Inhibition of Sphingosine Kinase 1 Reduces Sphingosine-1-Phosphate and Exacerbates Amyloid-Beta-Induced Neuronal Cell Death in Mixed-Glial-Cell Culture

Abstract

In Alzheimer's disease (AD) pathology, the accumulation of amyloid-beta (Aβ), a main component of senile plaques, activates glial cells and causes neuroinflammation. Excessive neuroinflammation results in neuronal dropouts and finally produces the symptoms of AD. Recent studies suggest that disorder in sphingosine-1-phosphate (S1P) metabolism, especially the decreased expression of sphingosine kinase (SK)1, followed by the reduction in the amount of S1P, can be a promotive factor in AD onset. Thus, we explored the possibility that dysregulated S1P metabolism affects AD through the altered function in glial cells. We evaluated the effect of PF-543, a pharmacological inhibitor of SK1, on the inflammatory responses by lipopolysaccharide (LPS)-activated glial cells, microglia, and astrocytes. The treatment with PF-543 decreased the intracellular S1P content in glial cells. The PF-543 treatment enhanced the nitric oxide (NO) production in the LPS-treated neuron/glia mixed culture. Furthermore, we found that the augmented production of NO and reactive oxygen species (ROS) in the PF-543-treated astrocytes affected the microglial inflammatory responses through humoral factors in the experiment using an astrocyte-conditioned medium. The PF-543 treatment also decreased the microglial Aβ uptake and increased the number of injured neurons in the Aβ-treated neuron/glia mixed culture. These results suggest that a decrease in the glial S1P content can exacerbate neuroinflammation and neurodegeneration through altered glial cell functions.

Keywords: astrocytes; microglia; neuroinflammation; sphingosine-1-phosphate.

Scheme 1

Sphingolipid metabolism.

Figure 1

Effect of PF-543 on NO production by LPS-treated BV-2 microglia and primary astrocytes. Cells were treated with 10 ng/mL LPS together with various concentrations (10, 20, 30 μM, (a,b); BV-2, or 10, 20 μM, (c,d); astrocytes) of PF-543 for 24 h. Nitrite concentrations in the medium were measured using a fluorescence assay with the DAN reagent (a,c). Cell viability was evaluated using the MTT assay (b,d). Data are means ± standard errors of 4–9 samples. ** p < 0.01, significantly different from control; ^# p < 0.05, ^## p < 0.01, significantly different from LPS by one-way ANOVA followed by Tukey’s multiple-comparison procedure.

Figure 2

Effects of PF-543 on intracellular S1P content in LPS-stimulated BV-2 microglia and primary astrocytes. Cells were treated with 10 ng/mL LPS with or without 10 ((a); BV-2) or 20 ((b); astrocytes) μM of PF-543 for 1 h. Intracellular S1P content was measured by dot blot assay using anti-S1P antibody. Typical dots for S1P and β-actin protein are shown in the photograph. The graph shows S1P/β–actin ratio of the density of detection dots. Data are shown relative to control. Data are means ± standard errors of 4 (a) or 4–5 (b) samples. ** p < 0.01, significantly different from control; ^## p < 0.01, ^# p < 0.05, significantly different from LPS by one-way ANOVA followed by Tukey’s multiple-comparison procedure.

Figure 3

Effect of PF-543 on NO production in LPS-treated neuron/glia mixed culture. Neuron/glia mixed culture was treated with 10 ng/mL LPS in the presence or absence of 20 µM PF-543 for 24 h. Nitrite concentrations in the medium were measured using a fluorescence assay with the DAN reagent. Data are means ± standard errors of 4–5 samples. ** p < 0.01, significantly different from control; ^# p < 0.05, significantly different from LPS by one-way ANOVA followed by Tukey’s multiple-comparison procedure.

Figure 4

Effect of PF-543 on NO production by LPS-treated BV-2 microglia and primary astrocytes. Cells were treated with 10 ng/mL LPS with or without 10 ((a–c); BV-2) or 20 ((d–f); astrocytes) μM PF-543 for 24 h. Nitrite concentrations in the medium were measured using a fluorescence assay with the DAN reagent (a,d). Cell viability was evaluated using the MTT assay (c,f). Expression of iNOS protein was detected by immunoblotting (b,e). Typical bands for iNOS and β-actin protein are shown in the photograph. The graph shows iNOS/β–actin ratio of the density of detection bands. Data are shown relative to 10 ng LPS/mL in the immunoblotting analysis for BV-2 cells (d). Data are means ± standard errors of 4 samples. ** p < 0.01, significantly different from control (DMSO) (a,b) or control (d); ^## p < 0.01, significantly different from LPS (DMSO) (a,b) or LPS (d) by one-way ANOVA followed by Tukey’s multiple-comparison procedure; ⁺ p < 0.05, significantly different from LPS (f) by Student’s t test.

Figure 5

Effect of PF-543 on ROS generation in LPS-treated microglia and astrocytes. Cells were treated with 10 ng/mL LPS in the presence or absence of 10 ((a); BV-2, (b); primary microglia) or 20 ((c); astrocytes) μM PF-543 for 24 h. Intracellular ROS generation was determined using H₂DCFDA, a cell-permeable fluorescent dye. Data are shown relative to the control group. Data are mean ± standard errors of 4 (a), 3 (b), or 5 (c) samples. ** p < 0.01, * p < 0.05, significantly different from control (DMSO) (a) or control (b,c); ^## p < 0.01, ^# p < 0.05, significantly different from LPS (DMSO) (a) or LPS (b,c) by one-way ANOVA followed by Tukey’s multiple-comparison procedure.

Figure 6

Effect of astrocyte-conditioned medium from LPS- and PF-543-treated astrocytes on NO and ROS production in BV-2 microglia. Astrocyte-conditioned medium (ACM) was prepared from cultured astrocytes treated with 10 ng/mL LPS with or without 20 μM PF-543 for 24 h. BV-2 cells were incubated with ACM for 24 h. Nitrite concentrations in the medium were measured using a fluorescence assay with the DAN reagent (a). Intracellular ROS generation was determined using H₂DCFDA, a cell-permeable fluorescent dye (b). Data are shown relative to the control group in the ROS assay. Data are mean ± standard errors of 4 (a) or 5 (b) samples. ** p < 0.01, * p < 0.05, significantly different from control; ^# p < 0.05, significantly different from LPS by one-way ANOVA followed by Tukey’s multiple-comparison procedure.

Figure 7

Effect of PF-543 on glutamate uptake by astrocytes. Primary astrocytes were treated with 10 ng/mL LPS in the presence or absence of 20 μM PF-543 for 24 h. Glutamate uptake was assessed using [³H]-conjugated glutamate and liquid scintillation counter. Data are shown relative to the control group. Data are mean ± standard errors of 7 samples. ** p < 0.01, * p < 0.05, significantly different from control; ^## p < 0.01, significantly different from LPS by one-way ANOVA followed by Tukey’s multiple-comparison procedure.

Figure 8

Effect of PF-543 on Aβ uptake by microglia. (a) Primary microglia were treated with 10 ng/mL LPS with or without 10 μM PF-543 for 24 h. Aβ uptake by cells was measured using fluorescence-labeled Aβ_1-42. (b) Neuron/glia mixed culture was incubated with pre-aggregated Aβ_1-42 (100 nM at monomer concentration) in the presence or absence of 20 μM PF-543 for 48 h. The number of Aβ uptake microglia was counted after immunocytochemical staining. Arrowhead means microglia that engulfed Aβ. Data are shown relative to the control group (a), or as the ratio of the Aβ uptake microglia against the total microglia (b). Data are mean ± standard errors of 4 (a) or 3 (b) samples. (a) ** p < 0.01, significantly different from control; ^# p < 0.05, significantly different from LPS by one-way ANOVA followed by Tukey’s multiple-comparison procedure. (b) * p < 0.05, significantly different from Aβ by Student’s t test. Scale bar = 50 µm.

Figure 9

Effect of PF-543 on neuronal damage in neuron/glia mixed culture. Neuron/glia mixed culture was treated with pre-aggregated Aβ_1-42 (100 nM at monomer concentration) in the presence or absence of 20 μM PF-543 for 48 h. The number of damaged neurons (arrowhead) that had no axons was counted after immunocytochemical staining. Data are shown as the ratio of damaged neurons against total neurons. Data are mean ± standard errors of 4–7 samples. ** p < 0.01, significantly different from control; ^## p < 0.01, significantly different from Aβ by a Student’s t test. Scale bar = 50 µm.

Figure 10

Effect of PF-543 on microglial phagocytosis of neurons in neuron/glia mixed culture. Neuron/glia mixed culture was treated with pre-aggregated Aβ_1-42 (100 nM at monomer concentration) with or without 20 μM PF-543 for 48 h. The number of microglia that took in neurons or debris (arrowhead) was counted after immunocytochemical staining. Data are shown as the ratio of neuron-phagocytosed microglia against total microglia. Data are mean ± standard errors of 5–8 samples. * p < 0.05, significantly different from control; ^## p < 0.01, significantly different from Aβ by one-way ANOVA followed by Tukey’s multiple-comparison procedure. Scale bar = 50 µm.

Figure 11

Effect of PF-543 on mRNA expression of PPAR-γ and CD36 in BV-2 microglia. Cells were treated with pre-aggregated Aβ_1-42 (1 μM at monomer concentration) with or without 20 μM PF-543 for 24 h. The level of mRNA expression was determined by real-time quantitative PCR with the delta–delta Ct method. Data are shown relative to the control group. Data are mean ± standard errors of 6 (a) or 3 (b) samples. * p < 0.05, significantly different from control; ^## p < 0.01, ^# p < 0.05 significantly different from Aβ by one-way ANOVA followed by Tukey’s multiple-comparison procedure.