Address correspondence and reprint requests to Kaoru Sato, Division of Pharmacology, National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya-ku, Tokyo 158–8501, Japan. E-mail: email@example.com
We investigated the effects of estrogen-related compounds including xenoestrogens [17β-estradiol (E2), 17α-ethynylestradiol (EE), diethylstilbestrol (DES), p-nonylphenol (PNP), bisphenol A (BPA) and 17α-estradiol (17α)] on l-glu uptake by cultured astrocytes via glutamate-aspartate transporter (GLAST). After 24 h treatment, E2 inhibited the l-glu uptake at 1 µm and higher concentrations. EE and DES also inhibited the l-glu uptake at 1 nm and higher concentrations. The other four compounds had no effect. The effects of E2, EE and DES were completely blocked by 10 nm of ICI182 780 (ICI). β-Estradiol 17-hemisuccinate : bovine serum albumin (E2-BSA), a membrane-impermeable conjugate of E2, also elicited the inhibition of l-glu uptake at 1 nm and higher concentrations, and the effect was blocked by ICI. 16α-Iodo-17β-estradiol (16αIE2), an estrogen receptor α (ERα) selective ligand, revealed an inhibitory effect at 10 nm, while genistein, an ERβ selective ligand, failed to reveal such an effect at this concentration. Western blot analysis showed that the predominant ER of cultured astrocytes was ERα. The colocalization of ERα with GLAST on plasma membranes was immunohistochemically detected in these cells. From these results, we concluded that estrogens down-regulate l-glu uptake activity of astrocytes via membrane ERα.
Abundant expression of estrogen receptors (ERs) is observed in the central nervous system (CNS), and it has been clarified that estrogens play diverse roles in regulating structures and functions of neuronal systems (Weiland 1992; Wong and Moss 1992; Wooley and McEwen 1994; Foy et al. 1999). Xenoestrogens, man-made non-steroidal compounds including pesticides and industrial by-products, were reported to mimic the actions of estrogens through interactions with ERs. However, recent reports have shown that estrogens have non-genomic effects through unknown mechanisms (Beyer and Raab 1998; Mermelstein et al. 1996; Gu et al. 1999). We also reported that estrogens and xenoestrogens modified mossy fiber-CA3 synapses, and thereby caused CA3-selective hypersensitivity to glutamate through non-genomic mechanisms (Sato et al. 2002). Thus, it is possible that xenoestrogens have more various risks than ever estimated before.
L-Glu is a major excitatory neurotransmitter in the CNS. L-Glu transporters are the only significant mechanism for the removal of l-glu from extracellular fluid and the maintenance of low and non-toxic concentrations of l-glu (Balcar and Johnston 1972; Logan and Snyder 1972; Johnston 1981). A growing body of evidence has suggested the importance of l-glu transporters in synaptic events, i.e. the time course of neuronal activation and the number of neurons activated by l-glu (Oliet et al. 2001; Thompson 2003). In spite of the significant importance of this system, the implication of estrogens in l-glu uptake by l-glu transporters is still unknown. Therefore, we evaluated the effects of estrogens [17β-estradiol (E2); 17α-ethynylestradiol (EE), an estrogen used for oral contraceptive pills and diethylstilbestrol (DES), a synthetic estrogen for preventing miscarriages] and xenoestrogens [p-nonylphenol (PNP), the degradation product of surface active agents used as a supplement of resins and bisphenol (BPA), a content of canned food, dental sealants and composites) on l-glu uptake by cultured astrocytes. Here we report novel effects of estrogens: these compounds down-regulated l-glu uptake activity of astrocytes via membrane ERα.
Materials and methods
Eagle's minimal essential medium (EMEM) was purchased from Nissui (Tokyo, Japan). Donor horse serum (HS; gelding) was from C-C Biotech Corporation (Valley Center, CA, USA). L-Glu, glutamate dehydrogenase (NAD(p)+) (EC 126.96.36.199) (GLD), genistein, paraformaldehyde (PFA), polyoxyethylene (10) octylphenyl ether (TritonX-100) and dimethylformamide were from Wako Pure Chemical (Osaka, Japan). β-NAD, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), 1-methoxyphenazine methosulfate (MPMS), dihydrokainic acid (DHK) dl-threo-β-hydroxy-aspartic acid (THA), E2, EE, DES, PNP, 17α, E2-BSA, BSA, and peroxidase-conjugated anti-rabbit IgG antibody were from Sigma (St Louis, MO, USA). BPA and sodium dodecyl sulfate (SDS) were from Nacalai tesque (Kyoto, Japan). ICI was from Tocris (Ballwin, MO, USA). 16αIE2 (Hochberg and Rosner 1980) was a gift from Dr R. B. Hocheberg (Yale University, New Haven, CT, USA). Rabbit polyclonal IgG to ERα (MC-20) was from Santa Cluz Biotechnology, Inc. (Santa Cruz, CA, USA). Rabbit polyclonal IgG to ERβ (Ab-1) was from Oncogene Research Products (Cambridge, MA, USA). Mouse IgG to glial fibrillary acidic protein (GFAP) and enhanced chemiluminescence (ECL) kit were from Amersham Biosciences (Arlington Heights, IL, USA). The guinea pig antiserum to GLAST was from Chemicon international (Temecula, CA, USA). All second antibodies used for immunohistochemical analysis were from Molecular Probes (Eugene, OR, USA).
All procedures in this study were in accordance with the guidelines of the National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya-ku, Tokyo, Japan. Primary cultures of astrocytes were prepared from the cerebral cortices of 3-day-old neonates of Wistar rats, as described previously (Suzuki et al. 2001). Briefly, dissociated cortical cells were suspended in modified EMEM containing 30 mm glucose, 2 mm glutamine, 1 mm pyruvate and 10% fetal bovine serum. They were then plated on uncoated 75 cm2 flasks at the density of 600 000 cells/cm2. A monolayer of type I astrocytes was obtained 12–14 days after the plating. Non-astrocytes such as microglia were detached from the flasks by shaking and were removed by changing the medium. Astrocytes in the flasks were dissociated by trypsinization and reseeded on uncoated 96 well micro titer plates or cover glasses at 20 000 cells/cm2. When astrocytes became confluent (approximately 9–10 days after the reseeding), the medium was switched to modified EMEM containing 10% HS (gelding), followed by 7 day-incubation. In this culture, > 98% of the cells were identified as type I astrocytes by positivity for GFAP and by their flattened, polygonal appearance.
Measurement of the extracellular l-glu concentration
The extracellular l-glu concentration was measured by the colorimetric method modified by Abe et al. (2000). Briefly, 50 µL of culture supernatant was transferred to a 96-well micro titer plate and was mixed with 50 µL of substrate mixture consisting of 20 U/mL GLD, 2.5 mg/mL β-NAD, 0.25 mg/mL MTT, 100 µm MPMS and 0.1% (v/v) Triton X-100 in 0.2 m Tris-HCl buffer (pH 8.2). After 10 min-incubation at 37°C, the reaction was stopped by adding 100 µL of solution containing 50% (v/v) dimethylformamide and 20% (wt/vol) SDS (pH 4.7). In this reaction, MTT (yellow) is converted into MTT formazan (purple) in proportion to the l-glu concentration (Fig. 1b). The amount of MTT formazan was determined by measuring the absorbance with a microplate reader at 570 nm (test wavelength) and at 655 nm (reference wavelength). The concentration of l-glu was estimated from the standard curve, which was constructed in each assay using cell-free media containing known concentrations of l-glu.
L-Glu was dissolved at 1 mm in phosphate buffered saline (PBS) and diluted to 100 µm with the culture medium. DHK and THA were freshly dissolved at 1 mm and diluted to final concentrations with the culture medium. E2, EE, DES, PNP, BPA and 17α were dissolved at 10 mm in ethanol and diluted to final concentrations with the culture medium. E2-BSA was freshly dissolved at 100 µm and diluted to final concentrations with the culture medium. ICI was dissolved at 1 mm in ethanol and coapplied with E2, EE, DES and E2-BSA at final concentrations. 16αIE2 and genistein were dissolved at 1 mm in ethanol and diluted to final concentrations with the culture medium.
Western blot analysis
Cells were washed twice with ice-cold PBS and then harvested. After intense sonication (23 kHz, 1 min × 3), the cell suspension was centrifuged at 800 g for 5 min at 4°C. An aliquot of this supernatant was removed for protein assay. Another aliquot was diluted in SDS sample buffer. Molecular mass markers and protein samples that contain equal amounts of protein were separated by electrophoresis on 10% polyacrylamide-SDS gels and transferred onto polyvinyliden difluoride membranes. The membranes were incubated with PBS containing 0.5% (v/v) Tween20 and 5% (wt/vol) skim milk for 1 h at 25°C, followed by overnight incubation at 4°C with rabbit polyclonal IgG to ERα (1 : 1000) or rabbit polyclonal IgG to ERβ (1 : 1000). The membranes were then washed for 30 min and incubated with peroxidase-conjugated anti-rabbit IgG (1 : 1000) for 1 h at 25°C. Immunoreactive proteins were visualized by the ECL kit. Specificities of the primary antibodies were confirmed by using the lysate of ovary cells in which strong expression of ERα and ERβ had already been confirmed. We also confirmed that the primary antibodies did not show any non-specific signals by using the lysate of COS7 cells that express neither ERα nor ERβ (Fig. 7a).
Cells were washed 3 times with PBS for 5 min and fixed with 4% PFA in 0.1 m phosphate buffer (PB) for 30 min at 4°C. After washing with Tris-buffered saline (TBS, pH 7.6), cells were treated with TBS containing 10% (wt/vol) BSA, 0.1% (v/v) TritonX-100 and 0.2% (v/v) Tween20 for 1 h at room temperature. For immunostaining of GFAP, cells were incubated overnight at 4°C with mouse IgG to GFAP diluted at 1 : 1000 in TBS containing 1% (wt/vol) BSA. After washing with TBS 3 times for 5 min, they were incubated with Alexa Fluor 488 rabbit antimouse IgG (1 : 200) for 3 h at room temperature (25°C). After washing with TBS 3 times for 5 min, fluorescent images were obtained by µ-Radiance laser scanning confocal system (Bio-Rad, Hercules, CA). For immunostaining of ERα or ERβ, cells were incubated overnight at 4°C with rabbit polyclonal IgG to ERα (1 : 500) or rabbit polyclonal IgG to ERβ (1 : 500). After washing with TBS 3 times for 5 min, cells were incubated with Alexa Fluor 488 goat anti-rabbit IgG (1 : 200) for 3 h at room temperature. After washing with TBS 3 times for 5 min, fluorescent images were obtained by the laser scanning confocal system using a 60 × oil immersion objective. For double-immunostaining of ERα and GLAST, cells were incubated overnight at 4°C with a mixture of rabbit polyclonal IgG to ERα (1 : 500) and the guinea pig antiserum to GLAST (1 : 4000). After washing with TBS 3 times for 5 min, cells were incubated with the mixture of Alexa Fluor 488 goat anti-rabbit IgG (1 : 200) and Alexa Fluor 568 goat anti-guinea pig IgG (1 : 200) for 3 h at room temperature. After washing with TBS 3 times for 5 min, fluorescent images were obtained by the laser scanning confocal system using a 60 × oil immersion objective.
Data were obtained from 4 independent experiments (averaged values of 4 wells for each). Data are expressed as means of these data ± SEM values. Tests of variance homogeneity, normality and distribution were performed to ensure that the assumptions required for standard parametric anova were satisfied. Statistical analysis was performed by one-way repeated-measure anova and post hoc Tukey's test for multiple pairwise comparisons.
L-Glu uptake activity of cultured astrocytes was evaluated by the clearance of exogenously applied l-glu. The principle of the measurement of l-glu concentrations is shown in Fig. 1(b). When l-glu was applied to the culture medium at 100 µm, it decreased to below half within 60 min and then to below 1 µm within 4 h (Fig. 1a). We measured the clearance of l-glu 60 min after the application as an index in following experiments. When the extracellular Na+ was removed, the l-glu clearance was completely abolished (Fig. 1c), indicating that the extracellular l-glu was uptaken by Na+ dependent l-glu transporter(s). To identify the predominant l-glu transporter(s) of cultured astrocytes, we co-applied THA (Balcar et al. 1977; an inhibitor of both of GLT-1 and GLAST) or DHK (Johnston et al. 1978; a selective inhibitor of GLT-1) with l-glu. THA (1 mm) completely blocked the l-glu clearance, while DHK (1 mm) had no effect, indicating that l-glu was uptaken by GLAST.
We investigated the effects of E2, EE, DES, PNP, BPA and 17α on the l-glu uptake by cultured astrocytes (Fig. 3). The chemical structures of these compounds are shown in Fig. 2. After 24 h treatment, E2 inhibited the l-glu clearance at 1 µm and higher concentrations. EE and DES also inhibited the l-glu clearance at 1 nm and higher concentrations. At 100 µm, E2, EE and DES (1 µm) inhibited the l-glu clearances by 34, 40 and 22%, respectively. The other four compounds had no effect. When ICI, a pure anti-estrogen, was co-applied at 10 nm with E2, EE and DES, it completely blocked the inhibitory effects of these compounds (Fig. 4), indicating that the effects were mediated by classical ER(s). When the extracellular Na+ was removed or 1 mm THA was applied, the effect of E2 (1 µm) completely disappeared (Fig. 8), demonstrating that the inhibition of l-glu clearance was caused by down-regulation of l-glu uptake via GLAST.
Some reports have suggested that a part of classical ERs are translocated to plasma membranes and mediates acute effects of estrogens (Pappas et al. 1995; Razandi et al. 1999). Therefore, we next investigated the effect of E2-BSA, a membrane-impermeable conjugate of E2 and BSA, on the l-glu uptake by cultured astrocytes. As is shown in Fig. 5(a), l-glu clearance was inhibited dose-dependently by a 24 h treatment of E2-BSA. The effect was significant at 1 nm and higher concentrations, and the clearance was inhibited by 30% at 100 µm. BSA alone had no effect at the concentrations corresponding to those included in E2-BSA. ICI (10 nm) completely blocked the effect of E2-BSA (1 µm), implying that this effect was mediated by classical ERs on plasma membranes. When the extracellular Na+ was removed or 1 mm THA was added, the effect of E2-BSA (1 µm) disappeared (Fig. 8), demonstrating that the effect of E2-BSA was caused by down-regulation of GLAST.
The results mentioned above indicate that estrogens down-regulate GLAST via classical ER(s) locating on the plasma membranes of cultured astrocytes. Therefore, we tried to identify the receptor isotype(s) mediating these effects. First, we tested the effects of 16αIE2 (Shughrue et al. 1999; Singh et al. 2000; a selective ligand for ERα) and genistein (Witkowska et al. 1997; a selective ligand for ERβ) on the l-glu uptake by cultured astrocytes (Fig. 6). At 10 nm, 16αIE2 revealed a significant inhibitory effect, while genistein had no effect at this concentration. When the extracellular Na+ was removed or 1 mm THA was added, the effect of 16αIE2 (10 nm) disappeared (Fig. 8), demonstrating that the effect of 16αIE2 was caused by down-regulation of GLAST as in the case of E2 and E2-BSA. Western blot analysis of the lysate of cultured astrocytes detected a single band of ERα, while little or no signal of ERβ was observed (Fig. 7a). Thus, we compared the subcellular localization of ERα with that of GLAST immunohistochemically (Fig. 7b). Although ERα was distributed over the whole cell, the strong signals were observed on the plasma membrane and in the nucleus (left). Although weak signals of GLAST were distributed in the nucleus and the soma, much stronger signals scattered on the plasma membrane (center). When two fluorescent images were merged, co-localization of ERα with GLAST on the plasma membrane became apparent (yellow dots, right). Taken together, these results strongly suggest that ERα is expressed on plasma membranes of cultured astrocytes and its activation causes the down-regulation of GLAST.
We investigated the effects of estrogens and xenoestrogens on l-glu uptake activity of cultured astrocytes. Our results are summarized as follows: (1) estrogens (E2, EE and DES) inhibited the l-glu uptake of cultured astrocytes by 24 h treatment, while xenoestrogens had no effect, (2) ICI completely blocked the effects of E2, EE and DES, (3) E2-BSA inhibited the l-glu uptake by 24 h treatment and the effect was completely blocked by ICI, (4) 16αIE2 inhibited the l-glu uptake, while genistein had no effect, (5) the predominant ER of cultured astrocytes was ERα, (6) ERα was co-localized with GLAST on plasma membranes of cultured astrocytes.
L-Glu is a major excitatory neurotransmitter in the CNS. The only rapid way to remove l-glu from the extracellular fluid surrounding the receptors is by cellular uptake by l-glu transporters (Balcar and Johnston 1972; Logan and Snyder 1972; Johnston 1981). To evaluate l-glu uptake activity of cultured astrocytes, we monitored the changes in the concentration of exogenously applied l-glu by a modified colorimetric method. This assay has already been confirmed to be optimal for measuring 1–200 µm of l-glu (Abe et al. 2000). The precise l-glu concentration in the synaptic cleft following release is still controversial. Shikorski and Stevens 1997) reported the peak concentration of 12 µm based on the assumption that the vesicle emptying is fast and complete. On the other hand, Clements et al. (1992) estimated it 1.1 mm by EPSC recording technique.
Among estrogen-related compounds used in this study, estrogens (E2, EE and DES) inhibited the l-glu uptake by cultured astrocytes, while the other compounds had no effect. Because of the transactivities of these compounds (E2 = EE = DES > PNP > BPA) (Nishikawa et al. 1999), we speculated that the effects of E2, EE and DES were mediated by classical type of ER(s). This speculation was further supported by the finding that ICI, a pure inhibitor for both of ERα and ERβ, completely blocked the inhibitory effects of these compounds.
Some reports have suggested that a part of classical ERs are translocated to plasma membranes and mediate acute effects of estrogens. Razandi et al. (1999) reported that the transfection of CHO cells with classical ERs caused their expression in both nuclei and membranes. ERs on the plasma membranes of tumor cells were structurally similar to classical ERs (Pappas et al. 1995). In the present study, we tested E2-BSA, a membrane-impermeable conjugate of E2 and BSA. E2-BSA also inhibited the l-glu uptake by cultured astrocytes and its effect was completely blocked by ICI. E2-BSA is generally used for the discrimination between classical nuclear ERs and ERs locating on plasma membranes. However, Stevis et al. (1999) pointed out that 1/100 free E2 was present in the commercial preparations of E2-BSA. In our experiment, the inhibitory effect of E2-BSA was observed even at lower concentration than that of E2. Thus, we considered that E2-BSA itself revealed the effect through classical ER(s) locating on plasma membranes.
We tried to identify the receptor isotype(s) mediating these effects. To date, 2 isotypes of classical ERs (ERα and ERβ) have been cloned. We first investigated the effects of 16αIE2 and genistein on l-glu uptake. The affinity of 16αIE2 for ERα and ERβ are 5.03 nm and 146.7 nm, respectively, and the selectivity for ERα has been ensured below 10 nm (Shughrue et al. 1999; Singh et al. 2000). Genistein binds preferentially to ERβ in the low to mid nanomolar range (Witkowska et al. 1997). L-Glu uptake was significantly inhibited by 16αIE2, but not by genistein at 10 nm, suggesting that ERα mediates the inhibitory effects of estrogens. Western blot analysis showed that the predominant ER isotype of cultured astrocytes was ERα. Immunohistochemical analysis showed that ERα was co-localized with GLAST on plasma membranes. These results support the contribution of ERα to the l-glu uptake inhibition further. Although some reports have suggested the existence of membrane ERs in astrocytes (Beyer et al. 1999; Hosli et al. 2000), their isotypes are unknown. Recently, Toran-Allerand et al. (2002) reported that ER-X, a novel plasma membrane-associated ER, elicited rapid activation of mitogen-activated protein (MAP) kinases, ERK1 and ERK2, in organotypic neocortical cultures. Because the effects mediated through ER-X were not inhibited by anti-estrogens (Singh et al. 1999), we considered that ER-X did not contribute to the inhibitory effect of estrogens on l-glu uptake.
At present, several signal transduction pathways are known to down-regulate l-glu uptake via GLAST. Activation of cAMP pathway resulted in a transient decrease in GLAST mRNA (Espinoza-Rojo et al. 2000). Although PKC was also reported to interact with GLAST, its effect was more complicated. Pan-isoform PKC inhibition suppressed GLAST uptake activity (Bull and Barnett 2002; Lortet et al. 2002; Conradt and Stoffel 1997), while their activation decreased GLAST mRNA (Espinoza-Rojo et al. 2000). Cytokines were also reported to down-regulate GLAST uptake activity via NO production (Ye and Sontheimer 1996; Hu et al. 2000). Interestingly, estrogens have long been known as an important vasoprotective molecules that cause the rapid dilation of blood vessels (Guetta and Cannon 1996). Recent studies clarified that estrogen activated endothelial nitric oxide synthase (eNOS) through activation of PI3K via ERα locating on plasma membranes of endothelial cells and produced NO, thereby causing the rapid dilation of blood vessels (Chen et al. 1999; Haynes et al. 2000). We are currently investigating the mechanisms underlying the down regulation of GLAST via membrane ERα. We have confirmed that a 10 min-treatment of E2 or E2-BSA elicits a significant inhibitory effect via the activation of phophatidylinositol 3-kinase (PI3K) (unpublished data). These results raise the possibility that the inhibitory effects of estrogens are mediated by the mechanisms including NO production. Taken the co-localization of ERα with GLAST into consideration, there might exist the signal transduction unit comprised of ERα, PI3K, NOS, GLAST, etc. in the vicinity of the cell surface, thereby causing the rapid down-regulation of GLAST. Whether the down regulation of GLAST was caused by the decrease in the transport activity or in the number of functional GLAST remains unknown. Although we have confirmed that total expression of GLAST is not changed by E2 or E2-BSA (unpublished data), the translocation of GLAST is currently under investigation.
To our knowledge, this is the first report showing that estrogens down regulate l-glu uptake activity of astrocytes via ERα locating on plasma membranes. A growing body of evidence has suggested that complex interplay between astrocytes and neurons modifies the formation, maintenance and efficacy of synapses (Oliet et al. 2001; Thompson 2003). L-Glu transporters of astrocytes regulate extracellular l-glu concentrations in the synaptic clefts, thereby terminating synaptic transmissions and preventing the neuronal excitotoxicity. Oliet et al. (2001) reported that the degree of the astrocytic coverage of neurons determined the diffusion of l-glu in the extracellular space and that the reduction in l-glu uptake by astrocytes affected neuronal transmitter release. Ultrastructural analysis showed that in the astrocytes that contact with neurons, ERα was translocated to the cell surface that surrounded neuronal spines (Milner et al. 2001). Estrogens might modulate synaptic functions via ERα on plasma membranes of astrocytes.
We thank Dr. Abe (Hoshi University, Tokyo, Japan) for helpful advice on experimental procedures, Ms. Horikawa for the contribution to some experiments and Dr. Richard Hochberg for providing us with 16αIE2. This work was partly supported by Health and Labour Science Research Grants for Research on Advanced Medical Technology and Research on Environmental Health from the Ministry of Health, Labour and Welfare, Japan awarded to YO and KN and a Grant-in-Aid for Young Scientists from the Ministry of Education, Science, Sports and Culture, Japan (KAKENHI 15700280) awarded to KS.