Activation of Trk neurotrophin receptors by glucocorticoids provides a neuroprotective effect

Abstract

Glucocorticoids (GCs) display both protective and destructive effects in the nervous system. In excess, GCs produce neuronal damage after stress or brain injury; however, the neuroprotective effects of adrenal steroids also have been reported. The mechanisms that account for the positive actions are not well understood. Here we report that GCs can selectively activate Trk receptor tyrosine kinases after in vivo administration in the brain and in cultures of hippocampal and cortical neurons. Trk receptors are normally activated by neurotrophins, such as NGF and brain-derived neurotrophic factor, but the activation of Trk receptors by GCs does not depend on increased production of neurotrophins. Other tyrosine kinase receptors, such as EGF and FGF receptors, were not activated by GCs. The ability of GCs to increase Trk receptor activity resulted in the neuroprotection of neurons deprived of trophic support and could be modulated by steroid-converting enzymes. Pharmacological and shRNA experiments indicate that Trk receptor activation by GCs depends on a genomic action of the GC receptor. The ability of GCs to promote Trk receptor activity represents a molecular mechanism that integrates the actions of GCs and neurotrophins.

Glucocorticoids (GCs) are commonly used to treat allergic, inflammatory, and autoimmune diseases due to their antiinflammatory and immunosuppressive effects. GCs act through an intracellular glucocorticoid receptor (GR) that regulates gene transcription. In addition to their peripheral effects, GCs also exert effects in the CNS. GCs are neuroprotective, anticonvulsive, and anxiolytic and have been linked to depression, epilepsy, anxiety, and memory (1). Detrimental effects of GCs have been documented, particularly during stress, when prolonged high levels of GCs lead to neuronal cell death (2, 3). Beneficial effects of moderate increases in GCs include the modulation of synaptic plasticity and hippocampal-dependent cognition (1, 4). The plethora of actions of GCs in the brain suggests communication with neurotrophic-signaling systems.

The neurotrophin receptors in the CNS promote neuronal survival and synaptic plasticity (5, 6). Neurotrophins, such as NGF and brain-derived neurotrophic factor (BDNF), activate Trk receptor tyrosine kinases through receptor dimerization, followed by autophosphorylation and resultant intracellular signaling. Previous studies have documented that GCs influence the actions of growth factors, such as FGF (7) and neurotrophins (8, 9). However, the mechanisms by which GCs influence growth and trophic factors are not understood.

Here we report that GCs, including dexamethasone (Dex), a potent GR agonist, selectively activate Trk receptor tyrosine kinases in neural cells. Acute administration of Dex promotes TrkB activation in the CNS in vivo without an increase in neurotrophin levels. The activation of Trk receptors by GCs required genomic actions of GR. A consequence of GC-induced Trk activation is the induction of Akt and increased neuronal survival. Thus, GCs use a unique neurotrophic receptor activation mechanism that likely explains the trophic and high-order properties of steroid action in the brain.

Results

Acute Dex Promotes Trk Phosphorylation in the Rat Brain.

Acute administration of Dex has been reported previously to increase GR phosphorylation on residue S211 (Fig. 1A). This phosphorylation event represents a surrogate biomarker for activated GR (10, 11). After i.p. administration of Dex for 6 h to postnatal day 18 (P18) rats, we observed increased S211 phosphorylation of GR in hippocampal lysates. In addition, the activity of TrkB receptors also was increased as a result of Dex treatment.

Fig. 1.

Acute dexamethasone promotes phosphorylation of TrkB in P18 rat brain. (A) P18 male and female rats were administered i.p. Dex or vehicle (0.9% saline) for 6 h. Phosphorylation of hippocampal TrkB (Y816) and GR (S211) was detected in total lysates. Representative results from two different animals per group are displayed. (B) Normalized hippocampal phospho-TrkB/TrkB levels (mean ± SEM). Each sample represents individual animals (n ≥ 17 per group). (C) Correlation between TrkB and GR phosphorylation in the hippocampus of 24 controls and 16 Dex-treated rats from an independent experiment. (D) P18 rats were first administered i.p. vehicle or 150 mg·kg⁻¹ metyrapone. Then 10 mg·kg⁻¹ Dex or vehicle was coadministered i.p. 3 h postmetyrapone for 6 h. Phosphorylation of TrkB and GR was detected in hippocampal lysates. Males and females were analyzed separately. Representative results from two different animals per group are displayed. (E) Normalized hippocampal phospho-TrkB/TrkB levels (mean ± SEM). Each sample represents individual animals (females, n ≥ 14 per group; males, n ≥ 12 per group). (F) Correlation between hippocampal TrkB and GR phosphorylation in males from an independent experiment (n ≥ 6 animals per group).

To monitor TrkB activation, we developed an antibody that recognizes Y816 when phosphorylated in response to BDNF. The phospho-TrkB antibody detected both glycosylated (145 kDa) and unprocessed (110 kDa) TrkB products, but did not recognize TrkA when activated by NGF [supporting information (SI) Fig. 6]. After Dex administration, there was a dose-dependent increase of TrkB phosphorylation in the hippocampus when compared with vehicle treated animals (P < 0.05) (Fig. 1 A and B). Increased TrkB phosphorylation also was significantly associated with increased GR phosphorylation in Dex-treated animals despite interanimal variability (R² = 0.52) (Fig. 1C).

To test the effects of increasing endogenous GC levels, we injected metyrapone, a potent inhibitor of cortisol synthesis. At high doses (150 mg·kg⁻¹), metyrapone paradoxically can result in elevated levels of corticosterone, the major endogenous GC in rodent (12). Rats that received metyrapone for 9 h showed increased GR and TrkB phosphorylation when compared with vehicle-treated rats (P < 0.05 in males) (Fig. 1 D and E) and elevated corticosterone levels (SI Fig. 7). TrkB phosphorylation was more variable in females than in males. Accordingly, amplitudes of corticosterone levels were more variable after treatment with metyrapone in females, suggesting a more reactive HPA axis in females than in males (SI Fig. 7). Coadministration of Dex for 6 h further increased GR and TrkB phosphorylation in the hippocampus of both females and males when compared with metyrapone-treated rats (P < 0.05; R² = 0.74) (Fig. 1 E and F). Thus, exogenous and endogenous GCs promote Trk receptor phosphorylation in vivo.

Acute Dex Treatment Promotes TrkB Phosphorylation in Neurogenic Regions.

In the dentate gyrus (DG) of the hippocampus, both TrkB and GR proteins are expressed by granular cells. TrkB is found at high levels in dendrites and axonal projections of the hilus, and GR is found in the soma and nucleus (Fig. 2A). DG cells that stained with phospho-TrkB antibody also coexpress the neuronal marker NeuN, nestin, and the glia marker GFAP, but did not coexpress doublecortin (DBC), a marker for migrating neurons (Fig. 2B). After Dex administration for 6 h, phosphorylated TrkB was detected within the SGZ, the neurogenic region of the DG (SI Fig. 8). An increased number of phospho-TrkB-positive cells (NeuN and GFAP) was observed in the GCL and SGZ in response to metyrapone or dexamethasone or in combination (Fig. 2C). Similarly, phosphorylated TrkB also was found in another neurogenic region, the subventricular zone, of Dex-treated rats (SI Fig. 9).

Fig. 2.

Acute dexamethasone promotes TrkB phosphorylation in neurogenic regions. (A) Expression of hippocampal GR and TrkB proteins in P18 rats. (B) Colocalization of phospho-TrkB labeling with markers in the dentate gyrus. Large neuronal cells and glial processes are notably stained in the GCL and SGZ. (C) Representative phospho-TrkB staining in neurogenic region of the dentate gyrus from P18 males that received the indicated treatment. (D) Total neurotrophins levels (mean ± SEM) in the lysates from the cortex and hippocampus of Dex-treated rats detected by ELISA.

Because neurotrophins are expressed in DG granule cells, we measured the levels of NGF, BDNF, and NT3 in hippocampal lysates from vehicle and Dex-treated rats by ELISA. Administration of Dex for 6 h did not alter neurotrophin protein levels either in the hippocampus or the parietal cortex, which also express abundant GR, TrkB, and BDNF (Fig. 2D). These results suggest that TrkB activation in response to Dex is independent of neurotrophin production.

Acute Dex Activates TrkB Signaling and Trophic Effects in Vitro.

To further investigate the effects of Dex treatment upon TrkB, we evaluated the effect of GR activation in brain slices. Stimulation of cortical slices from P9–P10 rat brains with 100 ng·ml⁻¹ BDNF or 1 μM Dex resulted in TrkB, Akt, PLCγ, and Erk phosphorylation. The signaling effects of BDNF were rapid, robust, and durable (Fig. 3 A and B). In contrast, signaling effects of Dex were delayed and weaker, and they lasted for several hours (Fig. 3 A and B). These results confirmed that GC can stimulate TrkB in neural cells in an ex vivo setting.

Fig. 3.

GCs activate Trk signaling and prosurvival effects in vitro. (A) Cortical slices (200 μm) from P9–P10 rats were treated with 1 μM Dex or 100 ng·ml⁻¹ BDNF for the indicated time. Lysates were probed with the indicated antibodies. (B) Quantification of Trk signaling responses to BDNF and Dex (mean ± SEM). Data were normalized to the untreated controls. (C) Phospho-TrkB immunoreactivity induced by 1 μM corticosterone for 3 h in cultured cortical neurons starved from B27 for 5 h. (D) Cultured hippocampal neurons starved from B27 for 5 h were treated with 1 μM corticosterone for 3 h either alone or in combination with 10 μM mifepristone. Trk immunoprecipitates were probed with Y-P and Trk antibodies. (E) Cultured cortical and hippocampal neurons were starved from B27 supplement. Drugs (100 nM K252a, 10 μM LY294002, 10 μM mifepristone, 10 μM spironolactone, 50 ng·ml⁻¹ BDNF, 1 μM corticosterone, and 1 μM cortisone) were applied to the cells once as soon as B27 was removed. Complete cell death in untreated controls usually occurred within 48–72 h after B27 deprivation. The percent cell survival (mean ± SEM) was quantified by subtracting the number of apoptotic nuclei to a population of 2,000 counted cells per condition. (F) Neurotrophin levels (mean ± SEM) released in the media of cultured cortical and hippocampal neurons were detected by ELISA. Treatments included 50 mM KCL for 10 min, 1 μM Dex, and EtOH 0.1% vehicle for 4 h.

To further characterize the role of neuronal cells in this mechanism, we used primary neuronal cultures from rat embryonic brains in which glial contamination was eliminated by 5-fluorouracil, an antimitotic drug. Stimulation of cortical neuron cultures with corticosterone for 3 h resulted in phosphorylation of TrkB in the neuronal soma (Fig. 3C). Phosphorylation of TrkB receptor was also observed in hippocampal neurons after treatment with 1 μM corticosterone for 3h. This response was reduced when cells were cotreated with 10 μM mifepristone, a GR antagonist (Fig. 3D). Together these data indicate that GR activation also can promote TrkB signaling in neuronal cells.

Neurotrophins are potent prosurvival molecules for neuronal cells (5). To further characterize the potency of GC-induced TrkB signaling, we assessed the ability of Dex to maintain the survival of primary neurons during trophic deprivation. Cell death (monitored by counting apoptotic nuclei) occurred within 72 h after B27 supplement deprivation. Treatment with BDNF or Dex rescued >30% of hippocampal (Fig. 3E) and cortical (SI Fig. 10) neurons in a dose-dependent manner. Suboptimal doses of Dex (0.1 mM) and BDNF (5 ng·ml⁻¹) provided only a small additive survival effect (+5%) (Fig. 3E), suggesting a redundant mechanism. Mifepristone, an inhibitor of GR, abolished the trophic effects of Dex, whereas spironolactone, a mineralocorticoid receptor antagonist, had little effect, suggesting that GR was the target of Dex for neuronal survival. Corticosterone displayed similar survival properties in both cortical and hippocampal neurons (Fig. 3E). K252a, a Trk inhibitor, reduced the protective effects afforded by both BDNF and Dex, indicating that both trophic stimuli involved Trk receptor activity. Finally, inhibition of the PI3K–Akt pathway by LY294002 prevented the protective effects of Dex.

Increased neurotrophin production might explain the activation of Trk and trophic effects by GCs. Therefore, we measured the levels of NGF, BDNF, and NT3 proteins from cortical and hippocampal neuron culture media by ELISA (Fig. 3F). We observed a significant increase in BDNF and NT3 levels after depolarization with 50 mM KCl for 10 min as reported previously (13). In contrast, no significant changes were observed in the culture media of Dex-treated neurons (Fig. 3F). These results indicated that the observed GC-induced Trk activation was independent of neurotrophin release.

Specificity of GR-Trk Signaling Cross-Talk.

To investigate the mechanism of GCs on Trk activation, we used PC12 cells, a neuroendocrine cell line, that express endogenous GR and are transfected with TrkA, TrkB, or TrkC receptors. Corticosterone promoted phosphorylation of each Trk species (Fig. 4A). Importantly, other receptor tyrosine kinases, such as EGFR and FGFR, were not activated by GCs (Fig. 4B). This result further suggests that GCs selectively promote phosphorylation of Trk receptors.

Fig. 4.

In vitro characterization of GCs specificity on Trk activation. (A) Phosphorylation of transfected TrkA, TrkB, and TrkC isoforms in PC12 cells. Cells were starved from serum and treated with 1 μM corticosterone for 3 h. Trk immunoprecipitates were probed with Y-P and Trk antibodies. (B) Phosphorylation of endogenous EGFR and FGFR in serum-starved PC12 cells in response to 1 μM corticosterone for 3 h, 50 ng·ml⁻¹ EGF for 10 min, or 50 ng·ml⁻¹ FGF1 for 10 min with heparin. EGFR and FGFR immunoprecipitates were probed with Y-P and EGFR or FGFR antibodies. (C) Dose–response curves were obtained by measuring phospho-TrkA/TrkA levels (mean ± SEM) from serum-starved PC12-TrkA cells treated with the indicated GC for 3 h. Data were normalized to the untreated control. (D) GC biosynthetic enzymes HSD1 or HSD2 were overexpressed in PC12-TrkA cells by lentiviruses. Expression of HSD was monitored 5 days after infection in lysates of serum-starved cells with a Flag antibody; GFP immunoreactivity accounts for infectivity. TrkA phosphorylation in response to 0.1 μM GC stimulation for 3 h was detected by Y-P antibody in TrkA immunoprecipitates.

To characterize the potency of natural and synthetic GCs in clinical use, we performed a dose–response analysis of Trk phosphorylation in PC12 cells. Dex was the most potent (EC₅₀ = 4.6 ± 3 nM) (Fig. 4C). Prednisone and 6α-methylprednisolone were less potent (EC₅₀ = 38 ± 15 and 2.3 ± 2 nM, respectively). Hydrocortisone (or cortisol), the most abundant circulating corticosteroid in humans, also displayed subnanomolar potency (EC₅₀ = 11.3 ± 5 nM). Interestingly, Trk phosphorylation was not observed with cortisone, a structurally related GC. However, cortisone increased survival of hippocampal neurons, but not cortical neurons after trophic support deprivation (Fig. 3E).

Cortisone is an inactive precursor that is normally converted to corticosterone by 11β-hydroxysteroid hormone convertase-1 (HSD1) (14). Indeed, the expression of 11β-HSD1 in PC12 cells resulted in Trk phosphorylation by cortisone (Fig. 4D). Conversely, the expression of 11β-HSD2, which catalyzes the conversion of corticosterone to cortisone, reduced the effectiveness of exogenous corticosterone. These results indicate the specific requirement for liganded activation of GR to promote Trk phosphorylation and trophic abilities. Other steroid hormones such as β-estradiol, retinoic acid, thyroid hormone T3, testosterone, and nonsteroid hormones (vasopressin and exendin-4) did not induce Trk phosphorylation, although receptors for these ligands were expressed (data not shown). Together these results point to the specificity of GC action and indicate that corticosteroids are sufficient to elicit Trk phosphorylation.

Genomic Functions of GR Are Involved in GC-Trk Cross-Talk.

GCs interact with specific intracellular receptors that modulate gene transcription in target cells. In addition to the genomic actions of GR, there is increasing evidence for rapid nongenomic effects of steroids (15). The finding that GC activates Trk receptors raises the question of whether this effect is due to a genomic or nongenomic effect of GR.

To examine the involvement of GC membrane-bound receptors, we tested a nonpermeable GC consisting of BSA conjugated to corticosterone (Fig. 4C). Activation of Trk receptors was not observed by this treatment. In contrast, Trk phosphorylation was detected >1 h after initial stimulation with GCs. For example, a short pulse of 1 μM corticosterone for 5 min was sufficient to promote delayed phosphorylation of Trk (Fig. 5A), Akt, and PLCγ (data not shown) that lasted for many hours. Longer pulses promoted higher Trk phosphorylation in terms of magnitude and duration. In addition, GC-induced Trk phosphorylation was not observed when transcription or translation was blocked by actinomycin D and cycloheximide, respectively (SI Fig. 11). These results suggest a genomic mechanism.

Fig. 5.

Activation of Trk by GCs depend on GR genomic functions. (A) PC12-TrkA cells starved from serum were exposed to 1 μM corticosterone for the indicated pulse time. GC was washed away in PBS, followed by a chase in serum-free medium for the indicated time. TrkA was immunoprecipitated and probed for Y-P. Data were normalized to the untreated control. (B) Specific shRNA against the rat GR (shRNA-GR) or a mismatch shRNA (neg) were delivered to PC12-TrkA cells by lentivirus. The expression of GR was monitored 5 days after infection in lysates of serum-starved cells with the GR(M20) antibody. Treatments included 0.1 μM corticosterone for 3 h or EtOH 0.1%. Levels of phospho-TrkA were detected from Trk immunopecipitates and expressed as p-TrkA/TrkA ratio (mean ± SEM) normalized to the untreated control (lane 1). (C) Molecular replacement of endogenous GR by mutants. The GFP cassette in the virus backbone carrying the shRNA specific for the GR rat sequence was replaced by the GR human sequence that is resistant to the ShRNA. PC12-TrkA cells were infected with the indicated virus. The expression of GR was monitored 5 days after infection in lysates of serum-starved cells with the GR(M20), GR (human), and GR(P20) antibodies to recognize rat GR, human GR, and both rat plus human GR, respectively. The same strategy was applied to the GR-DBD and GR-AF1 mutants. Sgk1, a GR-responsive gene, was used as a marker to evaluate GR genomic potency. Treatments included 0.1 μM corticosterone for 3 h or EtOH 0.1%. (D) Levels of phospho-TrkA were detected from Trk immunopecipitates and expressed as p-TrkA/TrkA ratio (mean ± SEM) normalized to the untreated control (lane 1).

To verify GR's role, we designed a specific shRNA against the rat GR sequence (shRNA-GR, Fig. 5C, virus no. 2) and generated lentivirus that was used to infect PC12 cells. The down-regulation of GR by shRNA-GR abolished GC-induced Trk phosphorylation (Fig. 5 B and C). As a negative control, a mismatched shRNA (shRNA-neg, virus no. 1), which did not lower GR levels, permitted Trk activation. To validate this result further, we performed a rescue experiment with the human GR sequence. This sequence, which is resistant to the shRNA-GR, was subcloned in place of the GFP cassette into the same lentiviral vector backbone carrying the shRNA-GR to produce the virus no. 3 (Fig. 5C). Infection of PC12 cells with virus no. 3 resulted in the expression of human GR and eliminated the expression of the endogenous GR protein, thus restoring GR expression and rescuing Trk phosphorylation induced by GCs (Fig. 5D).

We took advantage of this strategy to test the effect of two different GR mutants: one lacking DNA-binding capabilities (GR-DBD) (16) and one lacking transcriptional capabilities through its transactivation domain, AF1 (GR-AF1) (17). The mutants GR-DBD and GR-AF1 were introduced into the same lentiviral vector backbone carrying the shRNA-GR to produce virus nos. 4 and 5, respectively. Expression of the human GR-DBD or GR-AF1 mutants did not compensate for the loss of the endogenous wild-type GR (Fig. 5D).

Expression of a well characterized GC-responsive gene, serum- and GC-induced kinase (SGK1), was next examined to assess a functional response. In contrast to the human or rat wild-type GR, the expression of the GR-DBD or GR-AF1 mutants did not increase the levels of SGK1 in response to corticosterone (Fig. 5C). Therefore, we conclude that the GR DNA-binding and AF1 domains are essential for the ability of GCs to induce Trk phosphorylation.

Discussion

We have shown that the activation of GR by GCs elicits Trk tyrosine phosphorylation, signaling, and trophic properties in vitro. In young rats, activation of GR through elevation of endogenous GCs levels by metyrapone or administration of a synthetic GC (dexamethasone) correlated with TrkB phosphorylation in the DG of the hippocampus and the subventricular zone of the striatum, two established sites for neurogenesis. Both GR and phospho-TrkB were detected in all hippocampal subfields, and the pronounced expression was found within the SGZ and GCL, where neural progenitors and mature DG neurons reside, respectively. Astroglia (GFAP), neurons (NeuN), neural precursors (nestin) coexpressed phosphorylated TrkB, suggesting a role for TrkB in newborn DG cells. Expression of full-length TrkB is observed in neurons and some glia, which is consistent with other reports (18).

Several lines of evidence indicate that Trk activation required GR genomic actions. First, a nonpermeable GC analog (BSA corticosterone) did not induce Trk phosphorylation, suggesting that a GC cell surface receptor is not involved. Second, the relatively slow time course of Trk activation and the sensitivity to gene expression inhibitors further excluded a rapid nongenomic effect of GCs. Third, pharmacological blockade and knockdown of the GR by shRNA prevented the activation of Trk receptors. Fourth, rescue of Trk activation depended on GR DNA binding and AF1 domains; both are typically required for genomic effects by GCs. Therefore, Trk activation occurred as a consequence of GR genomic effects.

We found that the levels of the neurotrophins NGF, BDNF, and NT3 were unchanged in the hippocampi of Dex-treated animals. Furthermore, the amounts of NGF, BDNF, and NT3 released in the media from cultured neurons in response to Dex also were unaffected. This finding suggests that TrkB activation by GCs is independent of neurotrophin production and release. Elevation of TrkB expression by GCs also might result in Trk phosphorylation. We therefore evaluated the effect of GCs on TrkB expression by using a TrkB luciferase-promoter assay in transfected cortical neurons. However, there was no effect of GCs on Trk promoter activity (SI Fig. 12). GC-induced Trk activation resembles Trk trans-activation by adenosine (19) and could potentially enhance the effects of adenosine to achieve Trk phosphorylation. The adenosine antagonist ZM241395 did not prevent Trk phosphorylation induced by GCs (data not shown). Therefore, Trk activation by GCs involves GR regulation of gene expression. Although beyond the scope of this article, we are pursuing the gene products regulated by GR to induce Trk activation.

In many ways, the protective effects afforded by GCs are unexpected. GCs are potent enhancers of apoptosis and can potentiate glutamate excitotoxicity in hippocampal neurons both in vitro and in vivo (20). Our experiments were conducted in the presence of MK801, a glutamate receptor antagonist, to minimize glutamate-induced cell death. In the absence of glutamate excitotoxicity, less than half of the GC trophic activity was suppressed by Trk inhibition, suggesting that additional pathways are involved. PI3K/Akt inhibition completely abolished GC trophic properties, indicating that other signaling pathways activated by GC converge to the PI3K–Akt pathway. Thus, trophic abilities of GCs may have been minimized. Of note, the expression of IAP2, an antiapoptotic protein, is up-regulated rapidly after the activation of GR (21). The activation of GR also can rescue granule cells from death during development (22). Because DG newborn neurons are not readily connected by excitatory inputs (23), we can speculate that DG newborn cells that express phosphorylated TrkB may be sensitive to GCs, but not to glutamate. An increase in TrkB activity by GCs may represent a mechanism to induce the differentiation of newborn cells in conjunction with the levels of GCs.

Stress, enriched living conditions, exercise, and training in learning paradigms are known to regulate GCs levels and BDNF signaling. Elevations of GCs from stressful or emotional experiences can alter plasticity in the nervous system. Exposure to prolonged high GC levels gives rise to impaired explicit learning and plasticity in the hippocampus (24), whereas transient elevation of GC levels may facilitate plasticity (25, 26). Indeed, BDNF also facilitates learning and memory (27). Receptors for neurotrophins and GCs have considerable overlap in their distribution in brain regions that are crucial for learning and memory (hippocampus and amygdala) (28, 29). In the amygdala, the blockade of GR impairs the storage and consolidation of new information (30), a response in which TrkB has been shown to participate (31). Neuronal proliferation, synaptogenesis, and dendritic arborization are regulated by both Trk and GR activation (20, 32–34). Therefore, the anatomical and functional overlap between GR and Trk receptors represents a basis for interplay between both systems.

The ability of GCs to activate Trk receptors provides one explanation for how these two signaling systems are interconnected. GCs also are commonly used for empirical treatment of spinal cord injury, Bell's palsy, and sudden sensorineural hearing loss, and it is interesting to speculate that Trk receptor activation by GCs could provide neuroprotective effects of this therapy. Future studies are necessary to determine the range of physiological responses that are regulated by this mechanism.

Materials and Methods

Animals.

Time-pregnant Sprague–Dawley rats (Charles River Laboratories) were housed individually, allowed ad libitum access to food and water, and maintained on a 12-h light/dark cycle. All protocols complied with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Young rats received a single systemic administration of dexamethasone 21-phosphate disodium salt (Sigma–Aldrich) or saline 0.9% as vehicle for 6 h at P18–P20. Injections were performed at 0700 for metyrapone (Sigma–Aldrich) or 40% propylene glycol (Sigma–Aldrich) in saline as vehicle and 1000 for dexamethasone. Animals were killed at 1600 hours.

Brain Tissue Processing.

The hippocampus and dorsal striatum were carefully dissected, frozen in liquid nitrogen, and homogenized using a tissue homogenizer. For explants, P9–P10 rat brains were sliced (200 μm) by using a vibratome. Cortexes were microdissected from several adjacent slices and pooled. One hemisphere of each brain served as an internal control to minimize interanimal variability. Slices were incubated at 35°C during treatment and then frozen and homogenized.

Reagents.

Hormones and antagonists were obtained from Sigma–Aldrich. K252a and LY294002 were purchased from Calbiochem and Biomol. NGF, NT3, and EGF were from PeproTech. FGF1 and heparin were a gift from C. Basilico (New York University). Rabbit polyclonal affinity purified phospho-TrkA and phospho-TrkB antibodies were developed against the phosphorylated residue Y794 and Y816, respectively (35, 36). Antibodies to the human GR and phospho-GR (S211-P) were obtained from M. Garabedian (New York University). Commercially available antibodies used were as follows: pan-Trk (C14, B3), rat GR (M20), p-tyrosine (Y99-P), Akt (B-1), MKP1 (M18), and Erk 1/2, DBC were from Santa Cruz Biotechnology; p-Trk (Y490-P), p-Akt (S473-P), p-ERK (T202/Y204-P), p-PLCγ1 (Y783-P), and PLCγ1 were from Cell Signaling Technology; Flag, BDNF, and actin were from Sigma–Aldrich; FGFR was from Zymed; GFP was from Molecular Probes; EGFR, TrkB, and SGK1 were from Upstate Biotechnology; NeuN, nestin, and GFAP were from Chemicon; and GAPDH was from Biodesign.

Plasmids.

The coding sequences for the human 11β-HSD isoforms 1 and 2 (obtained from A. Odermatt, University Hospital, Bern, Switzerland) and the rat GRα (a gift from D. Pearce, University of California, San Francisco) were subcloned into FCIV1, a lentiviral plasmid with a GFP reporter. We designed an shRNA targeting the rat GR sequence in the pLentiLox3.7 (ATCC) plasmid carrying a GFP reporter: forward sequence, 5′-tgcgggagaagatgatccattcttcaagagagaatggatcatcttctcccgcttttttgaattc; reverse sequence, 5′-tcgagaattcaaaaaagcgggagaagatgatccattctctcttgaagaatggatcatcttctcccgca. A mismatch shRNA was used as control: forward sequence, 5′-tcagtggagatcaacgtgcaagttttcaagagaaacttgcacgttgatctccactgttttttgaattc; reverse sequence, 5′-tcgagaattcaaaaaacagtggagatcaacgtgcaagtttctcttgaaaacttgcacgttgatctccactga. To replace the GR protein, the GFP sequence from the pLentilox3.7 backbone bearing the GR shRNA was replaced by the human GR wild-type sequence or mutants.

Cell Cultures.

Cultured neurons from embryonic day 18 rats were prepared from timed-pregnant Sprague–Dawley rats as described previously (19). Primary neurons were cultured on poly-d-lysine, maintained in Neurobasal media containing B27 supplement, 0.5 mM l-glutamine, 5-fluoro-uridine, and uridine (10 mM each). Experiments were conducted 7–12 days after plating. During starvation from B27, 1 μM MK-801 was added to decrease the contribution of N-methyl-d-aspartate-mediated cell death. For survival assays, cells were fixed in paraformaldehyde (PFA) and stained with Hoechst-33342 to count the number of pyknotic nuclei. PC12 cells were maintained in DMEM containing 10% FBS and 5% horse serum supplemented with 2 mM glutamine plus 200 μg·ml⁻¹ G418. Cells were starved from serum overnight before experiments. 293FT cells (Invitrogen) were grown in DMEM containing 10% FBS plus 200 μg·ml⁻¹ G418. Transfections were performed by using Lipofectamine 2000 (Invitrogen).

Lysis.

Tissues homogenates or cultured cells were lysed as described previously (35). Clarified lysates were subjected to immunoprecipitation with the indicated antibody overnight at 4°C, followed by incubation with protein A-Sepharose beads (Amersham). The beads were washed five times with lysis buffer. Samples were boiled in SDS buffer and loaded on SDS/PAGE gels for Western blot analysis.

ELISA.

E_max immunoassay system (Promega) using a protocol of direct acid treatment for protein samples to improve sensitivity (37) was used to determine neurotrophin protein levels. For culture media from primary neurons, large dense cultures (10 × 10⁶ cells per plate) were prepared. Collected media were concentrated from a pool of 15 ml to ≈700 μl by using Centriprep YM-10 (Millipore). Plasma corticosterone levels were measured by using an ELISA kit (AssayPro). Trunk blood was collected after decapitation in EDTA-coated tubes (BD Vacutainer). Optical absorbance was read at 450 nm with a microplate reader.

Immunocytochemistry.

Serum starved for 5 h, cells were fixed after treatments in PFA, blocked in 5% BSA/0.1% Triton X-100/PBS for 30 min at 25°C, and incubated in blocking buffer overnight at 4°C with primary antibodies. Fluorescent-coupled secondary antibodies (Jackson ImmunoResearch) were incubated in blocking buffer for 30 min at 25°C. Cells were mounted in Vectashield containing DAPI (Vector Laboratories). Epifluorescence was captured by using a Nikon Eclipse E800 microscope and imaged with a Zeiss AxioCam HRc digital camera.

Immunohistochemistry.

Rats were perfused with 4% PFA, and brains were postfixed for 1 h and equilibrated in 30% sucrose. Free-floating coronal sections rinsed in PBS were blocked in 5% normal goat serum/5% normal horse serum/PBS/0.1% triton X100 for 1 h at 25°C. Primary antibodies were incubated for 48 h at 4°C with shaking. Alexa Fluor-conjugated secondary antibodies (Molecular Probes) were incubated for 90 min at 25°C. The p-TrkB staining was detected with the biotin-streptavidin amplification system (Molecular Probes). Sections mounted in Prolong Gold Antifade reagent (Invitrogen) were imaged by using a Zeiss 510 confocal microscope.

Lentivirus.

Viruses were produced by transfecting packaging plasmids into 293FT cells. The set of plasmids used were Δ8.9, pCMV-VSVG, and FCIV1 for overexpression (38) and pLP1, pLP2, pLP/VSVG, and pLentiLox3.7 for shRNAs. Media were collected after 48 h and diluted 1:4 with regular culture medium to infect cells.

Statistical Analysis.

Optical densities were measured by using NIH ImageJ software. Quantified data are presented as mean ± SEM and analyzed by the nonlinear regression curve-fitting program, GraphPad PRISM. A one-way ANOVA with post hoc Bonferroni's and Newmann–Keuls's test was used (*, P < 0.05; **, P < 0.01; ***, P < 0.001).