astrocytes recognize intracellular polyinosinic-polycytidylic acid via mda-5

The FASEB Journal • Research Communication

Astrocytes recognize intracellularpolyinosinic-polycytidylic acid via MDA-5

Joari De Miranda, Kavitha Yaddanapudi, Mady Hornig, and W. Ian Lipkin1

Center for Infection and Immunity, Mailman School of Public Health, Columbia University,New York, New York, USA

ABSTRACT RNA virus replication results in expres-sion of double-stranded RNA (ds-RNA) molecules thattrigger innate immune responses through interactionswith both intracellular and extracellular receptors. Weinvestigated the contributions of the extracellular andintracellular pathways to innate immunity in murineastrocyte primary cultures using polyinosinic-polycyti-dylic acid (poly I:C), a synthetic ds-RNA moleculedesigned to mimic RNA virus infection. Whereas extra-cellular poly I:C (naked poly I:C) mainly induced theexpression of regulated on activation normal T-cellexpressed and secreted (RANTES), interleukin-8 (IL-8), and tumor necrosis factor � (TNF-�), intracellulardelivery of poly I:C (complexed poly I:C) chiefly in-duced expression of IFN-� and IL-6. Experiments withastrocytes from Toll-like receptor 3 (TLR-3) knockoutmice indicated that naked poly I:C signals via a TLR-3-dependent NF-�B pathway. Complexed poly I:C in-duced the expression of the intracellular ds-RNA sen-sor proteins, retinoic acid inducible gene I (RIG-I), andmelanoma differentiation-associated gene 5 (MDA-5).However, transfection of astrocytes with dominant neg-ative forms of the helicases implicated MDA-5, but notRIG-I, as the intracellular sensor of poly I:C. Com-plexed poly I:C-mediated MDA-5 stimulation transmit-ted “downstream” signals, resulting in activation of thetranscription factors NF-�B and IRF-3. Our resultsillustrate the intricacy of extracellular and intracellulards-RNA recognition in viral infections of the centralnervous system and indicate the importance of MDA-5helicase as an intracellular ds-RNA sensor in astrocytes.De Miranda, J., Yaddanapudi, K., Hornig, M., Lipkin,W. I. Astrocytes recognize intracellular polyinosinic-polycytidylic acid via MDA-5. FASEB J. 23, 000–000(2009)

Key Words: ds-RNA � TLR-3 � RIG-I � CNS � innate immu-nity

Immune responses in the central nervous system(CNS) are largely mediated by microglia and astrocytes.Astrocytes integrate information from the microvascu-lature and neuronal interfaces modulating neuronalexcitability, synaptic transmission, and cerebral bloodflow. In CNS viral infection, astrocytes induce excito-toxic neuronal apoptosis through a tumor necrosisfactor � (TNF-�)-mediated mechanism (1).

The innate immune system recognizes microorgan-isms via pattern-recognition receptors (PRRs) locatedin the plasma membrane and cytosol of the cells. PRRsbind to key microbial components known as pathogen-associated molecular patterns (PAMPs), including thedouble-stranded RNA (ds-RNA) associated with virusreplication, thereby engaging host cells in pathogen-specific cell response programs (2). Polyinosinic-poly-cytidylic acid (poly I:C), a ds-RNA mimic, triggers theinnate immune system to secrete the antiviral cytokinesIFN-� and IFN-� and proinflammatory cytokines. In theCNS, poly I:C is recognized by resident microglia andastrocytes (3–6).

The innate immune system has two pathways for therecognition of ds-RNA associated with viral infection. Inone pathway, poly I:C present in the extracellular spaceis internalized through endocytosis, resulting in activa-tion of Toll-like receptor 3 (TLR-3) signaling (7).TLR-3 recognizes poly I:C through its extracellulardomain and transduces the signal via the cytoplasmicToll/Interleukin-1 receptor (TIR) domain. The down-stream signaling occurs by recruiting the adaptor pro-tein, TIR domain-containing adaptor inducing IFN-�(TRIF). The signaling cascade culminates in the activa-tion and nuclear migration of NF-�B, resulting ininduction of several genes involved in innate andadaptive immunity (8, 9). A second, intracellular path-way detects ds-RNA via the cytosolic sensor proteinsmelanoma differentiation-associated gene 5 (MDA-5)and retinoic acid inducible gene I (RIG-I). Thesecytosolic proteins are members of the DexD/H RNAbox helicase family with a caspase recruitment domain(CARD) at the N terminus and a helicase domain at theC terminus (10, 11). On binding poly I:C, MDA-5/RIG-I associate with the adaptor protein, IFN-� pro-moter stimulator 1 (IPS-1; also known as MAVS) (12).IPS-1 then recruits both IKKε/TBK1 and IKK�/�/�complexes, resulting in activation of the transcriptionfactors interferon regulatory factor 3 (IRF-3), NF-�B,and activating protein 1 (AP-1) (13).

Here, we investigate the contributions of the path-

1Correspondence: Center for Infection and Immunity,Mailman School of Public Health, Columbia University, 722West 168th St., Rm. 1801, New York, NY, 10032 E-mail:[email protected]

doi: 10.1096/fj.08-121434

10892-6638/09/0023-0001 © FASEB

The FASEB Journal article fj.08-121434. Published online November 26, 2008.

ways activated on treatment of murine primary astro-cytes with either naked or complexed poly I:C. Ourdata indicate that naked poly I:C triggers TLR-3-medi-ated signaling responses, resulting in the release ofIFN-�, interleukin (IL) -6, TNF-�, IL-8, and RANTES(regulated on activation normal T-cell expressed andsecreted). Complexed poly I:C treatment induces theexpression of RIG-I and MDA-5 in murine astrocytes.Furthermore, our data show that recognition of cytoso-lic poly I:C is mediated through MDA-5, resulting inIRF-3 activation and a more robust release of IFN-� andIL-6.

MATERIALS AND METHODS

Primary astrocyte culture

Postnatal day 1 to 3 C57BL/6J (B6), B6129SF2/J, and B6;129S1-Tlrtm1Flv/J TLR3 knockout mice (Jackson Laborato-ries; Bar Harbor, ME, USA) were anesthetized and decapi-tated. Brains were removed directly into ice-cold Hank’sbuffered saline solution (HBSS) (14). The cortex was dis-sected, freed of the meninges, trimmed, and treated withtrypsin (0.25%) and DNase (2000U) at 37°C for 10–20 min.Cells were mechanically dissociated, plated in 75 cm2 flasksand incubated in Dulbecco’s modified Eagle medium plusF12 salts (DMEM-F12) supplemented with 10% fetal calfserum and 5 mM glutamine at 37°C for 15–20 days until themixed glia culture reached confluence. The flasks wereagitated in a rotary shaker at 180 rpm. Tissue culture mediawere removed after 24 h, and the cell monolayer was washedwith HBSS solution to remove the contaminant glial cells.The adherent astrocytes were trypsinized and plated in 24-well poly-d-lysine-coated plates at a density of 2 � 105

cells/well. Cells were plated for 24–48 h before poly I:Ctreatment. The cell culture consisted of 95% or more ofastrocytes, as confirmed by glial fibrillary acid protein (GFAP)immunofluorescence (anti-GFAP, 1:50; BD Biosciences, SanJose, CA, USA) and flow cytometric analysis (data not shown).

Poly I:C treatment

Polyinosinic-polycytidylic potassium salt (Sigma-Aldrich, St.Louis, MO, USA) dissolved in PBS was heated to 55°C for 5min and allowed to cool at room temperature. ds-RNAconcentration was measured by UV spectroscopy. For nakedpoly I:C treatment, the drug was dissolved in culture mediumat a concentration of 25 �g/ml. Complexed poly I:C treat-ment was performed by transfecting 4 �g/ml of ds-RNAcoupled to lipofectamine 2000 (2.5 �l/well; Invitrogen, Carls-bad, CA, USA).

RNA extraction and cDNA synthesis

The cell monolayer was washed twice with ice-cold PBS, andRNA was extracted using the RNAeasy kit (Qiagen, Valencia,CA, USA). Following extraction, total RNA was quantified byUV spectroscopy. cDNA was synthesized using Taqman re-verse transcription reagents (Applied Biosystems, Foster City,CA, USA) from 1 �g of RNA in a total 100 �l reaction volume.

Quantitative real-time polymerase chain reaction (PCR)

Intron/exon spanning gene-specific PCR primers and probeslabeled at the 5�-end with the fluorescent reporter dye FAM

(6-carboxyfluorescein) and at the 3�-end with the quencherdye TAMRA (6-carboxy-tetramethyl-rhodamine) for themouse homologue of the RIG-I and MDA-5 helicases werepurchased from Applied Biosystems. For the determinationof the target transcript copy number, PCR products amplifiedfrom the primers were subcloned into a pGEM-T Easy vector(Promega Corp., Madison, WI, USA). The plasmid was thenlinearized, and the concentration was determined by UVspectroscopy. Several 10-fold dilutions of the linearized plas-mid were used in the assay to generate standard curves.Samples were normalized by laminin gene expression in thesame sample replicates using intron/exon-spanning mouselaminin-specific PCR primers and probe containing the 5�-end reporter dye TET (tetrachloro-6-carboxy-fluorecein) andthe 3�-end quencher dye TAMRA (F-CCTCGCAGCGCAGCC,R-ACCATTCTGACGCCTGATCTG P-TGGAGGCAGCGTCAC-CAAAAAGC). For the quantitative real-time PCR assay, each25-�l amplification reaction contained 5 �l of the cDNAtemplate, 12.5 �l universal master mix (Applied Biosystems),1.25 �l of the gene of interest primers and probe mix(Applied Biosystems), 200 nM of the laminin probe, and 300nM of the laminin primers. The assay was performed on aModel 7700 Sequence Detector System Thermal Cycler (Ap-plied Biosystems) with a profile that consisted of 1 cycle at50°C for 2 min, 1 cycle at 95°C for 10 min, and 45 cycles at95°C for 15 s, followed by 60°C for 1 min.

Western blot analysis

Astrocytes were washed once with ice-cold PBS and resus-pended in the sample buffer (10 mM Tris-HCl, pH 7.5; 10mM EDTA, 20% v/v glycerol; 1% w/v SDS; 0.005% w/vbromphenol blue; 100 mM dithiothreitol) and size fraction-ated by 8–10% SDS-PAGE. Proteins were transferred to apolyvinylidene difluoride (PVDF) membrane according tostandard protocols. Blots were blocked in a 5% nonfat milk,0.1% Tween 20, and PBS solution and incubated with anti-IRF-3 (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA,USA), anti-phospho-IRF-3 (1:200; Ser-396; Cell Signaling,Danvers, MA, USA), anti-NF-�B2 p100 (1:200; Cell Signaling)or anti-I�-B� (1:200; eBioscience, San Diego, CA, USA) at 4°Covernight. The membranes were washed 3 times in a PBS-0.1% Tween-20 solution and incubated with peroxidase-conjugated goat anti-mouse immunoglobulin G (IgG) (1:10,000; Bio-Rad, Hercules, CA, USA) or goat anti-rabbit IgG(1:10,000; Bio-Rad) for 2 h at room temperature. The blotswere developed using ECL kit reagents (Amersham Bio-sciences; Piscataway, NJ, USA). The blots were stripped andreprobed with mouse anti-GAPDH (glyceraldehyde-3-phos-phate dehydrogenase) antibody (Ambion, Austin, TX, USA),as a housekeeping gene and for loading control.

Cytokine measurement

Cytokines were measured in tissue culture supernatants usingthe bead-based Luminex assay. The bead mix was customdesigned (Millipore, Billerica, MA, USA), and the assay wasperformed according to the manufacturer’s protocol. Theresults were measured on a Luminex 100 system, and thecytokine concentration values were calculated using the LMAT1.0 software (Luminex Corporation, Austin, TX, USA). IFN-�cytokine analysis was performed by ELISA, according to themanufacturer’s protocol (PBL Biomedical Laboratories, Piscat-away, NJ, USA).

Flow cytometric analysis

Astrocytes were untreated or treated with naked poly I:C (25�g/ml), lipofectamine alone, or complexed poly I:C (4

2 Vol. 23 April 2009 DE MIRANDA ET AL.The FASEB Journal

�g/ml). Cells were harvested after 6 h of culture and analyzedfor the expression of intracellular activated caspase-8 by flowcytometry. Cells were washed twice with RPMI 1640 mediumsupplemented with 0.5% FBS. Cells (106) were stained forcaspase expression using the vibrant FAM caspase-8 assay kit(Invitrogen). Cells were analyzed on a LSRII analyzer (BDBiosciences). Data were obtained using FACS DiVa acquisi-tion software (BD Biosciences), and analyzed using FlowJo6.1software (Tree Star, Ashland, OR, USA) after appropriategating to exclude dead cells and debris based on forwardscatter and side scatter.

Generation and transfection of RIG-I and MDA-5 dominantnegative plasmids

Empirical inactivating point mutations at the Walker-type ATP-binding site of RIG-I pUNO and MDA-5 pUNO (Invivogen, SanDiego, CA, USA) expressing plasmids were performed using thesite mutagenesis kit (Stratagene; La Jolla, CA, USA). A lysine 270to alanine substitution was introduced in the RIG-I sequence(K270A) using the primer: 5�-GTGCTCCTACAGGTTGTGGAG-CAACCTTTGTTTCACTGCTTA-3�. A lysine 335 to alanine sub-stitution was introduced in the MDA-5 sequence (K335A) usingthe primer: 5�-CTCCCTACAGGGAGTGGAGCAACCAGAG-TGGTCGTTTA-3�. Astrocytes were either singly or doubly trans-fected with 0.75 �g of each of dominant negative plasmids of

RIG-I (K270A) or MDA-5 (K335A) for 48 h. Transfection wasperformed using 2.5 �l of lipofectamine LTX and 3 �l of Plusreagent (Invitrogen). Transfection efficiency was quantitated inastrocytes by flow cytometry. Cells were transfected with plasmidexpressing the FLAG-tagged dominant negative mutant form ofMDA-5 (MDA-5 K335A). 106 cells were stained for MDA-5K335A expression using the FITC-conjugated anti-FLAG anti-body (Sigma- Aldrich). MDA-5 K335A protein expression wasobserved in 46.7% of the cells (Supplemental Fig. 1).

Immunofluorescence

Astrocytes were fixed with 4% paraformaldehyde at roomtemperature for 30 min, permeabilized with 0.1% TritonX-100 for 15 min, and blocked for 2 h in PBS with 10%normal goat serum (Sigma-Aldrich). Cells were incubatedovernight at 4°C with anti-IRF-3 (1:50; Santa Cruz Biotechnol-ogy) and anti-GFAP (1:50; BD Biosciences) antibodies. Cellswere then incubated for 2 h at room temperature with Cy-3conjugated goat anti-rabbit IgG and Cy-2 conjugated goatanti-mouse IgG (1:200; Jackson Immunoresearch Laborato-ries; West Grove, PA, USA). Cells were mounted with Pro-Long Gold antifade reagent with DAPI (Invitrogen). Imageswere captured with a Zeiss LSM 510 NLO MultiphotonConfocal Microscope and analyzed using Carl Zeiss Confocal

Figure 1. Poly I:C treatment of mu-rine astrocytes results in increasedrelease of IL-6, IFN-�, TNF-�, IL-8,and RANTES. Astrocytes isolatedfrom wild-type C57BL/6J mice were

untreated (control) or treated with either naked poly I:C (25 �g/ml) or complexed poly I:C (4 �g/ml) for 6 or 24 h.Cytokines were assayed in cell supernatants. Results indicate mean � se concentration (pg/ml). *P � 0.05, **P �0.001; one-way ANOVA with dose group as independent variable.

Figure 2. Cytokine induction by naked poly I:C but not complexed poly I:C is TLR-3 dependent. Astrocytes isolated fromwild-type B6129SF2/J or TLR3 knockout B6;129S1-Tlrtm1Flv/J mice were untreated (control) or treated with either naked polyI:C (25 �g/ml) or complexed poly I:C (4 �g/ml) for 6 h. Cytokines were assayed in cell supernatant. Results indicate mean �se concentration (pg/ml). *P � 0.05, **P � 0.001; Mann-Whitney.

3MDA-5 MEDIATES POLY I:C RECOGNITION IN ASTROCYTES

Microscope (AIM) Software (Carl Zeiss GmbH; Thornwood,NY, USA).

Statistical analysis

For values that presented with a normal distribution, one- ortwo-way ANOVA tests were performed (independent vari-ables, dose group, or dose group and genotype, respectively).Nonparametrical approaches were pursued for the analysis ofthe values in which a non-Gaussian distribution was observed.For the evaluation of differences involving independentvariables with �3 levels (e.g., controls vs. naked poly I:C vs.complexed poly I:C treatment) the Kruskal-Wallis test wasused. For differences involving independent variables with �2 levels (e.g., control vs. poly I:C), we used the Mann-WhitneyU test. All nonparametric tests were 2-tailed. When theKruskal-Wallis test was significant (nominal � � 0.05 as thethreshold of significance), individual Mann-Whitney two-group comparisons were pursued to identify the nature of thesignificant effect. All statistical analysis was performed usingStatview for Windows software (ver. 5.0.1; SAS Institute; Cary,NC, USA).

RESULTS

Cytokines and chemokines play important roles inviral infection and innate immune processes. Thus,we examined the influence of naked and complexedpoly I:C on cytokine production by cultured murineastrocytes. Naked and complexed poly I:C treatmentfor 6 and 24 h significantly increased the levels ofIFN-�, IL-6, TNF-�, IL-8, and RANTES (Fig. 1).Whereas complexed poly I:C treatment induced ele-vated levels of IFN-� and IL-6, naked poly I:C treat-

ment had more pronounced effects on TNF-�, IL-8,and RANTES (Fig. 1). Lipofectamine treatmentalone (negative control for complexed poly I:C) didnot alter cytokine levels at 24 h (data not shown).Taken together, these data implicate the involvementof different signaling cascades following exposure tonaked or complexed poly I:C.

TLR-3-deficient (TLR-3/) mice show reduced re-sponses to naked poly I:C (7). To elucidate the role ofTLR-3 in poly I:C recognition, we compared the cyto-kine profiles of astrocytes isolated from wild-typeB6129SF2/J and B6;129S1-Tlrtm1Flv/J TLR-3 knockoutmice following exposure to naked or complexed polyI:C. Whereas TLR-3/ astrocytes were deficient in

Figure 3. Poly I:C induces NF-�B activation in murineastrocytes. Astrocytes isolated from wild-type B6129SF2/J andTLR3 knockout B6;129S1-Tlrtm1Flv/J mice were untreated(control) or treated with either naked poly I:C (25 �g/ml),lipofectamine alone, or complexed poly I:C (4 �g/ml) for6 h. A) Protein extracts from astrocyte cultures were subjectedto immunoblot analysis using anti-NF-�B2 antibody. Inactive(p100) and active (p52) forms of the protein are indicated.B) Immunoblot analysis of astrocyte cell extracts using anti-I�B� antibody. Anti-GAPDH was used as loading control.

Figure 4. MDA-5 but not RIG-I is the intracellular sensor ofcomplexed poly I:C in astrocytes. Astrocytes isolated fromwild-type C57BL/6J mice were untreated (control, solid bar)or treated with complexed poly I:C (4 �g/ml, white bar) for6 h. A) Endogenous mRNA levels of RIG-I and MDA-5 weremeasured by quantitative real-time PCR. Results are ex-pressed as the mean numbers of mRNA copies of RIG-I andMDA-5 relative to the value obtained for laminin mRNA;**P � 0.001; Mann-Whitney U test. B) Wild-type C57BL/6Jastrocytes expressing dominant-negative forms of RIG-I(K270A), MDA-5 (K335A), RIG-I (K270A), and MDA-5(K335A) or GFP control protein were treated with complexedpoly I:C (4 �g/ml) for 6 h. Levels of IL-6 protein weremeasured in culture supernatants. Results indicate mean � seconcentration (pg/ml). *P � 0.05; one-way ANOVA with dosegroup as independent variable.


IFN-�, IL-6, IL-8, and RANTES responses after treat-ment with naked poly I:C, cytokine expression aftercomplexed poly I:C was the same as in wild-type astro-

cytes (Fig. 2). These results suggest that TLR-3 isinvolved in the recognition of naked poly I:C and maynot play a role in complexed poly I:C signaling.

Figure 5. Treatment of astrocytes with complexed poly I:C but not naked poly I:C induces IRF-3 activation. Astrocytes isolated fromwild-type C57BL/6J mice were untreated (control) or treated with either naked poly I:C (25 �g/ml), lipofectamine alone, orcomplexed poly I:C (4 �g/ml) for 6 h. A) Astrocyte cell extracts were subjected to immunoblot analysis using anti-phospho IRF-3antibody. B) Anti-IRF-3 immunoblot of astrocyte cell extracts separated by native PAGE. Protein monomer and dimer forms areindicated. C) Protein lysates from cytoplasmic (top) and nuclear cell fractions (bottom) were subjected to immunoblot analysis usinganti-IRF-3 antibody. Corresponding immunoblot signals for GAPDH are shown (loading control). D–G) Immunofluorescence analysisof IRF-3 protein in untreated astrocytes (D) or in astrocytes treated with naked poly I:C (E), lipofectamine alone (F), orcomplexed poly I:C (G) for 6 h. Insets: triple labeling of cells with anti-IRF-3 (red), anti-GFAP (green), and nuclear counterstain(blue). Scale bars 20 �m.

Figure 6. Complexed poly I:C treatment of astro-cytes induces caspase-8 activation. Astrocytes isolatedfrom wild-type C57BL/6J mice were untreated (con-trol) or treated with either naked poly I:C (25�g/ml), lipofectamine alone, or complexed poly I:C(4 �g/ml) for 6 h. After staining with caspase-8-FLICA and propidium iodide (PI; marker for apo-ptosis or necrosis), cells were analyzed by flow cytom-

etry. A) Dot plots of caspase-8 and PI fluorescence in untreated, lipofectamine-treated, or poly I:C-treated astrocytes.Numbers in quadrants represent percentages of each subpopulation. B) Percentages of caspase-8� cells in untreated,lipofectamine-treated, or poly I:C-treated astrocytes. Values are means � se. *P � 0.05; Mann-Whitney U test.


TLR-3 specifically signals for NF-�B activation inresponse to viral infections (7). In resting cells, NF-�Bis present as an inactive complex with its regulatorysubunit I�B. On ds-RNA stimulation, the complexbecomes a substrate for proteasome degradation, re-sulting in the degradation of I�B and liberation of thecleaved NF-�B. The active subunit of NF-�B then mi-grates to the nucleus where it acts as a transcriptionfactor (15). To elucidate the role of TLR-3 in nakedand complexed poly I:C-induced NF-�B activation, pro-tein extracts of astrocytes were isolated from wild-typeand TLR-3/ mice and analyzed by Western blotanalysis. In wild-type cells, naked poly I:C treatment didnot increase the protein levels of the active NF-�B p52subunit (Fig. 3A, lanes 1, 3). However, naked poly I:Cinduced significant degradation of the inhibitory pro-tein I�B (Fig. 3B, lanes 1, 3). Complexed poly I:Cinduced the activation of NF-�B and decreased theintensity of the protein signal corresponding to I�B(Fig. 3A, B; lanes 2, 4). These effects were not observedin TLR-3/ cells exposed to naked poly I:C (Fig. 3A,lane 7). In contrast, responses were similar in wild-typeand TLR-3/ cells after exposure to complexed polyI:C (Fig. 3A, B; lanes 4, 8).

The cytoplasmic proteins RIG-I and MDA-5 havebeen identified as intracellular sensors of ds-RNA (11).However, expression of RIG-I and MDA-5 has not beenpreviously reported in murine astrocytes. Quantitativereal-time PCR demonstrated that complexed poly I:Cinduced a 21-fold and 25-fold increase in RIG-I andMDA-5 mRNA levels, respectively (Fig. 4A). To furtheraddress the function of RIG-I and MDA-5 in complexedpoly I:C signaling, we examined their activities inastrocytes transfected with constructs that directed ex-pression of proteins, wherein inactivating point muta-tions were introduced at ATP-binding sites: RIG-I(K270A) and MDA-5 (K335A) (16). We also used cellsdoubly transfected with constructs to enable simulta-neous expression of MDA-5 mutant protein and RIG-Imutant protein. Whereas levels of IL-6 were signifi-cantly reduced in cells transfected with mutant MDA-5or with both mutant MDA-5 and mutant RIG-I, noreduction was seen in cells transfected with mutantRIG-I (Fig. 4B). These results are compatible withMDA-5 serving as an essential regulator of complexedpoly I:C signaling in astrocytes.

RIG-I and MDA-5 transduce signals that lead to type1 IFN production by triggering the phosphorylation,dimerization, and nuclear translocation of active IRF-3(12). To assess the role of MDA-5 in naked andcomplexed poly I:C-induced IRF-3 activation, proteinextracts of astrocytes isolated from wild-type mice wereanalyzed by Western blot analysis. Complexed poly I:Ctreatment induced phosphorylation of IRF-3, as de-tected by an antibody specific for phospho-IRF-3 (Ser-396) (Fig. 5A, lane 4). Constitutive formation of IRF-3dimers was augmented by complexed poly I:C (Fig. 5B,lane 4). Western blot analysis of cytoplasmic and nu-clear cell fractions demonstrated that complexed polyI:C induced the nuclear translocation of IRF-3 (Fig. 5C,

lane 4). IRF-3 phosphorylation, dimerization, and nu-clear translocation were not observed with control,naked poly I:C or lipofectamine alone (Fig. 5A–C, lanes1–3). Immunofluorescence analysis confirmed thatIRF-3 staining was restricted to the nuclear compart-ment of astrocytes treated with complexed poly I:C(Fig. 5G). IRF-3 was present in the cytoplasm, and, to alower extent, in the nucleus of control, naked, orlipofectamine treated astrocytes (Fig. 5D–F).

MDA-5/RIG-I signaling induces a caspase-8/10 me-diated NF-�B activation in a TLR-3-independentmanner (17). To test whether complexed poly I:C-induced NF-�B activation occurred through caspase-8/10 signaling, expression levels of activated

Figure 7. Activation of different poly I:C-mediated signalingpathways depends on the mode of delivery of poly I:C to thecells. A) When delivered to the extracellular milieu (nakedpoly I:C), ds-RNA is recognized by TLR-3 receptors presenton the endosomes and plasma membranes of astrocytes.Binding of poly I:C to TLR-3 results in NF-�B cleavage andactivation. Cleaved 52-kDa product of NF-�B translocates tothe nucleus and induces the expression of RANTES, IL-8, andTNF-�. B) Intracellular delivery of poly I:C (complexed polyI:C) engages the intracellular helicase MDA-5. Complexedpoly I:C-dependent MDA-5 signaling results in IRF-3 phos-phorylation. Phospho-IRF-3 forms homodimers and migratesto the cell nuclei. Activated IRF-3 binds to the ISRE elementsin the cell nuclei and stimulates the synthesis of IFN-� andIL-6. Binding of complexed poly I:C to MDA-5 also triggerscaspase-8 cleavage and activation. Cleaved caspase-8 inducesTLR-3-independent NF-�B activation observed after com-plexed poly I:C treatment of astrocytes.


caspase-8 were measured in astrocytes by flow cytom-etry. Complexed poly I:C treatment resulted in a1.5-fold increase in the percentage of cells positivefor activated caspase-8 (percentage of caspase-8�astrocytes exposed to complexed poly I:C,11.2�0.5%; lipofectamine-treated cells, 7.3�0.6%,Fig. 6). No significant induction of activatedcaspase-8 was observed in untreated cells or cellstreated with naked poly I:C (Fig. 6).

DISCUSSION

Astrocytes respond to viral infections with the induc-tion of type I interferons, cytokines, and chemokines(6). We examined the mechanisms responsible for theactivation of astrocytes by extracellular and intracellu-lar ds-RNA associated with viral infection using nakedor complexed poly I:C. Our findings corroborate re-cent data indicating that TLR-3 is important for re-sponses of astrocytes to naked poly I:C (3). TLR-3,expressed in astrocytes, has been identified as themolecule that participates in the innate immune re-sponse to West Nile virus by facilitating its entry into theCNS (18). Our experiments with TLR-3/ astrocytesdemonstrated that TLR-3 stimulation by naked poly I:Cresulted in NF-�B activation, which, in turn, inducedthe production of proinflammatory cytokines such asIL-8, RANTES (Fig. 7A).

TLR-3-mediated signaling mechanisms leading tothe production of IFN-� and IL-6 involve the activa-tion of the transcription factors IRF-3 and IRF-7 (19).Although IRF-3 and IRF-7 function as direct trans-ducers of TLR-3 signaling, we did not observe signif-icant IRF-3 activation after naked poly I:C treatment.We speculate, therefore, that TLR-3-dependentIFN-� and IL-6 production in naked poly I:C-stimulated astrocytes may involve an IRF-7 signalingpathway.

The cytoplasmic helicases RIG-I and MDA-5 recog-nize viral nucleic acids and are essential for theproduction of type I interferons in the context of aviral infection (20). In vivo studies have shown thatintracellular recognition of poly I:C is mediated byMDA-5 (10), but little is known about its role duringviral infections of the CNS. We found that complexedpoly I:C treatment of astrocytes induced the expres-sion of both RIG-I and MDA-5 proteins; however,only MDA-5 participated in intracellular poly I:Crecognition. Moreover, MDA-5-mediated signaling isa prerequisite for IRF-3 activation. IRF-3 can formhomodimers or heterodimers with IRF-7; thesedimers may be responsible for the induction of IL-6and IFN-� production observed with complexed polyI:C treatment. Alternatively, induction of IFN-� ob-served with complexed poly I:C may be mediatedthrough the interaction of transcription factorsIRF-3, IRF-7, and NF-�B in the nuclear compartment.Assembly of these transcription factors on the IFN-�gene promoter (enhanceosome complex) is known

to result in a more robust and prolonged inductionof IFN-� gene (13).

Gene-targeting studies in mice have shown that ds-RNA-dependent protein kinase (PKR), an intracellulards-RNA sensor, participates in poly I:C recognition inastrocytes (3). Activation of PKR by ds-RNA results inPKR autophosphorylation and the downstream activa-tion of eukaryotic initiation factor 2�-subunit (eIF-2�).In our model, poly I:C did not alter the phosphoryla-tion status of PKR and eIF-2� at 6 h post-treatment(data not shown). Although these results suggest thatPKR may not contribute to early cytokine expressionafter ds-RNA exposure, they do not exclude a role inlate and prolonged cytokine induction.

Our results with TLR-3/ astrocytes demonstratethat complexed poly I:C induced a TLR-3-independentNF-�B activation. Therefore, the observed NF-�B acti-vation could be mediated through alternate signalingmechanisms. Earlier reports demonstrated that recog-nition of ds-RNA by MDA-5 results in cleavage ofcaspase-8 and -10 and that the activated, cleavedcaspases interact with downstream effectors to activateNF-�B (17). Complexed poly I:C up-regulated activatedcaspase-8 in astrocytes, suggesting its involvement in com-plexed poly I:C-mediated NF-�B activation (Fig. 7B).

We conclude that differences in poly I:C treatmentlead to the recruitment of different, partially redun-dant intracellular pathways, engaging astrocytes in twodistinct cytokine profiles. Participation of MDA-5 in therecognition of poly I:C suggests a crucial role for thishelicase in the innate immune response to viral ds-RNAin the CNS and further highlights the importance ofastrocytes in innate immune pathogen recognition.

The authors thank Vishal Kapoor for technical assistanceand Robert Serge for help with statistical analysis. This workwas supported by a grant from the U.S. National Institutes ofHealth (Northeast Biodefense Center U54-AI057158-Lipkin).

REFERENCES

1. Volterra, A., and Meldolesi, J. (2005) Astrocytes, from brain glueto communication elements: the revolution continues. Nat. Rev.Neurosci. 6, 626–640

2. Akira, S., Uematsu, S., and Takeuchi, O. (2006) Pathogenrecognition and innate immunity. Cell 124, 783–801

3. Carpentier, P. A., Williams, B. R., and Miller, S. D. (2007)Distinct roles of protein kinase R and toll-like receptor 3 in theactivation of astrocytes by viral stimuli. Glia 55, 239–252

4. Town, T., Jeng, D., Alexopoulou, L., Tan, J., and Flavell, R. A.(2006) Microglia recognize double-stranded RNA via TLR3.J. Immunol. 176, 3804–3812

5. Scumpia, P. O., Kelly, K. M., Reeves, W. H., and Stevens, B. R.(2005) Double-stranded RNA signals antiviral and inflammatoryprograms and dysfunctional glutamate transport in TLR3-ex-pressing astrocytes. Glia 52, 153–162

6. Park, C., Lee, S., Cho, I. H., Lee, H. K., Kim, D., Choi, S. Y., Oh,S. B., Park, K., Kim, J. S., and Lee, S. J. (2006) TLR3-mediatedsignal induces proinflammatory cytokine and chemokine geneexpression in astrocytes: differential signaling mechanisms ofTLR3-induced IP-10 and IL-8 gene expression. Glia 53, 248–256

7. Alexopoulou, L., Holt, A. C., Medzhitov, R., and Flavell, R. A.(2001) Recognition of double-stranded RNA and activation ofNF-�B by Toll-like receptor 3. Nature 413, 732–738


8. Yamamoto, M., Sato, S., Hemmi, H., Hoshino, K., Kaisho, T.,Sanjo, H., Takeuchi, O., Sugiyama, M., Okabe, M., Takeda, K.,and Akira, S. (2003) Role of adaptor TRIF in the MyD88-independent toll-like receptor signaling pathway. Science 301,640–643

9. Yamamoto, M., Sato, S., Mori, K., Hoshino, K., Takeuchi, O.,Takeda, K., and Akira, S. (2002) Cutting edge: a novel Toll/IL-1receptor domain-containing adapter that preferentially activatesthe IFN-� promoter in the Toll-like receptor signaling. J. Immu-nol. 169, 6668–6672

10. Kato, H., Takeuchi, O., Sato, S., Yoneyama, M., Yamamoto, M.,Matsui, K., Uematsu, S., Jung, A., Kawai, T., Ishii, K. J., Yamagu-chi, O., Otsu, K., Tsujimura, T., Koh, C. S., Reis e Sousa, C.,Matsuura, Y., Fujita, T., and Akira, S. (2006) Differential roles ofMDA5 and RIG-I helicases in the recognition of RNA viruses.Nature 441, 101–105

11. Takeuchi, O., and Akira, S. (2008) MDA5/RIG-I and virusrecognition. Curr. Opin. Immunol. 20, 17–22

12. Kawai, T., Takahashi, K., Sato, S., Coban, C., Kumar, H., Kato,H., Ishii, K. J., Takeuchi, O., and Akira, S. (2005) IPS-1, anadaptor triggering RIG-I- and Mda5-mediated type I interferoninduction. Nat. Immunol. 6, 981–988

13. Honda, K., and Taniguchi, T. (2006) IRFs: master regulators ofsignalling by Toll-like receptors and cytosolic pattern-recogni-tion receptors. Nat. 6, 644–658

14. Giulian, D., Li, J., Bartel, S., Broker, J., Li, X., and Kirkpatrick,J. B. (1995) Cell surface morphology identifies microglia as adistinct class of mononuclear phagocyte. J. Neurosci. 15, 7712–7726

15. Niederberger, E., and Geisslinger, G. (2008) The IKK-NF-�Bpathway: a source for novel molecular drug targets in paintherapy? FASEB J. 22, 3432–3442

16. Yoneyama, M., Kikuchi, M., Natsukawa, T., Shinobu, N.,Imaizumi, T., Miyagishi, M., Taira, K., Akira, S., and Fujita, T.(2004) The RNA helicase RIG-I has an essential function indouble-stranded RNA-induced innate antiviral responses. Nat.Immunol. 5, 730–737

17. Takahashi, K., Kawai, T., Kumar, H., Sato, S., Yonehara, S., andAkira, S. (2006) Roles of caspase-8 and caspase-10 in innateimmune responses to double-stranded RNA. J. Immunol. 176,4520–4524

18. Wang, T., Town, T., Alexopoulou, L., Anderson, J. F., Fikrig, E.,and Flavell, R. A. (2004) Toll-like receptor 3 mediates West Nilevirus entry into the brain causing lethal encephalitis. Nat. Med.10, 1366–1373

19. Doyle, S., Vaidya, S., O’Connell, R., Dadgostar, H., Dempsey, P.,Wu, T., Rao, G., Sun, R., Haberland, M., Modlin, R., and Cheng,G. (2002) IRF3 mediates a TLR3/TLR4-specific antiviral geneprogram. Immunity 17, 251–263

20. Kato, H., Takeuchi, O., Mikamo-Satoh, E., Hirai, R., Kawai, T.,Matsushita, K., Hiiragi, A., Dermody, T. S., Fujita, T., and Akira,S. (2008) Length-dependent recognition of double-strandedribonucleic acids by retinoic acid-inducible gene-I and mela-noma differentiation-associated gene 5. J. Exp. Med. 205, 1601–1610

Received for publication September 5, 2008.Accepted for publication November 6, 2008.


astrocytes recognize intracellular polyinosinic-polycytidylic acid via mda-5

Documents