This study reports the detection of age-related changes in spatial reference memory performance in C57BL/6 male mice with the use of a two-day reference memory task and changes in the protein expression of NMDA receptors within crude synaptosomes during aging. A significant decline in spatial reference learning ability in the old mice as compared to the young was detected with a two-day task, regardless of whether the cued trials were performed before or after the place training. The young mice required two days in order to show significant improvement in the place-learning task. There were significant decreases with increased age in the expression of the ζ1 and ε2 subunits of the NMDA receptor in crude synaptosomes that were prepared from prefrontal/frontal cortex and these changes appeared to be related to declines in performance ability in the spatial reference memory task.
Previous work in our laboratory utilized a reference memory task in the Morris water maze that involved 12 days of testing in order to characterize learning abilities in different ages of C57BL/6 mice [17
]. Many other studies showing memory declines during aging in mice report the use of between 3 and 12 days to assess reference memory ability [55
]. Frick and colleagues have used a one-day task with 12 place trials to show significant declines in memory performance between young (7 months old) and middle-aged (18 months of age) male and female C57BL/6 mice [62
]. It should be noted, however, that Calhoun and co-workers found no effect of aging on reference memory performance in male C57BL/6 mice performing a reference memory task in the water maze for 5 days [63
]. The present study demonstrated that declines in spatial memory ability in male C57BL/6 mice between 3 and 26 months of age could be detected in both place learning trials and repeated probe trials with a 2-day protocol of place training. These deficits in the reference memory task were seen in animals that showed no age-related differences in control cued testing, thus reducing the possibility that the performances in the reference memory task were due to problems with vision, motor ability or motivation. The performance of the aged animals in the 2-day task that was performed before the cued task showed tendencies in the middle to late trials to plateau at a level higher than the young performance or to show a worsening of performance. This pattern has been seen with C57BL/6 mice in a 12-day reference memory task [17
] and in other studies using multiple training days for mice [55
] and age-impaired rats [54
]. Thus, this two-day task appears to be as sensitive to reference memory declines as tasks involving more days of training.
The young mice in the present study that underwent place training first showed a significant improvement in performance in probe trial performance at the end of day one as compared to the naïve probe trials, but required two days to show significant improvement in the place learning trials. The significant difference in performance between young and old mice was seen in the probe trials both during day 1 and across both days, but the age difference in place learning performance could only be detected across 2 days of training. This suggests that two days is necessary to detect the greatest differences in spatial memory ability between young and old C57BL/6 mice with a protocol of 8 place trials per day. Young female and male rats are able to show significant learning within eight place learning trials in a single day with the use of a similar protocol [20
]. Our laboratory and others have also seen this necessity for using more trials for mice than rats in order to obtain significant improvement in performance in a working memory task [19
]. These studies suggest that mice need more rehearsal than rats do in order to perform both reference and working memory tasks. Mice have also been shown to use different strategies to perform water maze tasks [69
Some aged mice showed deficits in performance of the cued trials as compared to the young when the cued trials were performed first, but these mice improved in performance by cued trial 5. The majority of aged mice (6 out of 10) that were tested in the cued trials first showed similar performances to the young mice throughout the cued trials. In previous studies, both young and old mice have also been excluded from analysis after showing poor performance in cued trials that followed place training [19
]. This suggests that there are some mice that can't perform well in multiple tasks in the water maze. Regardless of whether the cued trials were conducted after or prior to the place training trials, those aged mice that showed no problems in the cued task exhibited significant deficits in the place and probe trials. This suggests that the aged mice do have problems in performing the spatial reference memory task and it is not simply a problem with learning to acquire the platform, motivation, or sensory or motor abilities. What cannot be ruled out is a problem with acquiring the specific rules for the hidden platform task.
There was a significant decline between young and old mice in the protein expression of the ζ1 and ε2 subunits in crude synaptosomes from the prefrontal/frontal cortex in the present study. There was a 21% decrease in protein expression of the ζ1 subunit between 3 and 26 months of age in synaptosomes from prefrontal/frontal cortex in this study, but the protein expression of the ζ1 subunit showed a non-significant trend to decrease from 1.5 months to 25 months of age in homogenates from the caudal portion of prefrontal/frontal cortex in a previous study [52
]. The protein expression of the ε2 subunit also showed a greater decrease with increased age in synaptosomes from whole prefrontal/frontal cortex (47%) than previously reported in homogenates from the caudal portion of the prefrontal/frontal cortex (17%) [52
]. These results suggest that the synaptic pool of receptors is more susceptible to the effects of aging than the whole population of receptors within the neuron in the prefrontal/frontal cortex. This may reflect an increased turnover rate in the subunits either before or after they are inserted into the synaptic pool. These same subunits showed significant declines (19% decrease for ζ1 and 21% decrease for ε2) between young and old mice in homogenates of the whole cerebral cortex from the same strain of mice [49
]. The ε1 subunit showed a significant decline between 10 and 30 months of age in homogenates from whole cerebral cortex [49
], but those ages and extra cortical regions were not included in the present study.
The crude synaptosomes prepared from the hippocampus showed a trend for an increase in protein expression of the NMDA subunits examined. A similar pattern is seen in Wistar rat homogenates of hippocampus in which NR1 subunit proteins showed a significant increase and NR2A and NR2B subunit proteins showed a trend for an increase [70
]. This differs, however, from the results from homogenates prepared from the hippocampi of C57BL/6 mice in which the ζ1 subunit showed a significant decrease in expression between 3 and 30 months of age [49
]. These differences in the effects of aging on NMDA receptor expression in the synaptic pool versus the whole tissue are similar to those found in Fisher 344 rats, in which the basal surface expression of NMDA subunit proteins does not change between adult ages [71
], but expression of both NR1 (ζ1) and NR2B (ε2) expression is decreased in homogenates of hippocampus from the same strain of rats [50
]. This may represent maintenance of the synaptic pool of NMDA receptor subunits within the hippocampus during aging, but a reduction in the pool being produced. This could happen if the turnover rate at the synapse was reduced. There is also the possibility that the mice in the present study were influenced by the behavioral experience, although the rats in both of the studies mentioned above were naïve to behavioral training [50
]. Regardless of the cause, there is a very different effect of aging on the NMDA subunit expression patterns between the prefrontal/frontal cortex and the hippocampus in C57BL/6 mice. It should be noted that Long-Evans rats do not show significant effects of aging on NR1 protein expression either by Western blotting of homogenates or immunofluorescence on tissue sections [72
], so there appear to be some strain and species differences in the effects of aging on NMDA receptor expression in the hippocampus.
The protein expression of the ζ1 and ε2 subunits in the synaptosomal fraction of the prefrontal/frontal cortex correlated negatively with the proximity measurements in the place and the learning index score derived from the probe trials. Since high proximity measurements and scores indicate poor performance, the lower expression of ζ1 and ε2 in the old animals was associated with poorer performance, i.e., a positive correlation between protein and learning ability. This suggests that the changes during aging in these subunits of the NMDA receptor within the prefrontal/frontal cortex contribute to the memory problems seen in the older mice. In support of this, the administration of the ε2 subunit antagonist ifenprodil into orbital regions of the prefrontal cortex of young mice inhibited learning in the same task, suggesting that ε2 subunits in that region are involved in performing this spatial reference memory task [73
The prefrontal cortex has been shown in both human and animal studies to be important in applying different strategies to the handling of memories that have spatial and temporal properties, in retrieval of the sources of information, and in adjusting to changing environmental conditions (reviewed in [3
]). Lesioning studies in rats show that orbital and medial prefrontal regions contribute to performance in spatial reference memory tasks in the Morris water maze [74
]. The investigators provide arguments that this is due to a deficit in organization of the behaviours necessary to perform the task or problems with handling the spatial memory temporarily, as opposed to specific spatial guidance problems [74
]. Insular cortex injections of tetrodotoxin also impair the retention of platform position in the water maze [76
]. More recent studies do not see an affect of lesions of the medial prefrontal cortex on performance of the spatial reference memory task in the water maze [77
], but the tissue dissections in this study were not limited to medial prefrontal regions. They also included orbital and insular prefrontal regions, as well as some motor and somatosensory cortices. The literature thus provides evidence that some prefrontal regions play a role in the performance of spatial reference memory tasks. The current study suggests that declines in specific NMDA receptor subunits in the prefrontal/frontal cortex contribute to declines in performing the task by aged mice. The specific components of the performance that are affected will have to be determined.
The age-related declines seen in spatial reference memory performance of C57BL/6 mice in a 12-day protocol show a relationship with decreases in binding of the transmitter glutamate to the NMDA receptor in the prefrontal/frontal cortex and hippocampus [17
] and to the ratio of ε1/ζ1 and ε1/ε2 subunit mRNAs within subregions of the hippocampus [18
]. Changes in NR2B protein expression in homogenates from the hippocampus also show a correlation with spatial reference memory in aged Fisher 344 rats [78
]. NR1 levels in the CA3 region correlate with water maze performance during aging in Long-Evans rats [72
]. There was the suggestion of a relationship between the ε2 subunit expression in the hippocampus within each age group and reference memory ability, but it did not reach significance. There is a difference in how aging affects NMDA receptor binding and subunit mRNA expression between the dorsal and intermediate hippocampus in C57BL/6 mice [79
], so using the whole hippocampus in this study may have diluted out some effects.
There was a significant decline in the protein expression of syntaxin, a presynaptic terminal membrane protein [80
], in the prefrontal/frontal cortex between 3 and 26 months of age in the C57BL/6 mice in this study. For this reason, syntaxin was not used to correct for loading differences in the present study. The fact that there were differences in the effects of aging on the different NMDA receptor subunit and syntaxin suggest that the effects were not due to different amounts of protein loaded. Although we were only able to control for loading differences in this study by loading known amounts of protein, the presence of the standards on every gel do provide control for variability between gels and transfers. There was no significant effect of aging on syntaxin in the hippocampus, however, to be consistent in the reporting, the hippocampal subunits were also not normalized to syntaxin. The present results are in agreement with studies on Wistar rats that also show no declines in the protein expression of syntaxin in the hippocampus [82
], but differ from results of whole cerebral cortex from Wistar rats in which syntaxin showed no aging change [83
]. There was evidence though in this study that syntaxin expression within the hippocampus of the old mice may contribute to the reference memory deficits seen in the aged animals.