| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
APPETITE, OBESITY AND METABOLISM
Neurology Service, Veterans Affairs Medical Center, East Orange 07018; and Department of Neurosciences, New Jersey Medical School, Newark, New Jersey 07103
Submitted 6 February 2003 ; accepted in final form 17 June 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
metabolism; hypothalamo-pituitary-adrenal axis; corticotropin-releasing hormone; food intake; glucocorticoid receptor; corticosterone; norepinephrine; leptin receptor
The hypothalamo-pituitary-adrenal (HPA) axis plays a pivotal role in the response to stress and the control of body weight and fat distribution. Stress increases PVN CRH release (45). This increases pituitary release of ACTH with subsequent increase in adrenal cortical release of glucocorticoids (49). Glucocorticoids then feed back to inhibit CRH and ACTH release (16) by acting primarily on glucocorticoid (GR) and mineralocorticoid (MR) receptors (10, 44). GR are most responsive to feedback inhibition by high levels of corticosterone in rats (2) and are expressed prominently in the hippocampus, a brain area that plays a role in learning and memory (30, 57), as well as the adaptive response to stress (51). Repeated stress and chronic elevation of glucocorticoids are associated with remodeling of dendrites of hippocampal pyramidal neurons (51), and this may contribute to habituation, possibly by reducing GR expression or function (49). On the other hand, acute stress may sensitize animals such that they have an increased response to subsequent stressors. This might occur at the level of hippocampal GR and other stress-related neural systems to induce long-term alterations in body weight, neurohumoral function, and behavioral responses to a subsequent stressful event.
In light of our prior studies showing hyporesponsiveness and/or habituation to the effects of chronic repeated stress on energy homeostasis in selectively bred DIO rats (42), we postulated that outbred DIO-prone rats would have an inherent difference in their central pathways mediating acute stress responsivity, which would antedate their development of obesity. Because acute stress can have dramatic effects on subsequent body weight gain (28, 59), we also postulated that the response to a single acute stressor might lead to differential long-term effects on body weight gain between obesity-prone and -resistant rats. To test this hypothesis, we assessed the effect of a single stress in chowfed outbred DIO- and DR-prone rats on their subsequent regulation of energy homeostasis and stress responsivity when fed either chow or an HE diet. In this case, outbred DIO-prone and DR rats were used since their propensity to be obesity prone or resistant can be predicted in adulthood, before the introduction of HE diet and the development of their weight gain phenotypes (34). Thus any differences in stress responsivity would reflect a preexisting trait, unaffected by diet or the presence of obesity.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
Experimental protocol. At 1300, 16 DIO- and 16 DR-prone chow-fed rats were placed in a Plexiglas restrainer, and blood was quickly collected by tail nip for assay of basal plasma corticosterone (Basal). Restraint stress was performed in the middle of the light cycle because the long-term effects on energy homeostasis are greater when animals are stressed at this time (59). We also wished to match the timing of stress to that used in our prior chronic stress study (42). Thus the major goal here was to time the stress to have its major impact on body weight while expecting that timing to have relatively little impact on the plasma corticosterone response to stress (32). After 18 min of restraint, a second blood sample was drawn (Stress1), and the rats were released after a total 20 min of restraint. The remaining 16 DIO- and 16 DR-prone rats were left undisturbed and served as unstressed controls. Then eight of the stressed and eight of the control DIO- and DR-prone rats were continued on chow, while eight of each phenotype were switched to HE diet (Research Diets no. C11024 [GenBank] F, New Brunswick, NJ). HE diet contains 4.47 kcal/g with 21% of the metabolizable energy content as protein, 31% as fat, and 48% as carbohydrate, 50% of which is sucrose (41). Rats were weighed daily starting 2 days before the first stress. After 9 days, the previously stressed rats were again stressed (Stress2) at 1300 with blood drawn 18 min into the restraint. We did not draw basal corticosterone levels before Stress2 because of the concern that too much stress would have an additive effect on the terminal GR and CRH mRNA measures.
On the 10th day, previously stressed rats had their open-field activity measured, in a randomized order, over a 15-min period between 1000 and 1400, in a 40 x 40 x 30.5-cm 16-beam Digiscan open-field apparatus (Omnitech). This was done to evaluate their anxiety level since stress has been shown to reduce the activity level (23, 53). Immediately after open-field testing, rats were decapitated, and brains were removed and quickly frozen on dry ice. Control rats were also killed on the 10th day with time of death randomized across all eight experimental groups.
Assays. Urine NE was assayed by high-performance liquid chromatography with electrochemical detection as previously described (34). Tail blood samples were collected into heparinized tubes, and the plasma was removed for corticosterone assay using a double-antibody radioimmunoassay kit (ICN Biomedicals).
In situ hybridization for CRH, GR, and leptin receptor (Ob-R)
mRNA. Fresh frozen brains were cut in five sequential sets of 15-µm
sections. Brain sections were mounted onto gel-coated slides that were then
fixed for 20 min in paraformaldehyde (4%) and kept frozen at -80°C until
they were assayed. In situ hybridization was used to localize CRH, GR, and
Ob-R mRNA expression using modifications of previously described methods
(33,
60). Briefly, brain sections
were thawed for 15 min and then were placed in paraformaldehyde (4%) for 20
min, digested for 30 min at 37°C with proteinase K (10 µg/ml in 100 mM
Tris·HCl containing 50 mM EDTA, pH 8.0), acetylated with acetic
anhydride (0.25% in 0.1 M triethanolamine, pH 8.0), and dehydrated through
graded concentrations (50, 70,
95, and 100%) of alcohol. After vacuum drying for at least 2 h, 90 µl of
hybridization mixture containing an antisense 35S-labeled cRNA
probe (
107 cpm/ml) were spotted on each slide. The slides were
sealed under a coverslip and incubated overnight at 60°C in a water bath.
The next day, the coverslips were removed, and slides were rinsed four times
with 4x SSC (0.6 M NaCl, 60 mM sodium citrate buffer, pH 7.0), digested
for 30 min at 37°C with RNase A (20 µg/ml in 10 mM Tris·500 mM
NaCl containing 1 mM EDTA), washed in descending concentrations of SSC
(2x, 10 min; 1x, 5 min; 0.5x, 5 min; 0.1x, 30 min at
60°C), and dehydrated through graded concentrations of alcohol. After a
2-h period of vacuum drying, the slides were apposed on an X-ray film (Eastman
Kodak, Rochester, NY) for a period of 2-21 days depending on the probe.
Antisense 35S-labeled cRNA probes. The CRH cRNA probe
was generated from the 1.2-kb EcoRI fragment of rat CRH cDNA (D.
Richard, Université Laval, Québec, Canada) subcloned into a
pGEM4 vector (Stratagene, La Jolla, CA) and linearized with HindIII
and EcoRI for antisense and sense probes, respectively. The GR cRNA
probe was generated from EcoRI fragment of rat GR cDNA (D. Richard)
subcloned into a pGEM4 vector (Stratagene) and linearized with BamHI
and XBaI for antisense and sense probes, respectively. The probe for
all splice variants of the leptin receptor (Ob-R) was generated from
EcoRI fragment of rat Ob-R cDNA (J. Elmquist, Harvard Medical School,
Boston, MA) subcloned into a pGEM-IIZF vector (Stratagene) and linearized with
SacI and HindIII for antisense and sense probes,
respectively. The specificity of the probes was confirmed by the absence of a
positive signal in sections hybridized with sense probe. Radioactive
riboprobes were synthesized by incubation of 250 ng linearized plasmid in 10
mM NaCl, ATP/GTP/CTP (0.55-0.64 mM), 
-35S-UTP (1,000
Ci/mmol; New England Nuclear, Beverly, MA), 40 U RNasin (Promega, Madison, WI)
and 20 U SP6, T7, or T3 RNA polymerase for CRH antisense probes, for 60 min at
37°C. The DNA templates were treated with 100 µl of DNase solution (1
µl DNase, 5 µl of 5 mg/ml tRNA, 94 µl of 10 mM Tris·10 mM
MgCl2). The preparation of the riboprobe was completed through a
phenol-chloroform extraction and an ammonium acetate precipitation.
Quantitative analysis of the hybridization signals. The resulting autoradiograms were read using computer-assisted densitometry (Drexel) by an observer blinded to the experimental groups (38). The optical density for each specific region was corrected for the average background signal, which was determined by sampling unlabeled areas outside of the areas of interest. Sections (n = 2/rat) for hypothalamic PVN and central nucleus of the amygdala (CeA) CRH mRNA expression and hippocampal GR mRNA expression were read at the rostrocaudal level of the midportion of the PVN. Sections for hypothalamic arcuate (ARC) and dorsomedial (DMN) nuclei Ob-R expression (n = 2/rat) were read through the rostrocaudal extent of the compact portion of the DMN.
Statistics. Body weights, food intake, feed efficiency, plasma
leptin, and trunk blood corticosterone levels and mRNA expression data were
analyzed using three-way ANOVA (phenotype x stress/no stress x
diet). Corticosterone levels at the beginning and end of stress and from
Stress1 to Stress2 were compared by two-way ANOVA (genotype x diet) with
repeated measures. Open-field data for stressed rats were compared by two-way
ANOVA (genotype x diet). If significant differences for a given
parameter were found by ANOVA (P 
0.05), intergroup comparisons
were made by post hoc Bonferroni test. Because of the complexity of potential
interactions, if marginal significant differences among factors were found by
three-way ANOVA, significant difference for each factor alone was assessed
separately by one-way ANOVA with post hoc Bonferroni comparisons.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
0.001;
Fig. 1]. There were no
differences among the groups in initial body weight (DR = 473 ± 6 g;
DIO = 483 ± 6 g). Overall, DIO rats gained 33% more weight than DR rats
[F(1,49) = 7.69, P 0.008], and rats fed HE diet gained
133% more weight than chow-fed rats [F(1,49) = 51.71, P
0.001]. When stressed DIO rats were fed chow, they gained 66% less body weight
over 9 days than chow-fed, nonstressed DIO rats (P 0.044). When
stressed DIO rats were fed HE diet, they gained 77% more body weight than
unstressed DIO rats on HE diet over 9 days (P 0.001). In
contrast, stress had no effect on body weight gain in DR rats. None of these
differences in weight gain were due to altered caloric intake over the 9 days
poststress, since there were no significant main effects of phenotype, diet,
or stress on food intake (Fig.
2). Because there were no differences in food intake but body
weight gain was altered, there were significant main effects of phenotype,
diet, and stress on feed efficiency [g of body weight gain/kcal of food
intake; F(1,49) = 10.98, P 0.002;
Fig. 3]. Feed efficiency was
reduced by 67% in stressed chow-fed DIO rats compared with unstressed chow-fed
DIO rats (P 0.008). On the other hand, feed efficiency was
increased by 64% in stressed compared with unstressed DIO rats fed HE diet
(P 0.036).
|
|
|
Plasma leptin levels serve as an index for carcass fat content (15, 24). As was seen for body weight, among all DIO groups, leptin levels were significantly higher in stressed DIO rats fed HE diet [F(1,55) = 4.94, P = 0.001; Fig. 4]. On the other hand, while stressed DIO rats fed chow gained less body weight, their leptin levels did not differ significantly from other groups.
|
Leptin receptor mRNA was assessed because of its important role in energy
homeostasis and because out-bred DIO-prone rats have a reduced sensitivity to
the anorectic effects of leptin
(39). There were no
significant overall phenotype-diet-stress main effects among the groups, but
there was a significant effect of stress alone. Indeed, stressed rats had an
overall reduction in Ob-R mRNA expression in the DMN
(Fig. 5) compared with
unstressed groups [F(1,42) = 4.137, P 0.048]. There
were no significant phenotype-diet-stress main effects for Ob-R in the ARC,
nor was there any significant correlation between plasma leptin levels and
Ob-R expression in the DMN or ARC.
|
Open-field activity. Only rats previously subjected to
immobilization stress were tested in the open-field apparatus. This test was
performed to compare the anxiety levels of the stressed rats
(22). There was a main effect
of phenotype on open-field activity. Stressed DIO rats, regardless of their
diet, had more horizontal activity during 7 min in an open field
[F(1,28) = 4.23, P 0.049;
Fig. 6]. Overall, stressed DIO
rats had more vertical activity than stressed DR rats [F(1,28) =
11.53; P 0.002]. However, this was due to a main effect of diet
[F(1,28) = 8.91; P 0.001] because DIO rats on HE diet
had more vertical activity than the other groups
(Fig. 6).
|
Plasma corticosterone and brain CRH and GR mRNA expression. Basal
concentrations of plasma corticosterone did not differ among the experimental
groups (Table 1). As expected,
corticosterone levels rose significantly after 20 min of restraint compared
with basal levels [F(2,40) = 24.03, P 0.001], but there
were no phenotype-specific differences in this corticosterone response.
Although no basal corticosterone levels were drawn on the 9th day after the
first stress (Stress1), there were no intergroup differences in corticosterone
levels among the four stressed groups after Stress2
(Table 1). Because no basal
levels were drawn before Stress2, it was not possible to assess what effect
prior Stress1 might have had on the relative increase in corticosterone levels
after the second Stress2.
|
In the brain, there were no differences among the groups for PVN CRH mRNA
expression, regardless of phenotype, stress, or diet
(Fig. 7). Whereas PVN CRH
expression tends to be highly correlated with plasma corticosterone levels,
CeA CRH expression appears to be independent of plasma corticosterone levels
(48). In keeping with this,
CeA CRH mRNA expression showed a significant main effect of both diet and
stress [F(1,38) = 18.52, P 0.001]. In DIO rats on HE
diet, CeA CRH mRNA expression was 244% higher in unstressed than stressed DIO
rats (P 0.006) and 209% higher than unstressed DIO rats on chow
(P 0.025; Fig.
7). GR mRNA expression in the CA2 and CA3 areas of the hippocampus
was very low or not detectable, and there were no significant intergroup
differences. However, there were significant main effects of phenotype and
diet [F(1,46) = 5.32, P 0.026] on GR mRNA expression in
the CA1 area and dentate gyrus of the hippocampus
(Fig. 8). In the CA1 area, GR
expression was 31% higher in DIO than DR rats and was 32% higher in rats fed
HE diet than chow [F(1,46) = 4.11, P 0.048]. There was
also a tendency to 20% lower CA1 GR expression in stressed vs. control rats
[F(1,46) = 3.84, P 0.056]. There were, however, no
significant differences by post hoc testing between all groups. Dentate gyrus
GR mRNA also tended to be 125% higher in DIO than DR rats [F(1,46) =
3.59, P 0.064] and in rats fed HE diet than those fed chow
[F(1,46) = 3.43, P 0.071], while stress had no effect. Again,
there were no specific intergroup differences for dentate GR mRNA expression
by post hoc testing. Finally, there were no significant intergroup differences
in GR mRNA expression in the PVN by ANOVA (data not shown).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
In addition, DIO rats, as a phenotype, had elevated expression of GR mRNA in the hippocampal CA1 area and a stress- and diet-dependent reduction in CeA CRH mRNA expression. Finally, stressed DIO rats fed HE diet also showed increased open-field activity, suggesting that they were less anxious than other groups of stressed rats (22). Unlike the DIO rats, neither stress nor diet affected DR weight gain, feed efficiency, apparent adiposity, open-field activity, CRH, or GR mRNA expression. On the other hand, a single bout of restraint stress was associated with reduced DMN Ob-R mRNA and a tendency toward lower CA1 GR mRNA expression in both DIO and DR rats, regardless of dietary exposure. These data suggest that a single stress may have a long-lasting effect on neural systems involved in stress and energy homeostasis in both DIO and DR rats but that energy homeostasis is altered by the stress-diet interaction only in DIO rats.
This specific effect of a single stressor predominantly on DIO body weight gain is in distinct contrast to our previous study in selectively bred DIO rats (42). Those DIO rats were first made obese and then subjected to chronic unpredictable stress while they were fed sequentially on HE diet and then a highly palatable liquid diet. As opposed to selectively bred DR rats that lost weight and had chronic HPA activation, obese DIO rats had little response to chronic stress (42). The difference in outcome of these two studies demonstrates the complex interactions between the control of energy homeostasis and stress (42). There are several reasons why obese DIO rats on an HE diet might be hyporesponsive to the physiological and metabolic effects of chronic unpredictable stress, while obesity-prone DIO rats appear to be more readily sensitized to the diet-dependent effects of a single stressor on energy homeostasis. First, chronic stress is likely to alter the diurnal rhythm of animals, and the time at which stress is administered during the day can be a major determinant of the long-term effects of stress on body weight (1, 59). The chronically stressed DIO and DR rats in our prior study were stressed twice a day during the light phase (42), while the rats in the current study were stressed once at midlight cycle. Second, dietary factors (macronutrient content, caloric density, length of exposure) and/or the presence of obesity can have major effects on stress responsivity (61, 62). Finally, animals can habituate to chronic stress, but such habituation is very dependent on the strength, duration, frequency, pattern, prior exposure and age of subject (25). The methods and parameters used to assess the physiological response to habituation are also critical (52). While sensitization leads to increased responsivity to subsequent stress, it too is dependent on many of the same factors that modulate habituation and on the particular outcome measure assessed (7, 13, 26). Thus the common thread between our prior (42) and present studies is that DIO rats differ markedly from DR rats in their overall responsivity to stress. While DIO rats might be less responsive or habituate more readily to the weight-reducing effects of chronic stress, they are clearly more responsive to the sensitizing effects of a single stress on weight gain in the context of differing dietary composition.
Both the present and previous studies (42) demonstrate the important interactions between central pathways mediating stress responsivity and those regulating energy homeostasis. Brain areas such as the ARC and DMN contain neurons that play critical roles in the regulation of energy homeostasis (5, 17, 37). They also contain leptin receptors that allow the brain to monitor and control adipose stores (4, 21). Lesions that disconnect the ARC from the PVN alter both stress responsivity and energy homeostasis (5). It is well described that adrenalectomy reduces adiposity in various rodent models of obesity (12, 63). One potential explanation is that adrenalectomy leads to corticosterone-dependent, constitutive activation of the long form (signaling) of the leptin receptor (Ob-Rb) in the hypothalamus (46). In the current study, we found that prior stress was associated with reduced expression of mRNA coding for all splice variants of the leptin receptor in the DMN but not ARC. Thus the elevated levels of corticosterone that occurred during the first stress might be implicated in a long-term depression of DMN Ob-R expression. Importantly, the DMN, like the ARC, has also been implicated in both energy homeostasis (6, 8, 27) and stress responsivity (18). These findings must be interpreted with caution, however, as we did not directly assess the expression of the signaling form (Ob-Rb) of the leptin receptor and because mRNA expression does not necessarily reflect receptor function. In addition, rats were stressed a second time with both restraint and open-field testing before terminal assessment of Ob-R expression.
The interaction of stress with genotype and diet was also apparent in the central pathways involved in stress responsivity. DIO rats had a phenotype-dependent increase in hippocampal CA1 and a tendency for elevated dentate GR expression, which were independent of stress and diet. This was selective for the hippocampus as there were no intergroup differences in PVN GR expression. Thus increased hippocampal GR expression appears to be a phenotypic trait of obesity-prone DIO rats. Given the complexity of the feedback systems in which hippocampal GR systems are involved and the lack of difference in DIO and DR corticosterone responses to acute stress, it is difficult to predict how this underlying GR increase might alter DIO energy homeostasis in response to acute or chronic stress.
As with hippocampal GR expression, CRH mRNA expression showed site-specific differences among groups. There was a selective reduction in CeA CRH expression in stressed vs. nonstressed DIO rats on HE but not chow diet. As with PVN GR expression, PVN CRH expression was unaffected by genotype, stress, or diet. This differential regulation of CRH in the CeA vs. the PVN is well described. The CeA is associated with the psychological component of stress (46), and the behavioral changes resulting from the stimulation of the CeA appear to be independent of stress-induced HPA axis activation (11, 48). Psychological stress increases CeA but not PVN CRH mRNA (46), and CeA CRH mRNA expression can be increased by corticosterone, while PVN CRH is decreased (47). Behaviorally, stress is generally associated with reduced exploratory activity as a function of anxiety (23). Interestingly, stressed DIO rats on HE diet actually had lower CeA CRH expression than stressed DIO rats on chow. Because psychological stress would be expected to increase CeA CRH, this reduction might be due to the salutary effects of a relatively high-fat diet on the physiological and behavioral responses to stress (14), interacting with the genotype-specific neural circuitry of the DIO rats. However, it is uncertain whether the diet- and stress-dependent differences in CeA CRH expression played any role in the overall increase in open-field activity in DIO rats. Whatever the relationship, their increased open-field activity suggests reduced anxiety that may lead to a reduced stress responsiveness such as we previously documented in selectively bred obese DIO rats (42).
In summary, there is a complex interaction between the effects of stress and diet on energy homeostasis in DIO vs. DR rats. On the one hand, DIO rats appear to be less responsive or to quickly habituate to the effects of chronic unpredictable stress on energy homeostasis (42). On the other hand, DIO rats have a diet-dependent sensitization to the effects of a single stressor on energy homeostasis. It is unlikely that corticosterone was a critical determinant of their differences in energy homeostasis after a single stress. However, their differences in hippocampal GR expression suggest that DIO rats have a preexisting alteration in central pathways mediating stress responses. This finding may have important implications for the hyporesponsiveness of DIO rats to chronic stress (42). Of course it is important to keep in mind that mRNA expression does not necessarily predict the amount of neuropeptide released at a synapse or the function of a given receptor. Clearly, our studies leave a number of questions unanswered as to how these various findings fit together to explain why stress and diet have such critical effects on energy homeostasis in DIO and DR rats. Nevertheless, they do demonstrate that the DIO model is a fruitful one for the further exploration of the ways in which stress alters the regulation of energy homeostasis in obesity.
|
DISCLOSURES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  W. A. Cupples Physiological regulation of food intake Am J Physiol Regulatory Integrative Comp Physiol, June 1, 2005; 288(6): R1438 - R1443. [Full Text] [PDF]
|
|
|
|
|
|
B. E. Levin The drive to regain is mainly in the brain Am J Physiol Regulatory Integrative Comp Physiol, December 1, 2004; 287(6): R1297 - R1300. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
