Abstract
The purpose of this study was to identify factors from astrocytes that can regulate LHRH neurosecretion. Exposure of LHRH-secreting (GT1–7) cells to conditioned media (CM) from C6 glial cells and hypothalamic astrocytes (HA) stimulated LHRH release. Assays of C6 and HA CM revealed that transforming growth factor-β1 (TGF-β1) and 3α-hydroxy-5α-pregnane-20-one (3α,5α-THP), both known LHRH secretagogues, were present in CM and their levels increased in parallel to the LHRH-releasing activity of CM. In contrast, TGF-α was undetectable in C6 or HA CM. Ultrafiltration to remove peptides with molecular weights >10 kDa virtually abolished the LHRH-releasing ability of the HA CM. Furthermore, immunoneutralization with a panspecific THF-β antibody dose-dependently attenuated the LHRH-releasing activity of the CM. Rat hypothalamus and GT1–7 cells were demonstrated to express TGF-β receptors as well as furin, an enzyme that converts latent TGF-β1 to active TGF-β1. Estrogen receptor-α and ER-β mRNA and protein were also demonstrated in HAs by reverse transcription-polymerase chain reaction and double immunofluorescence, and treatment with 17β-estradiol (17β-E2) increased both active and latent TGF-β1 levels in HA CM. The effect of 17β-E2 was completely blocked by the ER antagonist ICI8280. As a whole, these studies provide evidence of a previously undescribed 17β-E2-TGF-β1-LHRH signaling pathway.
Introduction
The decapeptide, LHRH, is considered a primary neural signal involved in induction of the preovulatory LH surge [1–5]. As a releasing factor, LHRH is released by nerve terminals in the median eminence where it is transported to the anterior pituitary by hypophyseal portal veins to act on gonadotropes to stimulate gonadotropin secretion [6]. The factors that regulate LHRH secretion are thus of significant interest and have been a subject of intense investigation. Due to this intense investigation, it is now well accepted that 17β-estradiol (17β-E2) released by the preovulatory follicle is the primary trigger for the LHRH and LH surge [7–9]. Other work suggests that the other major ovarian steroid hormone, progesterone may also act to facilitate and enhance the magnitude of the preovulatory LH surge [8–12].
In general, the mechanism whereby 17β-E2 regulates LHRH secretion is believed to be due to an indirect effect on neighboring neurons that impinge upon and regulate LHRH neurons, although there have been some recent challenges to this hypothesis [7, 13, 14]. Much attention has been focused on 17β-E2 action to regulate the release of inhibitory or excitatory transmitters in the central nervous system (CNS) to control LHRH secretion [4, 15–18]. Thus, there is growing evidence that 17β-E2 acts in part to stimulate the LHRH and LH surge by turning off inhibitory neurons and turning on excitatory neurons in the CNS.
While transmitter-containing neurons are likely involved in the mechanism of the LHRH and LH surge, they are not the only cell types in the brain, and in fact, glia cells are present in approximately ninefold greater numbers as compared to neurons. Generally, glia cells have been viewed as serving only support or structural roles in the brain; however, there is increasing evidence that astrocytes also have a neuroregulatory role. Along these lines, conditioned-media (CM) from astrocytes has been shown to stimulate LHRH secretion from immortalized LHRH neurons [19–21]. However, there is a controversy concerning the identity of the active factor from astrocytes that is responsible for the LHRH-releasing activity of CM. Melcagni and coworkers have provided evidence that transforming growth factor (TGF)-β1 is the astrocyte active factor [19, 20], while Ojeda and coworkers propose that the astrocyte active factor is TGF-α [21]. Part of the discrepancy could be due to the fact that many, but not all, of the studies supporting a role for TGF-β1 were performed using cortical astrocytes [19, 20], whereas the study supporting a role for TGF-α was performed using hypothalamic astrocytes (HAs) [21]. This could be important, as it has been reported that the functions and responses of astrocytes vary in a brain region–specific manner [22]. Another potential reason for the lack of clarity on this issue is that endogenous TGF-β1 and TGF-α levels in astrocyte CM have never been measured, and thus direct evidence is lacking on whether the growth factors are produced and whether the levels correlate with the LHRH-releasing activity of the CM.
Transforming growth factor-β1 and TGF-α are not the only potential neuroregulatory factors present in astrocyte CM, as other enhancing and/or additive factors may also exist. One candidate molecule for such a function is the neurosteroid, 3α-hydroxy-5α-pregnane-20-one (3α,5α-THP) that is produced by astrocytes and that we and others have shown to possess LHRH-releasing activity [23–25]. Finally, to be important physiologically, it would be critical to show that HAs are targets for estrogen action through demonstration of the presence of estrogen receptor (ER)-α or -β in HAs and by demonstrating estrogen regulation of the astrocyte active factor that controls LHRH release. The studies described in this article were designed to address these issues, with the overall goal of enhancing our understanding of astrocyte-LHRH neuronal interactions.
Materials and Methods
Cell Culture
GT1–7 cells
Immortalized LHRH neuronal cells, GT1–7 cells, donated by Dr. Pamella Mellon (University of California, San Diego) were cultured in 75-cm2 flasks and grown in Dulbecco's modified Eagle medium (DMEM) containing 5% fetal calf serum, 5% horse serum, and 1% penicillin-streptomycin (Life Technologies, Grand Island, NY) under an atmosphere of 5% CO2-95% O2 at 37°C [26]. At 70% confluence, cells were recovered by trypsin-EDTA and seeded in poly-d-lysine-coated 24-well plates at 200 000 cells/well. Upon reaching 70% confluence, DMEM was replaced with Opti-Mem I reduced serum medium containing transferrin (15 μg/ml), insulin (15 μg/ml), and 1% penicillin-streptomycin. Cells were used for incubations 16 h later.
C6 glial cells
C6 glial cells were cultured in 75-cm2 flasks and grown in DMEM containing 5% fetal calf serum, 5% horse serum, and 1% penicillin-streptomycin under an atmosphere of 5% CO2-95% O2 at 37°C [27, 28]. Upon reaching confluence, cells were recovered by trypsin-EDTA and seeded in six-well plates at 400 000 cells/well. At 80% confluence DMEM was replaced with Opti-Mem I reduced serum medium containing transferrin (15 μg/ml), insulin (15 μg/ml), and 1% penicillin-streptomycin. This medium was collected following 6, 12, and 18 h of culture and referred to as C6 6 h, C6 12 h, and C6 18 h conditioned medium (C6 6h CM, C6 12h CM, C6 18h CM), respectively. The CM was stored at −20°C until time of assay for TGF-α, TGF-β1, and 3α,5α-THP content.
Hypothalamic astrocytes
Primary cultures of astrocytes were obtained from the hypothalami of 1- to 2-day-old Holtzman Sprague Dawley rats (Harlan Sprague Dawley Inc., Indianapolis, IN) by the method of McCarthy and de Vellis [29, 30]. Hypothalami were removed from newborn pups and immersed in saline solution (138 mM NaCl, 5.4 mM KCl, 1.1 mM Na2HPO4, 22 mM glucose, and 0.9 mM CaCl2) at 4°C. Tissue was partially disrupted and trypsin 250 (Difco, Detroit, MI) neutralized to pH 7.0 in saline solution was added to the suspension to a final concentration of 0.1%. Trypsinization was terminated by adding an equal volume of complete culture medium (DMEM–Ham's F-12 medium; DMEM/F-12, 1:1, v:v) containing 5% fetal calf serum, 5% horse serum, and 1% penicillin–streptomycin. Suspended cells were filtered through cell strainers, and cells in the filtrate were plated at high density in 75-cm2 flasks and grown in complete culture medium under an atmosphere of 5% CO2-95% O2 at 37°C. On Day 10, cell cultures were shaken for 18 h under an atmosphere of 5% CO2-95% O2 at 37°C (250 rpm, stroke diameter = 1.5 inches) to remove contaminating oligodendrocytes and neurons. Following shaking, astrocytes were recovered by trypsin-EDTA and replated at one-third of their confluent density. Upon reaching confluence, astrocytes were seeded on cell chamber slides for immunofluorescence cell staining (see below) and six-well plates at 400 000 cells/well for cell culture experiments. Upon reaching 80% confluence, complete culture medium was replaced with Opti-Mem I reduced serum medium containing transferrin (15 μg/ml), insulin (15 μg/ml), and 1% penicillin-streptomycin. This medium was collected following 6, 12, 18, and 24 h of culture and referred to as HA 6h, 12h, 18h, and 24h conditioned medium, respectively. The CM was stored at −20°C until time of assay for TGF-α, TGF-β1, and 3α,5α-THP content.
Immunofluorescence Cell Staining
Hypothalamic astrocytes were characterized by immunocytochemical procedures using a monoclonal antibody raised against the astrocytic marker glial fibrillary acidic protein (GFAP) [31]. Cultures were more than 95% pure as assessed by the number of cells containing GFAP. Cells were fixed in methanol at −10°C, washed in 0.01 M PBS (pH 7.3), and incubated for 1 h in blocking solution (10% normal goat serum and 0.1% Triton X-100 in 0.01 M PBS). Following washing in 0.01 M PBS, cells were incubated overnight in mouse anti-GFAP monoclonal antibody diluted 1:300 (Chemicon, Temecula, CA) and rabbit anti-ER-β polyclonal antibody at 10 μg/ml (Affinity Bioreagents Inc, Golden, CO) or rabbit anti-ER-α polyclonal antibody ER715 diluted 1:500 (provided by the National Hormone and Pituitary Program, the NIDDK, and the Center for Population Research of the NICHHD). The polyclonal antibody ER715 was raised against a recombinant peptide corresponding to the hinge region of the ER molecule [32]. Cells were washed for 30 min in 0.01 M PBS and incubated for 90 min in goat anti-rabbit IgG-TR (Texas Red–conjugated secondary antibody) diluted 1:200 (Jackson ImmunoResearch, West Grove, PA). Additional washes in 0.01 M PBS were followed by a 90-min incubation in goat anti-mouse IgG-fluorescein isothiocyanate (fluorescein-conjugated secondary antibody) diluted 1:50 (Jackson ImmunoResearch). After final washes in 0.01 M PBS, coverslips were mounted with FluorSave (Calbiochem, La Jolla, CA). Specificity of staining was demonstrated by omission of primary antisera (control).
Treatments
GT1–7 cells
Unless stated otherwise, GT1–7 cells were exposed to C6 glial cell CM or HA CM for a period of 1h, after which the media was assayed for LHRH content. In the filtration experiments, HA 24h CM was subjected to ultrafiltration through polyethersulfone membranes (nominal molecular weight limit: 10 kDa; Millipore, Bedford, MA), and then the filtered CM was incubated with GT1–7 cells to assess its LHRH releasing activity. In the immunoneutralization studies, various doses of a panspecific TGF-β neutralizing polyclonal antibody (Oncogene Research Products, Cambridge, MA) was added to HA 24h CM, and then the CM was incubated with GT1–7 cells and its LHRH-releasing activity assessed. To determine whether 3α,5α-THP may be involved in mediating the stimulatory effect of C6 glial cell conditioned medium on LHRH secretion, GT1–7 cells were exposed to C6 18h CM in the presence of increasing doses of picrotoxin (25, 50, 100, and 200 μM), an antagonist to the γ-aminobutyric acid (GABA)A receptor, the receptor that is thought to mediate the action of 3α,5α-THP. To determine whether the ability of TGF-β1 to modulate LHRH release may be enhanced by 3α,5α-THP, GT1–7 cells were exposed to TGF-β1 (25 and 50 ng/ml), 3α,5α-THP (1 μM), and TGF-β1 (25 ng/ml) + 3α,5α-THP (1 μM) for 1 h in Krebs–Ringer bicarbonate (KRB) buffer (pH 7.4; Sigma Chemical Co., St. Louis, MO). The GT1–7 incubations were carried out under an atmosphere of 5% CO2-95% O2 at 37°C. Following all incubations, GT1–7 cells were tested by incubation with 56 mM KCl. Release of LHRH was observed in response to KCl exposure.
C6 glial cells
To determine whether l-glutamate is able to stimulate TGF-β1 release, C6 glial cells were exposed to l-glutamate (100 and 500 μM) in KRB buffer (pH 7.4) under an atmosphere of 5% CO2-95% O2 at 37°C for 30 min, and the buffer was collected for TGF-β1 measurement. Following l-glutamate exposure, C6 glial cells were cultured in Opti-Mem I reduced serum medium containing transferrin (15 μg/ml), insulin (15 μg/ml), and 1% penicillin–streptomycin under an atmosphere of 5% CO2-95% O2 at 37°C. The medium was collected following 6, 12, and 18 h of culture and assayed for TGF-β1 content.
Hypothalamic astrocytes
To determine whether the steroids 17β-E2 and progesterone are able to regulate TGF-β1 release from HAs in vitro, astrocytes were exposed for 36 h to 17β-E2 (0.25 nM, 0.5 nM, 5.0 nM), the diluent ethanol (5 μl/ml), or 17β-E2 (0.5 nM) for 36 h + progesterone (200 nM) during the last 3 h of the 36-h culture period. Coincubation with the ER antagonist ICI-182,780 (7a-[9-[4,4,5,5,5,pentafluoropentyl]sulfinyl]nonyl]-estra-1,3,5(10)-trene-3,17β-diol) (Trocris Cookson Inc., Ballwin, MO) (10−6 M) was also used to confirm that estradiol action involved mediation by ERs. All incubations were carried out in Opti-Mem I reduced serum medium containing transferrin (15 μg/ml), insulin (15 μg/ml), and 1% penicillin-streptomycin under an atmosphere of 5% CO2-95% O2 at 37°C. The medium was collected following 36 h of culture and assayed for TGF-β1 content. In some experiments (Fig. 14), the astrocyte conditioned media was treated with 1.0 N HCl according to the manufacturer's protocol to activate latent TGF-β1, so that total TGF-β1 levels could be determined.
Radioimmunoassay
Luteinizing hormone releasing hormone
Levels of LHRH were determined by RIA as described by Brann and Mahesh [33]. The UZ-8 LHRH primary antibody was used at a final dilution of 1:100 000 with a sensitivity of 0.5 pg/tube. Goat anti-rabbit secondary antibody was used at a final dilution of 1:250. Hormone levels were measured in 100 μl samples and expressed in terms of pg/well.
Transforming growth factor-α
Levels of TGF-α in C6 6h, C6 12h, C6 18h, HA 6h, 12h, 18h, and 24h conditioned medium were determined by RIA using a kit purchased from Peninsula Laboratories (San Carlos, CA). The TGF-α was measured in 100-μl samples with a sensitivity of 2.0 pg/tube.
3α,5α-THP
3α,5α-THP levels in C6 6h, C6 12h, C6 18h, HA 6h, 12h, 18h, and 24h conditioned medium were determined by RIA as previously described [34]. The antibody for 3α,5α-THP measurements was purchased from Dr. R.H. Purdy (Veterans Medical Research Foundation, San Diego, CA) and used at a final dilution of 1:1000 with a sensitivity of 10.0 pg/tube. Hormone levels were measured in 100-μl samples and expressed in terms of ng/ml.
Enzyme-Linked Immunofluorescence Assay
TGF-β1
TGF-β1 levels in C6 6h, C6 12h, C6 18h, HA 6h, 12h, 18h, and 24h conditioned medium were determined by enzyme–linked immunosorbent assay using a kit purchased from Promega (Madison, WI). The system detects biologically active TGF-β1 in an antibody sandwich format. Plates were coated with monoclonal TGF-β1 antibody that binds soluble TGF-β1 from solution. Captured TGF-β1 was bound by a polyclonal antibody specific for TGF-β1 and following washing was detected using anti-rabbit IgG conjugated to horseradish peroxidase (HRP) and a chromogenic substrate for HRP. The TGF-β1 was measured in 100-μl samples with a sensitivity of 32 pg/ml.
Reverse Transcription-Polymerase ChainReaction Analysis
Hypothalamic tissue was removed from adult male and female rats (Holtzman, Sprague Dawley, Madison, WI) and pooled in separate groups. Total RNA was isolated from hypothalamic tissue, GT1–7 cells, and HAs using Tri-Reagent (Molecular Research Center Inc., Cincinnati, OH) according to manufacturer's recommendations. Reverse transcription-polymerase (RT-PCR) was performed as described previously by our laboratory [35]. Primers were designed to amplify coding regions of the genes of interest: TGF-β type I receptor: 5′-GGTTTATGAGCAGGGGAAGAT-3′; 5′-CACAGTTTTTGAGCGAGGTTA-3′ product length = 1513 bp, TGF-β type II receptor: 5′-GTGGAGGAAGAACAACAAGAA-3′; 5′-GGCAACAGGTCAAGTCGT-3′ product length = 923 bp, TGF-β type III receptor: 5′-TGGAGTCAGGCGATAATGGAT-3′; 5′-AGAGGAATGTTGCGGTGGTAA-3′ product length = 1483 bp, ER-α: 5′-AGTCCTGGACAAGATCAACGA-3′; 5′-ATGAAGACGATGAGCATCCAG-3′ product length = 220 bp, ER-α: 5′-AATGCTCACACGCTTCGAG-3′; 5′-AACTTGGCATTCGGTGGTAC-3′ product length = 292 bp, furin: 5′-TATGGCTACGGGCTGTTGGAT-3′; 5′-GACGCTGGCACGGATGAT-3′ product length = 651 bp [36–39].
Protein Isolation and Western Blotting
Protein isolation and Western blotting was performed as described previously by our laboratory [40, 41]. Liver and hypothalamic tissue were isolated from female rats (Holtzman, Sprague Dawley, Madison, WI). A polyclonal TGF-β RI antibody at 1 μg/ml (Santa Cruz Biotechnology, Santa Cruz, CA) or a polyclonal TGF-β RII antibody at 1 μg/ml (Santa Cruz Biotechnology) was used. Following washing in 1× TBS (20 mM Tris, 137 mM NaCl, 0.1% Tween 20), binding of primary antibody was detected using a horseradish peroxidase (HRP)-conjugated secondary antibody and enhanced chemiluminescence system (Amersham Pharmacia Biotech, Piscataway, NJ). Blots were exposed to Kodak Biomax film (Rochester, NY), and molecular weight determination was made using a digital imaging system (IS-1000; Alpha Innotech, San Leandro, CA).
Statistical Analysis
The effect of different treatments was analyzed using a one-way ANOVA followed by the Student Newman-Keuls multiple comparison test. For all groups, n = 5. The results were expressed as means ± standard error. P values < 0.05 were considered significant.
Results
In our initial studies we utilized the C6 glial cell line to validate whether it would be a useful model for studying glial cell active factors. As shown in Figure 1A, exposure of GT1–7 cells to C6 18h CM for 1 h significantly increased LHRH release compared to cells incubated in Opti-Mem I reduced serum medium that had not been exposed to C6 glial cells (Control). GT1–7 cells exposed to 6 h and 12 h C6 glial cell CM had mean LHRH levels that were not different from controls. To determine whether C6 glial cells are capable of producing growth factors and neurosteroids implicated in modulating LHRH neuronal function, TGF-α, TGF-β1 and 3α,5α-THP levels were measured in C6 glial cell CM. As shown in Figure 1B, C6 glial cell 6h, 12h, and 18h CM contained detectable levels of TGF-β1; however, TGF-β1 levels were nondetectable in Opti-Mem I reduced serum medium that had not been previously exposed to C6 glial cells (Control). The TGF-β1 levels were significantly higher in 18h CM as compared to 6h and 12h CM. Note that the significantly higher TGF-β1 levels in 18h CM correlated well with the LHRH releasing ability of the CM. In contrast to our findings with TGF-β1, TGF-α levels were nondetectable in any sample (Fig. 1B).
Effect of C6 6h CM, C6 12h CM, and C6 18h CM on LHRH release from GT1–7 cells (A). The TGF-α and TGF-β1 (B) and 3α,5α-THP (C) content in C6 6h CM, C6 12h CM, and C6 18h CM. ND = nondetectable. Groups with different subscripts are significantly different P < 0.05
Figure 1C shows the levels of the neurosteroid 3α,5α-THP in C6 glial cell CM. As shown in Figure 1C, 3α,5α-THP levels were detected in 6h, 12h, and 18h C6 glial cell CM with highest levels observed in the 12h and 18h CM. Note that the levels of 3α,5α-THP in C6 glial cell CM did not correlate exactly to the LHRH-releasing ability of the CM as equivalent levels of 3α,5α-THP were observed in 12h and 18h CM, but only 18h CM significantly elevated LHRH release from GT1–7 cells.
To clarify further whether 3α,5α-THP could be the active factor from C6 glial cells in regulating LHRH release from GT1–7 cells, we used an antagonist for the GABAA receptor (picrotoxin) to see if it would eliminate the LHRH-releasing activity of 18h C6 glial cell CM. The GABAA receptor was targeted because 3α,5α-THP has high affinity for this receptor and it's many actions are believed to be mediated through this receptor [25, 42, 43]. The doses of picrotoxin were selected based on their ability to block 3α,5α-THP stimulation of LHRH release [24] (unpublished data). The GT1–7 cells were exposed to C6 glial cell 18h CM for 1 h in the presence of increasing doses of picrotoxin. As shown in Figure 2, picrotoxin alone had no significant effect on LHRH release. As expected, C6 glial cell 18h CM significantly increased LHRH release from GT1–7 cells, and this effect was not reversed by the low doses of 25 and 50 μM of picrotoxin. Higher doses of picrotoxin (100 and 200 μM) yielded intermediate LHRH-releasing activity by the 18h CM such that the LHRH levels were not significantly different from either the control or 18h CM, but they were significantly different compared to the picrotoxin (200 μM) control. Thus, 3α,5α-THP would appear to have only a minor contribution to the LHRH-releasing activity of C6 glial cell CM.
Effect of the chloride channel blocker picrotoxin on C6 18h CM-induced LHRH release in GT1–7 cells. *P < 0.05 versus control, picrotoxin (100 μM), and picrotoxin (200 μM). **P < 0.05 versus picrotoxin (200 μM)
The amino acid neurotransmitter, glutamate, is a major regulator of LHRH release [44], although this effect is thought to be indirect. Because glutamate receptors are present on glial cells [45, 46], we wanted to examine whether glutamate could enhance TGF-β1 release from C6 glial cells, with the hypothesis that glutamate may be an important neuronal factor for recruiting astrocyte participation in enhancing LHRH release. However, this hypothesis appeared invalid as 100 μM and 500 μM glutamate treatment did not enhance TGF-β1 release from C6 glial cells, and in fact was inhibitory (Fig. 3).
The TGF-β1 content in C6 glial cell conditioned medium following a 30-min exposure to the excitatory amino acid l-glutamate. Levels of TGF-β1 were also measured 6, 12, and 18 h following vehicle (6h, 12h, and 18h post-KRB) and l-glutamate (6h, 12h, and 18h postglutamate 100 μM and 500 μM) treatment. ND = nondetectable. *P < 0.05 versus 18h post-KRB
While the C6 glial cell line proved to be a good model for studying astrocyte factors involved in regulating LHRH release, we next wanted to confirm the results using the more physiological HA model. As shown in Figure 4, A–C, HA CM caused a stepwise elevation of LHRH release beginning with a slight nonsignificant increase in mean LHRH levels with HA 12h CM, followed by significant elevations in LHRH levels from exposure to HA 18h and 24h CM (Fig. 4A). Similar to the results obtained with C6 glial cells, TGF-α levels were undetectable in all samples, while TGF-β1 levels showed a stepwise elevation in 12h, 18h, and 24h HA CM (Fig. 4B). The elevations of TGF-β1 showed a strong correlation with the LHRH-releasing activity of the HA CM (compare Fig. 4, A and B). As shown in Figure 4C, 3α,5α-THP was also detectable HA CM with a pattern for increasing levels in 12h, 18h, and 24h HA CM that also correlated with the LHRH-releasing activity of the CM.
A) Effect of HA 6h, 12h, 18h, and 24h CM on LHRH release from GT1–7 cells. *P < 0.05 versus control and HA 6h CM. **P < 0.05 versus control, HA 6h CM, HA 12h CM, and HA 18h CM. B) The TGFα and TGFβ1 content in HA 6h, 12h, 18h, and 24h CM. ND = nondetectable. Groups with different subscripts are significantly different (P < 0.05). C) The 3α,5α-THP levels in HA 6h, 12h, 18h, and 24h CM. ND = nondetectable. *P < 0.05 versus HA 12h CM
To determine whether the active factor in HA CM was TGF-β1 or 3α,5α-THP, two types of experiments were performed. In the first experiment, HA 24h CM was subjected to ultrafiltration with a filter that removes all proteins with molecular weights greater than 10 kDa (TGF-β1 = 25 kDa; 3α,5α-THP being a nonprotein would not be removed by the filter). The filtered CM was then used in GT1–7 incubations to determine its LHRH-releasing ability. As shown in Figure 5A, nonfiltered HA 24h CM stimulated LHRH release from GT1–7 cells as expected. However, filtered HA 24h CM was unable to enhance LHRH release significantly from GT1–7 cells, although mean LHRH levels were slightly higher in filtered HA 24h CM versus control. Thus, the active factor is a protein with a molecular weight greater than 10 kDa.
(A) Effect of ultrafiltration on the stimulatory effect of HA 24h CM on LHRH release from GT1–7 cells. Groups with different subscripts are significantly different (P < 0.05). (B) Effect of increasing doses of a panspecific TGF-β neutralizing antibody (Ab) on the stimulatory effect of HA 24h CM on LHRH release from GT1–7 cells. *P < 0.05 versus control. **P < 0.05 versus HA 24h CM and HA 24h CM + 0.2 μg/ml Ab
To confirm further that TGF-β1 is the astrocyte-active factor in regulating LHRH release from GT1–7 cells, a second experiment was conducted in which a pan-specific TGF-β antibody was used to immunoneutralize TGF-β1 in the HA 24h CM to determine whether this would eliminate the LHRH-releasing ability of the CM. As shown in Figure 5B, HA 24h CM stimulated LHRH release from GT1–7 cells as expected, and this effect was blocked in a dose-dependent manner by increasing concentrations of the TGF-β pan-specific antibody.
Because these studies suggested that TGF-β1 is the major astrocyte-active factor in CM responsible for stimulating LHRH release, we next sought to determine whether TGF-β type I, II, or III receptors are present in GT1–7 cells and the rat hypothalamus using RT-PCR and Western blot analysis. As shown in Figure 6A, mRNA for all three TGF-β receptor types were expressed in rat hypothalamus, while only TGF-β receptor type I and II were expressed in GT1–7 cells. No product was observed in the absence of Superscript II RT enzyme, demonstrating that the products were not due to genomic DNA contamination. Western blot analysis confirmed that the 59-kDa TGF-β receptor type I and 76-kDa TGF-β receptor type II protein were both expressed in the hypothalamus and GT1–7 cells (and in the positive control tissue, liver). Due to technical difficulties, we were unable to examine protein expression for the TGF-β receptor type III.
A) The RT-PCR analysis for the expression of TGF-β type I, II, and III receptor in GT1–7 cells and hypothalamic tissue of the female rat in the presence (+) or absence (−) of reverse transcriptase. The RT-PCR products of 1513 bp, 923 bp, and 1483 bp, representing TGF-β type I, II, and III receptor, respectively, were detected in hypothalamic tissue. The expression of TGF-β type I and II receptor was also demonstrated in the GT1–7 cell line. B) Western blot analysis for TGF-β type I and II receptor protein expression in GT1–7 cells, hypothalamus, and liver tissue (positive control) of the female rat. The 59-kDa and 76-kDa bands representing TGF-β type I and II receptors, respectively, were detected in all tissues/cells
Brain astrocyte CM has been shown previously to consist of two-thirds latent and one-third active TGF-β1 [47, 48]. Using RT-PCR analysis we thus examined whether the hypothalamus and GT1–7 cells express mRNA transcripts for furin, an enzyme that activates latent TGF-β1 [49–51]. As shown in Figure 7, both rat hypothalamus and GT1–7 cells were found to express the transcript for furin. Sequencing of the 651-bp amplified products confirmed that they correspond to furin. No product was observed if Superscript II RT enzyme was omitted, verifying that the product was not due to genomic DNA contamination.
The RT-PCR analysis for the expression of furin in GT1–7 cells and hypothalamic tissue of the male rat in the presence (+) or absence (−) of reverse transcriptase. An RT-PCR product of 651 bp representing furin was detected in GT1–7 cells and hypothalamic tissue of the male rat
The above studies provide evidence that the astrocyte-active factor for the regulation of LHRH release is TGF-β1 with a possible small contribution by 3α,5α-THP. The TGF-β1 and 3α,5α-THP have been shown previously to enhance LHRH release from GT1–1 cells when added exogenously, with effective doses ranging from 5–50 ng/ml for TGF-β1 and 0.1–10 μM for 3α,5α-THP. As shown in Figure 8, the lowest effective dose for TGF-β1 stimulation of LHRH release in our experiments was 50 ng/ml. The lowest effective dose for 3α,5α-THP in our studies was 10 μM (data not shown). As also shown in Figure 8, low doses of TGF-β1 and 3α,5α-THP were not found to be synergistic.
Effect of 3α,5α-THP and TGF-β1 on LHRH release from GT1–7 cells. *P < 0.05 versus vehicle, 3α,5α-THP (1 μM), and 3α,5α-THP (1 μM) + TGF-β1 (25 ng/ml)
To be physiologically relevant, it would be critical to show that HAs are targets for 17β-E2, the trigger of the LHRH and LH surge. Therefore, we examined whether HAs express ER-α and -β mRNA and protein by using RT-PCR and double-immunofluorescence staining. As shown in Figure 9, RT-PCR analysis revealed the expected 220-bp product for ER-α and 292-bp product for ER-β in HAs. No product was observed in the absence of murine Moloney leukemia virus RT enzyme, demonstrating that the product was not due to genomic DNA contamination. As shown in Figure 10, double immunofluorescence staining with antibodies to ER-α and the astrocyte-specific protein GFAP confirmed that HAs possess ER-α immunoreactivity. Similarly, double immunofluorescence staining with antibodies to ER-β and GFAP revealed that ER-β immunoreactive protein is also present in HAs, which agrees with the RT-PCR studies (Fig. 11). To determine if the ER in HAs was functional, we examined whether 17β-E2 could regulate TGF-β1 release from HAs. Toward this end, HAs were exposed to either vehicle or three different doses of 17β-E2 for 36 h and then the medium was collected and assayed for TGF-β1. The 36-h exposure period was selected because it is well known that an exposure period of at least 29–36 h is necessary for 17β-E2 to induce the LHRH and LH surge in vivo. The doses of 17β-E2 were selected because they represent physiological levels of 17β-E2 that brain cells would be expected to be exposed to on proestrus [52–56]. As shown in Figure 12, levels of TGF-β1 were nondetectable in Opti-Mem I reduced serum medium that had not been previously exposed to HAs (Control). In contrast, the vehicle group that was exposed to HAs had detectable TGF-β1 levels. Of significant interest, all three doses of 17β-E2 caused a significant elevation of TGF-β1 release from HAs. The stimulatory effect of 17β-E2 was maximal at 0.5 nM, with no additional enhancement observed with increasing the dose to 5.0 nM. As shown in Figure 13, the stimulatory effect of 17β-E2 was reproducible, as 0.5 nM 17β-E2 significantly elevated TGF-β1 release from HAs, and it was specific as the addition of progesterone (200 nM) with 17β-E2 did not enhance the stimulatory effect of 17β-E2. As shown in Figure 14, acid treatment of HA CM revealed that 17β-E2 enhanced release of both active TGF-β1 (Fig. 14A) and latent TGF-β1 (Fig. 14B). Furthermore, approximately 30–40% of the total TGF-β1 in HA CM was in the active form, while the remaining 60–70% was in the latent form. As also shown in Figure 14, A and B, the stimulatory effect of 17β-E2 on both active and latent TGF-β1 release by HAs appears to be mediated by ERs, as evidenced by the finding that coincubation with an ER antagonist, ICI-182,780 totally abolished the stimulatory action of 17β-E2.
The RT-PCR analysis for the expression of ER-α and ER-β in HAs in the presence (+) or absence (−) of reverse transcriptase. The RT-PCR products of 220 bp and 292 bp representing ER-α and ER-β, respectively, were detected in HAs
Double immunostaining for GFAP (A,C) and ER-α (B,D) in HAs in vitro. The cells were plated at low density so that double immunostaining would be more easily visualized. Immunoreactivity was not detected in the absence of primary antisera (data not shown). Magnification: ×400, ×800
Double immunostaining for GFAP (A,C) and ER-β (B,D) in HAs in vitro. The cells were plated at low density so that double immunostaining would be more easily visualized. Immunoreactivity was not detected in the absence of primary antisera (data not shown). Magnification: ×800
Effect of 36 h exposure to 17β-E2 (0.25, 0.5, 5.0 nM) on TGF-β1 release from HAs in vitro. ND = nondetectable. Groups with different subscripts are significantly different (P < 0.05)
Effect of treatment with 17β-E2 (0.5 nM) for 36 h + P4 (200 nM) during the last 3 h of the 36-h culture period on TGF-β1 release from HAs in vitro. ND, nondetectable. Groups with different subscripts are significantly different (P < 0.05)
A) Effect of the ER antagonist ICI-182,780 (10−6 M) on 17β-E2 (0.5 nM) stimulation of active TGF-β1 release from HAs in vitro. B) Effect of the ER antagonist ICI-182,780 (10−6 M) on 17β-E2 (0.5 nM) stimulation of total (active + latent) TGF-β1 release from HAs in vitro. *P < 0.05 versus all other groups; **P < 0.01 versus all other groups
Discussion
The present study demonstrates that C6 glial cells and HAs produce factors that can regulate LHRH secretion. We observed similarities between the C6 glial cell and HA models with respect to the factors secreted and their LHRH-releasing activity. We are not aware of any other previous studies that characterized the factors released by C6 glial cells or assessed their ability to regulate LHRH secretion. Based on our findings, we conclude that C6 glial cells provide a useful and convenient model for studying potential interactions between glial cells and neurons.
Using the C6 glial cell line and HAs, our studies revealed that both C6 glial cells and HAs secrete significant quantities of the growth factor, TGF-β1 and the neurosteroid, 3α,5α-THP. In contrast, TGF-α levels were undetectable in C6 glial cell and HA CM. Our finding of undetectable levels of TGF-α in C6 glial cell CM and HA CM is in good agreement with previous reports that showed that TGF-α is undetectable in brain astrocytes unless it is induced by secretogues such as prolactin [57]. Thus, TGF-α does not appear to be the active factor from C6 glial cells or HAs responsible for the LHRH-releasing ability of CM.
On the other hand, TGF-β1 remains a viable candidate as the active factor from C6 glial cells and HAs responsible for stimulating LHRH release from GT1–7 cells. The TGF-β1 was shown to be present in CM from C6 and HAs, and the levels of TGF-β1 in CM showed a strong correlation with the LHRH-releasing activity of the CM. Filtration of HA CM, which would remove TGF-β1, virtually eliminated the LHRH-releasing ability of the CM. Likewise, immunoneutralization with a panspecificTGF-β antibody also virtually abolished the LHRH-releasing activity of the CM. Based on these observations, we conclude that the majority of the LHRH-releasing activity of HA CM is due to TGF-β1.
In support of a direct effect of TGF-β1 on LHRH neurons, our studies revealed that TGF-β type I and II receptor are present on GT1–7 cells and in rat hypothalamus. Other investigators have reported that TGF-β type I receptor is present in LHRH neurons in vivo [58]. In further support of a direct action on LHRH neurons, TGF-β1 stimulated LHRH release directly from LHRH neurons in vitro at a dose of 50 ng/ml in our study. A dose of 5 ng/ml of TGF-β1 was reported to be effective in a previous study [19, 20]. It is unclear why a higher dose of TGF-β1 was required in our study to stimulate LHRH secretion from the LHRH neurons. The difference in sensitivities between the two studies could be due to use of different LHRH cell lines—we used the GT1–7 cell line while the other study used the GT1–1 cell line. In support of this possibility, other investigators also found that the two LHRH cell lines also have different sensitivities to another growth factor, basic fibroblast growth factor [59].
While ng/ml concentrations of exogenous TGF-β1 were required to stimulate LHRH release from GT1–7 cells, only pg/ml concentrations of active TGF-β1 were observed in HA CM. This discrepancy could be explained by two possibilities: 1) GT1–7 cells possess a mechanism for converting latent TGF-β1 into active TGF-β1, and/or 2) there are additive factors in CM. Acid treatment of HA CM revealed that approximately 30–40% of TGF-β1 exists in the active form in CM, with the remaining 60–70% of TGF-β1 existing in the latent form. This finding agrees well with previous studies by Caldwell and coworkers [47, 48] who found that only approximately one-third of the TGF-β1 in frozen brain astrocyte CM or retinal glial cell CM is in the active form. Admittedly, latent TGF-β1 is considered to be inactive and requires activation to be functional biologically. However, our study raises the intriguing possibility that LHRH neurons may be capable of activating latent TGF-β1 as we found that rat hypothalamus and GT1–7 cells express the transcript for furin, a key enzyme that activates latent TGF-β1 [49–51].
Our second possibility suggests that there could exist additive factors in astrocyte CM that work with TGF-β1 to stimulate LHRH release. Along these lines, 3α,5α-THP could function as a possible additive factor because it was present in significant concentrations in CM, and its levels correlated with the LHRH-releasing activity of CM. Furthermore, picrotoxin treatment of C6 glial cell CM did reduce the LHRH-stimulatory effect of the CM. However, 3α,5α-THP most likely plays a minor role in the LHRH-releasing activity of CM as both ultrafiltration and TGF-β immunoneutralization, procedures that would not remove 3α,5α-THP from CM, eliminated the majority of the LHRH-releasing activity of CM. Thus, other additive factors must exist. It should be mentioned that the TGF-β neutralizing antibody used in our studies would neutralize TGF-β2 and TGF-β3 in addition to TGF-β1; thus, TGF-β2 and TGF-β3 could be candidates as additive factors in CM. In support of this possibility, both TGF-β2 and TGF-β3 have been shown to be produced by astrocytes [60, 61]. Furthermore, TGF-β2 has been shown to enhance LHRH secretion [62]. Thus, these TGF-β family members may participate with TGF-β1 to enhance LHRH secretion. Additionally, TGF-β1 has also been shown to stimulate two other LHRH-regulatory factors from astrocytes, prostaglandin E2 and interleukin-6 [63, 64]. These factors may also participate with TGF-β1 to enhance LHRH secretion. Further work is ongoing in our laboratory to address these possibilities.
Finally, our double immunofluorescence staining studies suggest that HAs possess both ER-α and ER-β immunoreactive protein. Estrogen receptor-α had been previously reported to be present in HAs [65], but to our knowledge, our study is the first to suggest that ER-β is also present in HAs. The double immunofluorescence result for ER-β has to be interpreted with some degree of caution because ER-β antibodies have been reported to exhibit some cross-reactivity to ER-α in other studies. However, RT-PCR using primers specific for ER-β provided alternate confirmation that the ER-β is expressed in HAs. It should also be pointed out that ER-β colocalization has also been reported in hippocampal astrocytes [66], suggesting that ER-β may be expressed in astrocytes from several different parts of the brain.
A functional role for ERs in HAs was suggested by our finding that 17β-E2 can enhance TGF-β1 release from HAs in vitro. The enhancing effect of 17β-E2 was exerted upon both the active and latent forms of TGF-β1. The ability of 17β-E2 to regulate release of TGF-β1 is intriguing, as 17β-E2 is well known to be the trigger for the preovulatory LHRH and LH surge [9], and it thus provides a mechanism for recruitment of astrocytes into the surge mechanism. The stimulatory effect of 17β-E2 on TGF-β1 release was specific, as the neurotransmitter, glutamate did not stimulate TGF-β1 release and in fact was inhibitory. The steroid hormone progesterone was also without effect on TGF-β1 release. The effect of 17β-E2 upon TGF-β1 release appeared to be an ER-mediated effect, as evidenced by the fact it was completely blocked by the potent ER antagonist, ICI-182,780. It should be mentioned that 17β-E2 has also been shown to increase TGF-β1 release from bone [67, 68], uterus [69, 70], and skin during the process of wound-healing [71]. Thus, the 17β-E2–TGF-β1 pathway appears to be an endocrine signaling pathway common to many different systems and potentially important in many different physiological processes. As a whole, the present studies provide further evidence that astrocytes have important neuroregulatory capabilities that are subject to endocrine regulation. Thus, astrocytes should no longer be considered in the unidimensional light of possessing only structural, segregational functions in the brain.
Acknowledgments
We thank Ms. Lynn Chorich, Ms. Liesl De Sevilla, and Mrs. Marlene Wade for their technical assistance in immunostaining, furin RT-PCR analysis, and the ER antagonist studies, respectively. We also thank Dr. Jim O'Conner for his expertise in LHRH iodinations.
References
Author notes
This research was conducted as part of the requirements for completion of the Ph.D. degree and was supported by a research grant (HD-28964) from the NIH/NICHHD. C.B.'s research was also supported by an NRSA Institutional Training Grant (HD07253) from the NIH/NICHHD.
