Reproductive cycles are generated through a complex

Transcription

1 NEUROENDOCRINOLOGY Estrogen-Negative Feedback and Estrous Cyclicity Are Critically Dependent Upon Estrogen Receptor- Expression in the Arcuate Nucleus of Adult Female Mice Shel-Hwa Yeo and Allan E. Herbison Centre for Neuroendocrinology and Department of Physiology, University of Otago School of Medical Sciences, Dunedin 9054, New Zealand The location and characteristics of cells within the brain that suppress GnRH neuron activity to contribute to the estrogen-negative feedback mechanism are poorly understood. Using adenoassociated virus (AAV)-mediated Cre-LoxP recombination in estrogen receptor- (ER ) floxed mice (ER flox/flox ), we aimed to examine the role of ER -expressing neurons located in the arcuate nucleus (ARN) in the estrogen-negative feedback mechanism. Bilateral injection of AAV-Cre into the ARN of ER flox/flox mice (n 14) resulted in the time-dependent ablation of up to 99% of ER -immunoreactive cell numbers throughout the rostrocaudal length of the ARN. These mice were all acyclic by 5 weeks after AAV-Cre injections with most mice in constant estrous. Control wild-type mice injected with AAV-Cre (n 13) were normal. Body weight was not altered in ER flox/flox mice. After ovariectomy, a significant increment in LH secretion was observed in all genotypes, although its magnitude was reduced in ER flox/flox mice. Acute and chronic estrogennegative feedback were assessed by administering 17 -estradiol to mice as a bolus (LH measured 3 h later) or SILASTIC brand capsule implant (LH measured 5 d later). This demonstrated that chronic estrogen feedback was absent in ER flox/flox mice, whereas the acute feedback was normal. These results reveal a critical role for ER -expressing cells within the ARN in both estrous cyclicity and the chronic estrogen negative feedback mechanism in female mice. This suggests that ARN cells provide a key indirect, transsynpatic route through which estradiol suppresses the activity of GnRH neurons. (Endocrinology 155: , 2014) Reproductive cycles are generated through a complex interplay of gonadal steroid positive and negative feedback mechanisms that act within the brain and pituitary gland to modulate the episodic secretion of LH and FSH. Estradiol secreted from the ovary plays a critical role in both feedback mechanisms by coordinating the activity of the GnRH neurons with ovarian status. Considerable work has been undertaken over the last 40 years aimed at elucidating the nature of estrogen feedback mechanisms (1). Although the understanding of estrogen-positive feedback has advanced considerably over this period (2, 3), less is known about the pathways involved in the negative feedback mechanism. It has been appreciated for some time that estradiol acts within both the brain and anterior pituitary gland to bring about negative feedback (4 7). With respect to the brain, investigations in rodents, sheep, and primates have all indicated that estradiol inhibits the frequency and/or amplitude of GnRH pulses secreted into the median eminence (8 13). However, the brain regions, cell types, and estrogen receptors (ERs) involved in suppressing the activity of GnRH neurons remain obscure. Initial investigations in the 1970s revealed an important role for the pituitary gland in acute negative feedback alongside more limited evidence for a role of cells within the mediobasal hypothalamus (MBH). Rats that were ren- ISSN Print ISSN Online Printed in U.S.A. Copyright 2014 by the Endocrine Society Received February 11, Accepted April 18, First Published Online June 6, 2014 Abbreviations: AAV, adeno-associated virus; ARN, arcuate nucleus; AVPV, anteroventral periventricular nucleus; ER, estrogen receptor; MBH, mediobasal hypothalamus; OVX, ovariectomized; POMC, proopiomelanocortin; TBS, Tris-buffered saline; VMN, ventromedial nucleus endo.endojournals.org Endocrinology, August 2014, 155(8): doi: /en

2 doi: /en endo.endojournals.org 2987 dered acyclic after the complete or partial surgical deafferentation of the MBH were nevertheless still able to exhibit an acute suppression of LH after parenteral estradiol treatment (14, 15). Although the suppression of LH in those animals was mostly accounted for by actions of estradiol at the pituitary, it was argued that the MBH may also have a role. More refined studies using implants of estradiol crystals into specific hypothalamic brain regions showed that various sites within the MBH were able to reduce LH levels or, where examined, LH pulse amplitude (16 20). However, a major concern with some of these studies was that estradiol had simply diffused to the nearby pituitary gland to suppress LH secretion (5). Also, the MBH is not the only brain region implicated in the estrogen-negative feedback response, and different brain sites may even be involved in a state-dependent manner (18, 20). Nevertheless, recent interest in the kisspeptin control of GnRH neurons has served to refocus attention on the role of the arcuate nucleus (ARN) within the MBH in estrogen-negative feedback (21 23). With respect to the types of ERs involved, genetic studies in the mouse have shown that the ER isoform is critical for the negative feedback response (24 27). Again, this appears to be the case both at the pituitary gland (28, 29) and within neurons (30). Together the lesioning, implant, genetic, and kisspeptin studies suggest that ER within the ARN may be involved in the estrogen-negative feedback mechanism. Recent advances combining genetically engineered mice with stereotaxic injections of adeno-associated viruses (AAVs) have enabled the roles of specific genes located in defined brain regions to be investigated (31). Here we use this approach to ablate ER within the ARN of female mice to provide a definitive account of the role of ARN ER -expressing neurons in estrous cyclicity and the estrogen negative feedback mechanism. Materials and Methods Animals Adult female C57BL/6 wild-type, ER flox/flox (32) and RO- SA26-CAGS- GFP (er26- GFP) reporter (33) mice were housed under a 12-hour lighting schedule (lights on 6:00 AM and off at 6:00 PM) with ad libitum access to food and water. All experiments were approved by the University of Otago Welfare and Ethics Committee. AAV injections A previously characterized (34, 35) recombinant AAV1/2- CBA-WPRE-bGH harboring Cre recombinase (AAV-Cre; viral genomes per microliter; Associate Professor Deborah Young, Department of Pharmacology, University of Auckland, New Zealand) was injected bilaterally into the mouse ARN. In preliminary experiments, 0.5, 1.5, and 3.0 L volumes of AAV-Cre were injected into the ARN of er26- GFP reporter or ER flox/flox mice to assess the time course and extent of AAV transfection and Cre recombination. For actual experiments, 28 adult female ER flox/flox and 22 control C57BL/6 mice received bilateral 3.0 L AAV-Cre injections into the ARN. Animals were given a sc injection of Carprofen (5 mg/kg body weight), anesthetized with 2% isoflurane, and placed in a stereotaxic apparatus (Stoelting). A custom-made bilateral Hamilton syringe apparatus (Leo Van Rens, EMTech, University of Otago, Dunedin, New Zealand) consisting of two 29-gauge needles (Hamilton) 0.8 mm apart was used to perform bilateral injections. Each syringe was filled with 3.0 L of AAV-Cre. A small portion of the skull above the intended injection site was removed with a drill. The bilateral syringe needles were then lowered to the mid rostrocaudal aspect of the ARN [1.2 mm posterior to bregma, midline, 5.9 mm ventral to dura, according to the atlas of Paxinos and Franklin (2001)]. The bilateral injections of the AAV-Cre were performed over a 15-minute period and the syringes left in situ for a further 15 minutes before being removed from the brain. Estrous cycle monitoring and negative feedback regimen Daily vaginal smears were collected between 10:30 to 11:30 AM from all animals beginning 1 week after the AAV-Cre injections and continued for the next 4 weeks. To provide a temporal representation of cyclical activity in the mice, the mean number of estrous cycles exhibited by ER flox/flox and control animals was calculated on a weekly basis by counting the number of proestrous smear days occurring in each 7-day period. The total number of days with estrous smears over the entire 4-week period was also determined. Statistical analysis was performed by nonparametric Mann-Whitney tests. An experimental protocol that examined acute and then chronic estrogen-negative feedback in the same mice was used (Figure 1). One day prior to the AAV-Cre injections, mice were anesthetized with 2% isoflurane, weighed, and the first tail blood sample obtained. Six weeks after the AAV-Cre injections, mice were anesthetized with isoflurane, weighed, a second tail blood sample drawn, and the mice ovariectomized (OVX). One week after OVX, the third tail blood sample was collected. The mice were then allowed to recover for 2 days before testing acute negative feedback by administering 17 -estradiol (1 g per 20 g body weight, sc; Sigma-Aldrich) in the morning. This regimen suppresses the LH levels in the OVX animals within 3 hours (30, 36). Three hours after 17 -estradiol, the mice were anesthetized and fourth tail blood sample was collected before implanting a sc SILASTIC brand capsule containing 17 -estradiol (1 g per Figure 1. Schematic diagram indicating the experimental protocol used for assessing estrous cyclicity and estrogen feedback in mice. Arrows indicate when tail-tip blood samples were collected.

3 2988 Yeo and Herbison ER in ARN Mediates Estrogen-Negative Feedback Endocrinology, August 2014, 155(8): g body weight). Five days after capsule implantation, the mice were anesthetized, a blood sample obtained, and perfusion fixed with 4% paraformaldehyde (see below). Blood samples were centrifuged at 4 C for 10 minutes and plasma extracted and stored at 20 C until a RIA was performed. RIAs were undertaken using the anti-rlh-s-11 antiserum and mouse LH-RP provided by Dr A. F. Parlow (National Hormone and Peptide Program, Torrance, California). The sensitivity of the RIA was 0.1 ng/ml with intra- and interassay coefficients of variation being 9.8% and 9.5%, respectively. Statistical analysis was undertaken using a repeated-measures ANOVA with post hoc Dunn s multiple comparison tests. Immunohistochemistry Animals were anesthetized with an overdose of pentobarbital (3 mg per 100 L) and transcardially perfused with 15 ml of 4% paraformaldehyde. The brains and pituitaries were removed from the skull and postfixed in the same fixative at room temperature for 1 hour. Brains were cryoprotected at 4 C overnight in 30% sucrose/tris-buffered saline (TBS). Three sets of 50- mthick coronal brain sections were then cut from the level of the medial septum through to the supramammillary nucleus for freefloating immunohistochemistry. The pituitaries were washed with TBS and transferred to 70% ethanol overnight at 4 C. After dehydration, the pituitaries were cryosectioned at a 15 m thickness for slide-mounted immunohistochemistry. Single-label immunohistochemistry was conducted to detect Cre recombinase and/or ER expression in the brain or pituitary (Table 1). The specificity of the polyclonal rabbit antisera raised against Cre recombinase (Professor G. Schütz, German Cancer Research Centre, Heidelberg, Germany) has been described previously (37, 38) as has the polyclonal rabbit anti-er antibody (Millipore) (30, 39). The omission of either primary antisera from this protocol resulted in a complete absence of the respective immunoreactivity. Sections were treated with 3% hydrogen peroxide to quench endogenous peroxidases and then washed in TBS. One set of sections (free floating or slide mounted) was then incubated in the Cre (1:10 000) or ER (1:10 000) antisera for 48 hours at 4 C. For secondary antibody labeling, sections were incubated with biotinylated goat antirabbit immunoglobulins (1:200; Vector) at room temperature. Sections were then incubated with Vector Elite avidin-peroxidase (1:100; Vector Laboratories). Finally, the sections were rinsed and immunoreactivity was revealed with glucose-oxidase and nickel-enhanced diaminobenzidine hydrochloride. The brain sections were mounted on slides, air dried, dehydrated in ethanol followed by xylene, and then coverslipped with DPX. Quantitative analysis of ER immunoreactivity The ARN of the mouse is approximately 1.5 mm in length and for analysis was divided into three rostrocaudal levels: rostral, middle, and caudal (Figure 2). Brain sections were first analyzed for Cre and only those animals exhibiting bilateral Cre expression within the ARN were included in the study and analyzed further. Digital images of ER immunoreactivity were then obtained from two sections at each of the three rostrocaudal ARN levels from each mouse at 10 magnification on a BX51 Olympus microscope. Image J (Rasband, ; National Institutes of Health, Bethesda, Maryland) was used for processing and analysis of the captured images. A square box (0.25 mm 2 ) was overlaid on the ARN to quantify the number of immunoreactive nuclei on each side of the brain. These four values (two sections at each rostrocaudal level and from each side of the brain) were then combined to provide a mean value for each level of the ARN in that mouse. Values were combined from experimental groups to generate mean SEM values with statistical analysis performed using nonparametric Mann-Whitney tests. The numbers of ER -immunoreactive cells were also determined within the ventrolateral division of the ventromedial nucleus (VMN), anteroventral periventricular nucleus (AVPV), and pituitary. Two sections representing each of these brain regions or pituitary were analyzed in each mouse using the same method described above for the ARN. Results Preliminary evaluation of Cre and ER expression after unilateral AAV-Cre injections Initially we determined the transduction efficiency of the AAV-Cre by injecting either 0.5 Lor1.5 L volumes unilaterally into the ARN of er26- GFP reporter mice (n 6) and examined both Cre and GFP expression 4 weeks later. Prior studies have shown that AAV-mediated gene expression reaches optimal levels after 3 4 weeks, and it is stable for at least 18 months without incurring immunogenicity or inflammation (40). Both volumes of AAV-Cre resulted in the appearance of Cre and GFP immunoreactivity within the ARN. However, even with the largest volume (1.5 L), we estimated that transfection occurred only in cells located over a rostrocaudal distance of approximately 0.8 mm within the ARN. Because the ARN is approximately 1.5 mm in length, we subsequently Table 1. Details of Antisera Used in Study Peptide/Protein Target Antigen Sequence (If Known) Name of Antibody Manufacturer, Catalog #, and/or Name of Individual Providing the Antibody Species Raised in; Monoclonal or Polyclonal Dilution Used Estrogen receptor TYYIPPEAEGFPNTI Millipore Polyclonal rabbit 1 in 10,000 Cre recombinase Cre Prof. Gunther Scheutz, Polyclonal rabbit 1 in 10,000 Heidleberg, Germany

4 doi: /en endo.endojournals.org 2989 Figure 2. Expression of Cre and ER in control mice injected with AAV-Cre bilaterally into the ARN. Cre immunoreactivity (left panels) and ER immunoreactivity (right panels) in the ARN of a control mouse at rostral, middle, and caudal levels of the ARN are shown. Scale bars, 200 m. used 3.0 L volumes, and this was found to transfect the cells along the complete rostrocaudal length of the ARN in addition to cells adjacent to the ARN (see Figure 2). Next, unilateral AAV-Cre injections were made into the ARN of ER flox/flox mice to determine the postsurgical period for effective Cre-mediated ER ablation. Mice were perfused at different time points (2, 3, and 4 wk after AAV-Cre injections) to determine the duration for effective ER ablation. Two weeks after the AAV-Cre injection, Cre recombinase expression was observed in the ARN of the animals, but no ER ablation was detected (n 2). At 3 weeks a small reduction in ER immunoreactivity was observed (n 3). However, at 4 weeks after the AAV-Cre injections, effective ablation of most ER was observed throughout the ARN (n 2). Injections of AAV-Cre into control wild-type mice resulted in Cre expression but no ablation of ER (see Figure 2). Together these studies indicated that bilateral 3.0 L injections of AAV-Cre were required for a period of greater than 4 weeks to ablate most ER protein in the ARN. Analysis of Cre and ER in experimental mice receiving bilateral AAV-Cre injections Experimental mice underwent immunohistochemical analysis of Cre and ER 8 weeks after AAV-Cre injection and functional evaluation of their negative feedback. Fourteen of 28 ER flox/flox animals and 13 of 22 control wild-type animals exhibited bilateral Cre expression in the ARN. The remaining animals showed unilateral or no Cre expression in the ARN and were not analyzed further. In control mice, Cre expression was detected within medial aspects of the basal hypothalamus extending from the rostral to caudal aspects of the ARN alongside a normal distribution of ER immunoreactivity (Figure 2). The same pattern of Cre expression was detected in ER flox/flox mice (Figure 3, A and B), although in these mice the density of ER -expressing cells was markedly reduced throughout the entire rostrocaudal length of the ARN (Figure 3, A and B). Some mice exhibited differential bilateral Cre expression at the rostral or caudal levels of the ARN (for example, see Figure 3A, rostral ARN) in which the immunolableling for Cre and ER in the adjacent sections demonstrated the near perfect correspondence between Cre expression and the relative absence of ER. Quantitative analysis of ER expression revealed individual animals with 45%-99% decrements in ER -immunoreactive cell numbers at the three different levels of the ARN. Overall, mean 60% 16%, 78% 14%, and 91% 8% decreases in ER -immunoreactive cell numbers were detected in the rostral, middle, and caudal aspects of the ARN in ER flox/flox mice (n 14, P.001; Mann-Whitney tests; Figure 4A). To assess the degree to which ER was altered in ER -expressing cells near the ARN, the VMN, AVPV, and anterior pituitary were also analyzed. This revealed a significant 29% decrease in ER -immunoreactive nuclei in the VMN (n 14; P.01; Mann-Whitney tests; Figure 4B) and no effects in the AVPV (n 14; Figure 4C) or anterior pituitary (n 6; Figure 4D). Estrous cyclicity The estrous cycles of wild-type control (n 13) and ER flox/flox (n 14) mice with bilateral AAV-Cre injections were assessed by daily vaginal smear for 4 weeks, beginning 1 week after the AAV-Cre injection (Figure 1). Control animals exhibited normal 4- to 6-day estrous cycles throughout this period (Figure 5, A, and D). In contrast, ER flox/flox mice exhibited normal cycles for up to 2 weeks before becoming acyclic (Figure 5, B D). This was typified by prolonged periods of estrus in most mice (Figure 5B) with some (Figure 5C) showing an interspersed occasional cycle or reversal to constant diestrus. Quantitative analyses showed that control animals had cycles over the whole 4-week period with 45% 3% of time spent in estrus. ER flox/flox mice spent 51% 7% time in estrus and showed a significantly lower cycle number of days over the whole 4-week period (P

5 2990 Yeo and Herbison ER in ARN Mediates Estrogen-Negative Feedback Endocrinology, August 2014, 155(8): Figure 3. Ablation of ER expression in two ER flox/flox mice injected with AAV-Cre bilaterally into the ARN. A, Cre immunoreactivity (left panels) and ER immunoreactivity (right panels) in adjacent sections of the ARN at rostral, middle, and caudal levels. Note the inverse relationship between Cre and ER immunoreactivities in the rostral ARN of this mouse. B, Cre and ER immunoreactivities in another ER flox/flox mouse. Scale bars, 200 m..001, Mann-Whitney test). When examined on a weekly basis, it was apparent that the numbers of cycles per week exhibited by ER flox/flox mice steadily dropped from normal levels in week 2 to almost complete acyclicity in week 5 (Figure 5D). Body weight was not different between genotypes with the weight gain over the 6-week period between AAV-Cre injection and OVX being g for wild-type mice and g for ER flox/flox mice. Estrogen-negative feedback Estrogen-negative feedback was assessed in wild-type control (n 13) and ER flox/flox (n 14) mice using the protocol schematized in Figure 1. Control and ER flox/flox mice exhibited normal suppressed levels of LH before any manipulations ( vs ng/ml) as well as 6 weeks after the AAV-Cre injections ( vs ng/ml; Figure 6). One week after OVX, the LH levels were elevated significantly in both groups (controls, ng/ml; ER flox/flox, ng/ml; P.001; ANOVA with Dunn s multiple comparison test in both cases; Figure 6). The increment in LH was significantly greater in the control mice compared with the ER flox/flox mice (P.001; Bonferroni post hoc test). Acute estrogen-negative feedback was examined by treating mice with a single sc injection of 17 -estradiol and evaluating the LH levels 3 hours later. In both genotypes, this resulted in a significant return of LH to basal levels ( ng/ml; Figure 6). Chronic estrogen-negative feedback was assessed by then implanting mice with estradiol capsules and examining LH levels 5 days later. In control mice (n 7), LH levels remained suppressed ( ng/ml), whereas in ER flox/flox mice (n 6), the levels were significantly elevated ( ng/ml; P.001; ANOVA with Dunn s multiple comparison test; Figure 6). No significant correlation was found between the degree of ER depletion (range 74%-99%) in the ARN and increment in LH levels in individual mice. Discussion We report here that stereotaxic injection of AAV-Cre into the ARN of ER flox/flox mice results in a mean 60%-90% decrease in ER expression within this nucleus and generates mice that are acyclic with disrupted estrogen-neg-

6 doi: /en endo.endojournals.org 2991 Figure 4. ER ablation in control and ER flox/flox mice after bilateral AAV-Cre injections into the ARN. A, The mean ( SEM) number of ER -expressing cells per section from rostral, middle, and caudal ARN of control and ER flox/flox animals. B D, Histograms indicating the mean number ( SEM) of ER -immunoreactive cells per section in the VMN (B), AVPV (C), and anterior pituitary gland (D) of ER flox/flox and control animals. ***, P.001; **, P.01; Mann-Whitney tests. ative feedback. This provides direct evidence that ER expressing cells in the ARN are involved in the cyclical regulation and estrogen-negative feedback suppression of GnRH neuron activity. We find that the AAV1/2 used in the present study is very effective at transducing large numbers of ARN cells and driving Cre expression as soon as 2 weeks after injection. The CBA promoter used in this AAV construct is known to preferentially target neurons (34, 40, 41), but we cannot exclude the possibility that nonneuronal cells are also transfected. It took a further 2 weeks before most ER protein in the ARN was ablated. This time scale of ER depletion in ER flox/flox mice correlated well with the gradual loss of estrous cycles that first became evident at 3 weeks after AAV injection and reached complete acylcicity by 5 weeks. The cycles and reproductive physiology of control mice injected with the AAV were normal, indicating that the transfection of ARN cells by the AAV itself had no effect on their functioning. Because of the unusual anatomy of the ARN, we found that we needed to inject large volumes of AAV to achieve ER deletion throughout the rostrocaudal length of the nucleus. This resulted in cells adjacent to the ARN also being transfected and, indeed, caused an approximately 30% loss of ER -expressing cells within the ventrolateral VMN. However, it is very unlikely that the reproductive phenotype reported here results from the loss of ER within the VMN. First, there is no evidence that VMN neurons are involved in the GnRH neuronal network and, second, the 29% reduction in ER is unlikely to be sufficient to generate a phenotype. Previous studies have shown that 50%-70% reductions in ER in the VMN are required to alter the regulation of body weight and female reproductive behavior (42 44). We did not find any influence of ER loss on body weight. Most importantly, the present experimental strategy does not impact on ER -expressing cells located in the pituitary gland or the AVPV, both regions known to be critical for the control of gonadotropin secretion. The basal levels of LH secretion were observed here to be unchanged in intact acyclic mice with ARN ER deletion. This result is reminiscent of cellspecific ER knockout studies in mice in which the deletion of ER from the brain or pituitary gland also has no effect on basal LH secretion (28 30). It thought that the deletion of ER from one tissue can have no effect on LH levels because the other can compensate to maintain secretion in check. This is also likely to be the case here in which deletion of ER from the ARN leaves other brain regions and the pituitary intact to maintain suppressed LH secretion. After ovariectomy, however, when gonadal steroids are removed from all sites, a significant elevation in LH secretion was observed in both controls and ARN ER -deleted mice. Interestingly, and again as seen in other neuronal genetic ER ablation models (30), the increment in LH secretion after ovariectomy was reduced in ARN ER -deleted mice. The reasons for this are unclear but may represent a reorganization of the GnRH neuronal network. One possible explanation is that, in the absence

7 2992 Yeo and Herbison ER in ARN Mediates Estrogen-Negative Feedback Endocrinology, August 2014, 155(8): Figure 5. Estrous cycles of control and ER flox/flox mice after bilateral AAV-Cre injections into the ARN. A C, Time lines showing representative estrous cycle patterns for a control (A) and two ER flox/flox mice (B and C). D, diestrus; E, estrus; M, metestrus; P, proestrus. D, Estrous cyclicity, indicated by mean ( SEM) number of proestrous smears per week, from the second to fifth week after bilateral AAV injection into the ARN of wild-type (open boxes) and ER flox/flox mice (filled boxes). of restraint from ARN ER neurons, the network rearranges to use steroid-independent mechanisms to suppress GnRH secretion. In this case, the increment in GnRH secretion after OVX would be muted compared with the normal mice. It is notable that ARN ER -deleted mice exhibited normal basal levels of LH despite exhibiting acyclicity associated with, typically, constant estrus. The reasons for this are unknown at present. Unfortunately, it was not possible to examine the pulsatile profile of LH secretion in these mice because this may have revealed abnormalities in pulse dynamics that could underlie the acyclicity; this can not be assessed from single-point LH measurements. Although roles for the ARN in the estrogen-positive feedback mechanism have been discounted consistently (2), it remains possible that the acyclic, constant estrus of ARN ER -deleted mice could result from a positive feedback defect. The acute bolus injection of estradiol was found to suppress LH secretion to basal levels in both control and experimental mice. It has previously been argued that estrogen-negative feedback is a multimodal mechanism with several different pathways operating in different time frames to restrain the activity of GnRH neurons (45). For example, there is evidence that acute estrogen feedback uses nonclassical estrogen signaling mechanisms including the rapid phosphorylation of camp response elementbinding protein and other direct and indirect actions on GnRH neurons (27, 46 49). Thus, although ER expression within the brain is involved in the acute negative feedback mechanism (30), it seems that the key neurons involved in this process are not located in the ARN. In contrast to acute feedback, the ability of estradiol to suppress LH secretion several days after ovariectomy is absent in ARN ER -deleted mice. This demonstrates that ER -expressing neurons in the ARN are required for chronic estrogen suppression of LH secretion. Early data from the rat indicated that the estrogen-sensitive cells mediating negative feedback were located within the MBH (17, 18, 20), and more recent studies have shown a key role for the ER isoform in negative feedback in the mouse (24 27). The present study uses the latest technology to

8 doi: /en endo.endojournals.org 2993 Figure 6. Acute and chronic estrogen-negative feedback in control (open diamonds) and ER flox/flox (filled diamonds) mice after bilateral AAV-Cre injections into the ARN. Mean ( SEM) plasma LH levels are shown at each of the five blood sampling intervals (arrows). Acute estradiol indicates a single sc injection of 17 -estradiol with blood sample taken 3 hours later; estradiol capsule indicates a 17 -estradiolfilled capsule with blood sample taken 5 days later. ***, P.001; **, P.01, one-way ANOVA with Dunn s multiple comparison test. bring these two findings together to provide compelling evidence that it is ER -expressing cells located within the ARN that are critical for negative feedback. We noted that the levels of LH measured in ovariectomized ARN ER deleted mice implanted with an estradiol capsule were as high as those of a 1- or 2-week OVX wild-type mouse (30). This surprising observation suggests that ARN ER -expressing cells must play a dominant role in the chronic estrogen feedback mechanism because no other estrogensensitive brain or pituitary sites appear to be able to suppress LH secretion. The phenotype of the ARN cells with ablated ER expression responsible for altered cyclicity and negative feedback is unknown. The AAV strategy achieved a 60%- 90% ablation of ER throughout the ARN that clearly involved the key cells required for negative feedback. A recent study aimed at examining the effects of ablating ER from proopiomelanocortin (POMC) neurons found that LH and FSH did not rise after ovariectomy and that, although exhibiting cycles, a small decrease in the length of the estrous phase occurred (44). Unfortunately, that study used a POMC-Cre mouse line that expresses Cre throughout the brain and not just in ARN POMC neurons (50, 51), so it is very difficult to locate the cells responsible for the phenotype. This is emphasized by the absence of any body weight defects in ARN ER -deleted mice reported here, suggesting that the metabolic deficits reported in POMC-deleted mice result from ER deletion outside the ARN. A recent genetic investigation has reported that ER deletion from kisspeptin neurons generates acyclic mice with very advanced puberty and elevated juvenile LH levels (52). Although these mice are not able to inform on the adult estrogen-negative feedback mechanism, they indicate that ER -expressing kisspeptin neurons are involved in the pubertal restraint on LH release, and it has been argued that this depends specifically upon the ARN population of kisspeptin neurons. It is also important to note that the ablation of ARN kisspeptin neurons in adult rats results in a reduced increment in LH after ovariectomy (21), as found here with ER ablation from the mouse ARN, although estradiol remains able to suppress LH secretion in those rats. As yet, definitive evidence supporting a role for ARN kisspeptin neurons in the adult estrogen negative feedback mechanism has not been forthcoming (22, 53, 54) and is complicated, experimentally, by a putative role for these cells in GnRH pulse generation (4, 55). Thus, it is possible that part of the reproductive phenotype reported here is attributable to ARN kisspeptin neurons expressing ER, but this remains to be proven. In summary, we report here that mouse estrous cyclicity and chronic estrogen-negative feedback require ER expression in the ARN. In contrast, the acute negative feedback response is independent of these cells. These observations help establish the mechanism of estrogen negative feedback in the mouse by indicating that likely ER -expressing afferent inputs from the ARN play a key estrogendependent role in controlling GnRH neuron activity to maintain suppressed gonadotropin secretion. Acknowledgments Associate Professor Deborah Young (University of Auckland) is thanked for providing the AAV construct. Dr A. Parlow (National Hormone and Peptide Program, Torrance California) is thanked for the RIA reagents. Mr Leo Van Rens (EMTech, University of Otago) is thanked for expert technical assistance. Address all correspondence and requests for reprints to: Professor Allan E. Herbison, Centre for Neuroendocrinology, Department of Physiology, University of Otago School of Medical Sciences, PO Box 913, Dunedin 9054, New Zealand. allan.herbison@otago.ac.nz. This work was supported by the New Zealand Health Research Council. Disclosure Summary: The authors have nothing to disclose. References 1. Freeman MC. Neuroendocrine control of the ovarian cycle of the rat. In: Neill JD, ed. Knobil and Neill s Physiology of Reproduction. San Diego: Academic Press; 2006:

9 2994 Yeo and Herbison ER in ARN Mediates Estrogen-Negative Feedback Endocrinology, August 2014, 155(8): Herbison AE. Estrogen positive feedback to gonadotropin-releasing hormone (GnRH) neurons in the rodent: the case for the rostral periventricular area of the third ventricle (RP3V). Brain Res Rev. 2008;57: Christian CA, Moenter SM. The neurobiology of preovulatory and estradiol-induced gonadotropin-releasing hormone surges. Endocr Rev. 2010;31: Herbison AE. Physiology of the gonadotropin-releasing hormone neuronal network. In: Neill JD, ed. Knobil and Neill: The Physiology of Reproduction. San Diego: Academic Press; 2006, Goodman RL, Knobil E. The sites of action of ovarian steroids in the regulation of LH secretion. Neuroendocrinology. 1981;32: Shupnik MA. Gonadal hormone feedback on pituitary gonadotropin genes. Trends Endocrinol Metab. 1996;7: Petersen SL, Ottem EN, Carpenter CD. Direct and indirect regulation of gonadotropin-releasing hormone neurons by estradiol. Biol Reprod. 2003;69: Sarkar DK, Fink G. Luteinizing hormone releasing factor in pituitary stalk plasma from long-term ovariectomized rats: effects of steroids. J Endocrinol. 1980;86: Caraty A, Locatelli A, Martin GB. Biphasic response in the secretion of gonadotrophin-releasing hormone in ovariectomized ewes injected with oestradiol. J Endocrinol. 1989;123: Chongthammakun S, Terasawa E. Negative feedback effects of estrogen on luteinizing hormone-releasing hormone release occur in pubertal, but not prepubertal, ovariectomized female rhesus monkeys. Endocrinology. 1993;132: Evans NP, Dahl GE, Glover BH, Karsch FJ. Central regulation of pulsatile gonadotropin-releasing hormone (GnRH) secretion by estradiol during the period leading up to the preovulatory GnRH surge in the ewe. Endocrinology. 1994;134: Levine JE, Duffy MT. Simultaneous measurement of luteinizing hormone (LH)-releasing hormone, LH, and follicle-stimulating hormone release in intact and short-term castrate rats. Endocrinology. 1988;122: Meredith JM, Levine JE. Effects of castration on LH-RH patterns in intrahypophysial microdialysates. Brain Res. 1992;571: Blake CA, Norman RL, Sawyer CH. Localization of inhibitory actions of estrogen and nicotine on release of luteinizing hormone in rats. Neuroendocrinology. 1974;16: Blake CA. A medial basal hypothalamic site of synergistic action of estrogen and progesterone on the inhibition of pituitary luteinizing hormone release. Endocrinology. 1977;101: Ferin M, Carmel PW, Zimmerman EA, Warren M, Perez R, Vande Wiele RL. Location of intrahypothalamic estrogen-responsive sites influencing LH secretion in the female Rhesus monkey. Endocrinology. 1974;95: Smith ER, Davidson JM. Location of feedback receptors: effects of intracranially implanted steroids on plasma LH and LRF response. Endocrinology. 1974;95: Nagatani S, Tsukamura H, Maeda K. Estrogen feedback needed at the paraventricular nucleus or A2 to suppress pulsatile luteinizing hormone release in fasting female rats. Endocrinology. 1994;135: Caraty A, Fabre-Nys C, Delaleu B, et al. Evidence that the mediobasal hypothalamus is the primary site of action of estradiol in inducing the preovulatory GnRH surge in the ewe. Endocrinology. 1998;139: Akema T, Tadokoro Y, Kawakami M. Changes in the characteristics of pulsatile LH secretion after estradiol implantation into the preoptic area and the basal hypothalamus in ovariectomized rats. Eondocrinol Jap. 1983;30: Mittelman-Smith MA, Williams H, Krajewski-Hall SJ, et al. Arcuate kisspeptin/neurokinin B/dynorphin (KNDy) neurons mediate the estrogen suppression of gonadotropin secretion and body weight. Endocrinology. 2012;153: de Croft S, Piet R, Mayer C, Mai O, Boehm U, Herbison AE. Spontaneous kisspeptin neuron firing in the adult mouse reveals marked sex and brain region differences but no support for a direct role in negative feedback. Endocrinology. 2012;153: Popa SM, Clifton DK, Steiner RA. The role of kisspeptins and GPR54 in the neuroendocrine regulation of reproduction. Annu Rev Physiol. 2008;70: Couse JF, Yates MM, Walker VR, Korach KS. Characterization of the hypothalamic-pituitary-gonadal axis in estrogen receptor (ER) null mice reveals hypergonadism and endocrine sex reversal in females lacking ER but not ER. Mol Endocrinol. 2003;17: Dorling AA, Todman MG, Korach KS, Herbison AE. Critical role for estrogen receptor alpha in negative feedback regulation of gonadotropin-releasing hormone mrna expression in the female mouse. Neuroendocrinology. 2003;78: Wersinger SR, Haisenleder DJ, Lubahn DB, Rissman EF. Steroid feedback on gonadotropin release and pituitary gonadotropin subunit mrna in mice lacking a functional estrogen receptor. Endocrine. 1999;11: Glidewell-Kenney C, Hurley LA, Pfaff L, et al. Nonclassical estrogen receptor signaling mediates negative feedback in the female mouse reproductive axis. Proc Natl Acad Sci USA. 2007;104: Singh SP, Wolfe A, Ng Y, et al. Impaired estrogen feedback and infertility in female mice with pituitary-specific deletion of estrogen receptor (ESR1). Biol Reprod. 2009;81: Gieske MC, Kim HJ, Legan SJ, et al. Pituitary gonadotroph estrogen receptor- is necessary for fertility in females. Endocrinology. 2008; 149: Cheong RY, Porteous R, Chambon P, Abraham IA, Herbison AE. Effects of neuron-specific ER and ER deletion on the acute estrogen negative feedback mechanism in adult female mice. Endocrinology., 2014;155: Betley JN, Sternson SM. Adeno-associated viral vectors for mapping, monitoring, and manipulating neural circuits. Hum Gene Ther. 2011;22: Wintermantel TM, Campbell RE, Porteous R, et al. Definition of estrogen receptor pathway critical for estrogen positive feedback to gonadotropin-releasing hormone neurons and fertility. Neuron. 2006;52: Mayer C, Boehm U. Female reproductive maturation in the absence of kisspeptin/gpr54 signaling. Nat Neurosci. 2011;14: Klugmann M, Symes CW, Leichtlein CB, et al. AAV-mediated hippocampal expression of short and long Homer 1 proteins differentially affect cognition and seizure activity in adult rats. Mol Cell Neurosci. 2005;28: Monory K, Massa F, Egertova M, et al. The endocannabinoid system controls key epileptogenic circuits in the hippocampus. Neuron. 2006;51: Campbell RE, Ducret E, Porteous R, et al. Gap junctions between neuronal inputs but not gonadotropin-releasing hormone neurons control estrous cycles in the mouse. Endocrinology. 2011;152: Kellendonk C, Tronche F, Casanova E, Anlag K, Opherk C, Schutz G. Inducible site-specific recombination in the brain. J Mol Biol. 1999;285: Casanova E, Fehsenfeld S, Mantamadiotis T, et al. A CamKII icre BAC allows brain-specific gene inactivation. Genesis. 2001;31: Moffatt CA, Rissman EF, Shupnik MA, Blaustein JD. Induction of progestin receptors by estradiol in the forebrain of estrogen receptor- gene-disrupted mice. J Neurosci. 1998;18: Burger C, Gorbatyuk OS, Velardo MJ, et al. Recombinant AAV viral vectors pseudotyped with viral capsids from serotypes 1, 2, and 5 display differential efficiency and cell tropism after delivery to different regions of the central nervous system. Mol Ther. 2004;10: Cucchiarini M, Ren XL, Perides G, Terwilliger EF. Selective gene

10 doi: /en endo.endojournals.org 2995 expression in brain microglia mediated via adeno-associated virus type 2 and type 5 vectors. Gene Ther. 2003;10: Musatov S, Chen W, Pfaff DW, Kaplitt MG, Ogawa S. RNAi-mediated silencing of estrogen receptor in the ventromedial nucleus of hypothalamus abolishes female sexual behaviors. Proc Natl Acad Sci USA. 2006;103: Musatov S, Chen W, Pfaff DW, et al. Silencing of estrogen receptor in the ventromedial nucleus of hypothalamus leads to metabolic syndrome. Proc Natl Acad Sci USA. 2007;104: Xu Y, Nedungadi TP, Zhu L, et al. Distinct hypothalamic neurons mediate estrogenic effects on energy homeostasis and reproduction. Cell Metab. 2011;14: Herbison AE. Multimodal influence of estrogen upon gonadotropin-releasing hormone neurons. Endocr Rev. 1998;19: Kwakowsky A, Herbison AE, Abraham IM. The role of camp response element-binding protein in estrogen negative feedback control of gonadotropin-releasing hormone neurons. J Neurosci. 2012; 32: Cheong RY, Kwakowsky A, Barad Z, Porteous R, Herbison AE, Abraham IM. Estradiol acts directly and indirectly on multiple signaling pathways to phosphorylate camp-response element binding protein in GnRH neurons. Endocrinology. 2012;153: Romano N, Lee K, Abraham IM, Jasoni CL, Herbison AE. Nonclassical estrogen modulation of presynaptic GABA terminals modulates calcium dynamics in gonadotropin-releasing hormone neurons. Endocrinology. 2008;149: Chu Z, Andrade J, Shupnik MA, Moenter SM. Differential regulation of gonadotropin-releasing hormone neuron activity and membrane properties by acutely applied estradiol: dependence on dose and estrogen receptor subtype. J Neurosci. 2009;29: Morrison CD, Munzberg H. Capricious Cre: the devil is in the details. Endocrinology. 2012;153: Padilla SL, Reef D, Zeltser LM. Defining POMC neurons using transgenic reagents: impact of transient Pomc expression in diverse immature neuronal populations. Endocrinology. 2012;153: Mayer C, Acosta-Martinez M, Dubois SL, et al. Timing and completion of puberty in female mice depend on estrogen receptor -signaling in kisspeptin neurons. Proc Natl Acad Sci USA. 2010;107: Ruka KA, Burger LL, Moenter SM. Regulation of arcuate neurons coexpressing kisspeptin, neurokinin B, and dynorphin by modulators of neurokinin 3 and -opioid receptors in adult male mice. Endocrinology. 2013;154: Frazao R, Cravo RM, Donato J Jr, et al. Shift in Kiss1 cell activity requires estrogen receptor. J Neurosci. 2013;33: Lehman MN, Coolen LM, Goodman RL. Minireview: kisspeptin/ neurokinin B/dynorphin (KNDy) cells of the arcuate nucleus: a central node in the control of gonadotropin-releasing hormone secretion. Endocrinology. 2010;151: