and retrieving it [8].Hence, we analyzed moving and resting time in a similar fashion.All age-groups improved their performance during repeated exposure to the complex maze as shown by a significant effect of trial on moving time (two-way ANOVA; F 39,1120 = 9.206, p < 0.001) and resting time (two-way ANOVA; F 39,1120 = 2.797, p < 0.001) (Figure 2).Likewise, a significant effect of age was found for moving time (twoway ANOVA; F 2,1120 = 27.251,p < 0.001) and resting time (two-way ANOVA; F 2,1120 = 41.994,p < 0.001) (Table 1).Compared to mice at an age of 4 months, moving time was increased in animals at the age of 9 and 12 months.Resting time, however, did not increase until the age of 12 months.
To rule out that total or moving time are modulated by speed in an age-dependent manner we analyzed running speed.Running speed was lowest in animals at 9 months of age and higher at both, 4 and 12 months.To analyze the effect of age and trial on running speed we performed a two-way ANOVA with factors trial and age and found significant effects for both, trial (F 39,1120 = 15.570,p < 0.001) and age (F 2,1120 = 32.193,p < 0.001) (Figure 3, Table 1).Regardless of this effect of age on speed, the length of the path that animals use on their way to the goal zone shows an age-dependent increase (Figure 3, Table 1).This indicates that the cognitive processes of

DISCUSSION
The present study examined spatial learning in a complex dry-land maze in mice at 4, 9, and 12 months of age.Overall, spatial memory performance declined with aging.This finding is in good harmony with a rich literature on decline of spatial memory as assessed with different setups of water mazes [2][3][4][5] and dry-land mazes [19,20,53].To our knowledge, separate analyses of different cognitive processes while the animals rest and move have only been performed, recently [8][9][10]13,14].

Running Speed
Running speed is a confounding variable in using escape latency from start zone to the goal zone in dry-land mazes.In harmony with a previous study in female CD-1 mice [54], we observed a u-shape-like running speed with highest speed in mice at 4 and 12 months of age and slowest speed in animals 9 months of age.Regardless of the modulation of moving and total time by speed we also observed a difference in the length of the path from start zone to goal zone being lowest in the youngest animals and highest in the oldest animals.We interpret the latter finding such as to show that the overall performance for spatial learning declines steadily with aging.

Encoding and Recall of Spatial Information
Spatial information is encoded while animals move in the maze [8][9][10].Older animals in the present study encoded the spatial map of the maze less successfully than younger animals and needed more time and a longer path length to reach the goal zone.This is in good accordance with the literature showing a decline of encoding performance in middle-aged and aged humans [15,16,55,56] and experimental animals [21][22][23][24][25].While the spatial encoding performance is worse in 9 and 12 months old animals compared to 4 months old animals, no further decline is observed between 9 and 12 months.Retrieval of spatial information in the complex maze take place predominantly during resting time [8,13,14].Resting time increases with age in the present study.This is in good accordance with the literature showing a decline of retrieval performance in aged humans [17,18] and experimental animals [26].In contrast to encoding of Learning curves for total, moving and resting time that male animals needed to perform the complex maze over the whole experimental period.Values represent means ± standard errors for male 4 (squares), 9 (circles), and 12 months old (triangles) mice.spatial information with similar performance of 9 and 12 months old animals there is further decline of spatial retrieval performance between 9 and 12 months.

Sequence of Decline of Spatial Learning Processes with Aging-A Synthesis Considering the Literature
The effects of age on memory performance are determined by both the level of information processing at the encoding phase and the degree of environmental support during the retrieval phase.Possible memory mechanism failures in the elderly include poor encoding or acquisition of new information, faulty storage or retention of information in memory, and/or an inability to decode or retrieve material that has been adequately encoded and stored [57].
Overall, we observe a progressive decline of spatial retrieval performance with increasing age from 4 and 9 to 12 months in the present study.In contrast, spatial memory encoding performance declines from 4 to 9 months but remains alike between 9 and 12 months.We interpret this pattern such as to indicate that spatial memory encoding performance reaches a floor level already at or prior to 9 months of age in male CD-1 mice in the current paradigm while spatial memory retrieval is less sensitive to aging and further decline of performance is observed between 9 and 12 months in the current paradigm.This interpretation is supported by previously reported results on encoding and retrieval in humans with assessment of source memory, i.e. the ability to remember contextual information present at the moment an event occurs.
The failure to retrieve memory content correctly may arise from a failure to encode the representation sufficiently strongly [58]."Encoding" can be assessed by means of investigating "subsequent memory effects" [59].Source memory accuracy differed between young, middle-aged, and old adults [60] for reasons of less efficient encoding mechanisms.Brain regions which participate in the successful encoding of source information for persons in their 50 s are different than those utilized by young adults and change between middle age and old age [60].
In humans, recall performance between young and middle-aged groups has been reported to be alike while it declined in old age.Middle-aged adults may have preserved retrieval capacities, and/or enhanced encoding capacities.Old subjects enhanced but failed to normalize their retrieval performance, suggesting only partially successful compensatory mechanisms [61].In verbal recall the old subjects scored lower than both the young and middle-aged subject with young and middle-aged subjects performing alike [62,63].

CONCLUSION
We conclude that age-related deficits of spatial memory manifest with an encoding deficit prior to a retrieval deficit in aging rodents.Future studies need to further investigate the neurobiological and neuroanatomical reasons for this age-specific decline in different spatial memory processes.Since there are indications towards sex differences in spatial memory [47][48][49], female mice should be included in follow-up studies.

Figure 1 .
Figure 1.Aerial view of the complex maze with curtain and geometrical cues.Animals were placed in the start zone.A video-tracking system (cf.methods) registered the location of the animals' position at a frequency of 1 Hz.

Figure 2 .
Figure 2.Learning curves for total, moving and resting time that male animals needed to perform the complex maze over the whole experimental period.Values represent means ± standard errors for male 4 (squares), 9 (circles), and 12 months old (triangles) mice.

Figure 3 .
Figure3.Average running speed, total time and distance that male animals needed to perform the complex maze over the whole experimental period using two-way ANOVA with Fisher LSD multiple comparison testing.Values represent means ± standard errors for male 4 (black bar, 4 mo), 9 (gray bar, 9 mo), and 12 months old (white bar, 12 mo) mice.

Table 1 .
Two-way ANOVA for male CD-1 mice in age of 4 months (4 mo), 9 months (9 mo), and 12 months (12 mo) in the complex maze.Post hoc multiple comparison testing (Fisher LSD) with p-values for comparison of differences between groups.* indicate statistical significance.
*encoding and retrieval of the spatial information is more successful in 4 than in 12 months old animals.