AGING AND MEMORY IN ANIMALS

The passage of time produces changes in both the behavior and the brains of organisms. A number of useful animal models of learning and memory in normal aging have expanded the knowledge base and extended the prospects for ameliorating learning and memory deficits. Completion of the mapping of the human and mouse genome and the development of transgenic mouse models in the 1990s have accelerated insights about mechanisms of learning, memory, and aging. Since the mid-1990s, mouse models of neuropathology in Alzheimer's disease have become available for behavioral testing. Two features of animal models make them invaluable: First, the life spans of most animals are considerably shorter than the human life span, compressing the time required to observe processes of aging. Second, invasive or high-risk observations and experimental manipulations are feasible with animals but not with humans.

Aging is most typically associated with declines in functioning, both neural and behavioral. However, individual organisms age at different rates. One organism may show a steady decline in functioning, whereas another shows only slight changes over the years. An important goal toward an understanding of aging processes is to determine how changes in neural structures impact behavior. As such, behavioral paradigms that engage well-defined neural substrates are particularly valuable. Two such neurobiologically well-characterized paradigms will be highlighted: eyeblink classical conditioning and spatial learning and memory.

Eyeblink Classical Conditioning

The basic classical conditioning procedure, named the "delay" procedure by Ivan Pavlov, involves repeated trials in which the presentation of an initially neural stimulus, such as a tone (the conditioned stimulus, or CS), is followed after approximately half a second by a stimulus that evokes a reflexive eyelid closure, such as a corneal air puff (the unconditioned stimulus, or US). The CS turns on and remains on while the US is delivered and the two stimuli coterminate. Initially, eye blinks occur only after the US, and the blinks are a reflexive response (the unconditioned response, or UR). Eventually, eye blinks occur after the CS but before the US. This is a learned response (the conditioned response, or CR). Thus, learning is defined as the acquisition of CRs.

Richard F. Thompson suggested that the eye-blink classical conditioning paradigm might be the Rosetta Stone for brain substrates of age-related deficits in learning and memory (Thompson, 1988). Thompson's major point was that eyeblink classical conditioning is a simple form of learning that can be studied with little modification across a variety of species, including humans. There are at least four advantages to using eyeblink conditioning as an animal model of the effects of aging on learning and memory in humans. First, both animals and humans show age-associated deficits in conditioning, and these can be easily dissociated from age-associated changes in sensory systems (i.e., differences in CS thresholds) or motor systems (differences in UR amplitude). Second, both animals and humans show age-associated changes in the neural substrates critical for eyeblink conditioning, the cerebellum and the hippocampus. Third, age-associated deficits have been artificially induced with drugs in both young animals and young humans. Finally, age-associated deficits can be reversed in normal older animals using cognition-enhancing drugs.

The critical substrate for eyeblink conditioning is the cerebellum. The hippocampus plays a modulatory role in acquisition of CRs. Abnormal functioning of the hippocampus retards the rate of eyeblink conditioning in the delay procedure. In addition, the hippocampus is essential for eyeblink classical conditioning procedures involving greater complexity, such as trace conditioning. In the trace procedure, the CS is presented and then turned off, and a blank period ensues before the onset of the US. The blank period is called the "trace" and is shown in the left panel of Figure 1.

In eyeblink classical conditioning, intervals between the CS and US (called interstimulus intervals, ISIs) of 250 and 750 milliseconds are the most common, with the most rapid conditioning between 250 and 500 milliseconds. The trace procedure extends the ISI in addition to inserting the blank trace period. Thus, it increases difficulty in two ways. Holding the ISI constant at 750 milliseconds, researchers compared the delay and trace procedures in three-month-old and twenty-four-month-old rabbits. There were significant effects of age and procedure (see Figure 1). Older rabbits performed more poorly in both procedures, and both age groups performed more poorly in the trace than in the delay procedure. Comparison of a number of studies testing the delay and trace procedure in older rabbits indicated that age differences in conditioning appeared earlier when the trace procedure was used.

Cerebellar Substrates of Impaired Conditioning in Older Animals

Purkinje cells in the cerebellum integrate CS and US input and show patterns of engagement during eyeblink conditioning. In rabbits, Purkinje cell counts have been carried out using histological techniques after behavioral testing with eyeblink classical conditioning. The correlations between Purkinje cell number and eyeblink classical conditioning in rabbits were high and statistically significant (Woodruff-Pak, Cronholm, and Sheffield, 1990). The fewer Purkinje cells a rabbit had, the longer it took it to acquire CRs. Further analysis demonstrated that this relationship was relatively independent of age because there was a highly significant correlation between Purkinje cell number and conditioning when only young rabbits were included. Mutant mice without Purkinje cells condition slowly and produce fewer CRs, whereas their wild type littermates with Purkinje cells condition normally (Chen et al., 1996). Given the essential role of the cerebellum in the acquisition of CRs, Purkinje cell loss may account for a significant portion of the age-related difference in eyeblink classical conditioning.

Hippocampal Substrates of Impaired Conditioning in Older Animals

There is some evidence of age differences in hippocampal activity during eyeblink conditioning. Neuronal responses were recorded from the dorsal hippocampus of young (three-month-old), middle-aged (twenty-six-to-thirty-three-month-old), and older (thirty-nine-to-fifty-month-old) rabbits during eye-blink conditioning in the 750-millisecond trace procedure shown in Figure 1 (Woodruff-Pak, Lavond, Logan, and Thompson, 1987). Older rabbits were significantly impaired in acquiring CRs. Furthermore, older rabbits showed significantly less neural activity in the US period than young rabbits by the second session of training. Matthew McEchron and John Disterhoft also reported that older rabbits had less hippocampal responsivity in the US period.

In a series of experiments in young and older rabbits, Disterhoft and McEchron (2000) found that conditioning-related hippocampal pyramidal-cell activity varied across cells and that the different response profiles were differentially affected by aging. Patterns of hippocampal pyramidal cell activation associated with acquisition of trace eyeblink conditioning were different from activity recorded after CRs became asymptotic. Pyramidal-cell activity associated with acquisition was more sensitive to the effects of aging (McEchron, Weible, and Disterhoft, 2001). Various patterns of single-unit pyramidal-cell activity were identified, and three response patterns were different between young and older rabbits that learned and those aged rabbits that did not. The patterns showed significant changes during the first five days of conditioning in the young and aged learners, but the patterns showed no change in the aged nonlearning group. If these cells function to hold important information for consolidation in other neural structures, age-related deficits in conditioning may be ameliorated by enhancing the function of these cells.

Spatial Learning and Memory

Spatial behavioral tests evaluate the ability of the organism to know or to have a representation of its location in the environment and thus to navigate effectively. The intact functioning of the hippocampus is necessary for learning and remembering spatial tasks. In old age, spatial memory is less efficient in humans and animals.

When single cells are recorded in the hippocampus of a behaving rat, firing rate increases when the animal is in a particular place in the environment. These cells have been called "place cells," and the area over which these cells show increased firing rates are called the cells' "place fields." Carol Barnes (2001) described deficits that occur in the development and maintenance of hippocampal place fields in old rats. As rats explore the environment, there is a change in the pattern of hippocampal place-cell discharge that occurs as a consequence of experience. For example, when rats run laps around a track in one direction, there is an expansion of the place fields and a shift in their centers of mass toward the origin of the route. In old rats, there is a striking reduction in this experience-dependent form of plasticity in the old place cells. Barnes suggested that the lack of field broadening observed in old rats might be expected to lead to a loss of precision in the information transmitted as a consequence of experience. Ability to remember routes may also be impaired by this deficit in place-cell field broadening.

Barnes also found that the same memory-impaired old rats that showed deficits in experience-dependent place-field expansion also retrieved inappropriate hippocampal maps on some occasions. Even when these memory-impaired old rats were in familiar environments, they retrieved inappropriate hippocampal maps from time to time. When a young rat is exposed to a familiar environment on one day and then exposed to that environment again a second time later that day, the same place-field map will be recorded from hippocampal place cells on both sessions. Testing old rats in this two-session recording procedure, Barnes and her colleagues (1997) observed a bimodal distribution of responses. For two-thirds of the double-session recordings, the old rats retrieved the same map on both occasions, performing normally as young rats. However, on one-third of the double-session recordings, the old rats exhibited a complete rearrangement of the place-field map between the two sessions. They apparently retrieved the wrong map on one of the sessions. This failure to retrieve the correct map may explain why old rats are more likely to make behavioral map-retrieval errors. Older organisms, including rats, monkeys, and humans, have a greater tendency to become lost. Altered hippocampal plasticity mechanisms may underlie these changes in cognition that occur during normal aging.

Transgenic Mouse Models of Alzheimer's Disease

Severe memory loss is the most prominent cognitive symptom of Alzheimer's disease (AD), and a fundamental role in the pathogenesis of AD is brain deposition of β -amyloid (A β). Mutations in the amyloid precursor protein (APP) and presenilin-1 (PS1) genes are linked to forms of AD that are carried in families and called familial AD. By altering APP metabolism, these mutations result in increased brain levels of A β peptide. Transgenic mice harboring mutant forms of the APP and/or PS1 gene associated with AD in humans are valid tools in the study of pathophysiological and behavioral effects of those genes in AD. Due to the relatively short life span of mice (two to three years), a high overexpression of the transgene is necessary to achieve the development of AD-like symptoms in the animals. The first transgenic mouse models of AD were produced in the mid-1990s, and thereafter there were a number of mouse models of AD that developed A β -containing plaques in the hippocampus and neocortex, thus modeling human AD.

The hippocampus is engaged in eyeblink classical conditioning and in spatial learning and memory, and both of these behaviors are profoundly impaired in AD. Spatial learning and memory is the behavior most commonly tested in transgenic mouse models of AD, although eyeblink classical conditioning is a useful alternative behavioral measure that has direct parallels and can be tested in humans diagnosed with AD. A frequently used behavioral test for AD mouse models is the Morris water maze, in which mice are placed in a pool of water that is opaque (to hide a platform) and must learn the location of that platform to escape from the water. Impairment in this and other forms of spatial learning and memory are observed in various transgenic mouse models of AD.

Dale Schenk and colleagues (1999) made a remarkable discovery that vaccinations with A β peptide can dramatically reduce amyloid deposition in transgenic mouse models of AD. A β peptide vaccination prevents spatial learning and memory loss (Morgan et al., 2000). The long-term behavioral results of A β peptide vaccinations indicate that the behavioral protection of the vaccinations is task-specific, with preservation of hippocampal-associated spatial-memory tasks most likely to occur (Arendash et al., 2001).

Evidence suggests that normal aging affects mammalian eyeblink conditioning through age-related deficits in the cerebellum and hippocampus. Age-related changes in spatial learning and memory also rely on hippocampal mechanisms. AD exacerbates impairment in learning and memory and profoundly disrupts hippocampal function early in its course. Transgenic models of AD provide a means to test therapeutic interventions, such as vaccination with A β peptide, that might protect against the cognitive and neural impairment characteristic of this neurodegenerative disease.

See also:AGING AND MEMORY IN HUMANS

Bibliography

Arendash, G. W., Gordon, M. N., Diamond, D. M., Austin, L. A., Hatcher, J. M., Jantzen, P., DiCarlo, G., Wilcock, D., and Morgan, D. (2001). Behavioral assessment of Alzheimer's transgenic mice following long-term A β vaccination: Task specificity and correlations between A β deposition and spatial memory. DNA and Cell Biology 20, 737-744.

Barnes, C. A. (2001). Plasticity in the aging central nervous system. International Review of Neurobiology 45, 339-355.

Barnes, C. A., Suster, M. S., Shen, J., and McNaughton, B. L. (1997). Cognitive map multistability in aged rat hippocampus. Nature 388, 272-275.

Chen, L., Bao, S., Lockard, J. M., Kim, J. J., and Thompson, R. F. (1996). Impaired classical eyeblink conditioning in cerebellar lesioned and Purkinje cell degeneration (pcd) mutant mice. Journal of Neuroscience 16, 2,829-2,838.

Disterhoft, J. F., and McEchron, M. D. (2000). Cellular alterations in hippocampus during acquisition and consolidation of hippocampus-dependent trace eyeblink conditioning. In D. S. Woodruff-Pak and J. E. Steinmetz, eds., Eyeblink classical conditioning, Vol. 2: Animal models, pp. 313-334. Boston: Kluwer Academic Publishers.

McEchron, M. D., Weible, A. P., and Disterhoft, J. F. (2001). Aging and learning-specific changes in single-neuron activity in CA1 hippocampus during rabbit trace eyeblink conditioning. Journal of Neurophysiology 86, 1,839-1,857.

Morgan, D., Diamond, D. M., Gottschall, P. E., Ugen, K. E., Dickey, C., Hardy, J., Duff, K., Jantzen, P., DiCarlo, G., Wilcock, D., Connor, K., Hatcher, J., Hope, C., Gordon, M., and Arendash, G. W. (2000). A β peptide vaccination prevents memory loss in an animal model of Alzheimer's disease. Nature 408, 982-985.

Schenk, D., Barbour, R., Dunn, W., Gordon, G., Grajeda, H., Guido, T., Hu, K., Huang, J., Johnson-Wood, K., Kahn, K., Kholodenko, D., Lee, M., Liao, Z., Lieberburg, I., Motter, R. Mutter, L., Soriano, F., Shopp, G., Vasquez, N., Vandevert, C., Walker, S., Wogulis, M., Yednock, T., Games, D., and Seubert, P. (1999). Immunization with amyloid- β attenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature 400, 173-177.

Thompson, R. F. (1988). The Rosetta stone for brain substrates of age-related deficits in learning and memory? Neurobiology of Aging 9, 547-548.

Woodruff-Pak, D. S., Cronholm, J. F., and Sheffield, J. B. (1990). Purkinje cell number related to rate of eyeblink classical conditioning. NeuroReport 1, 165-168.

Woodruff-Pak, D. S., Lavond, D. G., Logan, C. G., and Thompson, R. F. (1987). Classical conditioning in three-, thirty-, and forty-five-month-old rabbits: Behavioral learning and hippocampal unit activity. Neurobiology of Aging 8, 101-108.

Diana S.Woodruff-Pak