of February 8, 2018.This information is current as

Primary Neurons and Infected MiceNeurotropic Measles Virus Challenge in STAT1-Independent Control of a

Balachandran and Glenn F. RallEmmanuelle Nicolas, Michael Slifker, Siddharth Lauren A. O'Donnell, Stephen Conway, R. Wesley Rose,

http://www.jimmunol.org/content/188/4/1915doi: 10.4049/jimmunol.1101356January 2012;

2012; 188:1915-1923; Prepublished online 13J Immunol 

MaterialSupplementary

6.DC1http://www.jimmunol.org/content/suppl/2012/01/13/jimmunol.110135

        average*  

4 weeks from acceptance to publicationSpeedy Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

?The JIWhy

Referenceshttp://www.jimmunol.org/content/188/4/1915.full#ref-list-1

, 27 of which you can access for free at: cites 53 articlesThis article

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2012 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on February 8, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on February 8, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

STAT1-Independent Control of a Neurotropic Measles VirusChallenge in Primary Neurons and Infected Mice

Lauren A. O’Donnell,* Stephen Conway,* R. Wesley Rose,† Emmanuelle Nicolas,*

Michael Slifker,* Siddharth Balachandran,* and Glenn F. Rall*

Neurons are chiefly nonrenewable; thus, cytolytic immune strategies to clear or control neurotropic viral infections could have

lasting neurologic consequences. IFN-g is a potent antiviral cytokine that is critical for noncytolytic clearance of multiple

neurotropic viral infections, including measles virus (MV); however, the downstream pathways through which IFN-g functions

in neurons have not been defined. Unlike most cell types studied to date in which IFN-g affects gene expression via rapid and

robust activation of STAT1, basal STAT1 levels in primary hippocampal neurons are constitutively low, resulting in attenuated

STAT1 activation and consequently slower kinetics of IFN-g–driven STAT1-dependent gene expression. Given this altered ex-

pression and activation of STAT1 in neurons, we sought to determine whether STAT1 was required for IFN-g–mediated protection

from infection in neurons. To do so, we evaluated the consequences of MV challenge of STAT1-deficient mice and primary

hippocampal neurons explanted from these mice. Surprisingly, the absence of STAT1 did not restrict the ability of IFN-g to

control viral infection either in vivo or ex vivo. Moreover, the canonical IFN-g–triggered STAT1 gene expression profile was not

induced in STAT1-deficient neurons, suggesting that IFN-g regulates neuronal STAT1-independent pathways to control viral

replication. The Journal of Immunology, 2012, 188: 1915–1923.

Central nervous system invasion is a rare, but typicallysevere, consequence of several human viral infections thatinclude most herpesviruses, poliovirus, measles virus,

flaviviruses (e.g., West Nile virus), and HIV (1). When CNS infec-tions do occur, complications typically result in lasting neuro-logic damage and often in death, especially in the young, elderly,and immunocompromised (2). The severity of neurotropic viruspathogenicity may be due to the delicate balance that must beachieved by the antiviral immune response: effective immunitymust control the infection while minimizing CTL-mediated lysis ofthis chiefly nonrenewable cell population. Whereas multiple ani-mal models have demonstrated that cytokines, including IFN-g,can limit viral spread in the brain via noncytopathic mechanisms(3–6), the downstream signaling pathways that result in clearanceremain largely undefined. This becomes clinically significant inthose instances in which viral clearance fails, as in many cases ofpediatric encephalitis including measles and rubella, and neuro-

tropic infections of the elderly, including West Nile virus and St.Louis encephalitis virus (1, 7–10).IFN-g is the only type II member of the IFN family, which also

consists of type I IFNs (IFN-ab) and type III IFNs. Unlike type IIFNs, which are expressed by most cells soon postinfection, IFN-gis chiefly produced by activated immune cells such as NK cellsand T cells. IFN-g initiates a cellular response by binding to theIFN-g receptor complex (consisting of a heterotetramer of IFN-gR1 and R2 subunits), which triggers activation of receptor-associated JAK-1/2 and subsequent tyrosine phosphorylation ofthe cytoplasmic tail of the IFN-gR1 subunits. STAT1 is recruitedto the phosphorylated R1 subunit, where it is phosphorylated,homodimerizes, and translocates to the nucleus. The phosphory-lated STAT1 homodimer binds to g-activated sequence elementswithin IFN-g–responsive genes to initiate transcription. Over 250genes are induced in this manner to inhibit viral spread (11).Whereas STAT1 is central to a classical IFN-g response, a sub-stantial number of studies has also demonstrated the presence ofIFN-g–dependent, STAT1-independent pathways (12–20).IFN-g is required for viral clearance of many neurotropic viral

and bacterial infections (4, 6, 21–26). Of note, however, distinctimmune strategies may be employed to resolve CNS infections,depending on the cell type that is infected. For example, mousehepatitis virus can infect CNS resident astrocytes, microglia, andoligodendrocytes (27); whereas perforin is sufficient to mediateviral clearance from astrocytes and microglia, IFN-g is sufficientfor mouse hepatitis virus control in oligodendrocytes (28, 29). Inaddition, certain subsets of neural cells, such as neural precursors inthe retina, preferentially use STAT3 instead of STAT1 in responseto IFN-g (30). Numerous studies have also shown that both STAT1-dependent and -independent pathways are likely to be importantfor IFN-mediated viral clearance in other regions of the body (13,18, 19, 31, 32). For example, STAT1 plays a biphasic role in controlof systemic dengue infection in mice: STAT1-dependent pathwaysare required for early viral control, but STAT1-independent path-ways are required for later control and eventual viral clearance (13,

*Program in Immune Cell Development and Host Defense, Fox Chase Cancer Cen-ter, Philadelphia, PA 19111; and †Department of Biology, Arcadia University, Glen-side, PA 19038

Received for publication May 10, 2011. Accepted for publication December 9, 2011.

This work was supported by the following sources: National Institutes of HealthGrants RO1 NS40500 (to G.F.R.) and F32 NS062519-01 (to L.A.O.), a grant fromAutism Speaks (to G.F.R.), a gift from the F.M. Kirby Foundation (to G.F.R.), andGrant P30CA006927 from the National Cancer Institute.

The dataset presented in this article has been submitted to the Gene ExpressionOmnibus (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE33057.

Address correspondence and reprint requests to Dr. Glenn F. Rall, Fox Chase CancerCenter, 333 Cottman Avenue, Philadelphia, PA 19111. E-mail address: glenn.rall@fccc.edu

The online version of this article contains supplemental material.

Abbreviations used in this article: DPBS, Dulbecco’s PBS; dpi, days postinfection;GFAP, glial fibrillary acidic protein; hpi, hours postinfection; IC, intracerebral; KO,knockout; MEF, mouse embryonic fibroblast; MOI, multiplicity of infection; MV,measles virus; NSE, neuron-specific enolase; PBS-T, PBS-Tween 20; RT, room tem-perature.

Copyright� 2012 by The American Association of Immunologists, Inc. 0022-1767/12/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1101356

by guest on February 8, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

18, 32). Thus, although IFN-g has been conclusively linked withnoncytolytic clearance of multiple neurotropic infections, the roleof STAT1, particularly in neurons, remains less well defined.We have developed a mouse model of neuron-restricted measles

virus (MV) infection, in which CNS neurons express the MVreceptor (CD46) under the control of the neuron-specific enolase(NSE) promoter (NSE-CD46+ mice) (33). Using this model, wepreviously showed that adult NSE-CD46+ mice clear MV infec-tion from CNS neurons without neurologic damage or neuronalloss in an IFN-g–dependent and T cell-dependent manner (3, 34).IFN-g can act directly upon neurons to induce an antiviral state, asdemonstrated by the direct antiviral ability of IFN-g to limit MVreplication in purified CD46+ neurons cultured ex vivo (3). Sur-prisingly, however, the neuronal signaling response to IFN-gtreatment was distinct from that observed in control mouse em-bryonic fibroblasts (MEFs) (35). Specifically, neurons respondedto IFN-g with delayed and attenuated STAT1 expression and ac-tivation kinetics in comparison with control fibroblasts. This wasreflected in the delayed and reduced expression of classical IFN-g–dependent genes. These previous studies suggest that IFN-gactivates a critical antiviral program in neurons, but that STAT1may play a subordinated role in this response.In this report, we examined the requirement for STAT1 in the IFN-

g–mediated control of MVinfection and spread using both an ex vivomodel (infected primary hippocampal neurons), as well as infectionof NSE-CD46+ mice on a STAT1-deficient background.We observedthat STAT1was dispensable for the resolution of aMVCNS infectionand that the canonical gene profile induced by IFN-g was not acti-vated in neurons. These data indicate that IFN-g–triggered STAT1-independent pathways are most likely operative in CNS neurons.

Materials and MethodsCells, viruses, mice, and infections

Primary hippocampal neurons were prepared from embryonic (E15–16)mice, as previously described (35–38). Neurons were plated on 15-mmglass coverslips or in 12-well plates coated with poly-L-lysine (Sigma-Aldrich) at a density of 2.5 3 105 cells/well, unless otherwise noted.Neuron cultures were routinely .95% MAP2 positive. Neurons wereplated and incubated at 37˚C in a humidified incubator with 5% CO2 for5 d prior to IFN-g treatment or infection to allow for full differentiation.

Primary astrocyte cultures were prepared from E16 cortices, as describedpreviously (35, 39), and were routinely.95% glial fibrillary acidic protein(GFAP) positive. After 7 d in culture, monolayers were trypsinized, re-plated in poly-L-lysine–coated 6-well plates (2 3 105 cells/well), and in-cubated for 48 h before treatment.

MV-Edmonston (vaccine strain) was purchased from American TypeCulture Collection and passaged and titered in Vero cells. Passages 2 or 3 ofthe MV stock were used for intracerebral (IC) injections and in vitro in-fection assays.

Inbred C57BL/6 (H-2b) and homozygous NSE-CD46+ transgenic mice(line 18; H-2b) (33) were maintained in the closed breeding colony of theFox Chase Cancer Center. All experimental protocols were reviewed andapproved by the Institutional Animal Care and Use Committee. Homo-zygous NSE-CD46+ and haplotype-matched homozygous immune knock-out (KO) mice were intercrossed for three or more generations to obtainNSE-CD46+ mice on the desired KO background. STAT1-KO mice (40)were originally obtained from The Jackson Laboratory. Genotypes of allmice used in these experiments were confirmed by either PCR analysis oftail biopsy DNA, or by flow cytometry of PBLs immunostained withfluorophore-conjugated Abs to mouse B and T cell markers.

Anafane-anesthetized mice were infected with MV via IC inoculation (23104 PFU in a volume of 20 ml, delivered along the midline using a 27-gneedle). Mice were monitored daily postinfection for signs of illness, includingweight loss, ruffled fur, ataxia, and seizures. Moribund mice were euthanizedin accordance with Institutional Animal Care and Use Committee guidelines.

Primary neuron infections and IFN-g treatments

Five days postplating, primary hippocampal neurons from NSE-CD46+

mice and NSE-CD46+/STAT1-KO mice were infected with MV (multi-plicity of infection [MOI] 1.0) for 1 h. Thereafter, the inoculum was re-

moved, the neurons were washed twice in Dulbecco’s PBS (DPBS), andthe cells were placed in conditioned Neurobasal media.

For cells treated with IFN-g, murine IFN-g (BD Transduction Labo-ratories) was diluted in B27-free Neurobasal media, added to the cultures(100 U/ml final concentration), and incubated for the indicated times priorto collection.

Immunofluorescence assay

At the indicated time points post–IFN-g addition, cells on coverslips werefixed in a 1:1 solution of methanol:acetone for 5 min at 220˚C andrehydrated in DPBS. To prevent nonspecific Ab binding, cells were in-cubated in a 10% goat serum/20% FBS solution in DPBS for 1 h at roomtemperature (RT). A primary Ab was then added in blocking solution for1 h at RT. Abs, sources, and concentrations used included the following:monoclonal anti–phospho-STAT1 (pY701), BD Pharmingen (1:500); rab-bit polyclonal anti-MAP2, Chemicon (1:250); rabbit polyclonal anti-GFAP,Novus Biologicals (1:250); and Keller anti-MV human serum, a gift of M.Oldstone (The Scripps Research Institute) (1:1000). After three washes inDPBS, the cells were incubated in the appropriate secondary Ab inblocking solution for 1 h at RT: anti-mouse AlexaFluor-555 (1:10,000) oranti-rabbit AlexaFluor-488 (1:10,000) with Hoechst 33342 (10 mM) nu-clear stain, all from Molecular Probes. After three washes in DPBS,coverslips were mounted onto glass slides using Citi-Fluor AF1 (ElectronMicroscopy Sciences). Images were captured using an Inverted TE2000with Nikon C1 confocal scanhead (original magnification 340 objective)in the Fox Chase Cancer Center Confocal Facility.

Immunoblots

Cells were treated with IFN-g and collected at the indicated times, andimmunoblots were performed, as previously described (35). Abs usedincluded the following: anti-STAT1 (1:1000); phospho-specific STAT1(pY701; 1:1000); Keller anti-MV (1:1000); and anti- GAPDH (1:10,000;Chemicon). All were diluted in PBS-Tween 20 (PBS-T) containing 5%milk. After three washes in PBS-T (10 min each), the blots were incubatedin secondary Ab solution for 1 h at RT. Abs included the following: goatanti-rabbit HRP (1:1000; Vector Laboratories) for anti-STAT1; goat anti-mouse HRP (1:2000; Santa Cruz) for anti-pSTAT1; and anti-human-HRPfor Keller anti-MV (1:10,000), all diluted in PBS-T with 5% milk. Forquantitative analysis of immunoblots, densitometric analysis was per-formed using National Institutes of Health Image software (v. 1.63). Whenreprobing was necessary, blots were stripped in an acidic glycine/SDSsolution for 2 h at RT.

Immunohistochemical analysis of mouse tissues

Brains from four to five mice/time point were removed, immersed in tissue-embedding compound (Fisher), snap frozen in a dry ice-isopentane bath,and stored at 270˚C. Horizontal cryosections (10 mm) were air driedand stored at 270˚C. Standard immunohistochemistry using diamino-benizidene precipitation (3, 41) was performed. To detect MV-infectedcells, Keller anti-MV serum was used at a dilution of 1:2000, followedby a biotinylated anti-human secondary Ab (1:300; Vector). Rat anti-mouse CD4 (clone RM4-5; 1:100; BD Pharmingen) or rat anti-mouseCD8 a/CD8 b Abs (clones 53-6.7 and 53-5.8, respectively; 1:100 each;BD Pharmingen) were used to identify T lymphocyte subsets. Sectionswere then sequentially incubated for 1 h at RT with a biotinylated anti-ratIgG secondary Ab (1:200; Vector Laboratories), 30 min with a streptavi-din-peroxidase conjugate (ABC Elite; Vector Laboratories), and, finally,for 5–10 min with diaminobenzidine (0.7 mg/ml in 60 mM Tris) and urea-H2O2 (0.2 mg/ml), purchased as preweighed tablets (Sigma-Aldrich). Allcells were counterstained with hematoxylin and mounted with an aqueousmounting medium. Uninfected tissues or omission of the primary Abserved as negative controls. For all histological analyses, at least threesections per brain were examined from three different horizontal levels,and at least four mice per experimental group were assessed.

Quantitative RT-PCR

Mouse brains were snap frozen in liquid nitrogen and stored at280˚C. RNAwas isolated by TRIzol, according to the manufacturer’s instructions(Sigma-Aldrich). Contaminating DNA was removed from RNA prepara-tions using DNase I treatment (Invitrogen). Purified RNA was quantifiedusing a Nanodrop instrument. RNAwas reverse transcribed using Moloneymurine leukemia virus reverse transcriptase (Ambion) and a mixture ofanchored oligo-dT and random decamers. For each sample, two reverse-transcriptase reactions were performed with inputs of 100 and 20 ng. Analiquot of the cDNA was used for 59-nuclease assays using TaqManchemistry. For IFN-g expression, Assay-on-Demand Mm00801778_m1

1916 STAT1-INDEPENDENT VIRAL CONTROL IN NEURONS

by guest on February 8, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

was used. ATaqMan set specific for the N gene of MV (GenBank sequenceAB046218) was used for detecting viral RNA. Sequences were as follows:forward, 59-CGCAGGACAGTCGAAGGTC-39; reverse, 59-TTCCGA-GATTCCTGCCATG-39; probe, 59-6Fam-TGACGCCCTGCTTAGGCTG-CAA-BHQ1-39. Assays were used in combination with Universal Mastermix and run on a 7900 HT sequence detection system (Applied Bio-systems). Cycling conditions were 95˚C, 15 min, followed by 40 (two-step) cycles (95˚C, 15 s; 60˚C, 60 s). The assay for MV-N was validatedwith a 4-fold five-points dilution curve of cDNA. The slope was 23.54,corresponding to a PCR efficiency of 95%. For each sample, the values areaveraged and SD of data are derived from two independent PCRs. Relativequantification to the control was done using the comparative cycle thresh-old method.

Flow cytometric analysis of brain infiltrates

On the indicated day postinfection (dpi), mice were deeply anesthetizedwith 400 ml 3.8% chloral hydrate in PBS, delivered i.p. Once animals wereconfirmed to be nonresponsive, the mice were perfused with 30 ml PBS.Following perfusion, each brain and spleen were removed and pressedthrough a nylon mesh cell strainer in PBS. Dissociated tissue was run overa 30/70% discontinuous Percoll gradient for 20 min at 4˚C. Mononuclearcells were recovered from the interface, washed with PBS, treated with0.84% ammonium chloride to remove contaminating RBCs, and washedagain. Collected mononuclear cells were counted using a standard hemo-cytometer and plated into a V-bottom 96-well plate for subsequent Abstaining for multicolor flow cytometry. The following Abs (eBioscience)were used: PE-Cy5-CD8a, allophycocyanin-Ax750-CD4, PE-CD3e, FITC-CD11c, allophycocyanin-CD161c (NK1.1), PB-CD19, PE-Cy5.5 Gr-1(Ly-6G), PE-Cy7-CD49b (DX5), and Ax700-CD11b. Cells were allowedto incubate with Ab for 1 h at 4˚C and then washed following the incu-bation period. Pelleted, stained cells were resuspended and read in a BDLSR II system. Percentages obtained from flow cytometry were combinedwith hemocytometer counts to calculate total cell numbers.

Microarray

Primary hippocampal neurons were plated in poly-L-lysine–coated 6-wellplates at a density of 13 106 cells/well. RNAwas purified from whole-cell

lysates using the RNeasy mini kit (Qiagen), and contaminating DNA wasremoved using a RNase-free DNase set. A quantity amounting to 500 ngtotal RNA was amplified and labeled using the low RNA input linearamplification kit (Agilent). Labeled cRNA targets were hybridized ontoAgilent 4344k mouse whole genome arrays. Microarray images wereprocessed using Agilent Feature Extraction software (version 9.5). Datawere background corrected using the normexp method (PMID: 17720982)implemented in the Bioconductor package limma (42, 43), and quantilenormalized. Identification of differentially expressed genes was performedwith empirical Bayes moderated t tests using limma. Biological pathwaysand networks were examined with Ingenuity Pathway Analysis software(www.ingenuity.com). The dataset is at the Gene Expression Omnibus(http://www.ncbi.nlm.nih.gov/geo/), accession number GSE33057.

ResultsIFN-g–mediated STAT1 signaling is attenuated in CNSneurons

We previously showed that IFN-g is required for control of MV ina transgenic mouse model (NSE-CD46+), in which infection isneuron restricted (3). To define the mechanism by which IFN-gfunctions, we determined the efficacy of IFN-g–mediated controlof MV replication in primary neurons. Explanted CD46+ hippo-campal neurons were treated with murine rIFN-g (100 U/ml) or anequivalent amount of heat-inactivated IFN-g; 18 h thereafter, theneurons were infected with MV-Ed (MOI = 1), and relative in-fection levels were determined by quantifying viral protein load.Note that standard plaque assays cannot be used to measure virallevels because no infectious virus is released from infected neu-rons (44). Neuronal lysates were harvested 48 hours postinfection(hpi) for Western blot analysis. As shown in Fig. 1A and 1B, viralprotein levels were .75% reduced in the presence of IFN-g,whereas heat-inactivated IFN-g had no effect. These results con-firm that IFN-g limits MV replication in primary neurons.

FIGURE 1. IFN-g treatment limits MV neuronal infection, despite low basal levels of STAT1 and delayed STAT1 phosphorylation. A, Hippocampal

neurons explanted from NSE-CD46+ embryos were treated with IFN-g (100 U/ml, 18 h) or heat-inactivated (DH) IFN-g (100 U/ml, 18 h) and infected with

MV-Ed (MOI = 1). At 48 hpi, neurons were lysed in protein solubilization buffer and subjected to Western blot for MV Ag and GAPDH as a loading

control. B, Western blots from A were quantified by densitometry using ImageJ software. MV signal was normalized to GAPDH as loading control.

Statistical analysis was applied using a paired t test (n = 4, *p , 0.05 versus MV infected). C and D, Hippocampal neurons and cortical astrocytes from

NSE-CD46+ embryonic mice were treated with IFN-g (100 U/ml) for the indicated times and collected for Western blot analysis (C) or fixed for im-

munofluorescence (D). For Western blots (C), whole-cell lysates were subjected to Western blot for phosphorylated STAT1 (STAT1-P; tyrosine 701), total

STAT1, and GAPDH. D, Coverslips were stained with a mAb against STAT1-P (red), and a polyclonal Ab against GFAP or MAP2 as astrocyte or neuronal

markers, respectively (green). All experiments were conducted in triplicate. Original magnification 3400.

The Journal of Immunology 1917

by guest on February 8, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Our previous work (35) suggested that the downstream signalingresponse of primary neurons following IFN-g engagement wasdistinct from classical IFN-g–activated STAT1-dependent signal-ing (11). To determine whether this altered signaling profile wastrue for other brain parenchymal cells, primary hippocampal neu-rons were compared with primary astrocytes, which are derivedfrom the same pool of neural precursor cells (45). Both pri-mary cultures were consistently .95% pure, as defined by MAP2and GFAP staining, respectively (Fig. 1D). Cultures were treatedwith IFN-g (100 U/ml) for the indicated times, and samples werethen analyzed for both total and phosphorylated STAT1 byWestern blot (Fig. 1C) and by immunofluorescence (Fig. 1D).Like MEFs (35), astrocytes expressed abundant endogenous levelsof STAT1, which was rapidly phosphorylated after IFN-g treat-ment (Fig. 1C, 1D, top row). In contrast, primary hippocampalneurons expressed low basal levels of STAT1, and phosphorylatedSTAT1 was detectable in neurons only after extended IFN-gstimulation (24 h), coincident with an increase in available STAT1protein (Fig. 1C). These data demonstrate that delayed STAT1phosphorylation is not a general characteristic of resident brainparenchymal cells, but rather is unique to neurons.

STAT1 is not required for MV control in primary neurons

Despite reduced expression and activation of neuronal STAT1, itwas possible that these low levels were still sufficient to initiatean effective antiviral signal. Thus, to determine what role STAT1played in IFN-g–mediated MV clearance, CD46+ and CD46+/

STAT1-KO neurons were treated with IFN-g and infected withMV, as described above; neurons were then harvested for Westernblot analysis for MVAg or fixed for immunofluorescence. Quan-tified results are shown in Fig. 2A, and representative images areshown in Fig. 2B. Without IFN-g, increasing levels of MV Agwere detectable in both CD46+ and CD46+/STAT1-KO neuronsfrom 24 to 72 hpi, evidenced by the increasing size of viral Ag-positive foci (Fig. 2B). In contrast, infected CD46+ and CD46+/STAT1-KO neurons treated with IFN-g showed equivalently lessMV Ag at each time point; there were no statistical differencesbetween STAT1-expressing and KO neurons. These data suggestthat STAT1 is not required for IFN-g–mediated control of MV inneurons.IFN-g could abate MV growth either by directly eliminating

MV from infected neurons, by limiting trans-synaptic spread (44),or by some combination of the two. To assess this, receptor-expressing CD46+ neurons were infected with MV, and treatedwith IFN-g either 24 h before, 24 h after, or coincident with MVinfection. IFN-g–mediated control was most effective when neu-rons were pretreated with IFN-g (Fig. 2C, compare lanes 2 and 1);the antiviral effect was negligible when IFN-g was added post-infection (compare lanes 4 and 1). Moreover, in the presence ofIFN-g, MV Ag-positive neurons were typically found as singleimmunopositive cells, in contrast to abundant multineuronal net-works in the absence of IFN-g (Fig. 2B, quantified in Fig. 2D),suggesting that IFN-g may play a role in blocking viral spreadrather than directly resolving the infection.

FIGURE 2. Primary neurons use STAT1-independent pathways during IFN-g–mediated viral control. A, NSE-CD46+ and NSE-CD46+/STAT1-KO

hippocampal neurons were treated with IFN-g (100 U/ml) and infected with MV (MOI = 1) for 24 or 48 h. Whole-cell lysates were harvested, subjected to

Western blot analysis, and quantified on ImageJ software, with MV signal normalized to GAPDH. The average MV signal is plotted with MV infection at

24 h set to 100%, and error bars represent SEM (n = 3). B, Immunofluorescence analysis of MVAg in CD46+ and CD46+/STAT1-KO neurons. Neurons

were treated as in A and stained for MVAg (red), MAP2 (green), and Hoescht (blue). Original magnification3400. C, Primary CD46+ neurons were treated

with IFN-g either 24 h before infection (lane 2), coincident with infection (lane 3), or 24 h postinfection (lane 4). Lysates were collected from infected (and

corresponding uninfected) neurons, and blots probed for MVAg and GAPDH, as described. D, CD46+ neurons were either treated with IFN-g (100 U/ml)

or left untreated and, 24 h later, infected with MV. At 48 hpi, coverslips were collected and immunostained for MVAgs, and nuclei within immunopositive

clusters were counted. Clusters were “binned” into groups of 1–3, 4–6, 7–10, or .10 (n = 3). SDs are shown from three such experiments.

1918 STAT1-INDEPENDENT VIRAL CONTROL IN NEURONS

by guest on February 8, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

STAT1-deficient mice survive viral infection in CNS neurons

To determine what role STAT1 plays in the resolution of a neuronalviral infection in vivo, we infected NSE-CD46+ transgenic mice thathad been backcrossed to selective immune KO mice. NSE-CD46+

mice lacking T and B cells (NSE-CD46+/RAG2-KO), IFN-g(NSE-CD46+/IFN-g–KO), or STAT1 (NSE-CD46+/STAT1-KO)were infected IC with MV and monitored daily for signs of illness(Table I) and mortality (Fig. 3). Consistent with our publishedresults (3), immunocompetent NSE-CD46+ mice remained diseasefree, and all survived, whereas .95% of NSE-CD46+/RAG-2 KOmice and ∼50% of NSE-CD46+/IFN-g KO mice died of unre-stricted MV infection between 7 and 17 dpi, during the period ofmaximal T cell presence within the brain parenchyma (dottedhorizontal line, Fig. 3) (34). Of note, surviving NSE-CD46+/IFN-g–KO mice showed lasting signs of neurologic damage, includingataxia, piloerection, and hunched posture (Table I). Unexpectedly,NSE-CD46+/STAT1-KO mice segregated into two groups fol-lowing MV infection. Consistently, a small proportion (∼25%; 9of 35 mice in the experiment shown) died between 4 and 6 dpi,earlier than any other KO mouse tested in this model system, andprior to the entry of CD4+ and CD8+ T cells into the CNS (34).This early death was accompanied by grade 5 “popcorn” seizures[based on a modified Racine scale (46)] and temporary paralysis,followed by death within 24 h. Importantly, however, the majorityof NSE-CD46+/STAT1-KO mice survived this initial 6-d period,and showed no signs of neurologic damage at any time pointthereafter (Table I). Pairwise analysis contingent on survival to 8dpi indicated no statistically significant difference between NSE-CD46+ mice and NSE-CD46+/STAT1 KO mice, although survivaldifferences did reach significance (p , 0.05) for the other geno-types tested. Based on these observations, we conclude thatSTAT1 plays a biphasic role during MV infection, and that STAT1is dispensable for MV clearance from neurons during the adaptiveimmune phase.

T cell infiltration into the CNS and control and clearance ofMV in neurons are STAT1 independent

Studies using STAT1 KO mice have implicated a role for STAT1 inboth lymphocyte proliferation and migration (47–49). To addresswhether the absence of STAT1 affected the timing or magnitudeof T cell infiltration into the infected brain parenchyma, immunecell profiles were determined in brain tissues of MV-infectedNSE-CD46+ and NSE-CD46+/STAT1-KO mice by flow cytom-etry and immunohistochemistry. At 7 dpi, the profile of immunecells (including both innate and adaptive cell types) in the spleensof NSE-CD46+, NSE-CD46+/IFN-g KO, and NSE-CD46+/STAT1KO mice was statistically similar (Fig. 4A, left panel). In brains,however, a lower percentage of CD4+ and CD8+ T cells was foundin the CNS of IFN-g–deficient mice despite similar overall

numbers of infiltrating cells within the CNS, confirming IFN-g asa chemoattractant (Fig. 4A, right panel). However, no differencesin either profile (Fig. 4A, right panel) or localization (Fig. 4B) ofintraparenchymal T cells were observed between MV-permissivewild-type and STAT1 KO mice.To compare relative levels of IFN-g, we quantified IFN-g RNA

in brain tissue at the peak of T cell infiltration (11 dpi) byquantitative RT-PCR (Fig. 5A) and compared MV RNA levels inthe CNS at various dpi (Fig. 5B). As expected, NSE-CD46+/IFN-g–KO mice did not produce IFN-g (data not shown), and wereunable to clear MV RNA to levels seen in NSE-CD46+ or NSE-CD46+/STAT1-KO mice (Fig. 5B). Brains of NSE-CD46+/STAT1-KO mice expressed significantly higher levels of IFN-g RNA at11 dpi than NSE-CD46+ mice, which correlated with the decreasein MV RNA seen in nonseizing NSE-CD46+/STAT1-KO brainsbetween 4 and 11 dpi (Fig. 5B). Importantly, NSE-CD46+ micehad consistently low levels of MV RNA, indicating that STAT1most likely plays a pivotal role in the type I IFN innate response atearly times postinfection. Note also that, although ∼20% of NSE-CD46+/STAT1-KO mice develop lethal seizures at 4 dpi, there wasno statistical difference in viral load between seizing and healthymice. Together, these data indicate that the adaptive immune re-sponse of NSE-CD46+/STAT1-KO mice is comparable to thatmade by wild-type mice, leading to viral resolution in the absenceof the canonical STAT1-mediated pathway.

Canonical IFN-g–dependent genes are not expressed inSTAT1-KO neurons

Finally, to address the possibility that IFN-g resulted in the ex-pression of a classic profile of genes via an alternative (non-STAT1–mediated) pathway, we used a microarray strategy toidentify the top 50 genes that were activated by IFN-g in MEFs,and then determined whether these same genes were induced inIFN-g–treated neurons, with and without STAT1. Microarrayswere performed on both primary MEFs and primary NSE-CD46+

and NSE-CD46+/STAT1 KO neurons following IFN-g exposurefor 0, 3, 6, or 24 h. Three individual cultures were analyzed foreach time point. Heat maps, reflective of expression levels, areshown in Fig. 6. Yellow represents the baseline for each gene in

Table I. Measles virus-induced CNS disease in mice of variousimmunodeficient backgrounds

Genotype

Average Illness Score (% Surviving)

Week 1 Week 2 Week 3 Week 4

NSE-CD46+ 0 (100) 0 (100) 0 (100) 0 (100)NSE-CD46+/RAG-2 KO 0.6 (100) 2.8 (30) 2.3 (20) 2.5 (10)NSE-CD46+/IFN-g KO 0 (100) 1.9 (60) 1.8 (60) 1.3 (50)NSE-CD46+/STAT1 KO 1.2 (70) 1.0 (70) 0.5 (70) 0 (70)

Mice were subjectively assessed for signs of morbidity every other day postin-fection. Average scores (0–3) of surviving mice for a 4-wk observation interval areshown (n = 10 for each genotype), as follows: 0 = healthy; 1 = ruffled fur/piloer-ection, ataxic; 2 = ruffled fur/piloerection, ataxic, plus seizures; 3 = moribund.

FIGURE 3. STAT1 is not required for T cell-mediated resolution of a

neuron-restricted MV infection. Kaplan–Meier survival plot of MV-chal-

lenged NSE-CD46+ mice on various KO backgrounds. NSE-CD46+, NSE-

CD46+/STAT1-KO, NSE-CD46+/IFN-g–KO, and NSE-CD46+/RAG2-KO

mice were infected, as described in Materials and Methods. Results from

three separate experiments were pooled (n = 10–35 mice/condition). Mice

were monitored daily for signs of morbidity and were euthanized when

signs of illness were apparent. The dotted horizontal line indicates the peak

of T cell entry into the CNS parenchyma of NSE-CD46+ mice. Pairwise

analysis, contingent on survival to 8 dpi, was performed between infected

NSE-CD46+ mice and the other three genotypes. NSE-CD46+ versus NSE-

CD46+/STAT1 KO: not significant. NSE-CD46+ versus NSE-CD46+/IFNg

KO: p , 0.05. NSE-CD46+ versus NSE-CD46+/RAG KO: p , 0.05.

The Journal of Immunology 1919

by guest on February 8, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

each cell type; progression to red indicates induction of geneexpression. The top 50 IFN-g–activated genes in MEFs are shownin the first three columns; as expected, the majority of these genesachieved maximal levels by 3 h posttreatment. Primary STAT1+

neurons induced a similar profile of genes, although some geneswere not induced at all, and most others showed delayed inductionkinetics, consistent with the delayed activation of STAT1 in neu-rons. Importantly, despite the capacity to control MVas effectivelyas STAT1+ neurons, STAT1 KO neurons did not induce the ma-jority of canonical IFN-g–driven genes. Thus, it is likely that al-ternative genes are induced in STAT1 KO neurons that mediateviral control. A complete list of genes that were induced or sup-pressed in neurons following IFN-g exposure, independent ofSTAT1 expression, is provided in Supplemental Figs. 1 and 2.

DiscussionIn this manuscript, we make three key observations, as follows: 1)distinct from other parenchymal CNS cell types, the IFN-g–trig-gered, STAT1-dependent signaling profile in neurons is delayedand restricted; 2) STAT1 is not required for IFN-g–mediatedcontrol of MV either in primary neurons or in a permissive mousemodel; and 3) alternative signaling pathways are most likely op-erative in STAT1-deficient neurons (and perhaps also in STAT1-positive neurons) to result in viral control.A number of points warrant further discussion. First, these data

add to a growing literature that shows that the cellular response toextracellular cytokines is not monolithic (15, 30, 35, 50–52).Whereas signaling pathways and the genes they regulate aregenerally maintained across distinct cell populations, subtle dif-ferences in expression levels can exert major changes in the an-tiviral response. Specifically, the delayed and muted neuronalresponse to IFN-g, in contrast to the rapid and robust signal seenin MEFs (35) and astrocytes (Fig. 1C), has a marked impact on thekinetics and magnitude of IFN-g–responsive genes expressed in

neurons. Relevant to published studies, these primary neurons are∼95% pure cultures, and, in this report, we assessed pSTAT1 levelswithin 3–24 h post–IFN-g treatment; whereas untreated neurons dohave detectable levels of unphosphorylated STAT1, and eventuallyrespond to IFN-g [as shown in other studies (53–55)], the signatureof the early neuronal response is appreciably different from thatobserved in control cells. Of note as well, whereas these data spe-cifically compare primary hippocampal neurons with primary astro-cytes, differences may also be present in neuronal subpopulationsas well. The primary neurons used in these experiments are mainlyderived from the embryonic hippocampus, although dissection ofthis substructure from day E15 to 16 embryos can be imprecise,and most likely includes some cortical tissue. Moreover, bothglutamatergic and GABAergic neurons are present in the hippo-campus at this stage of development (36); thus, these cultures aremost likely comprised of different neuronal subtypes. This may berelevant to the STAT1 activation profile seen in neurons since at24 h posttreatment, about one-half of the neurons have a detectablepSTAT1 signal, whereas the remainder appears nonresponsive (Fig.1D, lower right panel). Efforts are underway to determine whetherthere is a correlation between neuronal subtype in the CNS andIFN-g responsiveness.A second issue is the apparent dispensability of STAT1 in IFN-

g–mediated viral control. In both infected mice and infected pri-mary neurons, IFN-g plays a central role in protection; its absencefrom mice results in death of ∼50% of MV-challenged animals,and permanent neurologic impairment in the survivors. Similarly,addition of murine rIFN-g to infected neurons restricts viral load,and may afford a survival advantage to these vulnerable cells.Given the crucial role of IFN-g in survival from MV CNS disease,we were therefore surprised that, during the period of adaptiveimmune cell presence in the brain of infected mice (7–17 dpi),STAT1 played no role in mouse survival. This was in contrast toan extensive body of evidence that supports a central and pivotal

FIGURE 4. Comparable immune cell profiles in

brains of wild-type and STAT-deficient mice. NSE-

CD46+, NSE-CD46+/IFN-g–KO, and NSE-CD46+/

STAT1-KO mice were inoculated with MV IC (2 3104 PFU). A, Multicolor flow cytometry analysis of

total lymphocyte populations was performed on

spleens and brains of MV-infected mice at 7 dpi.

Each bar represents the average of the total lym-

phocyte population from each organ for each cell

type (n = 5). Error bars represent SD. Statistical

analysis was applied using the Student t test (*p ,0.05 versus NSE-CD46+; #p , 0.05 versus NSE-

CD46+/IFN-g–KO). N.D., not detected. B, Repre-

sentative, horizontal brain sections immunostained

for CD4+ T cells (top) and CD8+ T cells (bottom)

from NSE-CD46+ mice and NSE-CD46+/STAT1-KO

mice at 7 and 11 dpi. Original magnification 3100.

1920 STAT1-INDEPENDENT VIRAL CONTROL IN NEURONS

by guest on February 8, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

role for STAT1 in mouse survival following infection by manyviruses. In these papers, STAT1 KO mice demonstrated greaterviremia and mortality from both peripheral viral challenges (15,55, 56) as well as from CNS infections (55), leading to the con-clusion that STAT1 was a key feature of protection against mul-tiple infections. Yet, evidence for STAT1-independent antiviralpathways also exists for other viral infections (15, 18, 19), sug-gesting that particular viruses or target tissues may not be asreliant on STAT1-dependent pathways for viral clearance. Thedifference between our results and those that underscore the im-portance of STAT1 in CNS infections may be because the only celltype infected in NSE-CD46+ mice is CNS neurons, which mayallow neuronal STAT1-independent pathways to be effectiveagainst this limited viral tropism.Whereas STAT1 is not required during the 7- to 17-dpi window

when T cells are abundant in the CNS, there is a small, but re-producible effect of STAT1 deficiency at early stages postinfection(3–6 dpi), resulting in death of 20–30% of infected mice. Wepresume that this implies participation of STAT1 in the type I IFNresponse (mediated by IFNs a and b), in which, along with STAT-2 and IFN regulatory factor-9, it is part of the IFN-stimulated genefactor 3 complex that binds to IFN-stimulated gene promoterelements. Whereas this hypothesis awaits confirmation, one issuethat this model does not address is that neurons, in vivo, mayencounter IFNs sequentially; that is, it is likely that cellular ex-posure to type I IFNs precedes IFN-g exposure. As such, whena neuron encounters IFN-g, it may already have made a tran-scriptional response to type I IFN. Thus, an ultimate appreciation

of the role of IFNs and their subsequent signaling pathways inantiviral immunity will most likely require an integrated evalua-tion of the neuronal response to serial IFN encounters.

FIGURE 5. STAT1 is necessary for early MV control, but is not required

for MV clearance from CNS neurons. A, Brains were removed at 11 dpi

from NSE-CD46+ and NSE-CD46+/STAT1-KO mice and assessed by

quantitative RT-PCR for levels of IFN-g message (n = 8). B, MV RNA in

the brains of MV-infected mice. Animals were infected IC with MV (2 3104 PFU), and brains were harvested at 4, 11, and 40 dpi. RNA was

extracted and subjected to quantitative RT-PCR for MV-N (n = 6–10 mice

per time point). Statistical analysis was applied via unpaired t test (*p ,0.04 versus NSE-CD46+ mice on the same dpi; **p , 0.003 versus NSE-

CD46+ mice on the same dpi).

FIGURE 6. Microarray expression profiles following IFN-g stimulation.

MEFs and primary NSE-CD46+ and NSE-CD46+/STAT1 KO neurons were

incubated for the indicated times with IFN-g, followed by RNA purifica-

tion. RNAwas then subjected to microarray, as described in Materials and

Methods, and analyzed by Ingenuity software. In this heat map, yellow

represents basal expression, and shades from yellow to red indicate in-

creasing expression levels (log2 scale).

The Journal of Immunology 1921

by guest on February 8, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Finally, the uncoupling of STAT1 dependence from IFN-g de-pendence makes it likely that a different constellation of genes isinduced in neurons that influence viral control. As shown in Fig. 6,the standard IFN-g profile is not induced in STAT1 KO neurons,and thus our current efforts are aimed at identifying what roleIFN-g–inducible, STAT1-independent genes may play in viralcontrol.From these data, we propose a signaling factor density model, as

follows: in MEFs and astrocytes, the abundant cytoplasmic levelsof STAT1 facilitate rapid and complete occupation of binding siteson the intracellular domain of the IFN-g receptors, triggering aprimarily STAT1-driven response. In contrast, neurons with lowerconstitutive expression of STAT1 may allow for other signal trans-duction factors to bind and eventually trigger transcription of adistinct subset of genes, creating a unique neuronal response pro-file. Indeed, the relevance of these alternative signaling pathwaysis supported by the impairment-free survival of STAT1 KO mice,implying that the canonical STAT1 signaling pathway plays a mod-est, perhaps negligible, role in viral control in neurons.As the field of cell-specific cytokine responses progresses, one

can envision that the use of cytokines as antivirals will need to bereassessed in the context of the disease or pathogen that is targetedand the cell populations affected.

AcknowledgmentsWe gratefully acknowledge Kerry Campbell, Luis Sigal, Christine Matullo,

Kevin O’Regan, and Sarah Cavanaugh for critical input on this manuscript.

DisclosuresThe authors have no financial conflicts of interest.

References1. Johnson, R. T. 1998. Viral Infections of the Nervous System. Lippincott-Raven,

Philadelphia.2. Rall, G. F., and M. B. A. Oldstone. 1997. Viral persistence in the CNS. In In

Defense of the Brain. P. K. Peterson, and J. S. Remington, eds. Blackwell Press,Cambridge, MA, p. 273–289.

3. Patterson, C. E., D. M. Lawrence, L. A. Echols, and G. F. Rall. 2002. Immune-mediated protection from measles virus-induced central nervous system diseaseis noncytolytic and gamma interferon dependent. J. Virol. 76: 4497–4506.

4. Burdeinick-Kerr, R., and D. E. Griffin. 2005. Gamma interferon-dependent,noncytolytic clearance of sindbis virus infection from neurons in vitro. J.Virol. 79: 5374–5385.

5. Komatsu, T., Z. Bi, and C. S. Reiss. 1996. Interferon-gamma induced type Initric oxide synthase activity inhibits viral replication in neurons. J. Neuro-immunol. 68: 101–108.

6. Binder, G. K., and D. E. Griffin. 2001. Interferon-gamma-mediated site-specificclearance of alphavirus from CNS neurons. Science 293: 303–306.

7. Appler, K. K., A. N. Brown, B. S. Stewart, M. J. Behr, V. L. Demarest,S. J. Wong, and K. A. Bernard. 2010. Persistence of West Nile virus in thecentral nervous system and periphery of mice. PLoS One 5: e10649.

8. Stewart, B. S., V. L. Demarest, S. J. Wong, S. Green, and K. A. Bernard. 2011.Persistence of virus-specific immune responses in the central nervous system ofmice after West Nile virus infection. BMC Immunol. 12: 6.

9. Barton, L. L., and N. J. Hyndman. 2000. Lymphocytic choriomeningitis virus:reemerging central nervous system pathogen. Pediatrics 105: E35.

10. Marfin, A. A., D. M. Bleed, J. P. Lofgren, A. C. Olin, H. M. Savage, G. C. Smith,P. S. Moore, N. Karabatsos, and T. F. Tsai. 1993. Epidemiologic aspects of a St.Louis encephalitis epidemic in Jefferson County Arkansas, 1991. Am. J. Trop.Med. Hyg. 49: 30–37.

11. Stark, G. R., I. M. Kerr, B. R. G. Williams, R. H. Silverman, and R. D. Schreiber.1998. How cells respond to interferons. Annu. Rev. Biochem. 67: 227–264.

12. Kaur, S., A. Sassano, A. M. Joseph, B. Majchrzak-Kita, E. A. Eklund, A. Verma,S. M. Brachmann, E. N. Fish, and L. C. Platanias. 2008. Dual regulatory roles ofphosphatidylinositol 3-kinase in IFN signaling. J. Immunol. 181: 7316–7323.

13. Ramana, C. V., M. P. Gil, Y. Han, R. M. Ransohoff, R. D. Schreiber, andG. R. Stark. 2001. Stat1-independent regulation of gene expression in responseto IFN-gamma. Proc. Natl. Acad. Sci. USA 98: 6674–6679.

14. Soler, C., A. Felipe, J. Garcıa-Manteiga, M. Serra, E. Guillen-Gomez,F. J. Casado, C. MacLeod, M. Modolell, M. Pastor-Anglada, and A. Celada.2003. Interferon-gamma regulates nucleoside transport systems in macrophagesthrough signal transduction and activator of transduction factor 1 (STAT1)-dependent and -independent signalling pathways. Biochem. J. 375: 777–783.

15. Gil, M. P., E. Bohn, A. K. O’Guin, C. V. Ramana, B. Levine, G. R. Stark,H. W. Virgin, and R. D. Schreiber. 2001. Biologic consequences of Stat1-independent IFN signaling. Proc. Natl. Acad. Sci. USA 98: 6680–6685.

16. Lin, W., and Y. Lin. 2010. Interferon-g inhibits central nervous system myeli-nation through both STAT1-dependent and STAT1-independent pathways. J.Neurosci. Res. 88: 2569–2577.

17. Ramana, C. V., A. Kumar, and R. Enelow. 2005. Stat1-independent induction ofSOCS-3 by interferon-gamma is mediated by sustained activation of Stat3 inmouse embryonic fibroblasts. Biochem. Biophys. Res. Commun. 327: 727–733.

18. Shresta, S., K. L. Sharar, D. M. Prigozhin, H. M. Snider, P. R. Beatty, andE. Harris. 2005. Critical roles for both STAT1-dependent and STAT1-independent pathways in the control of primary dengue virus infection inmice. J. Immunol. 175: 3946–3954.

19. Mann, B. A., J. H. Huang, P. Li, H. C. Chang, R. B. Slee, A. O’Sullivan,M. Anita, N. Yeh, M. J. Klemsz, R. R. Brutkiewicz, et al. 2008. Vaccinia virusblocks Stat1-dependent and Stat1-independent gene expression induced by type Iand type II interferons. J. Interferon Cytokine Res. 28: 367–380.

20. Joshi, S., S. Kaur, B. Kroczynska, and L. C. Platanias. 2010. Mechanisms ofmRNA translation of interferon stimulated genes. Cytokine 52: 123–127.

21. Pearce, B. D., M. V. Hobbs, T. S. McGraw, and M. J. Buchmeier. 1994. Cytokineinduction during T-cell-mediated clearance of mouse hepatitis virus from neu-rons in vivo. J. Virol. 68: 5483–5495.

22. Geiger, K. D., D. Gurushanthaiah, E. L. Howes, G. A. Lewandowski, J. C. Reed,F. E. Bloom, and N. E. Sarvetnick. 1995. Cytokine-mediated survival from lethalherpes simplex virus infection: role of programmed neuronal death. Proc. Natl.Acad. Sci. USA 92: 3411–3415.

23. Hallensleben, W., and P. Staeheli. 1999. Inhibition of Borna disease virus mul-tiplication by interferon: cell line differences in susceptibility. Arch. Virol. 144:1209–1216.

24. Jin, Y., G. Lundkvist, L. Dons, K. Kristensson, and M. E. Rottenberg. 2004.Interferon-g mediates neuronal killing of intracellular bacteria. Scand. J.Immunol. 60: 437–448.

25. Fiette, L., C. Aubert, U. Muller, S. Huang, M. Aguet, M. Brahic, andJ. F. Bureau. 1995. Theiler’s virus infection of 129Sv mice that lack the inter-feron alpha/beta or interferon gamma receptors. J. Exp. Med. 181: 2069–2076.

26. Cantin, E., B. Tanamachi, and H. Openshaw. 1999. Role for gamma interferon incontrol of herpes simplex virus type 1 reactivation. J. Virol. 73: 3418–3423.

27. Wang, F. I., D. R. Hinton, W. Gilmore, M. D. Trousdale, and J. O. Fleming.1992. Sequential infection of glial cells by the murine hepatitis virus JHM strain(MHV-4) leads to a characteristic distribution of demyelination. Lab. Invest. 66:744–754.

28. Parra, B., D. R. Hinton, N. W. Marten, C. C. Bergmann, M. T. Lin, C. S. Yang,and S. A. Stohlman. 1999. IFN-gamma is required for viral clearance fromcentral nervous system oligodendroglia. J. Immunol. 162: 1641–1647.

29. Bergmann, C. C., T. E. Lane, and S. A. Stohlman. 2006. Coronavirus infection ofthe central nervous system: host-virus stand-off. Nat. Rev. Microbiol. 4: 121–132.

30. Zhang, S. S.-M., M. G. Liu, A. Kano, C. Zhang, X. Y. Fu, and C. J. Barnstable.2005. STAT3 activation in response to growth factors or cytokines participates inretina precursor proliferation. Exp. Eye Res. 81: 103–115.

31. Ramana, C. V., M. P. Gil, R. D. Schreiber, and G. R. Stark. 2002. Stat1-dependent and -independent pathways in IFN-gamma-dependent signaling.Trends Immunol. 23: 96–101.

32. Deb, A., S. J. Haque, T. Mogensen, R. H. Silverman, and B. R. Williams. 2001.RNA-dependent protein kinase PKR is required for activation of NF-kappa B byIFN-gamma in a STAT1-independent pathway. J. Immunol. 166: 6170–6180.

33. Rall, G. F., M. Manchester, L. R. Daniels, E. M. Callahan, A. R. Belman, andM. B. A. Oldstone. 1997. A transgenic mouse model for measles virus infectionof the brain. Proc. Natl. Acad. Sci. USA 94: 4659–4663.

34. Lawrence, D. M. P., M. M. Vaughn, A. R. Belman, J. S. Cole, and G. F. Rall.1999. Immune response-mediated protection of adult but not neonatal mice fromneuron-restricted measles virus infection and central nervous system disease. J.Virol. 73: 1795–1801.

35. Rose, R. W., A. G. Vorobyeva, J. D. Skipworth, E. Nicolas, and G. F. Rall. 2007.Altered levels of STAT1 and STAT3 influence the neuronal response to interferongamma. J. Neuroimmunol. 192: 145–156.

36. Banker, G., and K. Goslin. 1991. Culturing Nerve Cells. MIT Press, Cambridge.37. Pasick, J. M., K. Kalicharran, and S. Dales. 1994. Distribution and trafficking of

JHM coronavirus structural proteins and virions in primary neurons and theOBL-21 neuronal cell line. J. Virol. 68: 2915–2928.

38. Rall, G. F., L. Mucke, and M. B. A. Oldstone. 1995. Consequences of cytotoxicT lymphocyte interaction with major histocompatibility complex class I-expressing neurons in vivo. J. Exp. Med. 182: 1201–1212.

39. Chen, W., J. Sulcove, I. Frank, S. Jaffer, H. Ozdener, and D. L. Kolson. 2002.Development of a human neuronal cell model for human immunodeficiencyvirus (HIV)-infected macrophage-induced neurotoxicity: apoptosis induced byHIV type 1 primary isolates and evidence for involvement of the Bcl-2/Bcl-xL-sensitive intrinsic apoptosis pathway. J. Virol. 76: 9407–9419.

40. Meraz, M. A., J. M. White, K. C. Sheehan, E. A. Bach, S. J. Rodig, A. S. Dighe,D. H. Kaplan, J. K. Riley, A. C. Greenlund, D. Campbell, et al. 1996. Targeteddisruption of the Stat1 gene in mice reveals unexpected physiologic specificity inthe JAK-STAT signaling pathway. Cell 84: 431–442.

41. Makhortova, N. R., P. Askovich, C. E. Patterson, L. A. Gechman, N. P. Gerard,and G. F. Rall. 2007. Neurokinin-1 enables measles virus trans-synaptic spreadin neurons. Virology 362: 235–244.

1922 STAT1-INDEPENDENT VIRAL CONTROL IN NEURONS

by guest on February 8, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

42. Smyth, G. K. 2004. Linear models and empirical Bayes methods for assessingdifferential expression in microarray experiments. Stat. Appl. Genet. Mol. Biol.3: Article 3.

43. Gentleman, R. C., V. J. Carey, D. M. Bates, B. Bolstad, M. Dettling, S. Dudoit,B. Ellis, L. Gautier, Y. Ge, J. Gentry, et al. 2004. Bioconductor: open softwaredevelopment for computational biology and bioinformatics. Genome Biol. 5: R80.

44. Lawrence, D. M. P., C. E. Patterson, T. L. Gales, J. L. D’Orazio, M. M. Vaughn,and G. F. Rall. 2000. Measles virus spread between neurons requires cell contactbut not CD46 expression, syncytium formation, or extracellular virus production.J. Virol. 74: 1908–1918.

45. Alvarez-Buylla, A., J. M. Garcıa-Verdugo, and A. D. Tramontin. 2001. A unifiedhypothesis on the lineage of neural stem cells. Nat. Rev. Neurosci. 2: 287–293.

46. Stratton, S. C., C. H. Large, B. Cox, G. Davies, and R. M. Hagan. 2003. Effectsof lamotrigine and levetiracetam on seizure development in a rat amygdalakindling model. Epilepsy Res. 53: 95–106.

47. Barbi, J., H. M. Snider, N. Bhardwaj, C. M. Lezama-Davila, J. E. Durbin, andA. R. Satoskar. 2009. Signal transducer and activator of transcription 1 in T cellsplays an indispensable role in immunity to Leishmania major by mediating Th1cell homing to the site of infection. FASEB J. 23: 3990–3999.

48. Lee, C. K., E. Smith, R. Gimeno, R. Gertner, and D. E. Levy. 2000. STAT1affects lymphocyte survival and proliferation partially independent of its roledownstream of IFN-gamma. J. Immunol. 164: 1286–1292.

49. Xie, B., J. Zhao, M. Kitagawa, J. Durbin, J. A. Madri, J. L. Guan, and X. Y. Fu.2001. Focal adhesion kinase activates Stat1 in integrin-mediated cell migrationand adhesion. J. Biol. Chem. 276: 19512–19523.

50. van Boxel-Dezaire, A. H., and G. R. Stark. 2007. Cell type-specific signaling in

response to interferon-gamma. Curr. Top. Microbiol. Immunol. 316: 119–154.51. Koumbadinga, G. A., A. Desormeaux, A. Adam, and F. Marceau. 2010. Effect of

interferon-g on inflammatory cytokine-induced bradykinin B1 receptor expres-

sion in human vascular cells. Eur. J. Pharmacol. 647: 117–125.52. Ramana, C. V., N. Grammatikakis, M. Chernov, H. Nguyen, K. C. Goh,

B. R. Williams, and G. R. Stark. 2000. Regulation of c-myc expression by IFN-

gamma through Stat1-dependent and -independent pathways. EMBO J. 19: 263–

272.53. Mizuno, T., G. Zhang, H. Takeuchi, J. Kawanokuchi, J. Wang, Y. Sonobe, S. Jin,

N. Takada, Y. Komatsu, and A. Suzumura. 2008. Interferon-gamma directly

induces neurotoxicity through a neuron specific, calcium-permeable complex of

IFN-gamma receptor and AMPA GluR1 receptor. FASEB J. 22: 1797–1806.54. Chesler, D. A., C. Dodard, G. Y. Lee, D. E. Levy, and C. S. Reiss. 2004.

Interferon-gamma-induced inhibition of neuronal vesicular stomatitis virus in-

fection is STAT1 dependent. J. Neurovirol. 10: 57–63.55. Burdeinick-Kerr, R., D. Govindarajan, and D. E. Griffin. 2009. Noncytolytic

clearance of sindbis virus infection from neurons by gamma interferon is de-

pendent on Jak/STAT signaling. J. Virol. 83: 3429–3435.56. Bente, D. A., J. B. Alimonti, W. J. Shieh, G. Camus, U. Stroher, S. Zaki, and

S. M. Jones. 2010. Pathogenesis and immune response of Crimean-Congo

hemorrhagic fever virus in a STAT-1 knockout mouse model. J. Virol. 84:

11089–11100.

The Journal of Immunology 1923

by guest on February 8, 2018http://w

ww

.jimm

unol.org/D

ownloaded from