Lithium promotes neural precursor cell proliferation: evidence for the involvement of the non-canonical GSK-3β-NF-AT signaling

Abstract

Lithium, a drug that has long been used to treat bipolar disorder and some other human pathogenesis, has recently been shown to stimulate neural precursor growth. However, the involved mechanism is not clear. Here, we show that lithium induces proliferation but not survival of neural precursor cells. Mechanistic studies suggest that the effect of lithium mainly involved activation of the transcription factor NF-AT and specific induction of a subset of proliferation-related genes. While NF-AT inactivation by specific inhibition of its upstream activator calcineurin antagonized the effect of lithium on the proliferation of neural precursor cells, specific inhibition of the NF-AT inhibitor GSK-3β, similar to lithium treatment, promoted neural precursor cell proliferation. One important function of lithium appeared to increase inhibitory phosphorylation of GSK-3β, leading to GSK-3β suppression and subsequent NF-AT activation. Moreover, lithium-induced proliferation of neural precursor cells was independent of its role in inositol depletion. These findings not only provide mechanistic insights into the clinical effects of lithium, but also suggest an alternative therapeutic strategy for bipolar disorder and other neural diseases by targeting the non-canonical GSK-3β-NF-AT signaling.

Figures

Figure 1

Lithium induced cell number increase in neural precursor cultures. (A) Lithium chloride dose dependently stimulated cell number increase in RG3.6 cell cultures. Cell count analysis was performed on cultures of RG3.6 cells treated with the indicated doses of LiCl for 7 days. *: Bonferroni/Dunn, P < 0.002 compared to no LiCl control. (B) Lithium chloride but not sodium chloride induced cell number increase in RG3.6 cell cultures. Cell count analysis was performed on cultures of RG3.6 cells treated with control medium, 3 mM NaCl or 3 mM LiCl for 5 days. *: paired t test, P < 0.01 compared to 3 mM NaCl or control medium treatment. (C) Lithium stimulated cell number increase in rat primary neural precursor cell cultures. Cell count analysis was performed on cultures of rat primary neural precursor cells treated with 3 mM NaCl or 3 mM LiCl for 5 days. *: Bonferroni/Dunn, P < 0.002. (D) Lithium stimulated cell number increase in mouse primary neural precursor cell cultures. Cell count analysis was performed on cultures of mouse primary neural precursor cells treated with 3 mM NaCl or 3 mM LiCl for 5 days. *: Bonferroni/Dunn, P < 0.002. The data represent mean ± standard error (n = 6 for each condition).

Figure 2

Lithium had no significant effect on percentage of GFP expressing live cells in RG3.6 cell cultures. (A) GFP-expressing cells are negative for propidium iodide (PI) staining. RG3.6 neurospheres were dissociated by trypsinization. The dissociated cells were stained with PI, a red fluorescent nuclear dye that specifically stain dead cells. The green GFP signal and the red PI staining were visualized and photographed using fluorescent microscope. (B) RG3.6 cells treated with NaCl or LiCl had similar GFP histograms. Flow cytometry analysis was performed on RG3.6 cell cultures treated with 3 mM NaCl or 3 mM LiCl for 3 days. Cells grown in non-FGF2 containing medium for 6 days were used as a negative control to gate GFP signal, since most cells grown in this condition were dead. (C) Quantification of the data shown in B. Lithium treatment had no significant effect on the percentage of GFP-expressing live cells in RG3.6 cell cultures.

Figure 3

Lithium significantly increased percentage of BrdU+ cells in RG3.6 cell cultures. RG3.6 cells were grown in culture medium containing 3 mM LiCl or control NaCl on laminin-coated coverslips for 3 days followed by 4 hours of BrdU labeling and subsequent BrdU immuno-staining. The data represent mean ± standard error. *: Bonferroni/Dunn, P < 0.0001, n = 8 for each condition.

Figure 4

Lithium had no significant effect on RNA expression of neurotrophic factors CNTF, GDNF, LIF, NGFβ, NGFγ and NT-3 in rat primary neural precursor cells. Rat primary neural precursor cells were treated with 3 mM LiCl or control NaCl for 3 days. Then the cells were lysed for RNA extraction and subsequent quantitative real time PCR analysis. RNA levels of neurotrophic factors were normalized to peptidylprolyl isomerase A (Ppia). The data represent mean ± standard error. *: Bonferroni/Dunn, P < 0.05, n = 3 for each condition.

Figure 5

Inositol had no significant effect on lithium-induced cell number increase in RG3.6 cell cultures. Cell count analysis was performed on RG3.6 cells cultured for 5 days in medium with or without 3 mM LiCl, and with or without 1 mM or 10 mM myo-inositol (n = 6 for each condition). The data represent mean ± standard error, and significance was determined with Fisher's PLSD post hoc test following analysis of variance (ANOVA).

Figure 6

Lithium inhibited GSK-3β in RG3.6 cells, and other GSK3β inhibitors mimicked lithium's effect on RG3.6 cell growth. (A) RG3.6 cells were grown in culture medium containing 3 mM LiCl or 3 mM control NaCl for 3 days, followed by western blotting analysis on P-GSK-3β (Ser9) and GSK-3β expression. (B) The blotting results in A were analyzed using ImageJ software (http://rsb.info.nih.gov/ij/, 1997-2007), and the ratio of P-GSK-3β was plotted. (C) Other GSK-3β inhibitors, like lithium, also increased cell numbers in RG3.6 cell cultures. Cell count analysis was performed on RG3.6 cells grown for 6 days in culture medium without or with 3 mM LiCl, 0.1% DMSO, 5 μM SB216763, or 25 μM SB415286 (n = 6 for each condition). DMSO was used as the vehicle control for SB216763 and SB415286, since these two drugs were dissolved in DMSO. The data represent mean ± standard error, and significance was determined with Bonferroni/Dunn post hoc analysis following ANOVA.

Figure 7

Calcineurin/NFAT inhibitor cyclosporin A antagonized lithium-induced cell number increase in RG3.6 cell cultures. (A) Lithium did not significantly change β-catenin expression in RG3.6 cells. RG3.6 cells were grown for 3 days in culture medium containing 3 mM LiCl or 3 mM control NaCl. The cells were then lysed in RIPA for western blotting analysis on β-catenin and β-actin expression. (B) Lithium increased nuclear expression of NF-AT. Aliquot of cells from A were used for cytoplamic and nuclear fractionation, followed by western blotting assays for the subcellular expressions of NF-AT. (C) Lithium stimulated transcriptional activation of NF-AT. RG3.6 cells treated with LiCl or NaCl were used for gene reporter assays. The luciferase activity is presented as fold induction relative to that of NaCl-treated cells. *: paired t test, P < 0.01. (D) Effect of different doses of cyclosporin A (CsA) on RG3.6 cell growth. Cell count analysis was performed on RG3.6 cells grown for 5 days in culture medium containing various doses of CsA (0 (Control), 0.01 μM, 0.1 μM, 1 μM, 10 μM, 50 μM, n = 6 for each dose). *: paired t test, P < 0.005 compared to control, 0.01 μM, 0.1 μM, or 1 μM CsA condition. (E) Cyclosporin A antagonized lithium-induced RG3.6 cell growth. Cell count analysis was performed on RG3.6 cells grown for 5 days in culture medium with or without 3 mM LiCl, and with or without 1 μM CsA (n = 6 for each condition). The data represent mean ± standard error. *: paired t test, P < 0.02 compared to No CsA + No LiCl, or CsA + No LiCl condition.