Ethyl pyruvate inhibits HMGB1 phosphorylation and release by chelating calcium

Abstract

Ethyl pyruvate (EP), a simple aliphatic ester of pyruvic acid, has been shown to have antiinflammatory effects and to confer protective effects in various pathological conditions. Recently, a number of studies have reported EP inhibits high mobility group box 1 (HMGB1) secretion and suggest this might contribute to its antiinflammatory effect. Since EP is used in a calcium-containing balanced salt solution (Ringer solution), we wondered if EP directly chelates Ca(2+) and if it is related to the EP-mediated suppression of HMGB1 release. Calcium imaging assays revealed that EP significantly and dose-dependently suppressed high K(+)-induced transient [Ca(2+)]i surges in primary cortical neurons and, similarly, fluorometric assays showed that EP directly scavenges Ca(2+) as the peak of fluorescence emission intensities of Mag-Fura-2 (a low-affinity Ca(2+) indicator) was shifted in the presence of EP at concentrations of ≥7 mmol/L. Furthermore, EP markedly suppressed the A23187-induced intracellular Ca(2+) surge in BV2 cells and, under this condition, A23187-induced activations of Ca(2+)-mediated kinases (protein kinase Cα and calcium/calmodulin-dependent protein kinase IV), HMGB1 phosphorylation and subsequent secretion of HMGB1 also were suppressed. (A23187 is a calcium ionophore and BV2 cells are a microglia cell line.) Moreover, the above-mentioned EP-mediated effects were obtained independent of cell death or survival, which suggests that they are direct effects of EP. Together, these results indicate that EP directly chelates Ca(2+), and that it is, at least in part, responsible for the suppression of HMGB1 release by EP.

Figure 1

Effects of EP on high K⁺-induced [Ca²⁺]_i increase in primary cortical neurons. (A) Paired transient Ca²⁺ responses of isolated primary cortical neurons were evoked by repetitive applications of high K⁺. Intracellular Ca²⁺ concentrations were presented as fluorescence ratios at 340/380 nm. High K⁺ (30 mmol/L) was applied for 10 s at the time indicated by arrows and it was followed by 5 min superfusion with Tyroid’s solution to wash out the initial K⁺ before the second K⁺ (30 mmol/L) application (n = 29). (B) Paired transient Ca²⁺ responses were measured after treating EP (5 mmol/L) for 3 min (bold line) before the second K⁺ (30 mmol/L) application (n = 60). (C) Effects of EP (0.5, 1, or 5 mmol/L) on high K⁺-evoked transients Ca²⁺ responses were measured. Results are presented as mean ± SEM and values at the top of each bar represent the number of cells. One-way ANOVA, Newman-Keuls post hoc test. **, difference from Con at p < 0.01; ^##, differences between indicated group at p < 0.01.

Figure 2

Excitation spectra of mag-fura-2 in the presence of Ca²⁺ or EP. (A) Excitation spectra of mag-fura-2 (10 μmol/L) were measured in the presence of CaCl₂. Excitation spectra of mag-fura-2 (10 μmol/L) with CaCl₂ (1 mmol/L) were measured in the presence of (B) EDTA, (C) EP, and (D) sodium pyruvate under cell free conditions by spectrofluorometer. The excitation spectrum detected at 510 nm was collected (γ_em = 510 nm, Slit = 15/20 nm, 1% T).

Figure 3

A23187-induced HMGB1 secretion in the absence of cell death. (A) Cell viabilities were determined using a MTT assay 24 h after 30 min of A23187 (1, 2.5 or 5 μmol/L) treatment in the presence of 1%, 2.5% or 5% of FBS. Results are presented as means ± SEMs (n = 3). (B–C) BV2 cells were treated with 1 or 2.5 μmol/L of A23187 in the presence of 5% FBS for 30 min and (B) LDH release, (C) nitrite production and (C) iNOS expression were measured 24 h later. Results are presented as means ± SEMs (n = 4). (D) BV2 cells were treated with 1 or 2.5 μmol/L of A23187 in the presence of 5% FBS for 30 min and HMGB1 levels in culture media and in cell lysates were estimated by immunoblotting at 3, 6, 12 and 24 h after A23187 treatment.

Figure 4

Inhibition of A23187-induced HMGB1 secretion by EP. (A) BV2 cells were incubated with A23187 (2.5 μmol/L) and 2.5 or 5 mmol/L of EP for 30 min and HMGB1 levels in culture media, cytoplasm, and nuclei were estimated by immunoblotting at 24 h after A23187 treatment. Immunoblots obtained using anti-α-tubulin or anti-p62 antibodies were used as controls for cytosolic and nuclear fractions, respectively. (B) Similar experiments were carried out in RAW264.7 mouse macrophage cells using 5 mmol/L of EP.

Figure 5

Reduction of A23187-induced intracellular calcium influx by EP in BV2 cells. (A,B) Intracellular Ca²⁺ levels at 30 min, 60 min and 6 h after A23187 treatment (2.5 μmol/L, 30 min) in the presence or absence of 5 mmol/L EP were visualized by (A) confocal microscopy and were measured using a (B) fluorescent microplate reader. (C,D) BV2 cells were either co- or pretreated with EP and intracellular Ca²⁺ levels were measured. For pretreatment, BV2 cells were treated with 1, 2.5 or 5 mmol/L EP for 1 h. Media were then replaced with fresh DMEM containing A23187 (2.5 μmol/L) and incubated for 30 min. Intracellular Ca²⁺ levels at 1 h after EP treatment were visualized by (C) confocal microscopy and were measured using a (D) fluorescence microplate reader. (A,C) Representative images from four independent experiments are presented and data are presented as mean ± SEM (n = 12). One-way ANOVA, Newman-Keuls post hoc test. ^##, difference from Con at p < 0.01; ^§§, difference from A23187-treated group at p < 0.01; * and **, differences between indicated group at p < 0.05 and p < 0.01, respectively. Scale bar, 20 μm.

Figure 6

Inhibition of HMGB1 secretion by Ca²⁺ chelation in NMDA-treated primary cortical cultures and in A23187-treated BV2 cells. (A) Primary cortical cultures were cotreated with NMDA (100 μmol/L for 10 min or 50 μmol/L for 30 min) and EGTA (0. 5, 1 or 2 mmol/L) for 30 min or treated with NMDA after pretreating with BAPTA-AM (5, 10 or 20 μmol/L) for 30 min and HMGB1 levels in culture media were estimated by immunoblotting at 24 h after NMDA treatment. (B) BV2 cells were cotreated with A23187 (5 μmol/L) and EGTA (0.25, 0.5 or 1 mmol/L) for 30 min or treated with NMDA after pretreating with BAPTA-AM (5, 10 or 20 μmol/L) for 30 min and HMGB1 levels in culture media were estimated by immunoblotting at 24 h after A23187 treatment. Representative images from three independent experiments were presented.

Figure 7

Reductions in PKCα- and CaMKIV-mediated HMGB1 phosphorylations by EP in A23187-stimulated BV2 cells. (A) BV2 cells were treated with 2.5 μmol/L of A23187 (30 min) and nuclear levels of PCKα and CaMKIV were examined by immunoblotting at 1, 2, or 3 h after A23187 treatment. (B) BV2 cells were treated with 2.5 μmol/L of A23187 (30 min) with or without 5 mmol/L EP and nuclear levels of PCKα and CaMKIV were examined by immunoblotting at 30 min or 1 h after A23187 treatment. (C) Total cell lysates were collected at 1, 2 or 4 h after A23187 treatment (2.5 μmol/L, 30 min), immunoprecipitated for HMGB1, and then immunoblotted with anti-phospho-Ser antibody. (D) Total cell lysates were collected at 2 h after A23187 treatment (2.5 μmol/L, 30 min) in the presence of EP (1, 2.5 or 5 mmol/L), immunoprecipitated for HMGB1, and then immunoblotted with anti-phospho-Ser antibody. Amounts of input protein before immunoprecipitation were estimated by immunoblotting with anti-HMGB1 and anti-α-tubulin antibodies. (E) BV2 cells were treated with 2.5 μmol/L of A23187 and 5 mmol/L EP with 0.1 or 1 mmol/L of CaCl₂ for 30 min. Lysates were collected 2 h later, immunoprecipitated for HMGB1, and immunoblotted with anti-pSer antibody. Representative images from four independent experiments are presented (C–E) and data are presented as mean ± SEM (n = 4). One-way ANOVA, Newman-Keuls post hoc test. ** and ^##, difference from Con at p < 0.01; ^‡‡ and ^§§, differences between indicated group at p < 0.05 and p < 0.01, respectively.