A

gnJii©

K

Jccthclic�flai

tmvpr

f

0d

Neuroscience Letters 414 (2007) 45–50

Induction of IP-10 (CXCL10) in astrocytesfollowing Japanese encephalitis

Sourojit Bhowmick a, Rachna Duseja a, Sulagna Das a, Mohan Babu Appaiahgiri b,Sudhanshu Vrati b, Anirban Basu a,∗a National Brain Research Centre, Manesar, Haryana, India

b National Institute of Immunology, New Delhi, India

Received 7 September 2006; received in revised form 26 October 2006; accepted 13 November 2006

bstract

Chemokines and their receptors are important elements for the selective attraction and activation of various subsets of leukocytes. Interferon-amma inducible protein (IP-10 or CXCL-10) is a potent chemoattractant and has been suggested to enhance the severity of virus infection and

euronal injury. In order to assess functional importance of this chemokine in viral encephalitis, we have exploited an experimental model ofapanese encephalitis. We report for the first time that in Japanese encephalitis, astrocytes are the predominant source of IP-10. A progressivencrease in IP-10 induction following viral infection is concomitant with the increase in IFN-� a known inducer of IP-10. However, this increasen IFN-� level is not sufficient to confer protection as animals eventually succumb to the infection.

2007 Published by Elsevier Ireland Ltd.

i[

oetvvscCt

ati

eywords: Japanese encephalitis; IP-10; Astrocytes; IFN-�

apanese encephalitis (JE) is an acute viral infection of theentral nervous system caused by a mosquito-borne flavivirusalled Japanese encephalitis virus (JEV). JEV targets the cen-ral nervous system (CNS), clinically manifesting with fever,eadache, vomiting, signs of meningial irritation and alteredonsciousness leading to high mortality and neurological seque-ae in some of those who survive [12]. The inflammation resultsn an increased level of cytokines, such as macrophage-derivedhemotactic factor (MDF), tumor necrosis factor alpha (TNF-) and interleukin-8 (IL-8) in the serum and the cerebrospinaluid (CSF) [10,19]. The increased levels of inflammatory medi-tors appear to play a protective role or to initiate an irreversiblemmune response leading to cell death.

IP-10 (CXCL10) is a CXC-chemokine that acts on its recep-or; CXCR3, to attract activated T cells, NK cells and blood

onocytes, and appears to be an important participant in a

ariety of inflammatory conditions [3,16,21]. It is a 10 kDaolypeptide that was first identified as the product of an earlyesponse gene to IFN-� and is produced by a variety of cells,

∗ Corresponding author. Tel.: +91 124 2338921x225;ax: +91 124 2338910/28.

E-mail address: anirban@nbrc.res.in (A. Basu).

JIw

tPiw

304-3940/$ – see front matter © 2007 Published by Elsevier Ireland Ltd.oi:10.1016/j.neulet.2006.11.070

ncluding T cells, macrophages, endothelial cells and astrocytes3,16,21].

Recent studies suggest that chemokine induction can alsoccur independently of the adaptive immune response. Forxample, infection with RNA virus that may cause encephali-is in humans, such as HIV or lymphocytic choriomeningitisiruses (LCMV), or in rodent models, such as mouse hepatitisirus (MHV) and Theiler’s virus, can directly induce the expres-ion of chemokines by astrocytes and microglia and establishhemokine gradients that promote leukocyte trafficking withinNS [2,14,17]. We undertook the present study to identify both

he source and levels of IP-10 in the mouse brain following JE.All experiments were performed according to the protocol

pproved by the Institutional Animal Ethics Committee. Threeo four days old BALB/c mice of either sex were injectedntracerebrally with approximately 100 pfu of JE virus of strainaOArS982. Control animals received the same amount of PBS.nfected mice were sacrificed at 2 and 4 days post-infection,hereas control mice were sacrificed at 4 days post-infection.Animals at days 2 and 4 post-infection and age matched con-

rols used for immunohistochemistry were perfused with 1×BS containing 7 U/mL heparin, followed by a fixative contain-

ng 2.5% para-formaldehyde in phosphate buffer. The brainsere processed for cryostat sectioning and the sections were

46 S. Bhowmick et al. / Neuroscience Letters 414 (2007) 45–50

Table 1Primers used for RT-PCR experiments and expected size of amplified products

Forward primer Reverse primer Product size (bp)

IC

swwwTFaU(mCBsu

ddiclml1actcabCd

t

rocaSosRpHrOPeaauml

NllsSawaT

Fwt

P-10 5′-GCCGTCATTTTCTGCCTCAT-3′yclophiln 5′CCATCGTGTCATCAAGGACTTCAT-3′

tained as described previously [4]. Astrocytes were labeledith rabbit anti-GFAP (1:750) and microglia were labeledith rabbit anti-mouse Iba-1 (1:500,Wako Chemicals, Japan),hich were then double stained with goat anti IP-10 (1:25).he corresponding secondary antibodies used were anti-rabbitlourescein (1:100; Vector, CA, USA) for GFAP and Iba-1, andnti-goat Alexa Red (1:500; Molecular Probes Invitrogen, CA,SA) for IP-10. Neurons were labeled with mouse anti-NeuN

1:200, Chemicon, CA, USA), which were then stained with antiouse Fluorescein (1:200, Vector). We have used DAB (Sigmahemicals) a precipitating peroxidase substrate to detect IP-10.iotinylated anti mouse IP-10 antibody was used (1:25; R&D

ystems, Minneapolis, MN, USA). Vector Elite Kit (ABC) wassed for developing.

For protein and RNA isolation, pups were sacrificed byecapitation. Cortical region from both hemispheres wereissected and homogenized in buffer containing proteinasenhibitor (Roche, Basel, Switzerland). The homogenate was thenentrifuged and 20 �g of supernatant was separated on polyacry-amide gels, electrophoresed, transferred onto nitro cellulose

embranes and immunoblots were performed as reported ear-ier [4]. Following incubation with the primary antibody (IP-10,:500), the blots were extensively washed (4×) in PBS–Tweennd further incubated for 2 h at room temperature with HRPonjugated goat anti-rabbit secondary antibody (Vector labora-ories). The membranes were then washed and visualized byhemiluminescence. The blots were stripped and reprobed withnti-�-tubulin (1:1000) to determine equivalent loading. Anti-odies against IP-10 and �-tubulin were procured from Santa

ruz Biotechnology, CA, USA. Measurements of the opticalensities for IP-10 were normalized to the levels of �-tubulin.

Total cellular RNA was isolated from virus treated and con-rol mouse, on second and fourth day post-infection, using TRI

aT(c

ig. 1. (a–c) Increased IP-10 expression in cortex of JEV infected mouse brain. Cryostere processed for IP-10 staining with DAB. A robust increase in IP-10 expression w

o the control. Inset shows an astrocyte at higher magnification. Scale bar for the pan

5′-GCTTCCCTATGGCCCTCATT-3′ 1075′TTGCCATCCAGCCAGGAGGTCT-3′ 196

eagent [5]. Total RNA (1.0 �g) was reverse-transcribed usingligo dT. Oligonucleotide primer pairs against mouse IP-10 andyclophilin mRNAs were chosen, checked for specificity usingbasic BLAST search, and prepared from Microsynth (Balgach,witzerland) (Table 1). Amplification of the mRNA sequencesf IP-10 and Cyclophilin produced single bands that were of theize predicted from the reported sequences for these mRNAs.everse-Transcriptase Polymerase chain reaction (RT-PCR) waserformed using the onestep RT-PCR kit (Qiagen Biosciences;amburg, Germany) following the manufacturer’s protocol and

eactions were carried out on Genius Techne thermal cycler.ne microgram of the total RNA was used as template in 25 �lCR reactions, containing 5× PCR buffer (5 �l), dNTPs (1 �l),nzyme mix (1 �l) specific forward and reverse primers (5 �M)nd RNase free water. The PCR products were separated on 2%garose gel, stained with ethidium bromide, and photographedsing Chemigenius Bioimaging System, Syngene. Measure-ents of the optical densities for IP-10 were normalized to the

evels of cyclophilin.The BD cytokine bead array (CBA) kit (BD Biosciences,

J, USA) was used to quantitatively measure IFN-� expressionevel in control, 2 and 4 days post-infected mouse brain tissueysates. Human CBA was used to measure IP-10 level in cultureupernatant obtained from mock infected and JE virus infectedVG and CHME3 cell line. Using 50 �l of mouse inflammationnd human chemokine standard and sample dilutions, the assayas performed according to the manufacturer’s instructions and

nalyzed by flow cytometry (FACS Calibur, Becton Dickinson).his method quantifies soluble particles, in this case cytokines

nd chemokines using fluorescence based detection mechanism.he beads, coated with desired cytokine (IFN-�) and chemokine

IP-10) react with test lysates and standards, to which fluores-ence dyes are then added. Analysis was performed using CBA

at sections from control and infected mouse brains at days 2 and 4 post-infectionas observed in response to JEV infection on day 4 post-infection, as compared

el is 20 �m.

cience

si

fhSasc

wl

ao

FfwtcN

S. Bhowmick et al. / Neuros

oftware that allows the calculation of cytokine concentrationsn unknown lysates [20].

Human cell line of fetal microglial origin CHME3 (a giftrom Dr. Steve Levison, UMDNJ, New Jersey, USA) and auman fetal astrocyte cell line SVG (a gift from Dr. Pankaj

eth, National Brain Research Center, India) were grownt 37 ◦C in Dulbecco’s Modified Eagle’s Medium (DMEM)upplemented with 10% fetal bovine serum and 1% peni-illin/streptomycin. The SVG and CHME3 cells were treated

awo

ig. 2. Co-localization of IP-10 and GFAP in infected brain sections. Cryostat sectionor IP-10 (CXCL10) and GFAP (marker for astrocytes) Panel B, Iba1 (marker for miere almost absent in control (b) sections (Panel A). Increased expression of GFAP (d

o control (Panel B). Merging the pictures (d and e) showed extensive co-localizationo-localized cells. Arrows indicate IP-10 positive cells (h and k) both in Panel C andeuN (Panel D). Scale bar for the panel is 20 �m.

Letters 414 (2007) 45–50 47

ith virus (1:5 MOI) for 1 h and then, protein were isolated 24 hater.

All the experiments performed and the data generated werenalyzed statistically by paired two-tailed Student’s t-test andne-way ANOVA test.

Brain sections of 2 and 4 days post-infected animals andge-matched controls were stained for IP-10 and developedith DAB. In contrast to the insignificant expression of IP-10bserved in the controls, a dramatic increase in IP-10 expres-

s for control (Panel A) and infected animals (Panels B–D) were double stainedcroglia) Panel C and NeuN (marker for neurons) Panel D. IP-10 positive cells) and IP-10 (e) positive cells were observed at day 4 post-infection as comparedon the fourth day post-infection. Arrows indicate the GFAP and IP-10 positiveD, but no co-localization of IP-10 was seen either with Iba-1 (Panel C) or with

48 S. Bhowmick et al. / Neuroscience Letters 414 (2007) 45–50

Fig. 3. Induction of IP-10 both at the mRNA and protein levels in mouse brainfollowing JE. (A) Induction of IP-10 mRNA was observed in the infected brain.SQ-RTPCR was performed for IP-10 and cyclophilin. Bands show a significantincrease in IP-10 mRNA levels following infection. (B) Graphical representa-tion of SQ-RTPCR. Values represent mean ± S.E.M. from three mice in eachgroup. IP-10 mRNA showed significant increase at 2 and 4 days post-infectionas compared to the control. The change in IP-10 mRNA expression between2 and 4 days post-infection was also significant. Asterisk (*) denotes signifi-cant change from control p < 0.0005 (control, 2 days) and p < 0.0005 (control,4 days); symbol (ˆ ) denotes significant change between 2 and 4 days infectedsamples p < 0.05. (C) JEV infection induced IP-10 at protein level. Immunoblotanalysis demonstrates the induction of IP-10 in the mouse brain during JEVinfection. Blots were reprobed for �-tubulin to establish equal protein loading.Immunoblot was obtained from two individual animals in each group. (D) Den-sitometric analysis of the immunoblot. Values represent means ± S.E.M. fromthree mice in each group. The 10 kDa band of IP-10 increased significantly onday 4 post-infection (*p < 0.05) as compared to control. The change between the2n

sptt

bpB

Fig. 4. Induction of IFN-� following JE virus infection. CBA was performed forIFN-� and analysed by FACS. Graph shows an increase in expression of IFN-�.Values represent means ± S.E.M. from three mice in each group. Asterisk (*)indicates significant change between control and 2 days (p < 0.05) and 4 days(s

ntctIbwo

fT4tf(

tpoiwo

irt

tCedSdlC

and 4 days post-infection was statistically significant (ˆp < 0.05). There waso significant difference between control and 2 days post-infected samples.

ion was noted at 4 days post-infection. However, at 2 daysost-infection the induction of IP-10 was sparse as compared tohe 4 days post-infected sections and was almost comparable tohe control (Fig. 1).

To determine the primary source of IP-10 expression in mouse

rain following JEV, double colour immunohistochemistry waserformed for IP-10 with GFAP (marker for astrocytes) [Panel], Iba1 (marker for microglia) [Panel C] and NeuN (marker for

p1(

p < 0.01); symbol (ˆ ) denotes significant change between 2 and 4 days infectedamples p < 0.05.

eurons) [Panel D]. A marked increase in IP-10 immunoreac-ivity on the 4 days post-infection was observed as compared toontrol (Panel A: b). A comparison of the extent of colocalisa-ion of IP-10, with the cell specific markers, demonstrated that,P-10 is predominantly expressed by astrocytes (GFAP) and noty microglia or neurons (Fig. 2). GFAP staining (Panel A: a), asell as Iba-1 and NeuN (data not shown) staining was performedn brain sections from uninfected controls.

To assess the mRNA transcript level of IP-10 in the brainollowing JEV infection we performed semi-quantitative PCR.here was a 2- and 4-fold increase in IP-10 mRNA level at 2 anddays post-infection, respectively. This increase from control

o the second day was significant (p < 0.0005) and the changerom 2 to 4 days post-infection was also significant (p < 0.05)Fig. 3).

Immunoblot analysis revealed a significant up regulation inhe expression of IP-10 on days 2 and 4 post-infection as com-ared to the control animals. While the increase in the expressionf IP-10 between the control and second day post-infection wasnsignificant, the change between the days 2 and 4 post-infectionas found to be significant (p < 0.05). Data was analyzed byne-way ANOVA test (Fig. 3).

CBA results demonstrated a significant 1.7- and 3-foldncrease in IFN-� levels at the 2 and 4 days post-infection,espectively, as compared to control (p < 0.05 and 0.01, respec-ively). Data was analyzed by one-way ANOVA test (Fig. 4).

Western Blot analysis also revealed a distinct increase inhe production of IP-10 by the SVG cells as compared to theHME3 cells, following coculture with the virus. The differ-nce in the IP-10 level between SVG and CHME3 cells at 1ay post-infection (p < 0.0005) and between control and infectedVG cells were found to be significant (p < 0.0005). CBA resultsemonstrated a significant 3.7- and 6.4-fold increase in IP-10evels in SVG infected cells as compared to SVG control andHME3 infected cells, respectively (p < 0.001 for both). This

rovides clear evidence that following infection in vitro, IP-0 is produced predominantly by astrocytes and not microgliaFig. 5).

S. Bhowmick et al. / Neuroscience

Fig. 5. Differential induction of IP-10 in CHME3 and SVG cells following co-culture with JE virus. (A) Representative immunoblot analysis demonstrates theinduction of IP-10 in CHME3 and SVG cells following co-culture with JEV (1:5MOI). Blots were reprobed for �-tubulin to establish equal protein loading. (B)Densitometric analysis of the immunoblot. Values represent means ± S.E.M.from three independent experiments. Expression of IP-10 by SVG cells fol-lowing infection with JE virus was significantly greater than that of CHME3(ˆp < 0.0005). Also the difference in IP-10 levels between control and infectedSVG cells, was found to be significant (*p < 0.0005). (C) CBA was performedto determine the level of IP-10 in culture supernatant obtained from the aboveexperiments. Graphical representation of CBA results shows that there is anupregulation of IP-10 expression in SVG cells following co culture with JE virus.Induction of IP-10 in the culture supernatant obtained from JE virus infectedSVG cell is significantly greater than that obtained from virus infected CHME3cSt

coem4Tc

tiaCf

tCw[iipHsaOd

1tafiliitfttoigi

A

1W

R

ells (ˆp ≤ 0.001). The difference in IP-10 levels between control and infectedVG cells were significant (*p ≤ 0.001). Values represent means ± S.E.M. from

hree independent experiments.

The present study was undertaken to better understand theontributions of IP-10 in the development of neuropathologybserved in JE. To this end, we have found that there is a robustxpression of IP-10 following JE and that astrocytes are theajor source of IP-10. The elevated expression of IP-10 ondays post-infection is concurrent with the death of animal.

his suggests that marked increase in the expression of thishemokine might play crucial role in outcome of JEV infection.

Cerebral infection with JEV in mice is characterized athe pathological level by significant recruitment of immuno-

nflammatory cells to the sites of viral infection. Reports havelso shown that the CXC chemokine IP-10 and its receptorXCR3, play a vital role in the inflammatory response in CNS

ollowing viral infections such as LCMV [1], mouse hepati-

Letters 414 (2007) 45–50 49

is virus TV [9] and HIV [11]. It is well established that theNS response to viral infection and inflammation, manifestsith accompanied reactive gliosis largely attributed to astrocytes

7]. Several studies have shown elevation of CXCL10 in CSFn association with mononuclear pleocytosis, including in HIVnfection [11,13]. Presence of elevated levels of IP-10, producedredominantly by astrocytes, has been observed in the brains ofIV-1 infected patients [11]. As IFN-� strongly induces expres-

ion of IP-10 [18], an increase in the level of IP-10 should bessociated with a simultaneous increase in the levels of IFN-�.ur data clearly demonstrated increase in the level of IFN-�uring a progressive infection.

We hypothesize that although the over expression of IP-0 and the ensuing massive inflammatory response is aimedowards resolving infection, it overwhelms the CNS and infectednimals eventually succumb to this inflammation. While ourndings contradict reports that suggest that IP-10 plays a regu-

atory role in conferring protective immunity during infection, its in line with the finding that CXCL10 contributes to an increasen clinical disease severity [15]. Our result underscore the impor-ance of IP-10 in mediating inflammatory responses of the brainollowing JE and emphasizes the need for future studies to iden-ify mechanisms by which this response can be controlled, so aso provide protection. Several reports have shown that, the usef blocking antibody for IP-10, reduces the T-cell infiltration,n the affected tissue [6,8]. Thus, IP-10 can be a potential tar-et for drugs, which aims to block its activity thereby reducingnflammation.

cknowledgements

This work was supported by grant (BT/PR/5799/MED/4/698/2005) from the Department of Biotechnology to A.B.e thank Director, NBRC for her support and encouragement.

eferences

[1] V.C. Asensio, I.L. Campbell, Chemokine gene expression in the brainsof mice with lymphocytic choriomeningitis, J. Virol. 71 (1997) 7832–7840.

[2] V.C. Asensio, I.L. Campbell, Chemokines in the CNS: plurifunctionalmediators in diverse states, Trends Neurosci. 22 (1999) 504–512.

[3] A. Bajetto, R. Bonavia, S. Barbero, G. Schettini, Characterization ofchemokines and their receptors in the central nervous system: physiopatho-logical implications, J. Neurochem. 82 (2002) 1311–1329.

[4] A. Basu, J.K. Krady, M. O’Malley, S.D. Styren, S.T. DeKosky, S.W. Levi-son, The type 1 interleukin-1 receptor is essential for the efficient activationof microglia and the induction of multiple proinflammatory mediators inresponse to brain injury, J. Neurosci. 22 (2002) 6071–6082.

[5] P. Chomczynski, N. Sacchi, Single-step method of RNA isolation by acidguanidinium thiocyanate-phenol-chloroform extraction, Anal. Biochem.162 (1987) 156–159.

[6] J.H. Dufour, M. Dziejman, M.T. Liu, J.H. Leung, T.E. Lane, A.D. Luster,IFN-gamma-inducible protein 10 (IP-10; CXCL10)-deficient mice reveala role for IP-10 in effector T cell generation and trafficking, J. Immunol.168 (2002) 3195–3204.

[7] L.F. Eng, R.S. Ghirnikar, GFAP and astrogliosis, Brain Pathol. 4 (1994)229–237.

[8] B.T. Fife, K.J. Kennedy, M.C. Paniagua, N.W. Lukacs, S.L. Kunkel, A.D.Luster, W.J. Karpus, CXCL10 (IFN-gamma-inducible protein-10) controlof encephalitogenic CD4+ T cell accumulation in the central nervous sys-

5 cience

[

[

[

[

[

[

[

[

[

[

[lation of IL-10 and IL-12p40 in secondary progressive multiple sclerosis,J. Neuroimmunol. 146 (2004) 209–215.

0 S. Bhowmick et al. / Neuros

tem during experimental autoimmune encephalomyelitis, J. Immunol. 166(2001) 7617–7624.

[9] L.M. Hoffman, B.T. Fife, W.S. Begolka, S.D. Miller, W.J. Karpus, Cen-tral nervous system chemokine expression during Theiler’s virus-induceddemyelinating disease, J. Neurovirol. 5 (1999) 635–642.

10] N. Khanna, M. Agnihotri, A. Mathur, U.C. Chaturvedi, Neutrophilchemotactic factor produced by Japanese encephalitis virus stimulatedmacrophages, Clin. Exp. Immunol. 86 (1991) 299–303.

11] S.A. Kolb, B. Sporer, F. Lahrtz, U. Koedel, H.W. Pfister, A. Fontana,Identification of a T cell chemotactic factor in the cerebrospinal fluid ofHIV-1-infected individuals as interferon-gamma inducible protein 10, J.Neuroimmunol. 93 (1999) 172–181.

12] R. Kumar, A. Mathur, A. Kumar, S. Sharma, S. Chakraborty, U.C.Chaturvedi, Clinical features & prognostic indicators of Japaneseencephalitis in children in Lucknow (India), Indian J. Med. Res. 91 (1990)321–327.

13] F. Lahrtz, L. Piali, D. Nadal, H.W. Pfister, K.S. Spanaus, M. Baggiolini, A.Fontana, Chemotactic activity on mononuclear cells in the cerebrospinalfluid of patients with viral meningitis is mediated by interferon-gammainducible protein-10 and monocyte chemotactic protein-1, Eur. J. Immunol.

27 (1997) 2484–2489.

14] B.R. Lane, R.M. Strieter, M.J. Coffey, D.M. Markovitz, Human immunod-eficiency virus type 1 (HIV-1)-induced GRO-alpha production stimulatesHIV-1 replication in macrophages and T lymphocytes, J. Virol. 75 (2001)5812–5822.

[

Letters 414 (2007) 45–50

15] M.T. Liu, H.S. Keirstead, T.E. Lane, Neutralization of the chemokineCXCL10 reduces inflammatory cell invasion and demyelination andimproves neurological function in a viral model of multiple sclerosis, J.Immunol. 167 (2001) 4091–4097.

16] J. Melchjorsen, L.N. Sorensen, S.R. Paludan, Expression and function ofchemokines during viral infections: from molecular mechanisms to in vivofunction, J. Leukoc. Biol. 74 (2003) 331–343.

17] J.P. Palma, B.S. Kim, The scope and activation mechanisms of chemokinegene expression in primary astrocytes following infection with Theiler’svirus, J. Neuroimmunol. 149 (2004) 121–129.

18] A. Sauty, M. Dziejman, R.A. Taha, A.S. Iarossi, K. Neote, E.A. Garcia-Zepeda, Q. Hamid, A.D. Luster, The T cell-specific CXC chemokines IP-10, Mig, and I-TAC are expressed by activated human bronchial epithelialcells, J. Immunol. 162 (1999) 3549–3558.

19] A. Singh, R. Kulshreshtha, A. Mathur, Secretion of the chemokineinterleukin-8 during Japanese encephalitis virus infection, J. Med. Micro-biol. 49 (2000) 607–612.

20] S.S. Soldan, A.I. Alvarez Retuerto, N.L. Sicotte, R.R. Voskuhl, Dysregu-

21] D.D. Taub, K. Conlon, A.R. Lloyd, J.J. Oppenheim, D.J. Kelvin, Preferen-tial migration of activated CD4+ and CD8+ T cells in response to MIP-1alpha and MIP-1 beta, Science 260 (1993) 355–358.