Neuroscience
Imaging correlates of brain function in monkeys and rats isolates a hippocampal subregion differentially vulnerable to aging

Scott A. Small * {dagger}, Monica K. Chawla {ddagger}, Michael Buonocore §, Peter R. Rapp ¶, and Carol A. Barnes

The hippocampal formation contains a distinct population of neurons organized into separate anatomical subregions. Each hippocampal subregion expresses a unique molecular profile accounting for their differential vulnerability to mechanisms of memory dysfunction. Nevertheless, it remains unclear which hippocampal subregion is most sensitive to the effects of advancing age. Here we investigate this question by using separate imaging techniques, each assessing different correlates of neuronal function. First, we used MRI to map cerebral blood volume, an established correlate of basal metabolism, in the hippocampal subregions of young and old rhesus monkeys. Second, we used in situ hybridization to map Arc expression in the hippocampal subregions of young and old rats. Arc is an immediate early gene that is activated in a behavior-dependent manner and is correlated with spike activity. Results show that the dentate gyrus is the hippocampal subregion most sensitive to the effects of advancing age, which together with prior studies establishes a cross-species consensus. This pattern isolates the locus of age-related hippocampal dysfunction and differentiates normal aging from Alzheimer's disease.

Cross-species studies have documented that the hippocampal formation, a structure vital for learning new memories, is particularly vulnerable to the aging process (1–3). A complex structure, the hippocampus is divided into separate but inter-connected anatomic subregions: the entorhinal cortex, the dentate gyrus, the CA subfields, and the subiculum (4). Each hippocampal subregion contains a distinct population of neurons that express a unique molecular profile (5). This uniqueness can account for why each hippocampal subregion is differentially vulnerable to mechanisms of memory dysfunction (6). For example, transient hypoperfusion will cause hippocampal-dependent memory deficits by targeting the CA1 subregion, whereas early Alzheimer's disease causes an overlapping memory deficit by targeting the entorhinal cortex.

Although a number of studies suggest some cell loss with age (7, 8), compared to neurodegeneration, aging is characterized by a relative absence of frank cell death and lacks definitive histopathological markers (9). Rather, aging affects hippocampal performance by impairing normal neuronal physiology, expressed as synaptic dysfunction. This physiologic feature of aging accounts, to a large degree, for the difficulty identifying the primary hippocampal subregions most sensitive to normal aging. Not only is quantifying synaptic dysfunction difficult in postmortem tissue (10–12), but because of hippocampal inter-connectivity, dysfunction in one subregion affects physiologic properties in other hippocampal subregions (1) and even throughout the circuit as a whole. Thus, assessing the functional integrity of each hippocampal subregion individually and simultaneously in living subjects is an effective approach for pinpointing the primary site of age-related hippocampal dysfunction.

Most in vivo functional imaging techniques can assess global hippocampal function but do not have sufficient spatial resolution required to visualize individual hippocampal subregions. One study used a high-resolution variant of functional MRI (13) to image individual hippocampal subregions across the human lifespan (14). Although a decline in signal was observed throughout, the dentate gyrus was the dominant hippocampal subregion that had a pattern of decline most consistent with normal aging (14). These results remain inconclusive, however, because of the difficulty in excluding individuals in the earliest stages of Alzheimer's disease, even among nondemented elderly. Furthermore, although the MRI technique used in the study is sensitive to deoxyhemoglobin, a hemodynamic correlate of brain metabolism, it is also sensitive to other tissue elements (13), such as iron, which cannot be excluded as factors influencing the observed effect.

Our first experiment was designed to address both limitations. Rhesus monkeys were chosen as our experimental group, because they, like all nonhuman mammals, develop age-related hippocampal dysfunction but do not develop Alzheimer's disease (9). Animals were imaged with an MRI technique that generates a measure of regional cerebral blood volume (CBV). Among the three correlates of oxidative metabolism that can be measured with MRI, cerebral blood flow, cerebral blood volume, and deoxyhemoglobin content (15, 16), we chose to rely on MRI measures of CBV to assess hippocampal function. CBV can be measured with higher spatial resolution compared with cerebral blood flow (17), and it provides a purer measure of basal metabolism compared to maps sensitive to basal deoxyhemoglobin (13). MRI measures of CBV have been shown to tightly correlate with regional energy metabolism and cerebral blood flow (for example, see refs. 16 and 18), and can successfully map dysfunction in the brain (18–21), including the hippocampal formation.

Any hemodynamic measure is only an indirect correlate of brain metabolism and cannot exclude that an effect is influenced by primary vascular alterations (22). Furthermore, MRI does not have neuronal resolution. Our second experiment was designed to address these issues. Arc is an immediate early gene whose expression is correlated with spike activity. Arc also modulates long-term potentiation (23–25), a mechanism associated with memory and impaired by aging (1). The cellular localization of Arc mRNA can be used to determine which individual neurons participate in a specific behavioral experience. This new methodology, cellular compartment analysis of temporal activity by fluorescence in situ hybridization combined with high resolution confocal microscopy (or catFISH) was used to map the expression of Arc mRNA in single cells of young, middle-aged, and old rats, after exploration of a novel environment. This imaging approach can directly assess behaviorally induced physiological function with single-cell resolution.

**********

Discussion


The functional organization of the hippocampal formation presents an obstacle when attempting to isolate the primary site of memory dysfunction (6). The hippocampal subregions are interconnected, and, therefore, a lesion in an individual subregion leads to secondary alterations in downstream subregions by means of transsynaptic mechanisms (1). Therefore, techniques that assess the function of the hippocampus globally and techniques that assess the function of an individual subregion in isolation are not ideally suited to pinpoint the primary source of dysfunction. By relying on approaches that map the functional integrity of the hippocampal subregions in living brains, we have isolated subregions differentially sensitive and resistant to the aging process.

Although age-related hippocampal dysfunction is observed in all mammalian species, rats are the most popular animal model of normal aging (32), and monkeys most closely resemble the human aging process (33) and human hippocampal formation (34). Both species were investigated in this study by using different but complimentary imaging techniques. The hippocampus of aging rhesus monkeys was investigated with an MRI measure of CBV. CBV was the first hemodynamic variable to be successfully measured with MRI technology (35), and a wide range of studies has established its intimate relationship to blood flow and brain metabolism (for examples, see refs. 15 and 16). Most relevant to this study, MRI measures of CBV have been found to linearly correlate with basal glucose metabolism (18), and CBV has successfully detected the effects of aging (36) and hippocampal dysfunction in Alzheimer's and other diseases (20, 21). Here we used an approach that generates CBV images with high spatial resolution (17), allowing us to visualize and analyze individual hippocampal subregions. Our first imaging study found that the dentate gyrus is the hippocampal subregion most sensitive to the effects of advancing age in rhesus monkeys.

Despite the ability to visualize individual hippocampal subregions in living brains, MRI does not possess cellular resolution. Furthermore, because it is inherently an indirect measure of brain metabolism, CBV, like any hemodynamic variable, is prone to misinterpretations (22). For example, an age-related change in vascular biophysics may lead to hemodynamic changes independent of underlying brain metabolism (37, 38). On both counts, directly imaging behaviorally induced Arc expression in hippocampal neurons compliments MRI. Arc expression effectively reports on spike activity (24, 25), and, therefore, its expression is a direct correlate of neuronal physiology. Furthermore, Arc expression is associated with long-term potentiation (23), a mechanism of plasticity linked to memory and impaired in age-related memory decline (1). Our second imaging study found that the dentate gyrus is the hippocampal subregion most sensitive to the effects of advancing age in rats.

By demonstrating a consistent pattern of age-related decline in the dentate gyrus, in two species that do not develop Alzheimer's disease, these results also act to confirm a prior human imaging study (14) and prior findings in aging rats (39). Thus, despite a variety of functional imaging techniques applied to a range of species, results converge on the same finding: The dentate gyrus is the hippocampal subregion most vulnerable to advancing age, whereas the pyramidal cell subregions are relatively spared, in rats, monkeys, and humans.

By identifying the neuronal population most sensitive to aging, these results set the stage for isolating the biochemical factors within granule cells that underlie the physiologic defect. Prior studies have observed, for example, a reduction in glutamate receptors in the granule cells of aging rhesus monkeys, a finding that can account for our findings (40). Nevertheless, by explicitly comparing molecular profiles of granule and pyramidal hippocampal neurons across the lifespan, future studies can begin to identify molecules most relevant to the aging process.

Our results, furthermore, clarify a distinction between normal aging and Alzheimer's disease, a pathological process that also targets the aging hippocampal formation. In contrast to aging, Alzheimer's disease causes clear histological changes, and the pattern of cell loss in Alzheimer's disease and other markers of disease have been used to identify vulnerable subregions of the hippocampal formation. The vast majority of studies have documented that the pyramidal cells of the hippocampal formation are most vulnerable, whereas the granule cells are relatively resistant, to Alzheimer's disease pathology (41–48). When changes have been observed in the dentate gyrus (49), they are postulated to occur secondarily to upstream lesions in the entorhinal cortex. Thus, by establishing differential patterns of hippocampal dysfunction, these studies confirm that aging and Alzheimer's disease are indeed distinct entities. Furthermore, imaging the functional integrity of individual hippocampal subregions is an effective method to detect the earliest stage of Alzheimer's disease, when it behaviorally overlaps with normal aging.