IRF8 Is a Critical Transcription Factor for Transforming Microglia into a Reactive Phenotype (original) (raw)

Cell Rep. Author manuscript; available in PMC 2014 Sep 9.

Published in final edited form as:

PMCID: PMC4158926

NIHMSID: NIHMS618850

Takahiro Masuda

1Department of Molecular and System Pharmacology, Graduate School of Pharmaceutical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, Fukuoka 812-8582, Japan

Makoto Tsuda

1Department of Molecular and System Pharmacology, Graduate School of Pharmaceutical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, Fukuoka 812-8582, Japan

Ryohei Yoshinaga

1Department of Molecular and System Pharmacology, Graduate School of Pharmaceutical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, Fukuoka 812-8582, Japan

Hidetoshi Tozaki-Saitoh

1Department of Molecular and System Pharmacology, Graduate School of Pharmaceutical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, Fukuoka 812-8582, Japan

Keiko Ozato

2Program in Genomics of Differentiation, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD 20892, USA

Tomohiko Tamura

3Department of Immunology, Graduate School of Medicine, Yokohama City University, Yokohama 236-0004, Japan

Kazuhide Inoue

1Department of Molecular and System Pharmacology, Graduate School of Pharmaceutical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, Fukuoka 812-8582, Japan

1Department of Molecular and System Pharmacology, Graduate School of Pharmaceutical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, Fukuoka 812-8582, Japan

2Program in Genomics of Differentiation, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD 20892, USA

3Department of Immunology, Graduate School of Medicine, Yokohama City University, Yokohama 236-0004, Japan

#Contributed equally.

SUMMARY

Microglia become activated by multiple types of damage in the nervous system and play essential roles in neuronal pathologies. However, how micro-glia transform into reactive phenotypes is poorly understood. Here, we identify the transcription factor interferon regulatory factor 8 (IRF8) as a critical regulator of reactive microglia. Within the spinal cord, IRF8 expression was normally low; however, the expression was markedly upregulated in microglia, but not in neurons or astrocytes, after peripheral nerve injury (PNI). IRF8 overexpression in cultured microglia promoted the transcription of genes associated with reactive states; conversely, IRF8 deficiency prevented these gene expressions in the spinal cord following PNI. Furthermore, IRF8-deficient mice were resistant to neuropathic pain, a common sequela of PNI, and transferring IRF8-over-expressing microglia spinally to normal mice produced pain. Therefore, IRF8 may activate a program of gene expression that transforms microglia into a reactive phenotype. Our findings provide a newly observed mechanism for microglial activation.

INTRODUCTION

Microglial cells are the resident immune-related glial cells of the central nervous system (CNS) that are crucial for maintaining homeostasis and sensing pathological alterations in the nervous system, such as following infection and injury (Glass et al., 2010; Hanisch and Kettenmann, 2007; Perry et al., 2010; Ransohoff and Cardona, 2010). Under normal conditions, microglia survey the surrounding local environment by actively moving their branched processes. As a consequence of multiple types of damage in the nervous system, microglia transform to reactive states through a progressive series of cellular and molecular changes, including morphological hypertrophy, proliferation and expression of various genes. In particular, expression of cell-surface receptors (e.g., toll-like receptors [TLRs] and nucleotide receptors [P2X and P2Y receptors]) and proinflammatory cytokines (e.g., interleukin [IL]-1β) is a critical process for inducing reactive phenotypes of microglia linked to the pathogenesis of various CNS diseases such as multiple sclerosis, Alzheimer's disease, and neuropathic pain (Glass et al., 2010; Inoue and Tsuda, 2009; McMahon and Malcangio, 2009; Perry et al., 2010). However, the molecular mechanisms by which microglia switch to reactive phenotypes are poorly understood.

RESULTS AND DISCUSSION

As microglia can transform into reactive phenotypes through the activation of gene transcription, we considered that reactive states of microglia may be controlled by transcription factors. To investigate this possibility, we performed a genome-wide screen of mRNAs from the spinal cord of mice with or without peripheral nerve injury (PNI), a model of CNS pathology in which remote injury of a peripheral nerve (fourth lumbar [L4] spinal nerve) results in activation of microglia in the spinal dorsal horn where the injured nerve projects. In three independent DNA microarray analyses, we identified interferon regulatory factor 8 (IRF8) as a transcription factor whose expression was significantly upregulated in the spinal cord after PNI (p = 0.015, Figure S1A). IRF8 is a member of the IRF family (IRF1-9), and is expressed in immune cells such as lymphocytes and dendritic cells. In the periphery, IRF8 plays pivotal roles in the immune system (Honda and Taniguchi, 2006; Tamura et al., 2000, 2008), but its role in the CNS is entirely unknown. Thus, we determined the type of cells expressing IRF8 in the spinal cord after PNI using an immunofluorescence approach. Three days after PNI, on sections from the L4 spinal dorsal horn, we observed strong immunofluorescence of IRF8 protein dotted on the ipsilateral side (Figure 1A). In contrast, on the contralateral side where intact nerves project, IRF8 immunofluorescence was weak. The observed staining was not a nonspecific signal because no IRF8 immunofluorescence was detected in mice lacking IRF8 (_Irf8_–/–) (Figure S1B). At higher magnification, almost all IRF8 immunofluorescence colocalized with the nuclear marker DAPI (Figure 1B). We further performed double-immunolabeling with cell type-specific markers and found that almost all IRF8+ cells were positive for the microglial markers Iba1 (328/332 IRF8+ cells; Figure 1C) and OX-42 (363/368 IRF8+ cells; Figure 1D). IRF8+ cells were not double-labeled with markers for neurons (NeuN, MAP2, and NF200) or astrocytes (GFAP) (Figures 1E–1H). By triple labeling with IRF8/Iba1/DAPI, we confirmed the nuclear localization of IRF8 in microglia (Figure S1C). Spinal microglia with the high levels of IRF8 showed activated morphology (Figure 1C), and the intensity of IRF8 immunofluorescence per Iba1+ cell was markedly increased (Figures S1E and S1F). We therefore concluded that, in the dorsal horn following PNI, reactive microglia are the cell type which express IRF8 and that the level of IRF8 expression is dramatically increased in individual reactive microglia. In addition, microglia-specific IRF8 upregulation in the spinal cord was also induced in rats after PNI (Figures S1K–S1O). Furthermore, we examined the temporal pattern of IRF8 expression by real-time RT-PCR analysis and found that Irf8 mRNA levels, which were low in naive mice, were increased in the spinal cord ipsilateral to PNI, starting from postoperative day 1 and peaking on day 3 and persisting for more than 3 weeks (Figure 1I; Figure S1D). Correspondingly, western blot analysis demonstrated upregulation of IRF8 protein in the ipsilateral spinal cord after PNI (Figures 1J and 1K). Spinal IRF8 upregulation at both earlier (24–44 hr) and later (days 7–21) time points after PNI was also specific to microglia (Figures S1G–S1J), indicating that the microglia-specific upregulation of IRF8 persists for at least 3 weeks after PNI.

An external file that holds a picture, illustration, etc. Object name is nihms-618850-f0001.jpg

PNI Induces IRF8 Upregulation Exclusively in Microglia in the Spinal Cord

(A) Visualization of IRF8 protein in the dorsal spinal cord 3 days after PNI.

(B) Nuclear localization of IRF8.

(C–H) Double immunolabeling of IRF8 with Iba1 (C), OX-42 (D), GFAP (E), NeuN (F), MAP2 (G), and NF200 (H).

(I) Real-time PCR analysis of Irf8 mRNA in WT mouse spinal cord before (Naive) and after PNI. Values represent the relative ratio of Irf8 mRNA (normalized to Gapdh mRNA) to the contralateral side of naive mice (n = 6; **p < 0.01).

(J) Western blot analysis of IRF8 protein in the spinal cords of WT mice before (Naive) and after PNI.

(K) A histogram of the relative band density ratio of IRF8 (normalized to β-actin) to the contralateral side of naive mice at each time point (n = 5; **p < 0.01).

Values are means ± SEM for all groups.

Scale bars: 100 μm (A), 10 μm (B), 50 μm (C–H).

See also Figure S1.

To determine if IRF8 activates transcription in microglia, we transduced primary cultured microglia with a lentiviral vector encoding IRF8 tagged with green fluorescent protein (GFP) (IRF8–GFP) or a control vector encoding GFP alone, and examined the levels of gene transcripts in microglia. On the basis of current concepts that there are multiple activity states of reactive microglia (Hanisch and Kettenmann, 2007; Perry et al., 2010; Ransohoff and Cardona, 2010), we focused on genes involved in microglial innate responses (TLR2 [_Tlr2_] and TLR4 [_Tlr4_]) (Rivest, 2009), chemotaxis (the metabotropic nucleotide receptor P2Y12R [_P2ry12_] and the chemokine receptor CX3CR1 [_Cx3cr1_]) (Haynes et al., 2006; Honda et al., 2001; Tran and Miller, 2003) and inflammatory components (IL-1β [_Il1b_], IL-6 [_Il6_], cathepsin S [_Ctss_], brain-derived neurotrophic factor [_Bdnf_], the ionotropic nucleotide receptors P2X4R [_P2rx4_], and P2X7R [_P2rx7_]) (Burnstock, 2008; Clark et al., 2007; Coull et al., 2005; Inoue and Tsuda, 2009; Tsuda et al., 2003). First, we confirmed transduction efficacy of the two vectors (Figures S2A and S2B) and the Irf8 mRNA level in microglia with IRF8– GFP (Figure S2C). In IRF8-transduced microglia, the mRNA levels for Tlr2, P2ry12, Cx3cr1, Il1b, Ctss, and P2rx4 were increased (Figure 2). We confirmed similar effects of IRF8 using the microglial cell line BV2 (Figures S2D and S2E). Bdnf mRNA tended to be increased (Figure 2), and mRNAs for Tlr4, Il6, and P2rx7 were not altered by IRF8 overexpression (Figure 2). Furthermore, in microglia transduced with a mutant IRF8 in which lysine at amino acid position 79 (required for DNA-binding activity; Tamura et al., 2000) was replaced with Glu (IRF8[K79E]), expression of genes that were induced by wild-type IRF8 was not increased (Figure 2). The failure of the mutant protein to induce gene expression was not due to insufficient levels of mutant IRF8 (Figure S2C). Thus, forced IRF8 expression activates transcription of genes associated with a reactive state of microglia in a manner that depends on its ability to bind DNA.

An external file that holds a picture, illustration, etc. Object name is nihms-618850-f0002.jpg

Forced Expression of IRF8 in Microglia Promotes Gene Transcription

Real-time PCR analysis of the mRNA of genes of interest in cultured microglia 72 hr after transduction. Values represent the relative ratio of mRNA (of tested genes, normalized to Gapdh mRNA) to control microglia with GFP alone (n = 5–7, *p < 0.05, ***p < 0.001 versus GFP; #p < 0.05, ##p < 0.01, ###p < 0.001 versus IRF8-GFP).

Values are the mean ± SEM for all groups.

See also Figure S2.

To investigate the role of IRF8 in microglial gene expression in vivo, we subjected _Irf8_–/– mice and their wild-type (WT) littermates to PNI and examined the abundance of microglial gene transcripts in the spinal cord. Based on our findings in vitro (Figure 2), we focused on putative IRF8-regulated microglial genes. In WT mice, the transcript levels of Tlr2, P2ry12, Cx3cr1, Il1b, Ctss, and P2rx4 were significantly increased 7 days after PNI (Figure 3A). However, in _Irf8_–/– mice, the PNI-induced increase in expression of Tlr2, P2ry12, Cx3cr1, Ctss, and Bdnf was suppressed, and P2rx4 and Il1b mRNAs did not significantly increase after PNI (Figure 3A). These results indicate that IRF8 is necessary for PNI-induced expression of a set of genes in reactive microglia in vivo. In addition, we examined gene expression at a later time point. On day 21 post-PNI, we found, in WT mice, a significant increase for the genes Tlr2, Cx3cr1, and Ctss in addition to Iba1 (Aif1) (Figures S3D and S3E). However, like Irf8 mRNA on day 21 (Figure S1D), their levels were lower than those on day 7. Among them, the upregulation of Ctss and Aif1 mRNAs was significantly suppressed in _Irf8_–/– mice (Figure S3D). The lesser effects of IRF8 deficiency on microglial gene expression might be related to the lower level of IRF8 in microglia on day 21 post-PNI.

An external file that holds a picture, illustration, etc. Object name is nihms-618850-f0003.jpg

IRF8 Is Required for Microglial Gene Expression, but Not for Microglial Proliferation, in the Spinal Cord after PNI

(A) Real-time PCR analysis of mRNAs of microglial genes in the spinal cords of WT and _Irf8_–/– mice 7 days after PNI. Values represent the relative ratio of mRNA (of tested genes, normalized to Gapdh mRNA) to the contralateral side of WT mice. C, contralateral; I, ipsilateral. (n = 6–7, *p < 0.05, **p < 0.01, ***p < 0.001).

(B) Double immunofluorescence for Ki-67 or p-HisH3 (green) and OX-42 (red) in the ipsilateral dorsal horn of WT or _Irf8_–/– mice 2 days after PNI. A histogram of the numbers of Ki-67+OX-42+ or p-HisH3+OX-42+ cells in the ipsilateral (I) and contralateral (C) dorsal horn (n = 3, ***p < 0.001).

(C) OX-42 (red) and Iba1 (green) immunofluorescence in the spinal cord of WT and _Irf8_–/– mice 14 days after PNI.

Values are the mean ± SEM for all groups. Scale bars: 50 μm (B), 200 μm (C), 50 μm (C, insets).

See also Figure S3.

Robust proliferation of spinal microglia is an early event occurring after PNI and is an important process for transforming into a reactive state (Tsuda et al., 2009). Thus, it is possible that the impaired microglial gene expression in _Irf8_–/– mice may be due to a change in this early process. However, PNI-induced increases in the number of proliferating spinal microglia that were positive for the proliferation markers Ki-67 and phosphorylated-histone H3 (p-HisH3) were indistinguishable between WT and _Irf8_–/– mice (Figure 3B). Furthermore, increased immunofluorescence of the microglial marker OX-42 after PNI, due to a cellular alteration of microglia, was similarly observed in the ipsilateral dorsal horn of both genotypes after PNI (Figure 3C). These data indicate that lack of IRF8 does not result in a global defect in reactive processes of microglia. However, _Irf8_–/– mice, surprisingly, showed a marked reduction in Iba1 immunofluorescence on both sides of the spinal cord (Figure 3C), which was also verified by real-time RT-PCR (Figures S3A and S3D) and Western blotting (Figure S3B). By using Irf8+/– mice, we showed that the reduction was dependent on levels of the Irf8 gene (Figure S3C). The low level of Aif1 mRNA was also observed in cultured _Irf8_–/– microglia (Figure S3F) and, conversely, was increased in cultured microglia transduced with IRF8, but not with IRF8(K79E) (Figures S3G and S3H). As there are IRF elements in the promoter region of the Iba1 gene (Sibinga et al., 2002), IRF8 may directly control Iba1 expression in microglia.

Our results above show that IRF8 is crucial for activating gene expression in reactive microglia in the spinal dorsal horn after PNI, raising the possibility that IRF8 might contribute to neuronal pathologies following PNI. Injury to the nervous system arising from bone compression in cancer, diabetes, infection, autoimmune disease or physical injury is known to result in debilitating chronic pain states (referred to as neuropathic pain) (McMahon and Malcangio, 2009; Tsuda et al., 2005). A cardinal symptom of neuropathic pain is characterized by tactile allodynia, abnormal pain hypersensitivity evoked by innocuous stimuli. Therefore, we investigated whether IRF8 induced in spinal microglia after PNI contributes to neuropathic allodynia using _Irf8_–/– mice. In WT mice, PNI produced a profound, long-term decrease in paw withdrawal threshold (PWT) to mechanical stimulation of the hindpaw (Figure 4A). The decrease in PWT on day 1 was comparable between WT and _Irf8_–/– mice, but _Irf8_–/– mice failed to show a further decrease in PWT after 3 days, lasting until the final time point tested. _Irf8_–/– mice showed no change in basal mechanical sensitivity or contralateral PWT after PNI (Figure 4A), nor did it affect motor function (data not shown). Furthermore, suppressing upregulated expression of spinal IRF8 after PNI by intrathecal administration of a small interfering RNA (siRNA) targeting IRF8, given on day 5 and 6 post-PNI to WT mice that had developed allodynia, caused a significant recovery in PWT (Figure 4B). Another IRF8 siRNA also produced similar effects (Figure S4F). Together, these data indicate that IRF8 upregulation in the spinal cord is necessary for full development and maintenance of tactile allodynia after PNI.

An external file that holds a picture, illustration, etc. Object name is nihms-618850-f0004.jpg

Microglial IRF8 Is Necessary for Abnormal Pain Hypersensitivity Caused by PNI

(A) PWT of _Irf8_–/– and WT mice before (Pre) and after PNI (n = 4; *p < 0.05, **p < 0.01 versus Pre; #p < 0.05, ##p < 0.01, ###p < 0.001 versus the ipsilateral side of WT mice). C, contralateral; I, ipsilateral.

(B) Reversal of PNI-induced allodynia by intrathecal administration of IRF8 siRNA (20 pmol) once a day for 2 days (on 5 and 6 days post-PNI) in WT mice (n = 3–6, *p < 0.05). Upper, representative immunoblots of IRF8 and β-actin proteins in the spinal cords of mice treated with control and IRF8 siRNAs on day 7 post-PNI.

(C) Experimental protocol.

(D) PWT of WT mice intrathecally administered with cultured microglia overexpressing either GFP, IRF8–GFP or IRF8(K79E)–GFP (n = 5–6, **p < 0.01 versus Pre; #p < 0.05 versus WT mice with GFP microglia; §p < 0.05, §§p < 0.01 versus WT mice with IRF8–GFP microglia).

(E) Allodynia by IRF8-transduced cultured microglia was prevented by preincubating microglia with a cocktail of IL-1β neutralizing antibody (5 μg) and CatS inhibitor (5 pmol) for 15 min before the intrathecal injection of microglia (n = 6–7, *p < 0.05, **p < 0.01 versus Pre; #p < 0.05, ###p < 0.001 versus IRF8–GFP/ control group).

(F) PWT of WT and _Irf8_–/– mice before (Pre) and after intraplantar CFA injection (n = 5, **p < 0.01 versus Pre).

(G) Hot-plate test. Values represent the latency (s) for animals to lick their hindpaws or jump (n = 6).

(H) Tail-flick test. Values represent the latency (s) to flick their tail from the heat source (n = 4).

(I–K) Capsaicin (I) and formalin (J and K) test. Values are the duration (s) of nociceptive behaviors (I: n = 6, J: n = 8) (**p < 0.01 versus WT mice). (K) Total duration (s) of nociceptive behaviors for 0–5 min (first phase) and for 10–60 min (second phase).

Values are the mean ± SEM for all groups.

See also Figure S4.

To further investigate the impact of microglial IRF8 in modulating pain, we employed a model in which cultured microglial cells are transferred spinally to normal mice (Figure 4C) (Coull et al., 2005; Tsuda et al., 2003, 2009). Transferring microglia transduced with GFP alone did not produce any change in PWT (Figure 4D). By contrast, spinal transfer of IRF8-transduced microglia to normal mice produced a significant decrease in PWT (Figure 4D). However, IRF8(K79E)-transduced microglia failed to decrease PWT (Figure 4D). It is possible that transferring IRF8-transduced microglia may activate endogenous spinal IRF8 in recipient mice, leading to allodynia. However, IRF8-transduced microglia also produced behavioral allodynia in _Irf8_–/– mice (n = 3, pre: 1.50 ± 0.25 g; 5 hr: 0.58 ± 0.04 g). Thus, microglial IRF8 is sufficient to produce tactile allodynia, the behavioral outcome depends on the DNA-binding activity of IRF8. Microglia-mediated inflammatory components are implicated in enhancing neuronal excitability in the CNS including the dorsal horn (Inoue and Tsuda, 2009; Kawasaki et al., 2008; McMahon and Malcangio, 2009). Because gene transcripts up-regulated by IRF8 included IL1b and Ctss (Figure 2), we tested the effect of these two inflammatory molecules in behavioral allodynia elicited by IRF8-expressing microglia. We found that pretreating IRF8-transduced microglia with a cocktail containing a function-blocking antibody for IL-1β and an inhibitor of Ctss attenuated the decrease in PWT by IRF8-transduced microglia (Figure 4E). Thus, these results imply that inflammatory signals derived from microglia overexpressing IRF8 can produce behavioral allodynia.

Because _Irf8_–/– mice may have congenital defects in pain processing, we also tested behavioral responses in other pain assays. Intraplantar injection of complete Freund's adjuvant (CFA), a model of pain caused by local inflammation of the hind-paw, produced no detectable change in spinal IRF8 expression (Figure S4G) and caused pain hypersensitivity in both WT and _Irf8_–/– mice (Figure 4F). In acute pain assays, behavioral responses to heat and capsaicin were indistinguishable between the two genotypes (Figures 4G–4I). Likewise, pain responses related to tissue damage by injecting dilute formalin into the hindpaw did not differ, although _Irf8_–/– mice showed a delay in onset of the second phase (Figures 4J and 4K). By using several markers for DRG neurons, we confirmed that the phenotypes of DRG neurons are comparable between the two genotypes (Figures S4A–S4E). Thus, microglial IRF8 critically contributes to chronic pain, especially neuropathic pain, without causing substantial alterations in basal sensory thresholds, local tissue inflammation-induced pain or motor performance.

Here, we demonstrate a role for IRF8 in the CNS as a key transcription regulator of microglia involved in converting to a reactive state after PNI in vivo. To our knowledge, IRF8 is the first transcription factor specific to microglia, among CNS-resident cells, that controls their reactive processes and CNS pathology. We further found that IRF8 upregulation in brain microglia was also observed in other models of injury (hypoglossal nerve axotomy and kainic acid-induced neuronal injury) (Figures S1P–S1R), suggesting that upregulation of microglial IRF8 may be commonly induced following diverse types of neuronal injury. Interestingly, IRF8 deficiency prevented gene expressions in microglia without affecting their proliferative activity. Therefore, our findings suggest that IRF8 is not the sole transcription factor for microglial activation but rather activates a program of gene expression that determines signaling activity of reactive microglia, which provides a newly observed mechanism for microglial activation. For the IRF8 upregulation in microglia, it has been reported that interferon (IFN)-γ increases IRF8 expression in vitro (Politis et al., 1994) and IFN-γ receptors are expressed in spinal microglia (Tsuda et al., 2009). However, we found that IFN-γ receptor-deficient mice failed to prevent PNI-induced Irf8 expression in the spinal cord (Figure S1S), proposing the possible involvement of an IFN-γ-independent mechanism (e.g., TLRs; Barber et al., 1995). Although elucidating this issue requires further investigation, our present data suggest that inhibiting upregulated expression or function of IRF8 may provide a way to comprehensively control the expression of a set of genes in reactive micro-glia. Of importance to therapeutic development, _Irf8_–/– mice were resistant to allodynia, and suppressing spinal IRF8 by its siRNA reversed established allodynia. Although the possible involvement of other cell types expressing IRF8 in neuropathic pain defects seen in _Irf8_–/– mice cannot be excluded, our findings indicate that the development and maintenance of neuropathic pain requires ongoing activity of microglial IRF8. Therefore, our results may provide a new therapeutic strategy for neuropathic pain, for which there is currently no effective therapy. In addition, reactive microglia with altered expression of various genes by an as-yet-unknown mechanism play prominent roles in the pathogenesis of other CNS diseases (Glass et al., 2010; Hanisch and Kettenmann, 2007; Perry et al., 2010). A recent meta-analysis identified IRF8 as a susceptibility locus for multiple sclerosis (De Jager et al., 2009), a refractory neurodegenerative disease in which reactive microglia are implicated (Glass et al., 2010). Whether reactive microglia upregulate IRF8 expression and its transcriptional activity during the course of CNS diseases remains unknown. However, if so, then this transcription factor could provide a therapeutic target in disorders of the CNS in addition to neuropathic pain.

EXPERIMENTAL PROCEDURES

Details are further described in the Extended Experimental Procedures.

Animals

Male IRF8-deficient mice (_Irf8_–/–) (Holtschke et al., 1996) and their WT litter-mates, and C57BL/6 mice were used. All mice used were aged 9–12 weeks at the start of each experiment and were housed at 22 ± 1°C with a 12-hr light-dark cycle and fed food and water ad libitum. All experimental procedures were performed under the guidelines of Kyushu University.

Peripheral Nerve Injury

Under isoflurane (2%) anesthesia, a unilateral L4 spinal nerve of mice was cut as described previously (Tsuda et al., 2009).

Microarray Analysis

Total RNA from the L3–L5 spinal cord was converted to biotin-labeled cRNA, which was hybridized to the Mouse WG-6 V2.0 BeadChip (Illumina). Gene expression was analyzed according to BeadStudio Gene Expression Module User Guide (Illumina).

Quantitative Real-Time PCR

Quantitative PCR was performed with Premix Ex Taq (Takara) using a 7500 real-time PCR system (Applied Biosystems), and the data were analyzed using 7500 System SDS Software 1.3.1.

Western Blot and Immunohistochemistry

Western blot and immunohistochemistry were performed using standard methods. See Supplemental Information for details.

Microglial Culture

Mouse primary cultured microglia were prepared as described previously (Tsuda et al., 2003, 2009).

Lentiviral Transduction

The lentiviral CS2-EF-MCS vector expressing mouse IRF8-GFP, IRF8(K79E)-GFP (Tamura et al., 2000), or GFP alone were used. Viral particles were added onto primary cultured microglia (1.2 × 105 cells/well). Total RNA in cultured microglia 72 hr after the transduction were used for real-time PCR analysis. For behavioral tests, the transduced microglia were washed, harvested, and administering intrathecally to catheterized mice.

Intrathecal Administration

Intrathecal injection of IRF8 siRNA (20 pmol/5 μl) and cultured microglia (2 × 104 cells/5 μl) was followed by infusion of 3 μl of PBS. Fifteen minutes before the injection of transduced microglia, cells were mixed with a reagent cocktail (5 μL) containing an IL-1β neutralizing antibody (5 μg, R&D Systems) and the CatS inhibitor Z-FL-COCHO (5 pmol, Calbiochem), or with a control vehicle.

Behavioral Studies

To assess mechanical sensitivity, calibrated von Frey filaments (0.02–2.0 g) were applied to the plantar surfaces of the hindpaws of mice, and the 50% PWT was determined. Methods for all other tests are described in the Supplemental Information.

Statistical Analysis

Statistical analyses were performed using the Student's t test (Figures 4B, 4G–4I, and 4K), two-way ANOVA with post hoc Bonferroni test (Figures 1I, 1K, 4A, 4F, and 4J), one-way ANOVA with a post hoc Dunnett's test (Figures 4A and 4D–4F), or Tukey's test (Figures 2, 3A, 3B, 4D, and 4E) using GraphPad Prism 4.03 software. Differences were considered significant at p < 0.05.

Supplementary Material

IRF8

ACKNOWLEDGMENTS

This work was supported by grants from the JSPS through the NEXT Program initiated by the CSTP (M.T.) and the MEXT of Japan (T.M., M.T., K.I.) and from the Core-to-Core program of JSPS (K.I.). We thank Ms. Junko Kitano, Ms. Yukari Hasegawa, Dr. Shigeo Hasegawa and the Research Support Center (Graduate School of Medical Sciences, Kyushu University) for assisting with experiments, and Dr. Hiroyuki Miyoshi (RIKEN BioResource Center) for providing the lentiviral vector and its packaging plasmids. T.M. designed and performed most of the experiments, analyzed the data, and wrote the manuscript. M.T. conceived the study, supervised the overall project, performed immunohistochemical experiments, and wrote the manuscript. R.Y. assisted with the experiments. H.T.-S. optimized transduction of lentiviral vectors into primary cultured microglia. K.O. and T.T. provided critical materials including _Irf8_–/– mice and advice on data interpretation. K.I. supervised the overall project and wrote the manuscript.

Footnotes

SUPPLEMENTAL INFORMATION

Supplemental Information includes Extended Experimental Procedures and four figures and can be found with this article online at doi:10.1016/j.celrep. 2012.02.014.

REFERENCES