| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Departments of Reproductive Science (N.J.M., J.K.), Medicine (S.C., L.S., J.S., D.J.D.), Psychiatry (F.V.), and Laboratory Medicine and Pathobiology (S.L.A.), Toronto General Hospital (S.C., L.S., J.S., D.J.D.), Center for Addiction and Mental Health (F.V.), Mount Sinai Hospital (S.L.A.), and the Banting and Best Diabetes Center (S.C., L.S., J.S., D.J.D.), University of Toronto, Toronto, Ontario, Canada M5G 2C4
Address all correspondence and requests for reprints to: Dr. D. Drucker, Toronto General Hospital, 200 Elizabeth Street CCRW3–838, Toronto, Ontario, Canada M5G 2C4. E-mail: d.drucker{at}utoronto.ca
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Considerable evidence supports multiple complementary biological actions for GLP-1 in the control of glucose homeostasis. Nutrient intake promotes GLP-1 release from the intestine, which serves to stimulate glucose-dependent insulin secretion from the pancreatic ß-cell (3, 4, 5). GLP-1 also inhibits glucagon secretion and decreases gastric emptying, which further reduce glycemic excursion in vivo. Taken together, these peripheral actions of GLP-1 function promote regulated nutrient assimilation and glucose control in both normal subjects and patients with type 2 diabetes (1).
The role of the PGDPs in the central nervous system (CNS) is less well understood. Original observations demonstrating that GLP-1 stimulates adenylate cyclase activity in hypothalamic and pituitary membrane preparations (6) were followed by the detection of CNS proglucagon messenger RNA (mRNA) transcripts, predominantly in the brain stem (7, 8). After the cloning of distinct PGDP receptors for glucagon, GLP-1, and GLP-2, a combination of molecular techniques has localized the expression of all three PGDP receptors to different regions of the rodent and human CNS (9, 10, 11). These findings have stimulated additional interest in the putative roles of the PGDPs in neuroendocrine systems in vivo.
Considerable experimental data support one or more actions for GLP-1 in the CNS. Intracerebroventricular (icv) administration of GLP-1 inhibits food and water intake in rats and mice (12, 13). Conversely, administration of the GLP-1 receptor antagonist exendin-(9–39) promotes increased food intake in acute, short term studies and increased weight gain in more long term, chronic experiments in rats (12, 14). Whether GLP-1 acts directly to inhibit food intake or induces conditioned taste aversion leading to decreased satiety (15) remains unclear. However, GLP-1 does not appear to be essential for appetite regulation and weight control, as disruption of GLP-1 signaling in mice does not result in increased food intake or long term changes in body weight (13, 16).
More recent studies have suggested that GLP-1 may modulate the activity of the hypothalamic-pituitary axis. GLP-1 increased GnRH secretion from hypothalamic GT-1 cells, and icv injection of GLP-1 increased the levels of plasma LH in rats (17). Furthermore, icv injection of GLP-1 activated c-fos in CRH-positive neurons of the hypothalamic paraventricular nucleus in association with increased levels of plasma corticosterone in rats (18). These findings taken together with studies demonstrating GLP-1 regulation of pituitary TSH release (19) raise the possibility that GLP-1 may modulate hypothalamic-pituitary function at several levels in vivo.
Despite accumulating data suggesting a possible therapeutic role for GLP-1 in patients with type 2 diabetes, the physiological role of GLP-1 in the hypothalamic-pituitary axis remains uncertain. As the majority of previous studies examining this issue have used direct icv or iv administration of pharmacological amounts of GLP-1, the possibility cannot be excluded that such injections may exert nonphysiological effects, possibly via activation of related receptors or nonspecific anorectic or stress responses (20, 21). We have recently examined the phenotypic consequences of disrupting GLP-1 action in vivo, through generation and analysis of GLP-1 receptor knockout (GLP-1R-/-) mice. These mice are viable, but exhibit abnormal glucose homeostasis due to defective glucose-induced insulin secretion. Importantly, GLP-1R-/- mice do not respond to GLP-1 administration, and no GLP-1-binding sites have been detected in the pancreatic islet cells or CNS of GLP-1R-/- mice (13, 22). In the present study we have examined whether the specific loss of GLP-1 signaling is associated with perturbations in regulation of neuroendocrine function in vivo.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Body weights were measured from weaning until adulthood (
6 weeks of
age). Females were inspected daily for vaginal opening as a biological
index of the onset of the first estrous cycle. Vaginal smears were
taken daily at 0900 h for a period of 24 days (six complete
estrous cycles) to monitor cyclicity. Smears were obtained by flushing
50 µl deionized double distilled water gently into the vagina and
immediately retrieving the water with a plastic pipette tip. Droplets
of the recovered fluid were smeared onto precleaned glass slides and
dried for microscopic examination. Females showing regular 4-day cyclic
vaginal smear patterns were killed between 1000–1200 h on the
appropriate days of the cycle by asphyxiation in a chamber filled with
CO2 gas followed immediately by exsanguination
via cardiac puncture using a 23-gauge needle and decapitation. Blood
was allowed to clot in a cold room at 4 C, and serum was separated by
centrifugation for 1 min in a microfuge and stored frozen at -20 C. An
additional group of female mice was allowed to remain in the TGH animal
care facility until they were 90–100 days of age before vaginal
smearing was initiated, to determine whether cyclicity was maintained
in GLP-1R-/- animals. Approximately 1 month later (at
120–125 days of age) the animals were killed, serum was collected, and
the ovaries were processed for histological examination.
At death in the majority of cases, male and female gonads and steroid-dependent accessory sex glands (epididymis, prostate gland, seminal vesicle, and uteri) were dissected free of surrounding fatty tissue, frozen over a sheet of aluminum foil covering dry ice, and weighed on a Sartorius Supermicro Balance (Sartorius, Goettingen, Germany). Where tissues were to be examined histologically, the organs were placed immediately after removal in buffered formalin solution and fixed overnight at room temperature before embedding in paraffin. Sections (4 µm) were prepared, stained with hematoxylin and eosin, and examined under a light microscope at x100 magnification. Counts of numbers of antral follicles and corpora lutea were performed in full thickness sections prepared from ovaries from animals killed at the proestrous stage of the estrous cycle.
Because of the likelihood that circulating corticosterone levels would be influenced by the degree of stress to which the animals were exposed before death, groups of male mice were killed for serum corticosterone determination under four different conditions: 1) nonstressed: animals were withdrawn one by one from their home cages and killed immediately in the animal room; 2) environmental stress A: animals were removed from their home cages, transported in a holding cage approximately 50 ft down a corridor to the necropsy laboratory, and killed there, approximately 2–4 min elapsing between removal from the animal room and death; 3) environmental stress B: animals were removed from the animal care facilities in their home cages, transported up five floors in an elevator to laboratories on the fourth floor of the Max Bell Research Center of the Toronto Hospital Research Institute and killed there, approximately 1–2 h after arrival; and 4) anesthetic stress: mice were anesthetized with metofane, and 10 min later while they were still under anesthesia, blood was collected.
To examine the regulation of plasma osmolarity, age- and sex-matched GLP-1R-/- and wild-type+/+ mice were injected sc for 7 days with either 200 µl vehicle (saline) or hydrocortisone (10 mg) and subsequently either permitted access to drinking water ad libitum or denied access to water for the last 13 h before death. At death, blood was withdrawn via cardiac puncture into heparanized tubes, and plasma was separated by centrifugation. Plasma osmolarity was assessed using a vapor pressure osmometer (Wescor, Inc., Logan, UT).
RIAs
Serum T4 was assayed using a kit purchased
from ICN Biomedicals, Inc. (catalogue no.
06B254011; Montreal, Canada). Testosterone, progesterone, estradiol,
and corticosterone levels were measured using kits purchased from
Diagnostic Products (Los Angeles, CA; catalogue no. TKTT1,
KE2D1, TKPG1, and TKRC1, respectively). For the estradiol assay, 100-
to 200-µl aliquots of serum were extracted into 2 ml fresh anhydrous
diethyl ether, evaporated to dryness, and reconstituted in 0.1% human
serum albumin to eliminate interfering substances present in rodent
serum. The remaining hormones were assayed directly in serum, according
to the kit instructions.
Immunocytochemistry
Immunocytochemistry was performed on the anterior pituitaries of
GLP-1R-/- and GLP-1R+/+ mice to determine the
distribution and relative abundance of each of the major
adenohypophyseal cell types as follows. Paraffin-embedded sections
(4–5 µm thick) were stained with hematoxylin and eosin as well with
the Gordon-Sweet silver method to demonstrate the reticulin fiber
network. Immunocytochemical stains to localize adenohypophysial
hormones were performed using the streptavidin-biotin-peroxidase
complex technique. Primary antisera directed against rat pituitary
hormones were used at the specified dilutions with overnight
incubations: GH, 1:2500; PRL, 1:2500; TSHß, 1:3000; FSHß, 1:600;
LHß, 1:2500 (National Hormone and Pituitary Program, Rockville, MD);
and ACTH prediluted preparation, further diluted 1:20 (DAKO Corp., Carpinteria, CA). The reaction products were visualized
with the Ultrastreptavidin kit (Signet Laboratories, Inc., Dedham, MA)
and 3,3'-diaminobenzidine.
In situ hybridization histochemistry
In situ hybridization histochemistry for hypothalamic
CRH mRNA was performed on adult male GLP-1R-/- and
GLP-1R+/+ mice, 7–9 weeks of age. Four experimental groups
of animals were employed: fed (ad libitum access to chow and
water) wild-type mice, fed GLP-1R-/- mice, 24-h fasted
wild-type mice (ad libitum access to water only), and 24-h
fasted GLP-1R-/- mice. Antisense riboprobes were created
using an in vitro transcription kit (Promega Corp., Madison, WI) using complementary DNA for CRH (donated by
Dr. Joel Elmquist). Sense and antisense complementary RNA probes were
transcribed from 1 µg linearized DNA template, and in situ
hybridization was carried out using
[35S]UTP-labeled complementary RNA probes as
previously described (23).
Briefly, mice were deeply anesthetized with sodium pentobarbital (90 mg/kg BW, ip) at 1000 h. The animals were then perfused transcardially with 0.9% saline, followed by 10% neutral buffered formalin (Sigma, St. Louis, MO). The mice were fixed for 1 h at room temperature before the brains were removed and postfixed for additional 4 h in 10% neutral buffered formalin. Frozen sections (25 µm) in the coronal plane were obtained using a Reichert cryostat at -18 C. Consecutive sections through the paraventricular nucleus were collected from each brain and stored in cryoprotectant (5% neutral buffered formalin, 30% ethylene glycol, and 20% glycerol). Sections were mounted onto slides (SuperFrost Plus Slides, Fisher Scientific, Fairlawn, NJ) and stored at -20 C in an air-tight container until the in situ hybridization procedure (<1 week). Slides were postfixed in 37% (wt/vol) formaldehyde for 5 min, rinsed twice in 1 x PBS, incubated in 0.5 µg/ml proteinase K buffer [0.5 M EDTA, 1.0 M Tris (pH 8.0), and diethylpyrocarbonate-water] for 30 min at 37 C, rinsed with water and triethoalemine-HCl, acetylated with acetic anhydride (0.25%) in triethoalemine-HCl, dehydrated in an ethanol series (70%, 80%, 95%, and 100% ethanol; chloroform; and 100% and 95% ethanol), air-dried, and stored at -20 C until hybridization.
Sections were hybridized with [35S]CRH riboprobe in a hybridization buffer [50% formamide, 0.1% SDS, 0.01% thiosulfate, 0.1 M dithiothreitol, 0.6 M NaCl, 10 mM Tris (pH 7.5), 1 mM EDTA, 5% dextran sulfate, 0.01% sheared salmon sperm DNA, 0.05% total yeast transfer RNA, 0.01% yeast transfer RNA, and 1x Denhardt’s solution] in a hybridization chamber consisting of tissue culture trays (VWR, Bridgeport, NJ), lined with chromatography paper saturated with a 50% formamide-4 x SSC (standard saline citrate) solution. The chamber was placed in a 58 C oven for 18.5 h. The slides were then washed four times in 2 x SSC and treated with ribonuclease A (20 µg/ml) for 30 min at 37 C. The sections were rinsed in buffer containing 0.5 M NaCl, 40 mM Tris-HCl (pH 8.0), and 1 mM EDTA (pH 8.0), then washed in 300 mM NaCl, 30 mM citrate (2 x saline-sodium citrate solution or 2 x SSC) and 1 M dithiothreitol (DTT) for 1 h at 50 C; 0.2 x SSC and 1 M DTT for 1 h at 55 C; and 0.2 x SSC and 1 M DTT for 1 h at 60 C. The slides were dehydrated in ethanol containing 0.3 M ammonium acetate and 1 mM DTT, air-dried for 1 h, and exposed to Biomax MR film (Eastman Kodak Co., Rochester, NY) for 48 h together with a 14C-labeled standard (Amersham Pharmacia Biotech, Markham, Canada) to verify the linearity of the film response and the consistency of signal detection across films.
Slides were dipped in a 1:1 mixture of NTB2 emulsion-water (Kodak) in light-tight and high humidity conditions, dried overnight, and exposed at 4 C for 2 weeks. The slides were developed, counterstained with thionin for 2–3 min, dehydrated in a 50%, 70%, and 95% ethanol series with acetic acid, two changes of 95% ethanol, two changes of 100% ethanol, and xylenes before coverslipping with Permount medium (Fisher Scientific). The emulsion-coated slides were analyzed using a Leitz light microscope (Rockleigh, NJ). Photomicrographs were generated by capturing images with a digital camera mounted directly on the microscope, and Apple Macintosh Image editing software (Apple Corp., Cupertino, CA) was used to transfer the images to a Power Point program for display.
Elevated plus maze test
The elevated plus maze consisted of two opposite open arms
(50 x 10 cm) and two arms with walls (50 x 10 x 40
cm) attached to a central platform (10 x 10 cm) to form a cross
and elevated 65 cm above the ground. The maze floor was constructed
from white Plexiglas, and the walls were made of black Plexiglas. Black
lines on the maze floor defined the center square and divided both the
open and closed arms into two equal segments, designated close and far
based on proximity to the center square. Tests were carried out during
the light phase of the light-dark cycle. Testing commenced by placing
each mouse on the central platform facing an open arm. The
GLP-1R-/- and wild-type CD1 mice were tested
in alternating order. After each 5-min test, the maze was cleaned. The
behavior of the animals in the maze was videotaped under white light
illumination. The number and duration of closed arm and open arm
entries with both forepaws were recorded and analyzed as previously
described (24).
Acoustic startle responses
Amplitudes of startle reactions were determined using an
acoustic startle apparatus obtained from MED Associates (St. Albans,
VT). The mouse startle chamber was housed within a sound-attenuated
acoustic startle cubicle (54.6 x 50.8 x 30.5 cm, interior
dimensions) with a ventilation fan. The chamber was constructed of
stainless steel rods suspended between Plexiglas plates fixed to a
Plexiglas base. Interior dimensions of the chamber were 5.1 x
3.8 x 3.8 cm, providing the animal some movement. The magnitude
of the startle response was measured by the startle platform for a
period of 500 msec from presentation of the acoustic stimuli and
subsequently transduced and collected onto a PC using Startle Reflex
Software for Windows (MED Associates). Acoustic white noise
stimuli were presented through a programmable audio stimulator (MED
Associates, St. Alban’s, VT). Background white noise levels were
maintained at 70 dB throughout testing. Sixteen-week-old male
GLP-1R-/- (n = l0), age-matched wild-type
CD1 (n = 11), and heterozygous GLP-1R+/- (n =
11) mice were tested in alternating order within the 4 acoustic startle
chambers. Mice were placed in the startle chamber for a 5-min
acclimation period with a background noise level of 70 dB. After the
acclimation period, mice were exposed to acoustic startle intensities
of 83, 85, 90, 100, and 120 dB for a duration of 30 msec in random
order. Testing consisted of 40 trials of the 5 randomly presented
stimulus intensities. Each mouse was therefore exposed to a total of
200 acoustic stimuli, or 40 at each of the 5 intensities. The startle
amplitude at each stimulus intensity was defined as the average of 40
readings. The time interval between stimuli was varied, lasting from
12–18 sec. Five weeks after the initial startle testing, some of the
same experimental groups of GLP-1R-/- (n
= 8), GLP-1R+/- (n = 8), and GLP-1R+/+ (n = 8)
mice received 0.5 mg/ml hydrocortisone sodium succinate (Abbott Laboratories, Saint-Laurent, Canada) in ad libitum
drinking water for a period of 6 days, after which the acoustic startle
test was repeated. After this final test, the animals were killed in
the biochemistry laboratory, and serum was collected for corticosterone
assay under conditions identical to those described for the
environmental stress B group described above.
Statistical analysis
All data are presented as the mean ± SEM. For
most experiments, statistical analysis was performed using
PC-compatible microcomputer programs [SPSS for Windows (SPSS, Inc., Chicago, IL) and SigmaStat (Jandel Scientific, San Rafael,
CA)]. The data from the maze tests were analyzed on an Apple MacIntosh
computer using INSTAT 1.12 (GraphPad Software, Inc., San
Diego, CA). Two group comparisons were made using unpaired Student’s
t tests. Multiple group comparisons were made using ANOVA,
followed by the Newman-Keuls test for individual group comparisons. Log
transformation of the data were used to eliminate inhomogeneity of
variance where necessary.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, A
and B). In contrast, adult female GLP-1R-/- mice were
slightly but significantly lighter than GLP-1R+/+ female
controls (Fig. 1C), consistent with the results of previous studies
(13, 16, 25). The animals were fertile, with no obvious reproductive
deficits. Growth rates in littermate male and female mice, assessed
from weaning to adulthood, showed the expected postpubertal sex
differences, but were not significantly different between
GLP-IR-/- animals and wild-type controls reared under
identical conditions (Fig. 1A).
|
; P
< 0.05, +/+ vs. -/- mice). Adrenal weights were also
significantly reduced in female GLP-1R-/- mice. However, no
statistically significant differences were observed in the weights of
the ovaries or uteri of GLP-1R-/- vs.
GLP-1R+/+ mice (Fig. 1C). Similarly, pituitary weights were
comparable in male and female GLP-1R+/+ and
GLP-1R-/- animals (Fig. 1, B and C). Histologically, we did
not observe any qualitative changes in the hypothalamus (not shown) or
pituitaries of GLP-1R-/- mice (Fig 2).
|
) and gonadotrophs (not shown), in GLP-1R-/- mice.
Despite the significant decrease in adrenal weights, adrenal histology
was normal in GLP-1R-/- mice, with appropriate numbers of
cell types and normal zonal distribution observed in adrenal glands
from both male and female GLP-1R-/- mice (Fig. 2B).
Similarly, despite a significant decrease in testicular weight, the
histological architecture and cellular organization appeared comparable
in GLP-1R+/+ and GLP-1R-/- mice even at 120–125
days of age (Fig. 2B). Although ovarian histology appeared grossly
normal in GLP-1R-/- mice, a small, but significant,
decrease in the mean number of developing follicles was observed in
GLP-1R-/- ovaries (Fig. 3).
The mean number of corpora lutea was also slightly lower in
GLP-1R-/- mice, although this difference did not attain
statistical significance.
|
). Cyclicity in female GLP-1R-/- mice
was normal once vaginal opening had occurred and was indistinguishable
from that in female +/+ control mice. A total of 20 wild-type +/+ and
10 GLP-1R-/- females were studied. All exhibited regular
4-day estrous cycles throughout the 24 days of the initial smearing
period. Of the animals that were allowed to reach 120–125 days of age,
the great majority of both genotypes continued to exhibit regular 4-day
estrous cycles until the day of death
(GLP-1R-/- mice, 10 cycling and 1 acyclic;
GLP-1R+/+ animals, 12 cycling and 1 acyclic).
Both animals that stopped cycling exhibited persistent estrous vaginal
smears, consistent with previous studies on hypothalamic failure in
aging mice (26).
Hormone levels
In rats, GLP-1 may modulate LH, TSH, and CRH
secretion (17, 18, 19). Therefore, we assessed the levels of thyroid
hormone as well as the principal gonadal and adrenal steroids in
GLP-1R+/+ and GLP-1R-/- mice. No differences in
the levels of testosterone or thyroid hormone
(T4) were detected in age-matched
GLP-1R+/+ and GLP-1R-/- males (Fig 4). Serum estradiol and progesterone
measurements in cycling females were also comparable in
GLP-1R+/+ and GLP-1R-/- animals.
Although there was a trend toward lower progesterone levels in the
GLP-1R-/- mice, this difference was not
statistically significant on any individual day of the estrous cycle
(Fig. 4). No significant differences in basal morning circulating
corticosterone levels were observed between GLP-1R-/- and
GLP-1R+/+ control mice of either sex despite the small, but
significant, decrease in adrenal mass in GLP-1R-/- mice
(Fig. 5, A and B). In contrast,
corticosterone levels were significantly greater in male
GLP-1R-/- mice killed under different stress paradigms
(Fig. 5B).
|
|
). The highest density of
labeled cells was observed in the lateral and medial magnocellular
division, but many labeled cells were also seen in the medial
parvocellular division. There were no significant differences in the
intensity of the CRH signal in either fed (n = 4) or fasted
(n = 4) GLP-1R+/+ vs. GLP-1R-/-
mice (Fig. 6, A–D).
|
To exclude the confounding variable of dehydration impacting on results
of behavioral studies, plasma osmolarity was assessed. Osmolarity was
normal in GLP-1R-/- mice and was slightly reduced after
several days of glucocorticoid administration (Fig. 7A). Nevertheless, plasma osmolarity rose
appropriately, but not excessively, after 13 h of water
deprivation in both saline- and glucocorticoid-treated
GLP-1R-/- mice. These findings demonstrate that
GLP-1R-/- mice do not exhibit basal abnormalities in
osmoregulation and are able to sense and respond to fluid restriction,
indicative of a functionally intact vasopressin response in
vivo (Fig. 7A).
|
). However,
GLP-1R-/- mice spent a significantly greater
amount of time in the center square of the maze compared with wild-type
controls (P < 0.01; Fig. 7C). Significant differences
between the genotypes were also observed in terms of acoustic startle
response. At low stimulus intensities (<100 dB), the
GLP-1R-/- animals appeared to be less
responsive than the wild-type controls. However, at the highest
acoustic stimulus intensity (120 dB), this difference was reversed
(P < 0.01; Fig. 7D). The heterozygous +/- mice
exhibited responses midway between, but not significantly different
from, those for the -/- and +/+ animals. To test whether the mild
derangement in hypothalamic-pituitary-adrenal regulation, as
denoted by decreased adrenal weights and heightened glucocorticoid
responses to stress in GLP-1R-/- mice, might contribute to
the observed differences in acoustic startle, the same groups of mice
were retested after the addition of hydrocortisone to their drinking
water for 6 days. Glucocorticoid replacement significantly diminished
the difference between the genotypes in the startle response at high
stimulus intensities to below the limits of statistical significance
(Fig. 7D). |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As GLP-1 is being tested as a potential therapeutic agent for the treatment of diabetes, and peripheral administration of GLP-1 modulates vasopressin and cortisol secretion in vivo, the potential neuroendocrine actions of GLP-1 may have relevance for understanding the consequences of GLP-1 administration in vivo. The findings that male GLP-1R-/- mice exhibit modest decreases in gonadal weight and female mice exhibit a slight delay in the onset of puberty are consistent with a possible effect of GLP-1 on central regulation of the hypothalamic-pituitary-gonadal axis. Our previous findings demonstrated normal growth curves in adult male and female GLP-1R-/- mice, and studies of younger mice reported here demonstrate normal growth in GLP-1R-/- mice from birth to adulthood. Thus, the mild reproductive abnormalities observed in GLP-1R-/- mice cannot be ascribed, as previously suggested (17), to disturbances of energy homeostasis impacting on the reproductive axis. Furthermore, although pharmacological injection of icv GLP-1 inhibits feeding in rodents (12), GLP-1R-/- mice are not obese and do not exhibit short or long term disturbances in control of food intake or body weight (13, 16). Hence, it seems unlikely that reproductive abnormalities in GLP-1R-/- mice can be linked to dysregulation of nutrient intake.
Our analysis of reproductive function in GLP-1R-/- mice was motivated in part by the demonstration of GLP-1R on hypothalamic GT1 cells and the findings that icv GLP-1 stimulates LH secretion (17). These observations raised the possibility that GLP-1 may control pubertal development through a direct effect on hypothalamic GnRH secretion. We did not observe any overt changes in the reproductive function of GLP-1R-/- mice, and the number and distribution of anterior pituitary cell types, including corticotrophs and gonadotrophs, were normal in GLP-1R-/- mice. Although a consistent delay in the onset of the first ovarian cycle was found in the GLP-1R-/- animals, once cyclicity was established no differences between the -/- and wild-type +/+ mice were observed. The observation that GLP-1R-/- mice are fertile and do not exhibit impairment of reproductive capacity taken together with the normal circulating estradiol and progesterone levels clearly suggest that GLP-1 signaling is not essential for successful induction of ovulation, gonadal steroid production, or reproduction in vivo. In both +/+ and -/- mice, the majority of females exhibited regular 4-day cycles, which continued until the animals were at least 120 days of age. The slight reduction in circulating progesterone levels observed in the female GLP-1R-/- mice may reflect the marginal effect on the number of ovarian follicles observed at histological examination.
Although we did not specifically test sexual behavior, the fact that GLP-1R-/- mice breed normally suggests that loss of GLP-1 signaling does not result in a major reproductive behavioral deficit. In males, developmental masculinization apparently occurs normally, consistent with their normal growth curves and apparently normal reproductive performance in the breeding program. In studies not presented here, we found that sex differences in estrogen receptor distribution within the hypothalamus and preoptic area also develop normally in GLP-1R-/- mice; hence, loss of GLP-1 signaling does not interfere with sexual differentiation of the brain (Maclusky, N. J., J. Kim, and D. J. Drucker, unpublished observations).
Previous studies have yielded conflicting data on the interaction between GLP-1 and vasopressin in the rat CNS. The finding of GLP-1R and vasopressin RNA colocalization in magnocellular neurons of the paraventricular and supraoptic nuclei (27) taken together with the induction of c-Fos immunoreactivity in the paraventricular nucleus (PVN) after icv GLP-1 (18) implies an anatomical and functional relationship between these two peptides. Nevertheless, icv injection of GLP-1 stimulated vasopressin secretion (18), whereas iv GLP-1 infusion inhibited vasopressin secretion (27). Our studies did not reveal any major impairment of osmoregulation in normal or fluid-restricted GLP-1R-/- mice before or after glucocorticoid treatment. Although these findings do not exclude the possibility that hypothalamic GLP-1R signaling in the PVN may contribute to physiological regulation of vasopressin release, GLP-1R signaling does not appear essential for sensing and/or responding to changes in plasma osmolarity in vivo.
The finding that centrally administered GLP-1 activates c-fos activity in CRH hypothalamic neurons (18) raises the possibility that the GLP-1R may represent an upstream component of the systems regulating the CRH-ACTH-adrenal axis. The icv administration of GLP-1 increases corticosterone secretion in rats (18), and iv GLP-1 infusion increased both ACTH and plasma cortisol in human subjects (30), consistent with a role for GLP-1 in ACTH-dependent glucocorticoid secretion. Although disruption of GLP-1 signaling might conceivably affect development of the hypothalamic CRH+ neurons and/or pituitary corticotrophs, thereby decreasing adrenal weight, these components of the HPA axis appear to develop normally in GLP-1R-/- mice. Furthermore, although adrenal weights are slightly lower in male and female GLP-1R-/- mice, levels of basal and stress-stimulated plasma corticosterone were clearly not reduced in the absence of GLP-1 signaling. In fact, stress-induced adrenal glucocorticoid secretion was paradoxically increased in GLP-1R-/- animals compared with that in wild-type controls killed under the same conditions.
The mechanisms underlying these subtle differences in corticosterone responses and adrenal weights remain unknown. Our initial hypothesis, based on findings that icv GLP-1 injections stimulated glucocorticoid release (18), suggested that abrogation of GLP-1 signaling might reduce basal and/or stress-induced corticosterone secretion. This hypothesis is obviously no longer tenable. Our data are consistent with the possibility that GLP-1 is involved physiologically in the regulation of CRH release; however, it remains unclear why disruption of GLP-1R signaling augments, rather than decreases, stress-activated adrenal corticosterone secretion. In wild-type animals, many of the systems that provide input to the PVN, including brain stem catecholaminergic neurons, the medial septum, and the hippocampus, also normally contain GLP-1R (31). The effects of GLP-1 on some of these centers could indirectly inhibit CRH release, resulting in enhanced stress responses in animals lacking GLP-1R. It is also possible that the lack of stimulatory input from GLP-1 to the PVN could have resulted developmentally in a compensatory disinhibition of CRH release in GLP-1R-/- mice. For example, GLP-1R-/- may exhibit a modest reduction in corticosterone feedback sensitivity, thereby normalizing basal CRH and circulating corticosterone levels. Under nonstress conditions, this would not result in any detectable phenotype. Under conditions of mild stress, however, such disinhibition might result in exaggerated CRH, and therefore corticosterone, release.
Despite the hypothesis that CRH expression may be mildly perturbed in GLP-1R-/- mice, we did not observe any significant difference in the expression of CRH mRNA between wild-type and GLP-1R-/- animals, with or without food restriction. Although the data from the plus maze and acoustic startle tests might be interpreted as consistent with an increase in intracerebral CRH release in GLP-1R-/- mice, they are inconclusive, as subtle abnormalities in multiple neuroendocrine circuits may ultimately influence the results of these test paradigms. In rats, icv CRH administration enhances acoustic startle responses (32). In both rats (33) and mice (34), icv CRH injections are anxiogenic in the elevated plus maze, reducing the number of entries into the open compared with the closed arms.
In the present study no significant differences were observed between +/+ and -/- mice in the relative numbers of entries into the open and closed arms of the plus maze. The only statistically significant effect was a reduction in the time spent in either the open or closed arms, with a concomitant increase in the time spent in the center square. This difference could indicate a general effect of the GLP-1R-/- phenotype on general locomotor or exploratory behavior, as opposed to a specific response. In the acoustic startle test, an increase in response with increasing sound stimulus intensity was observed in the GLP-1R-/- mice, consistent with an enhanced anxiogenic effect in these animals. However, differences between the GLP-1R-/- and GLP-1R+/+ mice were observed at both the lowest and the highest sound stimulus intensities, suggesting that the GLP-1R-/- mice not only respond differently to increased sound intensity, but also exhibit a different baseline level of activity in this test.
Whatever the underlying mechanisms, loss of GLP-1 function, as exemplified by targeted disruption of the gene encoding the GLP-1 receptor, clearly produces a subtle, but detectable, neuroendocrine phenotype, consistent with a putative physiological role for GLP-1 in the brain. Although GLP-1R-/- mice appear completely normal with respect to basal regulation of growth rates, thyroid, adrenal, and reproductive function, they do not respond normally to the challenge of a novel environment. The enhancement of corticosterone release in response to stress as well as the changes observed in the acoustic startle and elevated plus maze responses, all suggest that loss of GLP-1 signaling impairs CNS-pituitary-adrenal function. Furthermore, recent evidence suggests that systemic stress also activates GLP-1 neurons in the rat central nervous system, providing additional support for linkage between GLP-1 action and central regulation of the stress response in vivo (21). Under normal physiological conditions, therefore, GLP-1 may represent a modulatory component of the mechanisms regulating central and peripheral neuroendocrine and behavioral responses to stress.
|
Footnotes |
|---|
2 Present address: Center for Reproductive Sciences, Columbia
University, Columbia-Presbyterian Medical Center, 622 West 168th
Street, New York, New York 10032.
3 Senior Scientist of the Medical Research Council of Canada.
Received September 14, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-TSH cells. Endocrinology 137:4130–4138[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  G. G. Gosman, H. I. Katcher, and R. S. Legro Obesity and the role of gut and adipose hormones in female reproduction Hum. Reprod. Update, September 1, 2006; 12(5): 585 - 601. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. C. Riddle and D. J. Drucker Emerging Therapies Mimicking the Effects of Amylin and Glucagon-Like Peptide 1 Diabetes Care, February 1, 2006; 29(2): 435 - 449. [Full Text] [PDF]
|
|
|
|
 |
|
 Genetically Modified Animals in Endocrinology Endocr. Rev., August 1, 2004; 25(4): 673 - 677. [Full Text] [PDF]
|
|
|
|
 |
|
 J. A. Lovshin, Q. Huang, R. Seaberg, P. L. Brubaker, and D. J. Drucker Extrahypothalamic Expression of the Glucagon-Like Peptide-2 Receptor Is Coupled to Reduction of Glutamate-Induced Cell Death in Cultured Hippocampal Cells Endocrinology, July 1, 2004; 145(7): 3495 - 3506. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B.K.C. Chow, K.H. Cheung, E.M.W. Tsang, M.C.T. Leung, S.M.Y. Lee, and P.Y.D. Wong Secretin Controls Anion Secretion in the Rat Epididymisin an Autocrine/Paracrine Fashion Biol Reprod, June 1, 2004; 70(6): 1594 - 1599. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. P. Kinzig, D. A. D'Alessio, J. P. Herman, R. R. Sakai, T. P. Vahl, H. F. Figueiredo, E. K. Murphy, and R. J. Seeley CNS Glucagon-Like Peptide-1 Receptors Mediate Endocrine and Anxiety Responses to Interoceptive and Psychogenic Stressors J. Neurosci., July 16, 2003; 23(15): 6163 - 6170. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. E. Mayo, L. J. Miller, D. Bataille, S. Dalle, B. Goke, B. Thorens, and D. J. Drucker International Union of Pharmacology. XXXV. The Glucagon Receptor Family Pharmacol. Rev., March 1, 2003; 55(1): 167 - 194. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. B. Speert, S. J. MCClennen, and A. F. Seasholtz Sexually Dimorphic Expression of Corticotropin-Releasing Hormone-Binding Protein in the Mouse Pituitary Endocrinology, December 1, 2002; 143(12): 4730 - 4741. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Perry, D. K. Lahiri, D. Chen, J. Zhou, K. T. Y. Shaw, J. M. Egan, and N. H. Greig A Novel Neurotrophic Property of Glucagon-Like Peptide 1: A Promoter of Nerve Growth Factor-Mediated Differentiation in PC12 Cells J. Pharmacol. Exp. Ther., March 1, 2002; 300(3): 958 - 966. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. P. H. Burbach, S. M. Luckman, D. Murphy, and H. Gainer Gene Regulation in the Magnocellular Hypothalamo-Neurohypophysial System Physiol Rev, July 1, 2001; 81(3): 1197 - 1267. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. J. Drucker Minireview: The Glucagon-Like Peptides Endocrinology, February 1, 2001; 142(2): 521 - 527. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Baggio, T. J. Kieffer, and D. J. Drucker Glucagon-Like Peptide-1, But Not Glucose-Dependent Insulinotropic Peptide, Regulates Fasting Glycemia and Nonenteral Glucose Clearance in Mice Endocrinology, October 1, 2000; 141(10): 3703 - 3709. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
,
P < 0.05 vs. results obtained in
the same genotype at the 83-db stimulus intensity (Newman-Keuls
post-hoc comparison). Acoustic startle responses were
repeated (lower panel, hydrocortisone treated) in the
same groups of mice after 6 days of treatment with hydrocortisone
sodium succinate (0.5 mg/ml) in the drinking water.