Cortisol levels during human aging predict hippocampal atrophy and memory
deficits
Sonia J. Lupien1, 2, Mony de Leon3, Susan de Santi3, Antonio Convit3, Chaim Tarshish3, N. P. V. Nair1, Mira Thakur1, Bruce S. McEwen4, Richard L. Hauger5
& Michael J. Meaney1
1 Aging Research Program, Douglas Hospital Research Center, Department of Psychiatry, McGill University, 6875 Boulevard Lasalle, Verdun (Québec), H4H-1R3, Canada
2 Human Psychoneuroendocrine Research Laboratory, Geriatric Institute of Montreal, 4565 Queen Mary, Montreal (Québec) H3W-1W5, Canada
3 Aging and Dementia Research Center, NY University Medical Center, 550 First Avenue, New York, New York 10016, USA
4 Laboratory of Neuroendocrinology, Rockefeller University, 1230 York Avenue, New York, New York 10021, USA
5 Department of Psychiatry, University of California and VA Medical Center, 3350 La Jolla Village Drive,
San Diego, California 92161, USA
Elevated glucocorticoid levels produce hippocampal dysfunction and
correlate with individual deficits in spatial learning in aged rats. Previously
we related persistent cortisol increases to memory impairments in elderly
humans studied over five years. Here we demonstrate that aged humans with
significant prolonged cortisol elevations showed reduced hippocampal volume
and deficits in hippocampus-dependent memory tasks compared to normal-cortisol
controls. Moreover, the degree of hippocampal atrophy correlated strongly
with both the degree of cortisol elevation over time and current basal cortisol
levels. Therefore, basal cortisol elevation may cause hippocampal damage and
impair hippocampus-dependent learning and memory in humans.
The hypothalamic-pituitary-adrenal system is highly sensitive to everyday
challenges in animals and humans1. Acute stress triggers release
of glucocorticoids, which then feed back onto specific brain regions to inhibit
further release2. In addition to pituitary and hypothalamic
sites, the limbic system, particularly the hippocampus, has been implicated
in the regulation of glucocorticoid activity2. Although glucocorticoid
responses to stress are essential for survival, prolonged glucocorticoid elevation
can present serious health risks, including diabetes, hypertension, hyperlipidemia,
hypercholesterolemia, arterial disease, amenorrhea, impairment of growth and
tissue repair and immunosuppression3.
Elevated glucocorticoids are linked with hippocampal pathology in aging
rodents4. Basal plasma corticosterone levels among aged rats
correlate significantly with hippocampal degeneration and spatial learning
deficits5,
6. Elevated plasma corticosterone levels are found
only in aged rats with spatial memory deficits and not in aged rats with normal
spatial memory7. Cumulative exposure to high glucocorticoid
levels throughout life disrupts electrophysiological function, leading to
atrophy and ultimately death of hippocampal neurons, all of which can cause
severe cognitive deficits4 in hippocampus-dependent learning
and memory8. Adrenalectomy at mid-life, with low-level glucocorticoid
replacement, attenuates hippocampal degeneration and cognitive decline in
rats5, suggesting that elevated glucocorticoid levels directly
contribute to the development of cognitive impairments. Together, these results
strongly suggest that glucocorticoid elevation partly accounts for individual
differences in age-related hippocampal damage and memory defects in rodents.
Recent studies in patient populations have suggested a similar relationship
in humans . In Cushing's syndrome, which causes prolonged cortisol elevation,
hippocampal volume correlates negatively with plasma cortisol levels, and
positively with scores on verbal memory tests9. Alzheimer's
patients show an inverse relationship between mean 24-hour cortisol levels
and severity of cognitive decline, which is associated with progressive hippocampal
degeneration10. Here we ask whether this relationship extends
to healthy elderly human subjects.
For five to six years, we measured basal plasma cortisol levels annually
over a 24-hour period in 51 aged healthy volunteers11. To estimate
the cumulative exposure to glucocorticoids, a simple regression analysis on
plasma cortisol levels for each subject was conducted using year as the independent
variable and the integrated 24-hour cortisol concentration as the dependent
variable. The slope of the regression line (termed "cortisol slope") was chosen
as the measure of the dynamic change in adrenal activity so as to differentiate
patients whose cortisol was decreasing from those whose cortisol was increasing
with time11. We found considerable variation in plasma cortisol
levels, as well as clear evidence for three subgroup patterns: progressive
increase in cortisol levels with currently high basal cortisol levels (termed
"increasing/high cortisol"), progressive increase in cortisol levels with
currently moderate cortisol levels ("increasing/moderate cortisol"), or progressive
decrease in cortisol levels with currently moderate cortisol levels ("decreasing/moderate
cortisol"). Both acute and chronic cortisol elevations induce cognitive deficits
in human populations12,
13.
We measured the environmental validity of our yearly laboratory cortisol
measures by taking salivary samples for cortisol analysis at the subject's
home, four times a day, over a 30-day period. These cortisol levels clearly
differentiated the subgroups over this longer period of time (
Int. Soc. Psychoneuroendocrinology Abs.
92,1996). In this study, subjects with increasing/high
cortisol also reported higher feelings of stress than the decreasing/moderate
cortisol group over the 30-day period. Finally, the increasing/high cortisol
group showed significant impairments in hippocampus-dependent forms of memory
compared to the other groups14. Performance on tests of hippocampus-independent
memory was similar for all three groups of subjects14. These
findings suggest that increased glucocorticoid levels can influence hippocampus-dependent
memory in aged humans. The present study tests whether prolonged cortisol
elevation and memory impairment in normal elderly humans correlate with a
significant decrease in hippocampal volume, as reported in the animal literature.
Results We performed magnetic resonance imaging (MRI) on a subgroup of subjects
from the increasing/high cortisol group, and the decreasing/moderate cortisol
group. Given the increased variability in cortisol secretion and cognitive
function during human aging11,
12,
14, we used these two extreme
groups to assess the magnitude of the difference in hippocampal volume in
conditions of normal versus impaired glucocorticoid activity during human
aging. As the hippocampus is implicated in performance of several other cognitive
tasks8, particularly those sensitive to the time-limited15 and spatial16 aspects of memory, we also measured
these two groups on an immediate versus delayed memory task and on spatial
memory with a human maze.
Memory Using a repeated measures ANOVA with group as the between-subjects factor
and immediate and delayed memory as the repeated measure variable, we observed
a main effect for group [F(1,7) = 7.9, p<0.05]
and a significant interaction effect [F(1,7) = 17.5, p<0.01
]. Simple effects performed on this interaction revealed that the increasing/high
cortisol group showed significant impairments on delayed recall (p<0.02;
see Fig. 1), although the groups did not differ
on tests of immediate memory.
Figure 1. Mean ( SEM) correct recall of the increasing/high and decreasing/moderate
cortisol groups on the immediate and delayed memory task.
Using a similar model for performance on the spatial memory test with time
[log(sec)] required to follow the simple versus complex path as the repeated
measure variable, we found main effects for both group [F(1,8) = 8.4,
p<0.05] and path complexity [F(1,8) = 64.6, p<0.001
] with no interaction between these two variables. The increasing/high
cortisol group took significantly longer to recall and follow either the simple
or the complex path compared to subjects from the decreasing/moderate cortisol
group (Fig. 2).
Figure 2. Mean (SEM) time [log(sec)] for the increasing/high and decreasing/moderate
cortisol groups to find their way through a human maze for a simple and a
complex path.
Hippocampal volume We compared the increasing/high cortisol and the decreasing/moderate cortisol
groups over the volume of the hippocampus, parahippocampal gyrus, fusiform
gyrus, the middle-inferior, and the superior temporal gyri. Two findings were
noteworthy. First, the total hippocampal volume of the increasing/high cortisol
group was significantly reduced by 14% in comparison to that of decreasing/moderate
cortisol group [t(9) = 25.1, p<0.001
, Table 1 and Fig. 3
]. Second, this effect was unique to the hippocampus; we found
no group differences in the volume of the parahippocampal and fusiform gyri,
nor in the other temporal lobe structures analyzed (Table
1).
Figure 3. Reformatted (perpendicular to the long axis of the hippocampus) coronal
T1-weighted gradient echo coronal MRI image of the hippocampus (arrows on
left side) at the level of the lateral geniculate body for representative
increasing/high and decreasing/moderate subjects.
Correlations among measures The relationship between cortisol and hippocampal volume was examined by
regression analyses. A significant correlation was found (r
= -0.80, r2 = 0.64, p<0.01
; see Fig. 4) between cortisol slope and
hippocampal volume (left+right sum/hemi sum), as well as a significant correlation
(r = -0.68, r2 = 0.46,
p<0.02) between current cortisol measure and hippocampal volume.
Figure 4. Correlation between cortisol slope and hippocampal volume in 11 elderly
human subjects.
To compare these results to other studies that have linked hippocampal
volume and delayed memory performance17, we examined our data
in a set of hierarchical regression models. No significant correlations were
found between the principal measures (brain regions and cortisol slope) and
either age, gender or education (p>0.1). Consequently, these measures
were not included in our regression models as covariates. In the first step,
we covaried out the effect of immediate memory as a baseline measure. We then
entered either the volume or slope measurement. Of these two measures, only
the hippocampal volume showed a significant relationship to delayed memory
(Fchange(1,6) = 5.8, r2change
= 0.30, p<0.05). We further tested the anatomic specificity
of the hippocampus relative to the other temporal lobe regions. The middle
inferior temporal lobe was the only other region to be significantly related
to delayed memory (Fchange (1,6) = 7.9, r2change = 0.34, p<0.05). In a final set of
regression models, we tested the unique variance attributed to the hippocampus
and middle inferior temporal lobe relative to delayed memory. In this model,
we entered immediate memory and either the hippocampus or middle inferior
temporal lobe in step 1 and the second region in step 2. We found that although
each region accounted for a significant proportion of variance in delayed
memory, neither added an appreciable amount of variance to the other region.
We then examined the relationship of cortisol slope to each of the temporal
lobe regions separately. Only the hippocampus exhibited a relationship to
cortisol slope (r = -0.80, p<0.01
, two-tailed).
Discussion The present study revealed that elderly human subjects showing increasing
cortisol levels over years with currently high cortisol levels are impaired
on hippocampus-dependent memory and have a 14% reduction in hippocampal volume,
compared to elderly subjects showing decreasing cortisol levels over years
with currently moderate cortisol levels. These results are in accordance with
animal studies showing that cumulative exposure to high glucocorticoids has
functional and structural effects on the hippocampus6,
7,
8.
Studies of brain changes associated with dementia in later life show an anatomically
specific relationship between hippocampal volume and memory. These observations
have been extended to elderly populations showing mild cognitive impairments
(MCI)18. Interestingly, the magnitude of the decrease in hippocampal
volume in the increasing/high cortisol group was comparable to that previously
reported for elderly subjects with age-related MCI18. As our
subjects were originally selected on the basis of differences in cortisol
secretion11, rather than on differences in cognitive performance,
we suggest that increases in cortisol secretion in later life may initiate
MCI and/or MCI-related hippocampal atrophy.
Our findings are consistent with the idea that exposure to progressively
elevated glucocorticoid levels eventually compromise hippocampal integrity
and thus performance on hippocampus-dependent cognitive tasks7,
8.
The decrease in hippocampal volume was correlated with both the current basal
cortisol levels and the cortisol slope, which reflected changes in plasma
cortisol levels over the past five years. Hence, hippocampus-dependent cognitive
impairments are associated with a profile of progressively increasing basal
cortisol and a currently elevated cortisol level.
The hippocampus is not only a target for glucocorticoids but is also involved
in their regulation2,
4. Lesions to the hippocampus are associated
with increased basal glucocorticoid levels19, and the hippocampus
has been implicated in regulating glucocorticoid release during stress20. Thus, hippocampal atrophy is both a result of and a contributory
cause of elevated basal glucocorticoid levels. This is embodied in the glucocorticoid
cascade hypothesis for hippocampal aging21, which proposes that
hippocampal damage leads to increased circulating glucocorticoid levels, which
in turn worsen the degree of hippocampal damage.
The effects of glucocorticoids on the hippocampus depend on concentration4. Normal basal levels of glucocorticoids facilitate hippocampal plasticity22 and promote the survival of dentate gyrus granule cells23,
whereas elevated glucocorticoid levels are clearly associated with hippocampal
dysfunction. These apparently contrasting effects can be understood in terms
of the known differences in corticosteroid receptor subtypes24.
Mineralocorticoid receptors bind glucocorticoids with a five- to tenfold higher
affinity than do glucocorticoid receptors24. Basal glucocorticoid
levels act on the brain via mineralocorticoid receptors, whereas the compromising
effects involve activation of a high percentage of glucocorticoid receptors4. Normal basal cortisol levels, such as those seen in the decreasing/moderate
cortisol group, primarily activate mineralocorticoid receptors and facilitate
hippocampal long-term potentiation22, a synaptic model of memory.
In contrast, the elevated cortisol levels of the increasing/high cortisol
group, which correspond to the normal circadian peak value and even approximate
those levels seen during stress, would activate a larger proportion of glucocorticoid
receptors, serving not only to negate the effects of mineralocorticoid receptor
activation4 but also to promote the debilitating glucocorticoid
receptor effects on hippocampal function, including a dampening of long-term
potentiation25,
26.
We propose that decreased hippocampal volume is associated with both the
current high cortisol level and the increases in cortisol levels over the
years. As in the rodent hippocampus, prolonged exposure to elevated glucocorticoid
levels, along with elevated excitatory amino acid activity, could directly
contribute to the decrease in hippocampal volume observed in the increasing/high
cortisol group. Glucocorticoids serve to enhance calcium-dependent afterhyperpolarization
in hippocampal neurons, which dampens responses to excitatory input27.
Chronic elevations in circulating glucocorticoid levels are associated with
decreased long-term potentiation or primed-burst potentiation in rat hippocampus,
and these effects are reversed by adrenalectomy or by administration of a
glucocorticoid receptor antagonist22,
27. Sustained exposure
to elevated glucocorticoid levels or repeated psychological or psychosocial
stress also produces atrophy of hippocampal pyramidal neurons in the rat and
tree shrew, and these effects are attenuated by NMDA receptor blockers and
by phenytoin, which blocks sodium and T-type calcium channels28.
In addition, chronic stress, which persistently elevates glucocorticoid levels,
attenuates the normally occurring neurogenesis in hippocampal dentate gyrus
of adult monkeys, a process that is thought to be necessary to sustain a constant
level of neuron density in this region29. Finally, in the extreme,
glucocorticoids have been shown to enhance hippocampal neuron loss following
treatment with excitotoxins, and again this effect has been associated with
increases in intracellular calcium levels30. Together, these
various effects could account for the decreased hippocampal volume observed
in the increasing/high cortisol group. It remains to be seen whether hippocampal
volume reduction in these subjects reflects a reversible atrophy or an irreversible
neuronal loss and whether these hippocampal changes will lead to further impairments
of cognitive function, including dementia, later in life.
Methods Population. Subjects for the Study for Aging of the Douglas Hospital in Montreal are
solicited from ads in the local media. The medical status of each subject
is determined annually by a complete physical examination including ECG, EEG,
CAT scan, and a battery of laboratory tests for kidney, liver, and thyroid
functions, hemogram, vitamin B12, folate levels, as well as a neuropsychological
assessment (for a complete description of these data, see refs. 11 and 14). Informed consent is obtained
from all subjects. Our previous studies demonstrate that there is no change
in the circadian rhythm nor CBG levels in these subjects11,
nor are there any differences between men and women with regard to cortisol
history or any other variables tested11.
The sample used in the present study was composed of 11 subjects. Six of
them were from the increasing/high cortisol group (mean age: 76.5
4.3; mean education level: 10.5 2.3 years; current cortisol levels
at the time of MRI testing: 12.8 g/dl 3.1; cortisol slope: 0.56 g/dl/year).
The other five subjects were from the decreasing/moderate cortisol group (mean
age: 70.8 7.2; mean education level: 12.0 4.3 years; current
cortisol levels at the time of MRI testing ; 9.1 g/dl 2.9; cortisol
slope: -0.95 g/dl/year). There were no significant differences between
age or education level in the two groups (p>0.1). The increasing/high
cortisol group showed significantly higher current cortisol levels (p<0.05)
and a significantly higher cortisol slope over years (p<0.05) compared
to the decreasing/moderate cortisol group.
Cortisol measurement. Subjects are tested annually for plasma cortisol levels over a continuous
24-hour period with sampling each hour. Blood samples are centrifuged at 2500
rpm for 10 min at 0-40°C, frozen, and stored at -200°C until assayed.
The validation of the subgroups has been described in detail11,
and subjects were selected based on the groupings we described14.
For plasma cortisol samples, a 300 l aliquot of the extract was assayed
in duplicate using [3H]cortisol as tracer and a highly specific
cortisol antibody (B-63 from Endocrine Sciences, Tarzana CA). This antibody
cross-reacts less than 4% with deoxycortisol or deoxycorticosterone, and less
than 0.5% with other adrenal steroids. Intra- and inter-assay variability
are 4% and 6% respectively.
Measures of Memory. Immediate and delayed memory was measured by presenting 15 non-complex
line drawings of everyday objects taken from a standardized set of pictures
controlled for name agreement, image agreement, familiarity, and visual complexity.
The set of images was moderately prototypic [21.56 4.52], and the
characteristics of the images extracted from the Snodgrass and Vanderwart
standardized set were as follows : Name Agreement, 90.33% 8.51; Image
Agreement, 3.78 0.67; Familiarity, 3.780.89; Complexity, 2.770.82.
The image set was developed by I. Lussier, Univ. Montreal (1992). The subject
was presented with the 15 line drawings for 3 seconds each and asked to name
the object. Subjects were then asked to verbally recall as many line drawings
as possible, immediately after the presentation or 24 hours later. The number
of pictures correctly recalled on each occasion served as the dependent measure.
One subject from the increasing/high cortisol group and one from the decreasing/moderate
cortisol group were not available for 24-h delayed memory testing, so results
of analyses (see below) are based on the remaining nine subjects.
Spatial memory function was measured in the same individuals using a human
maze that allows for control of the difficulty level of the task and limits
extraneous perceptual factors, which could interfere with the measure of spatial
cognition. This human maze was designed by Dr. Romedi Passini from the Dept.
of Architecture at the University of Montreal. The surface area of the maze
was 1,500 square feet and the walls were 6 feet high, with no extraneous cues,
either on the floor or on the ceiling. The passages corresponded to a small
domestic corridor of one meter in width. The subject was shown a path by following
the experimenter through the maze and was then required to reproduce the path
on his/her own by walking through the maze. The subjects had to learn a simple
and a complex path. The complexity of the path was determined by the number
of decision points in the path. A decision point is an intersection in the
maze at which the subject must make a decision (turn left, right or go straight
ahead). The simple path required three decisions, and the complex path required
five decisions. All subjects learned the notion of a point of decision using
a smaller maze of 500 square feet that was built beside the first maze, which
also served to reduce the novelty of the procedure. The time taken to find
the correct path served as the measure of spatial memory function. Subjects
presented equivalent walking pace when measured on a pilot study, using the
smaller human maze. One subject from the decreasing/moderate cortisol group
needed a cane for walking and was thus not tested on the human maze. The results
of the analyses are thus based on the remaining 10 subjects.
MRI measurements. We acquired 124 sagittal T1-weighted (TR = 24, TE = 5, FA = 45) gradient
echo images using a 1.5 GE Sigma imager. The MRI volumetric method is similar
to that reported24. Sagittal images were acquired with a 24
cm field of view using a slice thickness of 1.2 mm without gaps and a 256
X 192 matrix. Scans were reformatted at a 2 mm slice thickness in the coronal
plane for the anatomical work. In reformatting the images, we identified the
anterior-posterior plane corresponding to the band of gray matter at the junction
of the left hippocampus and parahippocampal gyrus and set the plane of coronal
acquisition perpendicular to this. Using MIDAS image analysis software (Tsui,
M.H. Multimodal Image Data Analysis System (MIDAS). Version 1.0, unpublished
manual, 1995) developed at NYU in-house with a Sun Sparc workstation (Sun
Microsystems, Mountain View, California), we drew regions of interest in threefold
enlarged images. All the image analyses were performed blind to group membership.
We outlined the temporal lobe structures with a mouse-driven cursor and then
excluded the pixels that fell in the lowest 33% of the intensity range between
gray matter and cerebrospinal fluid. Tissue volumes were then estimated by
counting the numbers of remaining parenchymal pixels (of known size) over
the slices measured. In this study, for all regions, the anterior limit was
the body of the amygdala, and the posterior limit was the crus of the fornix
adjoining the splenium of the corpus callosum. The hippocampus and the head
size were sampled every 2 mm, and in most cases this included 12 to 15 coronal
sections. The other regions, which were sampled on every other slice (4 mm),
included parahippocampal gyrus, fusiform gyrus, the middle-inferior, and the
superior temporal gyri. To standardize our measurements of the temporal lobe
regions, we used a reference point lateral to the hippocampus, in the middle
of the temporal horn, as an "anchor". From this reference point, radial lines
were drawn to the deepest point of the sulci and then extended around the
surface to define the relevant gyri. All data were corrected for head size.
The head-size estimate, used as a surrogate for the premorbid brain size,
was determined by tracing the outline of the supratentorial compartment following
the dural and tentorial and midline margins. All data are calculated as the
ratio of the region of interest to the head-size correction.
McEwen, B.S. Protective and damaging effects of stress mediators. New Engl. J. Med. 238, 171−179 (1998) | Article |
Jacobson, L. & Sapolsky, R.M.The role of the hippocampus in feedback regulation of the hypothalamic-pituitary-adrenal axis. Endo. Rev. 12, 118−134 (1991) | ISI | ChemPort |
Brindley, D.N. & Rolland, Y.Possible connections between stress, diabetes, obesity, hypertension and altered lipoprotein metabolism that may result in atherosclerosis. Clin. Sci.77, 453−461 (1989) | PubMed | ISI | ChemPort |
Sapolsky, R. Stress, the aging brain, and the mechanisms of neuron death. (MIT Press, Cambridge, 1992)
Landfield, P., Baskin, R. K. & Pitler, T. A.Brain-aging correlates: Retardation by hormonal-pharmacological treatments. Science214, 581−583 (1981) | PubMed | ISI | ChemPort |
Issa, A., Gauthier, S. & Meaney, M. J.Hypothalamic-pituitary-adrenal activity in aged cognitively impaired and cognitively unimpaired aged rats. J. Neurosci. 10, 3247−3254 (1990) | PubMed | ISI | ChemPort |
Squire, L. R.Memory and the hippocampus : A synthesis from findings with rats, monkeys, and humans. Psychol. Rev. 99, 195−231 (1992) | Article | PubMed | ISI | ChemPort |
Starkman, M. N., Gebarski, S. S., Berent, S. & Schteingart, D. E.Hippocampal formation volume, memory dysfunction, and cortisol levels in patients with Cushing's syndrome. Biol. Psychiatry32, 756−765 (1992) | Article | PubMed | ISI | ChemPort |
Davis, K. L. et al. Cortisol and Alzheimer's disease, I : Basal studies. Am. J. Psychiatry143, 300−305 (1986) | PubMed | ISI | ChemPort |
Lupien, S. J. et al. Longitudinal study of basal cortisol levels in healthy elderly subjects : Evidence for sub-groups. Neurobiol. Aging17, 95−105 (1995) | Article | ISI |
Lupien, S. J.. & McEwen, B. S.The acute effects of corticosteroids on cognition: Integration of animal and human model studies. Brain Res. Rev.24, 1−27 (1997) | Article | PubMed | ISI | ChemPort |
Wolkowitz, O. M., Reus, V. I., Canick, J., Levin, B. & Lupien, S.J. Glucocorticoid medication, memory and steroid psychosis in medical illness. Ann. NY Acad. Sci. 823, 81−96 (1997) | PubMed | ChemPort |
Lupien, S. et al. Basal cortisol levels and cognitive deficits in human aging, J. Neurosci. 14, 2893−2903 (1994) | PubMed | ISI | ChemPort |
Scolville, W. B. & Milner, B. Hippocampal damage and memory dysfunction. J. Neurol. Psychiatry20, 11−21 (1957)
O'Keefe, J. & Nadel, L. The Hippocampus as a Cognitive Map (Oxford, Clarendon, 1978)
de Leon, M. J. et al. Frequency of hippocampal formation atrophy in normal aging and Alzheimer's disease. Neurobiol. Aging18, 1−11 (1997) | Article | PubMed | ChemPort |
de Leon, M. et al. Abnormal cortisol response in Alzheimer's disease linked to hippocampal atrophy. Lancet2, 391−392 (1988) | Article | PubMed | ChemPort |
Fischette, C. T., Komisurak, B. R., Ediner, H. M., Feder, H. H. & Siegal, A.Differential fornix ablations and the circadian rhythmicity of adrenal corticosterone secretion. Brain Res. 195, 373−380 (1980) | Article | PubMed | ISI | ChemPort |
DeKloet, E. R.Brain corticosteroid receptor balance and homeostatic control. Front. Neuroendocrinol.12, 95−123 (1991) | ISI |
Sapolsky, R. M., Krey, L. C. & McEwen, B. S.The neuroendocrinology of stress and aging: The glucocorticoid cascade hypothesis. Endo. Rev. 7, 284−301 (1986) | ISI | ChemPort |
Joels, M. Steroid hormones and excitability in the mammalian brain. Front. Neuroendocrinol.18, 495−4498 (1997)
Sloviter, R. et al. Selective loss of hippocampal granule cells in the mature rat brain after adrenalectomy. Science243, 535−538 (1989) | PubMed | ISI | ChemPort |
Reul, J. M. H. M. & De Kloet, E. R.Two receptor systems for corticosterone in rat brain: microdistribution and differential occupation. Endocrinology117, 2505−2512 (1985) | PubMed | ISI | ChemPort |
Kerr, D. S., Campbell, L. W., Hao, S. -Y. & Landfield, P. W.Corticosteroid modulation of hippocampal potentials: Increased effect with aging. Science245, 1505−1509 (1989) | PubMed | ISI | ChemPort |
Diamond, D. M., Bennett, M. C., Fleshner, M. & Rose, G. M.Inverted-U relationship between the level of peripheral corticosterone and the magnitude of hippocampal primed burst potentiation. Hippocampus2, 421−430 (1992) | PubMed | ISI | ChemPort |
Joels, M. & De Kloet, E. R.Effect of glucocorticoids and norepinephrine on excitability in the hippocampus. Science245, 1502−1504 (1989) | PubMed | ISI | ChemPort |
Magarinos, A. M., McEwen, B. S., Flugge, G. & Fuchs, E.Chronic psychosocial stress causes apical dendritic atrophy of hippocampal CA3 pyramidal neurons in subordinate tree shrews. J. Neurosci. 16, 3534−3540 (1996) | PubMed | ISI | ChemPort |
Gould, E., Tanapat, P., McEwen, B. S., Flugge, G. & Fuchs, E.Proliferation of granule cell precursors in the dentate gyrus of adult monkeys is diminished by stress. Proc. Natl. Acad. Sci. USA95, 3168−3171 (1998) | Article | PubMed | ChemPort |
Landfield, P. W. & Elridge, J. C.Evolving aspects of the glucocorticoid hypothesis of brain aging: Hormonal modulation of neuronal calcium homeostasis. Neurobiol. Aging15, 579−588 (1994) | Article | PubMed | ISI | ChemPort |
Acknowledgments This research was supported by grants from the American Alzheimer's Association
and the National Institute of Aging (AG09488) to MJM, by a fellowship from
the Fonds de la recherche en santé du Québec (FRSQ) to SJL,
and was supported in part by NIH grants to MdeL (AG12101, AG13616), and as
a part of the NIA Alzheimer's disease core center grant (P30 AG08051). It
was also supported by a grant from the John D. and Catherine T. MacArthur
Foundation. RLH was a recipient of the VA Clinical Investigator career development
award and partially supported by the UCSD NIMH CRC (PHS MH30914-14). The Aging
Research Program of the Douglas Hospital is generously supported by ALCAN
Canada Ltd.
SEM) correct recall of the increasing/high and decreasing/moderate
cortisol groups on the immediate and delayed memory task.
g/dl 
