| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Psychology (A.J.), University of Michigan, Ann Arbor, Michigan 48109; and Department of Anesthesiology (S.B.), Children’s Hospital of Philadelphia and the University of Pennsylvania, Philadelphia, Pennsylania 19104
Address all correspondence and requests for reprints to: Seema Bhatnagar, Department of Anesthesiology, 3615 Civic Center Boulevard, Abramson Research Center, Suite 402, Philadelphia, Pennsylvania 19104-4399. E-mail: bhatnagars{at}email.chop.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We recently showed that lesions of the posterior paraventricular thalamus (pPVTh) attenuate dexamethasone-induced negative feedback in repeatedly but not acutely restrained rats (2), indicating that the pPVTh contributes to feedback inhibition in repeatedly stressed rats. The pPVTh is part of the limbic thalamus, intimately connected to the amygdala and prefrontal cortex, and potentially important in transducing emotionally salient information (13). In general, the pPVTh inhibits habituated HPA activity (14) and other aspects of HPA activity and behavior in repeatedly but not acutely stressed rats (15, 16, 17). We hypothesized that the pPVTh could exert its inhibitory effects partly by serving as a site for negative feedback effects of corticosterone that are important for habituation of HPA activity. In the present studies, we first showed that GR are located in the pPVTh and MR have already been localized in the pPVTh (18). We then conducted a series of experiments to determine the specific corticosteroid receptor type(s) involved through intra-pPVTh administration of MR and/or GR agonists and antagonists over different time frames during 8 d of restraint. Overall, our results indicate that activation of MR and GR in the pPVTh throughout the entire week of stress is critical for habituation. We further show (in experiment 5) that the mPFC and basolateral amygdala (BLA) are candidate structures through which corticosterone actions in the pPVTh could regulate habituation. Collectively, these experiments provide evidence for a previously unrecognized site of corticosterone-mediated negative feedback and advance our understanding of the central mechanisms by which environmental events, such as repeated stress, can alter subsequent processing of stress-related information in the brain.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Repeated-stress paradigm and blood sampling procedure
In the present experiments, we used a repeated-restraint paradigm to examine habituation of HPA activity. Restraint consisted of placing rats in a ventilated, cylindrical Plexiglas tube for 30 min. The two open ends were closed with two pieces of masking tape. Rats were able to move laterally but were not able to turn around in the restrainer. Rats in acute stress groups were exposed to a single 30-min restraint. Rats in repeated stress groups were exposed to eight consecutive days of 30 min restraint per day. In experiments 2–4, we also collected blood samples in response to the first exposure to restraint (acute-stress groups) or to the eighth exposure to restraint (repeated-stress groups). The designs of experiments 2–4, illustrated in Fig. 1
, compare responses on d 1 vs. d 8 of restraint (i.e. between-groups designs). Although we did not collect blood between d 1 and 8 during the week of repeated restraint in any of the present experiments, we know from previous data from our lab that animals in this repeated-restraint paradigm show habituation of corticosterone responses to restraint between d 3 and 5 (unpublished observations). Each animal was placed in the restrainer, and blood samples were taken from the tail vein at 0, 15, and 30 min during restraint. After collection of the 30-min samples, animals were removed from their restrainers and replaced in their home cages. At 60 min, rats were replaced in the restrainer, and blood samples were collected again and animals subsequently returned to their home cages. The 0- and 60-min samples were collected within 60 sec of opening the cage to ensure that ACTH and corticosterone levels in plasma do not rise in response to the restraint itself (19). This sampling method of tail nicking, used by other groups as well (7, 20, 21), produces little reactivity in the animal and results in consistent basal and stress levels of ACTH and corticosterone (2, 15, 16, 19).
|
Experiment 2: acute GR and/or MR blockade in the pPVTh before the first or eighth restraint
The purpose of this experiment was to examine whether blockade of GR and/or MR in the pPVTh before the first or eighth restraint alters HPA responses to acute or repeated restraint. We first determined a dose of commonly used GR and MR antagonists to inject. A previous study reported alterations in HPA activity after intra-hippocampal injections of 5 ng of the GR antagonist RU-38486 or the MR antagonist spironolactone (20). To confirm that local injections of this 5-ng dose of GR or MR antagonist are effective in altering HPA activity under our conditions, we conducted a pilot study. We injected 5 ng GR antagonist (RU-38486) or MR antagonist (spironolactone) into the dorsal hippocampus 60 min before an acute restraint and sampled as described above (n = 4–5 per time point for vehicle group; n = 4–5 per time point for GR antagonist group; and n = 3–5 per time point for MR antagonist group; data not shown). Variations in n values per time point are a result, as in all other experiments reported here, of a lack of a sufficient plasma sample available for assay or a loss of sample in the assay. We found that 5 ng RU-38486 was effective in decreasing acute-restraint-induced ACTH responses at 15 min (491 ± 61 pg/ml) compared with vehicle (1019 ± 249 pg/ml) and at 30 min (531 ± 113 pg/ml) compared with vehicle (1250 ± 332 pg/ml). RU-38486 also decreased restraint-induced corticosterone responses at 15 min (24 ± 2 µg/dl) compared with vehicle (37 ± 4 µg/dl). We found that 5 ng spironolactone was effective in increasing basal corticosterone responses at 0 min (15 ± 2 µg/dl) compared with vehicle (2 ± 0.8 µg/dl). These results were consistent with the notion that hippocampal MR are important in regulating basal HPA activity and that hippocampal GR primarily regulate stress-induced HPA activity (10, 20). Therefore, we used 5 ng each of RU-38486 and spironolactone into the pPVTh in this experiment.
After 5 d of recovery from stereotaxic implantation of cannulas in the pPVTh (as described below), half of all rats were randomly assigned to either 1 or 8 d of 30 min restraint/d. On d 1 or 8, rats within each group received intra-pPVTh injections of 5 ng GR antagonist alone (RU-38486), MR antagonist alone (spironolactone), combined GR + MR antagonists, or vehicle 60 min before the onset of restraint (per time point, n = 6 for repeated vehicle group, n = 5–6 for repeated GR antagonist group, n = 6 for repeated MR antagonist group, n = 6–7 for repeated GR + MR antagonist group, n = 5–6 for acute vehicle group, n = 5 for acute GR antagonist group, n = 5 for acute MR antagonist group, and n = 5–6 for acute GR + MR antagonist group). A total volume of 0.25 µl of drug(s) or vehicle was injected into the pPVTh 60 min before restraint exposure and blood collection for assessment of HPA activity. The same volume was used for injections of combined antagonists as well as for single antagonist. We have previously used a 0.25-µl injection volume in the pPVTh for ibotenic acid lesions and found that this volume of injection localizes drug spread primarily to the pPVTh (2). The selection of the 60-min time point for drug administration was based on the results of a previous study as well as our preliminary study showing that RU-38486 and spironolactone injected 60 min before restraint were effective in altering stress and basal HPA activity, respectively (20).
Experiment 3: corticosterone implants in the pPVTh
In experiment 2, acute GR and/or MR blockade in the pPVTh before the eighth restraint did not alter habituation to repeated restraint. In experiment 3, we used steroid implants that would affect GR/MR functioning in the pPVTh throughout the 8 d of stress, and not solely on d 8. Specifically, we examined the effects of implants of corticosterone or of the steroid control, cholesterol, in the pPVTh on HPA activity in acutely and repeatedly restrained rats. Before using our corticosterone implants in the pPVTh, we first confirmed the effectiveness of our implants in inhibiting HPA activity by examining their effects on ACTH responses to acute restraint after implantation in the hypothalamic PVN (n = 4–6 for all groups per time point; data not shown). All rats were adrenalectomized. These n values as well as those presented for experiment 3 represent the final n values after omission of rats with missed placements of implants, spread of corticosterone outside the area of the PVN (or pPVTh in experiment 3) as assessed by GR immunocytochemical staining, or incomplete adrenalectomies as indicated by a rise in corticosterone in response to restraint. The PVN is a documented site of negative feedback action as demonstrated by studies using local implants or local microinjections of glucocorticoids (11, 29, 30). If our corticosterone implants were effective, we expected to observe a decrease in HPA activity when placed in the PVN. The procedure for this experiment was identical to that described below for experiment 3 except for the region of steroid implantation. In all groups, we looked at integrated ACTH, which measures the net secretion of ACTH across 0, 15, 30, and 60 min. We found that integrated ACTH was decreased with corticosterone implants in the PVN (1094 ± 207 pg/ml) compared with cholesterol in the PVN (2022 ± 248 pg/ml). Corticosterone implants in the PVN produced only a partial blockade of HPA activity, which is likely because of the lack of inhibition by endogenous corticosterone in other important sites of negative feedback. These findings of partial blockade of the HPA response are consistent with the results of other studies that implanted corticosterone in negative feedback sites in adrenalectomized rats (12, 31). Because these intra-PVN corticosterone implants were effective in inhibiting ACTH responses to restraint, we proceeded with steroid implantation in the pPVTh.
In this experiment, half of all rats were randomly assigned to either corticosterone or cholesterol implant groups. Stereotaxic intra-pPVTh steroid implantation and adrenalectomy were performed as described below. We adrenalectomized rats to prevent endogenous stress-induced corticosterone from acting on the brain, as previously performed when implanting corticosterone into the brain (11, 12, 29, 31). It could potentially be difficult to observe substantial differences between corticosterone-implanted rats and cholesterol-implanted rats when the adrenals are intact. This is because endogenous corticosterone would still be acting at the pPVTh even in cholesterol-implanted rats, resulting in inhibited HPA responses to repeated stress. In corticosterone-implanted rats, the inhibition of HPA responses to repeated stress resulting from the addition of corticosterone in the pPVTh might be difficult to observe in addition to the already inhibited HPA activity caused by endogenous corticosterone in the pPVTh; that is, a floor effect might have been reached. Therefore, we think that negative-feedback functions of the pPVTh in repeatedly stressed rats are more clearly elucidated when the animals are adrenalectomized. This ensures that the observed effects on HPA activity are a result of corticosterone implants in the pPVTh and not of endogenous corticosterone acting in other potential sites of negative feedback All rats were replaced with corticosterone pellets to provide rats with corticosterone levels that are within the average basal range (see below for additional explanation). After 1 wk of recovery from surgery, rats underwent either 1 or 8 d of 30 min of restraint (n = 6–10 per time point for all corticosterone groups, and n = 6–9 for all cholesterol groups). Blood samples were collected on d 1 or 8 of restraint, as above. At the end of the experiment, brains were perfused and processed for GR immunocytochemistry to confirm spread of corticosterone from the implant, as described below.
Experiment 4: daily GR and MR blockade in pPVTh in repeatedly stressed rats
In experiment 3, our steroid implants provided the pPVTh with continuous exposure to corticosterone from d 1–8 of restraint. That is, repeatedly stressed rats were exposed to corticosterone throughout repeated restraint as well as on test day, d 8. Acutely stressed rats were exposed to 7 d of corticosterone in the pPVTh as well as to corticosterone during their first and only exposure to restraint. Experiment 4 aimed to control for this continuous exposure by manipulating GR and MR in the pPVTh only before restraint on d 1–7 of restraint in repeatedly stressed rats or on d 1–7 of no stress in rats that were exposed to acute restraint only on d 8. Drugs were not injected into the pPVTh on test day (first acute restraint or the eighth restraint in control and repeatedly stressed rats, respectively) in any group of rats. After 5 d of recovery from stereotaxic implantation of the cannulas in the pPVTh (as described below), half of all rats were randomly assigned to either 1 or 8 d of 30 min restraint. Because, in experiment 2, we observed significant effects on HPA activity with combined receptor blockade and not with blockade of either receptor type alone, we used only the combined antagonist treatment in this experiment. From d 1–7 of restraint, repeatedly stressed groups received intra-pPVTh injections of 5 ng of combined GR antagonist (RU-38486) and MR antagonist (spironolactone) or vehicle at 60 min before each restraint (n = 7–11 per time point for GR + MR antagonist group, and n = 6–8 per time point for vehicle group). From d 1–7, rats that went on to be acutely stressed also received intra-pPVTh injections of 5 ng of combined GR antagonist plus MR antagonist or vehicle but were not restrained from d 1–7 (n = 10–13 per time point for GR + MR antagonist group, and n = 6–8 per time point for vehicle group). The total volume of drugs or vehicle injected into the pPVTh was 0.25 µl. For all animals in this experiment, there were no injections of drug or vehicle into the pPVTh on d 1 or 8 of restraint. Blood was collected on d 1 or 8 of restraint for assessment of HPA activity, as described above.
Experiment 5: detection of GR-containing efferent projections of the pPVTh
The pPVTh does not directly project to the PVN (24, 32) but does project to the bed nucleus of the stria terminalis, central amygdala, BLA, and mPFC (24, 33). Therefore, pPVTh regulation of the PVN, and therefore HPA activity, requires involvement of intermediary structures. As an initial assessment of which efferent sites of the pPVTh might be important in mediating its negative feedback inhibition of HPA activity, we focused on the BLA and mPFC, two limbic structures important for regulation of HPA responses to repeated stress (12, 34, 35). We injected the retrograde tracer fluorogold into the mPFC or BLA and subsequently examined dual labeling of GR and fluorogold in the pPVTh. Fluorogold (Fluorochrome Inc., Englewood, CO) was diluted to 3% in 0.9% saline (16). Rats were anesthetized as described for stereotaxic surgery below. Fluorogold was injected unilaterally into the mPFC or BLA by micro-iontophoresis. A fluorogold-filled glass micropipette of tip diameter between 30 and 100 µm was lowered to the level of the mPFC (n = 10) at the following stereotaxic coordinates from bregma: anteroposterior (AP) + 3.0 mm, mediolateral (ML) ± 0.5 mm, and dorsoventral (DV) –3.5 mm. These coordinates were aimed toward injection into the prelimbic (dorsal) subregion of the mPFC. Although the pPVTh projects to both prelimbic and infralimbic mPFC (24), the prelimbic subregion projects more heavily than the infralimbic subregion to other limbic sites that reportedly affect cognition (36, 37). Because the HPA axis likely requires input from cortical regions involved in cognitive processing during habituation to a familiar stressor, the prelimbic mPFC is an appropriate subregion to examine in the present context. For the BLA (n = 6), the following coordinates from bregma were used: AP –3.1, ML +5.0, and DV –7.3. Fluorogold was injected at a 5 µA current for 5 min into each brain region (16). The micropipette was allowed to stay in place for another 10 min before removal. All rats were left undisturbed for 2 wk and then perfused on d 14. Brains were collected and processed for dual immunocytochemical labeling of fluorogold and GR as described below.
Immunocytochemistry for GR
At the end of experiments 1 and 3, all rats were perfused with saline followed by 4% formalin, and brains were collected. Perfused brains were sliced coronally at 30 µm on a sliding microtome. A one-in-six series of sections were stained immunocytochemically for GR using an immunocytochemistry protocol that was modified from one that has previously been used (12). Free-floating sections to be reacted with GR antibody were incubated with 10% normal goat serum in 0.1 M PBS solution for 20 min at room temperature. Sections were then incubated overnight at 4 C with primary antibody raised against the GR (PA1-510; Affinity Bioreagents, Golden, CO) diluted 1:1000 in 0.1 M PBS cocktail containing 1% normal goat serum (Vector Laboratories, Burlingame, CA), 0.3% Triton X-100 (Sigma Chemical Co., St. Louis, MO), and 0.25% BSA (Sigma). Sections were subsequently incubated with a biotinylated goat antirabbit IgG (Vector) diluted 1:200 in the above PBS cocktail for 2 h at room temperature, washed, and then incubated with an avidin-biotin-peroxidase complex (Elite kit; Vector) for 2 h at room temperature. To visualize the immunoreactivity, a nickel diaminobenzidine reaction catalyzed by glucose oxidase was used. Brain sections were then mounted onto Superfrost slides using tap water. Sections were coverslipped after immersion in water, ethanol (70, 90, 95, and 100%) and CitriSolv (Fisher Scientific, Boston, MA). All slides were visualized using a Leica microscope coupled to a Spot RT digital camera through NIH Image (W. Rasband, National Institutes of Health, Bethesda, MD) software. Brains were examined by an experimenter blind to the condition of the sections being analyzed.
Dual immunocytochemistry for fluorogold and GR
In experiment 5, rats were anesthetized before perfusion with a mixture of ketamine, xylazine, and acepromazine (77:1.5:1.5 mg/ml given im at 0.1 ml/100 g body weight), and brains were collected and sliced according to the procedure described above for experiments 1 and 3. For dual labeling of fluorogold and GR in the same tissue, we performed immunocytochemistry for fluorogold first (16, 38) and immunocytochemistry for GR second (based on protocol described above), as follows. Free-floating sections were rinsed in Tris/PBS (TPBS) buffer and then incubated for 48 h at 4 C with primary fluorogold antibody (Chemicon International Inc., Temecula, CA) diluted 1:8000 in TPBS cocktail containing 0.3% Triton X-100 (Sigma) and 0.25% BSA (Sigma). Sections were subsequently incubated with a biotinylated goat antirabbit IgG (Vector) diluted 1:200 in the above TPBS cocktail for 1 h at room temperature, washed, and then incubated with an avidin-biotin-peroxidase complex (Elite kit; Vector) for 1 h at room temperature. To visualize fluorogold immunoreactivity, we used an SG substrate kit (Vector) that yielded a dark gray product. Immediately after staining with SG substrate, immunocytochemistry for GR was performed on the same tissue. GR immunocytochemistry in our dual-labeling procedure is identical to the single-labeling procedure described above with the exception of the substrate used to visualize immunoreactivity. Instead of using nickel diaminobenzidine, we used a NovaRED substrate (Vector). Dual labeling of red-stained GR and dark gray-stained fluorogold in the same cells yielded a dark purple color. Sections were subsequently mounted and coverslipped.
Intracerebral cannula implantation
In experiments 2 and 4, rats were anesthetized with a mixture of ketamine, xylazine, and acepromazine (77:1.5:1.5 mg/ml given im at 0.1 ml/100 g body weight) and placed in a stereotaxic apparatus with the skull flat and the tooth bar at –3.3 mm. Unilateral guide cannulas (22 gauge) were lowered into the pPVTh according to the following coordinates from bregma: AP –2.8 mm, ML 0.0 mm, and DV –6.2 mm. Cannulas were then cemented into place using dental cement anchored by skull screws. Dummy cannulas were used to close the guide cannulas. At the time of experimentation, dummy cannulas were removed and replaced with an injector cannulas (28 gauge) connected to PE tubing attached to a Hamilton syringe. Antagonist(s) or vehicle was injected over a 1-min period, with the needle remaining in place for another 1 min before removal. We used unilateral injections for the pPVTh because the pPVTh is a midline structure. With bilateral injections, there would likely be spread of drug lateral to the pPVTh into other thalamic regions. After completion of the experiment, brains were collected and sectioned on a cryostat at 30 µm. The exact placement of the cannula was confirmed by staining sections with cresyl violet and visualizing the tips of the injector cannulas. Any animals in which the tip of the injector cannula was found to be outside of the posterior division of the PVTh were excluded from the study. A representative cannula placement in the pPVTh and a representative placement that was classified as a missed placement and excluded from analyses are shown in Fig. 2. As previously defined, the posterior division of the PVTh extends from –2.56 to –3.3 mm from bregma (24, 39). Any animals showing evidence of cannula tracks primarily in the anterior and medial subdivisions were excluded. Animals that were classified as having correct placements in the pPVTh showed histological evidence of clear damage of the ependyma of the ventricle overlying the pPVTh. This ensured that any animals with placements in the ventricle or lateral to the pPVTh were not included. Also, in any animals included in the analyses, minimal damage was observed in the hippocampus.
|
Previous studies that implanted steroids into the brain have not subsequently measured local CNS concentrations of the steroid (12, 29, 31, 44) with the exception of one study (40). Although we did not measure corticosterone concentrations in the pPVTh after corticosterone implantation, we did confirm the spread of corticosterone from the implant and verified the placement of the implant by staining sections immunocytochemically for GR, as performed by one other group that implanted corticosterone into the brain (12). Before using this technique, we observed GR immunoreactivity in sections of unstressed, intact rats and adrenalectomized rats. We observed that GR staining is nearly absent in rats that are adrenalectomized without corticosterone replacement compared with intact rats with endogenous corticosterone (Fig. 3), consistent with other reports that the GR antibody used primarily detects the activated receptor (12). We therefore used observations of increased localization of staining in cells of the pPVTh to confirm delivery and spread of steroid. Results from GR immunocytochemistry demonstrated staining of GR that was discrete, darker, and more concentrated in the area underneath the tip of the corticosterone implant compared with the area in cholesterol-implanted rats (Fig. 3). All rats were adrenalectomized and replaced with corticosterone pellets. In groups that received corticosterone implants, we included data only from those animals that showed spread of steroid in at least approximately two thirds of the posterior division of the PVTh. Any animals showing spread of steroid primarily in the anterior and medial subdivisions were excluded, although animals exhibiting staining in the medial subdivision were included if at least two thirds of the posterior subdivision also showed staining. We have previously used these criteria to determine size of excitotoxic lesions in the pPVTh (2, 14, 39).
|
Drugs
In experiments 2 and 4, GR antagonist (RU-38486 or mifepristone) and MR antagonist (spironolactone) (Sigma) were initially dissolved in 1% ethanol and brought up to volume in 0.9% saline. Saline/ethanol served as the vehicle injection. In experiment 3, corticosterone 21-acetate and cholesterol (Sigma) were used in the intracerebral steroid implants as well as in the 35% corticosterone pellets. Corticosterone 21-acetate has previously been used to examine the effects of corticosterone on CNS functions (45).
Statistical analyses
All data in experiments 2–4 were analyzed using ANOVA. In experiment 2, stress (acute or repeated stress) x treatment (vehicle, RU-38486, spironolactone, or combined RU-38486 plus spironolactone) ANOVAs were performed at each time point of blood sampling. In experiment 3, stress (acute or repeated stress) x treatment (cholesterol or corticosterone) ANOVAs were carried out at each time point. In experiment 4, stress (acute or repeated stress) x treatment (vehicle or combined RU-38486 plus spironolactone) ANOVAs were performed at each time point. All significant main or interaction effects were followed by Fisher’s post hoc tests. The significance levels in all tests were set at P 
0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, GR immunoreactivity is shown in representative sections of anterior, medial, and posterior PVTh. Considerable amounts of GR were present in all divisions of the PVTh, implicating the region as a likely target of corticosterone action. Slightly different intensities of GR immunoreactivity through the anterior-posterior extent of the PVTh were observed. Specifically, although similar intensities were observed within medial and posterior divisions of the PVTh, slightly stronger intensities were observed in anterior compared with medial and posterior divisions. To compare GR staining in adrenal-intact vs. adrenalectomized rats, we refer the reader to Fig. 3
, which shows GR immunoreactivity in the posterior PVTh in an adrenalectomized rat from experiment 3. Eight days of restraint did not substantially alter distribution or intensity of GR immunoreactivity in any division of the PVTh compared with 1 d of restraint. Therefore, the pattern of GR distribution and intensity of staining throughout anterior, medial, and posterior divisions of the PVTh were similar in repeatedly and acutely restrained groups. It is possible that perfusing brains before stress might have exposed differences in GR immunoreactivity between acute and repeated restraint groups. However, alterations in GR immunoreactivity have previously been shown in rats that are perfused 60 min after the last exposure to stress (26). Therefore, demonstrations in the literature along with evidence that the GR antibody used primarily detects activated GR led us to perform GR immunocytochemistry post stress (12, 26).
|
. At 0 and 15 min, no significant effects were observed for GR antagonist alone, MR antagonist alone, or combined GR + MR antagonist. At 30 min, there was a significant stress effect [F(1,37) = 40.18; P 0.001] in which repeatedly stressed groups had lower ACTH than acutely stressed groups and a significant treatment effect [F(3,37) =3.854; P 0.05]where groups that received combined RU-38486 plus spironolactone had higher ACTH than vehicle groups. At 30 min, there was also a significant stress x treatment interaction [F(3,37) = 5.264; P 0.01]. Post hoc tests showed that, among vehicle-treated groups, repeatedly restrained rats displayed significantly lower ACTH than acutely restrained rats at 30 min, indicating habituation to repeated restraint. Furthermore, in acutely restrained groups, combined GR + MR antagonist treatment significantly increased ACTH compared with vehicle (Fig. 5). In contrast, GR or MR antagonist alone did not significantly alter ACTH in any group. At 60 min, there were no significant effects on ACTH in repeatedly or acutely restrained rats. We also examined HPA activity in rats with missed injections or evidence of injection tracks outside of the pPVTh. Because the number of animals in these groups was small (n = 2–5; data not shown), we did not conduct statistical analyses of the data. We observed that the effects of combined GR and MR antagonists into sites lateral to the pPVTh (mediodorsal thalamic nucleus) in acutely restrained rats were not different from injections of vehicle into the pPVTh. These controls support the idea that the effects observed are specific to the actions of the antagonists in the pPVTh and not in other sites.
|
0.01] demonstrating that overall, repeatedly restrained groups had lower net ACTH secretion across all time points compared with acutely restrained groups (Fig. 5). There was also a significant treatment effect [F(3,35) = 4.380; P 0.01] showing that integrated ACTH in GR + MR antagonist-treated groups was higher than GR antagonist alone or MR antagonist alone. Last, there was a significant stress x treatment interaction [F(3,35) = 2.838; P 0.05] for integrated ACTH.
Post hoc tests indicated that in vehicle-treated groups, repeatedly restrained rats had significantly lower integrated ACTH than acutely restrained rats. Post hoc tests also indicated that in acutely restrained groups, integrated ACTH was significantly higher after GR + MR antagonist treatment compared with vehicle, GR antagonist alone, and MR antagonist alone, but no significant effects were observed for repeatedly restrained groups (Fig. 5). No significant differences were found at basal, stress, or recovery time points in plasma concentrations of corticosterone (Fig. 6).
|
Experiment 3: corticosterone implants in the pPVTh in repeatedly stressed rats
The effects of corticosterone or cholesterol implants in the pPVTh on plasma ACTH responses to restraint in acutely or repeatedly restrained rats are shown in Fig. 7. At 0 min, there were no significant effects on ACTH. At 15 min, there was a significant stress effect [F(1,27) = 11.5; P 0.01] in which repeatedly restrained groups had lower ACTH than acutely restrained groups and a significant treatment effect [F(1,27) = 5.2, P 0.05] where corticosterone-treated groups had lower ACTH than cholesterol-treated groups. At 30 min, there were no significant effects on ACTH. At 60 min, we observed a significant treatment effect [F(1,20) = 4.0; P 0.05] in which corticosterone-treated groups had lower ACTH than cholesterol-treated groups. At 60 min, there was also a significant stress x treatment interaction effect [F(1,20) = 7.2; P 0.01]. Post hoc tests showed that, in repeatedly restrained rats, corticosterone treatment significantly decreased ACTH responses to restraint compared with cholesterol treatment, but no significant effects were observed in acutely restrained groups (Fig. 7A). We also analyzed integrated ACTH for which there was a significant stress effect [F(1,20) = 5.817; P 0.05] in which repeatedly restrained groups had lower integrated ACTH compared with acutely restrained groups. We also observed a significant treatment effect [F(1,20) = 7.991; P 0.01] in which corticosterone-treated groups had lower net ACTH secretion compared with cholesterol-treated groups (Fig. 7B). In general, plasma ACTH concentrations in these adrenalectomized rats were higher than intact rats because of the absence of maximal negative feedback normally exerted by endogenous corticosterone. Previous reports have shown sustained ACTH hypersecretion to restraint in adrenalectomized rats despite having sc corticosterone pellets that provide low circulating levels of corticosterone of 3–6 µg/dl (46), a range that is consistent with corticosterone levels in the present experiment. Adrenalectomized rats with corticosterone pellets are thought to hypersecrete stress-induced ACTH because of an absence of circadian variations in corticosterone (47) as well as a lack of acute increases in corticosterone during stress (48). However, adrenalectomy was necessary in all groups because, in adrenal-intact groups, it could potentially be difficult to observe substantial differences between corticosterone-implanted rats and cholesterol-implanted rats. This is because endogenous corticosterone would still be acting at the pPVTh even in cholesterol-implanted rats. In corticosterone-implanted rats, the inhibition of HPA responses to repeated stress resulting from the addition of corticosterone in the pPVTh might be difficult to observe in addition to the already inhibited HPA activity caused by endogenous corticosterone in the pPVTh; that is, a floor effect might have been reached.
|
). This confirmed complete adrenalectomy because there was no stress-induced rise in corticosterone as well as effectiveness of the 35% pellet in releasing low basal levels of corticosterone (12). To summarize, repeatedly restrained groups overall displayed significantly lower ACTH responses to restraint than acutely restrained groups at 15 min, providing evidence of habituation. Corticosterone implants in the pPVTh significantly suppressed ACTH responses to restraint compared with cholesterol treatment at 60 min in repeatedly restrained rats but did not significantly alter ACTH in acutely restrained rats at any time point.
Experiment 4: daily GR and MR blockade in pPVTh in repeatedly stressed rats
In experiment 3, corticosterone implants that provided the pPVTh with continuous exposure to corticosterone from d 1–8 decreased ACTH in repeatedly, but not acutely, restrained rats. Experiment 4 controlled for this continuous exposure to corticosterone provided by the implants by examining HPA activity after injections of GR + MR antagonist only before restraint on d 1–7 or without stress in rats that went on to be acutely stressed. Plasma ACTH responses to restraint after daily intra-pPVTh injections of combined GR + MR antagonist or vehicle are shown for acutely and repeatedly restrained rats in Fig. 8. At 0 min, no significant effects were observed on plasma ACTH. There were significant stress effects at 15 min [F(1,41) = 4.102; P 0.05] and 30 min [F(1,39) = 11.495; P 0.01] in which repeatedly restrained groups had lower ACTH than acutely restrained groups. At 60 min, there was a significant treatment effect [F(2,43) = 3.019; P 0.05] in which groups that received GR + MR antagonist had higher ACTH than vehicle groups. At 60 min, there was also a significant stress x treatment interaction effect [F(2,43) = 3.199; P 0.05]. Post hoc tests revealed that, in repeatedly stressed groups, GR + MR antagonist increased ACTH compared with vehicle treatment, but no significant effects were observed in acutely stressed groups (Fig. 8A). We also analyzed integrated ACTH for which there was a significant stress effect [F(1,30) = 4.047; P 0.05] demonstrating that repeatedly restrained groups overall had lower net ACTH secretion across all time points compared with acutely restrained groups. For integrated ACTH, there was also a significant stress x treatment interaction effect [F(1,30) = 4.763; P 0.05]. Post hoc tests indicated that in vehicle-treated groups, repeatedly restrained rats had significantly lower integrated ACTH than acutely restrained rats. Furthermore, post hoc tests indicated that in repeatedly restrained groups, integrated ACTH was significantly higher after GR + MR antagonist treatment compared with vehicle, but no significant effects were observed for acutely restrained groups (Fig. 8B).
|
0.01], 30 min (F(1,50) = 8.027; P 0.01), and 60 min [F(1,50) = 5.636; P 0.05] in which repeatedly restrained groups had lower corticosterone than acutely restrained groups (Fig. 8C). To summarize, at 15 and 30 min, repeatedly restrained groups had significantly lower ACTH and integrated ACTH and significantly lower corticosterone at 15, 30, and 60 min compared with acutely restrained groups, evidence of habituation. Furthermore, daily GR + MR antagonist treatment in the pPVTh before restraint on d 1–7 significantly increased integrated ACTH and ACTH responses at 60 min to the eighth restraint. In rats that went on to be acutely stressed (but that were not stressed from d 1–7), daily GR + MR antagonist treatment in the pPVTh from d 1–7 did not significantly alter ACTH responses to the first restraint compared with vehicle treatment.
Experiment 5: detection of GR-containing efferent projections of the pPVTh
In 10 rats with injections of fluorogold in the mPFC, fluorogold labeling at the site of injection was concentrated in the prelimbic area of the mPFC and in the dorsal portion of the infralimbic mPFC. The spread of fluorogold at the site of injection is shown in a representative animal in Fig. 9A. Any animals that displayed fluorogold diffusion outside of the mPFC were omitted from the analysis. All animals that received fluorogold injections into the mPFC showed a similar and consistent distribution and intensity of fluorogold labeling in forebrain regions. Regions where moderate labeling of fluorogold was observed include the frontal cortex, lateral septum, claustrum, basolateral and basomedial amygdala, and lateral and midline thalamus. Of particular relevance to the present study were observations of low to moderate labeling throughout anterior, medial, and posterior divisions of the PVTh (Fig. 9C). Representative GR-immunoreactive cells in the pPVTh are shown in Fig. 9D. Of 68 total GR-immunoreactive cells counted in the pPVTh per animal, we estimated that approximately 38% of these were also labeled with fluorogold injected in the mPFC. This was measured by subjective counts of the number of dual-labeled cells in the pPVTh/total number of GR in the pPVTh. The cell counts were averaged across animals. These dual-labeled cells in the pPVTh are shown in a representative animal in Fig. 9, E and F.
|
. Any animals that displayed fluorogold diffusion in amygdaloid nuclei other than the BLA were omitted from the analysis. We observed low to moderate labeling of fluorogold in lateral thalamic, hypothalamic, and amygdaloid nuclei. Importantly, we observed fluorogold labeling throughout the PVTh, with low labeling in the anterior division and moderate to heavy labeling in medial and posterior divisions. Of 74 total GR-immunoreactive cells counted in the pPVTh per animal, we estimated that approximately 59% of these cells were also labeled with fluorogold injected in the BLA. To summarize, after fluorogold injection into the mPFC or BLA, we observed fluorogold labeling in several forebrain regions including the anterior, medial, and posterior divisions of the PVTh. Importantly, cells of the pPVTh were dual labeled with fluorogold and GR after fluorogold injection into the mPFC or BLA. Thus, moderate densities of GR containing cells in the pPVTh projected to the mPFC or BLA.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our finding that combined GR + MR blockade in the pPVTh is more effective in altering acute stress-induced HPA activity than sole blockade of either receptor type (Fig. 5) supports a model that posits a cooperative relationship between GR and MR in control of cellular homeostasis and neuroendocrine function (49, 50, 51, 52). Coordinated activity of GR and MR is necessary in constraining HPA activity when circulating corticosterone levels are elevated during the circadian peak (52) and during acute stress (21). Based on our findings that combined GR + MR blockade in the pPVTh inhibited HPA responses to acute restraint, we suggest that the pPVTh is a relevant site where corticosterone can exert negative feedback effects to inhibit the acute stress response. On the other hand, repeatedly stressed rats in experiment 2 displayed no change after acute GR + MR blockade in the pPVTh (Fig. 5). Therefore, GR + MR blockade in the pPVTh on the final day of stress was not sufficient to interfere with the neural changes that lead to habituation of HPA activity. The differences between the role of the pPVTh in negative feedback under acute vs. repeated stress conditions may relate to the time course of corticosterone’s actions at the pPVTh. It is possible that by the time we administered antagonists on d 8 of restraint, the neural processes that contribute to habituation had already developed and were not affected by one injection on d 8. In fact, habituation to repeated restraint can be expressed as early as d 5 (34), implying that the processes involved in producing habituation develop over time. If this is the case, then manipulation of MR and GR throughout the week of stress, and not just on d 8, could alter HPA responses to repeated restraint. Therefore, we hypothesized that corticosterone released during the entire week of stress acts at the pPVTh to exert negative feedback inhibition of HPA activity and that this is important for habituation. We tested this hypothesis by using continuous corticosterone implants in the pPVTh throughout d 1–8 of restraint (experiment 3) as well as by daily injections of GR + MR antagonists on d 1–7 of restraint before testing on d 8 (experiment 4).
In experiment 3, repeatedly restrained animals with corticosterone implants in the pPVTh displayed lower ACTH responses to restraint, and therefore enhanced habituation, compared with all other cholesterol-implanted groups (Fig. 7). The implants likely provided the pPVTh with a greater amount of corticosterone than the pPVTh is normally exposed to given the continual period of time that the implant is in place. Evidence that the implants are still releasing corticosterone on the day of blood collection came from our GR immunocytochemistry data in steroid-implanted rats (Fig. 3). However, by manipulating GR + MR functioning in the pPVTh only before restraint on 7 d (experiment 4), we provided an effective control for this continuous corticosterone exposure. This is because our design in experiment 4 allowed us to examine the effects of daily GR + MR blockade in the pPVTh on the development of habituation without interfering with the expression of habituation on d 8. We injected GR + MR antagonists before restraint on d 1–7, a time when habituation to the repeated stressor should be developing, but did not inject GR + MR antagonists on d 8, the final day of stress on which habituation is expressed. Repeatedly stressed rats that received GR + MR antagonists on d 1–7 did not habituate compared with acutely stressed rats injected with either vehicle or antagonists, although the repeatedly stressed rats with vehicle did, as expected (Fig. 8). Therefore, blocking GR + MR in the pPVTh throughout the week of stress prevented the development of habituation, but blocking GR + MR in the pPVTh from d 1–7 before the first restraint did not affect HPA responses in acutely restrained rats. When considered together, the findings from experiments 2–4 suggest that the actions of corticosterone at GR and MR in the pPVTh are important during the development of habituation and not for the expression of an already developed habituated response on the final day of stress.
It is interesting to note that in experiment 3, despite the lack of normal stress-induced negative feedback of HPA activity by endogenous corticosterone, adrenalectomized rats with cholesterol implants still displayed the capacity to habituate to repeated stress. This suggests that habituation of the HPA axis is not entirely dependent on negative-feedback mechanisms. One hypothesis to consider is that the reduction in HPA responses to repeated restraint could be a reflection of behavioral adaptation to restraint and not negative feedback. However, corticosterone-implanted rats did show significantly enhanced habitation compared with cholesterol-implanted rats, suggesting that negative feedback is at least one factor that contributes to inhibition of HPA responses to restraint during habituation. Maximal habituation may be the result of multiple independent and/or convergent processes, some of which may include behavioral and psychological adaptation, decreases in stimulatory drive to the HPA axis from stress-responsive brain regions, and as shown in the present studies, corticosterone-mediated negative feedback at the pPVTh. In addition to corticosterone-mediated negative feedback in the pPVTh, there may be other hormones/neurotransmitters in the pPVTh or its efferent sites that contribute to inhibition of HPA responses to repeated stress. For example, the PVTh is densely innervated by fibers containing various stress-sensitive neurotransmitters such as cholecystokinin (53) for which a functional role in the pPVTh has been shown (16).
None of our manipulations of GR and/or MR in the pPVTh produced significant changes in corticosterone responses to acute or repeated stress (Figs. 6 and 8). Although we cannot point to one explanation with certainty, several possibilities have been investigated by others to explain the commonly observed dissociations in ACTH and corticosterone throughout the HPA literature. For example, the sensitivity of the adrenal gland to ACTH can change during stress (54, 55, 56, 57, 58, 59). Glucocorticoid production by the adrenal cortex has been shown to be controlled by additional non-ACTH mechanisms that relate to innervations of adrenocortical cells by the thoracic splanchnic nerve and by catecholaminergic and peptidergic nerve fibers from the adrenal medulla (60, 61, 62, 63). It is possible that adrenocortical functioning in our animals may have been altered by one of these factors independent of ACTH activity.
Habituation to repeated psychological stressors, such as restraint, is likely enabled by an integration of multiple inputs to the HPA axis from brain regions processing sensory, cognitive, and emotional information relevant to the stress experience. As previously suggested (14), the pPVTh may be part of a larger circuit of brain regions that regulate HPA activity under repeated stress conditions, including regions such as the mPFC and BLA. These regions are important because of the lack of direct projections from the pPVTh to the PVN and because limbic structures are likely important for processing of information that is related to habituation. The mPFC plays substantial roles in memory, learning, and emotional behavior (64, 65). The BLA has received considerable attention for its involvement in memory of stressful stimuli because it has been shown to integrate the effects of corticosterone on memory consolidation (66, 67). Our immunocytochemical results from experiment 5 demonstrate that the mPFC and BLA are projection sites of the pPVTh, confirming previous findings (24). Furthermore, we found that a subset of the pPVTh cells that project to the mPFC and BLA are GR positive, suggesting that the effects of corticosterone’s actions at GR in the pPVTh on HPA activity may be mediated by the mPFC and BLA, which indirectly project to the PVN (42, 68). A role in regulation of HPA responses to repeated stress has been shown for both the mPFC (12, 34) and BLA (35). Our findings do not exclude the possibility that other brain regions, such as the bed nucleus of the stria terminalis to which the pPVTh also projects (24), may also play important roles in pPVTh regulation of habituation but do strongly suggest involvement of the BLA and mPFC.
In addition to its inhibitory role in habituation, the pPVTh also inhibits HPA responses to acute novel restraint in rats previously exposed to repeated cold stress (16). We have not studied whether the pPVTh would similarly inhibit HPA responses to novel stress in rats previously exposed to repeated mixed stressors. Nevertheless, it is possible that corticosterone-mediated negative feedback at the pPVTh may play some part in inhibiting not only homotypic but also heterotypic stressors in repeatedly stressed rats. Therefore, if we had conducted the present experiments in rats that were repeatedly exposed to novel or mixed stressors, corticosterone implants in the pPVTh might constrain HPA responses to the eighth novel stressor compared with the first novel stressor. This outcome would still be consistent with the idea that the pPVTh is a site at which previous stress information is processed. However, the role of corticosterone in the pPVTh in regulation of HPA responses to mixed stressors has yet to be elucidated, and we feel it was beyond the scope of this set of studies.
In sum, the central finding of the present studies is that corticosterone released in response to restraint can act at GR and MR in the pPVTh throughout the week of stress and that these actions are important in producing habituation. These studies are the first to provide evidence that a thalamic region can be a site of corticosterone-mediated negative feedback. Furthermore, the actions of corticosterone at the pPVTh are more critical for the development of habituation than for the expression of habituation on the final day of stress. Last, GR-containing projections from the pPVTh innervate the BLA and mPFC, and such projections may underlie the cognitive processing of information related to habituation. Collectively, these data indicate that negative feedback action of corticosterone in the pPVTh is an important component of the limbic circuitry that determines the impact of previous stress experiences on an individual’s responses to subsequent stress exposure.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure summary: A.J. and S.B. have nothing to declare.
First Published Online June 29, 2006
Abbreviations: AP, Anteroposterior; BLA, basolateral amygdala; DV, dorsoventral; GR, glucocorticoid receptor; HPA, hypothalamic-pituitary-adrenal; ML, mediolateral; mPFC, medial prefrontal cortex; MR, mineralocorticoid receptor; pPVTh, posterior paraventricular thalamus; PVN, paraventricular nucleus; TPBS, Tris/PBS.
Received November 2, 2005.
Accepted for publication June 19, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
