Mouse microglial cell line MG6 cells (RCB2403) were obtained from RIKEN BRC (Tsukuba, Japan) and cultured as described previously (Kawakami et al., 2018). Cells were maintained until 70% confluency in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum, 10 μg/ml insulin, 10 μM 2-mercaptoethanol, 100 U/ml penicillin, and 100 μg/ml streptomycin at 37^°C in a 5% CO₂ incubator. Phosphorothioate-modified oligodeoxynucleotide (ODN) 1668 and control ODN were synthesized by Hokkaido system science (Hokkaido, Japan). The ODN1668 sequence was TCCATGACGTTCCTGATGCT and used as CpG-ODN. The control ODN (CTL-ODN) sequence was TCCATGAGCTTCCTGATGCT. GRK2 inhibitor (GRK2i), methyl 5-[2-(5-nitro-2-furyl)vinyl]-2-furoate was purchased from Calbiochem (San Diego, CA, United States). Stattic was obtained from Abcam (Cambridge, United Kingdom). BAY11-7082 and MG-132 were purchased from Sigma (Sigma, St. Louis, MO, United States).

All small interfering RNAs (siRNAs) were purchased from Sigma-Aldrich (St. Louis, MO, United States). MISSION siRNA universal negative control (SIC-001) was employed as the negative control in this study. GRK2 siRNAs, signal transducers and activators of transcription 1 (STAT1) siRNAs, IFN regulatory factor 1 (IRF1) siRNAs, TIR-domain-containing adaptor-inducing interferon-β (TRIF) siRNAs, STAT3 siRNAs, interferon alpha and beta receptor subunit 1 (IFNAR1) siRNAs and IRF3 siRNAs were transfected at a final concentration of 60, 15, 50, 50, 40, 40, and 50 nM using lipofectamine RNAiMAX (Life Technologies, Carlsbad, CA, United States) according to the manufacturer’s protocol, respectively.

Cells were harvested and lysed in 300 μl of Radio-Immunoprecipitation Assay (RIPA) buffer (Thermo Fisher Scientific, Rockford, IL, United States) containing protease inhibitor cocktail (Nacalai Tesque, Kyoto, Japan) on ice. The lysates were centrifuged at 18,000 ×g for 10 min at 4°C and the resulting supernatants were reserved. The supernatant proteins were quantified using BCA Protein Assay Kit (Thermo Fisher Scientific). Samples (20 μg of protein) were run on 10% polyacrylamide gel and electrotransferred onto polyvinylidene fluoride filter membrane. The membrane was blocked for 60 min at room temperature in 1% bovine serum albumin in Tris-buffered saline containing Tween 20, followed by overnight incubation with primary antibody, anti-iNOS rabbit polyclonal antibody (1:1,000; Cell Signaling, Danvers, MA, United States), anti-STAT1 mouse monoclonal antibody (1:300; Santa Cruz Biotechnology, Santa Cruz, CA, United States), anti-phospho-STAT1 (Tyr-701) mouse monoclonal antibody (1:300; Santa Cruz Biotechnology), anti-GRK2 rabbit polyclonal antibody (1:500; Santa Cruz Biotechnology), anti-STAT3 mouse monoclonal antibody (1:1000; Cell Signaling), anti-phospho-STAT3 (Tyr-705) rabbit monoclonal antibody (1:1000; Cell Signaling), anti-IRF1 rabbit monoclonal antibody (1:1000; Cell Signaling), anti-lamin B1 rabbit polyclonal antibody (1:3000; Proteintech, Rosemont, IL, United States), or anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mouse monoclonal antibody (1:10000; Wako Pure Chemical, Osaka, Japan), at 4°C. Primary antibody detection was performed with horseradish peroxidase-conjugated secondary antibodies. Binding of the antibody was detected by an enhanced chemiluminescence (ECL) Plus chemiluminescent system (GE Healthcare, Tokyo, Japan) and levels of protein expression were quantified by a lumino image LAS-4000 analyzer (Fuji Film, Tokyo, Japan). Additional details are described by our laboratory (Sakata et al., 2015; Abdelzaher et al., 2016; Ohashi et al., 2017; Kawakami et al., 2018).

Total RNA was isolated from cells with the use of Sepazol-RNA I Super G (Nacalai Tesque) according to the manufacturer’s manual. ReverTra Ace qPCR RT Master Mix (Toyobo, Osaka, Japan) was used for the reverse transcription reaction, and quantitative PCR analyses were performed using PowerUp^TM SYBR^® Green Master Mix (Thermo Fisher Scientific), as described in the manufacturers’ instructions. Values were normalized to the housekeeping gene GAPDH according to the manufacturer’s protocol (MX3000P real-time PCR system; Agilent Technologies Inc., Santa Clara, CA, United States). The IFN-β primer sequences were 5′-CAGCTCCAAGAAAGGACGAAC-3′ (sense) and 5′- GGCAGTGTAACTCTTCTGCAT-3′ (antisense), the iNOS primer sequences were 5′-TAGGCAGAGATTGGAGGCCTTG-3′ (sense) and 5′-GGGTTGTTGCTGAACTTCCAGTC-3′ (antisense), the IRF1 primer sequences were 5′-ATGCCAATCACTCGAATGCG-3′ (sense) and 5′-TTGTATCGGCCTGTGTGAATG-3′ (antisense), the interferon-γ-inducible 10 kD protein (IP10) primer sequences were 5′-CCAAGTGCTGCCGTCATTTTC-3′ (sense) and 5′-GGCTCGCAGGGATGATTTCAA-3′ (antisense), the interleukin (IL) -6 primer sequences were 5′-CCACTTCACAAGTCGGAGGCTTA-3′ (sense) and 5′-GCAAGTGCATCATCGTTGTTCATAC-3′ (antisense), the IL-1β primer sequences were 5′-TCCAGGATGAGGACATGAGCAC-3′ (sense) and 5′-GAACGTCACACACCAGCAGGTTA-3′ (antisense), the IRF7 primer sequences were 5′-GAGACTGGCTATTGGGGGAG-3′ (sense) and 5′-GACCGAAATGCTTCCAGGG-3′ (antisense), and the GAPDH primer sequences were 5′-TGTGTCCGTCGTGGATCTGA-3′ (sense) and 5′-TTGCTGTTGAAGTCGCAGGAG-3′ (antisense). Additional details are described elsewhere (Kawakami et al., 2018; Yamashita et al., 2018).

Culture medium levels of IFN-β were measured by the use of commercially available enzyme-linked immunosorbent assay (ELISA) kit (Mouse IFNβ DuoSet ELISA kit, DY8234-05; R&D Systems, Minneapolis, MN, United States) according to the manufacturer’s instructions. The plate was read on a microplate reader (Molecular Devices, Menlo Park, CA, United States). Assays were performed in duplicate.

Microglial cells were stimulated with 100 ng/ml of LPS for 3 h. The cells were incubated with allophycocyanin (APC)-anti-TLR4 (Clone: SA15-21) for 15 min on ice and then stained with SYTOX-ADDvanced. Fluorescence-activated cell scanning (FACS) analysis by flow cytometry was performed on an Accuri C6 (Becton Dickinson, Franklin Lakes, NJ, United States). The FACS pattern obtained from SYTOX-ADD-negative cells is indicated.

Values are presented as mean ± SEM. Data were analyzed by the use of Prism software (version 6; GraphPad Software, San Diego, CA, United States). Statistical significance was calculated by Student’s unpaired t-test, where differences at p < 0.05 was considered statistically significant.

When LPS (100 ng/ml) was applied to the mouse microglial cell line MG6 cells, the expression levels of iNOS protein were increased in a time-dependent manner, reaching a peak at 15 h after LPS and declining thereafter (Figure 1A,B). The induction of iNOS expression in glial cells and macrophages involves the activation of the Janus tyrosine kinase (JAK)/STAT signaling pathway (Dell’Albani et al., 2001; Ganster et al., 2001). We thus ascertained whether STAT1 and STAT3 can be activated in MG6 cells following LPS challenge. Activation of STAT1 and STAT3 was assessed by Western blot analysis for phospho-STAT1 at Tyr-701 and phospho-STAT3 at Tyr-705, respectively. STAT1 phosphorylation was transiently but markedly increased in cells at 6 h after LPS, whereas a sustained rise in total STAT1 levels was also observed with LPS stimulation (Figure 1A,C). On the other hand, the increase in STAT3 phosphorylation showed a marked peak at 6 h after LPS and a less pronounced sustained response, but total STAT3 levels were unaffected by LPS challenge (Figure 1A,D).

The knockdown of STAT1 was conducted in MG6 cells using its specific siRNAs. Our transfection of STAT1 siRNAs effectively silenced STAT1 expression levels in LPS-stimulated microglial cells (Figure 2A). Transfection of STAT1 siRNAs evidently but incompletely prevented the LPS-induced increase in iNOS protein expression (Figure 2A). Stattic is known to be a small-molecule inhibitor of STAT3 activation, dimerization, and nuclear translocation (Schust et al., 2006). Treatment with stattic reduced iNOS protein expression in LPS-stimulated cells in a concentration-dependent manner (Figure 2B). At a concentration of 2 μM, stattic completely eliminated the LPS-induced iNOS upregulation, but also abolished the total and phosphorylation levels of STAT1 following LPS application (Figure 2B,C), suggesting that stattic at this concentration appears to act on STAT1 as well as STAT3. Combined treatment with STAT1 siRNAs and stattic at a lower concentration (1 μM) resulted in a complete abolition of iNOS expression in LPS-stimulated cells (Figure 2C). We also found that iNOS expression following LPS challenge was inhibited when STAT3 siRNAs were transfected (Figure 2D). These results indicate that both STAT1 and STAT3 play a role in iNOS expression triggered by LPS in microglial cells.

To determine whether GRK2 can be involved in STAT1 and STAT3 phosphorylation in LPS-stimulated microglial cells, GRK2 siRNAs were used to knockdown microglial expression of GRK2. The ablation of GRK2 by siRNAs resulted in a significant inhibition of phosphorylation levels of STAT1 and STAT3 (Figure 3A,B). Stimulation of MG6 cells with LPS led to the translocation of phosphorylated STAT1 and STAT3 into the nucleus (Figure 3C). The nuclear translocation of phosphorylated STAT1 and STAT3 was dampened when GRK2 siRNAs were transfected. These findings suggest that GRK2 positively regulates activation of STAT1 and STAT3 in microglial cells.

As presented above, both STAT1 and STAT3 are critical regulators of iNOS expression in LPS-stimulated microglial cells. In line with our recent report (Kawakami et al., 2018), the LPS-induced increases in iNOS mRNAs were strongly prevented by transfection of GRK2 siRNAs (Figure 3D). In addition, GRK2 siRNA transfection greatly reduced the LPS-induced upregulation of mRNA levels of IL-1β, IL-6, IP-10, and IRF7 (Supplementary Figure 1).

In MG6 cells stimulated with LPS, iNOS protein expression was strongly reduced by the NF-κB-specific inhibitor BAY 11-7082 (Supplementary Figure 2A). However, neither phosphorylation nor nuclear translocation of p65 in LPS-stimulated cells was affected when GRK2 siRNAs were transfected (Supplementary Figures 2B,C). These findings suggest that GRK2 plays no role in regulating NF-κB activation in LPS-stimulated microglial cells.

Type I IFNs, such as IFN-β, is a second major group of cytokines that are produced by LPS-activated immune cells (Noppert et al., 2007), and signal through the JAK/STAT pathway to stimulate nuclear gene expression (Horvath, 2004). Type I IFN can signal through forming a ternary complex with the type I IFN receptor, composed of its two subunits IFNAR1 and IFNAR2 (Lee and Ashkar, 2018). The ablation of IFNAR1 by siRNAs resulted in a disappearance of iNOS expression in LPS-challenged MG6 cells (Figure 4A). When LPS was applied to MG6 cells, gene expression levels of IFN-β were greatly upregulated (Figure 4B). Furthermore, when the amounts of IFN-β in culture media were measured by ELISA, LPS challenge resulted in a striking increase in IFN-β protein levels (Figure 4C). The increased mRNA and protein levels of IFN-β were significantly suppressed by transfection of GRK2 siRNAs (Figure 4A,B). These data suggest that GRK2 contributes to the production of IFN-β in LPS-stimulated microglial cells.

IRFs, a family of transcription factors, play a central role in controlling the type I IFN induction at the gene transcriptional level (Honda et al., 2006). In microglial cells activated with LPS and other inflammatory stimuli, IRF1 also appears to be an essential transcription factor for expression of inflammatory mediators, including iNOS (Lee et al., 2001; Jantaratnotai et al., 2013; Matsuda et al., 2015). In MG6 cells, LPS challenge led to the upregulation of IRF1 mRNAs and the translocation of IRF1 proteins into the nucleus (Figure 5A,B). Transfection of IRF1 siRNAs strikingly eliminated LPS-induced upregulation of IFN-β mRNA expression (Figure 5C). Furthermore, IRF1 siRNA transfection negated LPS-induced iNOS expression (Figure 5D). These findings imply that IRF1 plays a crucial role in induction of IFN-β and subsequent production of iNOS in LPS-stimulated microglial cells.

A previous study with IRF3-deficient mice has established a key role for IRF3 in LPS-induced IFN-β gene expression (Sakaguchi et al., 2003). We tested whether IRF3 could be involved in induction and/or activation of IRF1 in LPS-stimulated MG6 cells. Transfection of IRF3 siRNAs did not substantially alter the LPS-induced increases in expression and nuclear translocation of IRF1 (Supplementary Figures 3A,B).

TRIF is essential for TLR-3 or TLR-4-mediated MyD88-independent pathways and contributes to pro-inflammatory cytokine production and, most importantly, induces type I IFN production, particularly IFN-β (Takeda and Akira, 2005). However, transfection of TRIF siRNAs had little effect on LPS-induced upregulation of IRF1 mRNA expression (Supplementary Figure 3C), indicating that LPS-stimulated induction of IRF1 is independent of the TRIF pathway. On the other hand, BAY 11-7082 greatly inhibited the LPS-induced increase in both mRNA and protein expression (Supplementary Figures 3D,E), which suggests that IRF1 is transcriptionally regulated by NF-κB. When GRK2 siRNAs were transfected, the LPS-induced upregulation of IRF1 mRNA was not substantially affected (Figure 5E), but that of IRF1 protein was evidently reduced (Figure 5F). In addition, the nuclear translocation of IRF1 was strongly reduced by GRK2 siRNA transfection (Figure 5G). The ability of GRK2 siRNAs to reduce the LPS-induced upregulation of IRF1 protein was evident regardless of whether the proteasome inhibitor MG132 was present, although the upregulation of IRF1 protein by LPS was markedly increased in the presence of MG132 (Figure 5H). These results suggest that GRK2 participates in LPS-induced upregulation and activation of IRF1 protein in microglial cells in a manner independent of the proteasome degradation route.

LPS-induced TLR4 endocytosis is necessary to initiate the TRIF-dependent IFN-β expression (Kagan et al., 2008). We thus examined whether GRK2 can regulate endocytosis of TLR4 in LPS-stimulated MG6 cells. FACS analysis showed much lower surface expression of TLR4 when MG6 cells were stimulated with LPS (Supplementary Figure 3F). Transfection of GRK2 siRNAs was without effect on the LPS-induced decrease in TLR4 surface expression, suggesting that GRK2 plays no regulatory role in the endocytosis pathway that delivers TLR4 to endosomes.

IRF1 has been known to control TLR9-mediated IFN-β production in myeloid dendritic cells (Schmitz et al., 2007). When TLR9 was activated with CpG-ODN, a synthetic oligodeoxynucleotide containing specific unmethylated CpG motifs, the mRNA levels of IFN-β were strikingly upregulated in MG6 cells (Figure 6A). The CpG-ODN-induced increase in IFN-β mRNAs were significantly inhibited by transfection of GRK2 siRNAs or IRF1 siRNAs. Furthermore, the knockdown of GRK2 by siRNAs attenuated CpG-ODN-induced IRF1 protein expression (Figure 6B). These findings suggest that GRK2 can positively regulate TLR9-mediated IFN-β production by upregulating expression of IRF1.

Paclitaxel, an anti-microtubule agent with anti-tumoral activity, is identified as a ligand to activate TLR4 signaling (Byrd-Leifer et al., 2001; Zimmer et al., 2008). We examined whether GRK2 is involved in paclitaxel-derived TLR4 signaling for iNOS expression in microglial cells. MG6 cells were challenged with paclitaxel at concentrations of 5–10 μM. Phosphorylated levels of STAT1 and STAT3 were highly upregulated as seen in LPS-stimulated cells, all of which were hampered by transfection of GRK2 siRNAs (Figure 7A). Furthermore, GRK2 siRNA transfection significantly suppressed the paclitaxel-induced increase in iNOS synthesis (Figure 7B).

Polyinosinic-polycytidylic acid [poly(I:C)] is a synthetic analog of double-stranded RNA recognized by TLR-3 (Matsumoto and Seya, 2008). We examined whether GRK2 plays a regulatory role in TLR3-mediated signaling for iNOS synthesis in microglial cells. When poly(I:C) (50 μg/ml) was applied to MG6 cells, the translocation of IRF1 into the nucleus occurred in a time-dependent manner (Figure 8A). Transfection of GRK2 siRNAs blocked the nuclear translocation of IRF1 in poly(I:C)-stimulated cells, as seen in LPS-stimulated cells (Figure 8B). In poly(I:C)-challenged cells, GRK2 siRNA transfection also significantly declined the upregulation of IFN-β mRNA levels (Figure 8C) and reduced the increases in phosphorylated levels of STAT1 and STAT3 (Figure 8D). As expected from the involvement of STAT1/3 in iNOS expression in LPS-stimulated MG6 cells, poly(I:C) administration led to the production of iNOS mRNA, which was blocked by treatment with GRK2 siRNAs (Figure 8E). In addition, GRK2 siRNA transfection greatly reduced the poly(I:C)-induced upregulation of mRNA levels of IL-1β, IL-6, IP-10, and IRF7 (Supplementary Figure 4). These results suggest that GRK2 serves as a key player for TLR3 signaling to produce iNOS in microglial cells.

We also examined the role of GRK2 in STAT1/3-mediated iNOS production in microglial cells supplemented with exogenous IFN-β. IFN-β supplementation (10 ng/ml) caused time-dependent increases in phosphorylated levels of STAT1 and STAT3 in MG6 cells (Figure 9A). These changes were greatly attenuated by transfection of GRK2 siRNAs (Figure 9A). Only exposure to IFN-β failed to induce iNOS mRNAs (Figure 9B). This suggests that IFN-β by itself was not sufficient to stimulate induction of iNOS expression. However, exogenously applied IFN-β markedly augmented iNOS mRNA levels caused by a low dose of LPS (Figure 9B). GRK2 siRNA transfection blocked the iNOS mRNA expression in the presence of low-dose LPS and IFN-β (Figure 9B), providing evidence to support a role of GRK2 in IFN-β-induced activation of STAT1/STAT3 to augment iNOS production.

Methyl 5-[2-(5-nitro-2-furyl)vinyl]-2-furoate acts as a selective inhibitor of kinase activity of GRK2 (Iino et al., 2002). We examined whether this GRK2 inhibitor can modify the function or expression of STAT1/3 and IRF1 in LPS-stimulated MG6 cells. As demonstrated in our recent report (Kawakami et al., 2018), treatment with the GRK2 inhibitor abrogated the LPS-induced upregulation of iNOS protein (Figure 10A). Furthermore, GRK2 inhibitor treatment strongly eliminated phosphorylation of STAT1 and STAT3 after LPS stimulation (Figure 10B). The IFN-β-induced increases in STAT1 and STAT3 phosphorylation were also attenuated by the GRK2 inhibitor (Figure 10C). However, GRK2 inhibitor treatment did not substantially affect expression and nuclear translocation of IRF1 following LPS challenge (Figure 10D,E). These results suggest that STAT1/3, but not IRF1, in activated microglia can be regulated by GRK2 in a kinase-dependent manner.