Journal of Steroid Biochemistry & Molecular Biology 145 (2015) 85–93

Selective estrogen-receptor modulators suppress microglial activationand neuronal cell death via an estrogen receptor-dependent pathway

Yasuhiro Ishihara a,*, Kouichi Itoh b, Atsuhiko Ishida a, Takeshi Yamazaki a

a Laboratory of Molecular Brain Science, Graduate School of Integrated Arts and Sciences, Hiroshima University, Higashi-Hiroshima 739-8521, Japanb Laboratory for Brain Science, Kagawa School of Pharmaceutical Sciences, Tokushima Bunri University, Kagawa 769-2193, Japan

A R T I C L E I N F O

Article history:Received 29 July 2014Received in revised form 25 September 2014Accepted 5 October 2014Available online 8 October 2014

A B S T R A C T

Growing evidence shows that steroid hormones, especially 17b-estradiol (E2), protect neuronal cells byattenuating excess activation of microglia. However, the use of E2 in the clinic is controversial because ofits peripheral actions in reproductive organs and its potential to increase risk for endometrial cancer andbreast cancer. Selective estrogen-receptor modulators (SERMs) bind to estrogen receptors (ERs), but theireffects as ER agonists or antagonists are dependent on the target tissue. SERMs pose very little cancer riskas a result of their anti-estrogen action in reproductive organs, but their action in the brain is not wellunderstood. In this study, we investigated the effects of SERMs tamoxifen (Tam) and raloxifene (Rlx) onmicroglial activation and subsequent neuronal injury. Tam and Rlx suppressed the increasesin proinflammatory cytokines and chemokine expression that were induced by lipopolysaccharide(LPS) in rat primary microglia cultures. The microglial-conditioned media pretreated with Tam or Rlxsignificantly attenuated cellular injury in SH-SY5Y cells elicited by microglial-conditioned media treatedwith LPS alone. Rat primary microglia expressed ERa and ERb primarily in the nucleus, and thus weexamined the involvement of ERs in the suppressive action of Tam and Rlx on microglial activation using apure ER antagonist, ICI182,780. Pretreatment with ICI182,780 abolished the suppressive effects of SERMson microglial activation, as well as their protective action on SH-SY5Y cells. A luciferase assay using avector with three estrogen response elements (EREs) revealed that Tam and Rlx activatedERE-mediated transcription in rat primary microglia. Taken together, these results suggest that Tamand Rlx suppress microglial activation and subsequent neuronal cell death via an ER-mediatedtranscription pathway. SERMs could represent a novel therapeutic strategy for disorders of the centralnervous system based on their ability to suppress neuroinflammation.

ã 2014 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Journal of Steroid Biochemistry & Molecular Biology

journal homepage: www.else vie r .com/locate / j sbmb

1. Introduction

Microglia are the primary immune cells of the central nervoussystem (CNS) and are activated quickly in response to externalpathogens or cell debris, after which they act by releasinginflammatory factors or engulfing foreign bodies to mediate the

Abbreviations: CNS, central nervous system; E2, 17b-estradiol; ER, estrogenreceptor; ERE, estrogen response element; GAPDH, glyceraldehyde-3-phosphatedehydrogenase; IL1b, interleukin-1b; LPS, lipopolysaccharide; MCP-1, monocytechemoattractant protein 1; MIP-2a, macrophage inflammatory protein 2a; MTT, 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide; Rlx, raloxifene;SERM, selective estrogen-receptor modulator; Tam, tamoxifen; TNFa, tumornecrosis factor a.* Corresponding author at: Laboratory of Molecular Brain Science, Graduate

School of Integrated Arts and Sciences, Hiroshima University, 1-7-1, Kagamiyama,Higashi-Hiroshima, Hiroshima 739-8521, Japan. Tel.: +81 82 424 6500;fax: +81 82 424 0759.

E-mail address: ishiyasu@hiroshima-u.ac.jp (Y. Ishihara).

http://dx.doi.org/10.1016/j.jsbmb.2014.10.0020960-0760/ã 2014 Elsevier Ltd. All rights reserved.

inflammatory response. However, excessive activation of micro-glia may be harmful for host cells; microglia can promote thedevelopment of some neuronal diseases by producing largeamounts of cytokines and other inflammatory molecules such astumor necrosis factor-a (TNFa), interleukin-1b (IL-1b), nitricoxide, and reactive oxygen species. Indeed, activated microgliaare reported to be associated with the pathogenesis of Parkinson’sdisease [1] and Alzheimer’s disease [2]. In these diseases, a largenumber of activated microglial cells, which have the potential torelease inflammatory cytokines, gather around lesions, indicatingthat microglia-mediated inflammatory responses could be amechanism in a variety of neurodegenerative diseases. Inaddition, microglia with abnormal activity are reportedlyinvolved in brain ischemia–reperfusion injury, trauma, epilepsy,depression, and schizophrenia [3–5]. Therefore, the regulation ofmicroglial activity is crucial to maintain physiological function inthe brain and to prevent the onset and development of CNSdisorders.

Table 1Primer sequences used in this study.

Gene Forward primer Reverse primer

TNFa AGCCCTGGTATGAGCCCATGTA CCGGACTCCGTGATGTCTAAGTIL-1b CACCTCTCAAGCAGAGCACAGA ACGGGTTCCATGGTGAAGTCMCP-1 TGTCTCAGCCAGATGCAGTT CAGCCGACTCATTGGGATCAMIP-2a CCCTCCTGTGCTCAAGACTC CCACAACAACCCCTGTACCCGAPDH AACGACCCCTTCATTGACCT CCTTGACTGTGCCGTTGAACT

86 Y. Ishihara et al. / Journal of Steroid Biochemistry & Molecular Biology 145 (2015) 85–93

17b-Estradiol (E2),which is synthesized in, and secreted from,peripheral endocrine glands such as the ovary, the placenta, and theadrenal cortex, passes through the blood–brain barrier to performdiverse functions in the CNS. In addition, the brain possesses aninherent endocrine system and de novo synthesizes E2. Recently,increasing evidence has shown that E2 protect neurons from excessor prolonged inflammation in the brain. Treatment withE2 suppresses inflammatory cytokine expression and nitric oxideproduction induced by lipopolysaccharide (LPS) in microglia [6,7].Thesesuppressive effectsare mediatedvia theestrogenreceptor(ER)andactby blockingDNA bindingand transcriptional activityofNF-kBp65 by preventing its nuclear translocation [8]. E2 has also beenreported to inhibit neuroinflammation in an ER-dependent mannerin studies using in vivo models of CNS diseases [9,10]. However,although the neuroprotective and anti-inflammatory effects of E2 inthe brain are well documented in animal models of neurodegenera-tive disorders and other diseases, the use of E2 in the clinic iscontroversialbecauseof its peripheralactions inreproductive organsand its potential to increase risk of endometrial cancer and breastcancer. Therefore, alternative compounds that share some mecha-nismsofactionwith E2mightrepresenttreatmentsforCNSdisorderswith a better safety profile than E2.

Selective estrogen receptor modulators (SERMs) include com-pounds with mixed agonist/antagonist action at the ER. SERMsbind to ERs, but their action as an ER agonist or antagonist isdependent on the target tissue and cell types, and the nature of thisrelationship varies with SERM compounds [11]. Tamoxifen (Tam),which was the first SERM compound to be used clinically, has beenwidely applied in the treatment of breast cancer, in which itfunctions as an ER antagonist. In contrast, Tam has estrogen-likecharacteristics in skeletal tissue [12]. Raloxifene (Rlx) is a second-generation SERM that was developed to function as an ER agonistin bone and as an ER antagonist in reproductive tissues, and isprescribed for the prevention and treatment of postmenopausalosteoporosis [13].

Some groups have reported neuroprotective effects of Tam andRlx using in vivo experimental models. Treatment with Tamsuppressed experimental spinal cord injury through attenuationof TNFa and IL-1b levels [14], and induced regeneration of the ratsensory cortex after a penetrating brain injury [15]. Rlx decreasedthe number of microglia and astrocytes in aged mice [16]. Tamand Rlx reduced the number of microglia in rats withintraperitoneal administration of LPS [17] as well as brain trauma[18]. Liu et al. demonstrated that Tam attenuated microglialactivation and brain injury elicited by irradiation [19]. Further-more, in astrocytes, SERMs including Tam and Rlx suppressed theexpression of interleukin-6 and interferon-g-inducible protein-10 induced by LPS via attenuating nuclear translocation of NF-kB[20]. However, the effects of SERMs in the brain, especially inmicroglia, are still not well understood. In this study, weexamined the action of SERMs Tam and Rlx on microglialactivation induced by LPS, with a focus on ER in rat primarymicroglia.

2. Materials and methods

2.1. Materials

LPS from Escherichia coli 026:B6, ICI182,780, and E2 wereobtained from Sigma–Aldrich (St. Louis, MO, USA). Tamoxifencitrate was purchased from Wako Pure Chemical Industries, Ltd.(Osaka, Japan). Raloxifene hydrochloride was purchased fromCayman Chemical (Ann Arbor, MI, USA). All other chemicals wereobtained from Wako Pure Chemical Industries, Nacalai Tesque(Kyoto, Japan), or Sigma–Aldrich and were of reagent grade.

2.2. Animals

All animal procedures were performed in accordance with theFundamental Guidelines for Proper Conduct of Animal Experi-ments and Related Activities in Academic Research Institutionsunder the jurisdiction of the Ministry of Education, Culture, Sports,Science, and Technology (Japan), and the Animal Care and UseCommittee of Hiroshima University (Hiroshima, Japan). PregnantWistar rats were obtained from Kyudo (Kumamoto, Japan) andwere maintained in a temperature-controlled animal facility with12 h light–dark cycles.

2.3. Primary microglia culture

Primary microglia cultures were prepared from 1-2-day-oldWistar rats (both males and females were used for microgliapreparation) according to the well-established “shaking off”method [21]. Brains were excised and the meninges were carefullyremoved. The tissue was dissociated by passing it through a 250-mm nylon mesh with the aid of a rubber policeman. After washingwith Hanks’ balanced salt solution, the cell suspension wastriturated with a Pasteur pipette and plated in a poly-L-lysine-coated 75 cm2 plastic culture flask at a density of 1 brain per flask in10 mL tissue culture medium, which consisted of Dulbecco’smodified Eagle’s medium (DMEM) supplemented with 10% fetalbovine serum (FBS), 5 mg/mL insulin, and 0.5 ng/mL granulocyte-macrophage colony-stimulating factor. Cultures were maintainedin a humidified atmosphere of 5% CO2 and 95% air at 37 �C. Themedium was changed 2 or 3 times per week. After 9–12 days,microglia were harvested by shaking the flask at 150 rpm for15 min and seeded at a density of 1 �105 cells/cm2 in microgliaculture medium (DMEM with 10% charcoal-stripped FBS and 5 mg/mL insulin). Culture medium was changed to remove non-adheringcells 30 min after seeding. After culturing for 24 h, reagents wereadded to the microglia. The cultures of isolated microglia wereuniformly immunopositive for CD11b and contained greater than95% microglial cells.

2.4. Total RNA extraction and real-time PCR

Determination of mRNA levels was performed as previouslydescribed [22]. Briefly, total RNA was extracted from microgliausing a High Pure RNA Isolation Kit (Roche Diagnostics K.K., Tokyo,Japan). Single-stranded cDNA was synthesized from 0.5 mg of totalRNA following the ReverTra Ace protocol (Toyobo, Osaka, Japan)with a random primer (9-mer; Takara Bio Shiga, Japan). Real-timePCR was performed using a LightCycler instrument (RocheDiagnostics) with Sybr Green Real-time PCR master mix (Toyobo).The primer sequences used in this study are listed in Table 1. Thelevels of mRNA were normalized to the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA level, and the valuesof treated cells were divided by those of untreated cells to giverelative mRNA levels.

Fig. 1. Toxicity of SERMs, Tam and Rlx on rat primary microglia. Rat primarymicroglia were pretreated with 0.3, 1, 3, or 10 mM of Tam or Rlx for 24 h, and thencell viability was evaluated by MTT assay. The values are the means � S.E.M. of fiveseparate experiments. **P < 0.01 vs. the untreated group.

Y. Ishihara et al. / Journal of Steroid Biochemistry & Molecular Biology 145 (2015) 85–93 87

2.5. Enzyme-linked immunosorbent assay (ELISA)

The levels of TNFa and IL-1b present in the cultured super-natants of microglia were evaluated with the TNFa and IL-1b MiniELISA Development Kit (PeproTech, Rocky Hill, NJ, USA) accordingto the manufacturer’s instructions.

2.6. Microglia-conditioned media

The culture medium of the microglia was replaced with newmedium immediately before the LPS treatment. SERMs or otherinhibitors were added again in the culture. The conditioned mediawere collected 24 h after administration of LPS, centrifuged toremove cell debris, and stored at �80 �C until use.

2.7. SH-SY5Y cell culture

Human neuroblastoma SH-SY5Y cells (CRL-2266, AmericanType Culture Collection, Manassas, VA, USA) were placed intoculture dishes and cultured in DMEM supplemented with 10% FBS,100 units/mL penicillin, and 100 mg/mL streptomycin at 37 �C in ahumidified atmosphere of 5% CO2 and 95% air. The media werereplaced every 3–4 days. Cells were sub-cultured when theyreached 80–90% confluence.

2.8. Measurement of cell viability

Twenty-four hours after treatment of SH-SY5Y cells withconditioned media, cell viability was measured by the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide (MTT)assay. MTT was added to the cell-culture medium at a finalconcentration of 0.5 mg/mL. After incubating the plates at 37 �C for2 h in a 5% CO2 and 95% air atmosphere, the resulting purple pelletswere dissolved in a 40 mM HCl/isopropanol solution. The absor-bance was read at 570 nm with a microplate reader (Bio-RadHercules, CA, USA). The percentage of cell survival was calculatedwith the value of the untreated cells taken as 100%.

2.9. Immunocytochemistry

Microglia cultured on glass cover slips were fixed with 4%paraformaldehyde at room temperature for 30 min. After excessparaformaldehyde was removed by 50 mM glycine, cells werepermeabilized with 0.2% Triton X-100 for 15 min and then blockedusing 1% bovine serum albumin (BSA). Rabbit polyclonal anti-bodies directed against ERa (1:50, Abcam, Tokyo, Japan) and ERb(1:50, Abcom) in PBS with 1% BSA and 0.1% Triton X-100 were thenapplied for 3 h, followed by incubation for 1 h in Alexa488anti-rabbit IgG (1:250, Molecular Probes, Eugene, OR, USA). Nucleiwere stained with 1 mg/mL Hoechst 33342 for 5 min. Images wereobtained with an inverted fluorescent microscope (BZ-9000,Keyence, Osaka, Japan).

2.10. Immunoblotting

Microglia were collected and lysed with RIPA buffer (25 mMTris–HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate,and 0.1% SDS). Equal amounts of protein were loaded and separatedusing SDS-PAGE with a 10% or 12% (w/v) polyacrylamide gel andtransferred onto a polyvinylidene difluoride (PVDF) membrane.The blocked membranes were incubated with the primaryantibodies: anti-ERa (1:1000, Abcam), anti-ERb (1:1000, Abcam),and anti-a-tubulin (1:2000, Sigma–Aldrich, St. Louis, MO, USA).The membranes were incubated with peroxide-conjugated sec-ondary antibodies (Thermo Fisher Scientific, Waltham, MA, USA)and then visualized using peroxide substrates (SuperSignal West

Pico, Thermo Fisher Scientific). The band intensity was quantifiedusing Image J software (NIH, MD, USA).

2.11. Plasmid construction and luciferase assay

The sequence of the estrogen-responsive element (ERE) wastaken from the report of Ciana et al. [23]. The 2 oligonucleotidesincluding 3 canonical estrogen-responsive elements (3 � ERE)spaced by 8 bp, 50-AGGTCACAGTGACCTAGATCCGCAGGTCACAGT-GACCTAGATCCGCAGGTCACAGTGACCT-30 and 50-AGGTCACTGT-GACCTGCGGATCTAGGTCACTGTGACCTGCGGATCTAGGTCACTGT-GACCT-30, were annealed and the resulting double-strand oligowas ligated into the EcoRV site of the pGL4.24 vector (Promega,Madison, WI, USA), and the resulting construct was namedpGL4.24-3 � ERE. The pGL4.24 and pGL4.24-3 � ERE constructswere transfected into rat primary microglia using Lipofectamine2000 (Invitrogen, Carlsbad, CA, USA) according to the manufac-turer’s instructions. Twenty-four hours after transfection, cellswere treated with SERMs or E2. Luciferase reporter activity wasmeasured using the Luciferase Assay System (Promega) and theGloMax 20/20 Luminometer (Promega). Luciferase activity wasnormalized to total protein levels.

2.12. Statistical analyses

All data are expressed as the mean � standard error of the mean(S.E.M.). The statistical analyses were performed using one-wayanalysis of variance (ANOVA), followed by the Student’s t-test orDunnett’s test. Multiple comparisons were made using the Holm’sor Bonferroni correction methods. Probability (P) values of<0.05 were considered to be statistically significant.

3. Results

3.1. SERMs Tam and Rlx suppressed the expression of proinflammatorymolecules induced by LPS in rat primary microglia

When rat primary microglia were treated with variousconcentrations of Tam and Rlx for 24 h, decreases in the viabilityof primary microglia were observed by treatment with 10 mM ofTam or Rlx (Fig. 1). Therefore, we have used SERMs at theconcentration of 0.3, 1, and 3 mM in this study.

Fig. 2. Suppressive effects of Tam and Rlx on the expression of cytokines and chemokines induced by LPS. Rat primary microglia were pretreated with 0.3,1, or 3 mM of Tam orRlx for 24 h and then stimulated with 10 ng/mL LPS for 6 h. Levels of mRNA of (A) TNFa, (B) IL-1b, (C) MCP-1, and (D) MIP-2a were determined by real-time PCR. Amounts ofmRNA are represented as fold-change from those measured in untreated cells. The values are the means � S.E.M. of five separate experiments. **P < 0.01 vs. the untreatedgroup. #P < 0.05, ##P < 0.01 vs. the LPS-treated group.

88 Y. Ishihara et al. / Journal of Steroid Biochemistry & Molecular Biology 145 (2015) 85–93

After stimulation of rat primary microglia with LPS for 6 h,mRNA levels of proinflammatory cytokines TNFa and IL-1b,chemokines, monocyte chemoattractant protein 1 (MCP-1), andmacrophage inflammatory protein 2a (MIP-2a) were increased(Fig. 2A–D). In addition, levels of TNFa and IL-1b released fromthe microglia into the culture medium were also elevated,indicating that microglia were activated by treatment with LPS(Fig. 3A and B). Treatment of microglia with Tam (0.3, 1, and3 mM) or Rlx (0.3, 1, and 3 mM) suppressed the increase inmRNA expression of TNFa, IL-1b, MCP-1, and MIP-2a that wereelicited by treatment with LPS (Fig. 2A–D). The action of Tam onthe attenuation of microglial activation showed dose-depen-dency and was strongest at 3 mM, while the suppressive effectsof Rlx on microglial activation were strongest at the 1 mMconcentration. Treatment with Tam or Rlx also attenuated thesecretion of TNFa and IL-1b into the medium (Fig. 3A and B).The suppressive effects of Rlx on microglial activation tended tobe greater than those of Tam but there were no significantdifferences between the action of Tam and Rlx (Figs. 2 and 3).Treatment with Tam or Rlx alone did not affect mRNA levels ofinflammatory molecules (Fig. 2A–D) and the amounts ofinflammatory cytokines in the culture medium (Fig. 3A andB). These results indicate that SERMs Tam and Rlx suppressedthe activation of rat primary microglia.

3.2. SERMs inhibit neuronal injury by attenuating microglialactivation by LPS

The conditioned medium from rat primary microglia culturestreated with LPS for 24-h induced cell death in SH-SY5Y humanneuroblastoma cells 24 h after incubation (Fig. 4B). Becausetreatment of SH-SY5Y cells with LPS produced no cytotoxicity(Fig. 4A), factor(s) that were released from microglia in response toLPS mediated cell death in SH-SY5Y cells. When the microglialculture from which the conditioned medium was collected waspretreated with Tam or Rlx before the LPS addition, cell deathinduced by this conditioned medium was significantly lower thanthat induced by the conditioned medium with LPS alone (Fig. 4B).The inhibitory effects of Tam and Rlx on cell death were highlycorrelated with their suppression of microglial activation(Figs. 2 and 4B). Therefore, these results suggest that Tam andRlx suppressed SH-SY5Y cell death by inhibiting microglialactivation elicited by LPS.

3.3. ERa and ERb were primarily expressed in the nuclei of rat primarymicroglia

Immunocytochemistry showed that ERa immunoreactivitywas detected primarily in the nucleus in untreated and

Fig. 3. Attenuation of TNFa and IL-1b release from microglia by treatment withTam and Rlx. Rat primary microglia were pretreated with 1 mM Tam or Rlx for 24 hand then stimulated with 10 ng/mL LPS for 24 h. Levels of (A) TNFa and (B) IL-1b inthe culture medium were detected by ELISA. The values are the means � S.E.M. offour separate experiments. **P < 0.01 vs. the untreated group. ##P < 0.01 vs. the LPS-treated group.

Fig. 4. Protective action of Tam and Rlx on neuronal cell injury elicited by microglialconditioned medium with LPS treatment. (A) SH-SY5Y cells were pretreated with0.3, 1, or 3 mM Tam or Rlx for 24 h and then treated with 10 ng/mL LPS for 24 h. Cellviability was evaluated by the MTT assay. (B) Rat primary microglia were pretreatedwith 0.3, 1, or 3 mM Tam or Rlx for 24 h and then stimulated with 10 ng/mL LPS for24 h. SH-SY5Y cells were treated with microglial conditioned medium for 24 h,followed by the measurement of cell death by the MTT assay. The values are themeans � S.E.M. of seven separate experiments. **P < 0.01 vs. the untreated group.#P < 0.05, ##P < 0.01 vs. the LPS-treated group.

Y. Ishihara et al. / Journal of Steroid Biochemistry & Molecular Biology 145 (2015) 85–93 89

LPS-treated microglia, indicating that ERa is largely expressed inmicroglial nuclei (Fig. 5A). ERa was considered to be expressed atlow levels in the cytosol because a weak ERa signal was detected(Fig. 5A). ERb was also detected primarily in the nucleus andmodestly in the cytosol of rat primary microglia cultured cells(Fig. 5B). There was no change in the localization of ERa and ERbduring LPS stimulation (Fig. 5A and B). Immunoblotting to measureexpression levels of ERs showed that treatment with LPS did notaffect the expression of ERa and ERb in microglia (Fig. 5C and D).Together, these data indicate that rat primary microglia expressERa and ERb primarily in the nucleus, and the abundance of thesereceptors is not influenced by LPS treatment.

3.4. The suppressive effects of Tam and Rlx on microglial activationwere mediated by an ER-dependent transcription pathway

When rat primary microglia were pretreated with ER antagonistICI182,780 and then treated with Tam or Rlx for 24 h, andsubsequently stimulated with LPS for 6 h, ICI182,780 nearlycompletely abolished the suppressive effects of Tam and Rlx on

increased TNFa and IL-1bmRNA expression by LPS in microglia(Fig. 6A), indicating that ER antagonism counteracted microglialinactivation by Tam or Rlx. Furthermore, the conditioned mediafrom cells pretreated with ICI182,780 before treatment with Tamor Rlx and the addition of LPS also abolished the protective actionof Tam and Rlx on SH-SY5Y cell death induced by LPS-treated cellconditioned medium (Fig. 6B). Treatment with ICI182,780 alonedid not affect inflammatory molecule expression (Fig. 6A) as wellas cell viability (Fig. 6B).

We next examined whether Tam and Rlx activate ER-dependenttranscription. ERs interact with EREs in gene promoters to activatedownstream gene transcription [24]. Thus, we performed aluciferase assay using a luciferase-expressing vector cloned with3 EREs (pGL4.24-3 � ERE) at the upstream of the luciferase gene.Treatment of rat primary microglia transfected with pGL4.24-3 � ERE with 10 nM E2 increased luciferase activity compared withuntreated cells, indicating that pGL4.24-3 � ERE activates micro-glial ER-ERE-dependent transcription of the luciferase gene (Fig. 7).Treatment of the microglia with Tam or Rlx induced significantlyhigher luciferase activity than that observed in untreated cells. Rlxtended to induce stronger luciferase activity than Tam at the sameconcentration, but no significant difference was detected betweenthe treatments. Taken together, these results suggest that Tam andRlx suppress microglial activation elicited by LPS via activatingER-dependent transcription.

Fig. 5. ER expression in untreated and LPS-treated microglia. Rat primary microglia were treated with 10 ng/mL LPS for 24 h. (A, B) Localization of ERa (A) and ERb (B) wasexamined by immunocytochemistry. Hoechst33342 was used to stain nuclei. (C, D) Levels of ERa and ERb; were evaluated by immunoblotting. (C) Representative images ofimmunoblotting. (D) The band intensity was measured and the protein levels of ERa and ERb; were divided by those of a-tubulin to calculate relative protein levels. Thevalues are the means � S.E.M. of four separate experiments.

90 Y. Ishihara et al. / Journal of Steroid Biochemistry & Molecular Biology 145 (2015) 85–93

4. Discussion

SERMs, Tam and Rlx attenuated microglial expression ofproinflammatory cytokines and chemokines induced by LPS in

rat primary microglia, indicating that Tam and Rlx suppressed theactivation of rat primary microglia. Suuronen et al. reported thatTam attenuated the release of interleukin-6 and nitric oxide fromactivated primary microglia [25]. Thus, our results support their

Fig. 6. Involvement of ER in microglial inactivation and neuroprotection induced byTam and Rlx.(A) Rat primary microglia were pretreated with 1 mM ICI182,780 (ICI)for 20 min and then treated with 1 mM Tam or Rlx for 24 h, followed by 10 ng/mL LPSstimulation for 6 h. Levels of TNFa and IL-1bmRNA were determined by real-timePCR. Amounts of mRNA are represented as fold-change from those in untreatedcells. The values are the means � S.E.M. of five separate experiments. **P < 0.01 vs.the untreated group. ##P < 0.01 vs. the LPS-treated group, aaP < 0.01 vs. theLPS + Tam-treated group, bbP < 0.01 vs. the LPS + Rlx-treated group. (B) Rat primarymicroglia were pretreated with 1 mM ICI for 20 min and with 1 mM Tam or Rlx for24 h, and subsequently stimulated with 10 ng/mL LPS for 24 h. SH-SY5Y cells weretreated with microglial conditioned media for 24 h and then cell viability wasassessed by the MTT assay. The values are the means � S.E.M. of six separateexperiments. **P < 0.01 vs. the untreated group. ##P < 0.01 vs. the LPS-treatedgroup, aaP < 0.01 vs. the LPS + Tam-treated group, bbP < 0.01 vs. the LPS + Rlx-treatedgroup.

Fig. 7. Induction of ERE-mediated transcription elicited by Tam and Rlx. pGL4.24 orpGL4.24-3 � ERE was transfected into rat primary microglia. Cells were stimulatedby 1 mM Tam 1 mM Rlx or 10 nM E2 for 6 h and luciferase activity was measured. Thevalues are the means � S.E.M. of 5 separate experiments. *P < 0.05, **P < 0.01 vs. thegroup of untreated cells with pGL4.24-3 � ERE.

Y. Ishihara et al. / Journal of Steroid Biochemistry & Molecular Biology 145 (2015) 85–93 91

findings. In addition, we demonstrated for the first time that Tamand Rlx inhibited neuronal cell death induced by LPS-treatedmicroglial conditioned media, suggesting that the neuroprotectiveeffects of Tam and Rlx are mediated by the suppression ofmicroglial activation. Because activated microglia are known torelease harmful factors such as reactive oxygen species and nitricoxide in addition to proinflammatory cytokines and chemokines[26], Tam and Rlx are thought to protect neuronal cells bysuppressing the production and/or release of these harmfulmolecules.

Expression of ERs has been studied in primary microglia,microglial cell lines, and in vivo models. However, ER expression inmicroglia remains controversial. ERa and ERb mRNA have beendetected in rat primary microglia [6], but there are no reports on ERprotein expression in primary microglia. The expression of ERbwas upregulated in monkey microglia during ischemia [27],suggesting that the stress response regulated ER expression. Sierraet al. reported that mouse microglia isolated based on c-fmspromoter activity expressed ERa, but not ERb [28]. They alsodemonstrated that expression of ERa in microglia was decreasedby LPS stimulation. It has also been demonstrated that the mousemicroglia cell line BV-2 expresses only ERb [29]. Thus, theexpression of ERs in microglia is thought to depend on species andculture conditions, and is thought to be regulated by variousstressful stimuli. In this study, ERa and ERb were expressedprimarily in the nuclei and modestly in the cytosol of rat primarymicroglia, and the addition of LPS did not affect ER expression.Therefore, rat primary microglia are considered to express ERsconstitutively.

The pure ER antagonist ICI182,780 abolished the suppression ofmicroglial activation and neuroprotection mediated by Tam andRlx indicating that Tam and Rlx bind to ER to exert their effects.Tam has also been reported to block chloride channels, and thispharmacological action is exerted at concentrations on the order oftens of mM [30]. In this study, 1 mM of Tam attenuated microglialactivation, supporting that the action of Tam is mediated by ERactivation. ER regulates physiological functions primarily by notonly regulating the transcription of various genes in the nucleus(genomic action), but also inducing signal transduction in thecytosol independent of its transcriptional effects (nongenomicaction) [31]. Because the majority of ER expression in rat primarymicroglia is in the nucleus, we speculated that the effects of ERswere mediated by genomic action in our experiments. Indeed, weshowed that Tam and Rlx activated transcription via ER-EREinteraction. The degree of ERE-dependent transcription induced by1 mM Tam or Rlx was similar to that produced by 10 nM E2.Therefore, Tam and Rlx attenuated microglial activation andsubsequently protected neuronal cells through an ER-mediatedtranscriptional pathway. Suuronen et al. suggested that SERM-induced anti-inflammatory responses were not ER-mediated, butattributable to SERM-induced modulation of LPS-activated proin-flammatory signaling cascades [25]. Thus, SERMs could attenuatemicroglial activation via multiple mechanisms.

There are some reports on the mechanisms by whichE2 suppresses microglial activation. Transcription factor NF-kBmediates the expression of proinflammatory molecules in micro-glia, and inflammatory mediators initiate an NF-kB signalingcascade that leads to phosphorylation and activation of IKKb (asubunit of the inhibitor of IkB kinase (IKK) complex). ActivatedIKKb phosphorylates the IkB inhibitory protein IkBa, targeting itfor ubiquitination and proteasomal degradation, which allows therelease of NF-kB into the nucleus, where it binds to its cognate DNAsequences to induce gene expression [32]. Ghisletti et al. showedthat E2 inhibited the nuclear translocation of NF-kB and attenuatedmicroglial activation [8], demonstrating an interaction betweenthe ER pathway and the NF-kB signaling cascade. Baker et al.

92 Y. Ishihara et al. / Journal of Steroid Biochemistry & Molecular Biology 145 (2015) 85–93

demonstrated that E2 potentiated ERK activation, resulting in thesuppression of proinflammatory proteins and inducible nitric oxidesynthase expression [29], suggesting that alterations in MAPKpathways may be a potential mechanism by which E2 decreasedmicroglial activity. Further study is needed to identify the eventsdownstream of ER that are elicited by Tam and Rlx.

Interestingly, the suppressive effects of high concentrations ofRlx (3 mM) on microglial activation were weaker than those of lowconcentrations of Rlx (1 mM) (Fig. 2). We could not study Rlxconcentrations greater than 3 mM because of its toxicity to ratprimary microglia (Fig. 1). Divergent effects of different doses ofE2 have also been reported for macrophage and T cell activation,in addition to microglial activation [6,33]. The sensitivedose-dependence of the effects of E2 on immune cells may be areflection of the multiple signal transduction pathways that can beregulated by E2 under various circumstances. Likewise, differentcell types in the immune system do not respond in the same way toE2 treatment [33]. Because Rlx acts on the ER, Rlx could exhibitdivergent effects in a similar manner to E2.

Microglial activation has been recognized as an essentialmediator of neuroinflammation and is considered to be involvedin various CNS disorders, including neurodegenerative diseasesand epilepsy [5,34]. Therefore, therapies that suppress microglialactivation could represent treatment strategies for these disorders.Although E2 is a typical steroid hormone that has been reported toattenuate microglial activation [6,7], its effects on carcinogenicity,especially for breast cancer and ovarian cancer, remain a matter ofdebate [35–37]. SERMs, Tam and Rlx are thought to have lesspotential to induce carcinogenesis, because Tam and Rlx showanti-estrogen action in the mammary gland and the ovary [13]. Inaddition, in this study, Tam and Rlx at a concentration of 1 mMstimulated ERs expressed in rat primary microglia culture cells.The clinical therapeutic dose of Tam is approximately 1 mM[38,39]. Therefore, SERMs such as Tam and Rlx might be a potentialcandidate for therapeutic agents for CNS diseases based on theirsuppressive action of microglial activation.

In conclusion, SERMs such as Tam and Rlx protected neuronalcells, likely by attenuating microglial activation. These suppressiveeffects of Tam and Rlx on microglial activation were mediated by anER-dependent transcription pathway. Thus, SERMs might repre-sent a novel strategy to treat CNS disorders via suppression ofneuroinflammation.

Acknowledgments

This work was supported by grants (KAKENHI) from theMinistry of Education, Culture, Sports, Science, and Technology ofJapan to Y.I., K.I. and T.Y. (26740024, 30291149, and 22310041) anda grant from the Hiroshima University Education and ResearchSupport Foundation to Y.I.

References

[1] L. Marinova-Mutafchieva, M. Sadeghian, L. Broom, J.B. Davis, A.D. Medhurst, D.T. Dexter, Relationship between microglial activation and dopaminergicneuronal loss in the substantia nigra: a time course study in a6-hydroxydopamine model of Parkinson’s disease, J. Neurochem. 110 (2009)966–975.

[2] S.E. Hickman, E.K. Allison, J. El Khoury, Microglial dysfunction and defectivebeta-amyloid clearance pathways in aging Alzheimer’s disease mice, J.Neurosci. 8 (2008) 8354–8360.

[3] Y.C. Wang, S. Lin, Q.W. Yang, Toll-like receptors in cerebral ischemicinflammatory injury, J. Neuroinflammation 8 (2011) 134.

[4] S. Najjar, D.M. Pearlman, K. Alper, A. Najjar, O. Devinsky, Neuroinflammationand psychiatric illness, J. Neuroinflammation 10 (2013) 43.

[5] D.N. Xanthos, J. Sandkuhler, Neurogenic neuroinflammation: inflammatoryCNS reactions in response to neuronal activity, nature reviews, Neuroscience15 (2014) 43–53.

[6] E. Vegeto, C. Bonincontro, G. Pollio, A. Sala, S. Viappiani, F. Nardi, A. Brusadelli,B. Viviani, P. Ciana, A. Maggi, Estrogen prevents the lipopolysaccharide-induced inflammatory response in microglia, J. Neurosci. 21 (2001) 1809–1818.

[7] A.J. Bruce-Keller, J.L. Keeling, J.N. Keller, F.F. Huang, S. Camondola, M.P. Mattson,Antiinflammatory effects of estrogen on microglial activation, Endocrinology141 (2000) 3646–3656.

[8] S. Ghisletti, C. Meda, A. Maggi, E. Vegeto, 17beta-estradiol inhibitsinflammatory gene expression by controlling NF-kappaB intracellularlocalization, Mol. Cell. Biol. 25 (2005) 2957–2968.

[9] E. Vegeto, S. Belcredito, S. Etteri, S. Ghisletti, A. Brusadelli, C. Meda, A. Krust, S.Dupont, P. Ciana, P. Chambon, A. Maggi, Estrogen receptor-alpha mediates thebrain antiinflammatory activity of estradiol, Proc. Natl. Acad. Sci. U. S. A. 100(2003) 9614–9619.

[10] C.M. Brown, T.A. Mulcahey, N.C. Filipek, P.M. Wise, Production of proin-flammatory cytokines and chemokines during neuroinflammation: novel rolesfor estrogen receptors alpha and beta, Endocrinology 151 (2010) 4916–4925.

[11] C.L. Smith, B.W. O’Malley, Coregulator function: a key to understanding tissuespecificity of selective receptor modulators, Endocr. Rev. 25 (2004) 45–71.

[12] R.R. Love, R.B. Mazess, H.S. Barden, S. Epstein, P.A. Newcomb, V.C. Jordan, P.P.Carbone, D.L. DeMets, Effects of tamoxifen on bone mineral density inpostmenopausal women with breast cancer, N. Engl. J. Med. 326 (1992)852–856.

[13] B.S. Komm, S. Mirkin, An overview of current and emerging SERMs, J. SteroidBiochem. Mol. Biol. 143C (2014) 207–222.

[14] O. Ismailoglu, B. Oral, A. Gorgulu, R. Sutcu, N. Demir, Neuroprotective effects oftamoxifen on experimental spinal cord injury in rats, J. Clin. Neurosci.17 (2010)1306–1310.

[15] N.E. Franco Rodriguez, J.M. Duenas Jimenez, B. De la Torre Valdovinos, J.R.Lopez Ruiz, L. Hernandez Hernandez, S.H. Duenas Jimenez, Tamoxifenfavoured the rat sensorial cortex regeneration after a penetrating braininjury, Brain Res. Bull. 98 (2013) 64–75.

[16] D.L. Lei, J.M. Long, J. Hengemihle, J. O’Neill, K.F. Manaye, D.K. Ingram, P.R.Mouton, Effects of estrogen and raloxifene on neuroglia number andmorphology in the hippocampus of aged female mice, Neuroscience 121(2003) 659–666.

[17] S. Tapia-Gonzalez, P. Carrero, O. Pernia, L.M. Garcia-Segura, Y. Diz-Chaves,Selective oestrogen receptor (ER) modulators reduce microglia reactivity invivo after peripheral inflammation: potential role of microglial ERs, J.endocrinol. 198 (2008) 219–230.

[18] G.E. Barreto, M. Santos-Galindo, L.M. Garcia-Segura, Selective estrogenreceptor modulators regulate reactive microglia after penetrating braininjury, Front. Aging Neurosci. 6 (2014) 132.

[19] J.L. Liu, D.S. Tian, Z.W. Li, W.S. Qu, Y. Zhan, M.J. Xie, Z.Y. Yu, W. Wang, G. Wu,Tamoxifen alleviates irradiation-induced brain injury by attenuating micro-glial inflammatory response in vitro and in vivo, Brain Res. 1316 (2010)101–111.

[20] M. Cerciat, M. Unkila, L.M. Garcia-Segura, M.A. Arevalo, Selective estrogenreceptor modulators decrease the production of interleukin-6 and interferon-gamma-inducible protein-10 by astrocytes exposed to inflammatory challengein vitro, Glia 58 (2010) 93–102.

[21] A. Suzumura, S.G. Mezitis, N.K. Gonatas, D.H. Silberberg, MHC antigenexpression on bulk isolated macrophage-microglia from newborn mousebrain: induction of Ia antigen expression by gamma-interferon, J. Neuro-immunol. 15 (1987) 263–278.

[22] Y. Ishihara, T. Kawami, A. Ishida, T. Yamazaki, Tributyltin induces oxidativestress and neuronal injury by inhibiting glutathione S-transferase in ratorganotypic hippocampal slice cultures, Neurochem. Int. 60 (2012) 782–790.

[23] P. Ciana, M. Raviscioni, P. Mussi, E. Vegeto, I. Que, M.G. Parker, C. Lowik, A.Maggi, In vivo imaging of transcriptionally active estrogen receptors, Nat. Med.9 (2003) 82–86.

[24] S. Nilsson, S. Makela, E. Treuter, M. Tujague, J. Thomsen, G. Andersson, E.Enmark, K. Pettersson, M. Warner, J.A. Gustafsson, Mechanisms of estrogenaction, Physiol. Rev. 81 (2001) 1535–1565.

[25] T. Suuronen, T. Nuutinen, J. Huuskonen, J. Ojala, A. Thornell, A. Salminen, Anti-inflammatory effect of selective estrogen receptor modulators (SERMs) inmicroglial cells, Inflamm. Res. 54 (2005) 194–203.

[26] H. Kettenmann, U.K. Hanisch, M. Noda, A. Verkhratsky, Physiology of microglia,Physiol. Rev. 91 (2011) 461–553.

[27] N. Takahashi, A.B. Tonchev, K. Koike, K. Murakami, K. Yamada, T. Yamashima,M. Inoue, Expression of estrogen receptor-beta in the postischemic monkeyhippocampus, Neurosci. Lett. 369 (2004) 9–13.

[28] A. Sierra, A. Gottfried-Blackmore, T.A. Milner, B.S. McEwen, K. Bulloch, Steroidhormone receptor expression and function in microglia, Glia 56 (2008)659–674.

[29] A.E. Baker, V.M. Brautigam, J.J. Watters, Estrogen modulates microglialinflammatory mediator production via interactions with estrogen receptorbeta, Endocrinology 145 (2004) 5021–5032.

[30] D.J. Hines, R.M. Hines, S.J. Mulligan, B.A. Macvicar, Microglia processes blockthe spread of damage in the brain and require functional chloride channels,Glia 57 (2009) 1610–1618.

[31] N. Vasudevan, D.W. Pfaff, Non-genomic actions of estrogens and theirinteraction with genomic actions in the brain, Front. Neuroendocrin. 29 (2008)238–257.

[32] S. Ghosh, M. Karin, Missing pieces in the NF-kappaB puzzle, Cell 109 (Suppl)(2002) S81–S96.

Y. Ishihara et al. / Journal of Steroid Biochemistry & Molecular Biology 145 (2015) 85–93 93

[33] M. Cutolo, A. Sulli, B. Seriolo, S. Accardo, A.T. Masi, Estrogens, the immuneresponse and autoimmunity, Clin. Exp. Rheumatol. 13 (1995) 217–226.

[34] S. Lehnardt, Innate immunity and neuroinflammation in the CNS: the role ofmicroglia in Toll-like receptor-mediated neuronal injury, Glia 58 (2010)253–263.

[35] J.E. Rossouw, G.L. Anderson, R.L. Prentice, A.Z. LaCroix, C. Kooperberg, M.L.Stefanick, R.D. Jackson, S.A. Beresford, B.V. Howard, K.C. Johnson, J.M. Kotchen,J. Ockene, I. writing group for the women’s health initiative, risks and benefitsof estrogen plus progestin in healthy postmenopausal women: principalresults from the women’s health Initiative randomized controlled trial, JAMA:J. Am. Med. Assoc. 288 (2002) 321–333.

[36] V. Beral, C. Million Women Study, Breast cancer and hormone-replacementtherapy in the million women study, Lancet 362 (2003) 419–427.

[37] E. Weiderpass, J.A. Baron, H.O. Adami, C. Magnusson, A. Lindgren, R. Bergstrom,N. Correia, I. Persson, Low-potency oestrogen and risk of endometrial cancer: acase-control study, Lancet 353 (1999) 1824–1828.

[38] K. Dhandapani, D. Brann, Neuroprotective effects of estrogen and tamoxifen invitro: a facilitative role for glia? Endocrine 21 (2003) 59–66.

[39] S.H. Mehta, K.M. Dhandapani, L.M. De Sevilla, R.C. Webb, V.B. Mahesh, D.W.Brann, Tamoxifen, a selective estrogen receptor modulator, reduces ischemicdamage caused by middle cerebral artery occlusion in the ovariectomizedfemale rat, Neuroendocrinology 77 (2003) 44–50.