Chaperoning epigenetics: FKBP51 decreases the activity of DNMT1 and mediates epigenetic effects of the antidepressant paroxetine.

Epigenetic processes, such as DNA methylation, and molecular chaperones, including FK506-binding protein 51 (FKBP51), are independently implicated in stress-related mental disorders and antidepressant drug action. FKBP51 associates with cyclin-dependent kinase 5 (CDK5), which is one of several kinases that phosphorylates and activates DNA methyltransferase 1 (DNMT1). We searched for a functional link between FKBP51 (encoded by FKBP5) and DNMT1 in cells from mice and humans, including those from depressed patients, and found that FKBP51 competed with its close homolog FKBP52 for association with CDK5. In human embryonic kidney (HEK) 293 cells, expression of FKBP51 displaced FKBP52 from CDK5, decreased the interaction of CDK5 with DNMT1, reduced the phosphorylation and enzymatic activity of DNMT1, and diminished global DNA methylation. In mouse embryonic fibroblasts and primary mouse astrocytes, FKBP51 mediated several effects of paroxetine, namely, decreased the protein-protein interactions of DNMT1 with CDK5 and FKBP52, reduced phosphorylation of DNMT1, and decreased the methylation and increased the expression of the gene encoding brain-derived neurotrophic factor (Bdnf). In human peripheral blood cells, FKBP5 expression inversely correlated with both global and BDNF methylation. Peripheral blood cells isolated from depressed patients that were then treated ex vivo with paroxetine revealed that the abundance of BDNF positively correlated and phosphorylated DNMT1 inversely correlated with that of FKBP51 in cells and with clinical treatment success in patients, supporting the relevance of this FKBP51-directed pathway that prevents epigenetic suppression of gene expression.

Association of FKBP51 with priming of autophagy pathways and mediation of antidepressant treatment response: evidence in cells, mice, and humans.

BACKGROUND: FK506 binding protein 51 (FKBP51) is an Hsp90 co-chaperone and regulator of the glucocorticoid receptor, and consequently of stress physiology. Clinical studies suggest a genetic link between FKBP51 and antidepressant response in mood disorders; however, the underlying mechanisms remain elusive. The objective of this study was to elucidate the role of FKBP51 in the actions of antidepressants, with a particular focus on pathways of autophagy. METHODS AND FINDINGS: Established cell lines, primary neural cells, human blood cells of healthy individuals and patients with depression, and mice were treated with antidepressants. Mice were tested for several neuroendocrine and behavioral parameters. Protein interactions and autophagic pathway activity were mainly evaluated by co-immunoprecipitation and Western blots. We first show that the effects of acute antidepressant treatment on behavior are abolished in FKBP51 knockout (51KO) mice. Autophagic markers, such as the autophagy initiator Beclin1, were increased following acute antidepressant treatment in brains from wild-type, but not 51KO, animals. FKBP51 binds to Beclin1, changes decisive protein interactions and phosphorylation of Beclin1, and triggers autophagic pathways. Antidepressants and FKBP51 exhibited synergistic effects on these pathways. Using chronic social defeat as a depression-relevant stress model in combination with chronic paroxetine (PAR) treatment revealed that the stress response, as well as the effects of antidepressants on behavior and autophagic markers, depends on FKBP51. In human blood cells of healthy individuals, FKBP51 levels correlated with the potential of antidepressants to induce autophagic pathways. Importantly, the clinical antidepressant response of patients with depression (n = 51) could be predicted by the antidepressant response of autophagic markers in patient-derived peripheral blood lymphocytes cultivated and treated ex vivo (Beclin1/amitriptyline: r = 0.572, p = 0.003; Beclin1/PAR: r = 0.569, p = 0.004; Beclin1/fluoxetine: r = 0.454, p = 0.026; pAkt/amitriptyline: r =  -0.416, p = 0.006; pAkt/PAR: r =  -0.355, p = 0.021; LC3B-II/PAR: r = 0.453, p = 0.02), as well as by the lymphocytic expression levels of FKBP51 (r = 0.631, p<0.0001), pAkt (r =  -0.515, p = 0.003), and Beclin1 (r = 0.521, p = 0.002) at admission. Limitations of the study include the use of male mice only and the relatively low number of patients for protein analyses. CONCLUSIONS: To our knowledge, these findings provide the first evidence for the molecular mechanism of FKBP51 in priming autophagic pathways; this process is linked to the potency of at least some antidepressants. These newly discovered functions of FKBP51 also provide novel predictive markers for treatment outcome, consistent with physiological and potential clinical relevance. Please see later in the article for the Editors' Summary.

Hsp70 cochaperones HspBP1 and BAG-1M differentially regulate steroid hormone receptor function.

Hsp70 binding protein 1 (HspBP1) and Bcl2-associated athanogene 1 (BAG-1), the functional orthologous nucleotide exchange factors of the heat shock protein 70 kilodalton (Hsc70/Hsp70) chaperones, catalyze the release of ADP from Hsp70 while inducing different conformational changes of the ATPase domain of Hsp70. An appropriate exchange rate of ADP/ATP is crucial for chaperone-dependent protein folding processes. Among Hsp70 client proteins are steroid receptors such as the glucocorticoid receptor (GR), the mineralocorticoid receptor (MR), and the androgen receptor (AR). BAG-1 diversely affects steroid receptor activity, while to date the influence of HspBP1 on steroid receptor function is mostly unknown. Here, we compared the influence of HspBP1 and BAG-1M on Hsp70-mediated steroid receptor folding complexes and steroid receptor activity. Coimmunoprecipitation studies indicated preferential binding of Hsp40 and the steroid receptors to BAG-1M as compared to HspBP1. Furthermore, Hsp70 binding to the ligand-binding domain of GR was reduced in the presence of HspBP1 but not in the presence of BAG-1M as shown by pull-down assays. Reporter gene experiments revealed an inhibitory effect on GR, MR, and AR at a wide range of HspBP1 protein levels and at hormone concentrations at or approaching saturation. BAG-1M exhibited a transition from stimulatory effects at low BAG-1M levels to inhibitory effects at higher BAG-1M levels. Overall, BAG-1M and HspBP1 had differential impacts on the dynamic composition of steroid receptor folding complexes and on receptor function with important implications for steroid receptor physiology.

Comparison of glucocorticoid receptor and epigenetically regulated genes in proliferating versus growth-arrested Neuro-2a cells.

In recent years, it has been established that environmental stress leaves enduring traces at distinct sites on the chromatin, accompanied by permanent alterations of gene transcription. This process depends on duration and extent of the discharge of stress hormones. Here, we aimed at identifying genes that are both regulated by the glucocorticoid receptor (GR) and display epigenetic features of transcriptional control. We used neuronal Neuro-2a cells as model system; cells were transiently transfected with GR and exposed to dexamethasone (Dex) for 2 days, either under conditions of cell proliferation or after serum deprivation-induced growth arrest. In parallel, Neuro-2a cells were treated with the histone deacetylase inhibitor trichostatin A. Comparison of gene expression profiles obtained from whole-genome microarray analyses revealed a network of genes that were GR-dependent and under control of epigenetic factors. Gene set enrichment analysis was performed in order to obtain insight into functional mechanisms implicated in stress hormone physiology. Dex response varied between proliferating and growth-arrested cells; enrichment was found for genes associated with metabolic pathways in proliferating cells, and for genes linked to inflammation in growth-arrested cells. The set of genes that were regulated by Dex under both growth conditions (proliferation and arrest) as well as by trichostatin A - (under cell proliferation) was enriched in mRNA transcripts encoding proteins which play a role in development and homeostasis. In summary, this study introduces a conceptual approach and incipient proof-of-concept for the identification of candidate genes that might be epigenetically programmed by activated GR.

Antidepressants inhibit DNA methyltransferase 1 through reducing G9a levels.

The discovery of epigenetic processes as possible pivotal regulatory mechanisms in psychiatric diseases raised the question of how psychoactive drugs may impact the epigenetic machinery. In the present study we set out to explore the specificity and the mode of action of the reported inhibitory effect of the TCA (tricyclic antidepressant) amitriptyline on DNMT (DNA methyltransferase) activity in primary astrocytes from the rat cortex. We found that the impact on DNMT was shared by another TCA, imipramine, and by paroxetine, but not by venlafaxine or the mood stabilizers carbamazepine and valproic acid. DNMT activity in subventricular neural stem cells was refractory to the action of ADs (antidepressants). Among the established DNMTs, ADs primarily targeted DNMT1. The reduction of enzymatic DNMT1 activity was neither due to reduced DNMT1 expression nor due to direct drug interference. We tested putative DNMT1-inhibitory mechanisms and discovered that a known stimulator of DNMT1, the histone methyltransferase G9a, exhibited decreased protein levels and interactions with DNMT1 upon AD exposure. Adding recombinant G9a completely reversed the AD repressive effect on DNMT1 function. In conclusion, the present study presents a model where distinct ADs affect DNMT1 activity via G9a with important repercussions for possible novel treatment regimes.

The CpG island shore of the GLT-1 gene acts as a methylation-sensitive enhancer.

Astrocytic lineage commitment and brain region-dependent specialization of glia are partly ascribed to epigenetic processes. Clearance of glutamate is an essential task, which astrocytes assume in a temporal-spatial fashion by distinct glutamate transporter expression. Glutamate transporter subtype 1 (GLT-1) is predominant in cortex (CTX), while it plays an inferior role in cerebellum (CER). Here, we set out to identify regulatory elements that could account for the differences in brain region-specific activity as well as response to dexamethasone (DEX) or epigenetic factors. We found a distal promoter element at the shore of the CpG island exhibiting enhancer function in response to DEX in reporter gene assays. This shore region showed slight enrichment in repressive trimethyl-histone H3 (Lys27) and under-representation of acetyl-histone H4 (H4ac) marks in DEX nonresponsive CER astrocytes as determined by chromatin immunoprecipitation. In addition, CpG sites of the shore region displayed higher methylation in CER than in CTX cells. Targeted in vitro methylation of CpG sites within the shore abrogated the stimulatory effects of DEX. Interestingly, the shore was characterized by a pronounced epigenetic plasticity in CTX cells since DEX exposure elicited an increase of H4ac in CTX in comparison to DEX nonresponsive CER. The transcriptional activity of this region was also affected by histone deacetylase inhibitors in a methylation- and brain region-dependent manner. Together, our study highlights the impact of an epigenetically adaptive DNA element of the GLT-1 promoter being decisive for brain region-specific activity and reactivity.

Antidepressants encounter autophagy in neural cells.

The prevailing view of antidepressants' (ADs) mode of action primarily focuses on their impact on neurotransmitter circuits, since the corresponding transporters and receptors are common targets of ADs. However, mounting evidence points to additional target structures, which may either support the beneficial effects or account for undesired side effects of ADs. Recently, we analyzed the influence of three ADs of different classes on autophagy-related processes in primary astrocytes and neurons. While amitriptyline (AMI) and citalopram (CIT) upregulate the expression of autophagic markers such as LC3B-II or Beclin 1, venlafaxine fails to exert these effects. Autophagy triggered by AMI and CIT is functional in terms of autophagic flux, and is partially mediated by class III PtdIns 3-kinase- and ROS dependent-pathways. Together, our study's results highlight a novel mode of action of ADs beyond monoaminergic neurotransmission.

Antidepressant drugs diversely affect autophagy pathways in astrocytes and neurons--dissociation from cholesterol homeostasis.

In the search for antidepressants' (ADs') mechanisms of action beyond their influence on monoaminergic neurotransmission, we analyzed the effects of three structurally and pharmacologically different ADs on autophagic processes in rat primary astrocytes and neurons. Autophagy has a significant role in controlling protein turnover and energy supply. Both, the tricyclic AD amitriptyline (AMI) and the selective serotonin re-uptake inhibitor citalopram (CIT) induced autophagy as mirrored by pronounced upregulation and cellular redistribution of the marker LC3B-II. Redistribution was characterized by formation of LC3B-II-positive structures indicative of autophagosomes, which associated with AVs in a time-dependent manner. Deletion of Atg5, representing a central mediator of autophagy in MEFs, led to abrogation of AMI-induced LC3B-I/II conversion. By contrast, VEN, a selective serotonin and noradrenaline reuptake inhibitor, did not promote autophagic processes in either cell type. The stimulatory impact of AMI on autophagy partly involved class-III PI3 kinase-dependent pathways as 3-methyladenine slightly diminished the effects of AMI. Autophagic flux as defined by autophagosome turnover was vastly undisturbed, and degradation of long-lived proteins was augmented upon AMI treatment. Enhanced autophagy was dissociated from drug-induced alterations in cholesterol homeostasis. Subsequent to AMI- and CIT-mediated autophagy induction, neuronal and glial viability decreased, with neurons showing signs of apoptosis. In conclusion, we report that distinct ADs promote autophagy in neural cells, with important implications on energy homeostasis.

Valproate and amitriptyline exert common and divergent influences on global and gene promoter-specific chromatin modifications in rat primary astrocytes.

Aberrant biochemical processes in the brain frequently go along with subtle shifts of the cellular epigenetic profile that might support the pathogenic progression of psychiatric disorders. Although recent reports have implied the ability of certain antidepressants and mood stabilizers to modulate epigenetic parameters, studies comparing the actions of these compounds under the same conditions are lacking. In this study, we screened amitriptyline (AMI), venlafaxine, citalopram, as well as valproic acid (VPA), carbamazepine, and lamotrigine for their potential actions on global and local epigenetic modifications in rat primary astrocytes. Among all drugs, VPA exposure evoked the strongest global chromatin modifications, including histone H3/H4 hyperacetylation, 2MeH3K9 hypomethylation, and DNA demethylation, as determined by western blot and luminometric methylation analysis, respectively. CpG demethylation occurred independently of DNA methyltransferase (DNMT) suppression. Strikingly, AMI also induced slight cytosine demethylation, paralleled by the reduction in DNMT enzymatic activity, without affecting the global histone acetylation status. Locally, VPA-induced chromatin modifications were reflected at the glutamate transporter (GLT-1) promoter as shown by bisulfite sequencing and acetylated histone H4 chromatin immunoprecipitation analysis. Distinct CpG sites in the distal part of the GLT-1 promoter were demethylated and enriched in acetylated histone H4 in response to VPA. For the first time, we could show that these changes were associated with an enhanced transcription of this astrocyte-specific gene. In contrast, AMI failed to stimulate GLT-1 transcription and to alter promoter methylation levels. In conclusion, VPA and AMI globally exerted chromatin-modulating activities using different mechanisms that divergently precipitated at an astroglial gene locus.

XAP2 inhibits glucocorticoid receptor activity in mammalian cells.

XAP2 is member of a protein family sharing the TPR protein interaction motif. It displays close homology to the immunophilins FKBP51 and FKBP52 that act via the Hsp90 folding machinery to regulate the glucocorticoid receptor (GR). We show that XAP2 inhibits GR by reducing its responsiveness to hormone in transcriptional activation. The effect of XAP2 on GR requires its interaction with Hsp90 through the TPR motif. The PPIase-like region turned out to be enzymatically inactive. Thus, PPIase activity is not essential for the action of XAP2 on GR, similarly to FKBP51 and FKBP52.

DNA methylation dependent silencing of the human glutamate transporter EAAT2 gene in glial cells.

Glutamate is the major excitatory neurotransmitter in the CNS that is cleared from the extracellular space by a family of high-affinity glutamate transporters. The astroglial glutamate transporter EAAT2 is thought to carry out the uptake of the vast quantity of glutamate, and dysregulation of EAAT2 expression is involved in the pathogenesis of neurological disorders with marked excitotoxic components. Here, we present a novel epigenetic mechanism by which the human EAAT2 gene is kept in a silent state. Sequence inspection identified a classical CpG island at the EAAT2 promoter. Bisulfite analysis of the DNA methylation profile revealed that lack of EAAT2 expression in human glioma cell lines was associated with a densely methylated EAAT2 promoter. In contrast, EAAT2 positive normal human brain tissue used as reference displayed hypomethylation of the same promoter regions. In vitro methylation of EAAT2 promoter sequences indeed altered the binding properties of nuclear factors to the respective DNA sites as illustrated by electrophoretic mobility shift assay. Moreover, we observed a reduced activity of a methylated EAAT2 promoter construct as compared to the unmethylated control, both in a human glioma cell line and rodent primary astrocytes. Further supporting a role of DNA methylation for EAAT2 silencing, inhibition of DNA methyltransferases robustly enhanced EAAT2 mRNA transcription in several cell lines tested. In conclusion, the idea is put forward of an epigenetic mode of EAAT2 regulation based on the differential methylation of the gene promoter. (c) 2007 Wiley-Liss, Inc.

Differential promotion of glutamate transporter expression and function by glucocorticoids in astrocytes from various brain regions.

Steroids that activate glucocorticoid receptors (GRs) and mineralocorticoid receptors have important regulatory effects on neural development, plasticity, and the body's stress response. Here, we investigated the role of corticosteroids in regulating the expression of the glial glutamate transporters glial glutamate transporter-1 (GLT-1) and glutamate-aspartate transporter (GLAST) in rat primary astrocytes. The synthetic glucocorticoid dexamethasone provoked a marked increase of GLT-1 transcription and protein levels in cortical astrocytes, whereas GLAST expression remained unaffected. Up-regulation of GLT-1 expression was accompanied by an enhanced glutamate uptake, which could be blocked by the specific GLT-1 inhibitor dihydrokainate. The promoting effect of dexamethasone on GLT-1 gene expression and function was abolished by the GR antagonist mifepristone. A predominant role of the GR was further supported by the observation that corticosterone could elevate GLT-1 expression in a dose-dependent manner, whereas aldosterone, the physiological ligand of the mineralocorticoid receptor, exerted only weak effects even when applied at high concentrations. Moreover, we monitored brain region-specific differences, since all corticosteroids used in this study failed to alter the expression of GLT-1 in midbrain and cerebellar glia, although expression levels of both corticosteroid receptor subtypes were similar in all brain regions analyzed. Dexamethasone, however, modestly enhanced GLT-1 expression in cerebellar glia in combination with the DNA methyltransferase inhibitor 5-aza-2-deoxycytidine, suggesting that suppression of GLT-1 expression in cerebellar cultures may at least in part be epigenetically mediated by a DNA methylation-dependent process. Taken together, our data highlight a potential role for glucocorticoids in regulating GLT-1 gene expression during central nervous system development or pathophysiological processes including stress.

Caveolin and GLT-1 gene expression is reciprocally regulated in primary astrocytes: association of GLT-1 with non-caveolar lipid rafts.

Caveolae represent membrane microdomains acting as integrators of cellular signaling and functional processes. Caveolins are involved in the biogenesis of caveolae and regulate the activity of caveolae-associated proteins. Although caveolin proteins are found in the CNS, the regulation of caveolins in neural cells is poorly described. In the present study, we investigated different modes and mechanisms of caveolin gene regulation in primary rat astrocytes. We demonstrated that activation of cAMP-dependent signaling pathways led to a marked reduction in protein levels of caveolin-1/-2 in cortical astrocytes. Application of transforming growth factor-alpha (TGF-alpha) also resulted in a decrease of caveolin-1/-2 expression. Decreased caveolin protein levels were mirrored by diminished caveolin gene transcription. The repressive effect of TGF-alpha on caveolin-1 expression was MAP kinase-independent and partly mediated through the PI3-kinase pathway. Further downstream, inhibition of histone deacetylases abrogated TGF-alpha effects, suggesting that chromatin remodeling processes could contribute to caveolin-1 repression. Intriguingly, alterations of caveolin gene expression in response to cAMP or TGF-alpha coincided with reciprocal and brain-region specific changes in glial glutamate transporter GLT-1 expression. The reciprocal regulation of caveolin-1 and GLT-1 expression might be gated through a common PI3-kinase dependent pathway triggered by TGF-alpha. Finally, we showed that GLT-1 is located in non-caveolar lipid rafts of cortical astrocytes. In conclusion, this study highlights the occurrence of the reciprocal regulation of caveolin and GLT-1 expression during processes such as astrocyte differentiation via common signaling pathways. We also provide strong evidence that GLT-1 itself is concentrated in lipid rafts, inferring an important role for glial glutamate transporter function.

Brain region-specific neuroprotective action and signaling of corticotropin-releasing hormone in primary neurons.

CRH regulates the body's response to stressful stimuli by modulating the activity of the hypothalamic pituitary axis. In primary cultures and cell lines, CRH also acts as a potent neuroprotective factor in response to a number of toxins. Using primary neuronal cultures from the cerebellum, cerebral cortex, and hippocampus, we demonstrate that CRH exerts a brain region-specific neuroprotective effect on amyloid beta 25-35 toxicity. At low CRH concentrations (10(-8) M), neuroprotective effects can be observed only in cerebellar and hippocampal cultures, but a higher CRH concentration (10(-7) M) additionally led to the protection of cortical neurons. These neuroprotective effects were inhibited by H89, a specific protein kinase A inhibitor. Western blot analysis, carried out using phospho-specific antibodies directed against MAPK, cAMP response element-binding protein (CREB), and glycogen synthase kinase (GSK)3 beta also resulted in brain legion-specific differences regarding intracellular signaling. Correlating with cell survival, low CRH concentrations resulted in activation of the CREB pathway and inactivation of GSK3 beta in cerebellar and hippocampal cultures, but higher concentrations additionally resulted in activated CREB and inactivated GSK3 beta in cortical cultures. In contrast, MAPK activation occurred only in cortical neurons. Differences in signaling were found to be independent of receptor expression levels because RT-PCR analysis indicated no region-specific differences in CRHR1 mRNA expression.

The human SPANX multigene family: genomic organization, alignment and expression in male germ cells and tumor cell lines.

Multigenicity is one of the features of cancer/testis-associated genes. In the present study we analyzed the number and expression of genes of the SPANX(CTp11) family of cancer/testis-associated genes. Genomic database analysis, next to the four previously described SPANX genes, revealed the presence of a novel gene: SPANXE. Moreover, we detected an allelic variant of SPANXB resulting in one amino acid substitution in the encoded protein: SPANXB'. Most SPANX genes are present on contig NT_011574 located at Xq26.3-Xq27.1. Based on expressed sequence tag databases and RT-PCR analysis three additional novel SPANX sequences were identified, though not represented so far in the human genome sequence. Sequence alignments justify a subdivision of this gene family based on the absence (SPANXA-likes) or presence (SPANXB) of an 18 base pair sequence stretch in the open reading frame. The alignments also reveal an unusually high level (99%) of intron homology. Furthermore, the nucleotide variations in the open reading frame almost all lead to amino acid substitutions. Southern blot and database analyses indicate that SPANX sequences are exclusively present in primates. With RT-PCR analysis on human sperm cell precursors and tumor cell lines most family members could be detected. SPANXB was only found in sperm cell precursors and could not be detected in the tumor cell lines tested. Overall SPANXA was the most frequently expressed SPANX variant in melanoma and glioblastoma cell lines.

Functional interaction of estrogen receptor alpha and caveolin isoforms in neuronal SK-N-MC cells.

Estrogen receptors (ERs) are expressed in neuronal cells and exhibit a wide variety of activities in the central nervous system. The actions of ERs are regulated in a hormone-dependent manner as well as by a number of co-activators and -repressors. A recently identified co-activator of ERalpha is caveolin-1 which has been shown to mediate the ligand-independent activation of this steroid receptor. In the present study we have demonstrated that neuronal SK-N-MC cells lacking functional ERalpha show high levels of caveolin-1/-2 specific transcripts and proteins. Ectopic expression of ERalpha in SK-N-MC cells leads to the transcriptional suppression of caveolin-1 and -2 genes. This silencing event is accompanied by changes in the methylation pattern of the caveolin-1 promoter. Certain CpG dinucleotides were methylated in the caveolin-1 promoter region of the SK-ERalpha cells whereas the same sites were non-methylated in control SK-N-MC cells, implicating a gene silencing mechanism including hypermethylation of DNA. In addition, inhibitors of methyltransferases or histone deacetylases, enzymes involved in the establishment and maintenance of silenced chromatin status, partially restored caveolin transcription in SK-ERalpha cells. In conclusion, our observations provide a possible mechanism of negative feedback regulation of ERalpha co-activator caveolin by the steroid receptor itself in this cellular model.

Estrogen receptor alpha-mediated silencing of caveolin gene expression in neuronal cells.

Estrogen receptors (ER alpha/ER beta) are expressed in neuronal cells and exhibit a variety of activities in the central nervous system. ER activity is regulated in a ligand-dependent manner and by co-regulatory factors. Caveolin-1 is a recently identified co-activator of ER alpha mediating the ligand-independent activation of this steroid receptor. Here the influence of ERs on caveolin expression in human neuroblastoma SK-N-MC cells as well as in rodent brain was investigated. We found that ectopic expression of ER alpha in SK-N-MC cells (SK-ER alpha) leads to a ligand-independent transcriptional suppression of caveolin-1/-2 genes. This suppression is specifically mediated by ER alpha and not ER beta because ER beta counteracts the observed caveolin-silencing process. Interestingly, decreased caveolin expression in SK-ER alpha is accompanied by changes in the methylation pattern of caveolin promoters. The analysis of selected promoter regions of the human caveolin-1 gene showed that certain CpG dinucleotides were hypermethylated in SK-ER alpha cells, whereas the same sites were unmethylated in control, ER beta-, and ER alpha/beta co-expressing SK-N-MC cells. Inhibition of DNA methylation or histone deacetylation led to partial re-expression of caveolin-1/-2 genes in SK-ER alpha. In vivo analysis revealed a down-regulation of caveolin-1 expression after long term estrogen exposure in certain regions of the mouse brain. In conclusion, we have shown for the first time that ER alpha and not ER beta silences caveolin-1/-2 expression in an epigenetic fashion in neuronal cells. The observed mechanism of gene silencing by ER alpha may have implications for the transcriptional regulation of further ER alpha target genes.