[1] 13w Age-Dependent Loss of Corticosterone Modulation of Central Serotonin 5-HT 1A Receptor Binding Sites

[1] 259w A loss of endocrine and neurotransmitter system interactions, including corticosterone regulation of 5-HT 1A receptors, may underlie the age-related deficits in the hypothalamic-pituitary-adrenal (HPA) axis including adapting to stress. In this study, female Fischer 344 rats, (ages 3, 13, and 18 months), were bilaterally adrenalectomized and supplemented for 3 weeks with placebo or corticosterone (200 mg or 600 mg) containing 21 day sustained-release pellets implanted subcutaneously (LC, MC, or HC, respectively). Scatchard analysis using the 5-HT 1A receptor agonist [ 3 H]8-hydroxy-2-(di-N-propylamino) tetralin (8-OH-DPAT) demonstrated a significant decrease in hippocampal receptor density in 3 month and 13 month MC groups (-35.2 and -32.1%, respectively) as compared to age-matched LC groups; a significant decline in 5-HT 1A receptor density in 3 month and 13 month HC groups was found compared to agematched MC groups (؊16.7 and ؊22.0%, respectively). However, these hormone treatments (LC or HC) failed to alter hippocampal 5-HT 1A binding site density in the 18 month groups. Cortical 5-HT 1A receptor densities were altered in a similar agedependent manner. In contrast, the density of hypothalamic 5-HT 1A receptors in the 18 month LC group was significantly increased above that in the 3 month LC group. An additional indicator of the hippocampal response to corticosterone, the distribution of glial fibrillary acidic protein (GFAP), revealed an agerelated decline in responsiveness to hormone treatment in the oldest group. The present study has identified an age-associated deficit in the regulation of hippocampal 5-HT 1A receptors by corticosterone which may underlie the diminished capacity of the aging HPA axis to cope with stress.

[1] 207w Senescent individuals have difficulty adapting to stressors, such as changing environments (Van Eekelen et al., 1995) and surgical procedures (Blichert-Toft et al., 1975), due to impairment of the hypothalamic-pituitaryadrenocortical (HPA) axis (Sapolsky et al., 1986;Hauger et al., 1994;Cizza et al., 1994). The serotonin (5hydroxytryptamine, 5-HT) neuronal system plays a role in regulating the HPA axis response to stress (Lorens and Van de Kar, 1987;Calogero et al., 1990) and modulates the release of stress hormones, including corticosterone. Corticosterone mediates the physiological need for increased availability and utililization of energy in stressful situations and also terminates this homeostatic response via negative feedback at multiple brain regions. Sites of feedback regulation include the hypothalamus and hippocampus. Actions of corticosterone have been characterized at both mineralocorticoid (MR) and glucocorticoid receptors (Jo ¨els and De Kloet, 1990;Jo ¨els and De Kloet, 1992). Moreover drugs which increase brain 5-HT function also elevate serum corticosterone levels (Fuller et al., 1974;Fuller 1992). Serotonergic neurons also have direct synaptic contacts with corticotrophin releasing factor (CRF)-containing neurons located in the paraventricular nucleus of the hypothalamus (Fuller et al., 1976). Thus, the serotonergic system is poised to modulate the release of CRF (Gibbs et al., 1983) and participate in the regulation of the HPA axis response to stress.

[2] 93w Among the 5-HT receptors subtypes, the 5-HT 1A receptor plays the major role in stress-mediated responses. Administration of the 5-HT 1A receptor agonist 8-hydroxy-2-(di-N-propylamino) tetralin (8-OH-DPAT) increased plasma corticosterone levels and pretreatment with 5-HT 1A receptor antagonists blocked this effect (Gilbert et al., 1987;Lorens and Van de Kar, 1987). The 5-HT 1A receptor agonist-induced release of corticosterone was also potentiated by the administration of agents which depleted central 5-HT stores, such as p-chlorophenylalanine, and thus revealed an increased sensitivity of these receptors in potentiating the release of stress homones (Kelder and Ross, 1992).

[3] 258w The serotonergic system has also been shown to undergo significant alterations with aging. Enhanced 5-HT turnover (Moretti et al., 1987) and a reduction in 5-HT affinity (Brunello et al., 1988) have been reported. Relevant to the HPA axis, 5-HT receptors in the hypothalamus undergo changes in affinity with advanced age (Halpern et al., 1989). The density of the 5-HT 1A receptor subtype declines with aging in human cortical and hippocampal postmortem tissues (Burnet et al., 1994;Dillon et al., 1991). Functional declines with advancing age have also been identified for the 5-HT 1A receptormediated regulation of central circadian rhythms in the rat (Penev et al., 1995). Normally, enhanced levels of circulating corticosterone serve to terminate responses to stress through activation of GR in the hippocampus and other brain regions. The loss of hippocampal GR and the rise in basal plasma levels of corticosterone found with aging (Landfield et al., 1978;Eldridge, 1989;Sapolsky, 1991) have demonstrated an age-related deficiency in feedback processes. Thus, the interaction of glucocorticoids hormones and specific neurotransmitter systems in the hippocampus, including the 5-HT 1A receptor, may contribute to the decline in hippocampal GR function observed with age (Eldridge et al., 1989). Found in high density in the hippocampus (Palacios et al., 1990), the 5-HT 1A receptor upregulates density under conditions of low circulating levels of corticosterone (Mendelson et al., 1992) and, conversely, downregulates with high levels of circulating corticosterone in young animals (Mendelson et al., 1992). A possible disregulation of this corticosterone/ 5-HT 1A receptor interaction may underlie the increased vulnerability of the hippocampus with aging.

[4] 111w The present studies examined the age-associated ability of circulating levels of corticosterone to alter 5-HT 1A reeptor binding site characteristics. Adrenalectomized female Fischer 344 rats, ages 2, 12, and 17 months, were supplemented with subcutaneously implanted placebo or corticosterone containing pellets for a 3 week period. Radioligand assays were used to determine the density (B max ) and affinity (K d ) for 5-HT 1A receptor binding sites in the hippocampus, frontal cortex, and hypothalamus across age and hormone treatment. Additionally, the distribution of hippocampal glial fibrillary acidic protein (GFAP) immunoreactivity was assessed to monitor corticosterone-and age-dependent changes in reactive gliosis as a marker of neuronal damage in this brain region.

[1] 125w Initially, a model in which surgical treatment groups consisted of intact animals, sham ADX and bilateral ADX animals was used to provide different levels of circulating plasma corticosterone. However, plasma corticosterone levels with a 7-day hormone exposure period were highly variable within treatments (see Materials and Methods) and across ages (data not shown). The high degree of variability in the plasma corticosterone levels resulted in overlapping 5-HT 1A receptor density determinations between equivalent treatment groups across age. For ex-ample, sham ADX 5-HT 1A receptor densities in the hippocampus were 174.0 Ϯ 9.8, 174.8 Ϯ 12.3, and 183.2 Ϯ 7.3 fmol/mg protein (mean Ϯ SEM) for 3, 13, and 18 month groups, respectively. These results demonstrated the need for a more well-regulated paradigm of hormone exposure.

[2] 188w In the model of hormone replacement used for the present study, commercially manufactured sustainedrelease corticosterone-containing pellets were used to deliver a constant concentrationl of corticosterone exposure during a 21 day maintenance period following ADX at each age (Table I). It should be noted that low levels of plasma corticosterone (1-3 µg/dL) indicate a residual source of adrenal hormone following ADX in the female Fischer 344 rat. In LC groups low levels of circulating plasma corticosterone were found at each age (Table I). The MC treatment resulted in significant two-to four-fold increases in circulating plasma corticosterone as compared to LC groups (Table I). Likewise, the HC treatment resulted in a two-to three-fold increase in circulating plasma corticosterone levels as compared to levels in the MC groups (Table I). The levels obtained via LC, MC, and HC treatment groups are analogous to low, moderate, or normal physiological and high or stress levels of plasma corticosterone (Sapolsky et al., 1985). The corticosterone replacement animal model utilized herein provided sustained levels of circulating plasma corticosterone across the lifespan, encompassing periods of reproductive fitness and senescence (Wise et al., 1983;Finch et al., 1984).

[3] 14w While circadian variations in corticosterone concentrations are not present with this model of sustained

[4] 176w TABLE I. Plasma Corticosterone Levels in Aging Female Fischer 344 Rats Following Bilateral Adrenalectomy (ADX) and Treatment With 21 Day Sustained-Release Placebo (LC), 200 mg Corticosterone (MC), or 600 mg Corticosterone (HC) Containing Pellets, Implanted Subcutaneously a Age Plasma corticosterone levels (µg/dL) LC MC HC 3 months 2.3 Ϯ 1.1 12.1* Ϯ 2.1 22.0* Ϯ 1.4 13 months 3.7 Ϯ 1.8 13.9* Ϯ 2.3 -18 months 3.4 Ϯ 1.9 9.3* Ϯ 0.5 23.0* Ϯ 1.5 a No significant difference was found in circulating plasma corticosterone levels within treatment groups across age. Significant increases in circulating plasma corticosterone levels were observed in MC groups compared to LC groups. Significant increases in circulating plasma corticosterone levels were also observed in the HC groups compared to MC groups. Values for 13 month HC group were not available. Two-way ANOVA followed by Student's t-test was used for comparisons between treatment and across age groups. Values (µg/dL; mean Ϯ SEM represent data from five to seven animals per treatment group. *Significantly different from other age-matched corticosterone treatment groups (P Ͻ 0.05).

[5] 83w hormone exposure, the hippocampal 5-HT 1A receptor mRNA has been reported to be independent of circadian variation in circulating corticosterone levels (Holmes et al., 1995). However, to avoid any circadian variation in hormone or neurotransmitter function a constant time of day (1100 hrs) was used for obtaining CNS tissue. As determined by vaginal lavage, this model of ADX and corticosterone hormone replacement disrupted estrous cycles in young and middle-aged groups and thus resulted in an equivalent ovarian hormone environment at each age tested.

[6] 105w The results demonstrated a decrease in the ability of circulating plasma corticosterone to modulate hippocampal 5-HT 1A receptor density with aging. The density (B max value) of 5-HT 1A receptors in MC groups did not vary in the hippocampus across age (Fig. 1A). In comparison to the placebo-treated groups, age-matched MC groups showed a significant decline in the density of 5-HT 1A receptors in the 3 month (35.2%; P Ͻ 0.05) and 12 month (32.1%; P Ͻ 0.05) groups (Fig. 1A). However, in the 18 month MC group, no change in receptor density was observed as compared to the age-matched placebo treated group (Fig. 1A).

[7] 356w In treatment groups receiving high levels of corticosterone exposure (HC groups), a significant decrease in the density of 5-HT 1A receptors was found in the 3 month group (-16.7%; P Ͻ 0.05) and in the 13 month group (-22%; P Ͻ 0.05; Fig. 1A) in comparison with agematched MC groups. However, in the 18 month HC group, no change in receptor density was observed as compared to the 18 month MC group (Fig. 1A). Comparing hippocampal HC groups vs. LC groups, significant decreases in density were observed in the 3 and 13 month Fig. 1. Hippocampal 5-HT 1A receptor density and affinity in adrenalectomized 3, 13, and 18 month female Fischer 344 rats supplemented with 21 day release placebo (LC), 200 mg (MC), or 600 mg (HC) corticosterone-containing pellets implanted subcutaneously. A: The graph of hippocampal 5-HT 1A receptor density shows MC groups underwent significant decreases in receptor density compared to age-matched LC groups and HC groups underwent significant decreases as compared to agematched MC groups in the 3 and 13 month age groups; however, no changes were observed between treatments in the 18 month group. B: A significant increase in hippocampal 5-HT 1A receptor affinity in the 13 month MC group was observed as compared to the age-matched LC group. Also, significant increases in receptor affinity in the 3 and 13 month HC groups compared to age-matched MC groups. No changes in receptor affinity were observed across hormone treatment in the 18 month groups. Bars represent B max (fmol/mg protein; mean Ϯ SEM) and K d (nM; mean ϩ SEM) values from five to seven animals per group at each age (white, LC; hatched, MC; gray, HC). A two-way ANOVA followed by Student's t-test for planned comparisons was used for analysis between treatment and across age groups. *Significant difference (P Ͻ 0.05) between LC and MC treatments of same age group. # Significant difference (P Ͻ 0.05) between MC and HC treatments of same age group. Significant difference (P Ͻ 0.05) between LC and HC treatments of the same age. Significant difference (P Ͻ 0.05) between 18 month and younger groups of the same treatment.

[8] 112w groups (-38.4% and -40.5%, respectively; P Ͻ 0.05; Fig. 1A) and absent in the 18 month group. In the hippocampus there was no change in 5-HT 1A receptor density between the 3 and 13 month LC as well as no change between the 3 and 13 month HC groups. However, a significant age-dependent decrease in the density of 5-HT 1A receptors in the 18 month LC group compared to the 3 month LC group (21.1%; P Ͻ 0.05) as well as a significant increase in the density of 5-HT 1A receptors in the 18 month HC group compared to the 3 month HC group was observed (17.6%; P Ͻ 0.05; Fig. 1A).

[9] 137w Hippocampal 5-HT 1A receptor affinity did not change between 3 and 18 month MC and age-matched LC groups. The affinity for this receptor was significantly increased (33%; P Ͻ 0.05) in the 13 month MC group in comparison with the age-matched LC group (K d ; Fig. 1B). Additionally, hippocampal 5-HT 1A receptor affinity was significantly increased in the 3 and 13 month HC groups in comparison with the age-matched MC groups (39 and 56%, respectively; P Ͻ 0.05, Fig. 1B). However, this increase, as were all of the K d values obtained in this study of 5-HT 1A receptor affinity corresponded to the high affinity (nM) state of the 5-HT 1A receptor. In contrast to younger groups, no change in K d values were found between hormone treatments in the 18 month age groups (Fig. 1B).

[10] 112w The results demonstrated a limited dependence of cortical 5-HT 1A receptor density on varied levels of circulating plasma corticosterone with aging. The density of cortical 5-HT 1A receptors in MC treatment groups did not vary across age (Fig. 2A). No changes in receptor density were observed across hormone treatment groups in the 3 and 18 month age groups. In the 13 month group, the MC treatment produced a significant decrease in 5-HT 1A receptor density compared to the age-matched LC group (Ϫ16.7%; P Ͻ 0.05, Fig. 2A). Also the HC group showed a significant decrease in 5-HT 1A receptor density compared to the age-matched MC group (Ϫ16.1%; P Ͻ 0.05, Fig. 2A).

[11] 126w In the frontal cortex, a marked dependence on circulating corticosterone levels in maintaining 5-HT 1A receptor affinity was observed. In addition, an agedependent effect of corticosterone on frontal cortex affinity was observed for this receptor subtype. Across all age groups, in MC groups significant increases in affinity were found in comparison with age-matched LC groups (K d ; 49, 69, and 80%; 3, 13, and 18 month groups, respectively; P Ͻ 0.05; Fig 2B). A significant agedependent decline in 5-HT 1A receptor affinity was observed in the 18 month LC group as compared to the 12 month LC (55%, P Ͻ 0.05; Fig. 2B). The HC groups did not reveal any changes in affinity as compared to the MC groups across age in the frontal cortex.

[12] 79w The density (B max ) of 5-HT 1A receptors in MC treatment groups did not vary in the hypothalamus across age (Fig. 3A). A significant downregulation of receptor density was observed in both 3 and 18 month MC groups as compared to age-matched LC groups (Ϫ28 and Ϫ59%, P Ͻ 0.05, respectively; Fig 3A). Also, the 18 month LC group revealed a significantly higher density of binding sites than the 3 month LC (42%, P Ͻ 0.05; Fig. 3A).

[13] 124w No change in affinity for the 5-HT 1A receptor was observed in the 3 month LC group as compared to the 3 month MC group in the hypothalamus. A significant decrease in affinity was observed in the 18 month LC group as compared to the age-matched MC group (51%, P Ͻ 0.05) in this brain region. The decline in affinity in the 18 month LC group was also significantly greater than that observed in the 3 month LC group (47%, P Ͻ 0.05). This decrease in affinity found in the hypothalamus is similar to the observations in the frontal cortex; however, the upregulation of 5-HT 1A receptor density by low levels of corticosterone in the 18 month group was observed only in the hypothalamus.

[14] 152w Two levels of corticosterone were used to examine GFAP localization with age, the LC and the HC groups. Analysis of the hippocampus was chosen because adrenalectomy treatment has been shown to markedly increase GFAP expression in young animals (O'Callaghan et al., 1989). Also, corticosterone administration has been shown to decrease GFAP mRNA levels in intact animals (Nichols et al., 1990) and decrease the resultant GFAP expression following ADX treatment (O'Callaghan et al., 1989). In this study, no significant changes were observed in GFAP immunoreactivity between the 3 and 18 month HC groups (Fig. 5). The removal of corticosterone, LC, enhanced GFAP immunoreactivity in the 3 month animals as compared to age-matched HC groups. In the 18 month group, there was no change in GFAP distribution in the LC group as compared to the age-matched HC group (Fig. 5), revealing an age-associated change in GFAP-induced responses to low levels of circulating plasma corticosterone.

[1] 154w The regulation of 5-HT 1A receptor density by corticosterone exposure has been shown previously in the hippocampus of young animals (Mendelson et al., 1992;Frankfurt et al., 1994;Chalmers et al., 1993). Our studies have extended these observations with respect to agerelated changes. We have demonstrated a marked decline in the ability of varying plasma corticosterone concentrations to regulate hippocampal 5-HT 1A receptors with increasing age. Equilibrium binding assay data show that the increases and decreases in 5-HT 1A receptor density observed in response to low and high circulating plasma corticosterone, respectively, present in young animals is absent in old animals. A similar disfunction in the regulation resulted with an enhanced affinity of the 5-HT 1A receptor in young and middle-aged groups, but not in the old group, with high circulating corticosterone levels. In all cases the K d values obtained remained within a high affinity (nM) range and provided evidence of only one binding site.

[2] 371w The ongoing debate regarding age-related changes in cell number and morphology includes description of hippocampal atrophy and cell loss as components of normal aging (Deleon et al., 1997); conversely, minimal or no change in cell numbers (Wickelgren, 1996) or increased dendritic extensions (Pyapali et al., 1995) with aging have been reported. Studies involving stress paradigms also produce hippocampal CA3 pyramidal neuronal atrophy (Magarinos et al., 1996). The results of this present study failed to reveal changes in pyramidal cell number across ages and hormone treatment and verified that the changes we have identified in 5-HT 1A receptor receptor density shows that the 13 month MC treatment resulted in a significant decrease in 5-HT 1A receptor density compared to the age-matched LC treatment group. Similarly, the 13 month HC treatment resulted in a significant decrease in 5-HT 1A receptor density compared to the age-matched MC group. No change in 5-HT 1A receptor density was observed across treatment in the 3 and 18 month groups. B: The graph of cortical 5-HT 1A receptor affinity reveals a significant decrease in receptor affinity in LC groups compared to age-matched MC groups at all ages and a significantly greater decrease in affinity in the 18 month LC group compared to 3 and 13 month MC groups. Bars represent B max (fmol/mg protein; mean Ϯ SEM) and K d (mean nM Ϯ SE) values from five to seven animals per group at each age (white, LC; hatched, MC; gray, HC). A two-way ANOVA, followed by Student's t-test for planned comparisons, was used for analysis between treatment and across age groups. *Significant difference (P Ͻ 0.05) between LC and MC treatments of same age group. # Significant difference (P Ͻ 0.05) between MC and HC treatments of same age group. Significant difference (P Ͻ 0.05) between LC and HC treatments of the same age. Significant difference (P Ͻ 0.05) between 18 months and younger groups of the same treatment. density were independent of hippocampal pyramidal cell loss following an extended period of corticosterone exposure with increasing age. These observations are in agreement with previous studies using a similar model of hormone replacement in Fischer 344 rats for an extended duration of 4 months exposure (Bodnoff et al., 1995).

[3] 196w GFAP immunoreactivity was assessed as a marker of age-dependent neuronal damage in this study. Following damage of hippocampal neurons, an astrocytic response, known as reactive gliosis, results in an increase in GFAP expression (Landfield et al., 1977;Lindsay et al., 1986). GFAP immunoreactivity has also been shown to increase steadily with normal aging in male (O'Callaghan et al., 1991;Linnemann et al., 1994) and female rats (Dugar et al., 1998). Likewise, corticosterone replacement has been demonstrated to acutely decrease hippocampal GFAP expression in young animals (O'Callaghan et al., 1989). In the present study, a decrease in the level of GFAP expression with corticosterone exposure was observed in the young animals. Additionally, an increase in GFAP immunoreactivity in the old animals was above that observed in the young treatment-matched group was found. However, no changes were observed in GFAP immunoreactivity between LC and HC groups within the 18 month group. The results suggest that the neuronal protective role of corticosterone may be decreased in the aging hippocampus and exemplify the toxicity of this hormone to the aging brain (Sapolsky et al., 1985a). Also, these findings could be representative of an ageassociated general decline in corticosterone regulation in the hippocampus.

[4] 329w The 5-HT 1A receptor has been characterized by brain region-specific heterogeneity with aging including differences in density, affinity, and turnover (Keck and Lakoski, 1996;Keck et al., unpublished observations). This study has further elucidated age-dependent and region-specific changes with aging 5-HT 1A receptors in the hippocampus, frontal cortex, and hypothalamus. Marked differences were observed between these serotonergic receptor populations that may represent functional differences in these regions with respect to regulation of the HPA axis (hypothalamus), cognitive processes (frontal cortex), or both (hippocampus). In the frontal cortex, the changes in density of 5-HT 1A receptors due to corticosterone concentration in young animals correlated with previous findings (Crayton et al., 1996) but were not as pronounced as those in the hippocampus. In contrast to the limited corticosterone-mediated responses in the hippocampus and frontal cortex found in the oldest group, an upregulation of 5-HT 1A receptor density in the hypo-Fig. 3. Hypothalamic 5-HT 1A receptor characteristics in adrenalectomized 3 and 18 month female Fischer 344 rats supplemented with 21 day sustained-release placebo (LC) or 200 mg (MC) corticosterone-containing pellets implanted subcutaneously. A: Hypothalamic 5-HT 1A receptor density is increased in both 3 and 18 month LC groups as compared to age-matched MC groups; a significant age-dependent increase in 5-HT 1A density was observed in the 18 month LC group as compared to the 3 month LC group. B: Hypothalamic 5-HT 1A receptor affinity is decreased in the 18 month LC group compared to the age-matched MC group as well as compared to the 3 month LC group, the latter revealing an age-dependent change. Bars represent B max (mean fmol/mg protein Ϯ SEM) and K d (mean nM Ϯ SEM) values from five to seven animals per group at each age (hatched, MC; white, LC). A two-way ANOVA, followed by Student's t-test for planned comparisons, was used for analysis between treatment and across age groups. *Significantly different from age-matched treatment (P Ͻ 0.05). Significantly different from 2 month LC group (P Ͻ 0.05).

[5] 67w thalamus was revealed in the presence of low levels of plasma corticosterone in this age group. The sustained regulation of hypothalamic 5-HT 1A receptor binding sites in both young and old treatment groups suggests that deficiencies in adapting to stress and changing environments with increasing age may not be mediated solely at the level of the hypothalamus but, rather, includes actions at the level of the hippocampus.

[6] 301w A lack of a corticosterone-induced increase in affinity for 5-HT 1A receptors in the aging hippocampus under conditions of high levels of corticosterone may contribute to the neuronal damage from glucocorticoidinduced excitotoxicity observed with advancing age (Sapolsky et al., 1985b). Such cellular damage may be due to the loss of a protective physiologic inhibition in cellular activity produced by activation of 5-HT 1A receptors as mediated by a decreased cell firing rate (Beck, 1989). A recent study has indeed shown that glucocorticoid level and 5-HT-induced inhibition of hippocampal neuronal cell firing is linked (Hesen et al., 1996). The loss of these inhibitory physiological interactions can produce a calcium overload at the synapse which may result in an excited cellular physiological response that, in turn, culminates in neuronal death (Lesch et al., 1996). Coupled with the age-dependent decline in the density of hippocampal GR receptors (Roth et al., 1974;Sapolsky et al., 1983), a loss of corticosterone regulation of 5-HT 1A receptor expression may enhance vulnerability of the hippocampus to the excitotoxic damage under conditions of chronically high levels of corticosterone with aging (Sapolsky et al., 1985a). Thus, the combination of hormone and neurotransmitter receptor changes with aging is consistent with the increases in neuronal damage often observed in the hippocampus of elderly individuals (Sapolsky et al., 1985a;Talmi et al., 1993). This model of hormone receptor interaction may also extend to other 5-HT receptor subtypes; for example, the recently cloned 5-HT 7 receptor (Shen et al., 1993) may also be involved in the regulation of the HPA axis. Upregulation of this serotonin receptor subtype mRNA was observed in the CA1 and CA3 subfields of the hippocampus of rats following adrenalectomy (Corre et al., 1997;Yau et al., 1997) and demonstrates a putative role of this receptor subtype in the regulation of responses to stress.

[7] 272w The 5-HT 1A receptor plays an important role in stress-induced responses (Mendelson et al., 1992;McKittrick et al., 1995); however, the interaction between corticosterone and this receptor subtype is complex. Increased levels of corticosterone result in a decreased expression of 5-HT 1A receptors and reduced corticosterone levels result in an increased expression for this receptor subtype (Mendelson, 1992). Acute stressors also have differential effects on 5-HT 1A receptor expression based on the nature of the stressor (Raghupathi et al., 1996). Our study has identified an age-related deficit in the regulation of 5-HT 1A receptors by corticosterone which may underlie the observed loss of adaptation and ability to cope with stress found with advancing age (Landfield et al., 1978;Meaney et al., 1987;Handa et al., 1993;Spencer et al., 1997). Several mechanisms for this loss of regulation as a function of age include alteration in 5-HT 1A receptor turnover, translation, posttranslational modification, and/or coupling to its G protein effector systems. Changes in membrane dynamics and composition with aging (Roth et al., 1995) could also result in altered coupling of the receptors and to effector molecules such as the 5-HT 1A receptor and the associated G i protein (Weiss et al., 1983). Similar age-related deficits in changes in G i protein numbers could be a contributing factor to the age-associated deficit in cellular reponsiveness following stress. For example, corticosterone exposure has been shown to upregulate hippocampal G i and G o protein levels in young animals (Okahara, 1995). A loss of G protein-mediated signal transduction processes by corticosterone, including coupling to 5-HT 1A receptors with advancing age, may underlie the declines in serotonergic function we have observed.

[8] 199w An extension of the present investigation will be examination of the changes in the levels of 5-HT 1A receptor mRNA expression with aging and varied corticosterone exposure to determine if the loss of 5-HT 1A receptor density regulation in older animals is mediated at the level of gene transcription. Previous studies have reported decreased levels of 5-HT 1A receptor mRNA in the aged human hippocampus and cortical tissues (Dillon et al., 1991;Burnet et al., 1994) supporting the hypothesis that age-related declines in regulation at the level of transcription may occur for this binding site in an age-related manner. Marked increases in hippocampal levels of 5-HT 1A receptor mRNA expression following ADX in young animals (Chalmers et al., 1993) have also been shown. These increases in message expression were attenuated by the exogenous administration of the synthetic glucocorticoid dexamethasone at the time of ADX (Chalmers et al., 1994). Such investigations support a possible mechanism at the level of transcription for the interaction of 5-HT 1A receptors and corticosterone in the HPA axis. Likewise, our results may be mediated by a decline in corticosterone/5-HT 1A receptor interaction at the level of gene transcription processes as proposed by Zhong and Ciaranello (1995).

[9] 98w It is important to note that the differences specific to the serotonergic system in the aging hippocampus may also play a role in the developing HPA axis. For example, the activation of ascending serotonergic projections into the hippocampal region via thyroid hormone release produced by handling of young pups (Sapolsky, 1997) may increase GR number in hippocampal neurons and result in a subsequent inhibition of corticosterone release throughout the lifespan (Liu et al., 1997). As applied to a model of aging, such altered feedback could serve to decelerate brain aging by reducing exposure to excitotoxic levels of corticosterone.

[10] 134w In summary, elucidation of corticosterone and serotonin 5-HT 1A receptor interaction in the aging brain may help to explain deficiencies that are found in elderly individuals adapting to stress (Sapolsky et al., 1986;Van Eekelen et al., 1992, 1995;Cizza et al., 1994;Hauger et al., 1994), declines in cognitive function (deToledo-Morrell et al., 1988;Levy et al., 1994;Arbel et al., 1994;Bodnoff et al., 1995), and underlie the increased incidence of depression (Reynolds et al., 1994;Salzman, 1994;Slotkin et al., 1989), problems that greatly affect the quality of life in the elderly. Continued elucidation of the cellular and molecular mechanisms of the interaction between endocrine and neurotransmitter systems should provide insight into the processes of normal and pathologi-cal brain aging, as well as provide new therapeutic approaches to combat the interrelated effects of stress and aging on HPA function.

[1] 167w Virgin female Fischer 344 rats (National Institute of Aging Colony; Indianapolis, IN) were housed in groups under standard laboratory conditions with lights on at 0700-1900 hr. Animals were 2, 12, and 17 months (mo) old at initiation of experiments, corresponding to sexually mature, reproductively middle-aged (irregular estrous cycles), and reproductively senescent ages, respectively (Wise et al., 1983;Finch et al., 1984). Reproductive status was confirmed by using daily vaginal lavage, followed by methylene blue staining and microscopic examination of vaginal epithelium for hormone-dependent characteristics (Wells, 1964). Ongoing studies in our laboratory are utilizing female rats to address the role of gender in the decline adaptive capabilities of nervous system function during aging. An older group was not chosen for evaluation because this rodent strain has shown an increase in endocrine pathologies with aging, including pituitary adenomas in the female (Boorman, 1990). Recognizing the importance of utilizing an animal model that is free of disease to study normal aging processes, our experimental design omitted the use of senescent rats.

[2] 146w All animals were bilaterally adrenalectomized (ADX) under sodium pentobarbital anesthesia (40 mg/kg, i.p.) and implanted subcutaneously with a 21 day sustained-release placebo, 200 mg corticosterone, or three 200 mg corticosterone pellets (Innovative Research of America; Sarasota, FL). These sustained-release pellets provided levels of circulating corticosterone that were equivalent to low (LC), normal (MC), and high or stressed levels of corticosterone (HC), respectively. These concentrations of corticosterone reflect differences in occupation of corticosteroid receptors in the hippocampus such that MR and GR occupied. This model of corticosterone replacement was chosen because a previous model which incorporated treatment groups of ADX, sham ADX and unoperated control, all carried for 7 days, resulted in a high variance of circulating plasma corticosterone levels. For example, values for 3-month-old animals were 22.3 Ϯ 7.5 and 11.4 Ϯ 5.7 µg /dL (mean Ϯ SE) for unoperated control and sham ADX animals, respectively.

[3] 162w Following surgery and pellet implantation, animals were weighed twice weekly and maintained for 21 days under standard conditions with saline solution (0.9% NaCl) available ad libitum. No significant changes in body weight were observed in any treatment group. Following the postsurgical period, rats were rapidly decapitated using a small animal guillotine (0900-1100 hr) within a few second period and care taken to minimize stress during the procedure. Trunk blood was collected in 7 ml vacutainer tubes containing 10.5 mg ethylene diamine tetraacetic acid (# 6450, Becton Dickinson; Rutherford, NJ). Samples were allowed to clot on ice for 30 min, spun at for 10 min at 2,000 rpm using a centrifuge (Beckman TJ-6 tabletop; Fullerton, CA), supernatant aliquoted, and stored frozen at -80 o C for subsequent plasma corticosterone determination. Brains were rapidly removed, the hippocampus, frontal cortex, and hypothalamus were dissected on ice, wet weights determined and tissues frozen on dry ice, and stored at -80 o C until time of assay.

[4] 179w Saturation equilibrium binding assays were used for the hippocampus and frontal cortex of individual animals and the hypothalamus of two pooled treatment individuals using [ 3 H]8-hydroxy-2-(di-N-propylamino) tetralin ([ 3 H]8-OH-DPAT; 120-154 Ci/mmol, Dupont New England Nuclear; Boston, MA) according to the protocols of Hall et al. (1985) and Keck and Lakoski (1996). Briefly, tissues (Ϸ100 mg) were homogenized with a Brinkmann Polytron (setting 4, 10 s; Brinkmann Instruments, Inc.; Westbury, NY) in 25 vol of 50 mM Tris-HCl buffer (pH 7.4; Sigma Chemical Company Co., St. Louis, MO). The homogenate was centrifuged using a Beckman JA-20 rotor at 31,000g for 10 min at 4 o C. The supernatant was discarded and the pellet was resuspended in 25 vol of Tris-HCl buffer, resuspended by vortexing, and centrifuged as above. The resulting supernatant was discarded and the pellet resuspended in 25 vol Tris-HCl, incubated at 37 o C for 10 min and centrifuged. The supernatant was discarded and the pellet rehomogenized in 25 vol of Tris-HCl buffer containing 4 mM CaCl 2 , 0.1% ascorbic acid, and 10 µM pargyline.

[5] 184w Binding assays incorporated duplicates of 0.1 ml tissue suspension, 25 µl of [ 3 H]8-OH-DPAT (final concentrations ranging from 0.4-13.4 nM), 0.125 or 0.1 ml assay buffer and 0 or 25 µl of 8-OH-DPAT (100 µM). Specific binding was defined as total counts bound minus the nonspecific binding determined by addition of the excess displacing drug (100 µM 8-OH-DPAT). Tubes were vortexed, incubated at 37 o C for 10 min, and the reaction mixture was rapidly filtered under vacuum (Brandel M-24 cell harvester; Gaithersburg, MD) through GF/B glass fiber filters, and washed with two 5 ml volumes of ice cold Tris-HCl buffer. Filters were then deposited into Beckman Poly-Q mini vials and 5 ml Beckman Ready Safe scintillation fluid (Beckman Instruments, Inc.; Fuller-ton, CA) was added to each vial using a Brandel deposit/dispense apparatus. Radioactivity was determined using liquid scintillation spectroscopy (Beckman Model LS 3801) calculated at an efficiency of 53% (H-value ϭ 30). Final protein concentrations per tube were assessed by a Bio-Rad Microassay (Lowry et al., 1951; Bio-Rad Laboratories; Hercules, CA) using bovine serum albumin as a standard (Standard II; Bio-Rad Laboratories.

[6] 154w Brains were blocked on ice to include the region of the dorsal hippocampus. Blocked tissue was frozen immediately in isopentane (Ϫ30 to Ϫ40°C), maintained in dry ice, wrapped tightly in parafilm and aluminum foil to prevent desiccation and stored at Ϫ80°C until preparation of coronal sections. Tissue sections (20 µm thick) were prepared on a cryostat at Ϫ20°C, thaw-mounted on gelatin-coated slides, and stored at Ϫ20°C until staining. Slides were allowed to come to room temperature, dipped for 1 min in 0.1% cresyl violet solution (pH adjusted to 3.5 with acetic acid), and rinsed by dipping in fresh distilled water for 2 min. Sections were then viewed under a light microscope for total pyramidal cell counting and neuronal cell morphology observation within CA1, CA3, and dentate gyrus hippocampal regions. This study was limited to the visualization of any gross steroidand/or age-dependent changes in cellular number or morphology which was the goal of these experiments.

[7] 49w Corticosterone radioimmunoassays were performed using antiserum from Radioassay Systems Laboratories (Carson, CA) with sensitivity of 0.02 ng/tube. Intra-assay variability was 4.5% and inter-assay variability was 11.9%. Assays were performed courtesy of Dr. Louis D. Van de Kar, Loyola University School of Medicine as described by Van de Kar (1985).

[8] 94w Receptor density (B max ) and affinity (K d ) values via Scatchard analysis were determined on individual tissues using the program ''Ligand/EBDA'' (Version 2.0, Elsevier Biosoft; Amsterdam). Typical assays yielded Hill coefficients close to unity and Scatchard correlation greater than 0.93 was a criterion for accepting data for final evaluation. Two-way analysis of variance was performed across age and hormone treatment followed by a Students' t-test (P Ͻ 0.05 level of significance) for both binding assays and quantitative autoradiographic results. Statistical comparisons were made using SigmaStat (Version 1.0, Jandel Corporation, San Rafael, CA).

[9] 52w Light microscopic examination of cresyl violetstained sections adjacent to those used in the autoradiographic analysis were evaluated for cell number by manual counting cells in comparable anatomical regions across age and treatment. It was determined that no appreciable changes in cell number were evident across this paradigm in the hippocampus (Fig 4).

[1] 238w Coronal sections (30 µm) of the dorsal hippocampus were obtained using a vibratome from all treatment groups. Immunohistochemical localization of GFAP was performed as previously reported (Patanow et al., 1996). Briefly, slices were stored in Tris buffered saline (TBS; pH 7.4) at 4°C until ready for analysis. Sections were placed in cold TBS (4°C) containing 30% hydrogen peroxide (H 2 O 2 ; Fisher Scientific; Fair Lawn, NJ) and 1 mg/ml sodium azide in seven ml six-well plates for 10 min while continuously shaking on a rocker to quench endogenous peroxidase activity; all rinses and incubations occurred with sections placed on a rocker. Slices were rinsed in TBS at room temperature (RT; 5ϫ for 15 min) to remove all traces of azide. Sections used for immunohistochemical analysis with the GFAP mouse monoclonal primary (1°; 1:1,000) antibody (anti-GFAP; Boehringer-Mannheim, Indianapolis, IN) were washed in TBS containing 0.5% Triton X-100 (Sigma Chemical) overnight at 4°C. Sections were rinsed the following day in TBS (4ϫ for 15 min) at RT and then incubated in (1°) Ab for 48 hr at 4°C. Slices were then rinsed in TBS (4ϫ for 15 min) at RT. The sections were incubated in diluted rabbit anti-mouse (2°; 1:500; Jackson Immunoresearch Laboratories, Inc. West Grove, PA) for 2 hr at RT followed by rinses in TBS (4ϫ for 15 min) at room temperature. Brain slices not incubated with 1°Ab were carried as negative controls in all experiments.

[2] 77w Immunoreactive sites were visualized using diaminobenzadine (DAB; 0.5 mg/ml; Sigma Chemical Co.) activated with 0.02% H 2 O 2 . After a final rinse in TBS, the slices were mounted on gelatin-subbed slides and allowed to air dry overnight. Slides were dehydrated through a graded series of alcohol washes and a final xylene wash. Sections were coverslipped with Cytoseal (Stephens Scientific; Riverdale, NJ) and examined under a light microscope (Olympus Model BH-2, Olympus Co., Lake Success, NY).