During pregnancy, glucocorticoids transfer environmental signals to the growing brain and its associated neuroendocrine system to modulate their maturation and function during adolescence and adulthood. Increased in utero exposure to glucocorticoids is associated with impaired fetal growth resulting in low birth weight (LBW) and compromised neural development. The underlying molecular changes affecting brain development, however, are largely unknown. Here, we compared the relative mRNA expression of genes directly involved in glucocorticoid signaling in the hippocampus, amygdala, and cortex of female non-human primate neonates (Macaca fascicularis) of naturally occurring normal birth weight and LBW. We focused on the glucocorticoid receptor (GR) and mineralocorticoid receptor (MR) genes as well as that for 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) and found a significantly decreased MR:GR mRNA ratio in the hippocampus and lower expression of 11β-HSD1 in the amygdala associated with LBW. The MR:GR mRNA ratio in the amygdala and cortex was not associated with birth weight, reflecting tissue-specific effects. Protein quantification in the hippocampus confirmed our finding of a decreased hippocampal MR:GR ratio. Our data suggest that the MR:GR ratio in the hippocampus and the expression of 11β-HSD1 in the amygdala are associated with intrauterine growth restriction in non-human primates during early perinatal development.
The quality of fetal growth predicts that of health over the life span, with a substantial effect on brain-based disorders such as attention-deficit hyperactivity disorder (Bhutta et al. 2002) and depression (Thompson et al. 2001, Alati et al. 2007). Hippocampal structure and function associate with cognitive and mood disorders, and low birth weight (LBW) is associated with reduced hippocampal volume (Buss et al. 2007, de Bie et al. 2010). While human small for gestational age infants show reduced brain weight and cell number in the brain compared with normal birth weight (NBW) controls (de Bie et al. 2010), the hippocampus appears particularly vulnerable. Studies with appropriate animal models show that intrauterine growth restriction (IUGR) associates with reduced hippocampal volume (Mallard et al. 2000, Lister et al. 2005).
There is considerable evidence in humans and other species that impaired fetal growth is linked to increased in utero glucocorticoid exposure (Reinisch et al. 1978, French et al. 1999, Bloom et al. 2001, Nyirenda et al. 2001, Meaney et al. 2007). Cord blood samples obtained from LBW human babies reveal significant elevations in both corticotrophin-releasing factor (CRF) and cortisol (Goland et al. 1993). Interestingly, maternal stress, protein deprivation, tobacco, and alcohol predict impaired fetal growth and increase maternal adrenal glucocorticoid release, which increases placental CRF expression; hence the relationship between maternal and fetal glucocorticoids – placental CRF in IUGR babies (Meaney et al. 2007). Moreover, levels of placental 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2), a placental barrier enzyme that inactivates cortisol, are reduced in LBW babies (Seckl et al. 1999, McTernan et al. 2002). Glucocorticoid administration to pregnant female rodents reduces birth weight (Nyirenda et al. 2001). Likewise, increased fetal glucocorticoid exposure appears to mediate the effects of conditions that associate with impaired fetal growth, such as malnutrition. Protein malnutrition increases fetal glucocorticoid levels in mice as well as in sheep (Bloomfield et al. 2003, Cottrell et al. 2012) and maternal adrenalectomy blocks the effects on fetal growth.
Glucocorticoids are important for normal intrauterine and postnatal brain development, regulating processes such as glial proliferation and myelinogenesis, remodeling axonal and dentritic spines, and affecting cell survival (Meyer 1983, Lupien et al. 2009). Nevertheless, treatments that elevate glucocorticoid levels during late pregnancy compromise brain development and correlate with poorer cognitive performance in humans (Damsted et al. 2011). A reduction in hippocampal volume was found in primates exposed to exogenous glucocorticoids before birth (Uno et al. 1994). These findings suggest that the increased in utero exposure to glucocorticoids that defines the endocrine conditions associated with impaired fetal growth compromises neural development. While fetal growth restriction appears to be associated with increased glucocorticoid exposure, little is known about the expression of corticosteroid receptors in late fetal or early postnatal development. Glucocorticoid receptor (GR) expression is dynamically regulated over perinatal development in a highly tissue-specific manner (Kalinyak et al. 1989, Bohn et al. 1994). It is critical to assess corticosteroid receptor levels to evaluate potential tissue sensitivity to corticosteroids as a function of fetal growth. In this study, we compared the relative mRNA expression of genes directly involved in glucocorticoid signaling in the hippocampus, amygdala, and cortex of non-human primate neonates (cynomolgus macaque) of naturally occurring NBW and LBW, focusing on the glucocorticoid receptor and mineralocorticoid receptor genes as well as that for 11β-HSD1, which converts inactive corticosteroids to cortisol and associates with increased corticosteroid action (Wyrwoll et al. 2011). We report that the MR:GR mRNA ratio is decreased in the ratio of the hippocampus of LBW animals. A similar decrease in the corresponding proteins was observed in hippocampal samples analyzed by western blotting. 11β-HSD1 mRNA showed reduced expression in the amygdala of the same group of animals. The results point to an association of a decreased MR:GR ratio in the hippocampus and reduced expression of 11β-HSD1 mRNA in the amygdala with fetal growth restriction during perinatal development.
Materials and methods
Animals and collection of brain tissue
The animals were bred and killed at the Nafovanny facility in Vietnam. All animal procedures were approved by Nafovanny. Oversight of Nafovanny is provided by the Vietnamese Ministry of Forestry. Our studies were approved by the Singapore Health Institutional Animal Care and Use Committee and performed in accordance with the Guidelines on the Care and Use of Animals for Scientific Purposes set by the National Advisory Committee for Laboratory Animal Research (NACLAR, http://www.ava.gov.sg/AnimalsPetSector/CareAndUseAnimalsForScientificPurp/) of Singapore, which is based on the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes, NHMRC, Australia, as well as the Guide for the Care and Use of Laboratory Animals by the National Academy of Sciences, USA (courtesy of the National Academies Press, Washington, DC, USA), and The Good Practice Guide for the Use of Animals in Research, Testing and Teaching, National Animal Ethics Advisory Committee, New Zealand. Sixty-four pregnant cynomolgus macaque dams (Macaca fascicularis), sired naturally, were monitored before delivery at the Vietnam Primate Breeding and Development Corporation. These 64 pregnant dams gave birth to 65 male and female neonates (original cohort of neonates; Table 1). The normative birth weight range was assessed and 11 female neonates were selected based on their birth weights to comprise two groups (see Table 1 for details): i) LBW group with n=6 and a birth weight range of 316–326 g (4–27th birth weight percentile of the original female cohort) and ii) NBW group with n=5 and a birth weight range of 358–371 g (50–65th birth weight percentile of the original female cohort). The average age of the NBW group was 4.7 days and that of the LBW group was 5.1 days. All neonates were sedated between 0920 and 1115 h with an i.m. injection of ketamine–HCl (15 mg/kg) and exsanguinated under anesthesia. Whole brains were collected and stored at −80 °C.
Cohort and sample details of all animals (Macaca fascicularis)
|Original cohort||NBW group||LBW group|
|Number of females/males||33/32|
|Number of still births||4|
|Number of singletons/twins||63/2|
|Birth weight (g) (and age (days)) of each animal|
|Mean±s.d. birth weight (g)||358±40.7||361±8.2||319±4.2|
|Median (and range of) birth weighta (g)||351 (299–480)|
|Birth weight percentile of the original female cohort||50–65th||4–27th|
|Mean age (days)|
|Hippocampus||4.7 (n=5)||5.1 (n=6)|
|Amygdala||4.8 (n=4)||5.3 (n=5)|
|Cortex||4.1 (n=4)||5.1 (n=5)|
NBW, normal birth weight; LBW, low birth weight.
Median birth weight of females (excluding still births and twins) in grams.
Coronal sections from the right hippocampus and the right amygdala (including adjacent cortical tissue) were prepared at an interval of two sections of 300 and 20 μm respectively by cryosectioning and thaw-mounted on poly-lysine-coated slides. Slides were stored at −80 °C until further processing for Nissl staining and RNA extraction. Nissl staining facilitated the identification of the regions of interest, e.g. right anterior hippocampus, cortex, and right amygdala. We defined the anterior hippocampus as the hippocampal part that is present in the coordinates A9.6 to A5.6 of the stereotaxic brain atlas by Szabo & Cowan (1984) (Fig. 1A). The lateral geniculate nucleus was used as a landmark (Fig. 1A). The cortical tissue (containing the parietal area-associated area of the superior temporal sulcus and the intraparietal sulcus-associated area in the superior temporal sulcus) was taken from the same section as the hippocampus. The amygdala was identified by the presence of its paralaminar nucleus and/or its characteristic shape and location anterior to the hippocampus (Fig. 1B). A standard protocol was used for Nissl staining of 20 μm brain sections. Pictures were taken with an Axio Observer Z1 microscope (Zeiss, Oberkochen, Baden-Wuerttemberg, Germany) and stitched with TissueFAXS suite version 3 Software (TissueGnostics, Vienna, Austria).
RNA extraction and cDNA synthesis
Areas used for RNA extraction from the hippocampus, amygdala, and cortex are indicated in Fig. 1A and B. Total RNA was isolated from a single 300 μm section, using the All Prep DNA/RNA Micro Kit (Qiagen) following the manufacturer's protocol. RNA quantity was measured spectrophotometrically (Nanophotometer, Implen, Munich, Germany). First-strand synthesis of cDNA was carried out with oligo-dT primer using the Transcriptor First Strand cDNA synthesis kit (Roche). cDNAs were tested for genomic contamination with a genomic contamination primer assay (Qiagen).
Quantitative real-time PCR
Quantitative real-time PCR was performed on an LC480 II Real-Time PCR instrument with Light Cycler 480 SYBR Green I Mastermix (Roche). All primers were designed with published sequences of Macaca mulatta and are listed in Supplementary Table 1, see section on supplementary data given at the end of this article. Pre-designed primers were purchased from Qiagen. Primer pairs for the amplification of different first exons of the NR3C1 gene were designed with the forward primer targeting a specific first exon while the target sequence for the reverse primer was in exon 2. The specificity of these primers was confirmed by cloning and sequencing of the PCR products. Primer efficiencies were determined with CAmpER Software (www.cebitec.uni-bielefeld.de) using the DART method (Peirson et al. 2003). Quantification was performed with the Advanced Relative Quantification application of the Light Cycler 480 Software 1.5 (Roche). Samples were analyzed in triplicates and normalized to the expression of β2-microglobulin. Values are expressed in arbitrary units. Pilot assays showed significant differences in GAPDH expression as a function of birth weight, but no effect on β2-microglobulin.
Protein extraction and western blotting
Hippocampal proteins were isolated from two 300 μm sections from an area as indicated in Fig. 1A. Tissue was lysed in RIPA buffer (containing 1% (v/v) Triton X-100 and 1% (w/v) sodium deoxycholate) by vortexing and sonication with a Bioruptor (Diagenode, Liège, Belgium)). Protein concentration was determined with the Dc Protein Assay Kit (Bio-Rad). Twenty micrograms of proteins were analyzed by western blotting with the following antibodies: GFAP (Millipore), GR (Millipore), neurofilament-L (Cell Signaling Technology), and MR (Novus Biologicals). β-Tubulin and β-actin antibodies (both from Sigma) were used to demonstrate equal loading. Quantification of protein bands was carried out with the Odyssey Infrared Imaging System and Odyssey Software version 2.1 (LI-COR, Lincoln, NE, USA) according to the manufacturer's instruction. For the calculation of the MR:GR protein ratio, we added all the values for the different GR isoforms detected in a given sample.
Data are presented as means±s.e.m. using GraphPad Prism 5 Software, Version 5.04 (GraphPad Software, San Diego, CA, USA). Student's t-tests were used to statistically compare differences in gene expression between samples from LBW and NBW neonates with a significance cutoff value of P<0.05. We removed one outlying value (animal 5 from the NBW group) from the hippocampal MR:GR mRNA ratios that was more than 4 s.d.s from the mean. There were no outlying values in any of the remaining analyses.
The aim of this study was to characterize gene expression differences of candidate genes associated with glucocorticoid signaling in different brain regions early in development as a function of birth weight across the normal range in non-manipulated cynomolgus macaques. From an original cohort of 65 male and female neonates, we selected female neonates to form two groups (see Table 1 for details): a NBW group (n=5, birth weight range of 351–371 g, 50–65th birth weight percentile of the original female cohort) and a LBW group (n=6, birth weight range of 316–326 g, 4–27th birth weight percentile of the original female cohort). The respective mean birth weights differed significantly (t9=10.86, P<0.001; mean 361.2 g, s.e.m. 3.7 g for NBW vs mean 318.8 g, s.e.m. 1.9 g for LBW). The brain regions included in our study were the anterior hippocampus, cortex (adjacent to the hippocampus), and the amygdala, all dissected from the right temporal lobe (Fig. 1A and B). The hippocampi of all 11 samples were included in our study (n=5 for NBW group and n=6 for LBW group). Owing to impaired integrity of some of the brain samples, group sizes for the amygdala and cortex were n=4 for NBW and n=5 for LBW neonates. As a result, the mean age difference between the two groups varied from 0.4 to 1 day (see Table 1 for details). We analyzed the relative expression level of genes encoding the GR (NR3C1), the MR (NR3C2) and 11β-HSD1 (11β-HSD1). Although we did not detect any statistically significant differences for the three genes examined individually (Fig. 2A, B and C), the ratio of MR:GR mRNA expression was significantly decreased in hippocampal samples from LBW compared with NBW animals (t8=5.147, P=0.001; Fig. 2D).
The human GR gene contains at least ten alternative non-coding first exons (Turner et al. 2006, Presul et al. 2007). We analyzed the usage of the following first exons: 1A, 1B, 1C3, 1D, 1E, 1F, 1G, and 1H. Expression of exons 1A, 1D, 1E, and 1G was not detected. No significant differences in the transcript levels for exons 1B, 1C3, 1H, and 1F were found as a function of birth weight (data not shown).
In light of the difference of the hippocampal GR:MR ratio, we estimated the cellular composition in hippocampal samples to ensure that differences in corticosteroid receptor mRNA levels were not associated with differences in the relative inclusion of glial and neuronal cells in our assays. We thus examined hippocampal expression of the neuronal marker protein neurofilament-L and the glial marker protein GFAP by western blotting. Comparable expression of both proteins was detected across all samples (Fig. 3A), suggesting that up-regulation of GR expression in LBW hippocampi is not associated with variation in neuron and glial cell numbers.
Additional western blotting revealed the expression pattern for GR protein in the hippocampus of NBW and LBW neonates (Fig. 3B). The GR protein is expressed in several N-terminal variants. The bands included are indicated in Fig. 3B and quantified in Fig. 3C. Our blot contained four of the five protein isoforms described in humans (Sinclair et al. 2011): quantification of these isoforms revealed a higher expression of isoforms GRα-D1 and GRα-Dx, with no group differences in any of the single isoforms (data not shown and Fig. 3C). The latter result is consistent with the quantification of the GR mRNA in the hippocampus of the NBW and LBW groups (Fig. 2A). Our blot did not contain proteins smaller than circa 30 kDa; hence, we were unable to detect the fifth GR isoform of about 25 kDa (Sinclair et al. 2011). The blot was re-probed with an anti-MR antibody to detect MR expression in the hippocampus (Fig. 3B). As with the MR mRNA expression, quantification and statistical analysis did not reveal any difference in protein level between the two birth weight groups (Fig. 3E). However, calculation of the MR:GR protein ratio (see Material and methods section for details) showed a significantly reduced ratio in the LBW group compared with the control group (t9=2.636, P=0.03; Fig. 3F). This result is in agreement with the reduced MR:GR mRNA ratio presented in Fig. 2D.
Expression of the GR, the MR, and 11β-HSD1 in the NBW and LBW groups was analyzed in the amygdala and cortex. There was significantly lower expression of 11β-HSD1 in the amygdala of the LBW group (t7=3.03, P=0.02; Fig. 4C). No significant changes were detected for the other genes and tissues tested (Fig. 4A, B, D, E and F), although the difference in the MR mRNA expression in the cortex approached significance (t7=2.019, P=0.08; Fig. 4E). Likewise, there was no statistically significant change in the GR:MR mRNA ratio in the amygdala and cortex respectively (data not shown).
Fetal growth restriction associates with increased fetal exposure to glucocorticoids through either direct activity of the adrenal or indirectly through reduced expression of 11β-HSD2, a placental barrier enzyme that metabolizes cortisol (or corticosterone in rodents) to biologically inactive corticosteroids (Seckl et al. 1999, Meaney et al. 2007). Maternal stress or exposure to elevated levels of glucocorticoids compromises hippocampal development in non-human primates (Uno et al. 1989, Coe et al. 2003) and is a candidate mechanism for the compromised neural development associated with IGUR states. However, the status of the relevant corticosteroid signaling systems in the developing hippocampus was unknown. The results of our studies with a cynomolgus macaque model of fetal growth suggest that LBW associates with decreased hippocampal GR:MR mRNA and protein ratios, with levels of GR and MR protein that are at least comparable with animals of NBW. These findings reveal that LBW associates with sustained hippocampal sensitivity to glucocorticoids.
The MR/GR balance hypothesis (Joëls et al. 2007) suggests that the relative activity of the MR and the GR (and other mediators of the stress response) determine the sensitivity and response of the brain to glucocorticoid signaling and stress in general. A normal ratio of these receptors promotes health, homeostasis, and adaptation. The MR rapidly enhances the excitability of hippocampal neurons through a non-genomic action in response to elevated glucocorticoid levels (Karst et al. 2005, Qiu et al. 2010) indicating a role for this receptor in the acute behavioral response to stress. Downregulation of the hypothalamic–pituitary–adrenal (HPA) axis, the principle mediator of the stress response, is achieved via a negative feedback mediated by activation of GRs in the hippocampus and prefrontal cortex. Recently, the MR:GR ratio was found to control activity of the HPA axis under stressed but not basal conditions in mice (Harris et al. 2012). An imbalance of these two receptors early in life as detected in the hippocampus of LBW neonates in our study may alter the glucocorticoid sensitivity of this brain structure. In contrast to the hippocampus, we did not detect a difference in the GR:MR mRNA ratio in the amygdala. However, reduced expression of 11β-HSD1, an enzyme that converts inactive corticosteroids to cortisol, in the amygdala of LBW animals might also change its glucocorticoid sensitivity.
Studies of Dean & Matthews (1999) and Sloboda et al. (2008) demonstrated the impact of synthetic glucocorticoid administration on the expression of fetal MR and GR: repeated maternal dexa/betamethasone injection enhanced fetal expression of both receptors in the hippocampus. By contrast, fetal corticosterone exposure or prenatal stress stably decreases MR as well as GR expression and increases basal corticosterone levels in adult rats (McCormick et al. 1995, Levitt et al. 1996, Welberg et al. 2001). Likewise, early-life stress in the common marmoset leads to long-term mild hippocampal decreases in GR as well as MR expression at adolescence (Arabadzisz et al. 2010). These studies suggest a dynamic regulation of corticosteroid receptor transcription over development by glucocorticoids induced by environmental factors like stress or food restriction. In our study, we did not detect a significant change in hippocampal corticosteroid receptor expression between the two birth weight groups. This might be explained by the fact that we used non-manipulated animals where changes might be more subtle compared with experimental conditions. Nevertheless, the change in the MR:GR ratio in the LBW group was highly significant (and not detected in the two other brain structures analyzed), suggesting indeed a dynamic regulation by environmental factors. Further studies are necessary to clarify whether the alterations in hippocampal MR:GR ratio in the hippocampus of LBW cynomolgus macaques are sustained into later life.
Our findings suggest differential relative hippocampal expression of the GR and MR as a function of birth weight. Importantly, activation of these receptors leads to contrasting effects on neural function (Joëls et al. 2007). MR activation maintains neuronal excitability and promotes neurogenesis (Gass et al. 2000, Joëls et al. 2007); MR knockout mice show profound reductions in hippocampal development. By contrast, activation of the GR dampens neuronal excitability and may inhibit neurogenesis and generally compromises synaptic plasticity (McEwen 1999). While a matter of some speculation, the results presented here suggest the potential for differential MR/GR activation in the hippocampus as a function of fetal growth. Similarly, downregulation of 11β-HSD1 in the amygdala in the LBW group seems to be a result of impaired fetal growth. Like the hippocampus, the volume of the amygdala increases postnatally (Lupien et al. 2009). We hypothesize that the changed GR:MR ratio in the hippocampus and the lower expression of 11β-HSD1 in the amygdala are manifestations of an altered developmental trajectory for the LBW neonates that might lead to impaired brain function and changes in the HPA axis adjustment.
This is linked to the online version of the paper at http://dx.doi.org/10.1530/JME-12-0218.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
This study was supported by the Agency of Science, Technology and Research (A*STAR), Singapore.
The authors thank Dr Makoto Yawata for his assistance in taking pictures. They also thank Drs Joanna Holbrook, Walter Stünkel, and Judy Sng for critically reading the manuscript.
ArabadziszDDiaz-HeijtzRKnueselIWeberEPilloudSDettlingACFeldonJLawAJHarrisonPJPryceCR2010Primate early life stress leads to long-term mild hippocampal decreases in corticosteroid receptor expression. Biological Psychiatry671106–1109. (doi:10.1016/j.biopsych.2009.12.016).
GassPKretzOWolferDPBergerSTroncheFReichardtHMKellendonkCLippHPSchmidWSchützG2000Genetic disruption of mineralocorticoid receptor leads to impaired neurogenesis and granule cell degeneration in the hippocampus of adult mice. EMBO Reports1447–451. (doi:10.1093/embo-reports/kvd088).
McCormickCMSmytheJWSharmaSMeaneyMJ1995Sex-specific effects of prenatal stress on hypothalamic–pituitary–adrenal responses to stress and brain glucocorticoid receptor density in adult rats. Brain Research. Developmental Brain Research8455–61. (doi:10.1016/0165-3806(94)00153-Q).
McTernanCLDraperNNicholsonHChalderSMDriverPHewisonMKilbyMDStewartPM2002Reduced placental 11β-hydroxysteroid dehydrogenase type 2 mRNA levels in human pregnancies complicated by intrauterine growth restriction: an analysis of possible mechanisms. Journal of Clinical Endocrinology and Metabolism864979–4983. (doi:10.1210/jc.86.10.4979).
QiuSChampagneDLPetersMCataniaEHWeeberEJLevittPPimentaAF2010Loss of limbic system-associated membrane protein leads to reduced hippocampal mineralocorticoid receptor expression, impaired synaptic plasticity, and spatial memory deficit. Biological Psychiatry68197–204. (doi:10.1016/j.biopsych.2010.02.013).
SlobodaDMMossTJLiSMatthewsSGChallisJRNewnhamJP2008Expression of glucocorticoid receptor, mineralocorticoid receptor, and 11β-hydroxysteroid dehydrogenase 1 and 2 in the fetal and postnatal ovine hippocampus: ontogeny and effects of prenatal glucocorticoid exposure. Journal of Endocrinology197213–220. (doi:10.1677/JOE-07-0375).