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.

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).

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.

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.

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.

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).