Hypothalamic Astrocytes Respond to Transforming ... - Semantic Scholar

0013-7227/97/$03.00/0 Endocrinology Copyright © 1997 by The Endocrine Society

Vol. 138, No. 1 Printed in U.S.A.

Hypothalamic Astrocytes Respond to Transforming Growth Factor-a with the Secretion of Neuroactive Substances That Stimulate the Release of Luteinizing Hormone-Releasing Hormone* YING J. MA, KARIN BERG-VON DER EMDE, FLORENCE RAGE†, WILLIAM C. WETSEL, AND SERGIO R. OJEDA Division of Neuroscience, Oregon Regional Primate Research Center, Oregon Health Sciences University (Y.J.M., K.B.E., F.R., S.R.O.), Beaverton, Oregon 97006; and the Laboratory of Cell and Molecular Pharmacology, National Institute of Environmental and Health Science (W.C.W.), Research Triangle Park, North Carolina 27709 ABSTRACT Previous studies demonstrated the involvement of transforming growth factor-a (TGFa), a member of the epidermal growth factor (EGF) family, in the developmental regulation of hypothalamic LHRH release. Although both TGFa and EGF stimulate LHRH release, they do not appear to act directly on LHRH neurons, as no EGF/TGFa receptors are detected on these cells in vivo. Instead, the stimulatory effect of TGFa on LHRH release seems to require a glial intermediacy. The present study identifies one of the glial molecules involved in this process. In vitro exposure of purified hypothalamic astrocytes to TGFa or EGF in a defined medium led to activation of the cyclooxygenase-mediated pathway of arachidonic acid metabolism, as indicated by an increase in PGE2 release, but failed to affect lipooxygenase-mediated metabolism, as assessed by the lack of increase in leukotriene C4 production; addition of TGFa- (T-CM) or EGF-conditioned medium to cultures of LHRH-producing GT1-1 cells stimulated LHRH release. In contrast, direct exposure of GT1-1 cells

to the growth factors was ineffective. Incubation of the cells in medium conditioned by untreated astrocytes (CM) was also ineffective. Blockade of either EGF receptor signal transduction or cyclooxygenase activity in the astrocytic cultures prevented both TGFa-induced PGE2 formation in astrocytes and the stimulatory effect of T-CM on LHRH release. Immunoneutralization of PGE2 actions or selective removal of the PG from T-CM also prevented T-CM-induced LHRH release. Addition of exogenous PGE2 restored the effect. Thus, PGE2 is one of the glial molecules involved in mediating the stimulatory effect of TGFa on LHRH release. The effectiveness of PGE2 in eliciting LHRH release was, however, greatly reduced when PG was delivered to GT1-1 cells in astrocyte-defined medium instead of CM. Thus, astrocytes appear to produce a yet to be identified substance(s) that facilitates the stimulatory effect of PGE2 on LHRH output. We postulate that the ability of TGFa to enhance LHRH release depends on the potentiating interaction of PGE2 with these additional glialderived molecules. (Endocrinology 138: 19 –25, 1997)

A

only the release of LHRH (5), but also the morphological and biochemical differentiation of the LHRH neuronal network (6 – 8). For instance, heat-labile molecules produced by astrocytes have been shown to enhance the neuronal phenotype, induce LHRH release, and promote the proliferation of a LHRH neuronal cell line (7). Other experiments have demonstrated that astrocytes can also produce heat-resistant bioactive substances that are able to stimulate LHRH output. One of these LHRH-releasing substances has been tentatively identified as transforming growth factor-b1 (TGFb1) (8). The results of these two studies are consistent with the suggestion proposed earlier that the cell-cell regulatory mechanisms used by glial cells to influence LHRH neuronal function involve trophic molecules of a peptidergic nature (3, 4). An important group of glial regulatory molecules is represented by growth factors acting via tyrosine kinase receptors, such as basic fibroblast growth factor (bFGF) (6, 9) and TGFa (5, 10). The former acts directly on LHRH neurons, via activation of specific receptor molecules, to enhance neuronal differentiation and to promote processing of the LHRH prohormone, without stimulating release of the mature decapeptide (6, 9, 11). TGFa, on the other hand, is an effective stimulator of LHRH release (5, 6), but it does not appear to

N ENHANCED secretion of the hypothalamic neuropeptide LHRH is essential for the initiation of mammalian puberty. Over the years, a variety of experimental approaches have led to the conclusion that this activation results from the interplay of a multiplicity of transsynaptic inputs of both an inhibitory and an excitatory nature (for review, see Refs. 1 and 2). In addition to such neurotransmitter-mediated regulation, a body of evidence is beginning to emerge in support of the concept that LHRH neuronal function is also influenced by molecules of astrocytic origin (3, 4) that are able to affect not Received July 15, 1996. Address all correspondence and requests for reprints to: Dr. Sergio R. Ojeda, Division of Neuroscience, Oregon Regional Primate Research Center, Oregon Health Sciences University, 505 NW 185th Avenue, Beaverton, Oregon 97006. * This work was supported by NIH Grant HD-25123 (to S.R.O.), the NIEHS Intramural Research Program (W.C.W.), P30 Population Center Grant HD-18185, and Grant RR-00163 for the operation of the Oregon Regional Primate Research Center. This is publication 2010 of the Oregon Regional Primate Research Center. † Postdoctoral Research Fellow supported by INSERM, France, and Grant HD-25123. Present address: Lab Neurobiologie, Endocrinologique URA, 1197 CNRS Montpellier II, place Eugene Bataillon, 34090 Montpellier, France.

19

20

GLIA-MEDIATED LHRH RELEASE

act directly on LHRH neurons to exert this stimulatory effect (5, 6, 12). Instead, the receptors that recognize TGFa have been detected on astrocytes (6, 13), suggesting that the stimulatory effect of TGFa on LHRH release involves the intermediacy of glial cells (5, 6, 10, 13). As hypothalamic astrocytes express the genes for both TGFa and its receptor (12, 13) and because they respond to TGFa with up-regulation of its own gene expression (12), we postulated that TGFa acts in a paracrine/autocrine fashion to stimulate the astrocytic production of bioactive molecules able to enhance LHRH release (5, 12, 13). The findings that TGFa stimulates both LHRH and PGE2 release from median eminence fragments and that this increase in LHRH release can be suppressed by blocking PG synthesis (5) led to the suggestion that PGE2 is one of the glial molecules involved in the astrocyte-dependent regulation of LHRH secretion (4, 5). The present study provides experimental evidence in support of this concept. Partial reports of these findings have appeared previously (14, 15). Materials and Methods Animals Pregnant Sprague-Dawley rats were purchased from Bantin and Kingman (Fremont, CA). They were housed in a room with a controlled photoperiod (14 h of light and 10 h of darkness) and temperature (23–25 C). Two or 3 days after birth, the pups were used to prepare astrocyte cultures.

Cell culture Astrocytes. Hypothalamic astrocytes were isolated by the method of McCarthy and de Vellis (16), as previously reported (12). Briefly, after culturing hypothalamic cell dispersates to confluency (8 –10 days), contaminating cells (neurons and oligodendrocytes) were removed by shaking the cultures at 250 rpm for 6 h, replacing the medium, and shaking again for 18 h. The astrocytes were then seeded in 6-well plates at 800,000 cells/well and grown in DMEM-Ham’s F-12 medium (1:1, vol/vol) supplemented with 10% calf serum until reaching 90% confluence (;4 days). At this time, the growth medium was replaced by an astrocytedefined medium (ADM; see below), and the cells were cultured for an additional 48 h before treatment. We have previously shown that the cultures are more than 95% pure, as assessed by the number of cells containing the astrocytic marker glial fibrillary acidic protein (12). ADM consisted of DMEM (lacking glutamate and phenol red) supplemented with l-glutamine (2 mm), HEPES (15 mm), insulin (5 mg/ml), and putrescine (100 mm). GT1-1 cells. These cells were seeded in 24-well plates at 200,000 cells/well and grown in DMEM containing 10% FCS, penicillin (100 U/ml), and streptomycin (100 mg/ml) under an atmosphere of 5% CO2-95% air at 37 C. Upon reaching 50 – 60% confluence, the medium was replaced with a neuronal defined medium (6) consisting of glutamate-free DMEM supplemented with transferrin (100 mg/ml), putrescine (100 mm), lglutamine (2 mm), sodium selenite (30 nm), and insulin (5 mg/ml). The cells were used for experiments 24 h later.

Treatments Astrocytes. In a previous study (12) we reported that both TGFa and epidermal growth factor (EGF) at 50 ng/ml and the phorbol ester 12O-tetradecanoyl phorbol-13-acetate (TPA) at 10 ng/ml were highly effective in inducing TGFa gene expression in cultures of hypothalamic astrocytes. We also reported that similar doses of TGFa and EGF stimulate PGE2 release from median eminence fragments in vitro (5). The same doses were, therefore, used in the present study to determine whether hypothalamic astrocytes respond to activation of EGF receptors (EGFR) by TGFa or EGF or to activation of protein kinase C-mediated

Endo • 1997 Vol 138 • No 1

pathways by TPA with the release of substances able to stimulate LHRH release. As previous results implicated PGE2 as one of these bioactive substances (5), it was used as the end point in culture medium collected after 4, 8, 16, and 24 h of treatment. Based on the release profile of the PG, the 16-h interval was selected for collection of medium conditioned by untreated astrocytes (CM) or by astrocytes treated with TGFa- (TCM) or EGF (E-CM). Because 16-h T-CM treatment effectively released LHRH from GT1-1 cells, experiments were conducted to verify that both the effect of TGFa on PGE2 release and the conditioning effect of the growth factor on astrocyte culture medium involve activation of EGFR. In these experiments, astrocytes were simultaneously treated with TGFa (50 ng/ml) and tyrphostin RG-50864 (60 mm), a selective inhibitor of EGFR tyrosine kinase activity (17) (Rhoˆne-Poulenc Rorer Central Research, Horsham, PA); we previously showed that treatment of hypothalamic astrocytes with this dose of RG-50864 blocks the effect of TGFa on its own gene expression (12). Additional experiments were performed to determine whether the conditioning effect of TGFa involves only the formation of products in the cyclooxygenase pathway of arachidonic acid metabolism or if it also includes the formation of lipooxygenase metabolites. To examine this issue, leukotriene C4 (LTC4) and PGE2 were measured in the culture medium of astrocytes treated with TGFa for 16 h. Other cultures were treated with TGFa plus different doses of indomethacin (Id), a preferential cyclooxygenase inhibitor, or nordihydroguaiaretic acid (NDGA), a preferential lipooxygenase inhibitor. In addition to measuring LTC4 and PGE2, the effectiveness of the treatments was determined by assessing the ability of the different media to stimulate LHRH release from GT1-1 cells. GT1-1 cells. To determine whether T-CM or E-CM was able to stimulate LHRH release, GT1-1 cells were exposed to CM, E-CM, or T-CM for 30 min, and the medium was collected for LHRH measurement. A direct stimulatory effect of TGFa and EGF on GT1-1 cells was assessed by treating GT1-1 cells, cultured in neuronal defined medium, with either growth factor. As the signal transduction pathway that mediates the effects of TGFa and EGF on cellular function involves activation of protein kinase C (18, 19), parallel cultures were treated with TPA to ensure that the cells were responsive to the direct, nonreceptor-mediated, activation of this pathway. T-CM from astrocytes in which the cyclooxygenase pathway had been inhibited by Id failed to stimulate LHRH release, suggesting that the effect of T-CM requires the presence of cyclooxygenase products. As TGFa induces PGE2 release from hypothalamic astrocytes (this study), and PGE2 is the major cyclooxygenase product involved in the control of LHRH release (20), two series of experiments were carried out to define the role of PGE2 in T-CM-induced LHRH release. In the first, T-CM was incubated with the highly specific PGE2 antiserum C51011/23 (21) for 30 min at room temperature (10 ml antiserum/ml T-CM) before addition to GT1-1 cells. Control tubes were incubated with a similar volume of normal rabbit serum. In the second series, various dilutions of the PGE2 antiserum were added to T-CM and incubated for 1 h at room temperature, followed by the addition of 300 ml of a 1:1 slurry of protein A-Sepharose (Sigma Chemical Co., St. Louis, MO) in PBS and incubation of the mixture for 1 h at room temperature with constant tipping. Antibody-PGE2-protein A-Sepharose complexes were then removed by centrifugation at 13,000 3 g for 1 min. The effectiveness of this immunoneutralization was assessed by measuring PGE2 by RIA. A 1:750 ratio of antibody to T-CM (vol/vol) effectively removed all measurable PGE2 from the culture medium. GT1-1 cells were then treated with this medium, and the ability of exogenous PGE2 to restore the effect of T-CM on LHRH release was determined by adding a known amount (2 ng/ml) of PGE2 to the PGE2-depleted T-CM.

RIAs PGE2. PGE2 released by cultured astrocytes was measured as previously described (21). The tritiated PGE2 ([5,6,8,11,14,15-N-3H]PGE2; 171 Ci/ mmol) used in the assay was obtained from DuPont-New England Nuclear (Wilmington, DE); the PGE2 antiserum (C510-11/23) was a generous gift from W. B. Campbell (Department of Pharmacology, Medical College of Wisconsin) and was used at a 1:8,000 dilution. The sensitivity of the assay was 3.6 pg PGE2/tube.


21

LHRH. LHRH released by GT1-1 cells into the culture medium was measured as previously reported (22). The assay employs 125I-labeled LHRH and the polyclonal antibody HFU60 (5) at a 1:25,000 dilution. The sensitivity of the assay was 2 pg/tube.

LTC4 An ELISA kit (Neogen Corp., Lexington, KY) was employed to measure the release of LTC4 from hypothalamic astrocytes. This assay is based on the competition between the unknown levels of LTC4 in experimental samples and a LTC4-horseradish peroxidase conjugate to bind a limiting amount of LTC4 antibody. The reaction is developed by measuring the degree of peroxidation of the substrate 3,39,5,59 tetramethylbenzidine in the presence of hydrogen peroxide. One-milliliter aliquots of culture medium were acidified with 1 n HCl (150 ml) and applied to C18 Sep-Pak columns (Waters Corp., Milford, MA) preconditioned by washing with 2 ml ethanol followed by 2 ml water. Upon application of the acidified sample, the column was washed with 1 ml water, followed by 1 ml petroleum ether (Sigma). The eicosanoids were then eluted with 1 ml formic acid methyl ester (Sigma). After evaporation of the solvent under a stream of nitrogen at room temperature (;30 min), the residue was dissolved in 200 ml assay buffer (provided by the manufacturer), and LTC4 was quantitated by ELISA in 50-ml samples. The concentrations of LTC4 were expressed as picograms per ml. The sensitivity of the assay was 2 pg/tube.

Statistics The effect of different treatments was analyzed using a one-way ANOVA followed by the Student-Newman-Keuls multiple comparison test for unequal replications.

Results Stimulation of LHRH release by TGFa requires a glial intermediacy

As previously seen using GT1-7 cells (6), direct exposure of GT1-1 cells to either TGFa or EGF failed to affect LHRH release (Fig. 1, left panel). That the cells were responsive to activation of one of the intracellular pathways used by these growth factors was shown by the ability of TPA, an activator of protein kinase C, to significantly stimulate LHRH release in parallel cultures. In contrast to their lack of direct response to TGFa, GT1-1 cells responded to T-CM or E-CM with a 2-fold increase in LHRH release compared to that in control cells treated with CM (Fig. 1, right panel). Medium conditioned by treating the astrocytes with TPA was also effective, but it did not result in a LHRH response greater than that caused by TPA alone. Astrocytes respond to TGFa with increased formation of PGE2

As a previous study showed that exposure of median eminence fragments to TGFa results in PGE2 release (5), the ability of TGFa and EGF to induce PGE2 release from hypothalamic astrocytes was examined. Figure 2 shows that both TGFa and EGF stimulate PGE2 production, with levels of PG increasing almost linearly between 4 –24 h of treatment. TPA was more effective than either growth factor in stimulating PGE2 release (Fig. 2). Blockade of glial EGFR signal transduction or PG synthesis prevents the stimulatory effect of T-CM on LHRH release

Inhibition of EGFR tyrosine kinase activity by the selective EGFR blocker RG-50864 (17) abolished the stimulatory effect

FIG. 1. Hypothalamic astrocytes mediate the action of TGFa and EGF on LHRH release from GT1-1 cells. Left panel, Inability of TGFa or EGF (50 ng/ml each) to directly stimulate LHRH release from GT1-1 cells compared to untreated control cells (C). In contrast, TPA (10 ng/ml) directly elicits LHRH release. Right panel, Ability of astrocyte culture medium conditioned by treatment of the cells for 16 h with TGFa (50 ng/ml; T-CM), EGF (50 ng/ml; E-CM), or TPA (10 ng/ml; TPA-CM) to stimulate LHRH release. Bars represent means, and vertical lines above the bars show the SEM. The numbers above the bars are the number of cell culture wells per group. *, P , 0.05; **, P , 0.01 (vs. C or CM).

of TGFa on PGE2 release and significantly reduced basal PGE2 output (Fig. 3, upper left panel). Moreover, the blocker eliminated the stimulatory effect of T-CM on LHRH release (Fig. 3, lower left panel). This inhibition was not due to a direct action of the receptor blocker on GT1-1 cells, because addition of RG-50854 to the cells failed to alter basal LHRH release (Fig. 3, open circles, lower left panel). Instead, the loss of LHRH-releasing activity in T-CM appeared to be related to the inability of TGFa to stimulate PG release upon blockade of glial EGFR. To verify this assumption, we treated astrocyte cultures with TGFa in the presence of Id, an inhibitor of PG synthesis and measured both PGE2 levels in the astrocyte culture medium and the LHRH response to Id-treated T-CM. Figure 3 shows that Id suppressed both basal and TGFa-stimulated PGE2 formation (upper right panel) and eliminated the stimulatory effect of T-CM on LHRH release (lower right panel). Direct addition of Id to GT1-1 cells did not affect LHRH release (Fig. 3, open circles, lower right panel). Products of the lipooxygenase pathway of arachidonic acid metabolism do not mediate the stimulatory effect of T-CM on LHRH release

In contrast to the clear-cut elevation in PGE2 levels observed after 16 h of TGFa treatment, levels of LTC4, the major product of the lipooxygenase pathway (23), were not increased over control values (Fig. 4, middle panel). Low doses (0.5–1 mm) of NDGA, a blocker of lipooxygenase activity,

22

Endo • 1997 Vol 138 • No 1


FIG. 2. Ability of TGFa (50 ng/ml), EGF (50 ng/ml), or TPA (10 ng/ml) to stimulate the production of PGE2 from hypothalamic astrocytes. The data are presented as the mean 6 SEM. Each group represents the mean of six to nine independent observations. * and **, First time points significantly greater than control values (P , 0.05 and P , 0.01, respectively).

affected neither the glial PGE2 response to TGFa nor the increase in LHRH induced by T-CM (Fig. 4). Higher doses (5 and 10 mm) suppressed the LHRH response to T-CM (Fig. 4, lower panel), but they also inhibited the stimulation of glial PGE2 by TGFa (Fig. 4, upper panel), without affecting the basal release of LTC4. Thus, the ability of NDGA to block the effect of T-CM on LHRH release is due to inhibition of PG synthesis and not to suppression of lipooxygenase activity. Glial PGE2 is a mediator of the stimulatory effect of T-CM on LHRH release

Neutralization of PGE2 actions by incubating T-CM with a specific PGE2 antiserum before addition to GT1-1 cells abolished the stimulatory effect of the conditioned medium on LHRH release (Fig. 5). Treatment of T-CM with normal rabbit serum was ineffective. Basal LHRH release was not altered by the culture medium of astrocytes treated with the PGE2 antiserum, but not exposed to TGFa, indicating that the antiserum by itself is unable to affect LHRH secretion. Glial PGE2 does not act alone to mediate T-CM-induced LHRH release

To determine whether the addition of exogenous PGE2 to T-CM devoid of the PG was able to restore the LHRH-

FIG. 3. Top panels, Inhibitory effect of RG-50864 (RG; 60 mM) and Id (10 and 25 mM) on PGE2 release from hypothalamic astrocytes treated with TGFa (50 ng/ml, 16 h). Bottom panels, Effect of T-CM and lack of effect of T-CM plus RG or T-CM plus Id on LHRH release from GT1-1 cells. Bars represent means, and vertical lines show the SEM. Numbers above bars are the number of observations. Open circles are mean LHRH released from GT1-1 cells that were directly treated with either RG (60 mM) or Id. *, P , 0.05; **, P , 0.01 (vs. all other groups).

releasing capability of T-CM, PGE2 in the astrocyte culture medium was immunoprecipitated before testing the T-CM on GT1-1 cells. As when the PG was bound to the antibodies but not removed from the medium (see Fig. 5), PGE2-deficient T-CM failed to induce LHRH release (Fig. 6). The addition of PGE2, at a dose (2 ng/ml) that is ineffective when administered in ADM, completely restored the response (Fig. 6), indicating that T-CM contains a substance(s) able to facilitate the stimulatory effect of the PG on LHRH release. In fact, this substance(s) also appears to be present in CM (the culture medium of control astrocytes not treated with TGFa), because CM allowed this normally ineffective dose of PGE2 to significantly enhance LHRH release (Fig. 6). Discussion

Morphological examination of the neuroendocrine hypothalamus under different physiological and experimental conditions has revealed the existence of strong structural


23

FIG. 5. Immunoneutralization of PGE2 in T-CM (see text for details) prevents T-CM-induced LHRH release. Bars shown the mean 6 SEM, and numbers above bars are the number of culture wells per group. Ab, PGE2 antibody; NRS, normal rabbit serum. *, P , 0.05 vs. all other groups.

FIG. 4. Top panel, Stimulatory effect of TGFa on PGE2 release from astrocyte cultures and inhibitory effect of different concentrations of NDGA on TGFa-induced PGE2 release. Middle panel, Absence of changes in LTC4 levels after 16 h of exposure to TGFa and inability of different concentrations of NDGA to inhibit basal LTC4 release. Bottom panel, Effect of T-CM plus NDGA on LHRH release. Bars show means, and vertical lines represent the SEM. Numbers above bars are the number of cell culture wells per group. **, P , 0.01 vs. all other TGFa-treated groups.

links between LHRH neurons and glial cells of astrocytic lineage (24). The glial apposition of LHRH neurons is not only substantial (25, 26), but also sexually dimorphic (27) and affected by gonadal steroids (25, 28), suggesting that hypothalamic glial cells may contribute to regulating LHRH neuronal function via actual physical changes in the glial-neuronal relationship. In addition to this morphological substrate, it now appears clear that hypothalamic astrocytes use trophic molecules of a peptidergic nature to influence the development and secretory activity of LHRH neurons. The glial growth factors

shown to affect LHRH neuronal function to date include TGFa (5), bFGF (6, 11), TGFb1 (8), and, very recently (29), neuregulins, members of the EGF family that initiate their actions by interacting with EGFR-related receptors. Perhaps the most well characterized of these polypeptides is TGFa, which has been demonstrated to be involved in the activation of LHRH release that occurs during both normal puberty and sexual precocity induced by hypothalamic lesions (10, 30). Rather surprisingly, LHRH neurons do not contain measurable EGFR able to transduce the TGFa signal (6, 13), implying the participation of an intermediate cell type in the process by which TGFa stimulates LHRH release. As hypothalamic astrocytes contain EGFR and respond to TGFa with upregulation of the TGFa gene (12), we postulated that astrocytes themselves are the cells that mediate the effect of TGFa on LHRH neurons (4 – 6, 12). The present results support this concept, as they demonstrate the inability of TGFa (and EGF) to act directly on GT1-1 cells to elicit LHRH release and the LHRH-releasing effectiveness of astrocyte culture medium conditioned by exposure of the cells to either growth factor. The lack of a direct TGFa effect on LHRH release may not be attributed to the short duration of the treatment (30 min), as a similar cell line (GT1-7) does not respond to a much longer (4, 8, and 24 h) exposure to the growth factor (6). It is also unlikely that higher doses of the growth factor would be effective, as a 1-mm concentration (6 mg/ml) of either TGFa or EGF failed to induce EGFR phosphorylation in GT1-7 cells, in contrast to their effectiveness when administered to astrocyte cultures (6). We cannot, however, rule out the possibility that LHRH neurons contain functional EGFR at very specific developmental windows (for instance, before or during migration) or that the receptors are present in a (changing) subpopulation of mature LHRH neurons. The former

24


FIG. 6. Ineffectiveness of T-CM-PGE2 [T-CM-PGE2(2)], in which PGE2 was immunoprecipitated, to release LHRH from GT1-1 cells, and restoration of the effect [T-CM-PGE2(1)] by reintroduction of PGE2 (2 ng/ml) to the PGE2-depleted medium. The ability of CM from untreated astrocytes to facilitate the effect of PGE2 on LHRH release is also depicted [CM-PGE2(1)]. Bars represent the mean 6 SEM. Numbers above bars are the number of culture wells per group. *, P , 0.05; **, P , 0.01 (vs. respective controls).

possibility is suggested by a recent report showing the presence of bioactive EGFR in a LHRH neuronal cell line different from GT1 cells (31). The results also demonstrate that one of the glial intermediates used by TGFa to elicit the release of LHRH is PGE2, and that the effect of TGFa on glial PGE2 synthesis requires the intermediacy of EGFR. That products of the cyclooxygenase-dependent metabolism of arachidonic acid are involved in mediating the effect of TGFa was demonstrated by the ability of indomethacin to prevent the stimulatory effect of the astrocyte-conditioned medium on LHRH release. The suppressive effect of specific immunoneutralization of PGE2 in the conditioned medium and the ability of exogenous PGE2 to restore the response identifies this PG as the major cyclooxygenase product mediating the TGFa effect on LHRH release. The relevance of these findings to the normal hypothalamus is indicated by the fact that median eminence fragments in vitro also respond to TGFa with PGE2 release and to indomethacin with suppression of the TGFa effect (5). It does not appear that lipooxygenase products of glial arachidonic acid metabolism are involved in the glia to neu-

Endo • 1997 Vol 138 • No 1

ron signaling mechanism by which TGFa promotes LHRH secretion. No changes in LTC4, the major product of the lipooxygenase-mediated pathway (23), were detected in astrocyte-conditioned medium 16 h after the addition of TGFa compared to that in untreated cultures. NDGA, an inhibitor of lipooxygenase activity, did not affect these basal levels of LTC4 and prevented the effect of the TGFa-conditioned medium on LHRH release only at doses that also suppressed PGE2 levels. Ligand-dependent activation of EGFR in other cell systems does result in the rapid synthesis of both cyclooxygenase and lipooxygenase products (32, 33), suggesting that, if astrocytes behave similarly, LTC4 levels may have already returned to basal values after 16 h of exposure to TGFa. Whatever the case, the absence of elevated LTC4 levels in the 16-h TGFa-conditioned medium in the face of increased PGE2 values strongly argues against a role for leukotrienes in mediating the effect of TGFa on LHRH release. Although the present results provide clear evidence for an involvement of astrocytic PGE2 in the glial-neuronal mechanism by which TGFa stimulates LHRH release, they also suggest the participation of additional unidentified molecules in this process. PGE2 at a concentration 3-fold higher than that detected in TGFa-conditioned medium failed to elicit LHRH release from GT1-1 cells, but it became effective in the presence of control CM. Melcangi et al. (8) showed that medium conditioned by immature astrocytes in culture contains a thermostable factor able to release LHRH. The factor was tentatively identified as TGFb. Although the ability of TGFb to potentiate the effect of PGE2 on LHRH release remains to be demonstrated, we noticed in preliminary experiments that a general blocker of excitatory amino acid receptors eliminates the potentiating ability of astrocyte-conditioned medium on PGE2-induced LHRH release. Thus, it is possible that excitatory amino acids released by astrocytes may be able to interact with PGE2 in the control of LHRH secretion. Hypothalamic astrocytes cultured in l-glutaminecontaining ADM, do, in fact, release the excitatory amino acids, aspartate and glutamate (Lee, B. J., et al., unpublished data). Further experimentation is required to identify the additional bioactive substance(s) released from astrocytes that contributes to the TGFa-mediated glial control of LHRH release. It should also be mentioned that in the present experiments, we did not observe a significant stimulatory effect of control CM on LHRH release. As this effect has been detected after exposing GT1 cells to CM for 1 h or longer (7, 8), our results may be explained by the short exposure time (30 min) used. At this point, it is important to recognize the emerging complexity of the glial-neuronal interactions that appear to control the secretory activity of LHRH neurons. Other growth factors acting via tyrosine kinase receptors, such as bFGF and neuregulins, have been recently shown to affect LHRH neuronal function (6, 11, 29). Although bFGF acts directly on LHRH neurons to facilitate neuronal differentiation, promote survival, and stimulate processing of the LHRH prohormone (6, 11), neuregulins stimulate LHRH release by inducing PGE2 formation in glial cells (29). Still other growth factors, such as TGFb, and insulin-like growth factors I and II have been shown to stimulate LHRH secretion (8, 34, 35). Although TGFb1 (8) and insulin-like growth factor II (35)


appear to act directly on LHRH neurons, the stimulatory effect of IGF-I shown using median eminence fragments (34) may be indirect, as the peptide is unable to stimulate LHRH release from GT1 cells (35). In addition to this growth factormediated regulation, LHRH neurons themselves appear to regulate glial proliferation via releasable factors (7), suggesting the existence of an intricate bidirectional communication between LHRH neurons and glial cells, mediated by a variety of intercellular signaling molecules. Although the results of the above-described in vitro experiments need to be interpreted with caution, they do suggest that such an intercellular signaling process may also function in vivo. It is likely that the biochemical communication between glial cells and LHRH neurons is intimately associated with the well known changes in cell to cell contacts that occur between the two cell types during reproductive life (24, 27, 28). The biological actions of PGE2 are initiated by binding of the PG to several membrane-anchored receptors (36). In recent experiments we demonstrated the presence of at least four types of PGE2 receptors in GT1-1 cells (37, 38). Among them, one subtype linked to calcium mobilization (EP1) has been identified by in situ hybridization techniques in LHRH neurons of immature animals (Rage, F., et al., unpublished data). Thus, it appears that LHRH neurons are endowed with the necessary recognition molecules to effectively transduce PGE2 signals derived from anatomically and functionally associated glial and neuronal networks.

13. 14. 15.

16. 17. 18. 19. 20.

21. 22. 23. 24.

25.

Acknowledgments We thank Janie Gliessman and Diane Hill for editorial assistance.

26.

References

27.

1. Ojeda SR, Urbanski HF 1994 Puberty in the rat. In: Knobil E, Neill JD (eds) The Physiology of Reproduction, ed 2. Raven Press, New York, vol 2:363– 409 2. Terasawa E 1995 Mechanisms controlling the onset of puberty in primates: the role of GABAergic neurons. In: Plant TM, Lee PA (eds) The Neurobiology of Puberty. Journal of Endocrinology, Bristol, pp 139 –151 3. Ojeda SR, Ma YJ, Berg-von der Emde K 1994 Neurotrophic factors and female sexual development. In: Sizonenko PC, Aubert ML, Vassalli J-D (eds) Frontiers in Endocrinology. Ares Serono Symposia, Rome, vol 6:111–123 4. Ojeda SR 1994 The neurobiology of mammalian puberty: has the contribution of glial cells been underestimated? J NIH Res 6:51–56 5. Ojeda SR, Urbanski HF, Costa ME, Hill DF, Moholt-Siebert M 1990 Involvement of transforming growth factor a in the release of luteinizing-hormone releasing hormone from the developing female hypothalamus. Proc Natl Acad Sci USA 87:9698 –9702 6. Voigt P, Ma YJ, Gonzalez D, Fahrenbach WH, Wetsel WC, Berg-von der Emde K, Hill DF, Taylor KG, Costa ME, Seidah NG, Ojeda SR 1996 Neural and glial-mediated effects of growth factors acting via tyrosine kinase receptors on LHRH neurons. Endocrinology 137:2593–2605 7. Gallo F, Morale MC, Avola R, Marchetti B 1995 Cross-talk between luteinizing hormone-releasing hormone (LHRH) neurons and astroglial cells: developing glia release factors that accelerate neuronal differentiation and stimulate LHRH release from GT1–1 neuronal cell line and LHRH neurons induce astroglia proliferation. Endocr J 3:863– 874 8. Melcangi RC, Galbiati M, Messi E, Piva F, Martini L, Motta M 1995 Type 1 astrocytes influence luteinizing hormone-releasing hormone release from the hypothalamic cell line GT1–1: is transforming growth factor-b the principle involved?. Endocrinology 136:679 – 686 9. Wetsel WC, Hill DF, Ojeda SR 1996 Basic fibroblast growth factor regulates the conversion of pro-luteinizing hormone-releasing hormone (LHRH) to LHRH in immortalized hypothalamic neurons. Endocrinology 137:2606 –2616 10. Ma YJ, Junier M, Costa ME, Ojeda SR 1992 Transforming growth factor alpha (TGFa) gene expression in the hypothalamus is developmentally regulated and linked to sexual maturation. Neuron 9:657– 670 11. Tsai PS, Werner S, Weiner RI 1995 Basic fibroblast growth factor is a neurotropic factor in GT1 gonadotropin-releasing hormone neuronal cell lines. Endocrinology 136:3831–3838 12. Ma YJ, Berg-von der Emde K, Moholt-Siebert M, Hill DF, Ojeda SR 1994

28.

29. 30.

31.

32. 33. 34. 35.

36.

37. 38.

25

Region-specific regulation of transforming growth factor a (TGFa) gene expression in astrocytes of the neuroendocrine brain. J Neurosci 14:5644 –5651 Ma YJ, Hill DF, Junier M, Costa ME, Felder SE, Ojeda SR 1994 Expression of epidermal growth factor receptor changes in the hypothalamus during the onset of female puberty. Mol Cell Neurosci 5:246 –262 Berg-von der Emde K, Ma YJ, Costa ME, Ojeda SR 1993 Hypothalamic astrocytes respond to transforming growth factor alpha (TGFa) with prostaglandin E2 (PGE2) release. Soc Neurosci Abstr 19:1682 Ma YJ, Berg-von der Emde K, Hill DF, Costa ME, Ojeda SR, Hypothalamic astrocytes respond to transforming growth factor alpha (TGFa) with secretion of neuroactive substances that stimulate the release of luteinizing hormonereleasing hormone (LHRH). 10th International Congress of Endocrinology, 1996, p 220 McCarthy KD, de Vellis J 1980 Preparation of separate astroglial and oligodendroglial cell cultures from rat cerebral tissue. J Cell Biol 85:890 –902 Yaish P, Gazit A, Gilon C, Levitzki A 1988 Blocking of EGF-dependent cell proliferation by EGF receptor kinase inhibitors. Science 242:933–935 Carpenter G, Cohen S 1979 Epidermal growth factor. Annu Rev Biochem 48:193–216 Massague J 1990 Transforming growth factor-a. A model for membraneanchored growth factors. J Biol Chem 265:21393–21396 Ojeda SR, Urbanski HF 1988 Intracellular regulatory mechanisms of LHRH secretion and the onset of female puberty. In: Imura H (ed) Neuroendocrine Control of the Hypothalamic-Pituitary System. Japan Scientific Press, Tokyo, pp 49 – 64 Campbell WB, Ojeda SR 1987 Measurement of prostaglandins by radioimmunoassay. Methods Enzymol 141:323–350 Ojeda SR, Urbanski HF, Katz KH, Costa ME 1986 The activation of estradiol positive feedback at puberty: estradiol sensitizes the LHRH releasing system at two different biochemical steps. Neuroendocrinology 43:259 –265 Samuelsson B 1982 The leukotrienes: an introduction. In: Samuelsson B, Paoletti R (eds) Leukotrienes and Other Lipoxygenase Products. Raven Press, New York, pp 1–17 Silverman A-J, Livne I, Witkin JW 1994 The gonadotropin-releasing hormone (GnRH), neuronal systems: immunocytochemistry and in situ hybridization. In: Knobil E, Neill JD (eds) The Physiology of Reproduction, ed 2. Raven Press, New York, vol 1:1683–1709 Witkin JW, Ferin M, Popilskis SJ, Silverman A 1991 Effects of gonadal steroids on the ultrastructure of GnRH neurons in the rhesus monkey: synaptic input and glial apposition. Endocrinology 129:1083–1092 Witkin JW, Silverman A 1985 Synaptology of luteinizing hormone-releasing hormone neurons in rat preoptic area. Peptides 6:263–271 Chen W-P, Witkin JW, Silverman AJ 1990 Sexual dimorphism in the synaptic input to gonadotropin releasing hormone neurons. Endocrinology 126:695–702 King JC, Letourneau RL 1994 Luteinizing hormone-releasing hormone terminals in the median eminence of rats undergo dramatic changes after gonadectomy, as revealed by electron microscopic image analysis. Endocrinology 134:1340 –1351 Ojeda SR, The neurobiology of mammalian puberty: Studies in rodents and primates. 10th International Congress of Endocrinology, 1996, p 701 (Abstract) Junier M, Ma YJ, Costa ME, Hoffman G, Hill DF, Ojeda SR 1991 Transforming growth factor alpha contributes to the mechanism by which hypothalamic injury induces precocious puberty. Proc Natl Acad Sci USA 88:9743–9747 Zhen S, Zakaria M, Su E, Radovick S, The role of immediate-early genes in gondotropin-releasing hormone (GNRH) neurons on the control of the onset of puberty. 10th International Congress of Endocrinology, 1996, p 221 (Abstract) Peppelenbosch MP, Tertoolen LGJ, Hage WJ, de Laat SW 1993 Epidermal growth factor-induced actin remodeling is regulated by 5-lipoxygenase and cyclooxygenase products. Cell 74:565–575 Peppelenbosch MP, Tertoolen LGJ, den Hertog J, de Laat SW 1992 Epidermal growth factor activates calcium channels by phospholipase A2/5-lipoxygenase-mediated leukotriene C4 production. Cell 69:295–303 Hiney JK, Ojeda SR, Dees WL 1991 Insulin-like growth factor-I as a metabolic signal involved in the regulation of female puberty. Neuroendocrinology 54:420 – 423 Olson BR, Scott DC, Wetsel WC, Elliot SJ, Tomic M, Stojilkovic S, Nieman LK, Wray S 1995 Effects of insulin-like growth factors I and II and insulin on the immortalized hypothalamic GT1–7 cell line. Neuroendocrinology 62:155–165 Coleman RA, Kennedy I, Humphrey PPA, Bunce K, Lumley P 1990 Prostanoids and their receptors. In: Hansch C, Sammes PG, Taylor JB, Emmett JC (eds) Comprehensive Medicinal Chemistry. Pergamon Press, New York, vol 3:643–714 Rage F, Ma YJ, Ojeda SR 1994 Different prostaglandin E2 (PGE2) receptor subtypes are expressed in immortalized luteinizing hormone releasing hormone (LHRH) neurons. Soc Neurosci Abstr 20:637 Wetsel WC, Daniel K, Collins S, Stimulation of LHRH secretion by activation of the EP3 subtype of the prostaglandin E2 receptor. 76th Annual Meeting of The Endocrine Society, Anaheim CA, 1994, p 376 (Abstract)