Intrahippocampal colchicine alters hypothalamic corticotropin-releasing hormone and hippocampal steroid receptor mRNA in rat brain.

The hippocampus appears to be an important modulator of the negative feedback effects of glucocorticoids on the hypothalamic-pituitary-adrenal axis. It is not known if hippocampal subfields CA1-4 or the dentate gyrus differentially alter gene expression of corticotropin-releasing hormone (CRH) in the paraventricular nucleus (PVN) of the hypothalamus. We, therefore, examined the effects of selective destruction of dentate gyrus granule cells, which send excitatory glutaminergic inputs to subfields CA4, CA3 and CA2, on CRH expression in the PVN. To determine the possible involvement of steroid receptors in the regulation of CRH expression, we examined the effects of intrahippocampal colchicine on gene expression of the mineralocorticoid (MR; type I) and glucocorticoid (GR; type II) receptors in hippocampal CA fields and dentate gyrus. Colchicine produced a selective loss of dentate gyrus granule cells without affecting pyramidal cells in CA1-4 as early as 1 day after injection; granule cells were completely destroyed after 3 days. CRH mRNA levels were reduced by 38-48% in the PVN 2-14 days after colchicine. MR mRNA levels were decreased in dorsal and ventral CA fields 1-7 days after colchicine. GR mRNA levels were relatively unchanged, showing a slight decrease only in dorsal CA fields on days 2-7. Unexpectedly, CRH was transiently expressed in dorsal and ventral CA fields 1-3 days after colchicine. In the same time period, mRNA levels of inositol 1,4,5-trisphosphate kinase were decreased, suggesting that increases in neural metabolic activity, indicated by this marker, are not responsible for the transient CRH effect. The results suggest that the dentate gyrus is important for maintenance of steroid hormone receptor mRNA levels in the hippocampus and CRH expression in the hypothalamic PVN, and that CRH gene expression is differentially regulated in the hypothalamus and hippocampus.     

Species-specific topography of corticosteroid receptor types in rat and hamster brain

In vivo and in vitro autoradiography with radiolabeled corticosteroid analogs as well as immunocytochemistry with monoclonal antibodies raised against the rat liver glucocorticoid receptor were used to determine the presence and the topography of two corticosteroid receptor systems (type I and type II) in hamster and rat brains. In the rat, the in vivo autoradiograms clearly revealed the retention by the type I receptor of tracer amount of [3H]corticosterone, primarily in the CA1 and CA2 cell field, dentate gyrus and lateral septum. In the hamster, tracer doses of [3H]cortisol were retained not only in the CA1, CA2, dentate gyrus and lateral septum, but also at high level in the CA3 and CA4 areas. In both species, immunocytochemistry showed the widespread distribution of the type II receptor sites in areas such as the hippocampus, lateral septum, hypothalamus (particularly in the paraventricular nucleus), thalamus and cortex (these results were also reflected in the in vitro autoradiography). Strong cell nuclear glucocorticoid immunoreactivity (type II-IR) was observed in the CA1 and CA2 (as well as CA3 and CA4 in the hamster) pyramidal neurons. In the hippocampus of intact animals, type II-IR was seen in the neuronal cell nuclei. Adrenalectomy caused a depletion of the type II-IR signal from the cell nucleus, which returned 1 h following subcutaneous administration of RU 28362 to adrenalectomized animals.(ABSTRACT TRUNCATED AT 250 WORDS)     

Blockade of calcium-permeable AMPA receptors protects hippocampal neurons against global ischemia-induced death

Transient global or forebrain ischemia induced experimentally in animals can cause selective, delayed neuronal death of hippocampal CA1 pyramidal neurons. A striking feature is a delayed rise in intracellular free Zn(2+) in CA1 neurons just before the onset of histologically detectable cell death. Here we show that alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-type glutamate receptors (AMPARs) at Schaffer collateral to CA1 synapses in postischemic hippocampus exhibit properties of Ca(2+)/Zn(2+)-permeable, Glu receptor 2 (GluR2)-lacking AMPARs before the rise in Zn(2+) and cell death. At 42 h after ischemia, AMPA excitatory postsynaptic currents exhibited pronounced inward rectification and marked sensitivity to 1-naphthyl acetyl spermine (Naspm), a selective channel blocker of GluR2-lacking AMPARs. In control hippocampus, AMPA excitatory postsynaptic currents were electrically linear and relatively insensitive to Naspm. Naspm injected intrahippocampally at 9-40 h after insult greatly reduced the late rise in intracellular free Zn(2+) in postischemic CA1 neurons and afforded partial protection against ischemia-induced cell death. These results implicate GluR2-lacking AMPA receptors in the ischemia-induced rise in free Zn(2+) and death of CA1 neurons, although a direct action at the time of the rise in Zn(2+) is unproven. This receptor subtype appears to be an important therapeutic target for intervention in ischemia-induced neuronal death in humans.     

Serotonin 5-HT(1A) receptor mRNA expression in dorsal hippocampus and raphe nuclei after gonadal hormone manipulation in female rats

Female ovarian steroids influence mood and cognition, an effect presumably mediated by the serotonergic system. A key receptor in this interplay may be the 5-HT(1A) receptor subtype. We gave adult ovariectomized female rats subcutaneous pellets containing different dosages of 17 beta-estradiol alone or in combination with progesterone, or placebo pellets, for 2 weeks. 5-HT(1A) receptor mRNA levels were analyzed by in situ hybridization in the dorsal hippocampus, dorsal and median raphe nuclei, and entorhinal cortex. Estradiol treatment alone reduced 5-HT(1A) gene expression in the dentate gyrus and the CA2 region (17 and 19% decrease, respectively). Estradiol combined with progesterone supplementation increased 5-HT(1A) gene expression versus placebo in the CA1 and CA2 subregions of the dorsal hippocampus (16 and 30% increase, respectively). Concomitantly, 5-HT(1A) mRNA expression was decreased by 13% in the ventrolateral part of the dorsal raphe nuclei, while no changes were found in the median raphe nucleus and entorhinal cortex. Chronic effects of ovarian hormones on 5-HT(1A) receptor mRNA expression appear tissue-specific and involve hippocampal subregions and the raphe nuclei. Modulation of 5-HT(1A) receptor gene expression may be of importance for gonadal steroid effects on mood and cognition.     

M-15: high-affinity chimeric peptide that blocks the neuronal actions of galanin in the hippocampus, locus coeruleus, and spinal cord.

The 20-amino acid peptide M-15 binds with high affinity (IC50 approximately 0.1 nM) to 125I-labeled galanin (125I-GAL) binding sites in membranes from the ventral hippocampus, midbrain, and rat spinal cord. Receptor autoradiographic studies show that M-15 can displace 125I-GAL from all labeled sites. M-15 acts as a reversible high-affinity antagonist in blocking the inhibitory effects of GAL on the evoked release of acetylcholine in vivo in the hippocampus and on the GAL-induced hyperpolarization of locus coeruleus neurons in slices. M-15 also blocks the facilitatory effects of GAL on the spinal flexor reflex. Thus, the chimeric peptide M-15 [GAL-(1-13)-substance P-(5-11)amide] represents the first antagonist to the neuronal actions of GAL.     

Epinephrine mediates the growth hormone-releasing effect of galanin in infant rats

The mechanism underlying the GH-releasing effect of galanin (GAL), a novel 29-amino acid peptide, was investigated in the neonatal rat. The effect of galanin was compared to that of clonidine (CLO), a drug known to release GH via endogenous GHRF. GAL administration (5-25 micrograms/kg BW, sc) induced in 10-day-old pups a clear-cut and dose-related rise in plasma GH 15 min postinjection. CLO (50-450 micrograms/kg BW, sc) induced a marked rise in plasma GH, but no dose-related effect was evident. Inhibition of hypothalamic norepinephrine and epinephrine biosynthesis by DU-18288 (6 mg/kg BW, ip) or selective inhibition of epinephrine biosynthesis by SKF-64139 (50 mg/kg BW, ip) completely abolished the GH-releasing effect of GAL (25 micrograms/kg, sc), but left unaltered the GH rise induced by CLO (150 micrograms/kg, sc). Passive immunization with an anti-GHRF serum decreased basal GH levels and prevented the GH-releasing effect of either GAL or CLO, whereas in pups pretreated with an antisomatostatin serum, CLO, but not GAL, increased the already elevated plasma GH titers. In all these data indicate that in the infant rat 1) GAL is a potent GH secretagogue; 2) the action of GAL is not exerted directly on GHRF- or somatostatin-secreting structures, but requires the intervention of catecholaminergic neurons; 3) the GH-releasing effect of GAL is ultimately exerted via GHRF release, although a mechanism operating to inhibit hypothalamic somatostatin release cannot be ruled out; and 4) differently from GAL, CLO releases GH via postsynaptic stimulation of GHRF-secreting neurons.     

Reduced hippocampal MT2 melatonin receptor expression in Alzheimer's disease

The aim of the present study was to identify the distribution of the second melatonin receptor (MT2) in the human hippocampus of elderly controls and Alzheimer's disease (AD) patients. This is the first report of immunohistochemical MT2 localization in the human hippocampus both in control and AD cases. The specificity of the MT2 antibody was ascertained by fluorescence microscopy using the anti-MT2 antibody in HEK 293 cells expressing recombinant MT2, in immunoblot experiments on membranes from MT2 expressing cells, and, finally, by immunoprecipitation experiments of the native MT2. MT2 immunoreactivity was studied in the hippocampus of 16 elderly control and 16 AD cases. In controls, MT2 was localized in pyramidal neurons of the hippocampal subfields CA1-4 and in some granular neurons of the stratum granulosum. The overall intensity of the MT2 staining was distinctly decreased in AD cases. The results indicate that MT2 may be involved in mediating the effects of melatonin in the human hippocampus, and this mechanism may be heavily impaired in AD.     

An immunohistochemical study of parvalbumin containing interneurons in the gerbil hippocampus after cerebral ischemia

We investigated postischemic changes of non-pyramidal neurons in the gerbil hippocampus 1 h - 7 days after 10 min of cerebral ischemia, with parvalbumin and microtubule-associated protein 2 (MAP2)-immunohistochemistry. Parvalbumin-immunoreactive interneurons in the hippocampus were unaffected up to 24 h after ischemia. A slight reduction of the immunoreactivity in neuronal processes was seen in the hippocampal CA1 sector 48 h after ischemia. Seven days after ischemia, a marked loss of parvalbumin-immunoreactive interneurons was observed in the hippocampal CA1 and CA3 sectors. Furthermore, reduced staining in the dentate granular and molecular layers was observed. MAP2-immunoreactive pyramidal neurons in the hippocampus were unchanged up to 48 h after ischemia. Seven days after ischemia, a severe loss of MAP2 immunoreactivity was found in the hippocampal CA1 and CA3 neurons and dentate hilar neurons. However, scattered CA1 neurons, most likely interneurons, preserved MAP2 immunoreactivity. The results demonstrate that transient cerebral ischemia can cause a loss of parvalbumin-immunoreactive interneurons in the hippocampus. Furthermore, some interneurons seem to lose parvalbumin synthesis. Although dentate granule cells are resistant to ischemia, considerable reductions of afferent input was suggested by parvalbumin staining.     

Galanin fragments and analogues: effects on glucose-stimulated insulin secretion from isolated rat islets.

Galanin occurs in intrapancreatic nerves and inhibits insulin secretion both in vivo and in vitro. To investigate which part of the galanin molecule accounts for this inhibition, we studied the effect of porcine galanin, four galanin fragments and three galanin analogues with substitutions of the 2nd amino acid in galanin, on glucose-stimulated insulin secretion from isolated rat islets cultured overnight. The islets were incubated for 1 h at 8.3 mM glucose. Porcine galanin1-29 inhibited insulin secretion at dose levels from 10(-8) M to 10(-6) M (p less than 0.001). Also, the galanin fragments GAL1-15, GAL2-29, and GAL3-29 significantly inhibited insulin secretion at and above concentrations of 10(-7) M (p less than 0.001), 10(-8) M (p less than 0.001), and 10(-7) M (p less than 0.001), respectively. Galanin analogues where the 2nd N-terminal amino acid had been changed from tryptophan to tyrosine (GAL-TYR2) or isoleucine (GAL-ILE2) inhibited insulin secretion, as did porcine galanin1-29, whereas after substitution with phenylalanine (GAL-PHE2) no effect was observed. Furthermore, the C-terminal fragment GAL10-29 did not influence insulin secretion. We conclude that the inhibitory action by galanin on glucose-stimulated insulin secretion from normal islets resides in the N-terminal part of the molecule. In contrast, the C-terminal part of galanin apparently does not influence insulin secretion.     

Immunoreactivity of calcium binding protein secretagogin in the human hippocampus is restricted to pyramidal neurons

Disturbed calcium homeostasis plays a crucial role in the aetiology of Alzheimer's disease (AD) and the aging process. We evaluated immunoreactivity of secretagogin, a recently cloned calcium binding protein, in hippocampus and adjacent entorhinal cortex of 30 neuropathologically examined post mortem brains (m:f=12:18; mean age, 79.8+/-15.1 years). The study group consisted of 15 cases fulfilling the criteria for high probability of AD according to the NIA-Reagan Institute Criteria and 15 cases with no to medium probability. Sections were incubated with secretagogin-specific antibodies and the number of immunoreactive neurons as well as staining intensities in both neurons and neuropil were assessed. Both cellular and neuropil immunoreactivity were restricted to subiculum and Ammons horn. Cellular immunoreactivity was further restricted to pyramidal neurons and showed a hierarchical distribution: the mean percentage of immunoreactive neurons was highest in sector CA3 (64.41%), followed by CA2 (44.09%), CA4 (34.38%), CA1 (10.9%), and the subiculum (2.92%; P<0.001, except CA2-CA4, P>0.05), while it did not differ significantly between groups with different degrees of AD pathology. The pattern of secretagogin immunoreactivity resembles that of calcium sensor proteins as it is restricted to a subset of neurons and therefore secretagogin could serve highly specialized tasks in neuronal calcium signalling.     

大鼠海马神经元及其线粒体增龄性改变的形态计量分析

目的探讨大鼠海马神经元及其线粒体形态结构增龄性改变的程度及性质。方法应用组织化学及电镜技术，通过计算机图像分析系统对海马神经元及其线粒体结构进行形态计量分析。结果海马ＣＡ１和ＣＡ３区神经元随增龄出现细胞皱缩，ＣＡ３区神经元数密度在老年组较青年组显著减少（Ｐ＜０．０５），ＣＡ１区神经元数密度各月龄组之间差异无显著性（Ｐ＞０．０５）。ＣＡ３区神经元胞体内线粒体体密度、数密度、比表面及嵴膜密度随增龄而减少，线粒体平均体积及平均截面积随增龄而增大。结论海马神经元及其线粒体形态结构随增龄发生显著性改变，这些形态结构的改变可能是大鼠海马老化的指征。     

Transient receptor potential vanilloid subtype 1 channel mediated neuropeptide secretion and depressor effects: role of endoplasmic reticulum associated Ca2+ release receptors in rat dorsal root ganglion neurons

OBJECTIVE: This study tests the hypothesis that the transient receptor potential vanilloid subtype 1 channel induced neuropeptide secretion and depressor response are mediated by, at least in part, activation of endoplasmic reticulum associated Ca release receptors, leading to increased cytosolic Ca in dorsal root ganglion neurons. METHODS/RESULTS: Bolus injection of capsaicin (10 or 50 microg/kg), a selective transient receptor potential vanilloid subtype 1 channel agonist, into anesthetized male Wistar rats caused a dose-dependent decrease in mean arterial pressure (P < 0.05). Capsaicin (50 microg/kg)-induced depressor effects and increase in plasma calcitonin gene related peptide (CGRP) levels (-29 +/- 2 mmHg, 82.2 +/- 5.0 pg/ml) were abolished by a selective transient receptor potential vanilloid subtype 1 channel antagonist, capsazepine (3 mg/kg, -4 +/- 1 mmHg, 41.8 +/- 4.4 pg/ml, P < 0.01), and attenuated by a selective ryanodine receptor antagonist, dantrolene (5 mg/kg, -12 +/- 1 mmHg, 57.2 +/- 2.6 pg/ml, P < 0.01), but unaffected by an inhibitor of endoplasmic reticulum Ca-ATPase, thapsigargin (50 microg/kg, -30 +/- 1 mmHg, 73.8 +/- 2.3 pg/ml, P > 0.05), or an antagonist of the inositol (1,4,5)-trisphosphate receptor, 2-aminoethoxydiphenyl borate (3 mg/kg, -34 +/- 5 mmHg, 69.0 +/- 3.7 pg/ml, P > 0.05). CGRP8-37 (1 mg/kg), a selective CGRP receptor antagonist, also blocked capsaicin-induced depressor effects. In contrast, dantrolene had no effect on CGRP (1 microg/kg)-induced depressor effects. In vitro, capsaicin (0.3 micromol/l) increased intracellular Ca concentrations and CGRP release from freshly isolated sensory neurons in dorsal root ganglion (P < 0.01), which were blocked by capsazepine (10 micromol/l) and attenuated by dantrolene but not thapsigargin or 2-aminoethoxydiphenyl borate. CONCLUSION: Our results indicate that transient receptor potential vanilloid subtype 1 channel activation triggers ryanodine receptor but not inositol (1,4,5)-trisphosphate receptor dependent Ca release from endoplasmic reticulum in dorsal root ganglion neurons, leading to increased CGRP release and consequent depressor effects.     

Ⅱ组代谢型谷氨酸受体阻断剂对大鼠海马神经元凋亡的影响

目的观察Ⅱ组代谢型谷氨酸受体阻断剂对大鼠海马神经元凋亡的影响。方法 SD大鼠24只随机分为假手术组、痴呆组和阻断剂组,每组8只。大鼠脑室注射凝聚肽Aβ_(25-35)5建立痴呆大鼠模型,阻断剂组1周后行阻断剂脑室注射,另2组等量注射人工脑脊液4周行Morris水迷宫测试;测试1周后取材行病理观察,采用TUNEL法检测海马CA1区锥体细胞凋亡情况。结果与假手术组比较,痴呆组大鼠平均潜伏期明显延长,平台象限滞留时间明显缩短(P0.05);与痴呆组比较,阻断剂组大鼠上述指标明显提高(P0.05)。与假手术组比较,痴呆组大鼠海马CA1区可见大量的凋亡锥体细胞,阳性锥体细胞数、总面积、平均吸光度(A)值明显升高(P0.05);与痴呆组比较,阻断剂组大鼠海马CA1区可见明显的阳性细胞和核固缩,但其阳性染色锥体细胞数、总面积、平均A值明显降低(P0.05)。结论Ⅱ组代谢型谷氨酸受体阻断剂可明显抑制痴呆引起的海马CA1区锥体细胞的凋亡,提示其可能通过影响细胞凋亡,参与痴呆的发病过程。     

Contextual learning requires synaptic AMPA receptor delivery in the hippocampus

The hippocampus plays a central role in learning and memory. Although synaptic delivery of AMPA-type glutamate receptors (AMPARs) contributes to experience-dependent synaptic strengthening, its role in hippocampus-dependent learning remains elusive. By combining viral-mediated in vivo gene delivery with in vitro patch-clamp recordings, we found that the inhibitory avoidance task, a hippocampus-dependent contextual fear-learning paradigm, delivered GluR1-containing AMPARs into CA3-CA1 synapses of the dorsal hippocampus. To block the synaptic delivery of endogenous AMPARs, we expressed a fragment of the GluR1-cytoplasmic tail (the 14-aa GluR1 membrane-proximal region with two serines mutated to phospho-mimicking aspartates: MPR-DD). MPR-DD prevented learning-driven synaptic AMPAR delivery in CA1 neurons. Bilateral expression of MPR-DD in the CA1 region of the rat impaired inhibitory avoidance learning, indicating that synaptic GluR1 trafficking in the CA1 region of the hippocampus is required for encoding contextual fear memories. The fraction of CA1 neurons that underwent synaptic strengthening positively correlated with the performance in the inhibitory avoidance fear memory task. These data suggest that the robustness of a contextual memory depends on the number of hippocampal neurons that participate in the encoding of a memory trace.     

Regional, cellular, and ultrastructural distribution of N-methyl-D-aspartate receptor subunit 1 in monkey hippocampus.

The regional, cellular, and subcellular distributions of N-methyl-D-aspartate (NMDA) receptor subunit 1, NMDAR-1, were investigated in monkey hippocampus by using a monoclonal antibody directed against a fusion protein corresponding to aa 660-811 of NMDAR-1. The data indicate that many neurons in each subfield of the hippocampus contain NMDAR-1 protein, although the intensity and distribution of immunoreactivity varied across regions, strata, and cellular compartments. In stratum lucidum of CA3, mossy fiber axons were immunoreactive for NMDAR-1, which may correspond to previously hypothesized presynaptic receptors. NMDAR-1-labeled postsynaptic profiles were present in stratum radiatum of CA3 but were largely absent from stratum lucidum. Such intraneuronal segregation of glutamate receptor subunits or classes may be spatially correlated with afferent systems that exhibit laminar segregation and terminate in different portions of the postsynaptic dendritic tree. For example, in CA3 pyramidal cells, NMDA receptors are postsynaptic in distal apical dendrites (stratum radiatum) where NMDA-dependent long-term potentiation in rats is mediated by associational/commissural afferents, and are absent from proximal apical dendrites (stratum lucidum), where NMDA-independent long-term potentiation is mediated by the mossy fiber input.     

Lack of decrease in hypothalamic and hippocampal glucocorticoid receptor mRNA during starvation

We have shown in a previous study that high corticosterone levels during repeated immobilization stress result in a reduction of glucocorticoid receptor (GR) mRNA in the hypothalamic paraventricular nucleus (PVN) and the hippocampus. The reduction of GR presumably accounts for loss of or decrease in glucocorticoid-negative feedback, and thus hyperfunction of the hypothalamic-pituitary-adrenocortical (HPA) axis persists during chronic stress. Starvation is a stress state in which the counterregulatory responses against the loss of food occur in the central nervous system. We explored the impact of starvation on the HPA axis, GR and mineralocorticoid receptor (MR) mRNAs in the hippocampus, the PVN, and the anterior pituitary (AP) of rats. Rats were starved for 4 days and sacrificed in the morning. Starved rats showed high levels of plasma corticosterone, whereas neither plasma corticotropin (ACTH), AP proopiomelanocortin (POMC) mRNA nor AP type-1 corticotropin-releasing hormone (CRH) receptor mRNA was altered in the starved rats. In the presence of high corticosterone, starvation resulted in a decrease in both CRH mRNA and type-1 CRH receptor mRNA in the PVN. Consistently, the starved rats did not show any changes in GR mRNA in the hippocampus (CA1-2, CA3, and dentate gyrus), the PVN or the AP despite the elevation of plasma corticosterone. A significant decrease in MR mRNA was seen in the dentate gyrus and the AP, but not in CA1-2, CA3 or PVN. The lack of reduction of GR may be one of the organism's counterregulatory responses during starvation, which allows an intact glucocorticoid negative feedback, thereby resulting in decreased anorectic neuropeptide levels, namely CRH, in the PVN. The results also indicate that GR mRNA in the hippocampus and other brain regions is not solely regulated by circulating glucocorticoids. The mechanism underlying the regulation of GR mRNA in the central nervous system remains to be clarified.     

短暂脑缺血对大鼠海马脑区锥体神经元外向整流氯通道功能的影响

目的 研究短暂前脑缺血对大鼠海马CA1和CA3脑区锥体神经元外向整流氯通道功能的影响.方法 采用膜片钳全细胞技术,在成年大鼠海马脑区锥体神经元上记录到可以被氯通道阻断剂DIDS阻断,具有外向整流特性的氯通道.结果 15 min前脑缺血再灌注6h和24 h后,海马CA1区锥体神经元氯通道电流持续性增强,而CA3区锥体神经元活动无明显改变.结论 氯通道功能增强可能参与海马CA1区锥体神经元在脑缺血后的迟发性死亡过程,并且为治疗缺血性脑损伤提供了新的手段.     

Local neurosteroid production in the hippocampus: influence on synaptic plasticity of memory

In neuroendocrinology, it is believed that steroid hormones are synthesized in the gonads and/or adrenal glands, and reach the brain via the blood circulation. In contrast to this view, we are in progress of demonstrating that estrogens and androgens are also synthesized locally by cytochrome P450s in the hippocampus, and that these steroids act rapidly to modulate neuronal synaptic plasticity. We demonstrated that estrogens were locally synthesized in the adult hippocampal neurons. In the pathway of steroidogenesis, cholesterol is converted to pregnenolone (by P450scc), dehydroepiandrosterone [by P450(17alpha)], androstenediol (by 17beta-hydroxysteroid dehydrogenase, 17beta-HSD), testosterone (by 3beta-HSD) and finally to estradiol (by P450arom) and dihydrotestosterone (by 5alpha-reductase). The basal concentration of estradiol in the hippocampus was approximately 1 nM, which was greater than that in blood plasma. Significant expression of mRNA for P450scc, P450(17alpha), P450arom, 17beta-HSD, 3beta-HSD and 5alpha-reductase was demonstrated by RT-PCR. Their mRNA levels in the hippocampus were 1/200-1/5,000 of those in the endocrine organs. Localization of P450(17alpha) and P450arom was observed in synapses in addition to endoplasmic reticulum of principal neurons using immunoelectron microscopy. Different from slow action of gonadal estradiol which reaches the brain via the blood circulation, hippocampal neuron-derived estradiol may act locally and rapidly within the neurons. For example, 1 nM 17beta-estradiol rapidly enhanced the long-term depression (LTD) not only in CA1 but also in CA3 and dentate gyrus. The density of thin spines was selectively increased within 2 h upon application of 1 nM estradiol in CA1 pyramidal neurons. Only ERalpha agonist propyl-pyrazole-trinyl-phenol induced the same enhancing effect as estradiol on both LTD and spinogenesis in the CA1. ERbeta agonist hydroxyphenyl-propionitrile suppressed LTD and did not affect spinogenesis. Localization of estrogen receptor ERalpha in spines in addition to nuclei of principal neurons implies that synaptic ERalpha can drive rapid modulation of synaptic plasticity by endogenous estradiol.