- Modifications of Glucocorticoid Receptors mRNA Expression in the Hypothalamic‐Pituitary‐Adrenal Axis in Response to Early‐life Stress in Female Japanese Quail
- Targeting glucocorticoid receptors prevents the effects of early life stress on amyloid pathology and cognitive performance in APP/PS1 mice
- Early life stress
- Maternal behaviour
- Barnes maze
- Stress response
- Tissue preparation
- Diaminobenzidine immunohistochemistry
- Imaging and quantification
- Western blotting
- SDS-soluble Aβ levels
- Mifepristone treatment
- Statistical analysis
- Glucocorticoid receptor-mediated amygdalar metaplasticity underlies adaptive modulation of fear memory by stress
- What and how much we eat might change our internal clocks and hormone responses
Modifications of Glucocorticoid Receptors mRNA Expression in the Hypothalamic‐Pituitary‐Adrenal Axis in Response to Early‐life Stress in Female Japanese Quail
Volume 26, Issue 12
Stress exposure during early‐life development can programme individual brain and physiology. The hypothalamic‐pituitary‐adrenal (HPA) axis is one of the primary targets of this programming, which is generally associated with a hyperactive HPA axis, indicative of a reduced negative‐feedback.
This reduced feedback efficiency usually results from a reduced level of the glucocorticoid receptor (GR) and/or the mineralocorticoid receptor (MR) within the HPA axis. However, a few studies have shown that early‐life stress exposure results in an attenuated physiological stress response, suggesting an enhance feedback efficiency.
In the present study, we aimed to determine whether early‐life stress had long‐term consequences on GR and MR levels in quail and whether the effects on the physiological response to acute stress observed in prenatally stressed individuals were underpinned by changes in GR and/or MR levels in one or more HPA axis components.
We determined GR and MR mRNA expression in the hippocampus, hypothalamus and pituitary gland in quail exposed to elevated corticosterone during prenatal development, postnatal development, or both, and in control individuals exposed to none of the stressors.
We showed that prenatal stress increased the GR:MR ratio in the hippocampus, GR and MR expression in the hypothalamus and GR expression in the pituitary gland. Postnatal stress resulted in a reduced MR expression in the hippocampus. Both early‐life treatments permanently affected the expression of both receptor types in HPA axis regions.
The effects of prenatal stress are in accordance with a more efficient negative‐feedback within the HPA axis and thus can explain the attenuated stress response observed in these birds. Therefore, these changes in receptor density or number as a consequence of early‐life stress exposure might be the mechanism that allows an adaptive response to later‐life stressful conditions.
Stress exposure during early life has long lasting impacts on both the structure and function of several tissues associated with a higher risk of developing health pathologies and behavioural disorders 1-4. The concept of developmental programming (i.e.
permanent changes in physiology and neural systems following early‐life experiences) has been suggested to explain the negative consequences of environmental stress exposure during early development.
During this programming, environmental adversity experienced by the mother initiates maternal responses, which in turn affect the development of her offspring, including the organisation and functioning of specific tissues, especially the brain 2-4.
The main focus of research on these organisational effects has been their role in facilitating higher risk to later diseases or syndromes (e.g. cardiovascular disease, depression).
However, the environmental matching hypothesis proposes that developmental programming may prime the offspring to cope better with stressful conditions and thus may be adaptive when environmental conditions in later life match those experienced during early stages. The negative consequences of early‐life adversity may therefore simply result from a mismatch between conditions between environmental conditions at different life stages 5-7.
One fundamental physiological system that links an individual to changes in its environment is the hypothalamic‐pituitary‐adrenal (HPA) or stress axis 1, 7, 8. This axis is activated during in both development and adulthood when a stressor is perceived in the brain ultimately resulting in the release of glucocorticoid hormones from the adrenal cortex.
The increase in glucocorticoid levels facilitates a switch of physiological processes and behaviours from non‐essential activities to those that promote short‐term survival, such as increased locomotion and mobilisation of energy stores.
This response is tightly regulated by a negative‐feedback loop at the level of the hippocampus, hypothalamus and anterior pituitary to shut the HPA axis down and to return to a homeostatic point, avoiding the negative consequences of chronically elevated glucocorticoids 3, 9.
The effects of glucocorticoids in the brain, as well as the tight regulation of the axis, are mediated by two intracellular receptors: the mineralocorticoid receptor (MR) and the glucocorticoid receptor (GR) 10-12. In both mammals and birds, GR occur everywhere in the brain but are most abundant in the hippocampus and hypothalamus and are also found in the pituitary gland.
MR are mostly found in the hippocampus, and also in the hypothalamus, and bind glucocorticoids with a five‐ to ten‐fold higher affinity than GR 11-15. Consequently, MR remain activated during periods of basal secretion and are involved in the maintenance of integrity and stability of the HPA axis, primarily determining stress sensitivity of this axis.
MR can also boost the initial acute stress response and promote behavioural adaptation to stress during novel situations 11, 12, 16, 17. GR are additionally recruited when the glucocorticoid levels rise further, preventing the initial reaction from overshooting by bringing back cells to baseline levels via inhibition of the HPA axis.
They also facilitate the recovery from stress and are involved in the mediation of memory consolidation by glucocorticoids 10-12, 16. Because both MR and GR are involved in the regulation of the stress response and in the negative‐feedback loop, it has been proposed that the GR:MR balance is crucial to maintain homeostasis and its disruption may compromise stress resilience and affect behaviour 11-13, 15-17.
The HPA axis is one of the primary targets of early‐life stress programming in the brain 1, 2, 4, 8, 18.
High glucocorticoid levels during early development can permanently alter HPA axis functioning via alteration of baseline or stress‐induced glucocorticoid levels and/or negative‐feedback during stress recovery.
It has been generally shown that an increase in pre‐ or postnatal stress programmes a hyperactive HPA axis, indicative of a reduced negative‐feedback 1-4, 8, 19-22.
However, a few studies have shown that early‐life stress exposure can result in an attenuated physiological stress response later in life, suggesting enhanced negative‐feedback efficiency 23-26.
This programming of the HPA axis has been ascribed to modifications of GR and/or MR expression in the hippocampus and other feedback sites 1-3, 18. It has been recently shown in the Japanese quail (Coturnix japonica) that exposure to early‐life stress can permanently program physiology and behaviour in a potentially adaptive way 26. At the physiological level, prenatal stress altered the HPA axis functioning in a way that attenuated the acute stress response suggesting an increased negative‐feedback efficiency within the HPA axis 26.
In the present study, we aimed to determine whether early‐life stress had long‐term effects on GR and MR levels and whether the effects on the physiological response to acute stress observed in prenatally stressed individuals were underpinned by permanent modifications of the level of both glucocorticoids receptors (MR and GR) in feedback sites within the HPA axis. In the same quail, as used previously 26, we measured the relative mRNA expression of MR and GR in the hippocampus, hypothalamus and pituitary gland using a quantitative real‐time polymerase chain reaction (PCR). Because prenatal stress resulted in an attenuated stress response, we predicted that GR expression should be enhanced in the hippocampus, hypothalamus and pituitary gland, resulting in a more efficient negative‐feedback in prenatally stressed quail. In the hippocampus, the balance between GR and MR is crucial for resilience from a stressor and an increase in GR level or a decrease in MR level in the hippocampus could lead to more pronounced effects of GR, facilitating the recovery from stress 10-12, 17. We consequently hypothesised that the GR:MR ratio should be reduced in the hippocampus of quail exposed to prenatal stress.
Prenatal stress was manipulated by injecting eggs with 10 μl of corticosterone (CORT) (concentration: 850 ng/ml; Sigma Aldrich, Poole, UK) dissolved in sterile peanut oil at the egg apex under sterile conditions on day 5 of incubation (B). This increased the endogenous CORT concentration in the yolk within 1.
8 SD above control yolks, which was determined by radioimmunoassay and liquid chromatography‐mass spectroscopy and is similar to previous studies that have increased CORT levels within physiologically relevant ranges 24, 27. Control eggs (C) were injected with peanut oil alone.
Chicks of each prenatal treatment were subsequently randomly allocated to one of two postnatal food treatments: either food removal for 25% of daylight hours (3.5 h) on a random daily schedule for 15 days from post‐hatching day 4 (F−) or ad lib. food at all times (C).
Random removal of food has been shown to increase stress hormones in birds, without causing food restriction 28, 29. After this postnatal treatment, all birds had access to ad lib. food (Standard Layer Pellet, BOCM, Wherstead, UK) 26, 30.
The present study was part of a large experiment looking at the long‐term and trans‐generational effects of pre‐ and/or postnatal stress exposure in the Japanese quail. For GR and MR expression in this F1 generation, we focused on females because one of the aims of the project was to look how the maternal developmental environment affects offspring phenotype.
We thus had four treatment groups: prenatal control/postnatal control (CC; n = 6); prenatal control/postnatal food‐ (CF−; n =6), prenatal CORT/postnatal control (BC; n =12); and prenatal CORT/postnatal food‐ (BF−; n =6). All experimental procedures were carried out under Home Office Animals (Scientific) Procedures Act project licence 60/4068 and personal licence 70/1364 and 60/13261.
At the end of the experiment, when females were 246.5 ± 1.4 (SEM) days old, they were sacrificed by injection of an overdose of Dolethal (Vetoquinol, Buckingham, UK).
Brains were quickly removed (within 1 min) then pituitary glands that lie on the underside of the brain in the centre of the floor of the cranium were also removed using forceps (within 1 min after brain removal) and placed on dry ice until frozen, then stored at −80 °C. To perform the dissections, the brains were placed ventral side up into a brain matrix (Roboz Surgical Instrument Co.
, Gaithersburg, MD, USA) with a 1‐mm graduated scale placed on a mixture of dry and wet ice to keep the brain frozen and a 2‐mm thick coronal section was cut using two razor blades positioned approximately 4 mm from the rostral pole and 2 mm from the cerebellum. The plane of cutting was adjusted to match as closely as possible the plane of the chicken brain atlas 31.
Then, when still frozen, two equivalent bilateral punches (1 mm in diameter each) were obtained from the hippocampus and a single punch was obtained from the medial hypothalamus that spanned the third ventricle. Each sample was stored separately at −80 °C.
Total RNA was extracted and purified using Absolutely RNA Miniprep kits (Agilent Technologies, Santa Clara, CA, USA) in accordance with the manufacturer's instructions.
The quantity and integrity of RNA were assessed with a RNA 6000 Pico assay kit for hippocampus and hypothalamus and a RNA 6000 Nano kit for pituitary gland using the Agilent 2100 Bioanalyzer (Agilent Technologies) in accordance with the manufacturer's instructions. The mean RIN number of these samples is 8.2 ± 0.1 (range 5.2–10).
First‐strand cDNA was synthesised using Affinity Script Multiple Temperature cDNA Synthesis kits (Agilent Technologies) and diluted to obtain a final concentration of 25 pg/μl.
This resulting cDNA was used to perform quantitative real‐time PCR (qPCR) for the genes of interest [GR and MR and the house‐keeping gene β‐actin (BA)] for the different brain regions using gene‐specific primers. BA was determined as the best candidate house‐keeping gene for our samples (M =0.30, other candidates M ≥ 0.
34) using a chicken (Gallus gallus) GeNorm kit (Primerdesign, Southampton, UK). Specific PerfectProbe™ primers (Primerdesign) were designed published chicken nucleotide gene sequences and were validated using quail cDNA by Primerdesign. These primers amplified single products with no dimer pairs.
GR sense primer: TAATGACCGTGGTGACCTTTTA, anti‐sense primer: TTTCTTGCTTTATGCCAGGAGTA (GenBank accession number NM_001037826). MR sense primer: GTAGAATAGAGGACAGATGAACTTTT, anti‐sense primer: ACCCAGAGAGAACACTACAGAT (GenBank accession number NM_001159345).
All qPCR reactions were run in duplicate and were performed in 20‐μl reactions containing 10 μl of 2 × Brilliant III Ultra‐Fast QPCR Master Mix (Agilent technologies), 1 μl of specific PerfectProbe™ primer (Primerdesign) at a working concentration of 300 nm, 0.3 μl of reference dye, 3.
7 μl of RNAse/DNAase‐free water and 5 μl of appropriate cDNA along with no‐template controls and blanks. Reactions were carried out on a Stratagene MX 3005P (Agilent Technologies) at 95 °C for 3 min, then 50 cycles of 95 °C for 20 s and 60 °C for 20 s. From standard curves generated with known concentration of cDNA, we determined that the amplification efficiency [Eff =10(−1/slope)−1] was higher than 95% for GR, MR and BA. Therefore, we used the delta Ct method (ΔCt) to quantify the relative expression of GR and MR relative to BA: 2−(Ct GR/MR – Ct BA) 32.
To compare the relative expression of both receptors in the different regions of the HPA axis, we used a generalised linear model (GLM) fitted with a gamma law because the residuals of linear models were not normally distributed, using the genmod procedure in sas, version 9.
4 (SAS Institute Inc., Cary, NC, USA). Receptor type (GR or MR), tissues (hippocampus, hypothalamus and pituitary gland) and their interaction were used as fixed factors.
To determine the consequences of exposure to early‐life stress on the GR:MR ratio, MR and GR relative expression, we also used GLMs fitted with a gamma law because the residuals of linear models were not normally distributed.
Pre‐ and postnatal treatments and their interactions were specified as fixed factors. For multiple comparisons, Tukey–Kramer adjustment was applied to obtain a corrected value. P 3.39, P 2.05, P
Targeting glucocorticoid receptors prevents the effects of early life stress on amyloid pathology and cognitive performance in APP/PS1 mice
Exposure to chronic stress or elevated glucocorticoid hormone levels in adult life has been associated with cognitive deficits and an increased risk for Alzheimer’s disease (AD).
Since exposure to stress during early life enhances stress-responsiveness and lastingly affects cognition in adult life, we here investigated; (i) whether chronic early life stress (ELS) affects AD pathology and cognition in middle-aged APPswe/PS1dE9 mice, and (ii) whether it is still possible to rescue these late effects by briefly blocking glucocorticoid receptors (GRs) at a translationally relevant, middle age. Transgenic APPswe/PS1dE9 mice were subjected to ELS by housing dams and pups with limited nesting and bedding material from postnatal days 2–9 only. In 6- and 12-month-old offspring, this resulted in enhanced hippocampal amyloid-β (Aβ)-40 and -42 levels, and in reduced cognitive flexibility, that correlated well with the Aβ42 levels. In parallel, CORT levels and BACE1 levels were significantly elevated. Surprisingly, blocking GRs for only 3 days at 12 months of age reduced CORT levels, reduced hippocampal Aβ40 and -42, and β-site APP-cleaving enzyme 1 (BACE1) levels, and notably rescued the cognitive deficits in 12-month-old APPswe/PS1dE9 mice. These mouse data demonstrate that exposure to stress during the sensitive period early in life influences later amyloid pathology and cognition in genetically predisposed, mutant mice, and as such, may increase AD vulnerability. The fact that a short treatment with a GR antagonist at middle age lastingly reduced Aβ levels and rescued the cognitive deficits after ELS, highlights the therapeutic potential of this drug for reducing amyloid pathology.
The mechanisms that underlie sporadic Alzheimer’s disease (AD) remain largely elusive, which hampers the development of successful intervention strategies for AD.
While familial forms of AD can be explained by genetic causes—often related to changes in amyloid-β (Aβ) (e.g.
refs 1,2)—sporadic AD ly has a multifactorial aetiology, in which, next to amyloid, also lifestyle factors play an important role2,3,4,5.
Stress is an important environmental risk factor that has been implicated in AD progression6,7.
Clinical observations suggest that stressful life events can reduce the age of onset in AD6, while stress-related disorders depression can promote AD symptoms and neuropathology8.
Glucocorticoid hormones (GCs; cortisol in humans and corticosterone (CORT) in rodents) are powerful steroids released in response to stress.
They are often increased in AD, notably already in early stages of the disease9,10, and dysregulation of the hypothalamus–pituitary–adrenal (HPA) axis is also associated with a higher AD risk8,10,11. Rodent studies further demonstrate that exposure to stress and/or elevated GC levels at an adult age impairs cognition and enhances Aβ levels both in mutant12,13,14,15,16 and in wild type (WT) animals17,18.
The early postnatal period is a particularly sensitive time window that determines sensitivity to stress and cognitive function in later life.
As exposure to early life stress (ELS) in WT mice is well-known to accelerate cognitive decline19,20, we tested the hypothesis that ELS—induced by housing mice with limited nesting and bedding material from postnatal days (PND) 2–921,22,23,24—increases the development of AD pathology and cognitive decline in APPswe/PS1dE9 mice, a classic mouse model for amyloid pathology25.
Secondly, in order to study whether glucocorticoids are instrumental, we tested whether briefly targeting glucocorticoid receptors (GRs) could rescue late effects of ELS on AD-related pathology and cognitive performance.
We therefore treated animals with mifepristone, which is a Food and Drug Administration-approved drug that selectively blocks GRs at high concentrations and is prescribed to treat Cushing’s disease. It has further been tested in preliminary studies on (aspects of) AD26 and psychotic depression27,28 (see ref. 29 for a review about the function and applicability of mifepristone in humans).
In this study, we conducted experiments under Dutch national law as well as under European Union directives on animal experiments. The animal welfare committee of the University of Amsterdam approved all experiments. Mice were housed at a temperature of 20–22 °C. Humidity was between 40 and 60%, and animals were fed ad libitum with standard chow and water.
Lights were on between 8.00 a.m. and 8.00 p.m. unless stated otherwise. WT and APPswe/PS1dE925 male littermates of 6 and 12 (±1) months of age were used. To obtain mice, two 10-week-old C57Bl/6J virgin WT females (Harlan Laboratories B.V., Venray, The Netherlands) and one heterozygous male APPswe/PS1dE9 mouse were housed together for 1 week to allow mating.
Individually housed pregnant females were monitored daily for the birth of pups. For litters born before 10.00 a.m., the day of birth (PND 0) was considered the previous day, after which the ELS paradigm was initiated from PND 2 to 9. At PND 21, mice were weaned and ear tissue was collected for identification and genotyping.
Littermates were housed with 2–6 mice per cage. All animals were undisturbed (except for cage cleaning once a week) until the start of the experimental procedures. Number of mice—6 months: Ctrl-WT, 12; Ctrl-APPswe/PS1dE9, 10; ELS-WT, 11; and ELS-APPswe/PS1dE9, 14; 12 months: Ctrl-WT, 16; Ctrl-APPswe/PS1dE9, 11; ELS-WT, 19; and ELS-APPswe/PS1dE9, 12.
Group sizes were chosen to ensure sufficient statistical power.
Early life stress
At PND 2, litters were culled to six pups per litter, and dams and their litters were randomly assigned to the ELS or control (Ctrl) condition until PND 9, after which all mice were treated equally, as described before21,22,23,24.
Briefly, the Ctrl condition consisted of cages with standard amounts of nesting and bedding material (one piece of nesting material (5×5 cm; Technilab-BMI, Someren, the Netherlands)).
In the ELS condition, a fine-gauge stainless steel mesh was placed in the cage, with a small amount of sawdust bedding and ½ piece of nesting material.
Maternal behaviour was observed daily from PND 3 until PND 8 at 9.00 a.m. and 8.30 p.m as described previously30. Briefly, the level of activity of the dam was scored in sixteen 1-min epochs per day, spread over 48 min observation sessions. The behaviours that were scored were: licking and grooming behaviour; nursing behaviour; and the time that the dam spent off the pups.
Six- and twelve-month-old APPswe/PS1dE9 and WT male mice were tested for spatial memory in the spatial Barnes maze task. Testing occurred during the dark, active phase in the afternoon (1 p.m.).
A classic setup was used (110 cm diameter, 12 exit holes) in which mice were placed in the centre of the maze twice daily (inter-trial interval of 30 min) for 4 consecutive days and allowed to navigate to the exit hole leading to the home cage (acquisition learning). A probe trial (all holes closed) was conducted 24 h after the last trial.
Following the probe trial, behavioural flexibility was tested by relocating the exit hole to another location on the maze (150°) for 4 days (reversal learning), followed by a probe trial.
Cages containing used bedding material were placed at equal distance under the maze to avoid guidance by odour cues, the board was rotated after each trial and the maze was cleaned with 25 % EtOH to dissipate odour cues. Distal extra-maze cues were always fixed relative to the exit hole. Performance of the mouse was assessed by an observer blind to the experimental condition of the mouse.
Two acute stressors were used to determine stress responsiveness; a 6-min forced swim test (original experiment) or an acute 0.4 mA foot shock (mifepristone experiment).
Blood samples were collected by a tail cut at 30 min (peak stress response) and 90 min (stress recovery) after the stressor.
A commercially available radioimmunoassay kit (MP Biomedicals, Eindhoven, The Netherlands) was used to measure plasma CORT levels.
Following behavioural testing, mice were sacrificed by quick decapitation, between 8.00 and 9.00 p.m. (beginning of the inactive phase), i.e., when plasma CORT levels are low. Blood plasma was collected and CORT levels were determined as described above.
For diaminobenzidine (DAB) immunohistochemistry, pre-mounted sections (40 µm) on glass slides (Superfrost Plus slides, Menzel) were dried overnight. Endogenous peroxidase activity was blocked using 0.3% H2O2 for 15 min, after which sections were boiled in a microwave (±95 °C) in citrate buffer (0.01 M, pH 6.
0, 15 min). To block non-specific staining, slices were incubated for 1 h in blocking mix (0.05 M Tris-buffered saline (TBS) containing 1% bovine albumin serum (BSA) and 0.1% triton), and primary antibody mix (6E10 (1:1500, Bioline) in blocking mix) was applied for 2 h at room temperature and overnight at 4 °C.
Secondary antibody (1:200, sheep anti-mouse biotinylated (GE Healtcare)) was applied for 2 h, after which sections were treated with avidin-biotin complex (ABC, 1:800, Vectastain elite ABC-peroxidase kit, Brunschwig Chemie). Chromagen development was conducted by incubation in 0.05 M Tris buffer containing 0.
01% H2O2 and 0.2 mg/ml DAB.
Imaging and quantification
Plaque load was quantified for all APPswe/PS1dE9 mice by an experimenter who was blind to the experimental conditions.
Using a Nikon DS-Ri2 microscope, representative images (×10 magnification) were captured and analysed using ImageJ. A fixed intensity threshold was applied to 8-bit binarised pictures to define the DAB staining.
The percentage of area covered by DAB staining was analysed as described previously31.
Following quick decapitation and dissection, hippocampi were snap-frozen and homogenised in RIPA buffer (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, pH = 6.8). Homogenates were sonicated for 2 × 30 s (max intensity), and centrifuged for 1 min (10 000 × g, 4°C).
Protein concentration was determined in the supernatant by BCA Protein Assay (Pierce, The Netherlands). 15 µg protein was separated on 12.5% polyacrylamide-SDS gels using electrophoresis, and proteins were transferred to polyvinylidene difluoride membranes at 75 V in Towbin buffer (25 mM Tris, 192 mM glycine, 20% methanol, pH = 8.3).
Five per cent BSA in TBST (0.1 M TBS + 0.1% Tween-20) was used to block the membranes, and membranes were incubated overnight at 4 °C with primary antibody.
Proteins studied were as follows: amyloid precursor protein (APP; 6E10, 1:1500, BioLegend, 100 kDa); β-site APP-cleaving enzyme 1 (BACE1; D10E5, 1:1000, Cell Signalling, 70 kDa); actin (A2066, Sigma-Aldrich, 42 kDa); and GAPDH (14C10, 1:3000, Cell Signalling, 37 kDa).
Secondary antibodies were incubated for 2 h, and signal was developed (Licor Odyssey FC; Leusden, the Netherlands). Using ImageJ (NIH; Bethesda) signal intensities were measured and normalised against GAPDH or actin as internal marker. Three independent replications were made, and protein levels were analysed as the mean of these replicates, and expressed as % of Ctrl-APPswe/PS1dE9 levels.
SDS-soluble Aβ levels
Aβ40 and -42 peptide levels were determined in whole hippocampal homogenates using a sandwich enzyme-linked immunosorbent assay (ELISA) as described previously32.
Mifepristone (Sigma; 40 mg/ml dissolved in 99.9% EtOH and diluted 20× in arachide oil) was injected intraperitoneally for 3 consecutive days (final dose: 10 mg/kg; injection volume: 5 µl/g body weight) between PND 339–341 (±7 days)23. The appropriate vehicle solution was administered accordingly.
Data were analysed using SPSS 22.0 (IBM software). Data are expressed as mean ± standard error of the mean. Data were considered statistically significant when p
Glucocorticoid receptor-mediated amygdalar metaplasticity underlies adaptive modulation of fear memory by stress
Glucocorticoid receptor (GR) is crucial for signaling mediated by stress-induced high levels of glucocorticoids. The lateral nucleus of the amygdala (LA) is a key structure underlying auditory-cued fear conditioning.
Here, we demonstrate that genetic disruption of GR in the LA (LAGRKO) resulted in an auditory-cued fear memory deficit for strengthened conditioning. Furthermore, the suppressive effect of a single restraint stress (RS) prior to conditioning on auditory-cued fear memory in floxed GR (control) mice was abolished in LAGRKO mice.
Optogenetic induction of long-term depression (LTD) at auditory inputs to the LA reduced auditory-cued fear memory in RS-exposed LAGRKO mice, and in contrast, optogenetic induction of long-term potentiation (LTP) increased auditory-cued fear memory in RS-exposed floxed GR mice.
These findings suggest that prior stress suppresses fear conditioning-induced LTP at auditory inputs to the LA in a GR-dependent manner, thereby protecting animals from encoding excessive cued fear memory under stress conditions.
Stress activates the hypothalamus–pituitary–adrenal (HPA) axis, which results in the release of glucocorticoid hormones (cortisol in humans and corticosterone in rodents) from the adrenal cortex.
Glucocorticoid hormones can readily enter the brain and bind to specific receptors in regions crucial for memories of stressful experiences, such as the hippocampus and amygdala, thereby enhancing the consolidation of emotionally arousing events (de Quervain et al., 2009; Roozendaal et al., 2009).
Two types of receptors mediate the effects of glucocorticoids: type I mineralocorticoid receptor (MR, Nr3c2) and type II glucocorticoid receptor (GR, Nr3c1).
Compared with MR, GR has a lower binding affinity for glucocorticoids and is largely unoccupied at basal levels, and is thus considered to be particularly important in signaling mediated by stress-induced high levels of glucocorticoids (de Kloet et al., 2005; Reul and de Kloet, 1985).
The lateral nucleus of the amygdala (LA) is a key structure underlying auditory-cued fear conditioning (AFC) (LeDoux, 2000). Auditory information, which is critical for AFC, reaches the LA either from the medial geniculate nucleus (MGN) or from the auditory cortex (AC).
Long-term potentiation (LTP) in the two auditory inputs to the LA is essential for the acquisition and expression of auditory-cued fear memory (Blair et al., 2001; Rogan et al., 1997; Tsvetkov et al., 2002).
In response to stress, GR activation plays a central role in the formation of long-term memory, which is an essential mechanism for learning from stressful events and respond adaptively to similar demands in the future (Finsterwald and Alberini, 2014).
In an electrophysiological study using brain slices, glucocorticoid prolonged excitatory synaptic responses in the basolateral complex of the amygdala (BLA) by binding to GR (Karst et al., 2010). In contrast, after an acute stress exposure, application of glucocorticoid suppressed excitatory synaptic responses in the BLA in a GR-dependent manner.
This switch in synaptic response to glucocorticoid is referred to as metaplasticity (Abraham and Tate, 1997; Schmidt et al., 2013). Prior delivery of behavioral stress has also been shown to suppress subsequent induction of LTP (Kavushansky and Richter-Levin, 2006). However, whether stress-induced amygdalar metaplasticity occurs in vivo in a way relevant to the strength of the auditory-cued fear memory, and what role LAGR plays in this neural process are unknown.
In this study, we generated LA-selective GR knockout (LAGRKO) mice to investigate the region-specific role of LAGR in mediating the modulatory effects of stress on fear memory. We first compared contextual and auditory-cued fear memory conditioned at different strengths between floxed GR (control) and LAGRKO mice at basal condition.
We then investigated the effect of LAGR disruption on the adaptive modulation of fear memory after exposure to acute stress.
Finally, using an optogenetic technique to induce LTP and long-term depression (LTD), we investigated how GR-dependent metaplasticity in the LA influenced auditory-cued fear memory in response to prior stress exposure.
To selectively knockout the GR gene in the LA, we generated two mutant mouse lines. The knock-in line expressed improved Cre (iCre) recombinase (Shimshek et al., 2002) under the control of the gastrin-releasing peptide gene (Grp) promoter, and the animals are referred to as Grp-iCre mice (Figure 1A,B).
Grp is abundant in the LA and absent or present at low levels in other subnuclei of the amygdala (Shumyatsky et al., 2002). Thus, Grp promoter could be used for driving the expression of Cre recombinase selectively in the LA.
To confirm its usefulness, we crossed a Grp-iCre mouse with a CAG-CAT-Z reporter mouse, which carried the chloramphenicol acetyltransferase gene (CAT) flanked by two loxP sites and the β-galactosidase gene in sequence (Araki et al., 1995).
Cre-mediated recombination between the two loxP sites resulted in the expression of β-galactosidase, which was detected by X-gal staining.
As shown in Figure 1C, strong β-galactosidase expression was detected in the LA and hippocampal CA3 region of Grp-iCre/CAG-CAT-Z mice, indicating the presence of robust Cre-loxP recombination in these brain regions. In addition, sparse and weak expression of β-galactosidase appeared in the accessory basal nucleus of the amygdala and in layer 6 of the cerebral cortex.
(A) Schematic diagram of the gene targeting strategy. iCre and pgk-neo cassettes flanked by two FRT sites were inserted into the gastrin-releasing peptide gene (Grp) locus. Met is the translation initiation site of Grp. The location of PCR primers (Gic1, Gic2, Gic3, NeoP1, and NeoP2) used for genotyping are indicated.
DT, diphtheria toxin gene; pBSK, pBluescriptII SK. The chimeric mouse obtained was crossed with a CAG-FLPe mouse to delete the pgk-neo cassette and establish the Grp-iCre (Gic) mouse line. (B) Genotyping PCR of genomic DNA prepared from WT; Gic+/−, Neo+; Gic+/−, and ΔNeo mice.
(C) Cre activity in Grp-iCre mice was examined by crossing Grp-iCre mice with lacZ reporter mice (CAG-CAT-Z) mice. β-galactosidase expression in a Grp-iCre/CAG-CAT-Z mouse brain stained with X-gal.
X-gal staining revealed robust Cre-loxP recombination in the lateral nucleus of the amygdala (LA) and the hippocampal CA3 region, with sparser recombination in the accessary basal nucleus of the amygdala (AB), and in layer 6 of the cerebral cortex (VI).
Next, we generated a floxed GR (Nr3c1loxP/loxP) mouse line in which exon 3 of the Nr3c1 gene (encoding the DNA-binding domain) was flanked by two loxP sites (Figure 1—figure supplement 1). Grp-iCre mice were crossed with floxed GR mice to establish the LAGRKO mouse line (Nr3c1loxP/loxP, GrpiCre+/−).
Immunofluorescence staining with a specific antibody against GR revealed a selective disruption of GR proteins in the LA of LAGRKO mice (Figure 2A). There was no significant difference between the two genotypes in the expression level of GR in the central nucleus and basal nucleus of the amygdala.
Double immunofluorescence staining with anti-GR and anti-NeuN (a neuronal marker) antibodies showed that GR was undetectable in approximately 70% of LA neurons in LAGRKO mice (floxed GR, 90.88 ± 1.39%; LAGRKO, 20.27 ± 0.95%; Figure 2A,B, Figure 2—figure supplement 1A).
The expression level of GR in the hippocampal CA3 region of floxed GR mice was very low, and it was slightly decreased in LAGRKO mice (Figure 2C).
There was no significant difference between the two genotypes in the expression level of GR in the cerebral cortex (Figure 2C, Figure 2—figure supplement 1B), hippocampal CA1 and CA2 regions, and dentate gyrus (Figure 2C). Collectively, these results indicate the successful establishment of a novel LAGRKO mouse line.
(A) The LAGRKO mouse line (GRflox/flox, Grp-iCre+/-) was established by crossing floxed GR (GRflox/flox) and Grp-iCre mice.
Double immunofluorescence staining of GR (green, left panels) and NeuN (magenta, middle panels) in coronal brain sections from floxed GR and LAGRKO mice.
The overlap of green and magenta signals (white, right panels) indicates the expression of GR in LA neurons in floxed GR mice (upper), which was apparently reduced in LAGRKO mice (lower). Magnified images of the boxed areas are shown in the insets.
LA, lateral nucleus of the amygdala; BA, basal nucleus of the amygdala; CeA, central nucleus of the amygdala. (B) Quantification of GR+ and NeuN+ cells in the LA of floxed GR and LAGRKO mice (n = 9 sections from three mice). Data are presented as mean ± S.E.M. **p
What and how much we eat might change our internal clocks and hormone responses
For the first time, a study shows how glucocorticoid hormones, such as cortisol, control sugar and fat levels differently during day and night, feeding and fasting, rest and activity, over the course of 24 hours.
The research conducted in mice found that the time-of-day dependent metabolic cycle is altered by high caloric diet.
Since glucocorticoids are widely used drugs for the treatment of inflammatory diseases, these findings published in Molecular Cell suggest that lean and obese patients might respond differently to steroid therapy.
Finally, it reveals the biological function of daily rhythms of hormone secretion (high before awakening and feeding, low when sleeping and fasting) as well as daily cycles of sugar and fat storage or release by the liver.
Each cell in the human body is driven by an internal clock which follows the circadian rhythm of 24 hours. It is synchronized with the natural cycle of day and night mainly by sunlight, but also through social habits. In a healthy system, glucocorticoid stress hormones, are produced every morning by the adrenal gland.
The secretion of glucocorticoidpeaks before awakening, prompting the body to use fatty acids and sugar as sources of energy, and enabling us to start our daily activities. When the circadian rhythm is disrupted (e.g. through shift work or jetlag) and/or when the glucocorticoid level alters (e.g.
through Cushing syndrome or long-term clinical application), profound metabolic dysregulation can be caused — obesity, type 2 diabetes, and fatty liver disease.
The researcher's goal therefore was to understand the relevance of these daily peaks of stress hormone secretion, the impact of these hormones on our “internal clock” and their role for daily cycles of metabolism.
Glucocorticoids' metabolic actions in the liver
To study glucocorticoids' metabolic actions in the liver, the researchers characterized the activity of their receptor, called the glucocorticoid receptor, using novel high throughput techniques. They analyzed mouse livers every 4 hours during day and night. The mice were either in normal condition or fed with high-fat diet.
They then used cutting-edge technologies in genomics, proteomics, and bioinformatics to picture when and where the glucocorticoid receptor exerts its metabolic effects. The researchers dissected the impact of daily surges of glucocorticoid release in the 24-hour-cycle of liver metabolism.
They could illustrate how glucocorticoids regulate metabolism differently during fasting (when the mice sleep) and during feeding (when they are active), by time-dependent binding to the genome. Furthermore, they showed how the majority of rhythmic gene activity is controlled by these hormones.
When this control is lost (in so-called knockout mice), blood levels of sugar and fat are affected. This explains how the liver controls blood levels of sugar and fat differently during day and night.
In a next step, as the glucocorticoid receptor is a widely-used drug target in immune therapies, they investigated its genomics effects after the injection of the drug dexamethasone, a synthetic glucocorticoid that also activates this receptor.
“With this experiment,” explains Dr. Fabiana Quagliarini, “we found that the drug response was different in obese mice compared to lean mice. It is the first time to show that diet can change hormonal and drug responses of metabolic tissues.
New insights for Chronomedicine and metabolic disease therapy
Glucocorticoids are a group of natural and synthetic steroid hormones such as cortisol. They have potent anti-inflammatory and immunosuppressive properties which can control the activity of the immune system.
This is why they are widely exploited in medicine.
The major drawback is that glucocorticoids also cause severe side effects by virtue of their ability to modulate sugar and fat metabolism: Patients may develop obesity, hypertriglyceridemia, fatty liver, hypertension or type 2 diabetes.
“Understanding how glucocorticoids control 24-hour-cycles of gene activity in the liver and consequently blood levels of sugar and fat, provides new insights into 'Chronomedicine' and the development of metabolic disease.
We could describe a new link between lifestyle, hormones and physiology at the molecular level, suggesting that obese people may respond differently to daily hormone secretion or to glucocorticoid drugs.
These mechanisms are the basis for the design of future therapeutic approaches,” highlights Prof. Henriette Uhlenhaut.
make a difference: sponsored opportunity