NOX2 inhibition enables retention of the circadian clock in BV2 microglia and primary macrophages

Abstract

Introduction: Sustained neuroinflammation is a major contributor to the progression of neurodegenerative diseases such as Alzheimer's (AD) and Parkinson's (PD) diseases. Neuroinflammation, like other cellular processes, is affected by the circadian clock. Microglia, the resident immune cells in the brain, act as major contributors to neuroinflammation and are under the influence of the circadian clock. Microglial responses such as activation, recruitment, and cytokine expression are rhythmic in their response to various stimuli. While the link between circadian rhythms and neuroinflammation is clear, significant gaps remain in our understanding of this complex relationship. To gain a greater understanding of this relationship, the interaction between the microglial circadian clock and the enzyme NADPH Oxidase Isoform 2 (NOX2) was studied; NOX2 is essential for the production of reactive oxygen species (ROS) in oxidative stress, an integral characteristic of neuroinflammation.

Methods: BV2 microglia were examined over circadian time, demonstrating oscillations of the clock genes Per2 and Bmal1 and the NOX2 subunits gp91phox and p47phox.

Results: The BV2 microglial clock exerted significant control over NOX2 expression and inhibition of NOX2 enabled the microglia to retain a functional circadian clock while reducing levels of ROS and inflammatory cytokines. These trends were mirrored in mouse bone marrow-derived primary macrophages.

Conclusions: NOX2 plays a crucial role in the interaction between the circadian clock and the activation of microglia/macrophages into their pro-inflammatory state, which has important implications in the control of neuroinflammation.

Keywords: BV2 microglia; NOX2; circadian rhythm; neuroinflammation; oxidative stress; primary macrophages.

Figure 1

Clock genes Per2 and Bmal1 display circadian oscillations in resting BV2 microglia. (A) BV2 microglia were synchronized using the serum-shock protocol and samples were collected every 2 h starting at 16 h post serum-shock (HPS16) for RT-qPCR and Western blotting experiments. (B, C) ECHO fitted plots for mRNA expression of clock genes Per2 and Bmal1 in BV2 microglia. Data represented as fold change in expression using Hprt1 as a reference gene and HPS0 as a reference sample for the ΔΔCt method of data analysis (n = 3) (D) Time course samples obtained from serum-shock synchronized BV2 cells analyzed using Western blotting to measure the levels of PER2 and BMAL1. Amido black staining was used to normalize for protein loading. For complete blot images refer to Supplementary Information ( Supplementary Figures 3 , 4 ) (E, F) ECHO fitted plots for PER2 and BMAL1 protein levels in BV2 cells (n = 3). For all ECHO plots, the bold lines depict the ECHO fitted model and the shaded region represents ±1 standard deviation of model at each time point. Statistical significance was determined using ECHO. All plots had p < 0.05 for ECHO significance fit.

Figure 2

BV2 circadian clock oscillations are affected by the cell’s inflammatory status. ECHO fitted plots for mRNA expression of clock genes Per2 and Bmal1 in BV2 microglia (n = 3) under (A) LPS (1 µg/mL) (pro-inflammatory) and (B) IL-4 (20 ng/mL) (anti-inflammatory) activation. Data represented as fold change in expression using Hprt1 as a reference gene and HSP0 as a reference sample for the ΔΔCt method of data analysis. Bold line represent model fit with shaded region representing the standard deviation of model at each time point. All plots had p<0.05 for ECHO significance fit.

Figure 3

NOX2 sub-units p47^phox and gp91^phox are under the control of the BV2 microglial circadian clock. ECHO fitted plots for mRNA expression of NOX2 components gp91^phox and p47^phox in BV2 microglia (n = 3) under (A) resting, (B) LPS (1 µg/mL) (pro-inflammatory) and (C) IL-4 (20 ng/mL) (anti-inflammatory) activation. Data represented as fold change in expression using Hprt1 as a reference gene and HSP0 as a reference sample for the ΔΔCt method of data analysis. Bold line represent model fit with shaded region representing the standard deviation of model at each time point. All plots had p < 0.05 for ECHO significance fit.

Figure 4

NOX2 inhibition in BV2 microglia results in the reduction of ROS levels and affects phosphorylated-p47^phox levels. ROS levels in BV2 microglia with and without LPS and, NOX2 inhibitors apocynin and GSK2795039 determined using (A) Amplex Red Assay to measure extracellular hydrogen peroxide levels and (B) DCFDA Assay to measure intracellular ROS levels. (C) Western blotting analysis of phosphorylated-p47^phox levels in BV2 microglia in the presence of LPS and NOX2 inhibitors apocynin and GSK2795039 (n = 3 to 4). For all experiments, BV2 cells with no additives were used as control sample. Data are represented as mean ± SEM and analyzed using single-factor ANOVA test. ** denotes p<0.01, **** denotes p<0.0001 and *** denotes p<0.001. For complete blot images, refer to Supplementary Information ( Supplementary Figure 5 ).

Figure 5

Inhibition of NOX2 by GSK2795039 under LPS activation results in the retention of circadian oscillations in BV2 microglia. ECHO fitted plots for mRNA expression of clock genes (n = 3) (A) Per2 and (B) Bmal1, and NOX2 components (C) gp91^phox and (D) p47^phox in BV2 microglia in the presence of 1 μg/mL LPS and 25 μM GSK2795039. Data represented as fold change in expression using Hprt1 as a reference gene and HSP0 as a reference sample for the ΔΔCt method of data analysis. Bold line represent model fit with shaded region representing the standard deviation of model at each time point. All plots had p<0.05 for ECHO significance fit.

Figure 6

Inhibition of NOX2 by GSK2795039 under LPS activation conserves PER2 oscillation in mouse bone marrow-derived macrophages (BMDMs). Luminescence traces obtained luciferase measurement of serum-shock synchronized mouse BMDMs with and without LPS, and NOX2 inhibitors. (A) Cells were exposed to LPS with and without NOX2 inhibitors apocynin and GSK2795039 during the starve stage of the synchronization protocol. (B) Cells were exposed to LPS with and without NOX2 inhibitors apocynin and GSK2795039 post serum-shock stage of the synchronization protocol. Luciferase measurements were obtained by the LumiCycle HPS0 to HPS120 (n = 3). Data are represented as mean ± SEM.

Figure 7

NOX2 inhibition in BV2 microglia creates an inflammatory environment similar to anti-inflammatory activation. (A) TNF-α and (B) IL-6 (pro-inflammatory cytokines) levels in BV2 cells (n = 3) with and without LPS and NOX2 inhibitors apocynin and GSK2795039 measured using ELISA. (C) TNF-α and (D) IL-6 (pro-inflammatory cytokines) levels in BV2 cells (n = 3) under IL-4 activation with and without LPS measured using ELISA. Data are represented as mean ± SEM and analyzed using single-factor ANOVA test. * denotes p < 0.05 and **** denotes p < 0.0001.

Figure 8

NOX2 inhibition affects the levels of inflammation associated cytokines in mouse bone marrow-derived macrophages. (A) TNF-α and (B) IL-6 (pro-inflammatory cytokines), and (C) IL-4 (anti-inflammatory cytokine) levels in BMDMs (n = 3) with and without LPS and NOX2 inhibitors apocynin and GSK2795039 measured using ELISA. Based on ELISA data, we did not see any IL-4 secretion from BV2 microglia. This could be because of the immortalization process. Data are represented as mean ± SEM and analyzed using single-factor ANOVA test. ** denotes p < 0.01, *** denotes p < 0.001 and **** denotes p < 0.0001.