The use of animals was approved by the Animal Care and Use Committee of the Southern Medical University and all protocols were carried out in accordance with the NIH Guidelines for the Care and Use of Laboratory Animals. Primary hippocampal neurons were isolated from neonatal Sprague-Dawley rats as previously described [13]. Briefly, rats (male:female = 1:1) on postnatal day 1 were obtained from the Experimental Animal Center of Southern Medical University. The pups were euthanized by decapitation and the head was separated from the body. The intact brain was immediately removed and quickly placed into dissection medium. The hippocampus was carefully separated under a microscope and placed in pre-chilled D-Hank’s solution (8 g NaCl, 0.4 g KCl, 0.06 g KH${}_2$PO${}_4$, 0.35 g NaHCO${}_3$, 0.12 g Na${}_2$HPO${}_4$•12H${}_2$O in 1 L ddH${}_2$O; Guangdong Guanghua Sci-Tech Co., Ltd., Guangzhou, China). Subsequently, the hippocampus was cut into small pieces and digested with 0.25% trypsin-EDTA (1:5 v/v; Gibco, Grand Island, NY, USA) at 37 ${}^{\circ }$C for 10 min. DMEM/F12 medium (Gibco) containing 10% fetal bovine serum (FBS; Gibco) and 10% newborn calf serum (NBCS; Gibco) were added to stop digestion. Cells were centrifuged at 80$\times$ g for 5 min at room temperature. The supernatant was discarded, and cell pellets were resuspended in Neurobasal-A medium (Gibco, Grand Island, NY, USA) supplemented with 2% B27 (Gibco), 1% 100$\times$ GlutaMAX-I (Gibco), 25 $\mu$M Glutamate (Genview, League City, TX, USA), and 0.5 $\mu$g/L Plasmocin (InvivoGen, San Diego, CA, USA). All cells were cultured in a 5% CO${}_2$ incubator at 37 ${}^{\circ }$C.

Primary rat hippocampal neurons were transfected with designated plasmid vectors using the Amaxa™ Rat Neuron Nucleofector Kit (Cat. # VPG-1003, Lonza, Basel, Switzerland) as previously reported [14]. Plasmid vectors carrying a constitutively active cdc42 mutant (pIRES-EGFP-cdc42F28L) were gifts from Prof. Andreas Püschel (Westfälische Wilhelms-Universität Münster, Germany). Plasmid vectors carrying a dominant negative cdc42 mutant (pIRES-EGFP-cdc42N17) were gifts from Prof. Lin Zhang (Southern Medical University, China). Plasmid vectors carrying a wild-type GSK-3$\beta$ (pcDNA3.1-GSK-3$\beta$ wild type) or a constitutively active GSK3$\beta$ mutant (pcDNA3.1-GSK-3$\beta$ S9A) were gifts from Prof. James R. Woodgett (University of Toronto, Canada). These plasmids were validated in our previous study [10]. Cells transfected with empty vectors (pIRES-EGFP or pcDNA3.1) were used as the controls. The culture medium was replaced with Neurobasal-A medium containing 2% B27, 1% 100$\times$ GlutaMAX-I, 25 $\mu$M Glutamate, and 0.5 $\mu$g/L Plasmocin at 4 and 24 h after transfection.

At 24-h post-transfection, rat hippocampal neurons were treated with or without phosphate-buffered saline (PBS), 2 mmol/L NaCl (diluted in cultured medium), or 2 mmol/L LiCl (diluted in cultured medium) for an additional 48 h.

GSK-3$\beta$ activity in primary rat hippocampal neurons following different treatments was measured using the GENMED kit (Genmed Scientifics Inc., Wilmington, DE, USA) as reported previously [15]. In brief, cells were harvested and centrifuged at 16,000$\times$ g for 5 min at 4 ${}^{\circ }$C. The supernatant was collected, and the protein concentration was determined using the BCA assay (Pierce, USA). Protein samples in triplicates (10 $\mu$g each) were added to a 96-well plate and incubated with the reagents from the kit. Optical density was measured at 340 nm using the SpectraMax 190 microplate reader (Molecular Devices, San Jose,CA, USA). GSK-3$\beta$ activity in the control group (transfected with empty pIRES-EGFP vectors) was set as 1.

After treatment, rat hippocampal neurons were washed with PBS, fixed in 4% paraformaldehyde (Sigma-Aldrich, St. Louis, MO, USA) for 90 min, and then permeabilized in 1% Triton X-100 (Sigma-Aldrich) for 30 min. After blocking with 3% bovine serum albumin (Genview, League City, TX, USA) in Tris-buffered saline (Sigma-Aldrich) at room temperature for 1 h, neurons were probed with anti-Tau-1 antibody (1:200 dilution; cat no. MAB3420, Merck Millipore, Billerica, MA, USA) and anti-microtubule-associated protein 2 (MAP2; 1:1000 dilution; cat no. AB5622, Merck Millipore, USA) at 4 ${}^{\circ }$C overnight. Subsequently, cells were incubated with rhodamine Red™-X goat anti-mouse IgG (1:700 dilution; Invitrogen, USA) and goat polyclonal secondary antibody to rabbit IgG-H&L (Cy5), pre-adsorbed (1:1000 dilution; cat no. ab6564, Abcam, New Temtories, HK, China) for 2 h. The nuclei were stained with Hoechst 33258 (0.5 $\mu$g/mL in PBS, Sigma-Aldrich) for 3 min at room temperature. The fluorescent signal was observed under a confocal microscope (FV500; Olympus, Japan). Processes with Tau-1 immunoreactivity in their distal segments were counted as axons. The longest axon of individual neurons in each group (100 GFP-positive neurons per group) was traced to determine axon length. The number of axons per cell and the percentages of neurons with no axon, a single axon, or multiple axons in individual groups were calculated. The longest axon of each neuron was used to determine the axon length using the Image-Proplus software (Version 6.0; Media Cybernetics, Silver Springs, MD, USA). One hundred GFP-positive neurons per group were analyzed.

Data were analyzed using the SPSS software (Version 19.0; IBM, Armonk, NY, USA) and presented as mean $\pm$ standard deviation (SD) from three independent experiments. One-way ANOVA followed by Tukey’s test was used to compare data among groups. A p-value of less than 0.05 was defined as statistically significant.

Hippocampal neurons were isolated from neonatal rats and transfected with vectors expressing cdc42F28L (a constitutively active cdc42 mutant) alone or together with vectors carrying GSK-3$\beta$ wt, GSK-3$\beta$ S9A (a constitutively active GSK3$\beta$ mutant), or empty vectors (GSK-3$\beta$ vector). Cells transfected with empty pIRES-EGFP vectors served as the controls (the Vector group). After 72 h, GSK-3$\beta$ activity in different groups of cells was measured. Overexpression of cdc42F28L significantly decreased GSK-3$\beta$ activity in hippocampal neurons compared with the Vector group (p 0.01). Transfection with vectors carrying either GSK-3$\beta$ wt or GSK-3$\beta$ S9A significantly increased GSK-3$\beta$ activity in neurons overexpressing cdc42F28L (p 0.01). The delivery of empty GSK-3$\beta$ vector did not change GSK-3$\beta$ activity in cdc42F28L-overexpressing neurons (Fig. 1). These findings suggest that transfection of GSK-3$\beta$ wt or GSK-3$\beta$ S9A restored GSK-3$\beta$ activity in hippocampal neurons overexpressing cdc42F28L.

Fig. 1.

Effects of GSK-3$\beta$ wt or GSK-3$\beta$ S9A on GSK-3$\beta$ activity in rat hippocampal neurons overexpressing cdc42F28L. Primary rat hippocampal neurons were transfected with pIRES-EGFP vectors expressing cdc42F28L alone or together with pcDNA3.1 vectors carrying GSK-3$\beta$ wt, GSK-3$\beta$ S9A, or empty pcDNA3.1 vectors (GSK-3$\beta$ vector). Cells transfected with empty pIRES-EGFP vectors served as the controls (Vector). At 72-h post-transfection, GSK-3$\beta$ activity was measured. The result of the Vector group was set as 1. Data are shown as mean $\pm$ SD. **p 0.01 vs. the Vector group; ##p 0.01 vs. the Cdc42F28L group. a.u., absorbance unit.

To investigate whether transfection of GSK-3$\beta$ wt or GSK-3$\beta$ S9A could affect axonogenesis in hippocampal neurons with excessive expression of cdc42F28L, we quantified the number and length of neuronal axons using immunofluorescence. At 72 h after transfection, different groups of neurons were stained with Tau-1 for axons, MAP2 for dendrites, and Hoechst 33258 for nuclei. Cells were then observed under a laser confocal microscope. The distal of the axons and the proximal segments of the dendrites were positively stained with Tau-1 (red). The dendrites and the proximal segments of the axons were stained with MAP2 (purple). As neurons were cultured for only 3 days, the dendrites were relatively short (Fig. 2A). Compared with the Vector group, neurons overexpressing cdc42F28L showed significantly lower numbers of neurons with no axon (14.33% vs. 5.66%; p 0.01) or a single axon (74.55% vs. 51.67%; p 0.01) but had a higher number of neurons with multiple axons (11.67% vs. 42.67%; p 0.01) (Fig. 2B). The number of axons per neuron and the mean axon length of the Vector group were significantly lower than those of the Cdc42F28L group (0.98 vs. 1.58 and 105.47 $\mu$m vs. 159.80 $\mu$m, respectively; both p 0.01) (Fig. 2C,D). In neurons overexpressing cdc42F28L, transfection with vectors carrying GSK-3$\beta$ wt or GSK-3$\beta$ S9A significantly increased the number of neurons with no axon (21.00% and 45.67%, respectively; p 0.01 vs. the Cdc42F28L group) but decreased the number of neurons with multiple axons (6.33% and 2.66%, respectively; p 0.01 vs. the Cdc42F28L group) (Fig. 2B). GSK-3$\beta$ wt or GSK-3$\beta$ S9A also significantly decreased the number of axons per neuron (0.92 and 0.57, respectively; p 0.01 vs. the Cdc42F28L group) and the mean axon length (66.88 $\mu$m and 38.03 $\mu$m, respectively; p 0.01 vs. the Cdc42F28L group) compared with the Cdc42F28L group (Fig. 2C,D). There was no significant difference in the percentage of different subtypes of neurons, the number of axons per cell, or mean axon length between the Cdc42F28 and Cdc42F28 + GSK-3$\beta$ vector groups (p $\mathrm{>}$ 0.05) (Fig. 2B–D). These results support that GSK-3$\beta$ wt or GSK-3$\beta$ S9A attenuated Cdc42F28L overexpression-induced axonogenesis of rat hippocampal neurons.

Fig. 2.

Effects of GSK-3$\beta$ wt or GSK-3$\beta$ S9A on axonogenesis of rat hippocampal neurons with excessive cdc42 activity. Primary rat hippocampal neurons were transfected with pIRES-EGFP vectors expressing cdc42F28L alone or together with pcDNA3.1 vectors carrying GSK-3$\beta$ wt, GSK-3$\beta$ S9A, or empty pcDNA3.1 vectors (GSK-3$\beta$ vector) at a ratio of 1:3 [24]. Cells transfected with empty pIRES-EGFP vectors served as the controls (Vector). (A) Immunofluorescence analysis of neurons. Different groups of cultured neurons were stained with Tau-1, MAP2, and fluorescent secondary antibodies, followed by nuclei staining with Hoechst 33258. Images were captured under a laser confocal microscope. Green, GFP; red, Tau-1; purple, MAP2; blue, Hoechst. Bar = 50 $\mu$m. (B) Percentage of neurons with no axon, a single axon, and multiple axons. (C) Number of axons per cell. (D) Mean axon length. Data are shown as mean $\pm$ SD. **p 0.01 vs. the Vector group; ##p 0.01 vs. the Cdc42F28L group.

To determine if cdc42 promoted axonogenesis by inhibiting GSK-3$\beta$, we transfected primary rat hippocampal neurons with vectors expressing cdc42N17 (a dominant negative cdc42 mutant) or empty control vectors, and then treated the cells with or without PBS, NaCl, or LiCl (a GSK-3$\beta$ inhibitor) for 48 h. GSK-3$\beta$ activity assays showed that transfection with cdc42N17 significantly induced the kinase activity in hippocampal neurons. Treatment with LiCl, however, significantly decreased GSK-3$\beta$ activity in neurons overexpressing cdc42N17 (p 0.01). No significant difference was observed in the Cdc42N17, Cdc42N17 + PBS, and Cdc42N17 + NaCl groups (p $\mathrm{>}$ 0.05) (Fig. 3).

Fig. 3.

Effect of LiCl treatment on GSK-3$\beta$ activity in rat hippocampal neurons overexpressing cdc42N17. Primary rat hippocampal neurons were transfected with pIRES-EGFP vectors expressing cdc42N17 or empty pIRES-EGFP vectors (Vector) for 24 h, followed by treatment with or without PBS, NaCl, or LiCl for 48 h. GSK-3$\beta$ activity in different groups of neurons was measured. Data are shown as mean $\pm$ SD. **p 0.01 vs. the Vector group; ##p 0.01 vs. the Cdc42N17 group. a.u., absorbance unit.

We further analyzed the effects of GSK-3$\beta$ inhibition on the axonogenesis of rat hippocampal neurons. After transfection with vectors carrying cdc42N17 or empty control vectors, neurons were treated with or without PBS, NaCl, or LiCl for 48 h. Immunofluorescence staining of neurons with Tau-1 (for axons), MAP2 (for dendrites), and Hoechst (for nuclei) is shown in Fig. 4A. Compared with the Vector group, transfection with cdc42N17 significantly increased the number of neurons with no axon (14.00% vs. 35.34 %; p 0.01) and decreased the number of neurons with a single axon (71.00% vs. 60.33%; p 0.01) or multiple axons (15.00% vs. 4.33%; p 0.01) (Fig. 4B). The number of axons per neuron and the mean axon length of the Cdc42N17 group were significantly lower compared to the controls (0.68 vs. 1.05 and 40.81 $\mu$m vs. 113.33 $\mu$m, respectively; both p 0.01) (Fig. 4C,D). Inhibition of GSK-3$\beta$ by LiCl treatment effectively reduced the number of axon-free neurons (2.00%) and neurons with a single axon (46.00%) but increased the number of neurons with multiple axons (52.00%) compared with the Cdc42N17 group (all p 0.01) (Fig. 4B). Neurons exposed to LiCl also exhibited a significantly higher number of axons per cell (1.60) and mean axon length (147.10 $\mu$m) compared to the Cdc42N17 group (both p 0.01) (Fig. 4C,D). There was no significant difference in the percentage of different subtypes of neurons, the number of axons per cell, or mean axon length between the Cdc42N17, Cdc42N17 + PBS, and Cdc42N17 + NaCl groups (p $\mathrm{>}$ 0.05) (Fig. 4B–D). The above data demonstrate that inhibition of GSK-3$\beta$ by LiCl abolished the inhibitory effect of cdc42N17 overexpression on axonogenesis of rat hippocampal neurons.

Fig. 4.

Effect of LiCl treatment on the inhibitory effect of cdc42N17 overexpression on axonogenesis. Primary rat hippocampal neurons were transfected with pIRES-EGFP vectors expressing cdc42N17 or empty pIRES-EGFP vectors (Vector) for 24 h, and then treated with or without PBS, NaCl, or LiCl for 48 h. (A) Immunofluorescence analysis of neurons. Different groups of cultured neurons were stained with Tau-1, MAP2, and fluorescent secondary antibodies, followed by nuclei staining with Hoechst 33258. Images were captured under a laser confocal microscope. Green, GFP; red, Tau-1; purple, MAP2; blue, Hoechst. Bar = 50 $\mu$m. (B) Percentage of neurons with no axon, a single axon, and multiple axons. (C) Number of axons per cell. (D) Mean axon length. Data are shown as mean $\pm$ SD. *p 0.05 vs. the Vector group; **p 0.01 vs. the Vector group; ##p 0.01 vs. the Cdc42N17 group.