Ovariectomy and Subsequent Treatment with Estrogen Receptor Agonists Tune the Innate Immune System of the Hippocampus in Middle-Aged Female Rats

The innate immune system including microglia has a major contribution to maintenance of the physiological functions of the hippocampus by permanent monitoring of the neural milieu and elimination of tissue-damaging threats. The hippocampus is vulnerable to age-related changes ranging from gene expression to network connectivity. The risk of hippocampal deterioration increases with the decline of gonadal hormone supply. To explore the impact of hormone milieu on the function of the innate immune system in middle-aged female rats, we compared mRNA expression in the hippocampus after gonadal hormone withdrawal, with or without subsequent estrogen replacement using estradiol and isotype-selective estrogen receptor (ER) agonists. Targeted profiling assessed the status of the innate immune system (macrophage-associated receptors, complement, inhibitory neuronal ligands), local estradiol synthesis (P450 aromatase) and estrogen reception (ER). Results established upregulation of macrophage-associated (Cd45, Iba1, Cd68, Cd11b, Cd18, Fcgr1a, Fcgr2b) and complement (C3, factor B, properdin) genes in response to ovariectomy. Ovariectomy upregulated Cd22 and downregulated semaphorin3A (Sema3a) expression, indicating altered neuronal regulation of microglia. Ovariectomy also led to downregulation of aromatase and upregulation of ERα gene. Of note, analogous changes were observed in the hippocampus of postmenopausal women. In ovariectomized rats, estradiol replacement attenuated Iba1, Cd11b, Fcgr1a, C3, increased mannose receptor Mrc1, Cd163 and reversed Sema3a expression. In contrast, reduced expression of aromatase was not reversed by estradiol. While the effects of ERα agonist closely resembled those of estradiol, ERβ agonist was also capable of attenuating the expression of several macrophage-associated and complement genes. These data together indicate that the innate immune system of the aging hippocampus is highly responsive to the gonadal hormone milieu. In ovariectomized female rats, estradiol replacement exerts potent immunomodulatory effects including attenuation of microglia sensitization, initiation of M2-like activation and modulation of complement expression by targeting hippocampal neurons and glial cells through ERα and ERβ.


Introduction
The hippocampus plays a crucial role in learning and memory, and it is essential for planning and creativity [1]. This brain area is vulnerable to age-related changes ranging from gene expression to neuronal network connectivity [2,3]. Ongoing neurogenesis, which helps to maintain normal hippocampal function, is impaired dramatically during aging [4]. Age-related alterations are implicated in the development of neurodegenerative diseases [5]. Changes in gene expression start at midlife [6], which overlaps with the time of cessation of estrogen and progesterone synthesis by the ovaries [7]. In rodents, alterations of the hippocampal transcriptome also begin at midlife [8]. Estrogen and progesterone receptors are expressed in the hippocampus [9][10][11][12]. ERa and ERb are of particular interest as they mediate the effects of 17bestradiol (E2) on gene expression [13,14], intracellular signaling [15,16] and synaptic plasticity [17,18]. Besides synaptic plasticity, E2 modulates glial functions and regulates the immune response [19][20][21][22][23]. By the time of menopause, declining gonadal hormone supply is associated with increased risks of hippocampal deterioration and memory disturbances [24,25].
Microglial cells, which form part of the innate immune system, play an important role in the maintenance of hippocampal functions via constant monitoring of the neural milieu to eliminate tissue-damaging threats. In addition, surveying microglia acutely modulate synaptic activity by regulating the number of functional synapses in organotypic slices [26]. Microglia also remodel neuronal circuits in the postnatal retinogeniculate system in an activity and complement-dependent manner [27]. Microglial cells express steroid hormone receptors [28]. While there is consensus about the expression of ERa, expression of ERb is rather controversial [28][29][30][31]. Independent studies have detected only ERa at mRNA and protein level in microglia isolated from adult mice [28,30]. Anti-inflammatory effects of E2 are mediated primarily through ERa [19], but a recent study has demonstrated that the androgen metabolite 5-androsten-3b,17b-diol can suppress inflammatory processes of microglia and astrocytes via acting through ERb [29]. Another laboratory has provided immunohistochemical data that support ERb expression in mouse microglia [31]. Progesterone and its metabolite allopregnanolone modify the action of E2 and exert immunomodulatory effects [32]. Microglia continuously survey the local environment via receptors for neurotransmitters, danger signals, chemokines, cytokines and complement activation fragments among others [33]. Receptors for danger signals and inhibitory neuronal ligands represent the accelerator and the break for the transition from surveying to effector microglia, respectively [33].
In rat models of normal aging, molecular and functional alterations of microglia indicate the elevation of microglia reactivity. Microglial expression of MHCII [34,35] and other cell surface markers such as CD11b [36], CD68 [37], CD80, CD86 [38] increases in the normal aged brain. The initial shift in microglia phenotype may occur at midlife [6] coinciding with the menopausal transition. Therefore, we hypothesize that menopause upsets the delicate balance between activation signals and inhibitory neuronal ligand control resulting in a shift of the microglia phenotype in the aging female hippocampus.
Middle-aged ovariectomized rats are widely used to study behavioral [39,40], cellular [41] and molecular [17,42] changes related to menopause. Here, we studied the hippocampal transcriptome of this animal model in order to explore the effects of ovariectomy and subsequent estrogen replacement; transcripts analyzed were related to the innate immune system including microglia, complement (especially components of the alternative pathway) and cytokines. In replacement studies we used E2 and isotype-selective ER agonists. As the expression of macrophageassociated genes is low in surveying microglia at this age [35], we followed mRNA expression, which allowed us to quantitatively typify gene expression. Results of the present study provide evidence that ovariectomy of middle-aged rats increases the expression of macrophage-associated receptors, components of the alternative activation pathway of complement and inhibitory neuronal ligands for microglia. We show that estrogen replacement attenuates the shift in the microglia phenotype, partly restores neuronal control of microglia and modulates complement expression. These data suggest that ovariectomy-dependent changes in the innate immune system can be attenuated by timely estrogen replacement.

Ethics Statement
All studies were carried out with permission from the Animal Welfare Committee of the Institute of Experimental Medicine, Hungarian Academy of Sciences (Permission Number: A5769-01) and in accordance with the legal requirements of the European Community (Decree 86/609/EEC). Animal experimentation described was conducted in accord with accepted standards of animal care.

Treatments of Experimental Animals
Female Harlan-Wistar rats were purchased from Toxicoop (Budapest, Hungary) and were housed under standard laboratory conditions with unrestricted access to phytoestrogen-free rodent diet (Harlan Teklad Global Diets, Madison, WI, USA). The following experimental groups were applied: middle-aged (13 month old) female rats (M group), middle-aged ovariectomized (OVX) rats (M/OVX group) and M/OVX rats with estrogen replacement. Replacement was performed with E2 (M/OVX+E2 group), ERa agonist LE2 (M/OVX+LE2 group) and ERb agonist DPN (M/OVX+DPN group). Middle-aged (n = 26) rats underwent bilateral ovariectomy except for the sham-operated M group (n = 8). All rats were housed individually after surgery, and ten days later received treatments with vehicle or ER agonists. Treatments in middle-aged OVX rats were carried out as described earlier [43]. In brief, Alzet 2004 osmotic minipumps filled either with E2 (0,33 mg/ml in propylene-glycol, n = 8), LE2 (3,33 mg/ml in propylene-glycol, n = 5), DPN (3,33 mg/ml in propylene-glycol, n = 5), or the vehicle only (n = 8, M/OVX group) were implanted subcutaneously for 29 days. Concentrations of E2 and isotype-selective ER agonists were calculated to produce a release rate of 2.0mg/d and 20 mg/d, respectively. Treatments with E2 and LE2 at these doses result in a similar increase in uterine weight, while DPN causes no change [43]. Body weights were measured to test efficacy of the treatments.

Sample Preparation
On the day of sample preparation, animals were deeply anesthetized and perfused transcardially with 100 ml of cold fixative solution containing 10% RNAlater (Qiagen, Heidelberg, Germany) in phosphate buffered saline. In all experiments, the same procedure was followed for the preparation of the hippocampal formation.

Quantitative Real-time PCR
The hippocampi were prepared and total RNA was isolated using the RNeasy Lipid Tissue Kit (Qiagen, Hilden, Germany). RNA analytics included A260 nm/A280 nm readings using a Nanodrop Spectrophotometer and capillary electrophoresis using Nano RNA Chips on 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA). All RNA samples displayed high RNA integrity numbers (RIN .8.2).
Custom TaqMan microfluidic cards were designed to study in depth the regulation of ninety-six genes by quantitative real-time PCR. Microfluidic cards (Applied Biosystems, Foster City, CA, USA) were preloaded by the manufacturer with selected inventoried assays for genes of our interest (Table 1). We used glyceraldehyde-3-phosphate dehydrogenase and hypoxanthine guanine phosphoribosyl-transferase as housekeeping genes. Expression of these genes did not vary among treatment groups of the study. Each assay consisted of a FAM dye-labeled TaqMan MGB probe and two PCR primers. Reverse transcription and real-time PCR were run as described earlier [43]. The ViiA7 RUO (Applied Biosystems) software and relative quantification against calibrator samples (DDCt) were used for data analysis. A computed internal control corresponding to the geometric mean of cycle threshold (Ct) values of the selected housekeeping genes was used for DCt calculation [44]. The use of TaqMan chemistry allowed comparison of DCt values which correlated with transcript levels. Based on DCt values, macrophage-associated genes were ranked into three categories based on their abundant (0,DCt,1), moderate (1,DCt,5) or low (DCt.5) level of mRNA expression. Relative quantity (RQ = 2 2DDCt ) was used to characterize gene expression in the various experimental groups. PCR experiments conformed to minimum information for publication of quantitative real-time PCR experiments (MIQE) guidelines [45].

Analysis of Human Microarray Data
The files of data set GSE11882 [6], deposited in Gene Expression Omnibus, contained microarray data from the hippocampus. Sample size (n), average age (age) in years and standard deviation of ages (SD) were used to characterize the premenopausal (n HC = 9, age HC = 39.4, SD HC = 8.0) and postmenopausal (n HC = 5, age HC = 71.2, SD HC = 4.4) hippocampus data. Pre-and postmenopausal groups consisted of nine and five individuals, respectively. Only changes with fold change (FC).1.5 (upregulation) and FC,0.66 (downregulation) were considered as FC is the most reliable parameter in the case of low sample numbers [46]. Raw microarray data were pre-processed for analysis by GC robust multi-array average (GCRMA) [47]. From the expression set, probesets were selected based on the relevance to rat data.

Statistical Analysis
In PCR data evaluation, group data were expressed as RQ(mean)6standard deviation (SD). Statistical significance of the changes in gene expression was analyzed using ANOVA followed by Newman-Keuls post-hoc test (Statistica software version 11.0, StatSoft Inc., Tulsa, OK). Homocedasticity and normality satisfied the criteria for running ANOVA. In correlation analysis Pearson's coefficient was estimated.
In microarray data evaluation, difference analysis of gene expression was performed by linear models combined with Bayesian methods [48], p was adjusted by the false discovery rate-based method [49]. In statistical and data mining work, Bioconductor packages [50] in R-environment were used.

Expression of Immune-related Genes in the Hippocampus of Middle-aged Female Rats
We investigated mRNA expression of macrophage-associated genes (Table 1) in the hippocampus. Under physiological conditions, these genes are expressed predominantly in microglial cells, although in the case of toll-like receptors astroglial expression has also been reported [51]. Among macrophage-associated genes, we measured moderate to low mRNA level of microglia response factor Iba1 and Fcgr2a, the rest of the genes tested had low mRNA levels indicating a very weak macrophage character of microglia in the hippocampus of middle-aged female rats. On the other hand, microglial receptors for inhibitory neuronal ligands showed significant expression. We detected high mRNA level of signalregulatory protein a (Sirpa), moderate levels of neuropilin 1 (Nrp1), fractalkine receptor (Cx3cr1), low levels of leukocyte common antigen (Cd45) and very low levels of Cd200r (Table 1).
We also investigated the expression of complement and proinflammatory cytokine genes ( Table 2). Both neurons and glial cells express complement components and their regulators. We found moderate mRNA levels of C3, Cfh, Crry and Cd59, other complement genes had low expression levels. We measured very low levels of proinflammatory cytokine mRNAs ( Table 2).

Impact of Ovarian Hormone Depletion and Subsequent Estrogen Treatment on the Expression of Macrophageassociated Genes
We examined the effect of ovarian hormones on mRNA expression of macrophage-associated genes that encode microglial markers, phagocytic receptors, recognition molecules and receptors for inhibitory neuronal ligands.
Only inhibition of Iba1 expression by DPN reached statistical significance compared to M/OVX animals. Similar to E2, LE2 enhanced the expression of Mrc1 and Cd163 compared to M and M/OVX animals ( Fig. 1). We also explored hormonal regulation of two MHC genes, RT1-EC2 and Cd74. Ovariectomy evoked 2.1-fold increase in mRNA expression of RT1-EC2 (p = 0.031) compared to middleaged controls. Estrogen treatments did not attenuate the alteration. Cd74 showed no hormonal regulation in the hippocampus.
In addition, we investigated the expression of four astroglial markers including glial fibrillary acidic protein, colony stimulating factor 1, and glial high affinity glutamate transporters (Slc1a2, Slc1a3). These genes showed negligible alterations (data not shown).

Effect of Ovarian Hormones and Isotype-selective ER Agonists on the Expression of Complement Genes
Neurons and glial cells express complement components and their regulators [54]. The classical and alternative complement activation pathways converge at C3 activation resulting in the release of C3a and generation of iC3b on the surface of cells rendered to phagocytosis [55]. Expression of C3, the central component of complement increased 1.9-fold (p = 0.001) in ovariectomized rats compared to controls (Fig. 5A). Factor B (Cfb), subunit of the alternative pathway C3 convertase C3bBb, showed similar but smaller (1.3-fold, p = 0.037) increase in mRNA expression than C3 (Fig. 5B). Factor D (Cfd) showed no alteration following ovariectomy. Negative regulators of C3bBb such as factor H (Cfh) (Fig. 5C), complement regulatory protein (Crry) and Cd55 did not change. Expression of positive regulator of C3bBb, properdin (Cfp) increased 1.5-fold (Fig. 5D), but the change did not reach statistical significance (p = 0.116). Cd59, an inhibitor of homologous complement lysis showed unaltered expression.
Chronic treatments with E2 and DPN reversed upregulation of C3, while LE2 did not (Fig. 5A). In contrast, E2 and LE2 increased mRNA expression of genes encoding components of the alternative pathway (Fig. 5B-D). In the case of Cfb, DPN showed a character different from E2, i.e. restored mRNA expression (Fig. 5B). each group. ANOVA identified statistically significant treatment effects for each gene (p = 0.009 for Iba1, p = 0.014 for Cd68, p,0.001 for M2 markers). Turquoise, black and red asterisks mark statistically significant group differences compared to M, M/OVX and M/OVX+E2 animals, respectively, by Newman-Keuls post hoc test. * corresponds to 0.01, P,0.05 and *** to P,0.001. M, middle-aged female rat; M/OVX, middleaged OVX rat; M/OVX+E2, M/OVX rats treated with E2; M/OVX+LE2, treated with 16a-LE2; M/OVX+DPN, treated with DPN. P, p value. doi:10.1371/journal.pone.0088540.g001 We examined mRNA expression of Tnfa, Il1b, Il6 and Il12b, which showed no regulation by ovarian hormones (data not shown). We also tested expression of anti-inflammatory cytokines Il4 and Il10, but their mRNA levels were below the detection limit of TaqMan-based real-time PCR.

Ovarian Hormone-dependent Regulation of Aromatase and Estrogen Receptor Genes
In the absence of circulating E2, local synthesis determines the actual E2 concentration in the hippocampus of OVX rats. In this process, cytochrome P450 aromatase (encoded by Cyp19a1) carries out hydroxylation of the A ring of androgen precursors, androstendione and testosterone [56]. Cyp19a1 expression showed 1.4-fold (p = 0.025) decrease in middle-aged OVX rats compared to controls (Fig. 6A). E2 and isotype-selective ER agonists did not restore mRNA expression of Cyp19a1, LE2 even decreased it further (Fig. 6A). It is known that aromatase activity has been involved in many functions, from synaptogenesis to neuroprotection [57]. In accord with decreased Cyp19a1, we found downregulation of growth associated protein 43 (RQ = 0.827, p = 0.018) and synaptophysin (RQ = 0.876, p = 0.022). We examined the correlation between the expression of Gap43, Cyp19a1 and macrophage-associated genes. In M/ OVX animals, Pearson's correlation analysis revealed positive correlation between Gap43 and Cyp19a1 (R = 0.73, p = 0.162), and negative correlation between Gap43 and Cd11b (R = 20.92, p = 0.077), Cyp19a1 and Cd11b (R = 20.81, p = 0.094), Cyp19a1 and Cd22 (R = 20.88, p = 0.052). These results suggested that a positive correlation might exist between the expression of Gap43 and Cyp19a1 in the hippocampus of middle-aged OVX rats. In addition, strong negative correlation might exist between the expression of Gap43 and Cd11b, Cyp19a1 and Cd22.

Data Analysis Reveals Upregulation of Macrophageassociated Genes and Complement in Postmenopausal Women
In order to address the effect of menopause on the innate immune system in the hippocampus, we compared the expression of selected genes in raw microarray data of pre-and postmenopausal women [6]. We identified a large series of changes related to menopause (and aging) that were strikingly similar to those induced by ovariectomy in middle-aged rats. These

Discussion
Focusing on the innate immune system in the hippocampus, we studied the impact of ovarian hormone deficiency and subsequent estrogen treatments with E2 and isotype-selective ER agonists on gene expression in middle-aged female rats. From the results we conclude that ovarian hormone deficiency alerts the innate immune system in the hippocampus, and subsequent E2 replacement is capable of counteracting several changes via ERa and ERb. Our findings are in line with previous reports that have demonstrated relationship between endogenous estrogen status and microglia reactivity [20], escape of microglia from inhibitory neuronal control [58], activation of microglia [34,35,59] and enhanced complement expression in the normal aged brain [6,8,60].

Ovariectomy Alerts the Innate Immune System in the Hippocampus
Ovariectomy sensitizes microglia. Microglia show plasticity and possess various priming states that determine the microglial response to subsequent stimuli. Most studied examples include the classical (M1) and alternative (M2) priming states that evolve in response to interferon-c and IL4 (or IL13), respectively. In response to ovarian hormone deficiency, we find elevated expression of several microglia markers (Iba1, Cd68, Cd80), phagocytic receptors (Cd11b, Cd18, Fcgr1a, Fcgr2b) and recognition molecules (Tlr3, Tlr9) in the hippocampus. Despite elevated expression of macrophage-associated genes, glial cells do not produce proinflammatory cytokines including tumor necrosis factor a (TNFa). This is in accord with the finding of a recent study demonstrating that while age-dependent microglial NF-kB activation and TNFa overproduction take place in the hypothalamus, only minor NF-kB activation occurs in the cerebral cortex of middle-aged mice [61]. On the other hand, several studies have reported elevated cytokine production in aging microglia isolated from rats and mice. Enhanced production of cytokines in aged microglia may represent age-related alterations. The apparent disagreement between the findings in aged and middle-aged animals suggests that in middle-aged female rats, elevation of cytokine production has not started yet.
The alterations in hippocampal gene expression suggest sensitization of microglia, characterized by moderate upregulation of Iba1, Cd68, Cd11b, Cd18, Fcgr1a, Fcgr2b and Cd80. Of note, mRNA expression of Cd74, Cd86, Cd163, Mrc1, Fcgr2b, Cd14, Tlr4 does not change after ovariectomy. Analysis of human microarray data and comparison between pre-and postmenopausal individuals reveal analogous changes in the expression of these genes indicating that microglia sensitization may occur in postmenopausal women. This special phenotype suggests that microglia adapt and maintain homeostasis by increasing the expression of phagocytic receptors and co-receptors to possess enhanced phagocytic potential.
Ovariectomy alters the neuronal control of microglia reactivity. CD22, formerly known as B-lymphocyte cell adhesion molecule, is expressed in rodent neurons [62]. Neurons secrete a soluble form of CD22, which inhibits microglial activation and proinflammatory cytokine production in vitro via CD45 [62]. CD45 is the most abundantly expressed transmembrane protein-tyrosine phosphatase. Quiescent microglial cells constitutively express CD45, which elevates further during activation [63]. We find weak mRNA expression of Cd22 and Cd45 in the hippocampus of middle-aged female rats, both of which increase after ovariectomy. In agreement with elevated expression of Cd22 and Cd45, mRNA expression of proinflam-p = 0.009 for Tlr9). Turquise, black and red asterisks mark statistically significant group differences compared to M, M/OVX and M/OVX+E2 animals, respectively, by Newman-Keuls post hoc test. * corresponds to 0.01,P,0.05 and ** to 0.001,P,0.01. doi:10.1371/journal.pone.0088540.g002  matory cytokines does not change significantly either. Interestingly, these alterations are reminiscent to those we observed in the hippocampus of postmenopausal women. These results suggest that in OVX rats and postmenopausal women neurons enhance inhibitory CD22-CD45 signaling which prevents the transition from surveying to effector microglia.
Ovariectomy induces a moderate increase in Cd200r1 expression in middle-aged rats. This finding suggests that in the hippocampus of ovarian hormone deprived animals microglial cells enhance the expression of receptors for inhibitory ligands including Cd45 and Cd200r1. On the other hand, mRNA expression of Cx3cr1, Nrp1 and Sirpa does not change in middleaged ovariectomized rats.
Another characteristic change which may contribute to the alteration of neuron-microglia communication is the decrease of Sema3a expression. Of note, downregulation of SEMA3A takes place in postmenopausal woman as well. Neurons secrete Sema3A, which can navigate axons, regulate neuronal polarization [64] and facilitate axonal transport [65]. Stressed neurons also secrete Sema3A, which promotes microglial production of growth factors and anti-inflammatory cytokines in vitro [66]. Sema3A also induces apoptosis of activated microglia [67] indicating that stressed neurons can protect themselves from activated microglia. Therefore, decreased expression of Sema3a may modulate both neuronal and glial functions via the aforementioned mechanisms.
Ovariectomy upregulates C3, factor B and properdin. Complement is a powerful proteolytic cascade which participates in the elimination of invading pathogens, apoptotic cells, and in the regulation of the innate immune response and inflammatory reactions [68]. Age-dependent upregulation of C1q and C3 has been described in the rodent and human brain [6,8,60]. We show ovarian hormone-dependent upregulation of C3 in the hippocampus of middle-aged ovariectomized rats and postmenopausal women. In middle-aged rats, we find low level expression of components of the alternative pathway, and observe upregulation of Cfb and properdin after ovariectomy. Increased expression of alternative pathway components can result in elevated tick-over [69], properdin-dependent [70] and amplification mechanisms [71]. However, expression of complement regulators such as Crry and Cd55 does not change. It has potential functional significance as activation of the alternative pathway may be connected to the initial activation of microglial cells. In mice, deletion of Crry induces microglial priming, but inactivation of the alternative pathway in Crry2/2 suspends the shift in the phenotype suggesting that the alternative pathway may play a role in microglia activation [72]. Apparent coincidence in the alterations of macrophage-associated and complement gene expression tempts us to speculate that elevated expression of C3, Cfb and Cfp may contribute to the initial activation of microglial cells after the cessation of ovarian hormone synthesis.
Ovariectomy alters expression of cytochrome P450 aromatase and ERa. Local estrogen synthesis in the hippocampus provides high E2 concentration [56,73]. We show that expression of Cyp19a1 aromatase decreases in the absence of circulating ovarian hormones. Earlier studies have demonstrated that expression of synaptic proteins is regulated by aromatase activity in the hippocampus [56]. In concert with decreased Cyp19 expression, we find down-regulation of synaptic marker genes including Gap43 and Syp. These findings suggest that local E2 level may decrease after the loss of circulating E2 and progesterone  from ovarian sources. We reveal that in middle-aged rats ovariectomy increases ERa expression, but has no effect on ERb and GPR30 in the hippocampus. It is important to note that these results reflect the changes of gene expression from a wide range of cell types including neurons, astrocytes and microglia. It is known that at hippocampal synapses ERa is predominantly concentrated presynaptically and highly enriched with synaptic vesicles [74]. Age decreases ERa level and/or localization in the hippocampus [11]. Taking together, following ovariectomy hippocampal neurons and glia respond to the sudden drop of serum levels of ovarian hormones by downregulation of aromatase, upregulation of macrophage-associated genes and receptors for inhibitory neuronal ligands. Altered expression of inhibitory neuronal ligands may contribute to maintain the tight control of microglia reactivity in middle-aged female rats.

E2 Replacement Modulates the Expression of Macrophage-associated Genes via ERa and ERb
Ovariectomy induces a mild change in the phenotype of microglial cells. E2 replacement restores the alterations in the case of Iba1, Cd68, Cd11b and Cd18. These genes are likely to be regulated only by E2, i.e. progesterone does not modulate them. Both ERa and ERb agonist treatments mitigate mRNA expression of these genes. Of note, E2 replacement does not reverse ovariectomy-evoked upregulation of Fcgr2b and Tlr3. These genes are likely to be regulated by E2 and progesterone. In addition, E2 and ERa agonist LE2 treatments trigger a marked increase in mRNA expression of Mrc1 and Cd163, which are widely used as M2 activation markers [75,76], indicating that E2 may differently regulates Mrc1 and Cd163 in the presence or in the absence of progesterone. These findings are in agreement with previous data demonstrating the role of ERa in the estrogenic regulation of microglia [19,28,30]. On the other hand, inhibition of Iba1, Cd68, Cd11b and Cd18 mRNA expression by ERb agonist DPN may support the view that microglia also express ERb [29,31]. However, we can not exclude the possibility that the effect of DPN is an indirect effect, i.e. the result of the interaction of microglia with other cell types.
Neuronal control of microglia reactivity has also significant impact on the innate immune system of the brain [77]. Of note, E2 and isotype-selective ER agonists enhance the expression of microglial receptors for inhibitory neuronal ligands including Cd45, Cd200r1 and Nrp1. The E2-like effects of DPN on the expression of Cd45 and Nrp1 can support the notion that microglial cells express low level of ERb. However, different effects of E2 and DPN on Cd200r expression do not confirm the presence of ERb in microglia. Estrogens affect the expression of inhibitory neuronal ligands as well. E2 replacement reverses Sema3a expression primarily via ERa, although ERb agonist DPN is also effective. Neither E2 nor isotype-selective ER agonists turn the elevation of Cd22 expression in OVX rats.
Summing up, our findings indicate that the innate immune system of the hippocampus in middle-aged rats is responsive to the gonadal hormone milieu. Ovariectomy triggers sensitization of microglia, elevation of complement expression especially components of the alternative pathway, and decline of aromatase level. Although E2 replacement does not influence mRNA expression of aromatase, E2 decreases the expression of microglia markers and phagocytic receptors in part, and increases M2 activation markers indicating a shift towards a protective microglia phenotype. E2 also modulates complement expression, attenuates C3 and enhances Cfb, Cfh expression. In addition, E2 partly restores Sema3A. The ERb agonist DPN attenuates both microglia sensitization and complement expression. Our results suggest that the innate immune system is responsive to estrogen replacement, at least for a short period of time after ovariectomy.