The Cytokine IL-1 Activates IFN Response Factor 3 in Human Fetal Astrocytes in Culture
http://www.100md.com
免疫学杂志 2005年第6期
Abstract
The cytokine IL-1 is a major activator of primary human fetal astrocytes in culture, leading to the production of a wide range of cytokines and chemokines important in the host defense against pathogens. IL-1, like TLR4, signals via the MyD88/IL-1R-associated kinase-1 pathway linked to activation of NF-B and AP-1. Recent studies have shown that TLR4 also signals independently of MyD88, resulting in the activation of IFN regulatory factor 3 (IRF3), a transcription factor required for the production of primary antiviral response genes such as IFN-. Using a functional genomics approach, we observed that IL-1 induced in astrocytes a group of genes considered to be IFN-stimulated genes (ISG), suggesting that IL-1 may also signal via IRF3 in these cells. We now show, using real-time PCR, that in astrocytes IL-1 induces the expression of IFN-, IRF7, CXCL10/IFN--inducible protein-10, and CCL5/RANTES. Chemokine expression was confirmed by ELISA. We also show that IL-1 induces phosphorylation and nuclear translocation of IRF3 and delayed phosphorylation of STAT1. The dependency of IFN-, IRF7, and CXCL10/IFN--inducible protein-10 gene expression on IRF3 was confirmed using a dominant negative IRF3-expressing adenovirus. The robust induction by IL-1 of additional ISG noted on the microarrays, such as STAT1, 2'5'-oligoadenylate synthetase 2, and ISG15, also supports an active signaling role for IL-1 via this pathway in human fetal astrocytes. These data are the first to show that IL-1, in addition to TLRs, can stimulate IRF3, implicating this cytokine as an activator of genes involved in innate antiviral responses in astrocytes.
Introduction
Astrocytes are the major glial cell population in the CNS. They reside ubiquitously in the brain parenchyma as highly ramified cells and play an important role in the maintenance of brain homeostasis. Injury to the CNS is inevitably accompanied by the rapid induction of a reactive astrogliosis, which is characterized by astrocytic hypertrophy, proliferation, and altered gene expression. Although historically the formation of a glial scar has been considered a major impediment to functional recovery of the CNS, reactive astrocytes also play an important role in repair mechanisms by removing toxic levels of factors such as glutamate, functioning as a source of growth factors and in restoration of the blood-brain barrier (1, 2, 3, 4).
A major factor that has been implicated in the induction of a reactive astrocytic phenotype is the cytokine IL-1 (5, 6). Using a functional genomics approach, we have shown that exposure of human fetal astrocytes to IL-1 for 24 h leads to the altered expression of 1000 genes, including the increased expression of multiple acute phase genes involved in the early phases of an inflammatory response, such as chemokines, cytokines, and adhesion molecules (7). The signaling cascade activated by IL-1 in most cell types has been well documented. After ligand binding to the type 1 IL-1R, the cytosolic adaptors MyD88 and Toll-interacting protein (Tollip) are rapidly recruited to the receptor complex, which then leads to binding of the IL-1R-associated kinase-4. IL-1R-associated kinase-4 is phosphorylated as a result of this interaction, mediating the recruitment of the TNF-associated factor-6. TNF-associated factor-6, through a series of other adaptor proteins, leads to activation of the downstream IB kinases (IKK and IKK), 3 which phosphorylate the NF-B repressor IB, resulting in the release of inhibition and translocation of NF-B complexes to the nucleus, with subsequent NF-B-dependent gene expression (8). A similar cascade of kinase-dependent events leads to activation of the MAPK pathway and JNK, resulting in the activation of additional transcription factors, including ATF and AP1 (9, 10). Signaling via these pathways ultimately leads to the production of proinflammatory cytokines and chemokines important in host defense as well as in the establishment of the acquired immune response.
The intracellular signaling domain of the IL-1R is homologous to that of the signaling domain of TLR4 and is termed the Toll-IL-1R homologous region (11, 12, 13). TLR family members are evolutionarily conserved proteins that are critically involved in host defense and are important mediators of inflammation and innate and adaptive immune responses. LPS derived from the cell wall of Gram-negative bacteria is a major ligand for TLR4 and is a potent activator of proinflammatory cytokines and chemokines such as IL-1, TNF, IL-6, and IL-8, which are dependent upon the MyD88 signaling pathway. However, studies of mice in which the gene for MyD88 had been inactivated, demonstrated that LPS could also signal independently of MyD88 to produce an alternative set of effector responses, leading to the production of IFN- and IFN-stimulated genes (ISG), such as CXCL10/IFN--inducible protein-10 (IP-10) (14, 15, 16). Similar cell signaling pathways have been found to be activated after binding of dsRNA to TLR3 (17, 18). IFN regulatory factor 3 (IRF3) and IRF7 are thought to be responsible for virus- or TLR ligand-induced type 1 IFNs (19). IRF3 is constitutively expressed in the cytoplasm of most cells, but upon activation by phosphorylation at multiple sites, it homodimerizes and translocates to the nucleus, where it functions as a required transcription factor in the initiation of type 1 IFN gene expression (20, 21). Autocrine and/or paracrine signaling through the type I IFN receptor then leads to phosphorylation of STAT1 and activation of a type 1 IFN-dependent gene program that includes the expression of IRF7 and the chemokines CXCL10/IP-10 and CCL5/RANTES as well as an autocrine positive feedback loop that boosts type 1 IFN production (16). In cells that show this response to LPS through TLR4, IL-1 was not found to activate this signaling pathway (22). These observations have led to the conclusion that although the activation of NF-B is a conserved response after activation of most TLRs and the IL-1R, the activation of IRF3 is a response unique to the TLR3 and TLR4 pathways as well as to infection by certain viruses. However, studies have also shown that the pathways involved in IRF3 activation may be both stimulus and cell type-specific (23, 24).
In this study we have addressed the question of whether the cytokine IL-1, which is a major activator of the NF-B and MAPK pathways in human fetal astrocytes, also leads to the activation of IRF3 in these cells. The results support the conclusion that IRF3 represents a target of IL-1 signaling in astrocytes, leading to the activation of IFN- and a group of genes considered to be secondary response genes to type 1 IFN signaling. These data suggest that astrocytes are endowed with a specific function in the brain resembling that of pathogen-activated brain macrophages/microglia. Activation of microglia leads to the production of IL-1, which then activates astrocytes, forming a second line of defense in the brain that amplifies the antimicrobial and antiviral innate immune responses of microglia.
Materials and Methods
Astrocyte cultures and cytokines
Enriched human fetal brain astrocyte cultures were established from second trimester abortuses as described previously (25). All tissue collection was approved by the institutional clinical review committee. To obtain cultures essentially free of microglia, cells were plated at 4 x 107 cells/10 ml medium/T75 tissue culture flask. After 14 days in vitro, floating microglia were removed, and astrocytes were subcultured by trypsinization for at least three passages. Culture purity was determined by immunohistochemistry for glial fibrillary acidic protein for astrocytes (BioGenex), microtubule-associated protein-2 for neurons (Sigma-Aldrich), and CD68 for microglia (DakoCytomation). Microglia, which are nondividing cells in this system, do not survive this subculturing process, and enriched astrocyte cultures typically contain <0.01% microglia contamination. Recombinant human IL-1 was a gift from the Biological Response Modifiers Program at National Cancer Institute or was purchased from PeproTech, and IFN- was purchased from Genzyme. Cytokines were diluted in medium containing 2 mg/ml endotoxin-free human serum albumin (Baxter Healthcare). Astrocytes were activated with IL-1 at 10 ng/ml (equivalent to 20 U/ml). Recombinant human CD14 was purchased from Biometec. All other reagents were purchased from Sigma-Aldrich unless otherwise indicated.
cDNA microarray
Total astrocyte (control and IL-1-treated) RNA was extracted using the RNeasy Mini Procedure (Qiagen). RNA samples tested for quality using the Agilent 2100 bioanalyzer were used for microarray analysis and were amplified once using the MessageAmp amplified antisense RNA (aRNA) kit (Ambion) before microarray analysis. Four different astrocyte cases were studied. Briefly, 5 μg of total RNA was reverse transcribed, followed by second-strand cDNA synthesis. After cDNA purification, aRNA was synthesized by in vitro transcription. Human cDNA arrays were obtained from the Albert Einstein College of Medicine cDNA Microarray Facility (relevant data available from http://microarray1k.aecom.yu.edu/). Each slide contained an unbiased, random collection of 27K cDNA probe elements derived from the sequence-verified GEM1 clone set (Incyte Genomics). Microarray procedures were performed according to protocols provided by the Microarray Facility (available from http://microarray1k.aecom.yu.edu/). In brief, after RT of aRNA (5 μg), IL-1-treated samples were labeled with Cy5-fluorescent nucleotides (Amersham Biosciences), and control samples were labeled with Cy3-fluorescent nucleotides. For each experiment, Cy5-labeled astrocyte cDNA was cohybridized with Cy3-labeled control cDNA. Hybridized slides were scanned using a GenePix 4000b scanner (Axon Instruments), and raw data files were generated containing measurements of signal and background fluorescence emissions. Data from spots containing artifacts were eliminated from further analysis. Data were also filtered, and only spots with average channel intensity over the average channel background intensity were analyzed further. The ratio of Cy5 (experimental values) and Cy3 (control RNA) was calculated for each spot to derive a relative expression value. Genes with differential expression in the experimental groups were compared by their spot ratio values.
Real-time quantitative PCR (Q-PCR)
Total RNA was isolated using the TRIzol method (Invitrogen Life Technologies) according to the manufacturer’s instructions. Briefly, confluent 100-mm cell culture dishes of astrocytes were lysed and scraped into microtubes. The lysis solution was then extracted with chloroform, and the aqueous phase was transferred to a fresh tube. RNA was precipitated with 2-propanol, washed once in 70% ethanol, and resuspended in a suitable volume of nuclease-free water. The RNA suspension was then treated to remove contaminating genomic DNA with DNase I, re-extracted, and precipitated. The RNA concentration was determined by measuring the OD values of the samples at 260 nm. Q-PCR was performed first by reverse transcribing 10 μg of total RNA in a reaction volume of 20 μl. Q-PCR was conducted in 384-well reaction plates. Each reaction sample comprised a 1:1 mixture of diluted (1/100) RT-PCR sample to SYBR Green Q-PCR Master Mix (Applied Biosystems) combined with a 1:1 mixture of gene-specific forward and reverse primers to the same SYBR Green Master Mix, resulting in a total volume of 8 μl/reaction. The plates were processed in an ABI PRISM 7000 light cycler (Applied Biosystems) using standard cycling conditions. All runs were accompanied by the two internal control genes 2-microglobulin (2M) and porphobilinogen deaminase (PBDA). Samples were normalized using a Ct-based algorithm to give arbitrary units representing a ratio of experimental to control. The following primers were used: IFN-: forward, cagcagttccagaaggagga; reverse, agtctcattccagccagtgc; CXCL10/IP-10: forward, gaatcgaaggccatcaagaa; reverse, gctcccctctggttttaagg; IRF7: forward, gagaagagcctggtcctggt; reverse, ctaggtcactcggcacag; CXCL8/IL-8: forward, gcagagggttgtggagaag; reverse, ggcatcttcactgattcttgg; CXCL5/RANTES: forward, tacaccagtggcaagtgctc; reverse, acacacttggcggttctttc; 2M: forward, cgagacatgtaagcagcatca; reverse, agcaagcaagcagaatttgg; PBDA: forward, acgatcccgagactctgcttc; reverse, gcacggctactggcacact.
Western blotting
Cells were lysed by resuspension in ice-cold hypotonic lysis buffer (10 mM HEPES (pH 7.9), 10 mM KCl, 1.5 mM MgCl2, 50 mM sodium fluoride, 1 mM Na3VO4, 1 μM PMSF, and a protease inhibitor mixture), incubated on ice for 30 min, sonicated, and spun at 16,000 x g for 10 min. Supernatants were aspirated and quantified for total protein concentration. Protein samples were separated on 10 or 12% polyacrylamide-SDS gels and transferred to polyvinylidene difluoride membranes, blocked with 5% milk in TBS-0.1% Tween 20 for 1 h, and incubated overnight with primary Ab (anti-IRF3, 1/1000 (Santa Cruz Biotechnology); anti-STAT1 Y701, 1/1000, and total STAT1, 1/1000, (Cell Signaling); and anti-Tank-binding kinase 1 (anti-TBK1), 1/500 (Gene Therapy Systems)). Membranes were washed with TBS-Tween 20 and incubated for 1 h with goat anti-rabbit HRP-linked secondary Ab (1/2000; Santa Cruz Biotechnology) in blocking buffer. Blots were washed and developed using the Supersignal West Pico chemiluminescence ECL kit (Pierce). Equivalence of loading was determined from Coomassie Blue-stained gels.
Immunoprecipitation and cell fractionation
Briefly, cells were lysed in ice-cold RIPA buffer (50 mM Tris-Cl, 1% Nonidet P-40, 0.25% deoxycholate, 150 mM NaCl, 1 mM EDTA, 1 mM PMSF, 1 mg/ml aprotinin and leupeptin, 1 mM sodium orthovanadate, and 1 mM sodium fluoride). Lysates were precleared with a 50% slurry of protein A agarose, quantified, and incubated with precipitating Ab and beads overnight at 4°C. Beads were collected by centrifugation at 14,000 rpm and washed three times, and immunocomplexes were dissociated by boiling for 5 min in SDS-PAGE sample buffer. Total lysate and lysate reacted with beads alone were also retained for analysis. Cell fractionation was performed using the NE-PER nuclear and cytoplasmic fractionation kit (Pierce) according to the manufacturer’s instructions. The purity of these fractions was determined using an affinity-purified rabbit polyclonal Ab raised against nonmuscle myosin IIB (1 μg/ml) as a cytoplasmic marker (gift from Dr. A. Bresnick, Albert Einstein College of Medicine), and a mouse mAb to proliferating cell nuclear Ag (1/2000; Sigma-Aldrich) was used as a nuclear marker.
Immunohistochemistry
Human fetal astrocytes were plated on glass-bottomed microwell dishes (MatTek) and treated with IL-1 for the indicated times. Cells were fixed either with 4% freshly prepared paraformaldehyde in PBS for 20 min on ice or in ice-cold 100% methanol for 20 min on ice. No difference was noted in the staining pattern after the two fixation procedures, and all subsequent experiments used paraformaldehyde. Cells were washed twice with PBS and once with PBS plus 0.01% glycine, blocked with 10% normal goat serum in PBS for 1 h at 25°C, and then incubated overnight at 4°C with an affinity-purified rabbit anti-IRF3 Ab (Santa Cruz Biotechnology) diluted 1/100 in blocking solution. Cells were then washed with 0.01% PBS-Tween 20 and incubated with goat anti-rabbit Alexa 488 (Molecular Probes) diluted 1/200 in PBS containing 2% BSA and 0.1% Triton. Cultures were washed in 0.01% PBS-Tween 20, nuclei visualized with 4',6-diamidino-2-phenylindole, dihydrochloride (Molecular Probes), and cells were covered with aqueous mounting medium (Biomeda). Control cultures were prepared by omitting the primary Ab.
Sandwich ELISA
ELISA was conducted on supernatants from 96-well plates. Briefly, 1 x 104 cells were infected with adenoviral constructs for 2 days before IL-1 treatment for 24 h. After 24 h of treatment with IL-1, supernatants were collected, and ELISA was performed. Samples were added to immunoadsorbent plates precoated with capture Ab to RANTES (1/125) or IP-10 (1/1000; R&D Systems) and blocked (1% BSA, 5% sucrose, and 0.01% sodium azide) for 2 h at room temperature. Plates were washed with 0.05% PBS-Tween 20, incubated for another 2 h with a biotinylated detection Ab (1/1000; R&D Systems), then again washed extensively and incubated with an HRP-streptavidin (1/1000) conjugate (E-Biosciences). A 3,3',5,5'-tetramethylbenzidine-peroxidase enzyme immunoassay substrate kit (Bio-Rad) was used to resolve the plates, and the exposure was stopped with the addition of 1 M H2SO4. The plate was read at 450 nm and referenced at 595 nm. Samples were diluted until values fell within the linear range of the ELISA detection limit.
Constructs and adenovirus
IRF3 N- and wild-type (WT)-expressing adenovirus was constructed by first excising from pCMV-BL cDNA corresponding to WT and N IRF3 at the EcoRV and XhoI sites (gift from J. Hiscott, McGill University, Quebéc, Canada). The insert was cloned into the EcoRV and XhoI sites in pBluescript, then excised using XbaI and KpnI. cDNA was subsequently ligated into the pShuttle vector (BD Biosciences). cDNA was excised according to the manufacturer’s instructions with PI-SceI and I-CeuI, then cloned into the BD-AdenoX vector. A PacI-digested linear piece of DNA containing the cDNA of N and WT IRF3 along with the adenovirus genome was transfected into HEK293 cells. At later times, supernatants were tested for production of recombinant adenovirus and expanded in culture.
Statistics
Student’s t tests were used to determine the significance of the data generated by Q-PCR.
Results
Microarray data demonstrate that IL-1-activates IRF3 response genes in primary human fetal astrocytes in culture
The initial evidence that the IFN system was up-regulated in IL-1-treated astrocytes came from a cDNA microarray screen. Primary human astrocyte cultures prepared from four cases were treated with the cytokines IL-1 (10 ng/ml) alone for 6 and 24 h or IL-1 plus IFN- (10 ng/ml) for the same time period. The gene expression patterns of these cultures were compared with those of untreated matched controls using microarrays of 28K cDNAs representing 18K unique human cDNAs. As shown in Fig. 1A, robust induction of several genes that are considered to be characteristic of ISGs was noted (26, 27). Of particular interest was evidence of induction of IFN-, IRF7, STAT1, CXCL10/IP-10, and CCL5/RANTES, because these genes have been implicated as targets of IRF3 activation after ligation of TLR3 or TLR4 (16, 28).
FIGURE 1. Microarray analysis of cytokine-induced gene expression in fetal human astrocytes, and confirmation of IRF3 dependent gene up-regulation in IL-1 treated cells. A, Cells were activated with IL-1 alone or IL-1 plus IFN- (I/I) for 6 or 24 h and processed for microarrays. Astrocytes from four different cases were studied. Data are presented as a pseudocolor map to facilitate interpretation using a log base 2 conversion factor and represent genes that were elevated for all cases analyzed. B, RT PCR was performed on 10 μg of total RNA isolated from IL-1 (10 ng/ml)-treated human primary astrocytes harvested at the times indicated. Q-PCR was then performed on the indicated IRF3-dependent genes. 2M and PBDA were both used as internal standards and for normalization of the data. Data shown are representative of five independent experiments. A.U., arbitrary units. C, Representative plate samples were resolved on a 2% agarose gel to access the specificity of the Q-PCR.
To validate the microarray data and to further substantiate activation of a set of ISGs by IL-1 in astrocytes, we performed Q-PCR for IFN-, IRF7, and the chemokines CXCL10/IP-10 and CCL5/RANTES. We also assessed activation of the chemokine CXCL8/IL-8, a known target of IL-1 signaling that is dependent upon NF-B activation, but not IRF3. As expected, IL-1 led to the robust activation of mRNA for IL-8 (Fig. 1B). In addition, the Q-PCR indicated that IFN- was activated rapidly in response to IL-1, peaking at 3 h postactivation, whereas the mRNA for CXCL10/IP-10 and CCL5/RANTES showed a more delayed response.
LPS does not activate IRF3-dependent genes in primary human astrocytes
In the human astrogliomal cell line U373, LPS has been shown to activate IRF3 as an immediate early gene via a pathway involving p38 (29). In primary human fetal astrocytes in culture, we have not found LPS to signal via the classical NF-B/MAPK-dependent pathways or to activate p38. However, signaling to IRF3 is known to be MyD88 independent, raising the possibility that LPS may indeed result in IRF3 activation in these cells. To test for this, we repeated the Q-PCR assays using LPS (1 μg/ml) as the stimulus, but detected no response (data not shown). Because TLR4 acts in concert with the LPS serum binding protein and CD14 (30), which is not expressed by astrocytes either in vitro or in vivo (31, 32), we also tested the effect of LPS (1 μg/ml) with the addition of recombinant soluble CD14 (0.1 μg/ml) and 10% serum. Again, we failed to find activation of the same set of genes, except for a low level response of IL-8 (8-fold compared with 800-fold by IL-1) at 6 h. A low level response to LPS in these cultures is usually due to minor contamination with microglia, which are sensitive to low doses of LPS due to the expression in vitro of TLR4 and CD14.
IL-1 leads to phosphorylation and translocation of IRF3 to the nucleus
In cells in which IRF3 has been activated by viral infection or ligands for TLR3, it has been shown that IRF-3 is phosphorylated as part of a macromolecular complex that contains TBK-1 and/or IB kinase-i (IKKi; also known as IKK) (33, 34, 35). To determine whether IRF-3 associates with either of these molecules in astrocytes, we attempted to coimmunoprecipitate them with an Ab to IRF-3 and probe Western blots of the precipitates with Abs to TBK-1 and IKKi. A clear signal was detected for TBK-1 on these blots, whereas no evidence for association with IKKi was observed (Fig. 2A and data not shown). Western blots also showed evidence of transient phosphorylation of IRF3, consistent with the formation of a functional signaling complex in IL-1-activated astrocytes (Fig. 2B). Because phosphorylation of IRF3 should result in the translocation of IRF3 to the nucleus, we then tested cytoplasmic and nuclear fractions of activated astrocytes by Western blotting (Fig. 3A). The data showed that IL-1 treatment of astrocytes led to increased localization of IRF3 in the nucleus that persisted through 24 h. The purity of these fractions for cross-contamination was determined using an Ab to nonmuscle myosin IIB for cytoplasmic and proliferating cell nuclear Ag for nuclear fractions (data not shown). In contrast, activation with LPS was without effect (Fig. 3B). Immunohistochemistry also confirmed translocation of IRF3 to the nucleus in activated cells (Fig. 3C).
FIGURE 2. Phosphorylation of IRF3 and its association with the kinase TBK-1 occurs in IL-1-treated human primary astrocytes. A, Human primary astrocyte cultures were treated with 10 ng/ml IL-1, and total cell lysates were harvested at the times indicated. Total protein was quantified and immunoprecipitated with Ab to IRF3 overnight at 4°C. Ab was captured on protein A agarose beads, and immunoblots were probed with Ab to TBK-1. The presence of TBK-1 in the starting lysates is shown in the lanes on the right. The bands at 50 kDa represent IgGh present in the immunoprecipitates and are shown as a loading control. B, Human astrocytes were treated with IL-1 at the times indicated, and total protein was isolated. Samples were resolved on 12% SDS-PAGE gels and immunoblotted for IRF3. Arrows indicate modified protein states. The gels were stripped and stained for -tubulin as a loading control (lower gel).
FIGURE 3. Nuclear localization of IRF3 upon IL-1 treatment in primary cultures of human astrocytes. A, Astrocytes were treated with IL-1 (10 ng/ml) for 1, 3, 6, and 24 h. Cells were lysed, and nuclear and cytoplasmic proteins were isolated. Both fractions were resolved on 12% SDS-PAGE gels and immunoblotted with anti IRF3 (1/1000). B, Astrocytes were treated with varying concentrations of LPS (0.01, 0.2, 1, and 10 μg/ml), and both nuclear and cytoplasmic proteins were harvested at 4 h, resolved using SDS-PAGE, and probed for IRF3. C, Astrocytes grown on MatTek dishes were stimulated with IL-1 for varying times. Cells were fixed, fluorescently stained for IRF3, and observed for nuclear translocation.
IFN- is an early response gene to IL-1 activation in primary human astrocytes
Activation of IRF3 is considered an essential component of the enhanceosome that leads to the production of IFN-. Signaling by IFN- through the type 1 IFN receptor then leads to an autocrine/paracrine amplification loop, resulting in transcriptional activation of additional IFN response genes that include IRF7 and, in some cells, CXCL10/IP-10 and CCL5/RANTES. The Q-PCR data showing that peak mRNA levels for IRF7, CXCL10/IP-10, and CCL5/RANTES lagged behind those for IFN- would be consistent with the possibility that these genes are downstream targets of IFN- gene expression. To test for this, cells were treated with cycloheximide (CHX) for 6 h, and mRNA expression was determined by Q-PCR. The data showed that mRNA levels for the tested genes in control or CHX-treated cells were: IFN-, 32.5 ± 10.2 vs 52.8 ± 14.5; IRF-7, 8.3 ± 2.5 vs 3.4 ± 0.7; CXCL10/IP-10, 643 ± 193 vs 48.6 ± 12.5; and CCL5/RANTES, 98.4 ± 30.7 vs 306.4 ± 140.7 (mean ± SEM; n = 4). The observation that CCL5/RANTES was enhanced by CHX is consistent with previously published data (36). Taken together, these results indicate that IFN- production occurs independently of translation after activation with IL-1 in human fetal astrocytes, whereas IL-1 induction of IRF7 and CXCL10/IP-10 is dependent upon protein synthesis.
IL-1 leads to delayed phosphorylation of STAT1
To confirm that IL-1 leads to type 1 IFN signaling in astrocytes, we assessed the phosphorylation status of STAT1 by Western blots using an Ab to phospho-STAT (Y701; Fig. 4). The results showed that in astrocytes activated with IL-1, STAT1 phosphorylation occurred in a time frame consistent with an autocrine response to IFN-. In contrast, phosphorylation of STAT1 was rapid after direct stimulation with IFN-. The Western blot also showed induction of STAT1 protein at 24 h by IL-1, consistent with the microarray data that showed increased mRNA for STAT1 in IL-1-treated astrocytes (see Fig. 1).
FIGURE 4. IL-1 induces STAT1 phosphorylation. Astrocytes were treated directly with either IL-1 (10 ng/ml) or IFN- (10 ng/ml) for the times indicated. Total cell protein was isolated and resolved by SDS-PAGE. The gels were probed with both anti-phospho-STAT-1 and total STAT-1 Ab.
Dominant-negative construct of IRF3 inhibits IL-1-induced IFN-, IRF7, CXCL10/IP-10, and CCL5/RANTES expression
We next observed what the effect of inhibiting IRF3 activity would have on the expression of IRF3-dependent ISG. To do this, we constructed an adenoviral vector harboring an IRF3 construct in which the N-terminal DNA-binding domain had been deleted. This construct has been shown previously to function as a dominant negative form of IRF3 (N) (37). Q-PCR of astrocytes pretreated with N-IRF3, then stimulated with IL-1, revealed a significant reduction in the expression of mRNA for IFN-, IRF7, CXCL10/IP-10, and CCL5/RANTES (Fig. 5A). In contrast, N-IRF3 had no effect on mRNA for CXCL8/IL-8.
FIGURE 5. Dominant negative IRF3 diminishes IRF3-dependent gene expression in IL-1-stimulated astrocytes, and IL-1 potentiates gene expression in WT-IRF3-expressing cells. A, Total RNA was isolated from astrocytes that had been infected with either a dominant negative adenoviral construct (multiplicity of infection, 10) or a WT adenoviral construct (multiplicity of infection, 10) 48 h before stimulation with IL-1 (10 ng/ml) for the indicated times. Q-PCR was performed on samples using SYBR Green. Data were normalized to the internal control standards 2M and PBDA and are presented as fold changes above control values. For the experiments using the dominant negative construct, astrocytes from at least three different cases were studied, and for each dataset the results are shown as the mean ± SD (n = 9). For the experiments using the WT construct, astrocytes from at least two different cases were studied, and for each dataset the results are shown as the mean ± SD (n = 3). *, p < 0.03 vs controls. B, Astrocyte cultures were pretreated with adenovirus for 2 days at an multiplicity of infection of 10. After pretreatment, fresh medium was added along with IL-1 (10 ng/ml) for 24 h. Supernatants were isolated and assayed for CCL5/RANTES, CXCL10/IP-10, and CXCL8/IL-8 by ELISA. C, Total cell lysates from adenovirus-infected cells were harvested and resolved on a 12% polyacrylamide gel by SDS-PAGE. The lower bands shown in the lane labeled N represent the IRF-3 construct lacking the N-terminal domain. Data shown are representative of two independent experiments.
In addition to an empty vector control, we generated a vector containing WT-IRF3, which was used in some experiments. Cells transfected with WT-IRF3 showed a superinduction after activation with IL-1 for the chemokines CXCL10/IP-10 and CCL5/RANTES, whereas induction of CXCL8/IL-8 was unaffected (Fig. 5A). ELISAs on supernatants harvested at 24 h from cells infected with both N-IRF3 and WT-IRF3 confirmed the reduction and enhancement of IRF3-dependent chemokine production in IL-1-stimulated astrocytes (Fig. 5B). This potentiation by IL-1 of chemokine expression in cells overexpressing IRF-3 (Fig. 5C) underscores our hypothesis that IL-1 signaling is capable of activating IRF3, leading to the induction of an antiviral response in primary human astrocytes in culture.
Discussion
In this report we show that treatment of human fetal astrocytes with the cytokine IL-1 leads to activation of the transcription factor IRF3 and induction of a set of immediate-early type 1 IFN-dependent genes that play a critical role in the host defense against pathogens. Key indicators of this activation of the IFN system by IL-1 in astrocytes include evidence of phosphorylation and translocation of IRF3 to the nucleus, up-regulation of mRNA for IFN-, evidence for downstream signaling via STAT1, induction of the secondary response gene IRF7 (38), and induction of the chemokines CXCL10/IP-10 and CCR5/RANTES. The robust induction by IL-1 of additional ISG, as noted on the microarrays such as STAT1, 2',5'-oligoadenylate synthetase 2, and ISG15, also support an active signaling role for IL-1 via this pathway in human fetal astrocytes.
In addition to viral infection, it is now well accepted that ligands for TLR3 and TLR4 activate IRF3, leading to the production of IFN- and the chemokines CXCL10/IP-10 and CCL5/RANTES (15, 16, 29). TLR play a critical role in the innate immune response by recognizing structurally conserved bacterial and viral elements termed pathogen-associated molecular patterns (39). Different TLRs recognize different ligands, but share common signal transduction pathways through expression of Toll-IL-1R homologous region domain-containing molecules that link to adapter proteins such as MyD88. For TLR4, additional studies in mice in which the gene for MyD88 had been inactivated demonstrated the presence of an alternative signaling pathway that directed downstream signaling in the absence of MyD88 (40, 41) and led to the activation of IRF3 (15, 16). Consistent with this alternative activation pathway for TLR4, the ligand LPS was found to activate IRF3 in the human astrogliomal cell line U373 (29). It could be argued, therefore, that in our astrocyte cultures it is LPS, rather than IL-1, that is providing a signal for activation of IRF3. However, as we have reported previously, we were unable to detect a response to LPS in our cultures of primary human astrocytes, including a failure of LPS to activate IL-8, which is exquisitely responsive to LPS in other systems (data not shown), suggesting that LPS does not signal via either the MyD88-dependent or -independent pathway in these cells. This is in stark contrast to human microglia isolated from these same tissues, which respond rapidly and robustly to LPS activation (25).
Upon activation by viral infection, IRF3 is phosphorylated by a virus-activated kinase at multiple serine and threonine sites, whereas activation by stress inducers leads to a more restricted phosphorylation pattern in the N-terminal region via a MAPK-related pathway (24). The transient shift in the migration pattern of IRF3 that we noted on our Western blots resembled that found after activation of IRF3 by stress inducers. The kinases that have been found to be involved in phosphorylation of IRF3 are the IKK-related kinases TBK-1 and IKKi (33, 34, 35). By coimmunoprecipitation studies we found that IRF3 associated with TBK1, but not IKKi, in IL-1-stimulated astrocytes. These data are, therefore, in agreement with results obtained in mouse macrophages and mouse fibroblasts, where it was found that TBK1 is the critical kinase required for IRF3 phosphorylation after activation via the nonviral pathway (42, 43). The presence of only one of these activating kinases associated with IRF3 in activated astrocytes could also account for the minimal and transient phosphorylation we observed 30 min after IL-1 treatment.
In LPS-stimulated murine B cells, IFN-, CXCL10/IP-10, and CCL5/RANTES have all been reported to be primary response genes after activation of IRF3 (16). However, in our system only IFN- gene expression was unaffected by CHX treatment, suggesting that CXCL10/IP-10 as well as possibly IRF7 are secondary response genes. This result would be consistent with the data derived from IFN--deficient mice, in which it has been shown that the production of CXCL10/IP-10 is mainly a consequence of IFN- production (44, 45). Similarly, after exposure to DNA-damaging agents, direct transcriptional activation of the CCL5/RANTES promoter was noted in one study (22), but not in another (46). Whether these discrepancies reflect differences in the activation of other transcription factors required for successful activation of CCL5/RANTES and CXCL10/IP-10 or represent cell-type differences in the specific kinases involved awaits further study.
In our experiments we observed a time lag between the onset of phosphorylation of IRF3 and the initial expression of IFN-. These data suggest that the initial induction of IFN- occurs as a result of IRF3 activation, but is later potentiated by the recruitment of other IRF family members as well as NF-B to the IFN- promoter (47). The IRF family consists of nine members, of which IRF1, -3, -7, and -9 have been implicated in the transcriptional regulation of IFN- genes (20, 48, 49). Of these IRFs, it has been suggested that IRF7 forms a key component of the late induction phase of IFN- (38, 50). In most cell types, IRF7 is not expressed constitutively, but is induced by IFN- via the JAK-STAT signaling pathway. This newly synthesized IRF7 is then thought to participate in the production of IFN- as well as IFN- (19).
The transcriptional induction of type 1 IFNs is central to the cellular antiviral response and plays an important role in the innate host defense against infection. However, type 1 IFNs also trigger a strong inflammatory response, inhibit cell proliferation, and induce apoptosis, suggesting a need for negative regulatory processes (51). Consistent with this idea, it has been shown that after activation, IRF3 is rapidly degraded via the ubiquitin-proteasome pathway (52), and IRF7 is unstable in most cell types (51). Our data showing that IFN- mRNA was only transiently elevated in astrocytes after activation by IL-1 probably also reflects the presence of a regulatory pathway. Specific cell types may also regulate downstream signaling effects by blocking phosphorylation of STAT1, as has been noted in alveolar vs peritoneal macrophages (53). In these cells, the authors argue that, given the vulnerability of lung tissue to excessive inflammation, activation of the innate immune response needs to be limited. Their data show that although the TLR-induced production of chemokines is intact in alveolar macrophages, the production of NO was not observed unless exogenous IFN was added. A similar phenomenon may exist in the CNS, where inflammation is also tightly regulated. In the inflamed CNS, astrocytes express CXCL10/IP-10, particularly in association with the glial endfeet that abut the vasculature (54). Astrocytes both in vitro and in vivo also express the inducible form of NO synthase (iNOS), but in vitro, IL-1, although necessary to initiate gene expression, leads to only low level production of NO unless IFN is present (55). We have shown previously that both IFN- and IFN- can enhance IL-1-induced iNOS expression when given alone; however, when IFN- and IFN- are added together, IFN- acts as a negative regulator of iNOS expression through a mechanism that involves inhibition of IFN--activating sequence-binding activity (56). In murine macrophages, IFN- has also been found to antagonize the activity of IFN- through inhibition of NF-B signaling (57) and down-regulating IFN--induced FcR expression (58) and IFN--induced MHC class II expression (59). Thus, the specific effects of IFN- at sites of inflammation are probably diverse and critically regulated by the presence or the absence of other cytokines in injured tissues. These data underscore the importance of understanding the complex pathways involved in inflammatory gene expression in astrocytes, which, it is to be hoped, will be of value in defining selective modes of intervention for inflammatory events in the CNS.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank Dr. John Hiscott (McGill University, Montreal, Canada) for the N and Wt IRF3 constructs, Dr. Brad Poulos for tissue acquisition, Wa Shen for tissue preparation, and Dr. Anne Bresnick (Albert Einstein College of Medicine) for affinity-purified rabbit polyclonal Ab to non-muscle myosin-IIB.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by U.S. Public Health Service Grants NS40137, NS11920, NS07098, MH55477, and NS046620, National Multiple Sclerosis Society Grant RG3444, and The Jayne and Harvey Beker Foundation (to G.R.J.).
2 Address correspondence and reprint requests to Dr. Mark A. Rivieccio, Department of Pathology, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461. E-mail address: mriviecc{at}aecom.yu.edu
3 Abbreviations used in this paper: IKK, IB kinase; aRNA, amplified antisense RNA; CXCL10/IP-10, IFN-induced protein (10 kDa); CHX, cycloheximide; IKKi, induced B kinase; iNOS, inducible form of NO synthase; IP-10, IFN--inducible protein-10; IRF, IFN regulatory factor; ISG, IFN-stimulated gene; ISRE, IFN-stimulated response element; 2M, 2-microglobulin; PBDA, porphobilinogen deaminase; Q-PCR, real-time quantitative PCR; TBK, Tank-binding kinase-1; WT, wild type.
Received for publication September 23, 2004. Accepted for publication January 7, 2005.
References
Janzer, R. C., M. C. Raff. 1987. Astrocytes induce blood-brain barrier properties in endothelial cells. Nature 325:253.
Anderson, C. M., R. A. Swanson. 2000. Astrocyte glutamate transport: review of properties, regulation, and physiological functions. Glia 32:1.
Liberto, C. M., P. J. Albrecht, L. M. Herx, V. W. Yong, S. W. Levison. 2004. Pro-regenerative properties of cytokine-activated astrocytes. J. Neurochem. 89:1092
Sofroniew, M. V., T. G. Bush, N. Blumauer, L. Kruger, L. Mucke, M. H. Johnson. 1999. Genetically-targeted and conditionally-regulated ablation of astroglial cells in the central, enteric and peripheral nervous systems in adult transgenic mice. Brain Res. 835:91.
Herx, L. M., V. W. Yong. 2001. Interleukin-1 is required for the early evolution of reactive astrogliosis following CNS lesion. J. Neuropathol. Exp. Neurol. 60:961.
John, G. R., L. Chen, M. A. Rivieccio, C. V. Melendez-Vasquez, A. Hartley, C. F. Brosnan. 2004. Interleukin-1 induces a reactive astroglial phenotype via deactivation of the Rho GTPase-Rock axis. J. Neurosci. 24:2837.
John, G. R., S. C. Lee, X. Song, M. Rivieccio, C. F. Brosnan. 2005. IL-1-regulated responses in astrocytes: relevance to injury and recovery. Glia. 49:161.
O’Neill, L. A., C. Greene. 1998. Signal transduction pathways activated by the IL-1 receptor family: ancient signaling machinery in mammals, insects, and plants. J. Leukocyte Biol. 63:650.
Zhang, P., B. S. Miller, S. A. Rosenzweig, N. R. Bhat. 1996. Activation of c-Jun N-terminal kinase/stress-activated protein kinase in primary glial cultures. J. Neurosci. Res. 46:114.
Ninomiya-Tsuji, J., K. Kishimoto, A. Hiyama, J. Inoue, Z. Cao, K. Matsumoto. 1999. The kinase TAK1 can activate the NIK-IB as well as the MAP kinase cascade in the IL-1 signalling pathway. Nature 398:252.
Radons, J., S. Dove, D. Neumann, R. Altmann, A. Botzki, M. U. Martin, W. Falk. 2003. The interleukin 1 (IL-1) receptor accessory protein Toll/IL-1 receptor domain: analysis of putative interaction sites in vitro mutagenesis and molecular modeling. J. Biol. Chem. 278:49145.
Dunne, A., M. Ejdeback, P. L. Ludidi, L. A. O’Neill, N. J. Gay. 2003. Structural complementarity of Toll/interleukin-1 receptor domains in Toll-like receptors and the adaptors Mal and MyD88. J. Biol. Chem. 278:41443.
Subramaniam, S., C. Stansberg, C. Cunningham. 2004. The interleukin 1 receptor family. Dev. Comp. Immunol. 28:415.
Takeda, K., T. Kaisho, S. Akira. 2003. Toll-like receptors. Annu. Rev. Immunol. 21:335.
Kawai, T., O. Takeuchi, T. Fujita, J. Inoue, P. F. Muhlradt, S. Sato, K. Hoshino, S. Akira. 2001. Lipopolysaccharide stimulates the MyD88-independent pathway and results in activation of IFN-regulatory factor 3 and the expression of a subset of lipopolysaccharide-inducible genes. J. Immunol. 167:5887.
Doyle, S., S. Vaidya, R. O’Connell, H. Dadgostar, P. Dempsey, T. Wu, G. Rao, R. Sun, M. Haberland, R. Modlin, et al 2002. Toll-like receptor 3 mediates a more potent antiviral response than Toll-like receptor 4. Immunity 17:251.
Smith, E. J., I. Marie, A. Prakash, A. Garcia-Sastre, D. E. Levy. 2001. IRF3 and IRF7 phosphorylation in virus-infected cells does not require double-stranded RNA-dependent protein kinase R or IB kinase but is blocked by vaccinia virus E3L protein. J. Biol. Chem. 276:8951.
Jiang, Z., T. W. Mak, G. Sen, X. Li. 2004. Toll-like receptor 3-mediated activation of NF-B and IRF3 diverges at Toll-IL-1 receptor domain-containing adapter inducing IFN-. Proc. Natl. Acad. Sci. USA 101:3533
Kawai, T., S. Sato, K. J. Ishii, C. Coban, H. Hemmi, M. Yamamoto, K. Terai, M. Matsuda, J. Inoue, S. Uematsu, et al 2004. Interferon- induction through Toll-like receptors involves a direct interaction of IRF7 with MyD88 and TRAF6. Nat. Immunol. 5:1061.
Mamane, Y., C., C. Heylbroeck, P. Genin, M. Algarte, M. J. Servant, C. DeLuca, H. Kwon, R. Lin, J. Hiscott. 1999. Interferon regulatory factors: the next generation. Gene 237:1.
Hiscott, J., P. Pitha, P. Genin, H. Nguyen, C. Heylbroeck, Y. Mamane, M. Algarte, R. Lin. 1999. Triggering the interferon response: the role of IRF-3 transcription factor. J. Interferon Cytokine Res. 19:1.
Kim, T., T. Y. Kim, Y. H. Song, I. M. Min, J. Yim, T. K. Kim. 1999. Activation of interferon regulatory factor 3 in response to DNA-damaging agents. J. Biol. Chem. 274:30686
Lopez-Collazo, E., S. Hortelano, A. Rojas, L. Bosca. 1998. Triggering of peritoneal macrophages with IFN-/ attenuates the expression of inducible nitric oxide synthase through a decrease in NF-B activation. J. Immunol. 160:2889.
Van Weyenbergh, J., P. Lipinski, A. Abadie, D. Chabas, U. Blank, R. Liblau, J. Wietzerbin. 1998. Antagonistic action of IFN- and IFN- on high affinity Fc receptor expression in healthy controls and multiple sclerosis patients. J. Immunol. 161:1568.
Lu, H. T., J. L. Riley, G. T. Babcock, M. Huston, G. R. Stark, J. M. Boss, R. M. Ransohoff. 1995. Interferon (IFN) acts downstream of IFN--induced class II transactivator messenger RNA accumulation to block major histocompatibility complex class II gene expression and requires the 48-kDa DNA-binding protein, ISGF3-. J. Exp. Med. 182:1517.(Mark A. Rivieccio,, Garet)
The cytokine IL-1 is a major activator of primary human fetal astrocytes in culture, leading to the production of a wide range of cytokines and chemokines important in the host defense against pathogens. IL-1, like TLR4, signals via the MyD88/IL-1R-associated kinase-1 pathway linked to activation of NF-B and AP-1. Recent studies have shown that TLR4 also signals independently of MyD88, resulting in the activation of IFN regulatory factor 3 (IRF3), a transcription factor required for the production of primary antiviral response genes such as IFN-. Using a functional genomics approach, we observed that IL-1 induced in astrocytes a group of genes considered to be IFN-stimulated genes (ISG), suggesting that IL-1 may also signal via IRF3 in these cells. We now show, using real-time PCR, that in astrocytes IL-1 induces the expression of IFN-, IRF7, CXCL10/IFN--inducible protein-10, and CCL5/RANTES. Chemokine expression was confirmed by ELISA. We also show that IL-1 induces phosphorylation and nuclear translocation of IRF3 and delayed phosphorylation of STAT1. The dependency of IFN-, IRF7, and CXCL10/IFN--inducible protein-10 gene expression on IRF3 was confirmed using a dominant negative IRF3-expressing adenovirus. The robust induction by IL-1 of additional ISG noted on the microarrays, such as STAT1, 2'5'-oligoadenylate synthetase 2, and ISG15, also supports an active signaling role for IL-1 via this pathway in human fetal astrocytes. These data are the first to show that IL-1, in addition to TLRs, can stimulate IRF3, implicating this cytokine as an activator of genes involved in innate antiviral responses in astrocytes.
Introduction
Astrocytes are the major glial cell population in the CNS. They reside ubiquitously in the brain parenchyma as highly ramified cells and play an important role in the maintenance of brain homeostasis. Injury to the CNS is inevitably accompanied by the rapid induction of a reactive astrogliosis, which is characterized by astrocytic hypertrophy, proliferation, and altered gene expression. Although historically the formation of a glial scar has been considered a major impediment to functional recovery of the CNS, reactive astrocytes also play an important role in repair mechanisms by removing toxic levels of factors such as glutamate, functioning as a source of growth factors and in restoration of the blood-brain barrier (1, 2, 3, 4).
A major factor that has been implicated in the induction of a reactive astrocytic phenotype is the cytokine IL-1 (5, 6). Using a functional genomics approach, we have shown that exposure of human fetal astrocytes to IL-1 for 24 h leads to the altered expression of 1000 genes, including the increased expression of multiple acute phase genes involved in the early phases of an inflammatory response, such as chemokines, cytokines, and adhesion molecules (7). The signaling cascade activated by IL-1 in most cell types has been well documented. After ligand binding to the type 1 IL-1R, the cytosolic adaptors MyD88 and Toll-interacting protein (Tollip) are rapidly recruited to the receptor complex, which then leads to binding of the IL-1R-associated kinase-4. IL-1R-associated kinase-4 is phosphorylated as a result of this interaction, mediating the recruitment of the TNF-associated factor-6. TNF-associated factor-6, through a series of other adaptor proteins, leads to activation of the downstream IB kinases (IKK and IKK), 3 which phosphorylate the NF-B repressor IB, resulting in the release of inhibition and translocation of NF-B complexes to the nucleus, with subsequent NF-B-dependent gene expression (8). A similar cascade of kinase-dependent events leads to activation of the MAPK pathway and JNK, resulting in the activation of additional transcription factors, including ATF and AP1 (9, 10). Signaling via these pathways ultimately leads to the production of proinflammatory cytokines and chemokines important in host defense as well as in the establishment of the acquired immune response.
The intracellular signaling domain of the IL-1R is homologous to that of the signaling domain of TLR4 and is termed the Toll-IL-1R homologous region (11, 12, 13). TLR family members are evolutionarily conserved proteins that are critically involved in host defense and are important mediators of inflammation and innate and adaptive immune responses. LPS derived from the cell wall of Gram-negative bacteria is a major ligand for TLR4 and is a potent activator of proinflammatory cytokines and chemokines such as IL-1, TNF, IL-6, and IL-8, which are dependent upon the MyD88 signaling pathway. However, studies of mice in which the gene for MyD88 had been inactivated, demonstrated that LPS could also signal independently of MyD88 to produce an alternative set of effector responses, leading to the production of IFN- and IFN-stimulated genes (ISG), such as CXCL10/IFN--inducible protein-10 (IP-10) (14, 15, 16). Similar cell signaling pathways have been found to be activated after binding of dsRNA to TLR3 (17, 18). IFN regulatory factor 3 (IRF3) and IRF7 are thought to be responsible for virus- or TLR ligand-induced type 1 IFNs (19). IRF3 is constitutively expressed in the cytoplasm of most cells, but upon activation by phosphorylation at multiple sites, it homodimerizes and translocates to the nucleus, where it functions as a required transcription factor in the initiation of type 1 IFN gene expression (20, 21). Autocrine and/or paracrine signaling through the type I IFN receptor then leads to phosphorylation of STAT1 and activation of a type 1 IFN-dependent gene program that includes the expression of IRF7 and the chemokines CXCL10/IP-10 and CCL5/RANTES as well as an autocrine positive feedback loop that boosts type 1 IFN production (16). In cells that show this response to LPS through TLR4, IL-1 was not found to activate this signaling pathway (22). These observations have led to the conclusion that although the activation of NF-B is a conserved response after activation of most TLRs and the IL-1R, the activation of IRF3 is a response unique to the TLR3 and TLR4 pathways as well as to infection by certain viruses. However, studies have also shown that the pathways involved in IRF3 activation may be both stimulus and cell type-specific (23, 24).
In this study we have addressed the question of whether the cytokine IL-1, which is a major activator of the NF-B and MAPK pathways in human fetal astrocytes, also leads to the activation of IRF3 in these cells. The results support the conclusion that IRF3 represents a target of IL-1 signaling in astrocytes, leading to the activation of IFN- and a group of genes considered to be secondary response genes to type 1 IFN signaling. These data suggest that astrocytes are endowed with a specific function in the brain resembling that of pathogen-activated brain macrophages/microglia. Activation of microglia leads to the production of IL-1, which then activates astrocytes, forming a second line of defense in the brain that amplifies the antimicrobial and antiviral innate immune responses of microglia.
Materials and Methods
Astrocyte cultures and cytokines
Enriched human fetal brain astrocyte cultures were established from second trimester abortuses as described previously (25). All tissue collection was approved by the institutional clinical review committee. To obtain cultures essentially free of microglia, cells were plated at 4 x 107 cells/10 ml medium/T75 tissue culture flask. After 14 days in vitro, floating microglia were removed, and astrocytes were subcultured by trypsinization for at least three passages. Culture purity was determined by immunohistochemistry for glial fibrillary acidic protein for astrocytes (BioGenex), microtubule-associated protein-2 for neurons (Sigma-Aldrich), and CD68 for microglia (DakoCytomation). Microglia, which are nondividing cells in this system, do not survive this subculturing process, and enriched astrocyte cultures typically contain <0.01% microglia contamination. Recombinant human IL-1 was a gift from the Biological Response Modifiers Program at National Cancer Institute or was purchased from PeproTech, and IFN- was purchased from Genzyme. Cytokines were diluted in medium containing 2 mg/ml endotoxin-free human serum albumin (Baxter Healthcare). Astrocytes were activated with IL-1 at 10 ng/ml (equivalent to 20 U/ml). Recombinant human CD14 was purchased from Biometec. All other reagents were purchased from Sigma-Aldrich unless otherwise indicated.
cDNA microarray
Total astrocyte (control and IL-1-treated) RNA was extracted using the RNeasy Mini Procedure (Qiagen). RNA samples tested for quality using the Agilent 2100 bioanalyzer were used for microarray analysis and were amplified once using the MessageAmp amplified antisense RNA (aRNA) kit (Ambion) before microarray analysis. Four different astrocyte cases were studied. Briefly, 5 μg of total RNA was reverse transcribed, followed by second-strand cDNA synthesis. After cDNA purification, aRNA was synthesized by in vitro transcription. Human cDNA arrays were obtained from the Albert Einstein College of Medicine cDNA Microarray Facility (relevant data available from http://microarray1k.aecom.yu.edu/). Each slide contained an unbiased, random collection of 27K cDNA probe elements derived from the sequence-verified GEM1 clone set (Incyte Genomics). Microarray procedures were performed according to protocols provided by the Microarray Facility (available from http://microarray1k.aecom.yu.edu/). In brief, after RT of aRNA (5 μg), IL-1-treated samples were labeled with Cy5-fluorescent nucleotides (Amersham Biosciences), and control samples were labeled with Cy3-fluorescent nucleotides. For each experiment, Cy5-labeled astrocyte cDNA was cohybridized with Cy3-labeled control cDNA. Hybridized slides were scanned using a GenePix 4000b scanner (Axon Instruments), and raw data files were generated containing measurements of signal and background fluorescence emissions. Data from spots containing artifacts were eliminated from further analysis. Data were also filtered, and only spots with average channel intensity over the average channel background intensity were analyzed further. The ratio of Cy5 (experimental values) and Cy3 (control RNA) was calculated for each spot to derive a relative expression value. Genes with differential expression in the experimental groups were compared by their spot ratio values.
Real-time quantitative PCR (Q-PCR)
Total RNA was isolated using the TRIzol method (Invitrogen Life Technologies) according to the manufacturer’s instructions. Briefly, confluent 100-mm cell culture dishes of astrocytes were lysed and scraped into microtubes. The lysis solution was then extracted with chloroform, and the aqueous phase was transferred to a fresh tube. RNA was precipitated with 2-propanol, washed once in 70% ethanol, and resuspended in a suitable volume of nuclease-free water. The RNA suspension was then treated to remove contaminating genomic DNA with DNase I, re-extracted, and precipitated. The RNA concentration was determined by measuring the OD values of the samples at 260 nm. Q-PCR was performed first by reverse transcribing 10 μg of total RNA in a reaction volume of 20 μl. Q-PCR was conducted in 384-well reaction plates. Each reaction sample comprised a 1:1 mixture of diluted (1/100) RT-PCR sample to SYBR Green Q-PCR Master Mix (Applied Biosystems) combined with a 1:1 mixture of gene-specific forward and reverse primers to the same SYBR Green Master Mix, resulting in a total volume of 8 μl/reaction. The plates were processed in an ABI PRISM 7000 light cycler (Applied Biosystems) using standard cycling conditions. All runs were accompanied by the two internal control genes 2-microglobulin (2M) and porphobilinogen deaminase (PBDA). Samples were normalized using a Ct-based algorithm to give arbitrary units representing a ratio of experimental to control. The following primers were used: IFN-: forward, cagcagttccagaaggagga; reverse, agtctcattccagccagtgc; CXCL10/IP-10: forward, gaatcgaaggccatcaagaa; reverse, gctcccctctggttttaagg; IRF7: forward, gagaagagcctggtcctggt; reverse, ctaggtcactcggcacag; CXCL8/IL-8: forward, gcagagggttgtggagaag; reverse, ggcatcttcactgattcttgg; CXCL5/RANTES: forward, tacaccagtggcaagtgctc; reverse, acacacttggcggttctttc; 2M: forward, cgagacatgtaagcagcatca; reverse, agcaagcaagcagaatttgg; PBDA: forward, acgatcccgagactctgcttc; reverse, gcacggctactggcacact.
Western blotting
Cells were lysed by resuspension in ice-cold hypotonic lysis buffer (10 mM HEPES (pH 7.9), 10 mM KCl, 1.5 mM MgCl2, 50 mM sodium fluoride, 1 mM Na3VO4, 1 μM PMSF, and a protease inhibitor mixture), incubated on ice for 30 min, sonicated, and spun at 16,000 x g for 10 min. Supernatants were aspirated and quantified for total protein concentration. Protein samples were separated on 10 or 12% polyacrylamide-SDS gels and transferred to polyvinylidene difluoride membranes, blocked with 5% milk in TBS-0.1% Tween 20 for 1 h, and incubated overnight with primary Ab (anti-IRF3, 1/1000 (Santa Cruz Biotechnology); anti-STAT1 Y701, 1/1000, and total STAT1, 1/1000, (Cell Signaling); and anti-Tank-binding kinase 1 (anti-TBK1), 1/500 (Gene Therapy Systems)). Membranes were washed with TBS-Tween 20 and incubated for 1 h with goat anti-rabbit HRP-linked secondary Ab (1/2000; Santa Cruz Biotechnology) in blocking buffer. Blots were washed and developed using the Supersignal West Pico chemiluminescence ECL kit (Pierce). Equivalence of loading was determined from Coomassie Blue-stained gels.
Immunoprecipitation and cell fractionation
Briefly, cells were lysed in ice-cold RIPA buffer (50 mM Tris-Cl, 1% Nonidet P-40, 0.25% deoxycholate, 150 mM NaCl, 1 mM EDTA, 1 mM PMSF, 1 mg/ml aprotinin and leupeptin, 1 mM sodium orthovanadate, and 1 mM sodium fluoride). Lysates were precleared with a 50% slurry of protein A agarose, quantified, and incubated with precipitating Ab and beads overnight at 4°C. Beads were collected by centrifugation at 14,000 rpm and washed three times, and immunocomplexes were dissociated by boiling for 5 min in SDS-PAGE sample buffer. Total lysate and lysate reacted with beads alone were also retained for analysis. Cell fractionation was performed using the NE-PER nuclear and cytoplasmic fractionation kit (Pierce) according to the manufacturer’s instructions. The purity of these fractions was determined using an affinity-purified rabbit polyclonal Ab raised against nonmuscle myosin IIB (1 μg/ml) as a cytoplasmic marker (gift from Dr. A. Bresnick, Albert Einstein College of Medicine), and a mouse mAb to proliferating cell nuclear Ag (1/2000; Sigma-Aldrich) was used as a nuclear marker.
Immunohistochemistry
Human fetal astrocytes were plated on glass-bottomed microwell dishes (MatTek) and treated with IL-1 for the indicated times. Cells were fixed either with 4% freshly prepared paraformaldehyde in PBS for 20 min on ice or in ice-cold 100% methanol for 20 min on ice. No difference was noted in the staining pattern after the two fixation procedures, and all subsequent experiments used paraformaldehyde. Cells were washed twice with PBS and once with PBS plus 0.01% glycine, blocked with 10% normal goat serum in PBS for 1 h at 25°C, and then incubated overnight at 4°C with an affinity-purified rabbit anti-IRF3 Ab (Santa Cruz Biotechnology) diluted 1/100 in blocking solution. Cells were then washed with 0.01% PBS-Tween 20 and incubated with goat anti-rabbit Alexa 488 (Molecular Probes) diluted 1/200 in PBS containing 2% BSA and 0.1% Triton. Cultures were washed in 0.01% PBS-Tween 20, nuclei visualized with 4',6-diamidino-2-phenylindole, dihydrochloride (Molecular Probes), and cells were covered with aqueous mounting medium (Biomeda). Control cultures were prepared by omitting the primary Ab.
Sandwich ELISA
ELISA was conducted on supernatants from 96-well plates. Briefly, 1 x 104 cells were infected with adenoviral constructs for 2 days before IL-1 treatment for 24 h. After 24 h of treatment with IL-1, supernatants were collected, and ELISA was performed. Samples were added to immunoadsorbent plates precoated with capture Ab to RANTES (1/125) or IP-10 (1/1000; R&D Systems) and blocked (1% BSA, 5% sucrose, and 0.01% sodium azide) for 2 h at room temperature. Plates were washed with 0.05% PBS-Tween 20, incubated for another 2 h with a biotinylated detection Ab (1/1000; R&D Systems), then again washed extensively and incubated with an HRP-streptavidin (1/1000) conjugate (E-Biosciences). A 3,3',5,5'-tetramethylbenzidine-peroxidase enzyme immunoassay substrate kit (Bio-Rad) was used to resolve the plates, and the exposure was stopped with the addition of 1 M H2SO4. The plate was read at 450 nm and referenced at 595 nm. Samples were diluted until values fell within the linear range of the ELISA detection limit.
Constructs and adenovirus
IRF3 N- and wild-type (WT)-expressing adenovirus was constructed by first excising from pCMV-BL cDNA corresponding to WT and N IRF3 at the EcoRV and XhoI sites (gift from J. Hiscott, McGill University, Quebéc, Canada). The insert was cloned into the EcoRV and XhoI sites in pBluescript, then excised using XbaI and KpnI. cDNA was subsequently ligated into the pShuttle vector (BD Biosciences). cDNA was excised according to the manufacturer’s instructions with PI-SceI and I-CeuI, then cloned into the BD-AdenoX vector. A PacI-digested linear piece of DNA containing the cDNA of N and WT IRF3 along with the adenovirus genome was transfected into HEK293 cells. At later times, supernatants were tested for production of recombinant adenovirus and expanded in culture.
Statistics
Student’s t tests were used to determine the significance of the data generated by Q-PCR.
Results
Microarray data demonstrate that IL-1-activates IRF3 response genes in primary human fetal astrocytes in culture
The initial evidence that the IFN system was up-regulated in IL-1-treated astrocytes came from a cDNA microarray screen. Primary human astrocyte cultures prepared from four cases were treated with the cytokines IL-1 (10 ng/ml) alone for 6 and 24 h or IL-1 plus IFN- (10 ng/ml) for the same time period. The gene expression patterns of these cultures were compared with those of untreated matched controls using microarrays of 28K cDNAs representing 18K unique human cDNAs. As shown in Fig. 1A, robust induction of several genes that are considered to be characteristic of ISGs was noted (26, 27). Of particular interest was evidence of induction of IFN-, IRF7, STAT1, CXCL10/IP-10, and CCL5/RANTES, because these genes have been implicated as targets of IRF3 activation after ligation of TLR3 or TLR4 (16, 28).
FIGURE 1. Microarray analysis of cytokine-induced gene expression in fetal human astrocytes, and confirmation of IRF3 dependent gene up-regulation in IL-1 treated cells. A, Cells were activated with IL-1 alone or IL-1 plus IFN- (I/I) for 6 or 24 h and processed for microarrays. Astrocytes from four different cases were studied. Data are presented as a pseudocolor map to facilitate interpretation using a log base 2 conversion factor and represent genes that were elevated for all cases analyzed. B, RT PCR was performed on 10 μg of total RNA isolated from IL-1 (10 ng/ml)-treated human primary astrocytes harvested at the times indicated. Q-PCR was then performed on the indicated IRF3-dependent genes. 2M and PBDA were both used as internal standards and for normalization of the data. Data shown are representative of five independent experiments. A.U., arbitrary units. C, Representative plate samples were resolved on a 2% agarose gel to access the specificity of the Q-PCR.
To validate the microarray data and to further substantiate activation of a set of ISGs by IL-1 in astrocytes, we performed Q-PCR for IFN-, IRF7, and the chemokines CXCL10/IP-10 and CCL5/RANTES. We also assessed activation of the chemokine CXCL8/IL-8, a known target of IL-1 signaling that is dependent upon NF-B activation, but not IRF3. As expected, IL-1 led to the robust activation of mRNA for IL-8 (Fig. 1B). In addition, the Q-PCR indicated that IFN- was activated rapidly in response to IL-1, peaking at 3 h postactivation, whereas the mRNA for CXCL10/IP-10 and CCL5/RANTES showed a more delayed response.
LPS does not activate IRF3-dependent genes in primary human astrocytes
In the human astrogliomal cell line U373, LPS has been shown to activate IRF3 as an immediate early gene via a pathway involving p38 (29). In primary human fetal astrocytes in culture, we have not found LPS to signal via the classical NF-B/MAPK-dependent pathways or to activate p38. However, signaling to IRF3 is known to be MyD88 independent, raising the possibility that LPS may indeed result in IRF3 activation in these cells. To test for this, we repeated the Q-PCR assays using LPS (1 μg/ml) as the stimulus, but detected no response (data not shown). Because TLR4 acts in concert with the LPS serum binding protein and CD14 (30), which is not expressed by astrocytes either in vitro or in vivo (31, 32), we also tested the effect of LPS (1 μg/ml) with the addition of recombinant soluble CD14 (0.1 μg/ml) and 10% serum. Again, we failed to find activation of the same set of genes, except for a low level response of IL-8 (8-fold compared with 800-fold by IL-1) at 6 h. A low level response to LPS in these cultures is usually due to minor contamination with microglia, which are sensitive to low doses of LPS due to the expression in vitro of TLR4 and CD14.
IL-1 leads to phosphorylation and translocation of IRF3 to the nucleus
In cells in which IRF3 has been activated by viral infection or ligands for TLR3, it has been shown that IRF-3 is phosphorylated as part of a macromolecular complex that contains TBK-1 and/or IB kinase-i (IKKi; also known as IKK) (33, 34, 35). To determine whether IRF-3 associates with either of these molecules in astrocytes, we attempted to coimmunoprecipitate them with an Ab to IRF-3 and probe Western blots of the precipitates with Abs to TBK-1 and IKKi. A clear signal was detected for TBK-1 on these blots, whereas no evidence for association with IKKi was observed (Fig. 2A and data not shown). Western blots also showed evidence of transient phosphorylation of IRF3, consistent with the formation of a functional signaling complex in IL-1-activated astrocytes (Fig. 2B). Because phosphorylation of IRF3 should result in the translocation of IRF3 to the nucleus, we then tested cytoplasmic and nuclear fractions of activated astrocytes by Western blotting (Fig. 3A). The data showed that IL-1 treatment of astrocytes led to increased localization of IRF3 in the nucleus that persisted through 24 h. The purity of these fractions for cross-contamination was determined using an Ab to nonmuscle myosin IIB for cytoplasmic and proliferating cell nuclear Ag for nuclear fractions (data not shown). In contrast, activation with LPS was without effect (Fig. 3B). Immunohistochemistry also confirmed translocation of IRF3 to the nucleus in activated cells (Fig. 3C).
FIGURE 2. Phosphorylation of IRF3 and its association with the kinase TBK-1 occurs in IL-1-treated human primary astrocytes. A, Human primary astrocyte cultures were treated with 10 ng/ml IL-1, and total cell lysates were harvested at the times indicated. Total protein was quantified and immunoprecipitated with Ab to IRF3 overnight at 4°C. Ab was captured on protein A agarose beads, and immunoblots were probed with Ab to TBK-1. The presence of TBK-1 in the starting lysates is shown in the lanes on the right. The bands at 50 kDa represent IgGh present in the immunoprecipitates and are shown as a loading control. B, Human astrocytes were treated with IL-1 at the times indicated, and total protein was isolated. Samples were resolved on 12% SDS-PAGE gels and immunoblotted for IRF3. Arrows indicate modified protein states. The gels were stripped and stained for -tubulin as a loading control (lower gel).
FIGURE 3. Nuclear localization of IRF3 upon IL-1 treatment in primary cultures of human astrocytes. A, Astrocytes were treated with IL-1 (10 ng/ml) for 1, 3, 6, and 24 h. Cells were lysed, and nuclear and cytoplasmic proteins were isolated. Both fractions were resolved on 12% SDS-PAGE gels and immunoblotted with anti IRF3 (1/1000). B, Astrocytes were treated with varying concentrations of LPS (0.01, 0.2, 1, and 10 μg/ml), and both nuclear and cytoplasmic proteins were harvested at 4 h, resolved using SDS-PAGE, and probed for IRF3. C, Astrocytes grown on MatTek dishes were stimulated with IL-1 for varying times. Cells were fixed, fluorescently stained for IRF3, and observed for nuclear translocation.
IFN- is an early response gene to IL-1 activation in primary human astrocytes
Activation of IRF3 is considered an essential component of the enhanceosome that leads to the production of IFN-. Signaling by IFN- through the type 1 IFN receptor then leads to an autocrine/paracrine amplification loop, resulting in transcriptional activation of additional IFN response genes that include IRF7 and, in some cells, CXCL10/IP-10 and CCL5/RANTES. The Q-PCR data showing that peak mRNA levels for IRF7, CXCL10/IP-10, and CCL5/RANTES lagged behind those for IFN- would be consistent with the possibility that these genes are downstream targets of IFN- gene expression. To test for this, cells were treated with cycloheximide (CHX) for 6 h, and mRNA expression was determined by Q-PCR. The data showed that mRNA levels for the tested genes in control or CHX-treated cells were: IFN-, 32.5 ± 10.2 vs 52.8 ± 14.5; IRF-7, 8.3 ± 2.5 vs 3.4 ± 0.7; CXCL10/IP-10, 643 ± 193 vs 48.6 ± 12.5; and CCL5/RANTES, 98.4 ± 30.7 vs 306.4 ± 140.7 (mean ± SEM; n = 4). The observation that CCL5/RANTES was enhanced by CHX is consistent with previously published data (36). Taken together, these results indicate that IFN- production occurs independently of translation after activation with IL-1 in human fetal astrocytes, whereas IL-1 induction of IRF7 and CXCL10/IP-10 is dependent upon protein synthesis.
IL-1 leads to delayed phosphorylation of STAT1
To confirm that IL-1 leads to type 1 IFN signaling in astrocytes, we assessed the phosphorylation status of STAT1 by Western blots using an Ab to phospho-STAT (Y701; Fig. 4). The results showed that in astrocytes activated with IL-1, STAT1 phosphorylation occurred in a time frame consistent with an autocrine response to IFN-. In contrast, phosphorylation of STAT1 was rapid after direct stimulation with IFN-. The Western blot also showed induction of STAT1 protein at 24 h by IL-1, consistent with the microarray data that showed increased mRNA for STAT1 in IL-1-treated astrocytes (see Fig. 1).
FIGURE 4. IL-1 induces STAT1 phosphorylation. Astrocytes were treated directly with either IL-1 (10 ng/ml) or IFN- (10 ng/ml) for the times indicated. Total cell protein was isolated and resolved by SDS-PAGE. The gels were probed with both anti-phospho-STAT-1 and total STAT-1 Ab.
Dominant-negative construct of IRF3 inhibits IL-1-induced IFN-, IRF7, CXCL10/IP-10, and CCL5/RANTES expression
We next observed what the effect of inhibiting IRF3 activity would have on the expression of IRF3-dependent ISG. To do this, we constructed an adenoviral vector harboring an IRF3 construct in which the N-terminal DNA-binding domain had been deleted. This construct has been shown previously to function as a dominant negative form of IRF3 (N) (37). Q-PCR of astrocytes pretreated with N-IRF3, then stimulated with IL-1, revealed a significant reduction in the expression of mRNA for IFN-, IRF7, CXCL10/IP-10, and CCL5/RANTES (Fig. 5A). In contrast, N-IRF3 had no effect on mRNA for CXCL8/IL-8.
FIGURE 5. Dominant negative IRF3 diminishes IRF3-dependent gene expression in IL-1-stimulated astrocytes, and IL-1 potentiates gene expression in WT-IRF3-expressing cells. A, Total RNA was isolated from astrocytes that had been infected with either a dominant negative adenoviral construct (multiplicity of infection, 10) or a WT adenoviral construct (multiplicity of infection, 10) 48 h before stimulation with IL-1 (10 ng/ml) for the indicated times. Q-PCR was performed on samples using SYBR Green. Data were normalized to the internal control standards 2M and PBDA and are presented as fold changes above control values. For the experiments using the dominant negative construct, astrocytes from at least three different cases were studied, and for each dataset the results are shown as the mean ± SD (n = 9). For the experiments using the WT construct, astrocytes from at least two different cases were studied, and for each dataset the results are shown as the mean ± SD (n = 3). *, p < 0.03 vs controls. B, Astrocyte cultures were pretreated with adenovirus for 2 days at an multiplicity of infection of 10. After pretreatment, fresh medium was added along with IL-1 (10 ng/ml) for 24 h. Supernatants were isolated and assayed for CCL5/RANTES, CXCL10/IP-10, and CXCL8/IL-8 by ELISA. C, Total cell lysates from adenovirus-infected cells were harvested and resolved on a 12% polyacrylamide gel by SDS-PAGE. The lower bands shown in the lane labeled N represent the IRF-3 construct lacking the N-terminal domain. Data shown are representative of two independent experiments.
In addition to an empty vector control, we generated a vector containing WT-IRF3, which was used in some experiments. Cells transfected with WT-IRF3 showed a superinduction after activation with IL-1 for the chemokines CXCL10/IP-10 and CCL5/RANTES, whereas induction of CXCL8/IL-8 was unaffected (Fig. 5A). ELISAs on supernatants harvested at 24 h from cells infected with both N-IRF3 and WT-IRF3 confirmed the reduction and enhancement of IRF3-dependent chemokine production in IL-1-stimulated astrocytes (Fig. 5B). This potentiation by IL-1 of chemokine expression in cells overexpressing IRF-3 (Fig. 5C) underscores our hypothesis that IL-1 signaling is capable of activating IRF3, leading to the induction of an antiviral response in primary human astrocytes in culture.
Discussion
In this report we show that treatment of human fetal astrocytes with the cytokine IL-1 leads to activation of the transcription factor IRF3 and induction of a set of immediate-early type 1 IFN-dependent genes that play a critical role in the host defense against pathogens. Key indicators of this activation of the IFN system by IL-1 in astrocytes include evidence of phosphorylation and translocation of IRF3 to the nucleus, up-regulation of mRNA for IFN-, evidence for downstream signaling via STAT1, induction of the secondary response gene IRF7 (38), and induction of the chemokines CXCL10/IP-10 and CCR5/RANTES. The robust induction by IL-1 of additional ISG, as noted on the microarrays such as STAT1, 2',5'-oligoadenylate synthetase 2, and ISG15, also support an active signaling role for IL-1 via this pathway in human fetal astrocytes.
In addition to viral infection, it is now well accepted that ligands for TLR3 and TLR4 activate IRF3, leading to the production of IFN- and the chemokines CXCL10/IP-10 and CCL5/RANTES (15, 16, 29). TLR play a critical role in the innate immune response by recognizing structurally conserved bacterial and viral elements termed pathogen-associated molecular patterns (39). Different TLRs recognize different ligands, but share common signal transduction pathways through expression of Toll-IL-1R homologous region domain-containing molecules that link to adapter proteins such as MyD88. For TLR4, additional studies in mice in which the gene for MyD88 had been inactivated demonstrated the presence of an alternative signaling pathway that directed downstream signaling in the absence of MyD88 (40, 41) and led to the activation of IRF3 (15, 16). Consistent with this alternative activation pathway for TLR4, the ligand LPS was found to activate IRF3 in the human astrogliomal cell line U373 (29). It could be argued, therefore, that in our astrocyte cultures it is LPS, rather than IL-1, that is providing a signal for activation of IRF3. However, as we have reported previously, we were unable to detect a response to LPS in our cultures of primary human astrocytes, including a failure of LPS to activate IL-8, which is exquisitely responsive to LPS in other systems (data not shown), suggesting that LPS does not signal via either the MyD88-dependent or -independent pathway in these cells. This is in stark contrast to human microglia isolated from these same tissues, which respond rapidly and robustly to LPS activation (25).
Upon activation by viral infection, IRF3 is phosphorylated by a virus-activated kinase at multiple serine and threonine sites, whereas activation by stress inducers leads to a more restricted phosphorylation pattern in the N-terminal region via a MAPK-related pathway (24). The transient shift in the migration pattern of IRF3 that we noted on our Western blots resembled that found after activation of IRF3 by stress inducers. The kinases that have been found to be involved in phosphorylation of IRF3 are the IKK-related kinases TBK-1 and IKKi (33, 34, 35). By coimmunoprecipitation studies we found that IRF3 associated with TBK1, but not IKKi, in IL-1-stimulated astrocytes. These data are, therefore, in agreement with results obtained in mouse macrophages and mouse fibroblasts, where it was found that TBK1 is the critical kinase required for IRF3 phosphorylation after activation via the nonviral pathway (42, 43). The presence of only one of these activating kinases associated with IRF3 in activated astrocytes could also account for the minimal and transient phosphorylation we observed 30 min after IL-1 treatment.
In LPS-stimulated murine B cells, IFN-, CXCL10/IP-10, and CCL5/RANTES have all been reported to be primary response genes after activation of IRF3 (16). However, in our system only IFN- gene expression was unaffected by CHX treatment, suggesting that CXCL10/IP-10 as well as possibly IRF7 are secondary response genes. This result would be consistent with the data derived from IFN--deficient mice, in which it has been shown that the production of CXCL10/IP-10 is mainly a consequence of IFN- production (44, 45). Similarly, after exposure to DNA-damaging agents, direct transcriptional activation of the CCL5/RANTES promoter was noted in one study (22), but not in another (46). Whether these discrepancies reflect differences in the activation of other transcription factors required for successful activation of CCL5/RANTES and CXCL10/IP-10 or represent cell-type differences in the specific kinases involved awaits further study.
In our experiments we observed a time lag between the onset of phosphorylation of IRF3 and the initial expression of IFN-. These data suggest that the initial induction of IFN- occurs as a result of IRF3 activation, but is later potentiated by the recruitment of other IRF family members as well as NF-B to the IFN- promoter (47). The IRF family consists of nine members, of which IRF1, -3, -7, and -9 have been implicated in the transcriptional regulation of IFN- genes (20, 48, 49). Of these IRFs, it has been suggested that IRF7 forms a key component of the late induction phase of IFN- (38, 50). In most cell types, IRF7 is not expressed constitutively, but is induced by IFN- via the JAK-STAT signaling pathway. This newly synthesized IRF7 is then thought to participate in the production of IFN- as well as IFN- (19).
The transcriptional induction of type 1 IFNs is central to the cellular antiviral response and plays an important role in the innate host defense against infection. However, type 1 IFNs also trigger a strong inflammatory response, inhibit cell proliferation, and induce apoptosis, suggesting a need for negative regulatory processes (51). Consistent with this idea, it has been shown that after activation, IRF3 is rapidly degraded via the ubiquitin-proteasome pathway (52), and IRF7 is unstable in most cell types (51). Our data showing that IFN- mRNA was only transiently elevated in astrocytes after activation by IL-1 probably also reflects the presence of a regulatory pathway. Specific cell types may also regulate downstream signaling effects by blocking phosphorylation of STAT1, as has been noted in alveolar vs peritoneal macrophages (53). In these cells, the authors argue that, given the vulnerability of lung tissue to excessive inflammation, activation of the innate immune response needs to be limited. Their data show that although the TLR-induced production of chemokines is intact in alveolar macrophages, the production of NO was not observed unless exogenous IFN was added. A similar phenomenon may exist in the CNS, where inflammation is also tightly regulated. In the inflamed CNS, astrocytes express CXCL10/IP-10, particularly in association with the glial endfeet that abut the vasculature (54). Astrocytes both in vitro and in vivo also express the inducible form of NO synthase (iNOS), but in vitro, IL-1, although necessary to initiate gene expression, leads to only low level production of NO unless IFN is present (55). We have shown previously that both IFN- and IFN- can enhance IL-1-induced iNOS expression when given alone; however, when IFN- and IFN- are added together, IFN- acts as a negative regulator of iNOS expression through a mechanism that involves inhibition of IFN--activating sequence-binding activity (56). In murine macrophages, IFN- has also been found to antagonize the activity of IFN- through inhibition of NF-B signaling (57) and down-regulating IFN--induced FcR expression (58) and IFN--induced MHC class II expression (59). Thus, the specific effects of IFN- at sites of inflammation are probably diverse and critically regulated by the presence or the absence of other cytokines in injured tissues. These data underscore the importance of understanding the complex pathways involved in inflammatory gene expression in astrocytes, which, it is to be hoped, will be of value in defining selective modes of intervention for inflammatory events in the CNS.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank Dr. John Hiscott (McGill University, Montreal, Canada) for the N and Wt IRF3 constructs, Dr. Brad Poulos for tissue acquisition, Wa Shen for tissue preparation, and Dr. Anne Bresnick (Albert Einstein College of Medicine) for affinity-purified rabbit polyclonal Ab to non-muscle myosin-IIB.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by U.S. Public Health Service Grants NS40137, NS11920, NS07098, MH55477, and NS046620, National Multiple Sclerosis Society Grant RG3444, and The Jayne and Harvey Beker Foundation (to G.R.J.).
2 Address correspondence and reprint requests to Dr. Mark A. Rivieccio, Department of Pathology, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461. E-mail address: mriviecc{at}aecom.yu.edu
3 Abbreviations used in this paper: IKK, IB kinase; aRNA, amplified antisense RNA; CXCL10/IP-10, IFN-induced protein (10 kDa); CHX, cycloheximide; IKKi, induced B kinase; iNOS, inducible form of NO synthase; IP-10, IFN--inducible protein-10; IRF, IFN regulatory factor; ISG, IFN-stimulated gene; ISRE, IFN-stimulated response element; 2M, 2-microglobulin; PBDA, porphobilinogen deaminase; Q-PCR, real-time quantitative PCR; TBK, Tank-binding kinase-1; WT, wild type.
Received for publication September 23, 2004. Accepted for publication January 7, 2005.
References
Janzer, R. C., M. C. Raff. 1987. Astrocytes induce blood-brain barrier properties in endothelial cells. Nature 325:253.
Anderson, C. M., R. A. Swanson. 2000. Astrocyte glutamate transport: review of properties, regulation, and physiological functions. Glia 32:1.
Liberto, C. M., P. J. Albrecht, L. M. Herx, V. W. Yong, S. W. Levison. 2004. Pro-regenerative properties of cytokine-activated astrocytes. J. Neurochem. 89:1092
Sofroniew, M. V., T. G. Bush, N. Blumauer, L. Kruger, L. Mucke, M. H. Johnson. 1999. Genetically-targeted and conditionally-regulated ablation of astroglial cells in the central, enteric and peripheral nervous systems in adult transgenic mice. Brain Res. 835:91.
Herx, L. M., V. W. Yong. 2001. Interleukin-1 is required for the early evolution of reactive astrogliosis following CNS lesion. J. Neuropathol. Exp. Neurol. 60:961.
John, G. R., L. Chen, M. A. Rivieccio, C. V. Melendez-Vasquez, A. Hartley, C. F. Brosnan. 2004. Interleukin-1 induces a reactive astroglial phenotype via deactivation of the Rho GTPase-Rock axis. J. Neurosci. 24:2837.
John, G. R., S. C. Lee, X. Song, M. Rivieccio, C. F. Brosnan. 2005. IL-1-regulated responses in astrocytes: relevance to injury and recovery. Glia. 49:161.
O’Neill, L. A., C. Greene. 1998. Signal transduction pathways activated by the IL-1 receptor family: ancient signaling machinery in mammals, insects, and plants. J. Leukocyte Biol. 63:650.
Zhang, P., B. S. Miller, S. A. Rosenzweig, N. R. Bhat. 1996. Activation of c-Jun N-terminal kinase/stress-activated protein kinase in primary glial cultures. J. Neurosci. Res. 46:114.
Ninomiya-Tsuji, J., K. Kishimoto, A. Hiyama, J. Inoue, Z. Cao, K. Matsumoto. 1999. The kinase TAK1 can activate the NIK-IB as well as the MAP kinase cascade in the IL-1 signalling pathway. Nature 398:252.
Radons, J., S. Dove, D. Neumann, R. Altmann, A. Botzki, M. U. Martin, W. Falk. 2003. The interleukin 1 (IL-1) receptor accessory protein Toll/IL-1 receptor domain: analysis of putative interaction sites in vitro mutagenesis and molecular modeling. J. Biol. Chem. 278:49145.
Dunne, A., M. Ejdeback, P. L. Ludidi, L. A. O’Neill, N. J. Gay. 2003. Structural complementarity of Toll/interleukin-1 receptor domains in Toll-like receptors and the adaptors Mal and MyD88. J. Biol. Chem. 278:41443.
Subramaniam, S., C. Stansberg, C. Cunningham. 2004. The interleukin 1 receptor family. Dev. Comp. Immunol. 28:415.
Takeda, K., T. Kaisho, S. Akira. 2003. Toll-like receptors. Annu. Rev. Immunol. 21:335.
Kawai, T., O. Takeuchi, T. Fujita, J. Inoue, P. F. Muhlradt, S. Sato, K. Hoshino, S. Akira. 2001. Lipopolysaccharide stimulates the MyD88-independent pathway and results in activation of IFN-regulatory factor 3 and the expression of a subset of lipopolysaccharide-inducible genes. J. Immunol. 167:5887.
Doyle, S., S. Vaidya, R. O’Connell, H. Dadgostar, P. Dempsey, T. Wu, G. Rao, R. Sun, M. Haberland, R. Modlin, et al 2002. Toll-like receptor 3 mediates a more potent antiviral response than Toll-like receptor 4. Immunity 17:251.
Smith, E. J., I. Marie, A. Prakash, A. Garcia-Sastre, D. E. Levy. 2001. IRF3 and IRF7 phosphorylation in virus-infected cells does not require double-stranded RNA-dependent protein kinase R or IB kinase but is blocked by vaccinia virus E3L protein. J. Biol. Chem. 276:8951.
Jiang, Z., T. W. Mak, G. Sen, X. Li. 2004. Toll-like receptor 3-mediated activation of NF-B and IRF3 diverges at Toll-IL-1 receptor domain-containing adapter inducing IFN-. Proc. Natl. Acad. Sci. USA 101:3533
Kawai, T., S. Sato, K. J. Ishii, C. Coban, H. Hemmi, M. Yamamoto, K. Terai, M. Matsuda, J. Inoue, S. Uematsu, et al 2004. Interferon- induction through Toll-like receptors involves a direct interaction of IRF7 with MyD88 and TRAF6. Nat. Immunol. 5:1061.
Mamane, Y., C., C. Heylbroeck, P. Genin, M. Algarte, M. J. Servant, C. DeLuca, H. Kwon, R. Lin, J. Hiscott. 1999. Interferon regulatory factors: the next generation. Gene 237:1.
Hiscott, J., P. Pitha, P. Genin, H. Nguyen, C. Heylbroeck, Y. Mamane, M. Algarte, R. Lin. 1999. Triggering the interferon response: the role of IRF-3 transcription factor. J. Interferon Cytokine Res. 19:1.
Kim, T., T. Y. Kim, Y. H. Song, I. M. Min, J. Yim, T. K. Kim. 1999. Activation of interferon regulatory factor 3 in response to DNA-damaging agents. J. Biol. Chem. 274:30686
Lopez-Collazo, E., S. Hortelano, A. Rojas, L. Bosca. 1998. Triggering of peritoneal macrophages with IFN-/ attenuates the expression of inducible nitric oxide synthase through a decrease in NF-B activation. J. Immunol. 160:2889.
Van Weyenbergh, J., P. Lipinski, A. Abadie, D. Chabas, U. Blank, R. Liblau, J. Wietzerbin. 1998. Antagonistic action of IFN- and IFN- on high affinity Fc receptor expression in healthy controls and multiple sclerosis patients. J. Immunol. 161:1568.
Lu, H. T., J. L. Riley, G. T. Babcock, M. Huston, G. R. Stark, J. M. Boss, R. M. Ransohoff. 1995. Interferon (IFN) acts downstream of IFN--induced class II transactivator messenger RNA accumulation to block major histocompatibility complex class II gene expression and requires the 48-kDa DNA-binding protein, ISGF3-. J. Exp. Med. 182:1517.(Mark A. Rivieccio,, Garet)