Screening inducers of neuronal BDNF gene transcription using primary cortical cell cultures from BDNF-luciferase transgenic mice

Abstract

Brain-derived neurotrophic factor (BDNF) is a key player in synaptic plasticity, and consequently, learning and memory. Because of its fundamental role in numerous neurological functions in the central nervous system, BDNF has utility as a biomarker and drug target for neurodegenerative and neuropsychiatric disorders. Here, we generated a screening assay to mine inducers of Bdnf transcription in neuronal cells, using primary cultures of cortical cells prepared from a transgenic mouse strain, specifically, Bdnf-Luciferase transgenic (Bdnf-Luc) mice. We identified several active extracts from a library consisting of 120 herbal extracts. In particular, we focused on an active extract prepared from Ginseng Radix (GIN), and found that GIN activated endogenous Bdnf expression via cAMP-response element-binding protein-dependent transcription. Taken together, our current screening assay can be used for validating herbal extracts, food-derived agents, and chemical compounds for their ability to induce Bdnf expression in neurons. This method will be beneficial for screening of candidate drugs for ameliorating symptoms of neurological diseases associated with reduced Bdnf expression in the brain, as well as candidate inhibitors of aging-related cognitive decline.

Figure 1

Construction of a screening assay to examine activity of Bdnf transcription using primary cultures of Bdnf-Luc mouse cortical cells in a 96-well format. (a) Schematic of wild-type Bdnf and Bdnf-Luciferase on a BAC. Detailed information was described previously^,. (b) Schematic for validating the screening assay system. In 96-well culture plates, Bdnf-Luc mouse (Tg) cortical cells and wild-type (Wt) mouse cortical cells were seeded into wells along lines A–D and lines E–H, respectively. At 13 DIV, cells in wells of columns 2-11 were treated with a high concentration (final 25 mM) of KCl for 6 h. Cells in the wells of columns 1 and 12 were treated with PBS for 6 h. (c) Luciferase activity of each well (left) and the average of luciferase activity (right). Means ± SEM (n = 8 (5 mM KCl), or 40 (25 mM KCl)), ****p < 0.0001 vs. 5 mM KCl (unpaired t-test).

Figure 2

Screening activators of Bdnf transcription from a dopaminergic library. (a) Representative result obtained using a commercially available dopaminergic library. Each compound was added into Bdnf-Luc cortical cells at 13 DIV at a final concentration of 10, 100, or 1000 nM, and luciferase activity was measured in each well 6 h later. Arrowheads show active compounds (that increased luciferase activity by more than 2-fold). For compound names, see Supplementary Table S1. (b) Changes in Bdnf expression in the presence of dopamine D₁ agonist A68930 in primary cultures of rat cortical cells. At 13 DIV, cells were treated with A68930 for 1 h, and then total RNA was prepared for RT-PCR analysis. Means ± SEM (n = 3), ***p < 0.001 and ****p < 0.0001 vs. 0 nM (one-way ANOVA with Dunnett’s multiple comparisons test). (c) Effect of APV or FK506 on A68930-induced Bdnf expression in cultured rat cortical cells. APV (200 μM) or FK506 (5 μM) was added 10 min before the addition of A68930 (100 nM). Means ± SEM (n = 3), *p < 0.05 and ****p < 0.0001 vs. DMSO/vehicle, ^††††p < 0.0001 vs A68930/vehicle (two-way ANOVA with Tukey’s multiple comparisons test).

Figure 3

Screening activators of Bdnf transcription from an herbal extract library. Representative results obtained using an herbal extract library consisting of 120 herbal extracts. Each extract was added into Bdnf-Luc cortical cells at 13 DIV at a final concentration of 500 μg/mL. Luciferase activity was measured in each well at 6 h (a), 24 h (b), or 48 h (c) after addition of extract. Means ± SEM (n = 3). Active extracts (that increased Bdnf induction by more than 2-fold) are shown in red.

Figure 4

GIN induces endogenous Bdnf expression in primary cultures of rat cortical cells. (a,b) Time-course (a) and dose-dependency (b) of changes in Bdnf mRNA in the presence of GIN in primary cultures of rat cortical cells. (a) Cells were treated with 500 μg/mL GIN and total RNA was prepared at the indicated time. Means ± SEM (n = 3), **p < 0.01, ***p < 0.001, and ****p < 0.0001 vs. water at the same time point (two-way ANOVA with Tukey’s multiple comparisons test). (b) Cells were treated with different concentrations of GIN, and total RNA was prepared 3 h after treatment. Means ± SEM (n = 3), ****p < 0.0001 vs. 0 μg/mL (one-way ANOVA with Dunnett’s multiple comparisons test). (c) Bdnf-Luc cortical cells were seeded into a 96-well culture plate and cultured for 13 days. Cells were then treated with different concentrations of GIN, and luciferase activity in each well was measured 6 h after treatment. Means ± SEM (n = 3), *p < 0.05 and ****p < 0.0001 vs. 0 μg/mL (one-way ANOVA with Dunnett’s multiple comparisons test). (d) Relationship between changes in endogenous Bdnf mRNA expression levels (b) and those in luciferase activity (c) was analysed using a correlation coefficient test. Statistical analysis was performed by Pearson’s correlation coefficient test. (e) The effect of GIN on the expression of 5′ exon-specific Bdnf mRNA in cultured rat cortical cells. Cells were treated with 500 μg/mL GIN and total RNA prepared 3 h after the treatment. Means ± SEM (n = 3), **p < 0.01, ***p < 0.001, and ****p < 0.0001 vs. water (unpaired t-test). N.D.: not detected.

Figure 5

No contribution of ginsenosides or LPA receptors to GIN-induced Bdnf expression. (a) Representative result obtained using a series of ginsenosides consisting of eight types: Rb1, Rb2, Rc, Rd, Re, Rf, Rg1, and Rg2. Each ginsenoside was added into Bdnf-Luc cortical cells at 13 DIV at the indicated concentrations, and luciferase activity was measured in each well 6 h after addition. Means ± SEM (n = 3). (b) Effect of the LPA_1/2 receptor agonist, oleoyl-L-α-lysophosphatidic acid (LPA), on Bdnf expression in primary cultures of rat cortical cells. Cells were treated with LPA at the indicated concentrations, and total RNA was prepared 3 h after treatment. Means ± SEM (n = 3). (c) Effect of the LPA_1/3 receptor antagonist Ki 16425 on the GIN-induced Bdnf expression. Ki 16425 was added to cells at the indicated concentration 10 min before the addition of GIN (500 μg/mL). Total RNA was prepared 3 h after treatment to examine changes in Bdnf expression by RT-PCR. Means ± SEM (n = 3), ****p < 0.0001 vs. water in the absence or presence of Ki 16425, (two-way ANOVA with Tukey’s multiple comparisons test). (d) Effect of a combination of ginsenosides and/or LPA on luciferase activity in cultured Bdnf-Luc cortical cells (left). Major ginsenosides (ginsenoside Rb1, Rc, Rd, Re, and Rg1) in GIN (Supplementary Table S2, No. 36 (refer to database URL)) were mixed (final concentrations of each ginsenoside; 10, 50, or 100 μM). Mixed ginsenosides (G-Mix) was added into Bdnf-Luc cortical cells at 13 DIV with or without LPA (final concentration; 0, 10, 50, or 100 μM). GIN was used as a positive control (right). Luciferase activity in each well was measured 6 h after addition. Means ± SEM (n = 6–8). ****p < 0.0001 vs. water (unpaired t-test).

Figure 6

GIN-induced Bdnf transcription via CREB-dependent pathways. (a) Effects of blocking Ca²⁺ entry sites and Ca²⁺ signalling inhibitors on GIN-induced Bdnf expression in primary cultures of rat cortical cells. APV (200 μM), nicardipine (Nica, 5 μM), U0126 (20 μM), KN93 (10 μM), or FK506 (5 μM) were added 10 min before the addition of GIN (500 μg/mL). Total RNA was prepared 3 h after treatment to examine changes in Bdnf expression by RT-PCR. Means ± SEM (n = 3), ****p < 0.0001 vs. water, ^††††p < 0.0001 vs. DMSO/GIN (two-way ANOVA with Tukey’s multiple comparisons test). (b) Changes in Bdnf promoter IV (Bdnf-pIV) activity in the presence of GIN. Transfected cortical cells were treated with 500 μg/mL GIN for 6 h, and luciferase activity measured. Means ± SEM (n = 3), **p < 0.01 vs. Wild type/water, ^††p < 0.01 vs. Wild type/GIN (two-way ANOVA with Tukey’s multiple comparisons test). (c,d) (left) Representative images of phosphorylated CREB at serine 133rd (c) or subcellular localization of CRTC1 (d), and (right) percentage of phospho-CREB-positive (c) or nuclear CRTC1-positive (d) neurons. Cultured rat cortical cells were treated with 500 μg/mL GIN for 30 min, then cells were fixed for immunostaining. Scale bar = 20 μm. Means ± SEM (n = 3), ****p < 0.0001 vs. water (unpaired t-test).