The Role of Glucocorticoid Receptors in Dexamethasone-Induced Apoptosis of Neuroprogenitor Cells in the Hippocampus of Rat Pups

Background. Dexamethasone (Dex) has been used to reduce inflammation in preterm infants with assistive ventilation and to prevent chronic lung diseases. However, Dex treatment results in adverse effects on the brain. Since the hippocampus contains a high density of glucocorticoid receptors (GCRs), we hypothesized that Dex affects neurogenesis in the hippocampus through inflammatory mediators. Methods. Albino Wistar rat pups first received a single dose of Dex (0.5 mg/kg) on postnatal day 1 (P1) and were sacrificed on P2, P3, P5, and P7. One group of Dex-treated pups (Dex-treated D1D2) was given mifepristone (RU486, a GCR antagonist) on P1 and sacrificed on P2. Hippocampi were isolated for western blot analysis, TUNEL, cleaved-caspase 3 staining for cell counts, and morphological assessment. Control pups received normal saline (NS). Results. Dex reduced the developmental gain in body weight, but had no effect on brain weight. In the Dex-treated D1D2 group, apoptotic cells increased in number based on TUNEL and cleaved-caspase 3 staining. Most of the apoptotic cells expressed the neural progenitor cell marker nestin. Dex-induced apoptosis in P1 pups was markedly reduced (60%) by pretreatment with RU486, indicating the involvement of GCRs. Conclusion. Early administration of Dex results in apoptosis of neural progenitor cells in the hippocampus and this is mediated through GCRs.


Introduction
Corticosteroids are used in preterm infants to suppress in�ammation, to facilitate extubation, and/or to prevent chronic lung diseases [1][2][3]. However, such early dexamethasone (Dex) therapy can result in an adverse neurodevelopmental outcome [4,5]. For example, Dex treatment reduces cerebral gray matter volume without affecting white matter and the basal ganglia, suggesting that Dex affects only certain cells in the brain [6][7][8].
Neurons in the dentate gyrus (DG) of the hippocampus continue to divide aer term and therefore remain vulnerable to the adverse effects of steroids during the early postnatal period [9][10][11]. Cells in the hippocampus have a high density of glucocorticoid receptors (GCRs) [12] suggesting that they could be affected by Dex [8][9][10][11][12][13]. Dex is known to change synaptic plasticity in the hippocampus [14]. As rat and human develop over different embryonic time scales [15], rat pups on P1 correspond to the human fetus at about week 22 to 24 of gestation. e e�uivalence in development is re�ected in brain weight, neurochemistry, electroencephalography, and synaptogenesis [16]. erefore, rat pups can be used as an animal model for human preterm infants undergoing drug exposure [17].
Here, we studied the effects of single-dose Dex therapy and the role of GCRs on hippocampal development.

Materials and Methods
2.1. Animals. e study was approved by the Animal Care and Use Committee of National Cheng Kung University. Time-dated pregnant Albino Wistar rats (body weight 250-300 g) were used. Food and water were provided ad libitum. e dams were allowed to deliver naturally on gestational day 21 ± 1. Animals were kept in a ventilated room at 22 ± 2 ∘ C under a 12/12 h light/dark cycle. e day of birth was designated P0.

Study Design.
In a preliminary study, each litter was divided into two groups: the Dex group received a dose of Dex phosphate (0.1, 0.2, or 0.5 mg/kg, i.p.) (Oradexon, 4 mg/mL, Organon, the Netherlands) and controls received an equal volume of normal saline (NS). Compared with the controls, changes in apoptosis were only observed with Dex 0.5 mg/kg, so this dose was used in the subsequent experiments. On day 1 (P1) pups received a single dose of 0.5 mg/kg Dex or NS. On P2, P3, P5, and P7, pups were anesthetized and euthanized with 10% chloral hydrate (W/V, 300 mg/kg, Riedel-de Haen, Germany) and then infused transcardially with NS. e body and isolated brain weights were measured to the nearest milligram. Brain tissues were processed histologically following standard protocols and cut serially at 5 m in the coronal plane.
2.4. TUNEL Assay. e sections were subjected to TUNEL (Oncogene Research, Cambridge, MA) and IF staining. e signal was detected by the streptavidin-horseradish peroxidase conjugate and diaminobenzidine reaction.

Counts of Cells
Stained for TUNEL, Cleaved-Caspase 3, and NeuN. Sections were examined by light microscopy and the images captured by a video camera coupled to a desktop computer (Eclipse 80i, Nikon, Japan). TUNEL-positive cells were identi�ed and counted at 400x magni�cation. e numbers of TUNEL-positive cells in the DG, cornu ammonias 1 (CA1), CA2, and CA3 in the hippocampus were counted. For comparison with the control, sections at the same level were analyzed. Four sections were counted for each pup and the results were averaged. Cleaved-caspase 3-and NeuN-positive cells were counted at 400x magni�cation in four, 1 mm 2 areas in the DG. Data were validated by TissueGnostics FACS-like Tissue Cytometry soware (Vienna, Austria). is method was also applied to IHC, IF, and double-stained cells in subsequent experiments.

Tissue Dissection and Western Blot
Analysis. e hippocampi were dissected from the pups on P2, one day aer Dex or NS treatment (D1D2). Western blots were performed on cytosolic and nuclear fractions of the hippocampus homogenates as described previously [19]. Brie�y, 20 g protein homogenate, determined by the Bradford protein assay (Bio-Rad, Hercules, CA), was separated by SDS-PAGE, blotted onto nitrocellulose (Hy-Bond, Amersham, Arlington Heights, IL) and blocked with nonfat dry milk. Blots were incubated with speci�c primary antibody, followed by incubation with horseradish-peroxidase-conjugated secondary antibodies (Abcam, Cambridge, UK) and detected by enhanced chemiluminescence (Bio-Rad). Samples were normalized to actin and proliferating cell nuclear antigen (pCNA) proteins (EMD Millipore). Each experiment was repeated at least three times.

Statistical Analysis.
Quantitative results are expressed as mean ± standard error (SE). Statistical analyses were performed using one-or two-way ANOVA with a multiple comparisons posttest or the Wilcoxon signed-rank test as appropriate.
values < 0.05 were considered statistically signi�cant. e body and brain weight results were analyzed by mixed-model ANOVA with age as the within-subject factor and Dex as the between-subject factor.

Body and Brain Weights.
Pups that had received Dex (0.5 mg/kg) showed reduced body weights. In contrast, no difference in the developmental growth of brain weight was found between the Dex and control groups (Table 1).

Apoptotic Cell Death.
More TUNEL-positive and cleaved-caspase 3-positive cells were found in the DG of P1 D1D2 Dex-treated group (Figures 1(b)  an increase to 2.7-to 3-fold (Figures 1(c) and 1(f)) with respect to the control (Figures 1(a) and 1(d)).

Glucocorticoid Receptors (GCRs) and Mifepristone (RU486) in Dex-Induced Apoptosis.
Western blot analysis showed that the nuclear fractions of GCRs in the hippocampus were upregulated in the Dex-treated D1D2 group ( Figure 4). e Dex-retarded developmental gain in body weight was blocked by RU486 while neither Dex nor RU486 affected the brain weight in this group (Table 2). Furthermore, Dex-induced apoptosis in the P1 D1D2 group was reduced by the preadministration of RU486 ( Figure  5). Dex treatment alone increased the apoptotic cell count (41.3 ± 0. 0) compared to the control (23.1 ± 0.18); 0.05); the number of apoptotic cells in pups treated with DMSO (22.4 ± 0.3) or DMSO plus Dex (41.9 ± 0.56) was similar to that in pups treated with NS or Dex alone ( § 0.05). RU486 treatment had no additional effect on apoptosis when compared with the NS or DMSO group. Pretreatment with RU486 followed by Dex reduced the apoptotic cell count (29.2 ± 0.45) ( 0.05) ( Figure 5).

3.�. Identi�cation of In�a��ator� Cells.
Eosin/hematoxylin or IHC staining did not reveal any in�ammatory cells in the NS and Dex-treated D1D2. Positive controls aer implantation of rat brain tumor cells revealed OX-6 positive cells (stained brown) which helped identi�cation of microglia in the brain (Figure 6).

Discussion
Steroids have long been used in the treatment of respiratory problems in preterm infants [21]. e complications of adverse neurological effects (such as increased risk of cerebral palsy and neurodevelopment impairment [4,22,23] demand the reevaluation of steroid-based therapeutic strategies in postnatal practice [23][24][25][26]. e safe timing and dosage of Dex remain undecided for preterm patients [4].
On the contrary, our data showed no such change. We attribute this discrepancy to the differences in timing, dosage regimes, and/or the preparation of Dex [28,31,32]. Among clinicians, the general consensus is that a lower dose of Dex (0.1-0.2 mg/kg/day) facilitates tracheal extubation and reduces the risk of chronic-lung disease. is study revealed no deleterious effects on the brain with single low doses of Dex (0.1-0.2 mg/kg), although it became harmful at a higher dose (0.5 mg/kg). e subgranular zone of the DG contains a reserve of neuroglial progenitor cells [33]. Apoptosis is crucial during neuronal development by eliminating excess cells and ensuring proper synaptic connectivity [34,35].
GCRs are known to be involved in Dex-induced apoptosis [36] and are present at high levels in the hippocampus where progenitor cells capable of dividing reside in the DG. Our results showed increased apoptosis throughout the DG, suggesting that perinatal development in the hippocampus is vulnerable even to a single dose of Dex when given at a critical time. How this effect is related to the GCR density, receptor types, and the proliferation of progenitor cells remains to be studied. It has been reported that administration of 3.0 mg/kg Dex to P7 mice increases the apoptosis of cerebellar progenitor cells and reduces the number of cerebellar neurons [37,38]. It is likely that the dosage and timing of Dex treatment and  vulnerability of neural progenitor cells to glucocorticoids together determine the effects on neonatal brain development.

Mediators of �n�ammation
Exposure to Dex in the neonatal period results in marked apoptosis among the nestin-expressing cells in the DG [39]. e importance of Dex-induced apoptosis in the hippocampus and the type of cells affected remain unknown. Our results from double staining in the Dex-treated pups showed higher ratios of cells coexpressing TUNEL and nestin to TUNEL-positive cells, indicating that Dex-induced apoptosis affects the neuroprogenitor cells. is result is consistent with previous reports that neuroprogenitor cells are sensitive to Dex during the early neonatal period [39,40]. Since the ratio of TUNEL and NeuN coexpressing cells to TUNEL-positive cells also increased, Dex treatment might also cause apoptosis in mature neurons and be associated with a transient and acute slowing of cell proliferation during hippocampal development [41,42]. Whether these neurons were derived from the proliferating progenitors cells or represent existing mature neurons in the DG remains to be determined. Mifepristone (or RU-486) is a synthetic steroid with both antiprogesterone and antiglucocorticoid properties. In a recent study, mifepristone was the only GCR antagonist found to increase both mineralocorticoid receptor (MCR) and GCR binding in the rat frontal cortex [43]. e effects of mifepristone on corticosteroid receptor expression could explain the neurocognitive improvement it is reported to induce [43,44]. RU486 has no effect on Dex regarding thymocyte composition and maturation [45]. MCRmediated responses to glucocorticoids, rather than GCRs, are important in steroid-responsive hearing disorders [46]. GCR activity levels are high in the hippocampus during the �rst postnatal week [40]. e anti-in�ammatory effects of glucocorticoids require the presence of GCRs [47]. Interestingly, results of these studies and those of our own (60% reduction in Dex-induced apoptosis by RU486) are consistent, suggesting that the type of receptor and the timing of Dex treatment determine the effects of glucocorticoids on the hippocampus. Examining pup brain tissues stained with eosin�hematoxylin or IHC revealed no in�ammatory cells in the NS and Dex-treated D1D2 groups. ese results suggest that while GCRs are likely to be key players in Dex-induced neuroprogenitor cell death, in�ammatory cells are not.
In summary, brain development is a dynamic process in which the growth spurt, differentiation, and cell responses to endogenous (and exogenous) steroids occur at critical times. Species differences add further complexity to the process. ese species-speci�c developmental schedules allowed the design of animal models to study the effects of drug treatment in preterm babies, as in this study. Here, we have demonstrated that timing is a major factor in determining Dex-induced apoptosis in the hippocampus, the vulnerable cells are neuronal precursors, and the process is partly regulated by GCRs. We also provided cytological evidence that the administration of a single dose of Dex can result in deleterious effects in the brain.  F 5: Dexamethasone-(Dex-) induced apoptosis in P1 D1D2 pups was reduced by preadministration of RU486. Dex treatment increased apoptotic cell counts in the hippocampus compared to control ( , * ); P1 D1D2 pups treated with DMSO or DMSO plus Dex showed results similar to those from pups treated with either normal saline (NS) or Dex alone ( , § ). RU486 treatment showed apoptotic cell counts similar to those from pups treated with NS or DMSO. Pups treated with RU486 plus Dex had lower apoptotic cell counts than those with Dex or DMSO plus Dex ( , † ). P1 D1D2: P1 pups receiving treatment on postnatal day 1 and sacri�ced on day 2.