| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Centre for Neuroendocrinology and Department of Anatomy and Structural Biology, University of Otago School of Medical Sciences, Dunedin 9054, New Zealand
Address all correspondence and requests for reprints to: Dr. Greg Anderson, Centre for Neuroendocrinology and Department of Anatomy and Structural Biology, University of Otago School of Medical Sciences, PO Box 913, Dunedin 9054, New Zealand. E-mail: greg.anderson{at}anatomy.otago.ac.nz.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Some (13, 14, 15), but not all (16, 17), experiments where pulsatile LH secretion was measured in animal models have shown that experimental hyperprolactinemia reduces pulse frequency and/or amplitude, whereas others have demonstrated that such treatments can inhibit GnRH release into the portal blood (18, 19). Although it is clear from these experiments that the GnRH system can be inhibited by prolactin either directly or via prolactin-sensitive afferent inputs, the interactions with estradiol have not been specifically investigated. Because estradiol is a key regulator of GnRH and LH secretion (20), and prolactin receptor expression and signaling is estrogen sensitive in some tissues (see e.g. Refs. 21 and 22), we hypothesized that prolactin and estradiol might interact to influence reproduction. The aim was to test this hypothesis by comparing the effects of chronic hyperprolactinemia, induced by treatment with the dopamine D2 antagonist and antipsychotic drug sulpiride, on LH pulse frequency in the presence and absence of physiological circulating concentrations of estradiol. Given that hyperprolactinemia may cause a diminished LH surge in women (23), we also tested whether sulpiride-induced hyperprolactinemia could inhibit the magnitude of the LH surge and the proportion of GnRH neurons activated during this event.
Prolactin receptors are expressed in specific hypothalamic nuclei and other brain regions. The medial septum (MS) and preoptic area (POA), where GnRH soma and much of their afferent input populations are located (24), express protein (25, 26) and mRNA (27, 28) for the long form of the prolactin receptor (PRL-R), the predominant form in the hypothalamus (29, 30). Modulation of prolactin’s effects in this and other brain regions is likely to be achieved in part through regulation of the level of PRL-R expression. PRL-R expression in various tissues is known to be positively influenced by estradiol (21, 31, 32). More limited evidence suggests that PRL-R is also regulated by prolactin itself in peripheral tissues (31, 33, 34, 35, 36). Thus, the actions of prolactin in the brain might be influenced by prevailing levels of these hormones via PRL-R regulation. For example, chronic hyperprolactinemia or high circulating estradiol levels might induce PRL-R within components of the GnRH system in the POA, thereby increasing the sensitivity of this system to the effects of prolactin. In the present study, we tested whether estrogen and prolactin interact to regulate prolactin receptor mRNA levels in the medial POA as well as two other brain sites: the arcuate nucleus (ARC, the location of the tuberoinfundibular dopaminergic neurons that control the negative feedback regulation of prolactin secretion) and the choroid plexus (ChP, a site containing a high abundance of prolactin receptors and where prolactin may gain entry into the brain) (37). To assess the effect of estrogen and prolactin on components of the PRL-R signaling pathway, we also measured the effects of these hormones on phosphorylation of the primary signal transducer for the PRL-R in the brain, signal transducer and activator of transcription 5 (STAT5), and on levels of mRNA for suppressors of cytokine signaling (SOCS). The SOCS family of proteins is a relatively recently described class of cytokine-inducible inhibitors of cytokine signaling that act as feedback inhibitors for a range of cytokines that use Janus kinase (JAK)/STAT pathways, including prolactin. In vitro studies of the action of these proteins indicate that they act as part of an intracellular feedback loop to inhibit the STAT phosphorylation (38), thereby moderating signal transduction. In vivo, prolactin has been shown to induce mRNA for SOCS-1, and -3 in ovary, adrenal, and mammary glands (39) as well as in the POA and ARC (40, 41), and their levels therefore reflect PRL-R signaling (42).
The aims of the current experiments, therefore, were to examine the interactive effects of prolactin and estradiol on tonic GnRH/LH pulsatile secretion and on the GnRH/LH surge. To test the hypothesis that estradiol or prolactin might alter the hypothalamic response to prolactin, we have also examined effects of hyperprolactinemia and estradiol on PRL-R mRNA levels and on two indices of prolactin responses in neurons: phosphorylation of STAT5 and induction of SOCS mRNA within discrete brain regions.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Experiment 1a
To examine the effects of sulpiride-induced hyperprolactinemia and estradiol on GnRH/LH pulsatility, rats were ovariectomized on d 0 of the experiment to remove most circulating endogenous sex steroids and the confounding effects of estrous cyclicity. Rats were randomly assigned to the following groups (n = 6): treatment with silicone rubber capsules (inner diameter 1.57 mm, outer diameter 3.18 mm, length 30 mm) containing 20 µg/ml estradiol (Sigma Chemical Co., St. Louis, MO) in sesame oil at the time of ovariectomy, treatment with twice-daily (0900 and 1700 h) sc injections of the D2 receptor antagonist sulpiride (1.25 mg in 0.25 ml 0.1 M tartaric acid) for the duration of the experiment to induce hyperprolactinemia, treatment with estradiol implants plus sulpiride injections, or no further treatment. Animals not injected with sulpiride received vehicle-only injections. Sulpiride and its dosage were chosen based on its specificity for the D2 receptor and absence of behavioral side effects (43), and the dose of estradiol was selected to produce low circulating concentrations that are similar to those seen in diestrous rats (unpublished observations). On d 7 of the experiment, all animals were fitted with an atrial cannula chronically implanted via the right jugular vein to facilitate repeated blood sampling. Beginning at 1000 h on d 9 of the experiment, a blood sample (0.4 ml) was drawn every 10 min for 3 h to enable measurement of pulsatile secretion of LH. Plasma was harvested and stored at –20 C for later RIA. Red blood cells were resuspended in sterile physiological saline and replaced into the animal after the subsequent sample. Additional samples were also collected at various times throughout the preceding day to enable a 24-h profile of circulating prolactin concentration to be plotted in response to the injections.
Experiment 1b
To examine the effects of sulpiride-induced hyperprolactinemia on the estradiol-induced GnRH/LH surge, the two non-estrogen-treated groups of rats from experiment 1a were subsequently subjected to an exogenous estrogen-induced GnRH/LH surge protocol. An additional group of six rats was ovariectomized at the same time as the other groups but received no further treatments; this group was used to provide basal LH concentration and GnRH neuron activity in the absence of surge induction. To induce the surge, rats had 200 µg/ml estradiol-filled capsules inserted sc immediately after blood sampling for LH pulse measurement in experiment 1a was completed. At 1200 h on d 11, they received an injection of 1.5 mg progesterone in 0.3 ml oil sc. The rats were perfused at the time of the expected peak of the LH surge (between 1600 and 1700 h) on d 11. A blood sample was collected by heart puncture at the time of perfusion, and the plasma was harvested and stored at –20 C for LH RIA.
Experiment 2
To test whether chronic estrogen exposure increases cellular responses to prolactin, four groups of rats (n = 4) were ovariectomized and chronically fitted with 22-gauge intracerebroventricular (icv) cannulas on d 0 to permit ovine prolactin injection into a lateral ventricle. The animals were treated with 20 µg/ml estradiol implants as described in experiment 1a and/or an acute icv prolactin injection before perfusion, such that different groups received estrogen only, estrogen plus prolactin, prolactin only, or neither hormone. Rats not treated with prolactin received an icv vehicle injection. All rats were injected with bromocriptine methanesulfonate (200 µg in 200 µl 10% ethanol; Research Biochemicals, Inc., Natick, MA) at 0900 h on d 9 of the experiment to reduce endogenous circulating prolactin levels to basal levels. Approximately 4 h later, the rats received a single injection of ovine prolactin (4 µg icv in 4 µl saline; Sigma) or vehicle and were perfused 30 min later for brain collection and phosphorylated STAT5 (pSTAT5) immunohistochemistry (40).
Experiment 3
To examine the effects of sulpiride-induced hyperprolactinemia and estradiol on PRL-R and SOCS mRNA levels in discrete brain regions, four groups of rats (n = 6) were ovariectomized on d 0 and treated for the duration of the experiment with 20 µg/ml estradiol implants and/or sulpiride injections, as for experiment 1a. On d 9 of the experiment, all rats were killed for brain microdissection by decapitation. The brains were rapidly removed, frozen on dry ice, and stored at –80 C for later microdissection of the POA, ARC, and ChP and real-time quantitative PCR measurement of PRL-R and SOCS mRNA.
Prolactin and LH RIA
Serum prolactin and LH concentration was measured in 10- and 50-µl sample volumes, respectively, by RIA. Values are expressed in terms of the rat standards NIDDK-rat PRL-RP-3 and NIDDK-rat LH-RP-3. Iodinated hormone (NIDDK-rat PRL-I-6 and NIDDK-rat LH-I-10) were used as tracers and primary antisera were NIDDK rabbit antirat PRL-RIA-9 (final dilution 1:100,000) and NIDDK-rabbit antirat LH-S11 (final dilution 1:500,000). The sensitivity of the prolactin and LH assays (95% confidence interval at 0 ng/ml on the standard curve) were 4 and 0.4 ng/ml, respectively. The intraassay coefficients of variation (CV) for serum pools falling in the middle of the standard curves were 15 and 14% in the prolactin and LH assays, and the interassay CV was 15% for the same LH sample. All prolactin values were obtained from a single assay.
Perfusion and measurement of GnRH, Fos, and pSTAT5 immunoreactivity
Animals were deeply anesthetized with 60 mg sodium pentobarbital (containing 1000 IU heparin) and perfused via the ascending aorta with 30 ml physiological saline (containing 120 IU heparin) and then 250 ml 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Brains were removed and postfixed in the same fixative overnight and then infiltrated with 30% sucrose for cryoprotection until they sank. Coronal (40 µm thick) sections containing the MS and the POA (experiment 1b) or the entire hypothalamus (experiment 2) were cut on a sliding microtome with a freezing stage from each brain to provide five similar series of consecutive sections (200 µm apart); these were stored in cryoprotectant at –20 C until immunohistochemistry. All immunohistochemistry steps were separated by three washes in 0.1 M phosphate buffer. Sections were dual-labeled using peroxidase immunohistochemistry. For pSTAT5 immunohistochemistry, an initial high temperature plus high pH antigen retrieval step (5 min in 0.1 M Tris HCl, pH 10, at 85 C) was required to unmask the antigen. After a 30-min blocking incubation in 5% normal goat serum and 1% hydrogen peroxide to quench endogenous peroxides, sections were incubated for 48 h in a polyclonal rabbit anti-Fos antibody (1:2500 rabbit anti-Fos, AB5; Oncogene Research Products, San Diego, CA) or polyclonal rabbit anti-pSTAT5 (tyr 694, 1:400; Cell Signaling Technology, Beverly, MA) containing 5% normal goat serum. The Fos antibody does not cross-react with Fos-related antigens or FosB, and the pSTAT5 antibody does not distinguish between pSTAT5a or pSTAT5b. This was followed by a 1-h incubation in biotinylated goat antirabbit IgG antibody (1:250; Chemicon, Temecula, CA), 1 h of incubation in Vector Elite ABC solution (Vector Laboratories, Burlingame, CA), and a 5-min incubation in a nickel-enhanced diaminobenzidine and hydrogen peroxide solution to visualize Fos immunoreactivity (blue-black nuclear staining). After this, sections for GnRH plus Fos immunohistochemistry were incubated for 48 h in a polyclonal rabbit anti-GnRH antibody (1:3000 rabbit anti-GnRH; SW1 kindly donated by Susan Wray) containing 5% normal goat serum and then biotinylated goat antirabbit IgG antibody and ABC solution as before. GnRH immunoreactivity was visualized by a brief incubation in an unenhanced diaminobenzidine and hydrogen peroxide solution until sufficient staining, without high background levels, was observed. Sections were mounted on (3-aminopropyl)-tirethoxy-silane-coated slides, dehydrated in graded alcohols, and coverslipped. Omission of GnRH, Fos, and pSTAT5 primary antibodies resulted in a complete absence of staining. GnRH neurons and Fos coexpression were counted by an operator blind to the treatment groups, using a light microscope at x200 magnification. All GnRH neurons from three sections per rat, which included the MS and region around the organum vasculosum of the lamina terminalis (OVLT), were counted. Neurons were counted as Fos positive if they had a distinct blue-black stained nucleus that was darker than the surrounding brown cytoplasmic staining. For each animal, the percentage of Fos-colocalized GnRH neurons was calculated to provide a single data point. For pSTAT5 immunoreactivity, numbers of immunoreactive cells were counted within the MS, OVLT, anteroventral periventricular nucleus (AVPV), periventricular nucleus (PeV), paraventricular nucleus (PVN), and ARC, using the image analysis software Image J (National Institutes of Health, Bethesda, MD) to identify and count darkly stained nuclei within standard-sized regions and the Paxinos and Watson brain atlas to define these regions. Counts are presented as total numbers of identified cells per region per section. For each region, one to three sections were counted and averaged to provide a single data point per region per animal.
Microdissection of the POA, ARC, and ChP and preparation of cDNA
Thick coronal brain sections (300 µm) were cut in a cryostat at –9 C, thaw-mounted onto glass slides, and refrozen. The POA was dissected from three consecutive sections (between approximately 0.0 and –0.9 mm relative to bregma) with four punches per section of a sterile 21-gauge micropunch needle (500 µm internal diameter), so that a block of tissue approximately 1 mm square was removed immediately dorsal to the optic chiasm. The ARC was dissected from five consecutive sections (between approximately –2.2 and –3.7 mm relative to bregma) with a single midline punch per section, centered around the ventral extent of the third ventricle. The ChP was dissected from three consecutive sections (between approximately 0.0 and –0.9 mm relative to bregma) with a single punch from each of the two lateral ventricles per section. Micropunched tissue was placed in 50 µl of a commercial lysis buffer (Cells-to-cDNA II Cell Lysis Buffer; Ambion, Austin, TX), sonicated in an ultrasonic cell disrupter, and then incubated at 75 C for 10 min to rupture the cells and inactivate endogenous RNases. The tissue lysate was then stored at –80 C until RT in a 20-µl volume (containing 56 ± 2 ng total RNA) as previously described (40, 41). Each run included a control tube with reverse transcriptase omitted, to demonstrate that the template for the PCR product was not genomic DNA.
Real-time quantitative PCR measurement of PRL-R, SOCS-1 and -3, and CIS mRNA
TaqMan probes and forward (sense) and reverse (antisense) primers to the rat genes encoding PRL-R, SOCS-1 and -3, and cytokine-inducible SH2-containing protein (CIS) (see Ref. 41 for details) were designed using Primer Express software (Applied Biosystems, Foster City, CA). Real-time PCR was conducted as described previously in a 22.5-µl reaction mix containing ABsolute qPCR Rox master mix (ABgene, Epsom, Surrey, UK), template cDNA (2.5 µl), and primers and probes at optimized final concentrations (40, 41). A stock of cDNA containing relatively high levels of PRL-R and SOCS mRNA was created by pooling multiple liver punches, and 4-fold dilution series of this were run on each plate as external standards. Standard curves of dilution factor vs. the cycle number at which fluorescence first exceeded a given threshold (CT) were linear (r2 < 0.98); these curves were used to calculate relative levels of PRL-R and SOCS mRNA from the sample CT values. These data were then normalized to total RNA content. Total RNA in the remaining tissue lysate was measured in 96-well plates using a Quant-iT RNA assay kit (Molecular Probes, Eugene, OR). The prolactin receptor and SOCS mRNA levels were then expressed as arbitrary (relative) units per nanogram total RNA. CV for the sample duplicates were almost always less than 1%.
Statistical analysis
The method for identifying LH pulses followed a modified version of that used by Goodman and Karsch (44). Basic criteria for identifying a pulse were 1) peaks of pulses must exceed both the preceding and following nadir concentrations by 2 SD of the peak value and by the sensitivity of the assay, and 2) a pulse must peak within two samples of the preceding nadir. To determine LH pulse amplitude, the difference between peak and nadir values for each identified pulse per animal was averaged. All significant treatment effects were identified using two-way ANOVA, with repeated-measures analysis for prolactin concentration. Where data failed equal variance or normality tests, they were log-transformed (base 10). This was followed by the Bonferroni t test for post hoc analysis to determine where significant effects occurred. Results are presented as mean ± SEM, and differences were considered significant at P < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
).
|
. Estradiol caused LH pulse frequency to decrease and LH pulse amplitude to increase compared with nonimplanted rats (main effect, P < 0.05). Although sulpiride-induced hyperprolactinemia was without effect on these parameters in the absence of estradiol, in the presence of estradiol implants, LH pulse frequency was reduced to 60% of that in vehicle-treated controls (Fig. 2, C and E; P < 0.05). There was no significant effect of sulpiride on LH pulse amplitude (Fig. 2F).
|
, A and B. In rats in which a GnRH/LH surge was induced, the plasma LH concentration averaged 29 ng/ml at the time of the peak of the surge, and 50% of GnRH neurons expressed Fos-positive nuclear staining (both P < 0.05 vs. rats that did not undergo the surge induction protocol) (Fig. 3C). There were no significant effects of sulpiride on either of these measures of surge intensity, although there was a tendency toward a reduced GnRH plus Fos colocalization in hyperprolactinemic rats (P = 0.1 vs. vehicle-injected controls).
|
. In rats receiving vehicle injections icv, no pSTAT5 immunoreactivity was detected regardless of estrogen treatment. Prolactin markedly induced pSTAT5 in the OVLT, PeV, PVN, ARC (main effect, P < 0.001), and AVPV (P < 0.01), but in the MS, staining tended to be fainter and the effect of prolactin was not significant (P > 0.05). In the AVPV, PeV, PVN, and ARC, but not the MS and OVLT, treatment with estradiol implants caused a marked (5- to 50-fold) increase in the level of pSTAT5 immunoreactivity induced by prolactin (P < 0.001) (Fig. 4 B). Where possible, we also evaluated pSTAT5 expression in the ChP. There was some evidence of induction of pSTAT5 in the ChP by prolactin plus estrogen treatment (see Fig. 4A), but because this tissue did not survive the antigen retrieval and free-floating immunohistochemistry procedures in most animals, the effect was not able to be quantified.
|
) as well as in the ChP (P < 0.001) and ARC (P < 0.05, SOCS-3 and CIS only) (Fig. 6). Sulpiride-induced hyperprolactinemia caused a smaller induction of PRL-R, SOCS-3, and CIS mRNAs in the ARC and ChP (Fig. 6) and of SOCS-1 in all three areas (Figs. 5 and 6) (main effects, P < 0.01). The effects of sulpiride on SOCS-3 and CIS mRNA levels in the ARC and ChP were greatly intensified by the presence of estradiol (Fig. 6, E–H) (interaction of estradiol and sulpiride, P < 0.001), but this effect was not significant in the POA (P = 0.2).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We used twice-daily injections of the dopamine D2 antagonist sulpiride because it models antipsychotic drug-induced hyperprolactinemia, because it requires fewer repeated injections than does treatment with exogenous prolactin itself and because the level of hyperprolactinemia can be controlled more easily than with techniques such as ectopic transplantation of pituitary or prolactin-secreting tumor cells. Using this pharmacological approach, the rats were hyperprolactinemic for at least 75% of each day (Fig. 1
). Because dopamine antagonists appear to have no effect on LH pulse frequency in chronically ovariectomized rats (45, 46), it seems unlikely that the effects of hyperprolactinemia on this parameter in our study were confounded by altered dopaminergic interactions with GnRH neurons. However, dopamine may inhibit the preovulatory LH surge (47) and dopamine antagonists can augment the surge magnitude (48), so we cannot rule out the possibility that inhibitory effects of hyperprolactinemia and stimulatory effects of central dopamine blockade cancelled each other out with respect to our measurements of the intensity of the GnRH/LH surge.
Suppression of LH pulse frequency by sulpiride-induced hyperprolactinemia provides good evidence for centrally mediated effects of prolactin on the reproductive system, because pulsatile LH secretion is a robust index of GnRH pulse frequency. This does not preclude the importance of actions of prolactin directly on the pituitary gonadotrophs, which express prolactin receptors (7) and respond to prolactin treatments by releasing less LH in response to exogenous GnRH (5, 6). However, because we observed no effect of hyperprolactinemia on LH pulse amplitude, the current study provides no support for a major effect at the pituitary level. A central mechanism of prolactin action is supported by the fact that prolactin causes suppression of GnRH mRNA levels (49) and GnRH release into portal blood (18, 19). In a number of human (8, 9) and animal studies (13, 14, 15), pathological or experimentally induced hyperprolactinemia has been shown to cause reduced LH pulse frequency and/or amplitude, whereas in other studies, no effect of hyperprolactinemia was observed on either parameter (16, 17). The varying results from these studies are likely due to the differing methods used to induce experimental hyperprolactinemia, the circulating and central prolactin concentration achieved by these treatments, and the different sexes and endocrine conditions of the animal models employed. To our knowledge, experiment 1a of the current study is the first to specifically test the effects of estrogen on hyperprolactinemia-induced LH pulse suppression. Our data provide evidence that the presence of estrogen is a prerequisite for suppression of LH pulse frequency by prolactin in the female rat. This estrogen effect appears to be facilitative rather than additive, because the low estradiol dose we used in this study did not exert an independent effect on LH pulse frequency. The efficacy of such low levels of estradiol in allowing prolactin to suppress GnRH secretion and fertility might explain how this effect is maintained long term in hyperprolactinemic subjects even after gonadal steroid secretion declines.
The pathways by which prolactin affects GnRH activity are poorly understood. Prolactin is known to influence a range of neurotransmitter systems that in turn can modulate GnRH neuronal activity, including β-endorphin (50), 
-aminobutyric acid (51), and dopamine and neuropeptide Y (52). Alternatively to this indirect mode of action, prolactin could act directly on GnRH neurons, as is suggested by the presence of PRL-R and inhibitory effects of prolactin on GnRH neuron-derived immortalized GT1 cells (53). We have recently provided evidence for potential direct actions of prolactin on GnRH neurons (12). Using single-cell nested PCR to measure mRNA for prolactin receptors within identified GnRH neurons, we showed that a relatively small subset (approximately 13%) of GnRH neurons expressed this receptor message. In agreement with this, we also demonstrated using in situ hybridization that a small proportion (5%) of GnRH neurons expressed PRL-R mRNA above background levels (unpublished data). In experiment 2, only weak prolactin-induced pSTAT5 was seen in the MS and the region surrounding the OVLT, where GnRH perikarya are primarily located. It remains to be tested whether GnRH neurons colocalize with prolactin-induced pSTAT5. We did, however, observe strong pSTAT5 expression in the AVPV and PeV, regions that contain numerous neurons known to directly regulate GnRH neurons (24). We have also observed strong PRL-R mRNA expression in this area using in situ hybridization (unpublished). These data, together with the lack of prolactin effects on PRL-R, SOCS-3, and CIS in the POA, allow for some possibility of direct prolactin actions of GnRH soma but also suggest that indirect inputs via prolactin-sensitive afferents in the AVPV and other regions are likely to occur. Such multimodal hormonal regulation of GnRH neurons is well documented for estrogen (20) and may be a common neuroendocrine phenomenon.
In experiment 1b of this study, we observed no effect of sulpiride-induced hyperprolactinemia on the magnitude of an exogenous steroid-induced LH surge. Although there was a tendency toward a reduction in the number of GnRH neurons activated in hyperprolactinemic rats, a small reduction in GnRH surge magnitude would be unlikely to be translated into an effect on LH surge magnitude because the GnRH surge appears to be far in excess of that required for a maximal LH surge (54). It is important to note that these findings do not preclude an effect of chronic hyperprolactinemia on the endogenously generated preovulatory surge; indeed, the steroid-dependent inhibition of LH pulse frequency we observed would be expected to prevent the preovulatory buildup of estradiol that is critical for surge occurrence (1). This is supported by other studies that showed a negligible effect of hyperprolactinemia on exogenous estradiol-induced LH surge levels (55) but a marked effect on the endogenously generated LH surge (56).
Given that estradiol is known to induce PRL-R mRNA and protein expression in the brain and liver (21, 31, 32, 57, 58), we then examined whether the requirement of estrogen for the suppression of LH pulse frequency by prolactin might be due to an estrogenic action within the POA on PRL-R mRNA expression or on responsiveness to prolactin. To address the latter endpoint, we measured both pSTAT5 and SOCS mRNA levels, because these are known to reflect PRL-R signaling (40, 42). PRL-R mRNA in the POA was potently stimulated by estrogen, raising the possibility that increased responsiveness of the GnRH system at the receptor level to prolactin is the mechanism underlying the steroid-dependent hyperprolactinemia-induced LH pulse frequency reduction shown in experiment 1a. PRL-R mRNA was also markedly induced by sulpiride-induced hyperprolactinemia in the present study. The evidence for induction of PRL-R by prolactin itself is much more limited than for estrogenic induction (33, 59, 60). We show here that the magnitude of this effect of prolactin was much greater in the ChP than in the ARC, consistent with the dramatic increase in PRL-R seen predominantly in the ChP during pregnancy and lactation (29), and that hyperprolactinemia was able to induce only SOCS-1 mRNA in the POA.
Marked effects of sulpiride-induced hyperprolactinemia were observed on the expression of multiple forms of SOCS mRNAs. The levels of these Janus kinase (JAK)/STAT signaling components is known to reflect the degree of PRL-R signaling (42). More recent data suggest that interactions of estradiol signaling with specific SOCS members also exist in cells (61, 62, 63). In the current study, whereas SOCS-1 mRNA was induced only by hyperprolactinemia, SOCS-3 and CIS mRNAs were also greatly induced by estradiol, and prolactin and estradiol were able to induce these two forms in an additive manner in the ARC and ChP (Fig. 6). Because estradiol potently stimulates prolactin release from the lactotrophs (64) (as reflected by the higher circulating prolactin concentration in estrogen-implanted vs. nonimplanted rats not injected with sulpiride), the ability to ascribe a given effect to either of these hormones or to their combined action is often confounded in this type of experiment. The effects of estrogen on SOCS mRNA expression in this study, however, were probably direct (rather than occurring via estrogen-induced prolactin), because the increased prolactin concentration in response to estradiol was relatively small compared with that induced by sulpiride, yet the effect of estradiol on SOCS mRNA was much larger. To our knowledge, these are the first results to show that estradiol can induce SOCS mRNA in the brain or to suggest that prolactin and estrogen can interact in central activation of STAT5 and SOCS induction. In vitro studies, however, have recently begun to shed light on mechanisms of steroid hormone and cytokine signaling interactions. Prolactin- and GH-induced STAT5 transcriptional activity is potently repressed by interaction of estrogen receptor-
and -β with STAT5 (22, 61, 65, 66). Although estrogen is thought to act by associating intracellularly with cytokine receptors or STAT molecules, estrogen was recently also shown to act via estrogen receptor- to induce SOCS-3 promoter activity in breast cancer cells (62) and SOCS-2 and -3 in liver (63). Estrogen is unable to inhibit activation of STAT5 by GH in fibroblasts from Socs2 knockout mice (61). Although our results are generally consistent with these in vitro data with regard to estrogenic SOCS induction, they differ markedly with regard to estrogen effects on prolactin-induced STAT5 activation. This difference may reflect tissue-specific actions or relate to the duration of estrogen exposure, because in our study, estradiol treatment lasted for 9 d, long enough to induce a robust increase in PRL-R expression. It would therefore be interesting to determine whether estrogen can induce SOCS expression and facilitate prolactin-induced STAT5 activation acutely (i.e. independently of PRL-R up-regulation).
There appears to be some tissue specificity in which SOCS genes are activated by estrogen; whereas there is no effect on SOCS-1 in brain (present results), kidney (61), or liver (63), SOCS-3 is stimulated in the liver and brain but not kidney. Only in the present study has CIS been shown to respond to estrogen, although to our knowledge, this has not yet been investigated in the kidney. Whether estrogen can stimulate SOCS-2 in the brain is unknown; we did not examine this in the current study because we had previously determined that SOCS-2 in the ARC is unaffected by prolactin (40). A functional estrogen-responsive element is present in the human (62) and murine (63) SOCS-3 promoter, and the fact that estrogen and GH had additive effects on the activity of this promoter in the latter study supports the idea that the two hormones activate it independently. A similar additive effect of estrogen and prolactin was evident on SOCS-3 and CIS mRNA levels in the current study, at least in the ARC and ChP.
Although enhanced STAT5-mediated prolactin signaling would be expected to increase the antifertility effects of prolactin, the consequences of central SOCS up-regulation for fertility regulation are unclear at present. Induction of SOCS by chronic hyperprolactinemia in regions controlling reproductive activity might serve as a mechanism that limits the infertility effects of this condition by preventing STAT5 overactivation; however, the ability of estrogen to induce the same effect appears contradictory to the fact that this hormone was required for suppression of LH pulse frequency by prolactin in experiment 1a. Central SOCS up-regulation would be expected to alter the reproductive effects of a range of cytokines, both inhibitory (e.g. prolactin) and stimulatory (e.g. leptin and insulin), with the net outcome being determined by the balance of these effects. Loss of SOCS function studies will be required to determine the role of these proteins in fertility regulation. Interestingly, any analysis of the overall consequences of SOCS induction by estrogen needs to take into account the fact that progesterone may exert the opposite effect (22, 67), allowing for complex endocrine interactions determined by the balance of the circulating steroidal milieu. It should be noted that in this study, we have used SOCS mRNA as a marker of prolactin action, and the functional consequences are not clear. Up-regulation of SOCS proteins might simply be a normal component of the prolactin signal transduction pathway, regulating levels of STAT5 phosphorylation and terminating acute signaling events, to ensure that the cells remain sensitive to subsequent prolactin signaling.
In conclusion, from these data, we show that the inhibition of GnRH pulse frequency by hyperprolactinemia is steroid dependent, possibly as a consequence of estrogen-induced PRL-R up-regulation in the POA. We also show that estradiol induces SOCS-3 and CIS in multiple brain regions, raising the possibility that a range of prolactin-regulated brain functions are also affected by cross-talk with this steroid. The implications of such altered responsiveness on central prolactin-mediated functions in addition to fertility regulation (such as modulation of maternal and anxiety-related behaviors, food intake, oxytocin and vasopressin secretion, and regulation of prolactin section and transport into the brain) (37) deserve further investigation.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online December 27, 2007
Abbreviations: ARC, Arcuate nucleus; AVPV, anteroventral periventricular nucleus; ChP, choroid plexus; CIS, cytokine-inducible SH2-containing protein; CV, coefficients of variation; icv, intracerebroventricular; MS, medial septum; OVLT, organum vasculosum of the lamina terminalis; PeV, periventricular nucleus; POA, preoptic area; PRL-R, long form of the prolactin receptor; pSTAT5, phosphorylated STAT5; PVN, paraventricular nucleus; SOCS, suppressor of cytokine signaling; STAT5, signal transducer and activator of transcription 5.
Received June 27, 2007.
Accepted for publication December 14, 2007.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
in mammary cells: a role in differentiated secretory epithelium. Regul Pept 121:163–170[CrossRef][Medline] regulates SOCS-3 expression in human breast cancer cells. Biochem Biophys Res Commun 335:168–174[CrossRef][Medline] and signal transducers and activators of transcription 5a. Endocrinology 145:1739–1749This article has been cited by other articles:
 |
|
 |
|
  G. M. Anderson, H.-L. Relf, M. Z. Rizwan, and J. J. Evans Central and Peripheral Effects of RFamide-Related Peptide-3 on Luteinizing Hormone and Prolactin Secretion in Rats Endocrinology, April 1, 2009; 150(4): 1834 - 1840. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Z. Rizwan, R. Porteous, A. E. Herbison, and G. M. Anderson Cells Expressing RFamide-Related Peptide-1/3, the Mammalian Gonadotropin-Inhibitory Hormone Orthologs, Are Not Hypophysiotropic Neuroendocrine Neurons in the Rat Endocrinology, March 1, 2009; 150(3): 1413 - 1420. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
) or vehicle (
). Injection times are denoted by arrows. Concentrations in the absence of estradiol implants were similar except that the diurnal surges in vehicle-injected rats were absent (data not shown). *, P < 0.001.
