A role for glucocorticoid-signaling in depression-like behavior of gastrin-releasing peptide receptor knock-out mice

Abstract

Background. The gastrin-releasing peptide receptor (GRPR) is highly expressed in the limbic system, where it importantly regulates emotional functions and in the suprachiasmatic nucleus, where it is central for the photic resetting of the circadian clock. Mice lacking GRPR presented with deficient light-induced phase shift in activity as well altered emotional learning and amygdala function. The effect of GRPR deletion on depression-like behavior and its molecular signature in the amygdala, however, has not yet been evaluated.

Methods. GRPR knock-out mice (GRPR-KO) were tested in the forced-swim test and the sucrose preference test for depression-like behavior. Gene expression in the basolateral nucleus of the amygdala was evaluated by micorarray analysis subsequent to laser-capture microdissection-assisted extraction of mRNA. The expression of selected genes was confirmed by RT-PCR.

Results. GRPR-KO mice were found to present with increased depression-like behavior. Microarray analysis revealed down-regulation of several glucocorticoid-responsive genes in the basolateral amygdala. Acute administration of dexamethasone reversed the behavioral phenotype and alterations in gene expression.

Discussion. We propose that deletion of GRPR leads to the induction of depression-like behavior which is paralleled by dysregulation of amygdala gene expression, potentially resulting from deficient light-induced corticosterone release in GRPR-KO.

Key words::

• Amygdala
• behavior
• depression
• glucocorticoid
• laser-capture microdissection
• microarray

Abbreviations
ACTH	=	adrenocorticotropin
arc	=	activity regulated cytoskeletal-associated protein
BLA	=	basolateral amygdala
CREB	=	cyclic-AMP response element binding
DEX	=	dexamethasone
egr1	=	early growth response 1
cfos	=	FBJ osteosarcoma oncogene
FST	=	forced-swim test
GRP	=	gastrin-releasing peptide
GRPR	=	gastrin-releasing peptide receptor
GRPR-KO	=	GRPR knock-out mice
HPA	=	hypothalamus–pituitary–adrenal
LCM	=	laser-capture microdissection
MAPK	=	mitogen-activated protein kinase
RT-PC	=	Rreverse transcriptase PCR
sgk-1	=	serum/glucocorticoid regulated kinase
SPT	=	sucrose preference test
WT	=	wild-type mice

Key messages

• Gastrin-releasing peptide receptor (GRPR) deletion induces depression-like behavior in mice.

• Depression-like behavior in GRPR knock-out mice is paralleled by dysregulation in the expression of glucocorticoid-responsive genes in the amygdala.

• The behavioral and gene expression phenotype can be rescued by acute administration of dexamethasone, a synthetic glucocorticoid.

Introduction

The gastrin-releasing peptide receptor (GRPR), which binds gastrin-releasing peptide (GRP)—belonging to the family of bombesin (BB)-like peptides—is highly expressed in the mammalian brain. GRPR has been found to be enriched in the hypothalamus, specifically in the suprachiasmatic nucleus (SCN) (1), suggesting a role in the regulation of circadian rhythms and in the pituitary gland (2), indicative of a role in the regulation of the hypothalamus–pituitary–adrenal (HPA) axis hormone secretion. Moreover, high densities of GRPR are found in the paraventricular nucleus of the hypothalamus which connects to amygdala (3), and GRP has been identified as amygdala-enriched gene (4). The anatomical distribution of GRPR is in line with its proposed function in emotional and stress-related behaviors (4–6) and its emerging role as molecular key player in mental disorders (see for review (7)). Specifically, a role for GRPR in amygdala-mediated learned fear has been suggested since GRPR-KO (knock-out) mice exhibited increased and more persistent long-term conditioned fear (4). Classical fear conditioning in experimental animals has proven a powerful tool for investigating the neural and molecular basis of emotional learning and the physiological and pathologically exacerbated expression of fear-related states, as present in human anxiety disorders. Enhanced and persistent fear memory in GRPR-KO mice thus suggests this mouse line as a suitable animal model for studying the neurobiological basis of mental disorders involving increased fear and anxiety. Although a participation of GRPR in several psychiatric diseases, including dementia, schizophrenia, autisms, and anxiety disorders, has been suggested (see for review (7)), its potential role in the pathophysiology of depression has not been studied yet. Given the high co-morbidity and partial overlap of symptoms and potentially underlying neural mechanisms of anxiety and depressive disorders (8,9), we decided to use GRPR-KO mice to explore the relevance of GRPR in depression-like behavior and to investigate its molecular basis in the basolateral nucleus of the amygdala, a key brain region responsible for emotional functions.

Materials and methods

Animals

Male mice, GRPR-KO and wild-type (WT) littermates (10–12 weeks old), were used for all experiments. Mice were kept single-housed in standard transparent laboratory cages on a 12 hour light/dark cycle (with lights on at 6.00 a.m.) in a temperature-controlled colony room (22 ± 1°C) and were provided with food and water ad libitum. All animal procedures described were carried out at Columbia University, approved by the Institutional Animal Care and Use Committees of Columbia University and the New York State Psychiatric Institute, and executed in accordance with National Institute of Health regulations. GRPR knock-out mice were described before and were found grossly normal (10). Mice used for the study were back-crossed to N10 or more to C57BL/6J strain, and genotypes were determined by PCR performed on tail DNA (10).

Pharmacological treatment

Dexamethasone 21-phosphate disodium (DEX-P; Sigma, St Louis, MO, USA) was dissolved in ethanol and subsequently diluted with saline to a final concentration of 2.5% ethanol and administrated according to a previously described dosage regime (11). Dexamethasone (0.3 mg/kg) and 2.5% ethanol in saline (0.9% NaCl) were injected via an intraperitoneal route (i.p.) in a final injection volume of 5 mL/kg 2 h prior to behavioral testing and brain dissection.

Behavioral testing

All behavioral experiments were conducted during the light phase between 9 a.m. and 12 a.m. Mice were allowed to acclimate to the experimental room for 1 h prior to the experiment. Animals were handled daily for 3 days prior to start of behavioral experiments. In the case of the pharmacology experiments, mice were injected i.p. with saline in the course of each of these handling sessions.

Forced-swim test (FST). The forced-swim test protocol followed published methods (12). Briefly, mice were placed into plastic buckets, 19 cm in diameter and 23 cm deep filled with 23–25°C water, and videotaped for 6 min by an overhead camcorder. The last 4 minutes were scored by analysis of the videotapes for immobility by an experienced experimenter blind to the experimental groups. Immobility was defined as passive floating of the animal in the water without any movements other than those necessary for the mouse to keep its head above the water.

Sucrose preference test (SPT). Animals were first habituated to the paradigm by giving a 48-h exposure to two bottles, one containing 2% sucrose solution, the other autoclaved water, 72 hours prior to the actual sucrose preference test. The bottles were counterbalanced across the left and right side of the feeding compartment after 24 h. During the test, mice were given a free choice between two bottles, one with 2% sucrose solution and the other with autoclaved water, during 24 h. The bottles were counterbalanced across the left and right side of the feeding compartment after 12 h. The consumption of water and sucrose solution was estimated simultaneously in control and experimental groups by weighing the bottles. Sucrose preference rate was calculated according to the formula: % preference = ((sucrose intake/total intake) × 100%).

Gene expression analysis

Brain dissection. Animals were sacrificed by cervical dislocation, and brains were rapidly taken out, snap-frozen, and stored at −80°C until needed; 8-μm brain sections were sliced with a cryostat and immediately stored at −80°C in a dry container. Blood was collected via cardial puncture into 1.5 mL Eppendorf tubes (containing 7.5 U heparin). Upon centrifugation (12,000 g) at 4°C for 5 min, plasma was collected and stored in aliquots at −80°C until used for further analysis.

Laser-capture microdissection (LCM). Per animal (n = 5 per group) a total of three coronal sections containing amygdala slices were subjected to LCM essentially as previously described (12). Brain sections were removed from −80°C and immediately dehydrated in a gradient alcohol series (75%, 90%, and 100% ethanol twice, 30 s each) and a final incubation in xylene for clearance (5 min). All reaction steps were performed in RNase-free solutions. Sections were then air-dried under laminar flow for 5 min and immediately used for LCM. LCM was carried out using a PixCell II system (Arcturus Bioscience Inc., Mountain View, CA, USA). Selected regions were lifted onto CapSure LCM plastic caps (Arcturus Bioscience Inc., Mountain View, CA) using a spot size of 15 μm, a laser power of 50 μV, and a duration of 2 ms. Caps with transfer films with the microdissected tissue were immediately placed into Eppendorf tubes containing 350 μL of lysis buffer, incubated at 37°C for 30 min, and stored in −80°C before RNA isolation. Total RNA was extracted from samples collected by LCM caps using RNAqueous^®-Micro Kit (Ambion Inc., Austin, TX) including DNAse treatment to remove potential genomic DNA contamination according to the manufacturer's instructions.

Microarray experiments. In order to obtain sufficient amounts of RNA for micorarray analysis two rounds of linear amplification were conducted using the GeneChip^® Two-Cycle Target Labeling kit (Affymetrix Inc., Santa Clara, CA) according to the supplier's instruction. Briefly, antisense RNA amplification was based on T7 RNA polymerase in-vitro transcription of cDNA synthesized by reverse transcription primed by oligo-dT attached to the t7 RNA polymerase promoter. The amplified cRNA was inspected for quality by agarose gel electrophoresis, and RNA concentrations were determined with a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). Amplified RNA was converted into biotinylated, fragmented cRNA using the IVT labeling kit (Affymetrix Inc., Santa Clara, CA). cRNA samples derived from single animals were hybridized in recommended buffers to microarrays (Affymetrix GeneChip^® Mouse Genome 430A 2.0 Array). The samples were stained and washed according to the manufacturer's protocol on a Fluidics Station 400 (Affymetrix Inc.) and scanned on a GeneArray Scanner (Affymetrix Inc., Santa Clara, CA). Primary data extraction was performed with Microarray Suite 5.0 (Affymetrix Inc., Santa Clara, CA). Data were filtered and sorted by a sequential analysis using the GeneSpring GX 7.3. software (Affymetrix Inc., Santa Clara, CA). Briefly, raw values were normalized using the GCRMA algorithm (13), filtered for 2-fold expressional changes, and subjected to statistical analysis using a non-parametric two-way analysis of variance. Genes with significant effects were selected by adjusting the resulting P values for multiple testing by means of the false discovery rate, using the linear step-up procedure (BH) of Benjamini and Hochberg (14).

Reverse transcriptase PCR (RT-PCR). Total RNA was diluted to a final concentration of 100 ng/μL and reverse-transcribed using the SuperScript™ First-Strand synthesis system for RT-PCR (Invitrogen Corporation, Carslbad, CA) following the supplier's manual. A total of 2 μL of the RT reaction was subjected to PCR amplification using the AccuPrime™ DNA polymerase system (Invitrogen Corporation, Carlsbad, CA) following the supplier's manual. The sequence of primer pairs for skg-1 (5′-CTGCAATGT GCCTTTTCTGA-3′ and 5′-ATGCTTCCCTCAA GCATCTG-3′), egr1 (5′-ATGGCAGCGGCCAAG GCC-3′ and 5′-GGGTACGGTTCTCCAGACCC T-3′), arc (5′-AGACACAGCAGATCCAGCTG-3′ and 5′-TGGCTTGTCTTCACCTTCAG-3′), cfos (5′-TACTACCATTCCCCAGCCG-3′ and 5′-TTGGC AATCTCGGTCTGCAA-3′) and GAPDH (5′-CTC ATGACCACAGTCCATGC-3′ and 3′-TTCAGCTC TGGGATGACCTT-5′) and respective amplification conditions were based upon published protocols (15–18). PCR products were separated by electrophoresis on a 1.5% agarose gel and stained with ethidium bromide. Band signals were analyzed and quantified by densitometry analysis using the Kodak Gel Logic 100 imaging system and software. Relative intensities were calculated by normalization to the band intensity levels of GAPDH.

ELISA

Frozen aliquots of plasma samples were thawed immediately before being used for ELISA experiments. ELISA analysis for plasma corticosterone levels was carried out using commercially available test kits (Assaypro, St Charles, MO, USA) following the manufacturer's instructions. All samples were analyzed in duplicates (n = 8–12 per group). Photometric measurements were performed on an ELISA plate-reader (Stat Fax-2100, Awareness Technologies Inc., Palm City, FL, USA). Calculation of results was obtained according to standard curves as provided in the test kits and analyzed together with the sample runs.

Statistical analysis

For statistical analyses of differences between two groups (GRPR-KO and WT) unpaired two-tailed Student's t tests were used. For experiments involving drug administration, two-way ANOVAs (2 (GRPR-KO or WT) × 2 (saline or dexamethasone)) followed by Scheffé's post-hoc tests for pairwise comparisons of significant ANOVA results were applied. An α-level of 0.05 was adopted in all instances. All analyses were carried out using BioStat 2009 professional software (AnalystSoft Inc., Alexandria, VA, USA).

Results

GRPR-KO mice present with increased depression-like behavior

To test the role of GRPR in depression-like behavior, GRPR-KO (n = 12) and WT (n = 10) litter-mate controls were subjected to the FST. Significantly increased time spent immobile in the FST was observed in GRPR-KO mice (P < 0.01) (Figure 1A). The sucrose preference test was then used to verify the depression-like behavioral phenotype in an independent paradigm. GRPR-KO animals were found to present with enhanced anhedonic behavior as displayed in a significant reduction in sucrose preference (P < 0.01) (Figure 1B).

Figure 1. Depression-like behavior in GRPR-KO mice. A: Immobility (percentage of total time) in the forced-swim test. B: Sucrose preference rate (percentage) in GRPR-KO and WT mice. All data are depicted as mean ± SEM. **P < 0.01.

Microarray analysis reveals reduced expression of glucocorticoid-responsive genes in the basolateral amygdala of GRPR-KO

To examine the molecular signature accompanying depression-like behavior in GRPR-KO mice, tissue of the basolateral nucleus of the amygdala was selectively collected using laser-capture microdissection (LCM) (Figure 2A). To obtain a representative gene expression profile and reduce variability induced by potential differences along the rostro-caudal extension of the amygdala or laterality effects, a total of three coronal sections of the amygdala from the rostral (Bregma −0.82 mm), the medial (Bregma −1.34 mm), and the caudal (Bregma −1.82 mm) amygdala were used to dissect the left and the right basolateral amygdala tissue by the means of LCM. Total RNA was isolated, amplified and used for oligonucleotide micorarray analysis in order to determine differences in amygdala gene expression between GRPR-KO and WT mice. Of 14,000 well characterized mouse genes represented on the Affymetrix GeneChip^® Mouse Genome 430A 2.0 Array, we found 84 probe sets to be significantly and at least 2-fold differentially expressed between GRPR-KO and WT mice (Table I). Remarkably, six of the top ten most down-regulated genes are known as glucocorticoid-responsive genes (Table II). Serum/ glucocorticoid regulated kinase (sgk-1), early growth response 1 (egr1), activity regulated cytoskeletal-associated protein (arc), and FBJ osteosarcoma oncogene (cfos) were selected for independent verification by semi-quantitative RT-PCR. Down-regulation of these transcripts in GRPR-KO mice was also confirmed in RT-PCR experiments using the same RNA isolated from LCM samples also applied in the microarray analysis (Figure 2B–E). To characterize the corticosterone profile of GRPR-KO, levels of circulating corticosterone were determined by ELISA analysis from plasma samples, taken around 5 p.m., shortly before the onset of the dark phase where the maximum rise in plasma corticosterone concentration can be expected (19). Corticosterone levels were significantly lower in GRPR-KO compared to WT mice (P < 0.01) (Figure 2F).

Table I. Genes significantly differentially expressed in the basolateral amygdala of GRPR-KO compared to WT mice.

Probe set ID	Name	Genbank ID	Fold-change	P value
1450260_at	Gastrin releasing peptide receptor	M57922	−6.5280	5.70 × 10^−05
1416041_at	Serum/glucocorticoid regulated kinase (sgk-1)	NM_011361	−5.1400	7.38 × 10^−05
1430195_at	RIKEN cDNA 2810043O03 gene	BF662697	−3.9220	0.0114
1417065_at	Early growth response 1	NM_007913	−3.8420	0.00196
1418687_at	Activity regulated cytoskeletal-associated protein	NM_018790	−3.7510	0.000103
1431182_at	Heat shock protein 8; similar to heat shock protein 8; similar to Heat shock cognate 71 kDa protein	AK004608	−3.6000	0.00348
1424671_at	Pleckstrin homology domain containing, family F (with FYVE domain) member 1	BC002120	−3.1680	0.000817
1423100_at	FBJ osteosarcoma oncogene	AV026617	−3.0420	0.0125
1445689_at	RNA binding motif, single stranded interacting protein 1, mRNA (cDNA clone MGC:25244 IMAGE:4527742)	AA717264	−3.0040	0.0068
1417065_at	Early growth response 1	NM_007913	−2.9090	0.0121
1437917_at	RIKEN cDNA D530037H12 gene	BG066910	−2.8680	0.00788
1425270_at	Kinesin family member 1B	BE199508	−2.8220	0.0239
1427371_at	ATP-binding cassette, subfamily A (ABC1), member 8a	BC026496	−2.8040	0.0259
1417765_a_at	Amylase 1, salivary	NM_007446	−2.7700	0.0158
1428306_at	DNA-damage-inducible transcript 4	AK017926	−2.7190	0.00434
1431725_at	Formin 2	AK013585	−2.6920	0.00761
1448890_at	Kruppel-like factor 2 (lung)	NM_008452	−2.6640	0.0131
1448123_s_at	Transforming growth factor, beta induced	NM_009369	−2.6330	0.0241
1452125_at	Thyroid hormone receptor associated protein 3	BG075035	−2.6160	8.55 × 10^−05
1448666_s_at	Transducer of ERBB2, 2	AV174616	−2.4870	0.0496
1436953_at	Wiskott-Aldrich syndrome protein interacting protein	C76969	−2.4690	0.0342
1426721_s_at	TCDD-inducible poly(ADP-ribose) polymerase	BB707122	−2.4200	0.0209
1427703_at	Platelet-activating factor acetylhydrolase, isoform 1b, beta1 subunit	L25109	−2.4160	0.0286
1422673_at	Protein kinase C, mu	NM_008858	−2.4150	0.035
1419055_a_at	Protein tyrosine phosphatase, non-receptor type 21	AW987375	−2.3970	0.00122
1455831_at	Fusion, derived from t(12;16) malignant liposarcoma (human)	BE985138	−2.3630	0.000904
1451527_at	Procollagen C-endopeptidase enhancer 2	AF352788	−2.3610	0.0117
1453147_at	Polymerase (RNA) III (DNA directed) polypeptide E	AV251549	−2.3570	0.00814
1426802_at	Septin 8	BB756379	−2.3270	0.0467
1418250_at	ADP-ribosylation factor 4-like	NM_025404	−2.3230	0.00111
1416505_at	Nuclear receptor subfamily 4, group A, member 1	NM_010444	−2.3150	0.000517
1439111_at	TSC22 domain family, member 1	BM230984	−2.2720	0.0143
1423233_at	CCAAT/enhancer binding protein (C/EBP), delta	BB831146	−2.2590	0.000548
1427809_at	Latrophilin 3, mRNA (cDNA clone IMAGE:30549106)	BC010480	−2.2390	0.00535
1418201_at	Pleckstrin homology domain containing, family G (with RhoGef domain) member 2	NM_138752	−2.2390	0.0382
1455812_x_at	Slit-like 2 (Drosophila)	BB530515	−2.2350	0.0156
1420088_at	Nfkbia	AI462015	−2.2060	0.000313
1434292_at	RIKEN cDNA E130013N09 gene	BI731047	−2.1740	0.0049
1416266_at	Prodynorphin	AF026537	−2.1660	0.0462
1419649_s_at	Myosin IC	NM_008659	−2.1650	0.0394
1434898_at	Trinucleotide repeat containing 6a	BI080625	−2.1090	0.0115
1448306_at	Nuclear factor of kappa light chain gene enhancer in B-cells inhibitor, alpha	NM_010907	−2.0960	0.000831
1456430_at	DNA segment, Chr 3, ERATO Doi 789, expressed	BB410567	−2.0890	0.00119
1426124_a_at	CDC-like kinase 1	U21209	−2.0810	0.00635
1448538_a_at	DNA segment, Chr 4, Wayne State University 53, expressed	BE447520	−2.0810	0.00967
1439506_at	Gene model 98, (NCBI); gene model 1804, (NCBI)	BB193557	−2.0790	0.0321
1427580_a_at	RNA imprinted and accumulated in nucleus	BB649603	−2.0650	0.00986
1416543_at	Nuclear factor, erythroid derived 2, like 2	NM_010902	−2.0520	0.046
1453864_at	Retinol dehydrogenase 14 (all-trans and 9-cis)	AK014097	−2.0450	0.0137
1418135_at	AF4/FMR2 family, member 1	NM_133919	−2.0410	0.0307
1436898_at	Splicing factor proline/glutamine rich (polypyrimidine tract binding protein associated)	BI738328	−2.0380	0.0233
1420048_at	Adult male hypothalamus cDNA, RIKEN full-length enriched library, clone: A230020L22 product: unclassifiable, full insert sequence	C78859	−2.0360	0.0105
1451577_at	Zinc finger and BTB domain containing 20	AW491109	−2.0300	0.0321
1430575_a_at	Tripeptidyl peptidase II	BB484264	−2.0270	0.0421
1425371_at	Polymerase (DNA directed), beta	BC006681	−2.0240	0.0392
1425281_a_at	TSC22 domain family 3	AF201289	−2.0240	0.0499
1419459_a_at	RIKEN cDNA 2610529C04 gene	NM_025952	−2.0230	0.0233
1459884_at	Cytochrome c oxidase, subunit VIIc	AA190297	−2.0180	0.014
1449731_s_at	Nfkbia	AI462015	−2.0170	0.00151
1437180_at	RIKEN cDNA 6530403A03 gene	BE824567	−2.0120	0.0223
1449851_at	Period homolog 1 (Drosophila)	AF022992	−2.0100	0.000983
1452376_at	Zinc finger protein 444	BE951951	2.0450	0.038
1427430_at	Expressed sequence AI848100	BB148987	2.0492	0.0484
1423495_at	2-4-dienoyl-Coenzyme A reductase 2, peroxisomal	BE952632	2.0534	0.0354
1427831_s_at	Zinc finger protein 260	L36316	2.0661	0.0166
1423856_at	Popeye domain containing 3	BC003896	2.0877	0.0379
1454622_at	Solute carrier family 38, member 5	BG066984	2.1097	0.0157
1420989_at	RIKEN cDNA 4933411K20 gene	NM_025747	2.1097	0.0386
1426166_at	Major urinary protein 5	M16360	2.1186	0.0159
1422145_at	Mannoside acetylglucosaminyltransferase 3	NM_010795	2.1231	0.0321
1421933_at	Chromobox homolog 5 (Drosophila HP1a)	NM_007626	2.1645	0.0195
1420401_a_at	Receptor (calcitonin) activity modifying protein 3	NM_019511	2.2422	0.00154
1418271_at	Basic helix-loop-helix domain containing, class B5	NM_021560	2.2472	0.0487
1419748_at	ATP-binding cassette, subfamily D (ALD), member 2	NM_011994	2.2779	0.00145
1426607_at	CDNA clone MGC:118117, IMAGE:6309338	BG068672	2.2831	0.00706
1450385_at	Karyopherin (importin) alpha 3	BM213828	2.3364	0.0261
1426152_a_at	Kit ligand	M64262	2.3810	0.0175
1423506_a_at	Neuronatin	AV218841	2.4814	0.0306
1415904_at	Lipoprotein lipase	BC003305	2.5575	0.0183
1416132_at	RIKEN cDNA C920006C10 gene, EFR3 homolog A	BB188557	2.6738	0.0274
1417975_at	Karyopherin (importin) alpha 4	BF018653	2.9674	0.0258
1424525_at	Gastrin releasing peptide	BC024515	4.5045	0.0248

Table II. Glucocorticoid-responsive genes significantly differentially expressed in the basolateral amygdala of GRPR-KO compared to WT mice.

Name	Fold-change	P value	Selected references
serum/glucocorticoid regulated kinase (sgk-1)	−5.14	0.0000738	Sato H, Horikawa Y, Iizuka K, Sakurai N, Tanaka T, Shihara N, et al. 2008; Buse P, Maiyar AC, Failor KL, Tran S, Leong ML, Firestone GL. 2007
early growth response 1 (egr1)	−3.842	0.00196	Revest JM, Di Blasi F, Kitchener P, Rouge-Pont F, Desmedt A, Turiault M, et al. 2005
activity regulated cytoskeletal-associated protein (arc)	−3.751	0.000103	McReynolds JR, Donowho K, Abdi A, McGaugh JL, Roozendaal B, McIntyre CK. 2010
pleckstrin homology domain containing, family F (with FYVE domain) member 1	−3.168	0.000817	Sato H, Horikawa Y, Iizuka K, Sakurai N, Tanaka T, Shihara N, et al. 2008;
FBJ osteosarcoma oncogene (cfos)	−3.042	0.0125	Shur I, Socher R, Benayahu D. 2005
nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha (Nfkbia)	−2.206	0.000313	Sato H, Horikawa Y, Iizuka K, Sakurai N, Tanaka T, Shihara N, et al. 2008;
period homolog 1 (Drosophila)	−2.01	0.000983	Conway-Campbell BL, Sarabdjitsingh RA, McKenna MA et al. 2010

Figure 2. Isolation of tissue from the basolateral amygdala for ensuing gene expression analysis by laser-capture microdissection. A: Tissue from the basolateral amygdala of a thin brain section is melted to a cap covered with thermoplastic film after laser firing onto the target region. After removing the cap from the brain section the targeted area is lifted and stays attached to the thermoplastic film from where it can be extracted and used for subsequent molecular analysis. mRNA expression of sgk-1 (B), egr1 (C), arc (D), and cfos (E) in basolateral amygdala samples of GRPR-KO and WT mice (n = 5 per group) evaluated by RT-PCR. Values normalized to the expression of GAPDH are represented in arbitrary values. F: Plasma corticosterone levels in GRPR-KO and WT mice determined by ELISA analysis. All data are depicted as mean ± SEM. **P < 0.01; ***P < 0.0001.

Acute administration of dexamethasone reduces depression-like behavior and restores amygdala expression of glucocorticoid-responsive genes in GRPR-KO mice

To test the hypothesis that deficient stimulation by glucocorticoids may account for the behavioral and molecular phenotype in GRPR-KO mice, we decided to evaluate whether treatment with dexamethasone (DEX) could ameliorate depression-like behavior and rescue aberrant expression of glucocorticoid-responsive genes. Acute administration of DEX 2 h prior to behavioral testing in the FST reduced the time spent immobile in GRPR-KO mice (n = 9–11 per group) to values comparable to WT controls (n = 10–11 per group) (significant main effect of genotype F_(1,40) = 7.91: P < 0.01; significant main effect of drug treatment F_(1,40) = 8.95: P < 0.01; significant effect of interaction between genotype and drug treatment F_(1,40) = 10.14: P < 0.01). Post-hoc analysis revealed no effect of DEX treatment on WT mice (P > 0.05) (Figure 3A). Animals were examined in the SPT 72 h after the FST, and DEX was administrated 2 h prior to behavioral testing. Diminished sucrose preference in GRPR-KO mice was restored to control levels after DEX treatment (significant main effect of drug treatment F_(1,40) = 9.05: P < 0.01; significant effect of interaction between genotype and drug treatment F_(1,40) = 9.98: P < 0.01). Post-hoc analysis demonstrated that DEX treatment did not affect sucrose preference in WT mice (P > 0.05) (Figure 3B). After a 72-h recovery period, mice were injected with DEX and 2 h later sacrificed and brains were dissected. Basolateral amygdala (BLA) tissue was harvested by means of LCM and RNA was extracted, amplified, and used for RT-PCR analysis (n = 5 per group). Levels of sgk-1 (main effect of genotype F_(1,20) = 2.65: P > 0.05; main effect drug treatment F_(1,20) = 28.74: P < 0.0001; effect of interaction F_(1,20) = 21.086: P < 0.0001)), egr1 (main effect of genotype F_(1,20) = 1.25: P > 0.05; main effect drug of treatment F_(1,20) = 8.72: P < 0.01; effect of interaction F_(1,20) = 10.72: P < 0.01)), arc (main effect of genotype F_(1,20) = 4.77: P < 0.05; main effect drug treatment F_(1,20) = 3.85: P > 0.05; effect of interaction F_(1,20) = 21.086: P < 0.01)), and cfos (main effect of genotype F_(1,20) = 2.65: P > 0.05; main effect drug treatment F_(1,20) = 45.42: P < 0.0001; effect of interaction F_(1,20) = 43.04: P < 0.0001)) in GRPR-KO mice were restored to control levels after DEX treatment. No effect of DEX treatment on the expressional levels of sgk-1, egr1, arc, and cfos in WT mice was observed (P > 0.05) (Figure 4A–D).

Figure 3. Application of dexamethasone reverses depression-like behavior in GRPR-KO mice. A: Immobility (percentage of total time) in the forced-swim test. B: Sucrose preference rate (percentage) in GRPR-KO and WT mice after dexamethasone (DEX) and saline control (sal) treatment. Significance levels resulting from post-hoc comparisons are indicated. All data are depicted as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 4. Application of dexamethasone reverses gene expression changes in GRPR-KO mice. mRNA expression of sgk-1 (A), egr1 (B), arc (C), and cfos (D) in basolateral amygdala samples of GRPR-KO and WT mice after dexamethasone (DEX) and saline control (sal) treatment (n = 5 per group) evaluated by RT-PCR. Values normalized to the expression of GAPDH are represented in arbitrary values. Significance levels resulting from post-hoc comparisons are indicated. All data are depicted as mean ± SEM. **P < 0.01; ***P < 0.0001.

Discussion

GRPR-KO mice present with increased depression-like behavior

Our results, demonstrating a depression-like behavioral phenotype in GRPR-KO mice are noteworthy and relevant in several aspects: First, GRPR, which has previously been proposed as candidate molecule involved in the pathogenesis of some psychiatric and neurological disorders (7), should now also be considered to be of relevance for the neurobiological disturbances underlying depressive disorders and as potential therapeutic target for the treatment of these diseases. Second, the observation that disruption of GRP-signaling induces both a depression-like behavioral phenotype and enhanced and persistent fear memory (4) proposes this mouse model as suitable tool for studying the biological basis underlying the co-morbidity and overlapping phenomenology among anxiety disorders and depression. Third, depression-like behavior resulting from deletion of GRPR, a protein highly expressed in the SCN and importantly involved in the photic entrainment/ phase-resetting pathway of the autonomous oscillating system (20,21), supports a dual role of this molecule, as modulator of the circadian rhythm and as regulator of mood states. Examining the neurobiological mechanisms mediating depression-like behavior in GRPR-KO mice may also constitute a valuable resource for investigating the so far poorly understood molecular pathways mediating the functional involvement of the circadian system in the regulation of emotional states.

Microarray analysis reveals reduced expression of glucocorticoid-responsive genes in the basolateral amygdala of GRPR-KO

To employ an unbiased method for investigating the molecular foundations underlying the depression-like behavioral phenotype observed in GRPR-KO we turned towards microarray analysis of basolateral amygdala (BLA) tissue, a target brain region in depression (see for review (22)) and an area where functional alterations in GRPR-KO have been previously found (4). LCM, a method earlier successfully employed to dissect mouse basolateral amygdala without contamination from surrounding tissue (12), was used to precisely isolate cells belonging to the basolateral amygdala and to apply RNA isolated from these samples to large-scale gene expression analysis. After stringent statistical analysis, genes significantly differentially expressed between GRPR-KO and WT samples were examined in a hypothesis-free approach. Particularly striking was the observation that six among the top ten genes most strongly down-regulated in GRPR-KO with respect to WT have been previously described as glucocorticoid-responsive genes. This pattern of gene expression changes suggested a common underlying regulatory mechanism potentially of relevance for the observed behavioral phenotype. The role of glucocorticoid-signaling in mediating the observed behavioral phenotype and in the modulation of the expression of four of these genes was therefore further studied.

sgk-1. With more than 5-fold reduced expression in GRPR-KO, sgk-1 is rapidly and strongly stimulated by glucocorticoids as a result of the functional glucocorticoid response element in the promoter (23). In the brain, sgk-1 has been identified as glucocorticoid-responsive gene e.g. in the rat hypothalamus (24). sgk-1 directly activates cyclic-AMP response element binding (CREB) (25) which may constitute the downstream cascade conferring the elaboration of dendritic branching and dendritic remodeling resulting from enhanced sgk-1 activity (26). Interestingly, alterations in dendritic morphology together with a reduction in the number of dendritic spines has been repeatedly found as a morphological correlate of depression, both in post-mortem samples from human patients and in experimental animals. Moreover, CREB regulates the expression of several genes potentially relevant for the pathophysiology of depression (27), and antidepressant treatment has been found to augment brain CREB levels (28–30).

egr1. The expression of egr1 has been found to be regulated by glucocorticoids in a mitogen-activated protein kinase (MAPK)-independent and a MAPK-dependent mechanism (31). The transcription factor egr1, which was more than 3-fold down-regulated in GRPR-KO mouse brain, is furthermore inducible by growth factors, release of specific neurotransmitters and depolarization, strongly suggestive of a plasticity-linked function of egr1 and its target genes (see for review (32)). Dysfunctional synaptic plasticity-related events resulting in morphological changes, including impaired growth, maturation, and survival of neurons, are one of the structural hall-marks observed in brains of depressed patients and in animal models of depression (see for review (33)). Moreover, egr1 expression has been found to be up-regulated in response to chronic antidepressant treatment (34), further supporting a role for this gene in the pathophysiology of depressive disorders and the neurobiological mechanisms mediating the response to antidepressant therapy.

arc. Known as an integrative modulator of synaptic plasticity (see for review (35)), arc has been found to be increased after corticosterone treatment in hippocampal synaptic tissue (36). We observed a more than 3-fold decreased arc expression in GRPR-KO mice compared to WT. Modulation of arc expression has also been observed in several animal models of depression (37,38), including glucocorticoid receptor-deficient mice, a genetic model of predisposition to depression (39) and resulting from antidepressant treatment (40,41). Therefore, a potential causal involvement of arc in the plasticity-related morphological neuronal alterations characteristic of depressive disorders can be hypothesized.

cfos. cfos is a member of the AP-1 transcription complex whose mRNA expression is regulated by glucocorticoids potentially through modulation of cfos splicing. cfos itself, on the other hand, transduces the response to dexamethasone (42) and regulates the expression of glucocorticoid receptors (43), suggesting a feedback loop mechanism which balances the transcriptional fine-tuning among the two systems. Similarly to egr1, cfos expression is induced by a variety of stimuli, and a role for cfos activation in plasticity-related processes has been suggested (see for review (44)). Thus, 3-fold reduced expression of cfos together with the observed decrease in egr1 and arc expression may point towards deficiencies of the molecular neuroplastic machinery in GRPR-KO, potentially resulting from aberrant glucocorticoid activation and possibly contributing to the observed depression-like behavioral phenotype.

More than 2-fold decreased peak plasma corticosterone levels in GRPR-KO provide first evidence for an effect of aberrant GRPR-signaling on circulating corticosterone. However, an extended analysis including data at trough, during stress and in a stress recovery period would be needed in order to provide a complete characterization of the plasma corticosterone profile of GRPR-KO and to definitively relate it to the behavioral phenotype.

Acute administration of dexamethasone reduces depression-like behavior and restores amygdala expression of glucocorticoid-responsive genes in GRPR-KO mice

Prolonged exposure to stress, resulting in an enhanced production of cortisol is known as predisposing factor for the development of stress-related mental disorders (45). Nevertheless, increasing lines of evidence support that—in addition to glucocorticoid excess—insufficient glucocorticoid signaling may play a significant role in the development and expression of neuropsychiatric disorders, including depression (45). Due to the aberrant expression of glucocorticoid-responsive genes in BLA tissue of GRPR-KO, we decided to evaluate a potential causal involvement of deficient glucocorticoid signaling in depression-like behavior and amygdala gene expression in GRPR-KO mice by testing the effects of acute DEX administration. Although no effect on WT mice was observed, DEX normalized depression-like behavior and deviant expression of glucocorticoid-responsive genes in GRPR-KO, suggesting a significant role for deficient glucocorticoid activity in the observed behavioral and molecular phenotype. This hypothesis is further supported by reports providing evidence that GRP, through binding to GRPR, induces release of corticosterone through stimulation of the secretion of adrenocorticotropin (ACTH) (46) and potentially also ACTH-independently, through the sympathetic adrenal nerve route, since bombesin-related peptides have been found to act on the autonomous nervous system (47,48). Interestingly, also light-induced corticosterone release is conferred through the SCN-adrenal sympathetic nervous pathway (49). It can be speculated that GRPR may mediate this effect since GRP, expressed in the ventral portion of the SCN, has been found to transmit photic information to the dorsal portion of the SCN where the core oscillating system resides (20). Hereby, GRP influences light-induced modulation of the autonomous circadian rhythmicity and corticosterone release, which, like many other physiological functions, is regulated through an endogenous rhythm that can be adjusted by external stimuli, mainly light. Evidence for an involvement of GRPR in glucocorticoid signaling has previously been suggested, since DEX administration reversed the effect of a pharmacological inhibitor of GRPR on memory in mice (11). Additionally the effect of DEX on cell proliferation has been shown to be inhibited by administration of a GRPR antagonist (22).

In summary, we propose a model in which inadequate glucocorticoid signaling, resulting from the deletion of GRPR and potentially also leading to insufficient light-induced secretion of adrenal corticosterone, may result in reduced expression of glucocorticoid-responsive genes in the basolateral amygdala and contribute to the development of a depression-like behavioral phenotype (Figure 5).

Figure 5. Model of biological mechanisms potentially underlying depression-like behavior and amygdala gene expression changes in GRPR-KO. Lack of GRPR (red triangles) in the suprachiasmatic nucleus (SCN) may lead to deficient stimulation of corticosterone release from the adrenal gland, possibly via reduced release of corticotropin-releasing factor (CRF) and adrenocorticotropin (ACTH) and potentially also through insufficient signaling via the sympathetic adrenergic route, also responsible for mediating the effects of light on corticosterone secretion. Insufficient corticosterone stimulation of the amygdala leads to a reduction in the expression of glucocorticoid-sensitive genes potentially inducing behavioral changes, such as depression-like behavior.

Acknowledgements

We thank Rae Silver, Department of Psychology, Columbia University, for enabling the use of the laser-capture microdissection facility and Yonghui Zhang, Institute for Cancer Genetics, Columbia University, for bioinformatic assistance with the analysis of microarray experiments.

Declaration of interest: Francisco J. Monje received a grant from the Hochschuljubiläumsstiftung der Stadt Wien. Daniela D. Pollak is supported by the Austrian Science Fund. The authors declare no other conflicts of interest.