| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* Institute of Experimental Endocrinology, Charité School of Medicine, Humboldt University, Berlin, Germany;
Neuroadaptations Group, Max Planck Institute of Psychiatry, Munich, Germany;
Corporate Research Gynecology and Andrology, Schering AG, Jena, Germany;
Section on Functional Neuroanatomy, National Institute of Mental Health, National Institutes of Health, Bethesda, Maryland, USA;
|| Institute de Génétique et de Biologie Moléculaire et Céllulaire (IGBMC), Illkirch, France
1Correspondence: Institute of Experimental Endocrinology, Charité School of Medicine, Humboldt University, Schumannstr. 20/21, Berlin 10117, Germany. E-mail: alexandre.patchev{at}charite.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: stress • adrenal cortex • steroidogenesis • hypothalamus • hippocampus
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Curtailment of the endocrine response to stress is executed by glucocorticoid receptor (GR)-mediated negative feedback mechanisms in the hippocampus, hypothalamus, and pituitary, manifested as reduced synthesis and secretion of pituitary ACTH and its hypothalamic secretagogues, corticotropin-releasing hormone (CRH), and vasopressin (AVP) (1
2
3)
.
GR signaling efficacy is determined by a variety of coactivators, including the transcription intermediary factor 2 (TIF-2/SRC-2/NCoA2/GRIP1). TIF-2 belongs to the p160 family of cellular proteins (e.g., SRC-1 and SRC-3), which contain a highly conserved motif (the so-called NR-box) that facilitates interaction with the ligand-dependent activation function (AF-2) of nuclear receptors and amplification of target gene transcription (4)
. Interactions between TIF-2 (5)
, and its mouse homologue GRIP1 (6)
, have been demonstrated for several nuclear receptors (7)
.
In the mouse, TIF-2 mRNA and protein are expressed ubiquitously (5
, 8
, 9)
, but reports differ regarding its presence in the rat brain (10
, 11)
. Targeted ablation of TIF-2 in the mouse is phenotypically marked by impaired gonadal function (12)
and altered energy balance and lipid metabolism (13)
. Since no other overt functional aberrations have been so far noted in TIF-2–/– mice, compensatory overexpression of the related coactivator SRC-1, implying partial functional redundancy of TIF-2 and SRC-1, has been presumed (14
15
16)
.
Given the role of TIF-2 as a GR coactivator, we tested the hypothesis that disruption of TIF-2 would result in attenuated GR signaling in the brain, measurable as alterations in the secretory dynamics of the HPA axis under basal (resting) conditions and following exposure to stressful stimuli. We expected that TIF-2 deficiency would become manifest as a "weakening of the grip" of GR-mediated restraint of the HPA axis, resulting in pituitary-adrenal hyperactivity under quiescent conditions, exaggerated and prolonged secretory response to stress, and increased expression of neuropeptides that trigger HPA activation, e.g., CRH and AVP. As reported here, our starting hypothesis was supported in several respects; however one unexpected but important finding was that TIF-2–/– mice demonstrate symptoms of adrenal insufficiency. The latter result suggests that impaired HPA axis regulation in TIF-2 deficient mice may result from insidious adrenocortical hypofunction rather than failure to amplify GR signaling.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Polymerase chain reaction (PCR) was done using the primers P1 [CTG CAC GGT GCC automatic gain control (AGC) AAA GC), P2 (GAC CAG GGC TTG CTC AGA AC) and P3 (CCC CTG GAT TGT TCC AAA GG]. The WT and –/– alleles resulted in products of 512 and 298 bp, respectively. Animals were housed under standard conditions (24°C; 60–70% relative humidity, 12h light/dark with lights on at 06:00; ad libitum access to chow pellets and water). All procedures complied with NIH guidelines on laboratory animal use and were approved by the local ethical board.
Stress protocol
Peak and nadir basal CORT levels were determined in tail vein blood (collected within 30 s of removal from home cage) at 18:00 and 06:00 h.
The first experiment, which examined the endocrine response to short-term acute stress, started immediately after the morning blood collection. Stress procedures encompassed transfer to the procedure room, placement into a glass container (diameter 15 cm), and exposure to low-frequency vibration (ca. 25 Hz, delivered by a Vortex) and loud music (80–90 dB) for 1 min, before returning to the home cage. Blood samples were taken to measure stress-induced (30 min) and recovery-phase (120 min) levels of CORT. One week later, animals were randomly assigned to subgroups for tissue collection. Between 08:00 and 11:00 h one subgroup was killed immediately on removal from the home cage, and the other 30 min after exposure to the above-described stress procedure. Tissue specimens were either snap-frozen or immersion-fixed in PBS/p-formaldehyde for subsequent analysis.
A different set of animals from the third cross-bred generation was used for examination of responses to prolonged physical stress. The mice were immobilized in plastic restrainers for 4 h between 08:00 and 12:00 h. Blood from tail vein incisions was collected for CORT determination 15, 120, and 240 min after the beginning of immobilization.
Hormone assays
Serum CORT was measured by RIA (MP Biomedicals, Orangeburg, NY, USA). Pituitary ACTH contents were determined by RIA (DSL, Webster, TX, USA) in 0.1 N acetic acid extracts, and results were normalized to tissue protein levels; the antibody (Ab) had 100% cross-reaction for the species-invariable, bioactive amino acid fragment ACTH (1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24)
. Coefficients of variation of the CORT and ACTH assays were 12 and 10%, respectively.
Nucleic acid probes and primers
AVP transcripts were detected with a 48-mer antisense probe (17)
complementary to bases 1493–1540 of the murine vasopressin gene (M_88345). The 40-mer probe for 3ß-HSD mRNA corresponded to bases 531–570 of the murine gene sequence (M_77015), with one mismatch to a rat-specific probe (18)
. Oligonucleotides were labeled with [35S]-dATP (NEN, Boston, MA, USA) using terminal deoxynucleotidyl transferase (TdT) (Invitrogen, Karlsruhe, Germany). Labeled ribonucleotide probes for the detection of CRH-, GR-, SRC-1, and StAR protein-encoding transcripts were produced from linearized plasmids using in vitro-transcription kits with T7, T3, and Sp6 RNA polymerases (Promega, Madison, WI, USA) and [35S]-dUTP. The CRH cDNA template was a 760 bp fragment comprising bases 260–1020 of the mouse CRH gene (AY_128673). The GR template included bases 81–528 of the rat GR mRNA sequence (M_14053). A 622 bp fragment corresponding to bases 1121–1743 of the human SRC-1 (U_90661) was used for SRC-1 mRNA detection. The probe for StAR protein transcripts was amplified by PCR from the clone IRAKp961L18177Q (RZPD, Berlin, Germany) using the primers 5'-ATGTTCCTCGCTACGTTCAAG-3' and 5'-CCAAGCGAAACACCTTGCC-3'. The resulting fragment of 364 bp, corresponding to bases 57–420 of the murine StAR protein sequence (NM_011485), was cloned into pTOPO and, on sequence verification and linearization with HindIII and XbaI, used for the generation of antisense and sense probes, respectively. The plasmid encoding 246 bp (bases 1914–2160) of the 3'UTR of the mouse SF-1 gene (NM_139051) was linearized with HindIII and EcoRI, for transcription of antisense and sense probe, respectively.
In situ hybridization histochemistry
Transcripts encoding AVP and CRH were quantified in the rostral extension of the PVN (bregma –0.75 to –0.85); those encoding GR were measured in the dorsal hippocampus (bregma –1.70 to –1.90) and the PVN (19)
. Expression of SRC-1 mRNA was examined in the hippocampus. Coronal 10-µm-thick brain cryosections were hybridized with oligo- and ribonucleotide probes according to established protocols (20
, 21)
. Autoradiographs were generated by film exposure (BioMax MR; Kodak, Rochester, NY, USA).
Semiquantitative densitometric analysis of brain autoradiograhs was performed using the NIH software Scion Image Beta 4.2.0. AVP transcript signal intensity was measured in fixed-size squares within the ventromedial compartment of the PVN, centered over its parvocellular division. CRH transcript densities were evaluated within the entire PVN. GR hybridization signals were measured in the pyramidal cell layer of the CA1–2 fields of the hippocampus and in the PVN. SRC-1 mRNA transcript intensities were compared in the granule cell layer of dentate gyrus (22)
.
Bilateral optical density (OD) measurements were made in two sections per animal. Individual averaged transmittance levels were converted to specific radioactivity by third-order polynomial equations generated from coexposed 14C standards (ARC; St. Louis, MO, USA).
Detection of transcripts encoding SF-1, StAR protein, and 3ß-HSD in the adrenal cortex was performed in 5 µm sections through the equatorial plane of paraffin-embedded glands. Following deparaffinization (xylene), sections were permeabilized, delipidated, hybridized, and stringently washed (40
, 41)
. Slides were exposed to photographic emulsion (NTB; Kodak, Rochester, NY, USA) for 3 d and lightly counterstained with cresyl violet.
The software package Image Pro Plus (Media Cybrnetics, Silver Spring, MD, USA) was used for the quantification of transcripts in adrenal sections. In each section, two rectangular frames of preset size were placed within the microscopic image of zona fasciculata at a magnification of x400. Identical illumination conditions and size thresholds for inclusion/exclusion of silver grains were applied to all sections. The number of signal-positive particles was subsequently related to the cell counts measured in each frame. Four measurements were averaged to produce individual means.
Reverse-transcriptase polymerase chain reaction (RT-PCR) detection of DAX-1
RNA was isolated using the RNeasy Plus Mini Kit (Qiagen, Hilden, Germany); integrity was confirmed with the Agilent System (Agilent Technologies, Palo Alto, CA, USA), according to the manufacturer’s protocol. From each individual extract, 500 ng were reverse-transcribed into cDNA using the Superscript III Platinum Kit (Invitrogen, Karlsruhe, Germany). For real-time PCR, cDNA (50 ng) was amplified using the TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA, USA) according to the protocol of the supplier. The oligonucleotide probe sets for mouse DAX-1 (Mm 00431729_m1) and ß-actin (Mm 00607939_s1) were purchased from Applied Biosystems. Normalization to ß-actin and calculation of relative DAX-1 expression was done by the 
Ct method (Applied Biosystems).
Adrenocortical morphometry
Cellular densities in the adrenocortical zona fasciculata were determined in hematoxylin-eosin-stained paraffin sections through the equatorial plane. Cell nuclei were counted using the Image Pro Plus software within fixed-size rectangular frames of 120 x 120 µm, positioned in two diametrically opposed sites of zona fasciculata. The threshold discrimination tool was uniformly applied to ensure recognition of only nuclear staining as positive signal. For each animal, counts from four frames placed at the opposite poles two sections were averaged.
Statistics
Numerical data (shown as mean ± SEM) were analyzed by 1-way ANOVA and Tukey’s test, where appropriate. Key ANOVA parameters (F-values, degrees of freedom) are given in the description of results. The level of significance was preset at P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
TIF-2 deletion is associated with increased neural drive on the HPA axis
Expression levels of hypothalamic CRH and AVP, and pituitary ACTH content in TIF-2 deficient animals indicated hyperactivation of the hypothalamo-pituitary unit. Under basal (stress-free) conditions, the levels of the mRNAs encoding CRH and AVP in the hypothalamic paraventricular nucleus (PVN) were significantly elevated in TIF-2–/– mice, as compared to WT and TIF-2+/– (Fig. 1
; F2,10=12.19 and 10.66; P=0.004 and 0.003, respectively). CRH and AVP mRNA transcript numbers were rapidly up-regulated by stress in WT and TIF-2+/–, but not TIF-2–/–, mice (Fig. 1
; F2,11=5.71 and 4.48; P=0.02 and 0.035, respectively). Pituitary ACTH content in TIF-2–/– mice was significantly higher than that measured in WT and TIF-2+/– animals (Fig. 2
; F2,10=8.55; P=0.007). Stress exposure failed to significantly change pituitary ACTH content in any genotype.
|
|
Genotype-associated changes in the expression of GR in the brain complemented the symptoms of HPA hyperactivity. Significantly decreased GR mRNA levels were observed in the CA1–2 subfield of the hippocampus of homo- and heterozygous TIF-2 deficient mice (Fig. 3
; F2,8=58.47; P<0.001), and a similar, albeit nonsignificant, trend was seen in the hypothalamic PVN. Expression of SRC-1 mRNA was seen throughout the hippocampal formation, without any significant genotype-related differences (Fig. 3)
.
|
Impaired adrenocortical output in TIF2–/– mice
Surprisingly, basal (morning) serum corticosterone (CORT) levels in TIF-2–/– mice were significantly lower than those in TIF-2+/– and WT animals (Fig. 4
; F2,22=5.82, P=0.01). At the diurnal zenith of adrenocortical activity CORT secretion was similar in all groups (140±22, 139±21 and 137±20 ng/ml). Exposure to a brief emotional stress significantly increased CORT levels over baseline in all genotypes; however, the secretory response measured 30 min post-stress was significantly attenuated in TIF-2–/– animals, as compared to that in TIF-2+/– and WT mice (Fig. 4
; F2,22=17.35; P<0.001). Although similar trends were observed in CORT levels measured after 120 min, they were not statistically significant.
|
An independent experiment corroborated the observation of decreased adrenocortical output in TIF-2-deficiency during the initial phase of the stress response. CORT levels, measured during protracted immobilization stress for 4 h, were significantly lower in TIF-2–/– and heterozygous animals at 15 min (Fig. 4
; F2,12=8.90; P<0.004), but these differences between genotypes became less pronounced with increasing stress duration. As before, resting (prestress) CORT levels were significantly lower in TIF-2 deficient mice (F2,12=4.35; P<0.04).
Cellular basis for diminished adrenocortical activity in TIF2–/– mice
The discrepancy between signs of increased hypothalamo-pituitary stimulation and modest adrenocortical secretory output prompted us to focus on the adrenal gland. Morphological examination revealed that homo- and heterozygous TIF-2 deficient mice display reduced cell densities in the zona fasciculata of the cortex (Fig. 5
; F2,24=30.09, P<0.001); these animals showed a loss of the usual columnar organization of cells and an abundance of hypertrophic cells and cells with pyknotic nuclei.
|
We next studied several functional descriptors of the adrenal cortex. Adrenal DAX-1 mRNA levels (measured by semiquantitative RT-PCR) were significantly higher in TIF-2–/– vs. WT animals (Fig. 6
; P<0.009). In contrast, steady-state levels of mRNA coding for the orphan nuclear receptor SF-1 were slightly but significantly decreased in TIF-2–/– vs. WT mice (Fig. 6
; P<0.04). Neither one of these parameters was altered 30 min after stress (data not shown).
|
Basal mRNA levels of steroidogenic acute regulatory (StAR) protein in the adrenal cortex were similar in WT and TIF-2–/– genotypes. Notably, however, TIF-2–/– mice failed to show an up-regulation of StAR protein mRNA levels 30 min after acute stress; moreover, these animals showed a markedly attenuated increase in this parameter following exposure to prolonged immobilization stress (Fig. 7
; F5,15=22.24; P<0.001). Similarly, the basal and stress-induced expression of 3ß-hydroxysteroid dehydrogenase (3ß-HSD) in TIF-2–/– mice was significantly lower than that in WT mice (Fig. 7
; F5,18=28.06; P<0.001).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, but its physiological relevance in GR-mediated regulation of the HPA axis has not been addressed. Thus, experiments were designed to test the hypothesis that ablation of TIF-2 would disrupt the efficacy of GR-mediated restraint on basal and stress-induced activity at various regulatory levels.
While TIF-2+/– mice showed profiles that largely resembled those found in WT mice, gene dosage proved an important factor in several respects. Under stress-free conditions, hypothalamic levels of the mRNAs encoding CRH and AVP, the two major corticotropin secretagogues, were elevated in TIF-2–/– mice. Predictably, pituitary reserves of ACTH were also significantly increased in TIF-2–/– animals. Given that the CRH and AVP genes, as well as the POMC gene (whose posttranslational products give rise to ACTH), are all subject to negative regulation by GR activation, these results indicate that deletion of TIF-2 leads to disruption of GR negative feedback at the hypothalamic and pituitary levels and, ultimately, hyperactivity of the central compartment of the HPA axis. Reduced GR mRNA expression in the hippocampus and PVN, both of which are crucially involved in the central control of the HPA axis, would also be expected to contribute to reduced restraint of HPA activity in TIF-2–/– mice. Increased basal adrenocortical activity with flattened diurnal oscillations, and exaggerated and prolonged response to stress would be the expected manifestations of impaired glucocorticoid negative feedback in the brain and pituitary of TIF-2–/– mice. However, none of these predictions was fulfilled: despite compelling evidence for hyperactivation of the hypothalamo-pituitary unit, in TIF-2 deficient mice: i) the amplitude of the nocturnal increase in CORT was commensurate with that seen in WT animals; ii) glucocorticoid levels at their diurnal nadir were significantly lower than in other genotypes; and iii) unlike the situation in WT and heterozygous TIF-2 mutants, the CORT response to stress in TIF-2–/– mice was markedly attenuated during the initial phase and generally showed sluggish dynamics. Thus, the most striking feature of aberrant HPA axis regulation in TIF-2 deficient animals is the dissociation between hyperactivity of the hypothalamo-pituitary compartment and symptoms of inadequate adrenocortical secretory output. Accordingly, alterations in the central mechanisms of HPA axis regulation, although indicative of decreased glucocorticoid restraint, cannot be solely ascribed to the lack of TIF-2, a major amplifier of GR transcriptional signaling (5
6
7
8
9)
.
The patterns of GR expression in the hippocampus of TIF-2–/– mice illustrate the confounds associated with interpretation of our studies in terms of TIF-2 as a GR coactivator only. Based on the prevailing view (23
, 24)
, homologous down-regulation by CORT excess would be the most plausible explanation of decreased GR expression in the hippocampus. However, this is hardly tenable in view of the absence of symptoms of hypercorticalism in TIF-2 deficient mice under either basal or stressful conditions. Our data also do not support previous suggestions that another GR coactivator, SRC-1, may compensate for TIF-2 deficiency (14
15
16)
, despite recent evidence that SRC-1 is involved in HPA axis regulation (25)
. Thus, as the importance of coregulators in the control of nuclear receptor expression has not been comprehensively elucidated, we considered it prudent to assume that TIF-2 may be involved in the control of GR expression through mechanisms that are independent of its role as a nuclear receptor coregulator.
Deficits in the endocrine response to stress could also originate from impaired hypothalamo-pituitary neurohumoral communication. Unlike WT and TIF-2+/– mice, CRH and AVP expression in TIF-2 null mutants was refractory to brief stressful stimuli. Accordingly, the inability of TIF-2–/– mice to up-regulate CRH and AVP expression in response to a stressor may be ascribed to either limited synthetic capacity of the (already hyperactive) hypothalamic neurons or autologous desensitization of pituitary CRH and AVP receptors (26
, 27)
, ultimately resulting in impaired translation of the peptidergic stimuli into adequate pituitary-adrenal secretory responses. This aspect of the neuroendocrine response of TIF-2–/– mice resembles the situation observed in adrenalectomized animals; the latter display increased basal and blunted stress-induced expression of ACTH, CRH, and AVP (28
29
30)
.
In light of the various above-mentioned findings that suggested deficits at the distal end of the HPA cascade, we investigated some key aspects of adrenocortical morphology and function. Previous studies described anomalous gonadal developmental and impaired reproductive function in TIF-2 deficient mice (12
, 15)
. In the present study, we observed that TIF-2 null mutants tend to have smaller adrenal glands, hinting at possible defects in adrenal organogenesis. Both TIF-2–/– and TIF-2+/– mice were found to have structural abnormalities in the glucorticoid-synthesizing zona fasciculata of the adrenal cortex, namely, reduced cell densities and loss of the columnar organization. These characteristics, together with the sporadic occurrence of pyknotic cells in the zona fasciculata, are reminiscent of the histological changes described in congenital adrenal hypoplasia (31)
. The view that adrenocortical development and function may be disrupted in TIF-2 null mutants was bolstered by our observations of high levels of DAX-1 expression in the adrenal cortex and, concomitantly, reduced abundance of SF-1. These two transcription factors have been shown to act in a reciprocal manner in the coordination of adrenocortical growth, maturation and steroidogenesis (32
33
34)
; their expression profiles in TIF-2–/– mice indicate that aberrant adrenocortical development belies the reduced steroidogenic capacity in these animals.
Since TIF-2 can coactivate SF-1 (35
, 36)
, it seems plausible that disruption of TIF-2 affects SF-1 dependent gene transcription which, in turn, impairs corticosteroidogenesis. Indeed, StAR protein and 3ß-HSD, which are positively regulated by SF-1 and crucially involved in glucocorticoid synthesis (37
, 38)
, were poorly induced in the TIF-2 null mutants. Interestingly, TIF-2 null mutants and SF-1-haploinsufficient mice show highly similar phenotypes in terms of adrenal histology and stress responses (39)
.
In summary, our data demonstrate that signs of a "loss of grip" of glucocorticoid feedback in the HPA axis in TIF-2 deficient mice are not likely to be solely ascribed to disrupted GR signal amplification. Preliminary data from ongoing experiments addressing the responsiveness of TIF-2 deficient mice to exogenous glucocorticoids are not indicative of glucocorticoid resistance in these animals. Symptoms of structural and functional impairment in the adrenal cortex of TIF-2 mutants suggest that alterations in the neural control of the HPA axis occur secondarily to insidious adrenocortical insufficiency. In view of earlier and recent demonstrations of GR expression in the fetal (40)
and adult adrenal cortex (41)
, and potentiation of corticotropin effects in adrenocortical cells by glucocorticoids (42)
, the intriguing question opened by these observations is whether the lack of TIF-2, which results in poor amplification of trophic GR signaling, may be responsible for compromised adrenocortical development and steroidogenic capacity. Meanwhile, the association between TIF-2 deletion and adrenal dysfunction (this study) and hypogonadism (12
, 15)
strongly suggest a novel role for this nuclear receptor coregulator in the morphogenesis and function of steroid-producing glands.
|
ACKNOWLEDGMENTS |
|---|
Received for publication July 24, 2006. Accepted for publication August 21, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
