Suppression of Prolactin-Induced Signal Transducer and Activator of ...

0013-7227/06/$15.00/0 Printed in U.S.A.

Endocrinology 147(10):4996 –5005 Copyright © 2006 by The Endocrine Society doi: 10.1210/en.2005-0755

Suppression of Prolactin-Induced Signal Transducer and Activator of Transcription 5b Signaling and Induction of Suppressors of Cytokine Signaling Messenger Ribonucleic Acid in the Hypothalamic Arcuate Nucleus of the Rat during Late Pregnancy and Lactation Greg M. Anderson, Paulien Beijer, Angela S. Bang, Mark A. Fenwick, Stephen J. Bunn, and David R. Grattan Centre for Neuroendocrinology and Department of Anatomy and Structural Biology, University of Otago School of Medical Sciences, Dunedin 9054, New Zealand During late pregnancy and lactation, the tuberoinfundibular dopamine (TIDA) neurons that regulate prolactin secretion by negative feedback become less able to produce dopamine in response to prolactin, leading to hyperprolactinemia. Because prolactin-induced activation of dopamine synthesis in these neurons requires the Janus kinase/signal transducer and activator of transcription 5b (STAT5b) signaling pathway, we investigated whether prolactin-induced STAT5b signaling is reduced during lactation and whether induction of suppressors of cytokine signaling (SOCS) mRNAs occur at this time and in late pregnancy. During lactation, the ability of exogenous prolactin to induce STAT5 phosphorylation and STAT5b nuclear translocation was markedly reduced when compared with diestrous rats. In nonpregnant female rats, acute treatment with ovine prolactin markedly increased levels of SOCS-1 and -3 and cytokine-inducible SH2-containing protein mRNA in arcuate nucleus micropunches. On gestation

T

HE CONCENTRATION OF prolactin in circulation is normally tightly regulated by a short-loop negative feedback system. Prolactin activates neuroendocrine dopamine neurons within the hypothalamus (1), presumably by acting directly on prolactin receptors expressed on these neurons (2, 3). Neuroendocrine dopamine neurons are therefore able to register a rise in the concentration of prolactin and respond with an increase in secretion of dopamine into the portal circulation, which in turn acts on the pituitary lactotrophs to inhibit more prolactin release. Three anatomically distinct populations of neuroendocrine dopaminergic neurons participate in the regulation of prolactin, the largest and best studied of which are the tuberoinfundibular dopaminergic (TIDA) neurons (4). Prolactin receptor-mediated First Published Online July 20, 2006 Abbreviations: CIS, Cytokine-inducible SH2-containing protein; CV, coefficients of variation; DOPAC, 3,4-dihydroxyphenylacetic acid; G0, gestational d 0; icv, intracerebroventricular; L0, lactational d 0; pSTAT5, phosphorylated STAT5; SOCS, suppressors of cytokine signaling; STAT5b, signal transducer and activator of transcription 5b; TIDA, tuberoinfundibular dopamine. Endocrinology is published monthly by The Endocrine Society (http:// www.endo-society.org), the foremost professional society serving the endocrine community.

d 22, SOCS-1 and SOCS-3 mRNA levels were 10-fold that on G20. SOCS-1 and -3 and cytokine-inducible SH2-containing protein mRNA levels were also elevated on lactation d 7. At these times, dopaminergic activity was decreased and the rats were hyperprolactinemic. The high levels of SOCS mRNA were prevented by bromocriptine pretreatment (gestation d 22) or pup removal (lactation d 7), which suppressed circulating prolactin to basal levels. These results demonstrate that around the end of pregnancy, prolactin loses the ability to activate STAT5b, associated with an increase in SOCS mRNAs. The loss of this stimulating pathway may underlie the reduced tuberoinfundibular dopamine neuron dopamine output and hyperprolactinemia that characterizes late pregnancy and lactation. The high maternal levels of SOCS mRNAs appear to be dependent on prolactin, presumably acting through an alternative signaling pathway to STAT5b. (Endocrinology 147: 4996 –5005, 2006)

signal transduction in these neurons appears to exclusively use signal transducer and activator of transcription 5b (STAT5b) to stimulate tyrosine hydroxylase (the rate-limiting enzyme for dopamine synthesis) gene transcription (5), although rapid effects of prolactin on tyrosine hydroxylase activity may be mediated through other kinase pathways (6). Consistent with this, transgenic STAT5b-deficient mice are hyperprolactinemic because of an inability of their TIDA neurons to respond appropriately to prolactin (7), and prolactin can induce phosphorylation of STAT5 in the arcuate nucleus (8). Acute surges of prolactin occur in many situations, such as during the afternoon and night of the first half of pregnancy in rodents (9, 10). Although prolactin’s negative feedback loop restores its circulating concentration to basal levels within a few hours in these situations, a few physiological circumstances have been described in which prolonged hyperprolactinemia is permitted because of reduced TIDA neuronal responsiveness to prolactin. The end of pregnancy and lactation are such cases. We have shown that TIDA neuronal activity, which is maintained at high levels throughout most of the second half of pregnancy because of stimulation by placental lactogens (11, 12), declines markedly 1 d before parturition despite the continued presence of lactogens. This

4996

Anderson et al. • Prolactin Signaling in Maternal Arcuate Nucleus

elicits a sustained surge of prolactin, albeit after a delay of about 12 h (13). Treatment with high levels of exogenous prolactin or placental lactogens does not attenuate this antepartum prolactin surge (14, 15), emphasizing the marked reduction in responsiveness of the TIDA neurons and negative feedback system to stimulation by these hormones. This loss of feedback sensitivity is maintained during lactation (16, 17). The neuroendocrine and cellular mechanisms by which TIDA neuronal responsiveness to prolactin declines to allow hyperprolactinemia at these times is unknown, but they do not appear to involve a down-regulation of prolactin receptor mRNA or protein (3, 18, 19) on TIDA neurons. Recently, a family of cytokine-inducible inhibitors of cytokine signaling has been described (20 –22). These proteins act as feedback inhibitors of signaling for a range of cytokines that use Janus kinase (JAK)/STAT pathways. In vitro studies of the action of these proteins, termed suppressor of cytokine signaling (SOCS) proteins, indicate that they act as part of an intracellular feedback loop to inhibit the phosphorylation of STATs (23), thereby restraining the level of signal transduction. In vivo, prolactin has been shown to induce mRNA for SOCS-1, -2, and -3 and cytokine-inducible SH2-containing protein (CIS) in ovary, adrenal, and mammary glands (24). The role played by these SOCS proteins in prolactin’s hypothalamic actions is virtually unknown, although CIS mRNA and protein levels have recently been shown to be increased in the hypothalamus in response to the suckling stimulus in lactating rats (25). The abrupt reduction in TIDA neuronal responsiveness to prolactin in late pregnancy, after a prolonged period of stimulation by placental lactogens, suggests a possible role for SOCS in inhibiting prolactinstimulated STAT5b signaling at this time and during the ensuing lactational period. In this study, we have investigated this hypothesis by testing whether lactation inhibits prolactin-induced STAT5 signaling as well as by quantifying levels of SOCS-1, -2, and -3 and CIS mRNA within the arcuate nucleus during late pregnancy and lactation. Materials and Methods Animals Female Sprague Dawley rats, aged 10 –12 wk and weighing 200 –250 g at the beginning of the experiments, were obtained from the University of Otago animal facility. Rats were group housed (except after surgery or parity) under conditions of controlled lighting (lights on from 0500 – 1900 h) and temperature (22 ⫾ 1 C) and had free access to pelleted food and water. For experiments 1, 3, and 4, estrous cycles were monitored by daily cytological examination of vaginal smears. Proestrous females were paired with males, and mating was confirmed by the presence of sperm in vaginal smears the following morning. This was designated gestation d 0 (G0), and parturition occurred on the morning of G22, also designated lactation d 0 (L0). Pup litters were normalized to 10 pups within 24 h of birth. All animal experimental protocols were approved by the University of Otago Committee on Ethics in the Care and Use of Laboratory Animals.

Perfusion and immunohistochemistry for phosphorylated STAT5 (pSTAT5), STAT5b, and tyrosine hydroxylase 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 of 4% paraformaldehyde in 0.1 m phosphate buffer (pH 7.4). Brains were removed and postfixed in the same fixative. For pSTAT5 immunohis-

Endocrinology, October 2006, 147(10):4996 –5005

4997

tochemistry (experiment 1a), the brains were microwave heated to 55 C for 5 min within 4 h of perfusion to reduce formation of paraformaldehyde cross-bridges before embedding overnight in paraffin wax. Coronal microtome sections (12 ␮m thick) were prepared at the level of the arcuate nucleus from each brain to provide several similar series of consecutive sections (240 ␮m apart), which were floated onto (3-aminopropyl)triethoxysilane-coated slides and stored at room temperature. Immediately before immunohistochemistry, the sections were dewaxed and rehydrated in xylene and graded ethanol solutions before being subjected to an antigen-retrieval procedure (heating in a 1000-W microwave oven at high power for 20 min in boiling 0.1 m sodium citrate buffer, then cooling for 15 min in the same solution) (26). The sections were washed three times in 0.1 m phosphate buffer and then incubated for 48 h at 4 C in 5% normal goat serum containing the following primary antibodies: rabbit polyclonal anti-pSTAT5 (tyr 694, 1:400; Cell Signaling Technology, Beverly, MA) and mouse monoclonal anti-tyrosine hydroxylase (MAB318, 1:10,000; Chemicon, Temecula, CA). Sections were washed again and incubated for another 3 h with goat antimouse Alexa 488 and goat antirabbit Alexa 568 (both 1:1000; Molecular Probes, Eugene, OR) in 1% normal goat serum and then washed again. Sections were coverslipped with Vectorshield (Vector Laboratories, Burlingame, CA) and viewed and photographed using fluorescence microscopy on an Olympus BX51 microscope at ⫻200 magnification. All tyrosine hydroxylase immunoreactive cells on both sides of the dorsomedial and ventrolateral regions of the arcuate nucleus from two to three evenly spaced sections were analyzed for colocalization with pSTAT5, which was expressed only in the cell nuclei. For each animal, the percentage of pSTAT5-colocalized TIDA neurons was calculated to provide a single data point. Omission of pSTAT5 and tyrosine hydroxylase primary antibodies resulted in an absence of specific staining. For STAT5b immunohistochemistry (experiment 1b), the brains were infiltrated until sinking in 30% sucrose for cryoprotection. Coronal sections (40 ␮m thick) were cut at the level of the arcuate nucleus on a sledge microtome with a freezing stage to provide a series of consecutive sections 240 ␮m apart. These were stored in cryoprotectant at ⫺20 C. The free-floating sections were subjected to the same immunohistochemistry procedure as outlined above for the mounted sections (without antigen retrieval) except that the primary antibodies were mouse monoclonal anti-STAT5b (1:3000; Zymed Laboratories, San Francisco, CA) and rabbit polyclonal anti-tyrosine hydroxylase (AB151, 1:6000; Chemicon). The sections were mounted on (3-aminopropyl)triethoxysilane-coated slides, coverslipped with Vectorshield, and viewed using fluorescence microscopy as outlined above for colocalization of tyrosine hydroxylase and STAT5b. STAT5b staining within identified TIDA neurons was categorized as either located primarily in the cytoplasm, located primarily in the nucleus, evenly distributed throughout the cell, or absent (5). For each animal, the percentage of nuclear STAT5b-colocalized TIDA neurons was calculated to provide a single data point. Omission of STAT5b and tyrosine hydroxylase primary antibodies resulted in an absence of specific staining.

Brain microdissection and serum collection Animals for experiments 2– 4 were killed for brain microdissection by decapitation. Trunk blood samples were collected and allowed to clot, and the harvested serum was stored at ⫺20 C until prolactin RIA. The brains were immediately removed, frozen on dry ice, and stored at ⫺80 C until microdissection of the hypothalamic arcuate nucleus for measurement of SOCS mRNA. Thick coronal brain sections (300 ␮m) were cut in a cryostat at ⫺9 C, thaw-mounted onto glass slides, and refrozen. The arcuate nucleus was dissected from four consecutive sections (between approximately ⫺2.2 and ⫺3.4 mm relative to Bregma) with a single midline punch of a sterile 21-gauge micropunch needle (500 ␮m internal diameter) (27). Micropunched arcuate tissue was stored at ⫺80 C in 50 ␮l of a commercial lysis buffer (Cells-to-cDNA II Cell Lysis Buffer; Ambion, Austin, TX) until RT. For experiment 4, the median eminence was also punched from three consecutive sections (approximately Bregma ⫺2.5 to ⫺3.4) using two overlapping 21-gauge needle punches (27). Median eminence tissue was placed in 150 ␮l of tissue buffer (0.05 m Na2HPO4; 0.03 m citric acid; 12% methanol, pH 3.0), sonicated in an ultrasonic cell disrupter, and centrifuged at 10,000 rpm for 5 min, and the supernatant was collected and stored at ⫺20 C until measurement of 3,4-dihydroxyphenylacetic acid (DOPAC) by HPLC.

4998



The tissue pellet was dissolved in 1 m NaOH and stored at ⫺20 C until measurement of protein content using an assay kit (Bio-Rad, Hercules, CA).

RNA. Coefficients of variation (CV) for the sample duplicates were almost always less than 1%.

Preparation of cDNA

Prolactin RIA

The crude tissue lysate in cell lysis buffer was sonicated in an ultrasonic cell disrupter and then incubated at 75 C for 10 min to rupture the cells and inactivate endogenous RNases. Next, the lysate was treated with 4 U DNase 1 (Ambion) according to the manufacturer’s instructions to degrade genomic DNA. After DNase 1 inactivation, RT of RNA into cDNA was performed in the Cells-to-cDNA II Cell Lysis Buffer (RNA isolation is not required with this procedure). A 20-␮l volume of lysate (containing 56 ⫾ 2 ng total RNA) was reverse transcribed into cDNA in a 33-␮l reaction with 30 U MultiScribe reverse transcriptase (Applied Biosystems, Foster City, CA) using random hexamer primers (Applied Biosystems). Each run, which was conducted in a 96-well PCR plate, included a control tube with reverse transcriptase enzyme omitted to demonstrate that the template for the PCR product was not genomic DNA.

Serum prolactin concentration was measured in 10-␮l sample volumes by RIA. Values are expressed in terms of the rat standard NIDDKrat PRL-RP-3. Iodinated rat prolactin (NIDDK-rat PRL-I-6) was used as tracer, and primary antiserum was NIDDK rabbit antirat PRL-RIA-9 (final dilution, 1:150,000). The sensitivity (95% confidence interval at 0 ng/ml on the standard curve) averaged 2 ng/ml (four assays). The intraassay CV averaged 15% for a serum pool falling in the middle of the standard curve, and the interassay CV was 8% for the same serum pool.

Real-time quantitative PCR measurement of SOCS-1, -2, and -3 and CIS mRNA TaqMan probes and forward (sense) and reverse (antisense) primers to the rat genes encoding SOCS-1, -2, and -3 and CIS (see Table 1) were designed using Primer Express software (Applied Biosystems). Realtime PCR was conducted as described previously (19) with minor modifications. A 20-␮l reaction mix was prepared containing primers and probes at optimized final concentrations (Table 1), ABsolute qPCR Rox master mix (ABgene, Epsom, UK), and RNase-free water. Template cDNA (2.5 ␮l) was added. A stock of cDNA containing relatively high levels of SOCS mRNA was created by pooling many punches of liver (20), and 4-fold dilution series of this were run on each plate as external standards. An ABI PRISM 7000 Sequence Detection System (Applied Biosystems) was used to detect fluorescence during each PCR cycle. 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 SOCS mRNA from the sample CT values. Initial attempts to normalize the sample data to the housekeeping gene 18S rRNA proved unsuccessful because of variation in the levels of this gene across treatments. We therefore normalized the sample data to total RNA content (which is advocated as a less error-prone method) (28). Total RNA in the remaining tissue lysate was measured in 96-well plates using a Quant-iT RNA assay kit (Molecular Probes), which has a lower detection limit of around 2 ng and tolerates a range of contaminating substances in the solution. The fluorescent signal was read on a POLARstar Optima fluorescence microplate reader (BMG Labtechnologies, Offenburg, Germany) at excitation and emission wavelengths of 640 and 680 nm, respectively. The SOCS mRNA levels were then expressed as arbitrary (relative) units per nanogram total

Estimation of TIDA neuronal dopamine output The content of the dopamine metabolite DOPAC in the median eminence was measured to provide an index of TIDA neuronal dopamine output. Levels of DOPAC are considered to be proportional to the amount of dopamine released from nerve terminals (29). DOPAC concentration was assayed by isocratic HPLC with electrochemical detection. The mobile phase, consisting of tissue buffer containing 0.1 mm EDTA and 0.4 mm sodium octyl sulfate (pH 3.0), was pumped at a flow rate of 600 ␮l/min. Catecholamines were separated on a reverse-phase Luna 3-␮m C18(2) column (150 mm; Phenomenex, Auckland, New Zealand) and detected using a conditioning cell (⫹350 mV) and a dualelectrode analytical cell (⫺150 mV and ⫹350 mV) using a Coulochem II detector (ESA, Bedford, MA). DOPAC was quantified by comparison with external standards. All values were corrected for total protein content in the tissue punches.

Experiment 1a: does lactation inhibit phosphorylation of STAT5 in TIDA neurons? Virgin diestrous and lactating rats were used. All animals were chronically fitted with 22-gauge intracerebroventricular (icv) cannulae 2– 4 d before perfusion to permit ovine prolactin injection into a lateral ventricle. At approximately 1000 h on the morning of diestrus or 1 wk postpartum, all rats were injected with bromocriptine methanesulfonate (200 ␮g in 200 ␮l 10% ethanol; Research Biochemicals, Inc., Natick, MA) (5), and the pups of lactating dams were removed (10) to reduce endogenous circulating prolactin levels to basal levels. Four hours later, the rats received a single injection of ovine prolactin (4 ␮g icv in 4 ␮l saline; Sigma Chemical Co., St Louis, MO) or vehicle and were perfused for brain collection 30 min later (n ⫽ 4 –5 per group). The dose and mode of administration of ovine prolactin were based on pilot experiments (data not shown).

TABLE 1. Details of oligonucleotides used for real-time quantitative PCR Oligonucleotide

SOCS-1 Sense Antisense Taqman probe SOCS-2 Sense Antisense Taqman probe SOCS-3 Sense Antisense Taqman probe CIS Sense Antisense Taqman probe

NCBI accession no. and sequence

Accession no.

Nucleotide no.

Final conc. (nM)

1304–1322 1371–1355 1329–1351

900 900 250

488–504 559–532 506–530

300 300 250

2065–2085 2147–2127 2091–2109

300 300 250

454– 472 516–500 474– 498

300 300 150

AJ243123 5⬘-CTTCTTGGTGCGCGACAGT-3⬘ 5⬘-GGCCCGAAGCCATCTTC-3⬘ 6FAM-CGGAACTGCTTCTTCGCGCTCAG-TAMRA NM_058208 5⬘-CAGCCGGGCCGACTAAC-3⬘ 5⬘-TGACACATATGATAGAATCCAATCTGAA-3⬘ 6FAM-TGCGGATTGAGTACCAAGACGGGAA-TAMRA AJ249240 5⬘-CAGCTCCAAGAGCGAGTACCA-3⬘ 5⬘-CGGTTACGGCACTCCAGTAGA-3⬘ 6FAM-TGGTGAACGCCGTGCGCAA-TAMRA XM_343472 5⬘-GCCTTCCCAGATGTGGTCA-3⬘ 5⬘-CCGGGTGTCAGCCGTAC-3⬘ 6FAM-CCTTGTGCAGCATTACGTGGCCTCC-TAMRA



4999

Experiment 1b: does lactation inhibit nuclear translocation of STAT5b in TIDA neurons? Virgin diestrous and lactating rats were used as for experiment 1a, but without icv cannulae or bromocriptime pretreatment. In addition to the lactating rats with 4 h of pup removal, lactating rats with pups removed for 24 h were included to examine long-term absence of suckling on prolactin-induced STAT5b signaling. All animals received a single injection of ovine prolactin (250 ␮g sc in 200 ␮l saline; Sigma) (5) or vehicle and were perfused for brain collection 40 min later (n ⫽ 4 –5 per group).

Experiment 2: does acute prolactin treatment increase SOCS mRNA expression in the arcuate nucleus? In this experiment, we aimed to determine which of the SOCS mRNAs were up-regulated by an acute increase in the circulating prolactin concentration. Twenty rats were ovariectomized 7 d before the experiment to eliminate cyclical variation in the endogenous prolactin concentration. All animals received a single injection of ovine prolactin (300 ␮g sc in 200 ␮l saline; Sigma) or vehicle at approximately 1000 h and were decapitated for brain collection 1, 2, or 4 h later (n ⫽ 5 per group). The choice of ovine prolactin dose was based on previous reports (25, 30). Vehicle-treated controls were decapitated 1 h after injection.

Experiment 3: is the expression of SOCS mRNAs in the arcuate nucleus elevated during nocturnal prolactin surges of early pregnancy? We used the twice-daily surges of prolactin exhibited by early pregnant rats to examine whether SOCS expression is induced during periods of increased endogenous prolactin secretion. On G6 –G7, rats were decapitated for brain collection at times corresponding to immediately before, during, and after the nocturnal prolactin surge (2400, 0200, or 1100 h; n ⱖ 10 per group) (10). To test whether changes in SOCS expression were induced by the high endogenous circulating prolactin levels, another group of rats (n ⫽ 10) was used, identical to the 0200-h group except that they were treated 5 h before decapitation with bromocriptine methanesulfonate (200 ␮g sc in 200 ␮l 10% ethanol) (Research Biochemicals) to block the nocturnal prolactin surge.

Experiment 4: is the expression of SOCS mRNAs in the arcuate nucleus elevated during the hyperprolactinemia of late pregnancy and lactation? We hypothesized that increased SOCS expression might be responsible for the loss of TIDA neuronal responsiveness to prolactin (and consequent hyperprolactinemia) that occurs during late pregnancy and lactation. On G20, G22, and L7, rats (n ⱖ 10 per group) were decapitated at 1200, 0300, and 1200 h, respectively (0300 h on G22 represents the approximate peak of the prolonged late-pregnancy prolactin surge (13). To test whether changes in SOCS expression were induced by the elevated prolactin levels, two additional groups of rats (G22 and L7) were pretreated 8 h before decapitation to reduce prolactin levels by administering bromocriptine (as for experiment 3) or removing pups, respectively. Brains were treated and analyzed for SOCS mRNA and DOPAC (a neurochemical index of TIDA neuronal activity) levels in microdissected arcuate nucleus and median eminence tissue, respectively.

Statistical analysis All significant treatment effects were identified using ANOVA, followed by the Bonferroni t test for post hoc analysis to determine where significant effects occurred. Where data failed equal variance or normality tests, they were log-transformed (base 10) before statistical analysis. Results are presented as the mean ⫾ sem, and differences were considered significant at P ⬍ 0.05.

Results Experiment 1a

Representative images of TIDA neurons costained for pSTAT5 are shown in Fig. 1. In control (vehicle-treated)

FIG. 1. Top, Representative examples of staining for tyrosine hydroxylase (green fluorescence) and pSTAT5 (red fluorescence) in the arcuate nucleus of vehicle-treated diestrous rats (left), ovine prolactin-treated (4 ␮g icv) diestrous rats (middle), and ovine prolactintreated lactating rats (right). Long arrows depict examples of TIDA neurons exhibiting pSTAT5 immunoreactivity; short arrows depict examples of pSTAT5-negative TIDA neurons. Arrowhead denotes pSTAT5 immunoreactivity in a tyrosine-hydroxylase-negative cell. Bottom, Percentage of TIDA neurons with pSTAT-immunoreactive nuclei in response to ovine prolactin treatment (n ⫽ 4 –5 per group). *, P ⬍ 0.05; **, P ⬍ 0.001 vs. the corresponding vehicle-treated rats; †, P ⬍ 0.05 vs. vehicle-treated diestrous rats.

diestrous rats, there was little or no detectable pSTAT5 in the nuclei of TIDA neurons. Control lactating rats had a significantly higher basal proportion of TIDA neurons colocalized with pSTAT5 compared with control diestrous rats (P ⬍ 0.05). Although treatment with prolactin significantly increased pSTAT5 coexpression in all rats, the magnitude of the effect was greatly reduced in lactating compared with diestrous animals (P ⬍ 0.05) (Fig. 1). The number of detectable TIDA neurons was not affected by any treatment (mean, 15.4 ⫾ 1.6 neurons per section). Most arcuate cells responding to prolactin were TIDA neurons. The number of nontyrosine-hydroxylase arcuate cells per section exhibiting pSTAT5 immunoreactivity went from 0.1 ⫾ 0.1 (vehicletreated) to 2.2 ⫾ 0.4 (prolactin-treated) in diestrous rats and from 3.8 ⫾ 1.4 (vehicle-treated) to 4.3 ⫾ 0.4 (prolactin-treated) in lactating rats. Experiment 1b

Representative images of TIDA neurons costained for STAT5b are shown in Fig. 2. In control (vehicle-treated) diestrous rats, less than 10% of colocalized STAT5b immunoreactivity was located predominantly in the nuclei of TIDA neurons. Similarly to the previous experiment, control lactating rats (with pups removed for 4 h) had a significantly higher basal proportion of TIDA neurons containing predominantly nuclear STAT5b than control diestrous rats (P ⬍

5000



FIG. 2. Top, Representative examples of staining for tyrosine hydroxylase neurons (left, red fluorescence) and STAT5b (right, green fluorescence) in the arcuate nucleus. Short arrows denote predominantly cytoplasmic colocalized STAT5b staining; long arrows denote predominantly nuclear colocalized STAT5b staining; open arrow denotes even STAT5b staining throughout the cytoplasm and nucleus, and arrowheads denote STAT5b immunoreactivity in tyrosine-hydroxylase-negative cells. Bottom, Percentage of TIDA neurons with STAT5b predominantly translocated to the nuclei in response to ovine prolactin treatment (n ⫽ 4 –5 per group). *, P ⬍ 0.01 vs. the corresponding vehicle-treated rats; †, P ⬍ 0.05 vs. vehicle-treated diestrous rats.

FIG. 3. Relative levels of mRNA for SOCS-1, ⫺2, ⫺3 and CIS in microdissected arcuate nuclei in ovariectomized rats in response to 300 ␮g ovine prolactin sc (n ⫽ 5 per group). *, P ⬍ 0.05 vs. 0 h.

0.05), although this difference was not seen in lactating animals with pups removed for 24 h. Prolactin treatment markedly increased nuclear STAT5b staining in TIDA neurons of diestrous rats (P ⬍ 0.01). In contrast, prolactin had no effect on this variable in either group of lactating animals (Fig. 2). The proportion of TIDA neurons in which STAT5b could not be detected in any compartment was not affected by any treatment (mean, 4.8 ⫾ 1.9% of TIDA neurons).

by 2400 h (start of the nocturnal surge). Pretreatment with bromocriptine abolished the nocturnal prolactin surge (P ⬍ 0.001 vs. untreated 0200-h animals) (Fig. 4). In contrast to the responses to exogenous prolactin seen in experiment 2, the levels of all four of the SOCS mRNAs did not increase significantly in association with the peak of the nocturnal prolactin surge, nor was there any significant decrease in response to suppression of endogenous prolactin by bromocriptine pretreatment (Fig. 5).

Experiment 2

Experiment 4

Levels of SOCS-1 and -3 and CIS mRNA in the arcuate nucleus were low/undetectable in vehicle-treated rats, and all increased significantly after treatment with prolactin. Peak levels were seen 2 h after the sc injection (Fig. 3). The magnitude of the response was greatest for CIS (17-fold induction) and least for SOCS-1 mRNA (7-fold induction). In contrast, levels of SOCS-2 mRNA were not affected by acute prolactin treatment and remained at intermediate levels at all time points (Fig. 3).

As expected in late-pregnant rats, serum prolactin concentration was low on G20 but high on G22 and L7 (P ⬍

Experiment 3

In early-pregnant (G6 –7) rats, serum prolactin concentration was lowest at 1100 h (the nadir between the diurnal and nocturnal prolactin surges) and highest at 0200 h (peak of the nocturnal surge) (P ⬍ 0.001 vs. 1100 h; Fig. 4). In a few (four of 12) rats, the serum prolactin concentration had increased

FIG. 4. Serum prolactin concentration in early-pregnant (G6 –7) rats (n ⱖ 10 per group). At 0200 h is the approximate time when the nocturnal prolactin surge peaks. *, P ⬍ 0.05; **, P ⬍ 0.001 vs. 1100 h; ††, P ⬍ 0.001 vs. 0200 h (untreated).



5001

G22 and L7 (P ⬍ 0.01), and CIS mRNA was increased 4-fold at the latter time (P ⬍ 0.05). These increases in SOCS mRNA levels appeared to be caused largely by the elevated circulating prolactin concentration, because the effects were prevented when endogenous prolactin was suppressed to basal levels by bromocriptine pretreatment on G22 (P ⬍ 0.01) or pup removal on L7 (P ⬍ 0.01) (Fig. 7). A small but significant prolactin-dependent elevation of SOCS-2 mRNA was also observed during lactation (P ⬍ 0.05) (Fig. 7). Discussion

FIG. 5. Levels of mRNA for SOCS-1, -2, and -3 and CIS in arcuate nuclei microdissected from early-pregnant (G6 –7) rats (n ⱖ 10 per group). At 0200 h is the approximate time when the nocturnal prolactin surge peaks. There were no significant changes over time or in response to bromocriptine pretreatment to suppress the nocturnal prolactin surge.

0.001). Bromocriptine on G22 or pup removal during lactation caused a return to basal prolactin levels (P ⬍ 0.001) (Fig. 6). As reported previously, the concentration of median eminence DOPAC in untreated rats followed an inverse pattern to prolactin (high TIDA activity on G20 and low activity on G22 and L7; P ⬍ 0.01) (Fig. 6). Levels of mRNA for all four SOCS proteins were low/undetectable on G20; however, SOCS-1 and SOCS-3 mRNAs were increased 9- to 15-fold on

FIG. 6. Serum prolactin concentration (top) and median eminence DOPAC content (an index of TIDA neuronal dopamine output, bottom) in late-pregnant (G20 –22) and lactating (L7) rats (n ⱖ 10 per group). The antepartum prolactin surge occurs on G21–22. *, P ⬍ 0.05; **, P ⬍ 0.001 vs. G20; ††, P ⬍ 0.001 vs. G22 (untreated); ‡‡, P ⬍ 0.001 vs. L7 (untreated).

These results show that prolactin-induced activation of STAT5b in TIDA neurons is significantly reduced during lactation. This is consistent with the previously described inability of prolactin to induce dopamine synthesis at this time (16, 17). However, mRNAs for other downstream prolactin-sensitive genes, SOCS proteins, are elevated and remain responsive to prolactin during late pregnancy and lactation. Acute prolactin treatment caused a rapid induction of SOCS-1 and -3 and CIS mRNAs. During the prolonged latepregnancy prolactin surge and during lactation, SOCS-1 and -3 were markedly elevated in a prolactin-dependent manner, whereas the prolactin surges of early pregnancy did not appear to produce this effect. These results extend recent reports that prolactin can induce SOCS mRNA in peripheral tissues (24, 31) and suggest that this is also a normal part of the signal transduction mechanism in neurons. The reduction in prolactin-induced STAT5 signaling during lactation is evidence that the change in prolactin negative feedback that permits hyperprolactinemia at this time is mediated at a post-receptor level. Consistent with this, prolactin

FIG. 7. Levels of mRNA for SOCS-1, -2, and -3 and CIS in arcuate nuclei microdissected from late-pregnant (G20 –22) and lactating (L7) rats (n ⱖ 10 per group). The antepartum prolactin surge occurs on G21–22. *, P ⬍ 0.05; **, P ⬍ 0.001 vs. G20; †, P ⬍ 0.05 vs. G22 (untreated); ‡, P ⬍ 0.05 vs. L7 (untreated).

5002


receptor mRNA and protein expression in the arcuate nucleus and on TIDA neurons are maintained at levels at least as high as those seen during nonpregnancy and early pregnancy (3, 18, 19). The experimental paradigm we used involved removal of pups for at least 4 h to ensure that the circulating prolactin concentration was basal (10), although the unresponsiveness of STAT5b to prolactin was maintained even after 24 h of pup removal. A recent similar study using lactating rats showed that in the presence of pups, nuclear translocation of STAT5 in TIDA neurons is inhibited compared with animals in which pups had been removed for 16 h (25). Our data add to this by showing that even in the absence of the suckling stimulus, prolactin-induced phosphorylation of STAT5 is inhibited by lactation and by showing that the STAT signaling isoform that is regulated by lactation is STAT5b. This is in agreement with our recent finding that STAT5b signaling mediates prolactin’s transcriptional actions in nonpregnant rats, despite the presence within TIDA neurons of the closely related STAT5a (5). Interestingly, we observed a considerable difference in the magnitude of the response to prolactin in experiments 1a and 1b (i.e. ⬎ 80% of TIDA neurons responding to prolactin in the pSTAT5 experiment during diestrus but only 25% in the STAT5b experiment; in addition, a small pSTAT5 response to prolactin was observed during lactation, whereas this was not seen for STAT5b nuclear translocation). These differences were likely because of differences in local prolactin concentration achieved by the two different routes of administration as well as limitations in the ability to completely distinguish nuclear-translocated from cytoplasmic STAT5b visually in experiment 1b. In agreement with our results, others (8) have recently reported that lactating rats retain at least some ability to phosphorylate STAT5 in response to prolactin, although these authors did not report comparisons with nonlactating rats. The refractoriness in the ability of prolactin to induce STAT5b nuclear translocation during lactation appears to last at least 24 h after pup removal. This is in agreement with other work showing that lactating rats with pups removed for 24 h are still unable to respond to exogenous prolactin with an increase in TIDA neuron dopaminergic activity (16). However, a recent experiment using Wistar rats (25) showed that prolactin-induced STAT5 nuclear translocation in TIDA neurons had returned to levels equivalent to those of diestrous controls by 16 h after pup removal. At present, the reason for this difference is unclear but may be related to differences in animal strain or the way nuclear translocation was measured. Because assessing changes in mRNA levels in whole hypothalami or large hypothalamic dissections could lead to misleading results through averaging of multiple different responses, we used the micropunch technique to specifically isolate the arcuate nucleus from the rest of the hypothalamus. Nevertheless, this area contains multiple cell types and is involved in numerous diverse functions (32). It is therefore possible that the responses to prolactin reported here represent the integrated response of multiple prolactin-responsive cell types, including the TIDA neurons. We and others have shown that more than 90% of TIDA neurons contain prolactin receptor mRNA (3) and protein (2), although these are also observed in a small number of other unidentified


arcuate cells (2, 3). It is probable that TIDA neurons would have contributed the majority of the prolactin-dependent increases in SOCS levels observed, because they account for 80 –90% of prolactin receptor-expressing cells in the arcuate nucleus (3). Our pSTAT5 immunohistochemistry results support this, because most of the pSTAT5 immunoreactivity observed after treatment with prolactin was coexpressed with TIDA neurons. A number of studies have confirmed the presence of SOCS3 mRNA within the arcuate nucleus (e.g. Refs. 33 and 34), and one has reported protein expression (35). In contrast, others have recently reported that SOCS-1 and -3 are undetectable in the arcuate nucleus (25). These conflicting findings and the paucity of data on SOCS protein expression in the hypothalamus suggest the technical difficulties involved in obtaining good SOCS staining using immunohistochemistry. Prolactin acutely increased SOCS mRNAs in the arcuate nucleus, showing that this is likely to be a part of normal homeostatic signaling responses in TIDA neurons. Although we did not observe any change in these genes in response to endogenous prolactin fluctuations during early pregnancy, in late pregnancy and lactation, the situation differed. At this time, a marked increase in SOCS-1 and -3 mRNA in the arcuate nucleus occurred (Fig. 7), temporally associated with increased serum prolactin concentrations (Fig. 6), which begin on G21 (13). Our results show that these high levels of circulating prolactin are required for the increased SOCS mRNA levels, because prolactin suppression by bromocriptine treatment prevented the effect. Furthermore, pup removal also prevented high SOCS mRNA levels in lactating rats, showing that this is dependent on prolactin and/or the suckling stimulus (which is known to also induce SOCS proteins) (25). Based on their modes of action described in vitro, these high levels of SOCS proteins would be expected to suppress STAT signaling in the TIDA neurons by binding to phosphotyrosine residues in Janus kinase 2 (in the case of SOCS-1) or prolactin receptors (in the case of SOCS-3 and CIS) (36 –38). Hence, prolactin activation of STAT5b-mediated tyrosine hydroxylase mRNA induction in TIDA neurons would be decreased, resulting in the suppression of dopamine output seen in this study (Fig. 6) and others (13). In support of this concept, prolactin-induced STAT5 phosphorylation and STAT5b nuclear translocation in TIDA neurons was markedly reduced by lactation in our study. The resulting suppression of tyrosine hydroxylase gene transcription in TIDA neurons would permit an increase in prolactin secretion, which could in turn stimulate more SOCS induction. Implicit in this argument is the idea that whereas STAT5b signaling is suppressed during late pregnancy and lactation, prolactin can still stimulate SOCS transcription, perhaps via an alternative signaling pathway such as MAPK (39), which is not itself inhibited by SOCS proteins (40). This cascade of events may be initiated by factors specific to late pregnancy, because the nocturnal prolactin surge on G7 was unable to induce arcuate nucleus SOCS mRNA expression (at least at 0200 h, which is 1–2 h after surge initiation), and the normal prolactin feedback response of TIDA neuronal dopamine output is maintained throughout the early-pregnancy prolactin surges (41). The differentiating mechanisms that cause SOCS to be


induced in late pregnancy but not early pregnancy remain to be elucidated, but it is unlikely to be the differing circulating prolactin concentrations, because at the times of sample collection, these were very similar. It is more likely that the divergent responses result from the differing endocrine environments during the periods leading up to the time of sample collection. In contrast to early pregnancy, the second half of pregnancy is characterized by constant stimulation of the neuroendocrine dopamine neurons by placental lactogens, which mimic prolactin’s negative feedback actions to produce chronic TIDA neuronal stimulation (11, 12, 42). Furthermore, the steroidal milieu differs greatly between early and late pregnancy. Although G7 is characterized by low levels of estrogen and medium-high levels of progesterone, G22 is characterized by high estrogen and low progesterone (43). A number of recent experiments have shown that these steroids are able to influence prolactin or GH-induced STAT5 transcriptional activity (44 – 47), and a recent study by Leung et al. (46) has shown that estrogen is able to induce SOCS-2 in vitro. TIDA neurons, in which estrogen and prolactin receptors and STAT5 are all expressed (2, 5, 48 –51), would appear to be likely candidates for such cross-talk between the estrogen receptor and prolactin signaling pathways, especially in late pregnancy when circulating estrogen and placental lactogens are markedly elevated. Hence, it is possible that rising estrogen, coupled with consistently elevated placental lactogens, would act as the trigger that promotes preferential SOCS expression in response to prolactin in TIDA neurons. Although the nocturnal surge of prolactin on G7 did not cause induction of SOCS mRNA, a single injection of ovine prolactin (which creates a similar surge of circulating prolactin) induced SOCS-1 and -3 and CIS mRNA up to 17-fold compared with control levels. The reason for this difference is unclear at present. It is possible that a sharp rise in prolactin after the very low levels characteristic of ovariectomized animals would have induced a more acute response of STAT5b activity (as seen in experiment 1) followed by SOCS induction than in intact, early-pregnant animals, which would have had fluctuating levels of prolactin over the preceding week. An alternative explanation could be that progesterone, which is elevated by G7 (43), inhibits SOCS expression as it has been shown to do for another suppressor of cytokine signaling, caveolin-1 (52). It is noteworthy that nonpregnant rats injected with prolactin have a much more marked suppression of TIDA activity (which occurs after several hours and hence was not measured in experiment 2) than is seen in response to early-pregnancy prolactin surges (41). However, the comparison of these two situations from our results is somewhat limited by the fact that SOCS mRNA was measured only at a single time point during the nocturnal prolactin surge on G7. It is possible that an acute rise in SOCS mRNA occurs but at a later time after the peak of the surge. The analysis of transgenic mice lacking SOCS genes has begun to shed light on their in vivo roles. This approach allows functional redundancies between different SOCS proteins to be assessed, in contrast to studies in which SOCS proteins are overexpressed. Although socs1⫺/⫺ and socs3⫺/⫺ mice die perinatally and embyonically because of immune


5003

deregulation and placental insufficiency, respectively, the socs1⫺/⫺ mice can be rescued if bred on an interferon-␥deficient background (53). These mice displayed enhanced mammary lobuloalveolar development during pregnancy and increased milk production, suggesting an important role for SOCS1 in limiting prolactin signaling. Moreover, deletion of a single copy of socs1 rescued lactational defects in mice heterozygous for a knockout of the prolactin receptor (53). On the other hand, recent studies have used cre-lox (tissuespecific) socs3⫺/⫺ mice to define the roles that SOCS3 plays in regulation of cytokine actions (54); however, no effects on prolactin signaling have been reported to our knowledge. These observations of genetically modified mice complement in vitro studies on the individual physiological roles of SOCS proteins, as does the current study in which we measured induction of a range of SOCS mRNAs within the arcuate nucleus in three different physiological situations. Clearly, a combination of these approaches will be required to fully understand the essential physiological roles of SOCS proteins. We did not observe any response of SOCS-2 mRNA levels to exogenous or endogenous prolactin surges and only a very minor response to lactational hyperprolactinemia. There appears to be considerable tissue specificity in the SOCS-2 response to exogenous prolactin, because induction has been reported in adipose tissue (31), ovary (24), and leukocytes (55) but not adrenal glands or mammary tissue (24). In vitro experiments have shown SOCS-2 to either have no biological effect on (56, 57) or, interestingly, potentiate (37, 58) prolactin signaling. CIS, on the other hand, was markedly induced by ovine prolactin treatment and to a lesser magnitude by lactational hyperprolactinemia but not by the late-pregnancy prolactin surge. At present, the reason for this variability in response to prolactin is unclear. Mice overexpressing a CIS transgene displayed reduced STAT5 phosphorylation levels in mammary glands and liver, and the females failed to lactate after parturition because of a failure in terminal differentiation of the mammary glands, suggesting a defect in prolactin signaling (59). On the other hand, cis⫺/⫺ mice have no obvious defects (54). This may indicate that the contribution of CIS to cytokine regulation is largely redundant. Other less-studied SOCS proteins may also be involved in modulating prolactin regulation in late pregnancy and lactation. In this regard, SOCS-7 was recently shown to inhibit prolactin-induced STAT5 signaling in vitro (60). In conclusion, these results show that prolactin’s ability to induce STAT5b signaling in TIDA neurons is greatly suppressed by lactation, independent of suckling effects. Furthermore, a marked up-regulation of SOCS-1 and -3 mRNAs during late pregnancy and lactation may contribute to this suppression of prolactin feedback actions and the consequent reduction in TIDA neuron dopamine output that maintains hyperprolactinemia at these times. The fact that this effect is itself largely dependent on high levels of prolactin suggests that SOCS proteins contribute to, but do not initiate, this change in neuronal function. A major challenge, therefore, will be to test whether removal of SOCS-1 and/or SOCS-3 (possibly by using small interfering RNA or transgenic conditional knockout technologies) restores normal prolactin-induced STAT5b signaling and TIDA neuron do-

5004 Endocrinology, October 2006, 147(10):4996 –5005


pamine output at these times. The role of late-pregnancy changes in the concentration of circulating estradiol in inducing of SOCS-1 and -3 also warrants additional investigation.

21.

Acknowledgments

22.

Prolactin RIA reagents were provided by Dr. A. F. Parlow at the National Institute of Diabetes and Digestive and Kidney Diseases National Hormone and Pituitary program.

23.

Received June 22, 2005. Accepted July 10, 2006. 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, P.O. Box 913, Dunedin, New Zealand. E-mail: greg.anderson@ anatomy.otago.ac.nz. This work was supported by the New Zealand Marsden Fund.

24. 25.

26. 27.

References 1. DeMaria JE, Lerant AA, Freeman ME 1999 Prolactin activates all three populations of hypothalamic neuroendocrine dopaminergic neurons in ovariectomized rats. Brain Res 837:236 –241 2. Lerant A, Freeman ME 1998 Ovarian steroids differentially regulate the expression of PRL-R in neuroendocrine dopaminergic neuron populations: a double label confocal microscopic study. Brain Res 802:141–154 3. Kokay IC, Grattan DR 2005 Expression of mRNA for prolactin receptor (long form) in dopamine and pro-opiomelanocortin neurons in the arcuate nucleus of non-pregnant and lactating rats. J Neuroendocrinol 17:827– 835 4. Ben-Jonathan N, Hnasko R 2001 Dopamine as a prolactin (PRL) inhibitor. Endocr Rev 22:724 –763 5. Ma FY, Anderson GM, Gunn TD, Goffin V, Grattan DR, Bunn SJ 2005 Prolactin specifically activates signal transducer and activator of transcription 5b in neuroendocrine dopaminergic neurons. Endocrinology 146:5112–5119 6. Ma FY, Grattan DR, Goffin V, Bunn SJ 2005 Prolactin-regulated tyrosine hydroxylase activity and messenger ribonucleic acid expression in mediobasal hypothalamic cultures: the differential role of specific protein kinases. Endocrinology 146:93–102 7. Grattan DR, Xu J, McLachlan MJ, Kokay IC, Bunn SJ, Hovey RC, Davey HW 2001 Feedback regulation of PRL secretion is mediated by the transcription factor, signal transducer, and activator of transcription 5b. Endocrinology 142:3935–3940 8. Cave BJ, Norman M, Flynn A, Townsend J, Wakerley JB, Tortonese DJ 2005 Prolactin-induced activation of STAT5 within the hypothalamic arcuate nucleus. Neuroreport 16:1423–1426 9. Smith MS, Neill JD 1976 Termination at midpregnancy of the two daily surges of plasma prolactin initiated by mating in the rat. Endocrinology 98:696 –701 10. Andrews ZB, Grattan DR 2003 Opioid receptor subtypes involved in the regulation of prolactin secretion during pregnancy and lactation. J Neuroendocrinol 15:227–236 11. Tonkowicz PA, Voogt JL 1983 Termination of prolactin surges with development of placental lactogen secretion in the pregnant rat. Endocrinology 113:1314 –1318 12. Arbogast LA, Soares MJ, Tomogane H, Voogt JL 1992 A trophoblast-specific factor(s) suppresses circulating prolactin levels and increases tyrosine hydroxylase activity in tuberoinfundibular dopaminergic neurons. Endocrinology 131:105–113 13. Andrews ZB, Kokay IC, Grattan DR 2001 Dissociation of prolactin secretion from tuberoinfundibular dopamine activity in late pregnant rats. Endocrinology 142:2719 –2724 14. Grattan DR, Averill RL 1995 Absence of short-loop autoregulation of prolactin during late pregnancy in the rat. Brain Res Bull 36:413– 416 15. Fliestra RJ, Voogt JL 1997 Lactogenic hormones of the placenta and pituitary inhibit suckling-induced prolactin (PRL) release but not the ante-partum PRL surge. Proc Soc Exp Biol Med 214:258 –264 16. Demarest KT, McKay DW, Riegle GD, Moore KE 1983 Biochemical indices of tuberoinfundibular dopaminergic neuronal activity during lactation: a lack of response to prolactin. Neuroendocrinology 36:130 –137 17. Arbogast LA, Voogt JL 1996 The responsiveness of tuberoinfundibular dopaminergic neurons to prolactin feedback is diminished between early lactation and midlactation in the rat. Endocrinology 137:47–54 18. Pi XJ, Grattan DR 1999 Increased prolactin receptor immunoreactivity in the hypothalamus of lactating rats. J Neuroendocrinol 11:693–705 19. Augustine RA, Kokay IC, Andrews ZB, Ladyman SR, Grattan DR 2003 Quantitation of prolactin receptor mRNA in the maternal rat brain during pregnancy and lactation. J Mol Endocrinol 31:221–232 20. Starr R, Willson TA, Viney EM, Murray LJ, Rayner JR, Jenkins BJ, Gonda

28. 29. 30. 31.

32. 33. 34.

35. 36. 37. 38. 39.

40. 41. 42. 43. 44. 45.

46.

TJ, Alexander WS, Metcalf D, Nicola NA, Hilton DJ 1997 A family of cytokine-inducible inhibitors of signalling. Nature 387:917–921 Endo TA, Masuhara M, Yokouchi M, Suzuki R, Sakamoto H, Mitsui K, Matsumoto A, Tanimura S, Ohtsubo M, Misawa H, Miyazaki T, Leonor N, Taniguchi T, Fujita T, Kanakura Y, Komiya S, Yoshimura A 1997 A new protein containing an SH2 domain that inhibits JAK kinases. Nature 387:921– 924 Naka T, Narazaki M, Hirata M, Matsumoto T, Minamoto S, Aono A, Nishimoto N, Kajita T, Taga T, Yoshizaki K, Akira S, Kishimoto T 1997 Structure and function of a new STAT-induced STAT inhibitor. Nature 387:924 –929 Starr R, Hilton DJ 1999 Negative regulation of the JAK/STAT pathway. Bioessays 21:47–52 Tam SP, Lau P, Djiane J, Hilton DJ, Waters MJ 2001 Tissue-specific induction of SOCS gene expression by PRL. Endocrinology 142:5015–5026 Anderson ST, Barclay JL, Fanning KJ, Kusters DH, Waters MJ, Curlewis JD 2006 Mechanisms underlying the diminished sensitivity to prolactin negative feedback during lactation: reduced STAT5 signaling and up-regulation of cytokine-inducible SH2 domain-containing protein (CIS) expression in tuberoinfundibular dopaminergic neurons. Endocrinology 147:1195–1202 Frost AR, Sparks D, Grizzle WE 2000 Methods of antigen recovery vary in their usefulness in unmasking specific antigens in immunohistochemistry. Appl Immunohistochem Mol Morphol 8:236 –243 Palkovits M, Brownstein MJ 1988 Maps and guide to microdissection of the rat brain. New York: Elsevier Bustin SA 2002 Quantification of mRNA using real-time reverse transcription PCR (RT-PCR): trends and problems. J Mol Endocrinol 29:23–39 Lookingland KJ, Jarry HD, Moore KE 1987 The metabolism of dopamine in the median eminence reflects the activity of tuberoinfundibular neurons. Brain Res 419:303–310 Selmanoff M 1985 Rapid effects of hyperprolactinemia on basal prolactin secretion and dopamine turnover in the medial and lateral median eminence. Endocrinology 116:1943–1952 Ling C, Billig H 2001 PRL receptor-mediated effects in female mouse adipocytes: PRL induces suppressors of cytokine signaling expression and suppresses insulin-induced leptin production in adipocytes in vitro. Endocrinology 142:4880 – 4890 Chronwall BM 1985 Anatomy and physiology of the neuroendocrine arcuate nucleus. Peptides 6(Suppl 2):1–11 Kasagi Y, Tokita R, Nakata T, Imaki T, Minami S 2004 Human growth hormone induces SOCS3 and CIS mRNA increase in the hypothalamic neurons of hypophysectomized rats. Endocr J 51:145–154 Tups A, Ellis C, Moar KM, Logie TJ, Adam CL, Mercer JG, Klingenspor M 2004 Photoperiodic regulation of leptin sensitivity in the Siberian hamster, Phodopus sungorus, is reflected in arcuate nucleus SOCS-3 (suppressor of cytokine signaling) gene expression. Endocrinology 145:1185–1193 Baskin D, Breininger J, Schwartz M 2000 SOCS-3 expression in leptin-sensitive neurons of the hypothalamus of fed and fasted rats. Regul Pept 92:9 –15 Ram PA, Waxman DJ 1999 SOCS/CIS protein inhibition of growth hormonestimulated STAT5 signaling by multiple mechanisms. J Biol Chem 274:35553– 35561 Dif F, Saunier E, Demeneix B, Kelly PA, Edery M 2001 Cytokine-inducible SH2-containing protein suppresses PRL signaling by binding the PRL receptor. Endocrinology 142:5286 –5293 Wormald S, Hilton DJ 2004 Inhibitors of cytokine signal transduction. J Biol Chem 279:821– 824 Gutzman JH, Nikolai SE, Rugowski DE, Watters JJ, Schuler LA 2005 Prolactin and estrogen enhance the activity of activating protein 1 in breast cancer cells: role of extracellularly regulated kinase 1/2-mediated signals to c-fos. Mol Endocrinol 19:1765–1778 Baetz A, Frey M, Heeg K, Dalpke AH 2004 Suppressor of cytokine signaling (SOCS) proteins indirectly regulate toll-like receptor signaling in innate immune cells. J Biol Chem 279:54708 –54715 Demarest KT, Moore KE, Riegle GD 1983 Role of prolactin feedback in the semicircadian rhythm of tuberoinfundibular dopaminergic neuronal activity during early pregnancy in the rat. Neuroendocrinology 36:371–375 Demarest KT, Duda NJ, Riegle GD, Moore KE 1983 Placental lactogen mimics prolactin in activating tuberoinfundibular dopaminergic neurons. Brain Res 272:175–178 Arbogast LA, Voogt JL 1991 Mechanisms of tyrosine hydroxylase regulation during pregnancy: evidence for protein dephosphorylation during the prolactin surges. Endocrinology 129:2575–2582 Stoecklin E, Wissler M, Schaetzle D, Pfitzner E, Groner B 1999 Interactions in the transcriptional regulation exerted by Stat5 and by members of the steroid hormone receptor family. J Steroid Biochem Mol Biol 69:195–204 Faulds MH, Pettersson K, Gustafsson J-A, Haldosen L-A 2001 Cross-talk between ERs and signal transducer and activator of transcription 5 is E2 dependent and involves two functionally separate mechanisms. Mol Endocrinol 15:1929 –1940 Leung KC, Doyle N, Ballesteros M, Sjogren K, Watts CKW, Low TH, Leong GM, Ross RJM, Ho KKY 2003 Estrogen inhibits GH signaling by suppressing GH-induced JAK2 phosphorylation, an effect mediated by SOCS-2. Proc Natl Acad Sci USA 100:1016 –1021


47. Cao J, Wood M, Liu Y, Hoffman T, Hyde J, Park-Sarge OK, Vore M 2004 Estradiol represses prolactin-induced expression of Na⫹/taurocholate cotransporting polypeptide in liver cells through estrogen receptor-␣ and signal transducers and activators of transcription 5a. Endocrinology 145:1739 –1749 48. Mitchell V, Loyens A, Spergel DJ, Flactif M, Poulain P, Tramu G, Beauvillain JC 2003 A confocal microscopic study of gonadotropin-releasing hormone (GnRH) neuron inputs to dopaminergic neurons containing estrogen receptor ␣ in the arcuate nucleus of GnRH-green fluorescent protein transgenic mice. Neuroendocrinology 77:198 –207 49. Grattan DR, Pi XJ, Andrews ZB, Augustine RA, Kokay IC, Summerfield MR, Todd B, Bunn SJ 2001 Prolactin receptors in the brain during pregnancy and lactation: implications for behavior. Horm Behav 40:115–124 50. Lerant A, Kanyicska B, Freeman ME 2001 Nuclear translocation of STAT5 and increased expression of Fos related antigens (FRAs) in hypothalamic dopaminergic neurons after prolactin administration. Brain Res 904:259 –269 51. Bennett E, McGuinness L, Gevers EF, Thomas GB, Robinson IC, Davey HW, Luckman SM 2005 Hypothalamic STAT proteins: regulation of somatostatin neurones by growth hormone via STAT5b. J Neuroendocrinol 17:186 –194 52. Park DS, Lee H, Frank PG, Razani B, Nguyen AV, Parlow AF, Russell RG, Hulit J, Pestell RG, Lisanti MP 2002 Caveolin-1-deficient mice show accelerated mammary gland development during pregnancy, premature lactation, and hyperactivation of the Jak-2/STAT5a signaling cascade. Mol Biol Cell 13:3416 –3430 53. Lindeman GJ, Wittlin S, Lada H, Naylor MJ, Santamaria M, Zhang JG, Starr R, Hilton DJ, Alexander WS, Ormandy CJ, Visvader J 2001 SOCS1 deficiency


54. 55.

56.

57. 58.

59.

60.

5005

results in accelerated mammary gland development and rescues lactation in prolactin receptor-deficient mice. Genes Dev 15:1631–1636 Greenhalgh CJ, Alexander WS 2004 Suppressors of cytokine signalling and regulation of growth hormone action. Growth Horm IGF Res 14:200 –206 Dogusan Z, Hooghe-Peters EL, Berus D, Velkeniers B, Hooghe R 2000 Expression of SOCS genes in normal and leukemic human leukocytes stimulated by prolactin, growth hormone and cytokines. J Neuroimmunol 109:34 –39 Helman D, Sandowski Y, Cohen Y, Matsumoto A, Yoshimura A, Merchav S, Gertler A 1998 Cytokine-inducible SH2 protein (CIS3) and JAK2 binding protein (JAB) abolish prolactin receptor-mediated STAT5 signaling. FEBS Lett 441:287–291 Tomic S, Chughtai N, Ali S 1999 SOCS-1, -2, -3: selective targets and functions downstream of the prolactin receptor. Mol Cell Endocrinol 158:45–54 Pezet A, Favre H, Kelly PA, Edery M 1999 Inhibition and restoration of prolactin signal transduction by suppressors of cytokine signaling. J Biol Chem 274:24497–24502 Matsumoto A, Seki Y, Kubo M, Ohtsuka S, Suzuki A, Hayashi I, Tsuji K, Nakahata T, Okabe M, Yamada S, Yoshimura A 1999 Suppression of STAT5 functions in liver, mammary glands, and T cells in cytokine-inducible SH2containing protein 1 transgenic mice. Mol Cell Biol 19:6396 – 6407 Martens N, Uzan G, Wery M, Hooghe R, Hooghe-Peters EL, Gertler A 2005 Suppressor of cytokine signaling 7 inhibits prolactin, growth hormone, and leptin signaling by interacting with STAT5 or STAT3 and attenuating their nuclear translocation. J Biol Chem 280:13817–13823

Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.