Gonadotropin-Releasing Hormone/Luteinizing Hormone Surge

Circadian Regulation of Kiss1 Neurons: Implications for Timing the Preovulatory
Gonadotropin-Releasing Hormone/Luteinizing Hormone Surge
Jessica L. Robertson, Donald K. Clifton, Horacio O. de la Iglesia, Robert A. Steiner and Alexander S. Kauffman
Endocrinology 2009 150:3664-3671 originally published online May 14, 2009; , doi: 10.1210/en.2009-0247
To subscribe to Endocrinology or any of the other journals published by The Endocrine
Society please go to: http://endo.endojournals.org//subscriptions/
Copyright © The Endocrine Society. All rights reserved. Print ISSN: 0021-972X. Online
NEUROENDOCRINOLOGY
Circadian Regulation of Kiss1 Neurons: Implications
for Timing the Preovulatory Gonadotropin-Releasing
Hormone/Luteinizing Hormone Surge
Jessica L. Robertson, Donald K. Clifton, Horacio O. de la Iglesia, Robert A. Steiner,
and Alexander S. Kauffman
Departments of Biology (J.L.R., H.O.d.l.I.), Obstetrics and Gynecology (D.K.C., R.A.S.), and Physiology and Biophysics
(R.A.S.), University of Washington, Seattle, Washington 98195; and Department of Reproductive Medicine (A.S.K.),
University of California San Diego, La Jolla, California 92093
The preovulatory GnRH/LH surge depends on the presence of estradiol (E2) and is gated by a
circadian oscillator in the suprachiasmatic nucleus (SCN) that causes the surge to occur within a
specific temporal window. Although the mechanisms by which the clock times the LH surge are
unclear, evidence suggests that the SCN is linked to GnRH neurons through a multisynaptic pathway that includes neurons in the anteroventral periventricular nucleus (AVPV). Recently, Kiss1
neurons in the AVPV have been implicated in the surge mechanism, suggesting that they may
integrate circadian and E2 signals to generate the LH surge. We tested whether Kiss1 neurons
display circadian patterns of regulation in synchrony with the temporal pattern of LH secretion.
Mice housed in 14 h light, 10 h dark were ovariectomized, given E2 capsules (or nothing), and
transferred into constant darkness. Two days later, the mice were killed at various times of day and
their LH and Kiss1 levels assessed. In E2-treated females, LH levels were low except during late
subjective day (indicative of an LH surge). Similarly, AVPV Kiss1 expression and c-fos coexpression
in Kiss1 neurons showed circadian patterns that peaked coincident with LH. These temporal
changes in Kiss1 neurons occurred under steady-state E2 and constant environmental conditions,
suggesting that Kiss1 neurons are regulated by circadian signals. In the absence of E2, animals
displayed no circadian pattern in LH secretion or Kiss1 expression. Collectively, these findings
suggest that the LH surge is controlled by AVPV Kiss1 neurons whose activity is gated by SCN signals
in an E2-dependent manner. (Endocrinology 150: 3664 –3671, 2009)
I
n female mammals, ovulation is induced by a surge of LH
secretion from the pituitary, an event which is itself stimulated
by a surge of GnRH secretion from neurons in the forebrain (1).
It is well established that estradiol (E2) is critical for triggering the
preovulatory GnRH/LH surge; in rodents, the surge normally
occurs only on proestrus, as E2 levels are peaking, and females
that are ovariectomized (OVX) do not display LH surges (2, 3).
Although E2 is an absolute prerequisite for generating the LH
surge, in rodents, the surge event is also gated by a circadian
oscillator in the suprachiasmatic nucleus (SCN). This gatekeeper
ensures that the LH surge is precisely timed to occur within a 2to 4-h window, near the onset of darkness (and coincident with
the initiation of locomotor activity in nocturnal animals). Intact
female rodents display an LH surge in the late afternoon of
proestrus but not at other times of that day, and OVX females
treated with constant steady-state E2 display a late afternoon
LH surge, which repeats daily at the same time (4). In addition,
lighting paradigms that alter the natural period of the circadian clock, or phase-shift the clock’s temporal output, modify
the timing of the LH surge (5– 8); in all cases, the surge onset
remains tightly coupled to the onset of daily locomotor activity, which is also timed by the SCN, implying that the timing
of the surge and activity rhythms are both gated by the SCN.
Finally, lesions of the SCN or genetic disruptions of the circadian molecular clockwork disrupt estrous cyclicity and prevent the E2-induced LH surge (9 –15). Collectively, these
observations argue that proper functioning of the SCN⬘s circadian clockwork is essential for the timing and generation of
ISSN Print 0013-7227 ISSN Online 1945-7170
Printed in U.S.A.
Copyright © 2009 by The Endocrine Society
doi: 10.1210/en.2009-0247 Received February 26, 2009. Accepted May 4, 2009.
First Published Online May 14, 2009
Abbreviations: AVPV, Anteroventral periventricular nucleus; CT, circadian time; DIG,
digoxigenin; E2, estradiol; ER␣, estrogen receptor-␣; OVX, ovariectomized; RNase, ribonuclease; SCN, suprachiasmatic nucleus; SSC, saline sodium citrate.
3664
endo.endojournals.org
Endocrinology, August 2009, 150(8):3664 –3671
Endocrinology, August 2009, 150(8):3664 –3671
the GnRH/LH surge in rodents, but precisely how this occurs
is poorly understood.
There are two possible routes by which the SCN could regulate GnRH neurons and hence the LH surge. One pathway
involves direct innervation of GnRH neurons by the SCN (16 –
18); however, anatomical and physiological evidence suggests
that a second route is perhaps more critical (4). This second
pathway involves an indirect circuit linking the SCN with GnRH
neurons through a relay station in the anteroventral periventricular nucleus (AVPV), a hypothalamic region that is critical
for the GnRH/LH surge in rodents (reviewed in Refs. 3, 19, and
20). Some neurons in the AVPV express estrogen receptor-␣
(ER␣), the primary receptor implicated in mediating positive
feedback effects of E2 on the LH surge (3); moreover, some ER␣expressing cells in the AVPV receive direct neural input from the
SCN (16, 21). Thus, an E2-sensitive population of neurons
within the AVPV could represent the cellular conduit linking the
SCN to GnRH neurons, and recent evidence suggests that these
particular AVPV cells may be Kiss1 neurons.
The Kiss1 gene encodes kisspeptin, a neuropeptide that plays
a critical role in the neuroendocrine regulation of reproduction.
In rodents, sheep, and primates (including humans), treatment
with kisspeptin elicits a rapid increase in gonadotropin secretion,
mediated by kisspeptin’s direct stimulation of GnRH neurons
(reviewed in Refs. 19 and 22). Kiss1 neurons are located in several regions of the hypothalamus (23), but in rodents, the population of Kiss1 cells in the AVPV has been implicated as the
critical component that drives the preovulatory GnRH/LH
surge. First, virtually all Kiss1 neurons in the AVPV express ER␣,
and Kiss1 mRNA in this region is strongly up-regulated by E2
(24, 25). Second, Kiss1 expression in the AVPV of cycling female
rats is induced, along with Fos in Kiss1 neurons, at the time of the
preovulatory LH surge (26). Third, infusions of antiserum to
kisspeptin prevent the LH surge in intact and E2-treated female
rats (27, 28). Finally, the expression of Kiss1 in the AVPV is
sexually differentiated, with females having significantly more
Kiss1 neurons than males (29, 30), corresponding with the ability of females (but not males) to display an LH surge.
Although these observations provide compelling evidence
that Kiss1 neurons in the AVPV participate in the E2-mediated
induction of the LH surge, it remains ambiguous whether Kiss1
neurons also represent part of the circadian circuit linking the
SCN to GnRH neurons. If Kiss1 neurons in the AVPV mediate
the circadian effects of the clock in the SCN on the daily LH
surge, then Kiss1 gene expression or transcriptional activity
within Kiss1 cells may display a circadian pattern of regulation.
To address this possibility, we evaluated the temporal patterns of
Kiss1 expression and induction of the immediate-early gene c-fos
(as a marker of neuronal activation) within Kiss1 neurons in the
AVPV of female mice. To eliminate the possible effects of fluctuating environmental conditions, all experiments were performed under constant environmental conditions (constant
darkness and temperature). In addition, because animals lacking
E2 do not display LH surges, we reasoned that this might reflect
the absence of E2-dependent circadian activation of Kiss1 cells in
the AVPV. As an alternative, we supposed that Kiss1 neurons
could well maintain their daily circadian activation, regardless of
endo.endojournals.org
3665
the presence or absence of E2, but could generate a preovulatory
LH surge only when E2-induced kisspeptin production was sufficiently high enough to drive GnRH neurons (e.g. on proestrus).
To distinguish between these possibilities, we evaluated whether
Kiss1 neurons display circadian activation in both the presence
and absence of E2 and propose a model of GnRH/LH surge
generation via an E2-dependent circadian activation of Kiss1
neurons in the AVPV.
Materials and Methods
Animals
Adult (2 month old) C57/BL6J female mice purchased from Charles
River Laboratories (Wilmington, MA) were individually housed with
access to a running wheel in a 14-h light, 10-h dark cycle (lights off at
2100 h). Mice had ad libitum access to standard rodent chow and water
throughout the study. Running wheel activity was continuously recorded
for the duration of the study using Clocklab (Actimetrics, Wilmette, IL)
and analyzed using El Temps software (http://www.el-temps.com/). All
experiments were approved by the University of Washington Animal
Care and Use Committee.
OVX and LH surge protocol
Ovaries were removed from isoflurane-anesthetized mice that were
pretreated with Buprenex analgesic (1.5 ␮g, sc). Briefly, the anesthetized
animal’s ventral surface was shaved and cleaned and the ovaries dissected
out through midline incisions in the skin and abdominal musculature.
After ovary removal, the muscle was sutured and the skin closed with
sterile wound clips.
To induce a predictably timed daily LH surge, mice in experiments
1 and 2 were sc implanted at the time of OVX with a SILASTIC brand
(Dow Corning, Midland, MI) capsule containing 0.625 ␮g E2 (in
sesame oil, following the protocol of Ref. 31). This E2-treatment paradigm results in constant serum levels of E2 (⬃30 pg/ml) between 2
and 5 d after implantation and reliably induces daily afternoon LH
surges (defined as ⱖ8-fold above baseline levels) beginning on d 2 (31,
32). All surgeries were performed in the morning during the first
several hours after lights on.
Blood and tissue collection
At the time of killing (see specific experiments for details), mice were
anesthetized with isoflurane and their blood collected via retroorbital
bleeding. Blood was assayed for LH concentrations via a sensitive mouse
LH RIA, performed by the University of Virginia Ligand Assay Lab as
described previously (33, 34). After blood collection, animals were rapidly decapitated and their brains immediately collected and frozen on dry
ice. Frozen brains were stored at ⫺80 C until sectioning on a cryostat.
Five sets of 20-␮m sections were cut in the coronal plane, thaw-mounted
onto Superfrost-plus slides, and stored at ⫺80 C. One set was used for
each in situ hybridization (ISH) assay.
Single-label ISH
Slide-mounted brain sections were processed for Kiss1 ISH, as previously described (23, 26, 30). Briefly, sections were fixed in 4% paraformaldehyde, rinsed in phosphate buffer, treated with acetic anhydride,
rinsed in 2⫻ saline sodium citrate (SSC), delipidated in chloroform, dehydrated in ethanols, and air dried. Antisense mouse Kiss1 probe, spanning
bases 76 – 486 of the mouse cDNA sequence, was generated using 33P, as
previously described (23, 24). The radiolabeled Kiss1 riboprobe was
combined with 1/20 volume yeast tRNA (Roche Biochemicals, Indianapolis, IN) in TE (0.1 M Tris/0.01 M EDTA, pH 8.0), heat denatured,
added to hybridization buffer at a ratio of 1:4, and applied to each slide
(100 ␮l/slide; 0.03 pmol probe/ml). Slides were then coverslipped and
3666
Robertson et al.
Circadian Regulation of Kiss1 Neurons
Endocrinology, August 2009, 150(8):3664 –3671
placed in a humidity chamber at 55 C for 16 h. After hybridization, the
slides were washed in 4⫻ SSC at room temperature and placed into
ribonuclease (RNase) [37 mg/ml RNase A (Roche Biochemicals) in 0.15
M sodium chloride, 10 mM Tris, 1 mM EDTA, pH 8.0] for 30 min at 37
C and then into RNase buffer without RNase at 37 C for 30 min. Slides
were then washed in 0.1⫻ SSC at 62 C, dehydrated in ethanols, and air
dried. Dry slides were dipped in Kodak NTB emulsion (VWR, West
Chester, PA), air dried, and stored at 4 C for 10 –12 d (depending on the
assay), after which they were developed and coverslipped.
Double-label ISH
For double-label ISH, slide-mounted brain sections were treated similarly to single-label ISH with the following modifications. Digoxigenin
(DIG)-labeled antisense Kiss1 probe was synthesized along with radiolabeled c-fos riboprobe (labeled using 33P), and both riboprobes were
dissolved in the same hybridization buffer along with tRNA and applied
to slides for overnight hybridization. The radiolabeled c-fos riboprobe
was generated by using 33P as described previously (35). After the 0.1⫻
SSC washes on d 2, slides were incubated in 2⫻ SSC with 0.05% Triton
X-100 and 2% sheep serum for 1 h at room temperature and then washed
in buffer 1 (100 mM Tris-HCl, pH 7.5, and 150 mM NaCl). Slides were
then incubated overnight at room temperature with alkaline-conjugated
anti-DIG antibody fragments (Roche Biochemical) that were diluted
1:350 – 400 in buffer 1 containing 1% sheep serum and 0.3% Triton
X-100. The next day, slides were washed with buffer 1 and incubated
with Vector Red alkaline phosphatase substrate (Vector Laboratories,
Burlingame, CA) for 2 h. The slides were then air dried, dipped in NTB
emulsion, stored at 4 C, and developed 10 –13 d later (depending on the
assay).
Experiment 1: evaluation of the circadian regulation of
Kiss1 mRNA in the AVPV of E2-treated animals
Kisspeptin signaling arising from the AVPV has been proposed to
mediate the generation of the LH surge. Because the LH surge is regulated
by the circadian clock, we hypothesized that Kiss1 neurons in the AVPV
are themselves under circadian control. To test this hypothesis, we used
single-label ISH to analyze levels of Kiss1 mRNA in female mice that were
housed under constant lighting conditions and experiencing constant E2
exposure. After 13–20 d acclimation to a 14-h light, 10-h dark cycle, mice
were OVX in the morning and implanted with E2-containing capsules
that induce a daily LH surge. On the night of surgery, mice were released
into constant dim light (⬍1 lux, equivalent to constant darkness). At least
36 h after transfer to constant darkness, animals were killed under dim,
red-light illumination at one of eight circadian time points throughout
the day: circadian time (cT) 0, CT 4, CT 8, CT 9, CT 11, CT 12, CT 16,
and CT 20 (n ⫽ 4 – 6 per time point). By convention in chronobiology,
the onset of locomotor activity, as determined by each animal’s running
wheel activity rhythm, was denoted as CT 12. We determined in advance
that the free-running period of the female mice housed in constant darkness, both in the presence or absence of E2, is significantly close to 24 h
and that the phase drift after 48 h in constant darkness is insignificant.
Thus, the onset of locomotor activity the day before killing (when animals were already in constant darkness) was used to determine each
animal’s specific CT 12 the following day. All kills were carried out ⫾15
min from the various calculated time points (see Fig. 1 for example).
Blood and brains were collected at the time of killing and used to measure
serum levels of LH and Kiss1 mRNA expression in the AVPV for each
circadian time point.
Experiment 2: circadian activation of Kiss1 neurons in
the AVPV of E2-treated animals
Experiment 1 determined that there is a circadian pattern of Kiss1
gene expression in the brains of E2-treated female mice housed in constant darkness and that this Kiss1 pattern coincided with the circadian
pattern of LH secretion. We next asked whether Kiss1 neurons also
undergo a circadian activation, as evidenced by the induction of the
FIG. 1. Representative locomotor actogram of a female mouse housed in 14-h
light, 10-h dark for approximately 1 wk. The animal was OVX and implanted
with an E2-containing capsule (indicated by an asterisk) on the morning of d 9
and then transferred into constant darkness later that night (indicated by two
asterisks). The mouse was killed 2 d after receiving the E2 capsule, which was at
CT 12 (indicated by the black arrow). Animals in other groups were killed at
various time points throughout the second day in constant darkness, as indicated
by the horizontal dashed line (see Materials and Methods for specific times).
Shaded areas represent the time of darkness (⬍1 lux dim red light). Note that CT
12 (indicated here by the black arrow) is slightly earlier than the extrapolated
time of lights off. Therefore, CT 12 was calculated individually for each animal
based on its own specific activity onset.
immediate-early gene c-fos in Kiss1 cells. To address this question, we
used double-label ISH to determine colocalization of Kiss1 mRNA and
c-fos mRNA in the brains of animals from experiment 1. One set of brain
tissue from E2-treated females that were housed in constant darkness was
processed for double-label ISH and the percentage of AVPV Kiss1 neurons coexpressing c-fos was compared across the eight circadian time
points.
Experiment 3: evaluation of the circadian regulation of
Kiss1 neurons in the AVPV of OVX animals
Experiments 1 and 2 suggested that Kiss1 neurons of E2-treated mice
are under circadian control and that this temporal regulation of Kiss1
cells mirrors that of LH secretion. Because OVX animals do not display
an LH surge, we next asked whether the circadian stimulatory regulation
of Kiss1 neurons is dependent on the presence of E2. We postulated that
in the absence of E2, Kiss1 cells would not show a circadian-induced
increase in either Kiss1 gene expression or c-fos induction, in contrast to
what happens in the presence of E2. To test this hypothesis, we replicated
experiments 1 and 2 but this time using OVX mice that were not E2
treated. Mice were housed in 14 h light, 10 h dark for 6 d and then OVX.
Six to 7 d after OVX, animals were released into constant dim light (⬍1
lux; constant darkness). At least 36 h later, the mice were killed under
dim, red-light illumination at one of eight circadian time points, as in
experiment 1 (n ⫽ 3–5 per time point). Blood and brains were collected
at the time of killing and processed to determine serum levels of LH (via
RIA) as well as Kiss1 and c-fos mRNA (via single- and double-label ISH).
Quantification of ISH assays and statistical analysis
Slides were analyzed with an automated image processing system by
a person unaware of the treatment group of each slide. For single-label
experiments, custom grain-counting software was used to unilaterally
Endocrinology, August 2009, 150(8):3664 –3671
count the number of radiolabeled cell clusters and the number of silver
grains in each cell (a semiquantitative index of mRNA content per cell)
(23, 26) for all AVPV sections. Cells were considered Kiss1 positive when
the number of silver grains in a cluster exceeded that of background by
3-fold. For double-label assays, DIG-containing cells (Kiss1 cells) were
identified under fluorescence microscopy, and the grain-counting software was used to quantify silver grains (representing c-fos mRNA) over
each cell. A cell was considered double labeled if it had a signal-tobackground ratio of at least 3 (26, 30). For each animal, the amount of
double labeling was calculated as a percentage of the total number of
Kiss1 mRNA-expressing cells and then averaged across animals to produce a group mean.
All LH and ISH data are expressed as the mean ⫾ SEM for each
group. One-way ANOVAs were used to assess variation among experimental groups (time points) in each experiment, and differences
in means were assessed by post hoc Fisher’s least significant difference
tests (with Staview 5.0.1; SAS Institute, Cary, NC). Significance level
was set at P ⬍ 0.05.
Results
Experiment 1: Kiss1 gene expression in the AVPV of
E2-treated female mice is under circadian control
This experiment tested whether Kiss1 expression in the AVPV
exhibits a circadian pattern under constant conditions and
whether this pattern mirrors that of LH secretion. Female mice
housed in constant darkness and treated with constant E2 displayed a significant circadian pattern of LH secretion, with serum LH levels being low to undetectable in the subjective morning (CT 0, CT 4, CT 8), high in the subjective late afternoon/early
evening (CT 11, CT 12), and low again in the late subjective night
(CT 20) (P ⬍ 0.05, Fig. 2A). LH levels were at intermediate levels
at CT 9 and CT 16, reflecting the presence of an LH surge in a
FIG. 2. A, Mean (⫾SEM) serum levels of LH in OVX, E2-treated mice housed in
constant conditions and killed at one of eight time points throughout the
circadian day. Values with different letters differ significantly from each other
(P ⬍ 0.05). B, Percentage of animals at each time point displaying an LH surge
(defined as an 8-fold or greater increase in LH values compared with mean CT 0
and CT 4 values) (31). n ⫽ 4 – 6 animals per group.
endo.endojournals.org
3667
small subset (⬃20%) of animals in these two groups (Fig. 2B).
Kiss1 gene expression in the AVPV of E2-treated females also
exhibited a similar circadian pattern, with higher expression visible in the late afternoon (CT 11, CT 12) vs. earlier or later time
points (Fig. 3A). Quantitative analysis determined that there was
a trend for increased number of Kiss1 cells at CT 11 and CT 12
(P ⬍ 0.09; Fig. 3B) and significantly more (24 –28%) grains per
Kiss1 cell (indicative of Kiss1 mRNA level per cell) in the late
afternoon/ early evening compared with earlier and later timepoints (P ⬍ 0.01; Fig. 3C). Similarly, total Kiss1 mRNA in the
AVPV, calculated as the product of Kiss1 cell number and number of grains per Kiss1 cell, showed a significant circadian pattern
with markedly higher levels (⬃45–55%) at CT 11 and CT 12 vs.
earlier and later time-points (P ⬍ 0.01, data not shown).
Experiment 2: Kiss1 cells in the AVPV of E2-treated mice
show a circadian pattern of neuronal activation
Experiment 1 identified a circadian pattern of Kiss1 expression in the brains of E2-treated female mice housed in constant
darkness, coinciding with a circadian pattern of LH secretion.
This experiment tested whether Kiss1 cells in E2-treated mice
also undergo a circadian pattern of neuronal activation, as determined by c-fos induction in Kiss1 cells. We found that the
expression of c-fos in Kiss1 neurons in the AVPV was low or
nondetectable in the subjective morning and late subjective
night, despite the presence of c-fos-expressing cells in other brain
regions at these times (e.g. thalamus and paraventricular nucleus). In contrast, c-fos expression in Kiss1 neurons was prevalent in subjective late afternoon/early evening, coincident with
the occurrence of LH surges at these times (see experiment 1) and
indicative of a circadian regulation (Fig. 4). Quantitatively, the
percentage of Kiss1 neurons expressing c-fos was no more than
10% at CT 0, CT 4, and CT 8, elevated to 40% at CT 11 and CT
12, and roughly 10% again at CT 20 (P ⬍ 0.01; Fig. 4). Similar
to the LH data in these E2-treated females (experiment 1), the
percentage of c-fos/Kiss1 coexpression was intermediate at CT 9
and CT 16 (Fig. 4), reflecting high coexpression in one animal in
each of these groups (the same animals with high LH).
Experiment 3: Kiss1 neurons in the AVPV of OVX mice
do not exhibit circadian patterns of regulation
Because OVX mice are incapable of displaying an LH surge,
this experiment tested whether the circadian patterns of Kiss1
gene expression and neuronal activation observed in experiments 1 and 2 are dependent on the presence of E2. Unlike our
findings in E2-treated animals, serum levels of LH in OVX mice
lacking E2 and housed in constant darkness showed no evidence
of a circadian pattern. Although the baseline concentrations of
LH were higher in all OVX animals (reflecting absence of negative feedback after OVX), there was no obvious circadian pattern or significant change in LH levels across the 24-h cycle (P ⬎
0.95; Fig. 5). Likewise, neither Kiss1 cell number in the AVPV
nor Kiss1 mRNA/cell were different among any of the circadian
time points (P ⬎ 0.90 for each; Fig. 6), nor was there any significant difference in the percentage of c-fos coexpression in
Kiss1 neurons across the circadian day (P ⬎ 0.90; Fig. 6D). Collectively, these observations demonstrate that there is little or no
3668
Robertson et al.
Circadian Regulation of Kiss1 Neurons
Endocrinology, August 2009, 150(8):3664 –3671
over, we show that the ability of Kiss1 neurons to exhibit circadian regulation of both
gene expression and neuronal activation is
contingent on the presence of E2, perhaps
explaining why an LH surge is absent in females with low or absent E2. Collectively,
our findings suggest that Kiss1 neurons in
the AVPV could receive and integrate both
hormonal and temporal cues to generate
and time the LH surge.
Considerable evidence suggests that the
precisely timed pattern of LH release in female rodents reflects the interaction of E2
with an obligatory signal that emanates
from the circadian clock in the SCN. First,
intact female rodents normally display an
LH surge only in the late afternoon of
proestrus; however, barbiturate treatment
during a critical window (2–3 h in the late
afternoon) not only blocks the LH surge
(and ovulation) but also delays its next appearance by exactly 24 h (36 –39). Second,
in OVX rodents, treatment with constant
high levels of E2 elicits a daily LH surge,
which occurs at the exact same time every
day (40 – 42). However, females lacking a
functional SCN (achieved by creating discrete lesions of SCN or genetic disruptions
of the molecular clockwork in the SCN) are
incapable of displaying an LH surge (at any
time) in response to an E2 challenge (9, 11,
14, 15, 43). This establishes the absolute
FIG. 3. A, Representative dark-field photomicrographs showing Kiss1 mRNA-expressing cells (as reflected
requirement of circadian signaling from the
by the presence of white clusters of silver grains) in the AVPV of OVX, E2-treated female mice that were
SCN to produce an LH surge. Finally, exhoused in constant conditions and killed at different times throughout the circadian day. 3V, Third
perimental paradigms that modify the cirventricle. B, Mean (⫾SEM) number of Kiss1-expressing cells in the AVPV across the circadian day displayed a
trend (P ⬍ 0.09) for more Kiss1 neurons during the late subjective afternoon/early evening. C, The amount
cadian clock’s period or phase shift the
of Kiss1 mRNA per cell, as indicated by the number of silver grains per cell, was significantly different
clock’s temporal output correspondingly alacross circadian time points, with highest values at CT 11 and CT 12 (P ⬍ 0.01); values with different
ter the timing of the LH surge in rodents.
letters differ significantly from each other. n ⫽ 4 – 6 animals per group.
However, in all such cases, the onset of the
LH surge remains tightly coupled to the onset of daily locomotor
circadian pattern of activation of Kiss1 neurons in the AVPV of
rhythm (5– 8), indicating that the temporal gating of the surge
mice lacking E2 and housed in constant conditions.
and behavioral rhythms share a common circadian pacemaker,
presumably the SCN. Although these findings argue persuasively
that the SCN regulates the LH surge, the neural mechanisms by
Discussion
which this occurs are poorly understood. Because the AVPV
plays a key role in generating the LH surge (reviewed in Refs. 3,
Since the pioneering work of Everett and Sawyer in 1950 (36 –
19, and 44), it seems plausible that the SCN signals to GnRH
39), we have known that the LH surge (and subsequently, ovuneurons through intermediaries in the AVPV. In support of this,
lation) is dependent not only on E2 but also on a neural signal that
tract tracing experiments demonstrate that SCN neurons innerreflects the time of day, so that the LH surge occurs only within
vate a subset of AVPV neurons (16) and these specific AVPV
a narrow temporal window (typically late afternoon in nocturnal
neurons also express ER␣, which mediates the E2-induced LH
rodents). Despite the fact that this LH surge phenomenon has
been extensively studied over the intervening years (4), the neusurge.
ronal circuitry and molecular mechanisms that represent the
The phenotypic identity of the E2-responsive neurons in the
point of convergence between the E2 and circadian signals have
AVPV that are targets for SCN-derived projections is unknown.
However, our results suggest that these neurons are Kiss1 neuremained a mystery. In the present study, we show that a poprons. In E2-treated female mice, we found that Kiss1 gene exulation of Kiss1 neurons in the AVPV, previously implicated in
the LH surge mechanism, is under circadian regulation. Morepression in the AVPV varied across the circadian day. The ex-
Endocrinology, August 2009, 150(8):3664 –3671
endo.endojournals.org
3669
A
CT0
CT11
3V
C 250
180
Grains per Kiss1 cell
Kiss1 cells in AVPV
CT20
3V
3V
3V
B
CT12
140
100
60
200
150
100
50
pression of Kiss1 increased (by ⬃25%) during the subjective late
afternoon and early evening and then diminished in late subjective night. Because the animals were housed in constant darkness, these changes in Kiss1 gene expression were unlikely to
have been produced by an hourglass mechanism that might be
initiated each day through a daily resetting of a light-dark cycle.
Furthermore, because all animals in all time points had the same
constant E2 treatment, it is unlikely that these temporal changes
in Kiss1 expression were caused by group differences or temporal
changes in circulating levels of E2. These data indicate that Kiss1
gene expression is regulated by both sex steroid signals (up-regulated Kiss1 expression compared with OVX levels, as previously determined) (24, 30) and circadian signals (up-regulated
Kiss1 expression at specific times) (present study). Moreover,
FIG. 5. Mean (⫾SEM) plasma LH in OVX mice housed in constant conditions and
killed at one of eight times throughout the circadian day. There was no
significant difference in LH levels between any of the time points. n ⫽ 4 –5
animals per group.
CT20
CT16
CT12
CT9
CT11
CT4
CT8
CT20
CT12
CT16
CT11
CT8
CT9
CT0
Time
Time
50
40
30
20
CT20
CT16
CT12
CT11
CT9
CT8
0
CT0
10
CT4
FIG. 4. A, Representative photomicrographs of Kiss1 mRNA and c-fos
mRNA coexpression in the AVPV of OVX, E2-treated female mice housed in
constant conditions and killed at different times throughout the circadian
day. Kiss1-containing neurons were visualized with Vector Red substrate, and
c-fos mRNA was marked by the presence of silver grains. White arrows denote
example Kiss1 cells lacking c-fos; yellow arrows denote example Kiss1 cells
coexpressing c-fos. B, Mean (⫾SEM) percentage of Kiss1 mRNA-containing
neurons in the AVPV that coexpress c-fos in OVX, E2-treated female mice killed
at one of eight times throughout the circadian day. There was a significant effect
of time (P ⬍ 0.01) with increased coexpression of Kiss1 and c-fos in the late
afternoon/early evening. Values with different letters differ significantly from
each other. n ⫽ 4 – 6 animals per group.
% Kiss1 cells
expressing c-fos
D
0
CT4
0
CT0
20
Time
FIG. 6. Lack of a circadian pattern in Kiss1 gene expression in OVX females. A,
Representative dark-field photomicrographs showing Kiss1 mRNA-expressing
cells in the AVPV of OVX mice housed in constant conditions and killed at
different time points throughout the circadian day. B, Mean (⫾SEM) number of
Kiss1 mRNA-expressing neurons in the AVPV of OVX female mice housed in
constant conditions and killed at one of eight time points throughout the
circadian day. C, Mean (⫾SEM) number of silver grains per Kiss1 cell in the AVPV
of OVX mice killed across the circadian day. D, Mean (⫾SEM) percentage of Kiss1
mRNA-containing neurons in the AVPV that coexpress c-fos in OVX female mice.
In A–C, there was no significant difference in the measures between any of the
circadian time points (P ⬎ 0.90 for all measures). n ⫽ 3–5 animals per group.
transcriptional activation of Kiss1 neurons, as measured by neuronal c-fos induction, also exhibited a strong circadian pattern in
constant conditions (in the presence of E2). Thus, in addition to
a timed increase in Kiss1 gene transcription, Kiss1 neurons are
also temporally induced by afferent circadian signals to increase
the expression of other genes (or perhaps alter neuronal firing).
Presumably, this temporal activation of Kiss1 neurons is related
to a timed stimulation of GnRH neurons, resulting in the circadian-gated LH surge. Whether the observed circadian pattern of
Kiss1 gene expression and Kiss1 neuronal activation are causally
related is unknown, but it is conceivable that the pattern of increased Kiss1 gene expression reflects a replenishing of Kiss1/
kisspeptin stores after the neuron has fired to release kisspeptin
(denoted by high neuronal c-fos induction).
The mechanism by which the SCN regulates Kiss1 neurons
has not been elucidated. Based on previous studies, it is clear that
the circadian regulation of the LH surge by the SCN involves
neuronal, not humoral, signaling from the clock to GnRH neurons (45, 46). Thus, if Kiss1 neurons in the AVPV serve as a
cellular conduit to relay input from the SCN to GnRH neurons,
these Kiss1 neurons would likely be regulated by the clock
through neuronal innervation. In rodents, there are direct neuronal connections between the SCN and the AVPV and also
between the AVPV and GnRH neurons (16, 21, 47, 48), sug-
3670
Robertson et al.
Circadian Regulation of Kiss1 Neurons
gesting that a multisynaptic neuronal circuit links the SCN with
GnRH neurons through the AVPV. Although neuroanatomical
evidence documenting a direct SCN-Kiss1 circuit has yet to be
reported, preliminary observations in mice suggest a direct,
monosynaptic innervation of Kiss1 neurons in the AVPV by vasopressin signaling arising from the SCN (49). These investigations also implicate vasopressin as the key neurotransmitter upregulating Kiss1 neurons, a conjecture supported by findings that
vasopressin can stimulate GnRH/LH secretion in animals lacking a functional SCN (50, 51). It is also worth noting that Kiss1
cells themselves may be circadian oscillators and that the observed temporal changes in Kiss1 activation could be generated
intrinsically by intracellular clocks. This possibility derives support from the fact that many non-SCN cells throughout the body
(e.g. GnRH or liver cells) express clock genes and in some cases
exhibit endogenous circadian rhythmicity in vitro, independent
of SCN input (52, 53). However, given the strong evidence that
SCN signaling is critical for both the generation and timing of the
LH surge, and the fact that the SCN projects to the AVPV (and
likely Kiss1 cells), the most parsimonious explanation is that the
circadian pattern of Kiss1 neurons is controlled by the SCN.
Animals lacking E2 cannot display an LH surge. The inability
to surge in the absence of E2 could reflect the absence of circadian
input to the surge-generating circuitry and/or some other missing
E2-dependent aspect of the surge-generating mechanism. Here,
we show that the circadian pattern of Kiss1 regulation observed
in E2-treated mice was completely absent in OVX animals. Thus,
despite a lower, yet significant, number of detectable Kiss1 neurons in the OVX condition, these Kiss1 cells were not temporally
activated in the absence of E2. This finding argues against a
model in which the SCN activates Kiss1 neurons every day but
generates an LH surge only when there is sufficient Kiss1 mRNA
and kisspeptin release to drive a GnRH surge. Instead, our results
imply that either 1) the SCN does not automatically signal to
Kiss1 neurons every day but, instead, does so only in the presence
of E2 or 2) the SCN sends temporal cues to Kiss1 neurons every
day, but the ability of Kiss1 neurons to receive/decode the circadian signal is E2 dependent. Both the SCN and Kiss1 neurons
express ERs (24, 54, 55), rendering either of these models plausible. It should also be noted that E2 may also act upstream of the
SCN at other E2-responsive neurons that project to the clock
(56). Regardless of mechanism, any of these E2-dependent models would result in Kiss1 neurons being up-regulated by circadian
input only when there is E2 present, thus explaining why the LH
surge occurs only during proestrus or conditions of elevated E2.
Notably, these models indicate that E2 has a dual role for inducing the LH surge: 1) to stimulate Kiss1 gene expression, presumably to generate abundant kisspeptin, regardless of time of day
(i.e. fewer AVPV Kiss1 neurons in OVX than E2-treated animals)
and 2) to permit the SCN to activate Kiss1 neurons, thereby
triggering them to signal to GnRH neurons. Although complementary and synergistic in their effects on the hypothalamopituitary-gonadal axis, these two actions of E2 on the Kiss1 system may involve independent processes.
In conclusion, we report a significant circadian regulation of
Kiss1 gene expression and transcriptional activation of Kiss1
neurons in the AVPV of female mice housed in constant condi-
Endocrinology, August 2009, 150(8):3664 –3671
tions and exposed to steady-state E2. These results implicate
Kiss1 neurons in the AVPV as critical integrators of both sex
steroids and circadian signals and suggest that the timing of
the LH surge involves temporal activation of Kiss1 neurons.
We also determined that this circadian pattern of Kiss1 neuron
regulation is E2 dependent, implying that the SCN is incapable
of activating Kiss1 neurons in the absence of the molecular/
cellular effects of E2.
Acknowledgments
Address all correspondence and requests for reprints to: Dr. Alexander
S. Kauffman, Department of Reproductive Medicine, University of California, San Diego, 9500 Gilman Drive 0674, La Jolla, California 920930674. E-mail: [email protected].
This research was supported by the Eunice Kennedy Shriver National
Institute of Child Health and Human Development/National Institutes
of Health (NIH) through cooperative agreements U54 HD12629 (to the
University of Washington Center for Research in Reproduction and Contraception) and U54 HD28934 (University of Virginia Ligand Assay
Core) as well as Grants K99/R00 HD056157 and R01 HD27142. Additional funding support was provided by NIH Grant R01 MH075016
and an National Science Foundation Graduate Research Fellowship
(to J.L.R.).
Disclosure Summary: The authors have nothing to disclose.
References
1. Herbison AE, Porteous R, Pape JR, Mora JM, Hurst PR 2008 Gonadotropinreleasing hormone neuron requirements for puberty, ovulation, and fertility.
Endocrinology 149:597– 604
2. Clarkson J, Herbison AE 2009 Oestrogen, kisspeptin, GPR54 and the preovulatory luteinising hormone surge. J Neuroendocrinol 21:305–311
3. Herbison AE 2008 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 57:277–287
4. de la Iglesia HO, Schwartz WJ 2006 Timely ovulation: circadian regulation of
the female hypothalamo-pituitary-gonadal axis. Endocrinology 147:1148 –
1153
5. Alleva JJ, Waleski MV, Alleva FR 1971 A biological clock controlling the
estrous cycle of the hamster. Endocrinology 88:1368 –1379
6. Fitzgerald K, Zucker I 1976 Circadian organization of the estrous cycle of the
golden hamster. Proc Natl Acad Sci USA 73:2923–2927
7. Moline ML, Albers HE 1988 Response of circadian locomotor activity and the
proestrous luteinizing hormone surge to phase shifts of the light-dark cycle in
the hamster. Physiol Behav 43:435– 440
8. Moline ML, Albers HE, Todd RB, Moore-Ede MC 1981 Light-dark entrainment of proestrous LH surges and circadian locomotor activity in female hamsters. Horm Behav 15:451– 458
9. Brown-Grant K, Raisman G 1977 Abnormalities in reproductive function
associated with the destruction of the suprachiasmatic nuclei in female rats.
Proc R Soc Lond B Biol Sci 198:279 –296
10. Miller BH, Olson SL, Levine JE, Turek FW, Horton TH, Takahashi JS 2006
Vasopressin regulation of the proestrous luteinizing hormone surge in wildtype and Clock mutant mice. Biol Reprod 75:778 –784
11. Miller BH, Olson SL, Turek FW, Levine JE, Horton TH, Takahashi JS 2004
Circadian clock mutation disrupts estrous cyclicity and maintenance of pregnancy. Curr Biol 14:1367–1373
12. Raisman G, Brown-Grant K 1977 The ‘suprachiasmatic syndrome’: endocrine
and behavioural abnormalities following lesions of the suprachiasmatic nuclei
in the female rat. Proc R Soc Lond B Biol Sci 198:297–314
13. Wiegand SJ, Terasawa E 1982 Discrete lesions reveal functional heterogeneity
of suprachiasmatic structures in regulation of gonadotropin secretion in the
female rat. Neuroendocrinology 34:395– 404
14. Wiegand SJ, Terasawa E, Bridson WE, Goy RW 1980 Effects of discrete lesions
of preoptic and suprachiasmatic structures in the female rat. Alterations in the
Endocrinology, August 2009, 150(8):3664 –3671
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
feedback regulation of gonadotropin secretion. Neuroendocrinology 31:147–
157
Lucas RJ, Stirland JA, Darrow JM, Menaker M, Loudon AS 1999 Free running
circadian rhythms of melatonin, luteinizing hormone, and cortisol in Syrian
hamsters bearing the circadian tau mutation. Endocrinology 140:758 –764
de la Iglesia HO, Blaustein JD, Bittman EL 1995 The suprachiasmatic area in
the female hamster projects to neurons containing estrogen receptors and
GnRH. Neuroreport 6:1715–1722
Van der Beek EM, Horvath TL, Wiegant VM, Van den Hurk R, Buijs RM 1997
Evidence for a direct neuronal pathway from the suprachiasmatic nucleus to
the gonadotropin-releasing hormone system: combined tracing and light and
electron microscopic immunocytochemical studies. J Comp Neurol 384:569 –
579
van der Beek EM, Wiegant VM, van der Donk HA, van den Hurk R, Buijs RM
1993 Lesions of the suprachiasmatic nucleus indicate the presence of a direct
vasoactive intestinal polypeptide-containing projection to gonadotrophin-releasing hormone neurons in the female rat. J Neuroendocrinol 5:137–144
Kauffman AS, Clifton DK, Steiner RA 2007 Emerging ideas about kisspeptinGPR54 signaling in the neuroendocrine regulation of reproduction. Trends
Neurosci 30:504 –511
Simerly RB 1998 Organization and regulation of sexually dimorphic neuroendocrine pathways. Behav Brain Res 92:195–203
Watson Jr RE, Langub Jr MC, Engle MG, Maley BE 1995 Estrogen-receptive
neurons in the anteroventral periventricular nucleus are synaptic targets of the
suprachiasmatic nucleus and peri-suprachiasmatic region. Brain Res 689:254 –
264
Popa SM, Clifton DK, Steiner RA 2008 The role of kisspeptins and GPR54 in
the neuroendocrine regulation of reproduction. Annu Rev Physiol 70:213–238
Gottsch ML, Cunningham MJ, Smith JT, Popa SM, Acohido BV, Crowley WF,
Seminara S, Clifton DK, Steiner RA 2004 A role for kisspeptins in the regulation of gonadotropin secretion in the mouse. Endocrinology 145:4073– 4077
Smith JT, Cunningham MJ, Rissman EF, Clifton DK, Steiner RA 2005 Regulation of Kiss1 gene expression in the brain of the female mouse. Endocrinology 146:3686 –3692
Smith JT, Dungan HM, Stoll EA, Gottsch ML, Braun RE, Eacker SM, Clifton
DK, Steiner RA 2005 Differential regulation of KiSS-1 mRNA expression by
sex steroids in the brain of the male mouse. Endocrinology 146:2976 –2984
Smith JT, Popa SM, Clifton DK, Hoffman GE, Steiner RA 2006 Kiss1 neurons
in the forebrain as central processors for generating the preovulatory luteinizing hormone surge. J Neurosci 26:6687– 6694
Adachi S, Yamada S, Takatsu Y, Matsui H, Kinoshita M, Takase K, Sugiura
H, Ohtaki T, Matsumoto H, Uenoyama Y, Tsukamura H, Inoue K, Maeda K
2007 Involvement of anteroventral periventricular metastin/kisspeptin neurons in estrogen positive feedback action on luteinizing hormone release in
female rats. J Reprod Dev 53:367–378
Kinoshita M, Tsukamura H, Adachi S, Matsui H, Uenoyama Y, Iwata K,
Yamada S, Inoue K, Ohtaki T, Matsumoto H, Maeda K 2005 Involvement of
central metastin in the regulation of preovulatory luteinizing hormone surge
and estrous cyclicity in female rats. Endocrinology 146:4431– 4436
Clarkson J, Herbison AE 2006 Postnatal development of kisspeptin neurons
in mouse hypothalamus; sexual dimorphism and projections to gonadotropinreleasing hormone neurons. Endocrinology 147:5817–5825
Kauffman AS, Gottsch ML, Roa J, Byquist AC, Crown A, Clifton DK, Hoffman
GE, Steiner RA, Tena-Sempere M 2007 Sexual differentiation of Kiss1 gene expression in the brain of the rat. Endocrinology 148:1774 –1783
Christian CA, Mobley JL, Moenter SM 2005 Diurnal and estradiol-dependent
changes in gonadotropin-releasing hormone neuron firing activity. Proc Natl
Acad Sci USA 102:15682–15687
Christian CA, Glidewell-Kenney C, Jameson JL, Moenter SM 2008 Classical
estrogen receptor ␣ signaling mediates negative and positive feedback on gonadotropin-releasing hormone neuron firing. Endocrinology 149:5328 –5334
Kauffman AS, Buenzle J, Fraley GS, Rissman EF 2005 Effects of galanin-like
peptide (GALP) on locomotion, reproduction, and body weight in female and
male mice. Horm Behav 48:141–151
Kauffman AS, Rissman EF 2004 A critical role for the evolutionarily conserved
endo.endojournals.org
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
3671
gonadotropin-releasing hormone. II. Mediation of energy status and female
sexual behavior. Endocrinology 145:3639 –3646
Finn PD, Steiner RA, Clifton DK 1998 Temporal patterns of gonadotropinreleasing hormone (GnRH), c-fos, and galanin gene expression in GnRH neurons relative to the luteinizing hormone surge in the rat. J Neurosci 18:713–719
Everett JW, Sawyer CH 1950 A 24-hour periodicity in the “LH-release apparatus” of female rats, disclosed by barbiturate sedation. Endocrinology 47:
198 –218
Siegel HI, Bast JD, Greenwald GS 1976 The effects of phenobarbital and
gonadal steroids on periovulatory serum levels of luteinizing hormone and
follicle-stimulating hormone in the hamster. Endocrinology 98:48 –55
Stetson MH, Gibson JT 1977 The estrous cycle in golden hamsters: a circadian
pacemaker times preovulatory gonadotropin release. J Exp Zool 201:289 –294
Stetson MH, Watson-Whitmyre M 1977 The neural clock regulating estrous
cyclicity in hamsters: gonadotropin release following barbiturate blockade.
Biol Reprod 16:536 –542
Legan SJ, Coon GA, Karsch FJ 1975 Role of estrogen as initiator of daily LH
surges in the ovariectomized rat. Endocrinology 96:50 –56
Legan SJ, Karsch FJ 1975 A daily signal for the LH surge in the rat. Endocrinology 96:57– 62
Norman RL, Blake CA, Sawyer CH 1973 Estrogen-dependent 24-hour periodicity in pituitary LH release in the female hamster. Endocrinology 93:965–
970
Stetson MH, Watson-Whitmyre M 1976 Nucleus suprachiasmaticus: the biological clock in the hamster? Science 191:197–199
Kauffman AS 2009 Sexual differentiation and the Kiss1 system: Hormonal and
developmental considerations. Peptides 30:83–93
de la Iglesia HO, Meyer J, Schwartz WJ 2003 Lateralization of circadian
pacemaker output: activation of left- and right-sided luteinizing hormonereleasing hormone neurons involves a neural rather than a humoral pathway.
J Neurosci 23:7412–7414
Meyer-Bernstein EL, Jetton AE, Matsumoto SI, Markuns JF, Lehman MN,
Bittman EL 1999 Effects of suprachiasmatic transplants on circadian rhythms
of neuroendocrine function in golden hamsters. Endocrinology 140:207–218
Gu GB, Simerly RB 1997 Projections of the sexually dimorphic anteroventral
periventricular nucleus in the female rat. J Comp Neurol 384:142–164
Simerly RB, Swanson LW 1987 The distribution of neurotransmitter-specific
cells and fibers in the anteroventral periventricular nucleus: implications for
the control of gonadotropin secretion in the rat. Brain Res 400:11–34
Kallo I, Vida B, Deli L, Kalamatianos T, Hrabovszky E, Caraty A, Coen CW,
Liposits Z, Evidence that kisspeptin neurons in the AVPV form part of a
bisynaptic pathway connecting suprachiasmatic vasopressin neurons with preoptic GnRH cells. Program of the 38th Annual Meeting of the Society for
Neuroscience, Washington, DC, 2008 (Abstract 618.4)
Palm IF, van der Beek EM, Wiegant VM, Buijs RM, Kalsbeek A 2001 The
stimulatory effect of vasopressin on the luteinizing hormone surge in ovariectomized, estradiol-treated rats is time-dependent. Brain Res 901:109 –116
Palm IF, Van Der Beek EM, Wiegant VM, Buijs RM, Kalsbeek A 1999 Vasopressin induces a luteinizing hormone surge in ovariectomized, estradioltreated rats with lesions of the suprachiasmatic nucleus. Neuroscience 93:
659 – 666
Chappell PE, Goodall CP, Tonsfeldt KJ, White RS, Bredeweg E, Latham KL
2009 Modulation of gonadotropin-releasing hormone (GnRH) secretion by an
endogenous circadian clock. J Neuroendocrinol 21:339 –345
Chappell PE, White RS, Mellon PL 2003 Circadian gene expression regulates
pulsatile gonadotropin-releasing hormone (GnRH) secretory patterns in the
hypothalamic GnRH-secreting GT1-7 cell line. J Neurosci 23:11202–11213
Shughrue P, Scrimo P, Lane M, Askew R, Merchenthaler I 1997 The distribution of estrogen receptor-␤ mRNA in forebrain regions of the estrogen
receptor-␣ knockout mouse. Endocrinology 138:5649 –5652
Shughrue PJ, Lane MV, Merchenthaler I 1997 Comparative distribution of
estrogen receptor-␣ and -␤ mRNA in the rat central nervous system. J Comp
Neurol 388:507–525
De La Iglesia HO, Blaustein JD, Bittman EL 1999 Oestrogen receptor-␣immunoreactive neurones project to the suprachiasmatic nucleus of the female
Syrian hamster. J Neuroendocrinol 11:481– 490