[1] 17w LONG-TERM GLUCOCORTICOID TREATMENTS DECREASE LOCAL CEREBRAL BLOOD FLOW IN THE RAT HIPPOCAMPUS, IN ASSOCIATION WITH HISTOLOGICAL DAMAGE

[1] 189w The present study examined the influence of a long-term treatment with glucocorticoid on local cerebral blood flow of the hippocampus in rats, estimated with the hydrogen clearance method. Either a cholesterol (100 mg, as a control) or corticosterone (100 mg) bead was implanted subcutaneously in rats for a period of three months, beginning at 12 weeks of age. The effects of the treatments on the local circulation of the hippocampus were evaluated three to four months after the termination of the treatments. Hippocampal cerebral blood flow in corticosterone-treated rats was significantly lower (P<0.05) than that in control rats, and fluctuated over a day in lower amplitude than the controls. Severe histological damage was observed in the CA1 and CA3 cell fields of the hippocampus in corticosteronetreated rats. These neuropathological changes were characterized by soma shrinkage and condensation, or nuclear pyknosis, as reported previously.We concluded that a long-term glucocorticoid exposure resulted in an impairment of the hippocampal functions, accompanied by neuronal damage similar to that found in aged hippocampus. The present results support the hypothesis that glucocorticoids accelerate age-related changes in the brain. 1997 IBRO. Published by Elsevier Science Ltd.

[1] 80w Figure 1 shows mean CBF values for rats in the control and the corticosterone groups with respect to the time of day. Two-way repeated measures ANOVA (treatment time) showed a significant fluctuation in mean CBF values over a day in both groups (F 23,360 =3.381, P<0.0001) and a significant glucocorticoid treatment effect (F 1,360 =69.914, P<0.0001), suggesting a diurnal variation in both groups, and lower values in the corticosterone group, respectively (Fig. 1). There was no significant treatment time interaction.

[2] 135w The overall mean CBF value for the 24 h period for the corticosterone group was lower than that for the control group (P<0.05 by Student's t-test). This was related to a smaller CBF in each of the light cycle (P<0.05 by Student's t-test) and the dark cycle (P<0.05 by Student's t-test) in rats in the corticosterone group (Table 1). The dark/light ratio, i.e., the mean increase rate during the dark cycle (19.00-05.00) compared to the light cycle (06.00-18.00), was not significantly different (control group, 1.20 0.05; corticosterone group, 1.13 0.03), but the (after lights off)/(before lights off) ratio, i.e., that during the 19.00-21.00 period (after lights off) compared to the 16.00-18.00 period (before lights off), was significantly lower in the corticosterone group (control group, 1.25 0.62; corticosterone group, 1.05 0.59, P<0.05 by ANOVA) (Fig. 1).

[1] 266w The present study demonstrated that long-term glucocorticoid treatments decreased local CBF in the hippocampus and influenced the diurnal rhythm of hippocampal CBF. There is a generally accepted concept that local CBF correlates with changes in energy metabolism, and, further, the latter is correlated with the level of neuronal activity. 23,29 For example, sensory stimulation causes appropriate increases in CBF and metabolism. 9 It has been shown that glucocorticoid has an inhibitory effect on hippocampal pyramidal cell firing 22 and field potentials, 33 but the mechanism underlying this effect may not account for the possible decrease in the hippocampal neuronal activity as shown in this study, because the effect of glucocorticoid was observed to be longlasting following the cessation of glucocorticoid treatments. Further, when local CBF was measured, circulating glucocorticoid levels might rather decrease due to down-regulation of the hypothalamopituitary-adrenal axis after long-term glucocorticoid treatment. Conversely, if we measured local CBF in the hippocampus during glucocorticoid treatments, local CBF values might also be lower by the mechanism of inhibitory effects of glucocorticoid, as noted above. Therefore, it is likely that the glucocorticoid treatment as employed in the present study was rather more toxic to the hippocampal neurons than it was inhibiting to the neuronal activity. Although there might be another possibility that CBF monitoring may be a stimulant to the hypothalamo-pituitaryadrenal axis, if so, local CBF in the hippocampus of rats in the corticosterone group could rather increase as compared with rats in the control group, for lower inhibitory influences due to lower corticosterone level in rats in the corticosterone group. This possibility cannot explain our results.

[2] 329w The most notable findings in the present study were decreases in local CBF in the CA3 cell field and neuropathological changes in the CA1 and CA3 cell fields of the hippocampus in rats given long-term glucocorticoid treatments. The glucocorticoid effects of damaging or destroying hippocampal neurons have been accounted for by glucocorticoid's ''toxic action''. 25 Due to such structural damage that might result in reduction of local metabolic demands, a decrease in local CBF in the hippocampus was considered as having occurred in glucocorticoid-treated rats in our study. Furthermore, changes in local CBF in other brain regions, if any, should be smaller, since regions other than the hippocampus are less sensitive to glucocorticoid. 21,25,34 Hippocampal CBF in glucocorticoid-treated rats fluctuated in lower amplitudes due to larger decrease in hippocampal CBF during the dark cycle, where normal rats actively respond to the environmental stimuli and thus have an increased hippocampal CBF. 4 Few studies have evaluated the effects of glucocorticoids on spontaneous behaviours. However, Ehlers et al. reported that exposure to corticosterone over a 10 week period did not effect locomotor activity. 3 Accordingly, we may speculate that hippocampal neurons of rats given long-term glucocorticoid treatments are less responsive to environmental stimuli and thus less active. This fact is in harmony with the possible impairment of learning abilities. 6,14 As it has been revealed that neural factors regulate local CBF independently of the local metabolic rate, i.e. septohippocampal projections, 2 there is a possibility that the septal neurons could provide hippocampal CBF with features of diurnal rhythm. With regard to the timing of cell death by glucocorticoid treatments, the parent cholinergic neurons in the septum start to degenerate prior to their target hippocampal neurons during long-term glucocorticoid exposure. 31 Although we did not quantify neuronal damages in the septum, we cannot deny a possibility that, due to degeneration of the septal neurons, local CBF in the hippocampus could significantly decrease and the diurnal rhythm of hippocampal CBF could be affected.

[3] 150w In three rats given long-term corticosterone treatments, the total number of pyramidal neurons in CA1 appeared to be reduced as compared with rats in the control group, but not different in CA3. Further, Glucocorticoids and local CBF in the hippocampus the number of damaged neurons in both CA1 and CA3 cell fields was increased, in partial agreement with findings of a recent report. 13,26 However, these changes differed from those reported by Sapolsky et al., 26 namely, the total cell number was decreased to 75% of the control in CA3, more substantial than in CA1. The reason for the discrepancy is unclear in detail, but this difference is probably due to a lower dose of glucocorticoid in our study (approximately 3 mg/day) than in theirs (5 mg/day), and possibly due to a long-term period between the cessation of glucocorticoid treatments and the histological analysis in 750 Y. Endo et al.

[4] 264w our study, about six months. These damaged neurons that appear dark in our study were probably irreversible, since these neurons were darkly stained with acidic dye, Eosin. 1 It has been shown that histological changes in aged hippocampus are characterized by (i) the decrease in the number of neurons of Ammon's horn, and (ii) glial proliferation and infiltration, with both characteristics being observed in long-term glucocorticoid exposure. 12,26 The results of these studies have led to the hypothesis that glucocorticoid plays an important role in promoting an age-related central neuron degeneration. 13,16,27 Our present results, in part, support such a hypothesis and provide further strong evidence that these neuropathological changes in the hippocampus caused by chronic exposure to glucocorticoids are associated with a decrease in local CBF in the hippocampus. Aged rats exhibit deficits in their spatial working memory in radial eight arm maze 8,35 and have significantly decreased glucose utilization in the limbic system, including the hippocampus. 7 In aged human, local CBF in the brain region including the hippocampus significantly decreases. 15 In patients with senile dementia of the Alzheimer type, CBF levels are maintained before the onset of the symptoms of dementia, and as cognition declined, CBF decreased rapidly and diffusely. 24,36 The changes of CBF are thought to be biologically significant. These phenomena are similar to those which we also observed in rats given long-term glucocorticoid treatment. 6 Furthermore, these influences of glucocorticoids appeared to be long-lasting or irreversible. As a result, long-term exposures to an excess of glucocorticoids could cause hippocampal dysfunctions similar to aged hippocampus, morphologically and physiologically.

[1] 66w In summary, a long-term glucocorticoid exposure caused a reduction of local CBF in the hippocampus and a lower amplitude of this fluctuation over a day as compared with age-matched control, accompanied by neuronal damage similar to that found in aged hippocampus. It is suggested that long-term glucocorticoid-treated rats could be a useful investigative model for the study of normal ageing or some types of senile dementia.

[1] 40w Adult male Wistar-Imamichi rats, obtained from Animal Reproduction Research Co. (Urawa, Saitama, Japan), were used. All animals were housed under standard conditions, with lights on 05.00-19.00 and controlled room temperature at 24 C, and received food and water ad libitum.

[2] 31w ‡To whom correspondence should be addressed. Present address: Department of Physiology, University of Occupational and Environmental Health School of Medicine, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu 807, Japan. Abbreviations: CBF, cerebral blood flow.

[3] 111w At 12 weeks of age, rats had pure corticosterone (Sigma Chemical Co., St. Louis, MO) beads (100 mg) implanted subcutaneously in the back under ether anaesthesia (corticosterone group). The beads dissolved at an average release rate of around 3 mg/day, at a dose sufficient to mimic the circadian peak level of the hormone. 19 Age-matched rats had cholesterol (Sigma Chemical Co., St. Louis, MO) beads (100 mg) implanted (control group). Rats of each group were reimplanted with corticosterone or cholesterol beads, respectively, at four week intervals over 12 weeks. Following the treatment period, rats were allowed recovery and the experiments described hereafter were done three to four months after the treatments.

[4] 110w At 12 months of age (six months after the cessation of corticosterone treatments), the additional groups of six rats (three rats of the control group, three rats of the corticosterone group) were anaesthetized with i.p. injection of sodium pentobarbital (40 mg/kg body weight) and transcardially perfused with 200 ml 0.9% saline and added heparin (5 U/ml), followed by perfusion with 2.0 l of fixative (10% neutral buffered formalin). Brains were removed carefully and post-fixed in the same fixative. Serial sections were cut at 15 µm and mounted alternately on two series of glass slides. One series was stained with Haematoxylin and Eosin and another series was counterstained with Cresyl Violet.

[5] 76w The number of the pyramidal neurons, expressed as number/200 µm, was derived by averaging cell counts from five successive sections taken over a 200 µm length of each cell field in the hippocampus. A neuron was counted as damaged if a neuron was atrophic or darkly stained. Neurons without the above changes were counted as undamaged. The total number of neurons, damaged or undamaged neurons, was calculated in each cell field of the hippocampus/animal and averaged/group.

[6] 25w Differences between means were analysed by two-way repeated measures analysis of variance (ANOVA) or by Student's t-test, with the significance level being set at P<0.05.

[7] 29w Histological examinations were performed in three rats of each group, six months after the cessation of the treatments. There was severe damage in the 746 Y. Endo et al.

[8] 119w hippocampus in all three rats in the corticosterone group. The most remarkable changes were found in the pyramidal neurons in the CA1 and CA3 cell fields, but not in CA2, CA4, and not in granule cells of the dentate gyrus (Figs 2, 3). On a quantitative analysis in the subfields of the hippocampus and the dentate gyrus, the total cell number of pyramidal neurons in CA1 in rats in the corticosterone group appeared to be reduced as compared with rats in the control group, but not different in the others. Further, the numbers of damaged neurons in the CA1 and CA3 cell fields in rats in the corticosterone group were greater than that in the control group (Fig. 4).

[9] 48w The neuropathological changes were characterized by soma shrinkage and condensation, or nuclear pyknosis, in agreement with other reports. 28 Almost all pyramidal neurons in the CA1 and CA3 had an irregularly-shaped perikarya associated with dispersed Nissl bodies. These neurons were specially visualized with acidic dye, Eosin (Fig. 3).

[1] 205w Chronic exposure to glucocorticoids, which are secreted by stress and induce various catabolic responses that mediate adaptation to stress, damages the hippocampal neurons, a principal neural target site for steroids. 17 For example, daily injections of glucocorticoids over a long period result in a reduction in the number of neurons in the hippocampus, with or without a depletion of glucocorticoid receptors, in a direction that accelerates ageing. 12,13,26,27 Exposure to an excess of glucocorticoids either through repeated stress or glucocorticoid injection caused a premature ageing change in the hippocampal morphology, on the one hand, 12,32 and the neurochemical and electrophysiological parameters on the other. 10,30 We have also found that long-term exposure to glucocorticoids resulted in an impairment of maze learning in rats, similar to that found in aged ones. 6 Conversely, reducing the exposure to glucocorticoids by means of adrenalectomy or behavioural manipulations protects the hippocampus from age-associated neuron loss and dysfunction. 12,18 In a previous study, we found that adrenalectomy increased local cerebral blood flow (CBF) in the hippocampus, in a direction that appeared to delay ageing. 5 Accordingly, there is a possibility that endogenous glucocorticoids, even at normal levels, could lead to a gradual accumulation of deleterious effects on the hippocampal neurons.

[2] 77w We hypothesized in the present study that longterm exposure to glucocorticoids, within a physiological range, could impair certain hippocampal functions, in association with neuropathological changes in the hippocampus. In the present study, we investigated whether this hypothesis was correct by measuring local CBF in the hippocampus. Therefore, the purpose of the present study was to delineate whether exposure to an excess of glucocorticoids resulted in an impairment of hippocampal functions that was often observed in aged rats.

[3] 70w Three to four months after the cessation of the treatments (at nine to 10 months of age), local CBF in the hippocampus was measured in eight rats of the control group and nine rats of the corticosterone group. One week prior to these measurements of local CBF, a platinum electrode with a 1 mm bare tip was implanted stereotaxically under ether anaesthesia in the dorsal hippocampus of the right hemisphere.

[4] 184w Measurements were performed at 1 h intervals over a day by the hydrogen clearance method in unanaesthetized, freely moving rats. 4,5,20 An upper compartment of the chamber, constructed from a metabolic cage for the CBF measurements, held the rat and the lower compartment contained a duct that served either to collect faeces and urine, or to introduce the gas mixture. A 15% H 2 /85% air mixture was metered into the chamber at 0.6 l/min for 3 min and the respired H 2 washed out. The wash-out curves were recorded on an X-Y recorder, and local CBF during the 2 min following the first 30 s of the wash-out curves was calculated, based on the blood-tissue exchange theory of Kety and Schmidt. 11 At the end of CBF measurement, the rats were perfused with 10% formalin under ether anaesthesia and the brains were removed for histological identification of the site of each measuring electrode tip. We used the data of rats for statistical analysis, only in cases where the electrode tip was found to be located within the CA3 cell field of the hippocampus.