Characterization of an In Vitro Model of Alphavirus Infection of Immature and Mature Neurons

Terminally differentiated, mature neurons are essential cellsthat are not easily regenerated. Neurotropic viruses, such asSindbis virus (SV), cause encephalomyelitis through their abilityto replicate in neurons. SV causes the death of immature neurons,while mature neurons can often survive infection. The lack ofa reproducible and convenient neuronal cell culture system hashindered a detailed study of the differences in levels of virusreplication between immature and mature neurons and the molecularevents involved in virus clearance from mature neurons. We havecharacterized SV replication in immortalized CSM14.1 rat neuronalcells that can be differentiated into neurons. During differentiation,CSM14.1 cells ceased dividing, developed neuronal morphology,and expressed neuron-specific cell markers. SV infection ofundifferentiated CSM14.1 cells was efficient and resulted inhigh levels of virus replication and cell death. SV infectionof differentiated CSM14.1 cells was less efficient and resultedin the production of 10- to 100-fold less virus and cell survival.In undifferentiated cells, SV induced a rapid shutdown of cellularprotein synthesis and pE2 was efficiently processed to E2 (ratioof E2 to pE2, 2.14). In differentiated cells, the SV-inducedshutdown of cellular protein synthesis was transient and pE2was the primary form of E2 in cells (ratio of E2 to pE2, 0.0426).We conclude that age-dependent restriction of virus replicationis an intrinsic property of maturing neurons and that the CSM14.1cell line is a convenient model system for investigating theinteractions of alphaviruses with neurons at various stagesof differentiation.

INTRODUCTION

Top

Introduction

Viral encephalitis due to infection with arthropod-borne virusesis an important cause of mortality and often results in significantlong-term neurological deficits in those that survive (8, 31,46). The enveloped, message-sense RNA virus Sindbis virus (SV)is the prototype alphavirus in the family Togaviridae and isrelated to viruses known to cause encephalitis and arthritisin humans (47). Neurons are the primary target cell for SV inmice (28), and the outcome of infection is determined by bothhost and viral factors (20, 21, 37, 50, 52). An important hostdeterminant is the maturity of the neuronal population infected.Immature animals replicate virus in the central nervous system(CNS) to higher titers than mature animals and are highly susceptibleto fatal disease, dying within 3 to 4 days after infection (20,29, 37). Adult animals infected with the same strain of SV surviveand, in the absence of a virus-specific immune response, developpersistent infection (20, 34). Immunocompetent mature animalscan recover from infection and clear virus from the CNS throughnoncytolytic mechanisms involving anti-viral glycoprotein antibody(34) and gamma interferon (3). However, an understanding ofthe mechanisms underlying these virus-host interactions hasbeen hampered by the lack of a convenient in vitro system forstudy.

The availability of a reproducible and convenient in vitro systemis essential to understand the molecular mechanism(s) of age-dependentvirus-neuron interactions and noncytolytic clearance. Severalcell culture systems have been used to study aspects of SV-inducedcell death and immunity-mediated virus clearance, such as neuroblastomacells, primary cortical neuron cultures, dorsal root ganglion(DRG) neurons, and cells overexpressing the antiapoptotic proteinBcl-2 (10, 34, 41, 51, 54). However, none of these systems offersthe possibility of studying large numbers of well-characterizedcells while truly mimicking the responses of immature and matureneurons to infection.

To develop a cell culture system that models the interactionof SV with neurons in vivo, we have used CSM14.1 cells (11,57), a temperature-sensitive, immortalized rat nigral neuroncell line that can be differentiated in vitro. CSM14.1 cellswere derived by immortalization of primary rat embryonic day14 mesencephalic neural cells with a retroviral vector containingtsA58, a temperature-sensitive mutant of the simian virus 40large tumor antigen (11). We have studied undifferentiated anddifferentiated CSM14.1 cells for expression of neuron-specificcell markers and susceptibility to SV infection and have shownthat they mimic the in vivo properties of SV infection of immatureand mature neurons.

MATERIALS AND METHODS

Top

Materials and Methods

Cell culture. The CSM14.1 cell line was obtained from Dale Bredesen (BuckInstitute for Age Research, Novato, Calif.) (11, 57, 58) andwas maintained and passaged in Dulbecco's modified Eagle medium(DMEM) (Gibco, Grand Island, N.Y.) supplemented with 10% heat-inactivatedfetal bovine serum (FBS) (Gibco), 2 mM L-glutamine (Gibco),50 µg of gentamicin (Quality Biological, Inc, Gaithersburg,Md.) per ml, 100 U of penicillin/ml, and 100 µg of streptomycin(Gibco) per ml in a humidified 5% CO₂ incubator at 31°C.For cell differentiation, CSM14.1 cells were plated in six-wellplates under the permissive conditions described above. Oncethe cells reached confluence, they were washed once with phosphate-bufferedsaline, pH 7.4 (PBS), and switched to the nonpermissive cultureconditions of DMEM-1% FBS and an incubation temperature of 37to 39°C. Two different stocks of CSM14.1 cells were usedin this study. One was previously infected with SV, cured ofthe infection, and tested negative for viral RNA by reversetranscriptase PCR (CSM14.1c), while the other had not been previouslyexposed to SV (CSM14.1). Essentially all experiments were performedwith both stocks of cells and exhibited no differences. Thetypes of cells used to generate the data in the figures areindicated in the figure legend for each experiment.

Baby hamster kidney 21 (BHK-21) cells grown in DMEM-10% FBSwere used for production of virus stocks and for virus quantificationby plaque formation.

Analysis of expression of neural proteins. For Western blot analysis of neuronal cell markers, cells werewashed three times with ice-cold Tris-buffered saline (TBS;10 mM Tris HCl [pH 7.5], 150 mM NaCl) and lysed with ice-coldlysis buffer (TBS containing 1% Igepal CA-630, 0.1% SDS, and0.5% sodium deoxycholate; protease inhibitors were added immediatelybefore use). Brain and liver tissues were homogenized in lysisbuffer using a Fast Prep machine (Bio 101, Carlsbad, Calif.).After cellular debris was removed by centrifugation, equal amountsof lysate proteins were separated by sodium dodecyl sulfate-polyacrylamidegel electrophoresis (SDS-PAGE) and transferred onto a nitrocellulosemembrane (Hybond-C Extra; Amersham, Piscataway, N.J.). Antibodies(Ab) specific for tyrosine hydroxylase (TH; polyclonal; Chemicon,Temecula, Calif.), microtubule-associated protein 2 (MAP-2;monoclonal; Sigma-Aldrich, St. Louis, Mo.), glial fibrillaryacidic protein (GFAP; polyclonal; Chemicon), and actin (monoclonal;Chemicon) were used according to the manufacturers' recommendations.Levels of expression of cell markers were quantified relativeto the amount of actin by using Image Gauge software (Fuji PhotoFilm Co., Ltd.). The intensity of each band over a specifiedsurface area was quantified in arbitrary units determined bythe software. Background intensity was subtracted from the totalintensity, and the resulting difference was divided by the surfacearea to obtain specific intensity over area (in square millimeters).The quantified intensity per surface area of each cell markerwas then divided by that of actin to get the relative levelof protein expression.

For immunofluorescence staining for TH, neuron-specific nuclearprotein (NeuN; monoclonal; Chemicon)-, neurofilament 68 (NF-68;polyclonal; Chemicon)-, neuron-specific enolase (NSE; polyclonal;Chemicon)-, and GFAP-stained CSM14.1 cells were plated on glasscoverslips in six-well plates. At various times after the initiationof differentiation, cells were washed with PBS, fixed with 4%paraformaldehyde, permeabilized, and blocked with permeablization-blockingbuffer (PBS plus 5% bovine serum albumin [BSA], 0.4% TritonX-100, and 1% normal goat serum [NGS] for TH and NeuN staining;or PBS plus 1% BSA, 0.4% Triton X-100, and 4% NGS for NF-68,NSE, and GFAP staining) and stained with Ab in dilution buffer(PBS containing 5% BSA and 1% NGS for TH and NeuN staining;PBS containing 1% BSA, 0.4% Triton X-100, and 1% NGS for NF-68and GFAP staining; or PBS containing 1% BSA and 1% NGS for NSEstaining). Ab specific for TH, NeuN, NF-68, NSE, or GFAP wereused according to the manufacturers' recommendations. Cellswere stained with a secondary Ab conjugated with Alexa-fluor488 (green; Molecular Probes, Eugene, Oreg.) for TH and NeuNor with fluorescein (Pierce, Rockford, Ill.) for NF-68, NSE,and GFAP. For TH and NeuN, cellular nuclei were also stainedwith Hoechst dye (Molecular Probes). The fluorescent signalswere visualized with a fluorescence microscope (Nikon, Melville,N.Y.).

Viruses. The 633 strain of SV, which contains glutamine at E2-55, wasused (52). Recombinant SV expressing enhanced green fluorescentprotein (eGFP) as a reporter gene was constructed using 633cDNA engineered to contain a second subgenomic promoter (27).The eGFP gene was inserted into the BstEII cloning site to create633-eGFP. Full-length viral RNA was transcribed in vitro withan SP6 promoter from an XhoI-linearized plasmid and transfectedinto BHK cells with Lipofectin (Invitrogen, Carlsbad, Calif.).At 24 to 48 h after transfection, supernatant fluid containingprogeny virus was collected, assayed by plaque formation onBHK cells, and stored in aliquots at –80°C.

In vitro SV infection, detection of virus, and determination of cell viability. Cells plated in six-well plates were infected with SV 633-eGFP(multiplicity of infection [MOI] of 5, determined on BHK cells)diluted in DMEM-1% FBS. After 1 h, the cells were washed oncewith PBS and fresh medium was added. Supernatant fluids werecollected at various times after infection and analyzed forvirus production by plaque assay. For virus growth curves, theamount of virus was normalized to the viable cell number atthe time of infection and expressed as the number of log₁₀ PFUper cell. Cell viability was determined by trypan blue exclusion.

Immunofluorescence and Western blot analyses of SV structural proteins. For immunofluorescent staining of the SV E2 protein, infectedand mock-infected CSM14.1 cells grown on glass coverslips insix-well plates were washed with PBS, fixed with 4% paraformaldehyde,and permeabilized with ice-cold acetone. After several washeswith PBS, cells were stained with anti-E2 monoclonal Ab (MAb)G5 (54) for 30 min on ice, washed six times with PBS, and stainedwith a secondary Ab conjugated with Alexa-fluor 594 (red; MolecularProbes) for 30 min on ice. Nuclei were then stained with Hoechstdye and visualized as before.

For Western blot analysis of SV structural proteins, cells werewashed three times with ice-cold PBS and lysed with ice-cold1% (wt/vol) CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate}lysis buffer (CHAPS in immunoprecipitation [IP] buffer [PBScontaining 5 mM EDTA, 5 mM EGTA, 1 mM Na₃VO₄, 10 mM Na pyrophosphate,and 50 mM NaF]) as described above. Equal amounts of proteinper sample were separated by SDS-PAGE and analyzed by Westernblotting with a rabbit polyclonal Ab specific for SV (28) orthe anti-E2 MAb, G5 (54), and the Supersignal picogram or femtogramquantity detection system (Pierce).

Metabolic labeling and analysis of intracellular proteins. At various times after infection, SV-infected and mock-infectedcells were washed twice with PBS, incubated in Met- and Cys-freemedium for 1 h, and labeled with Tran³⁵S-label (100 µCi/ml;ICN Biomedicals, Irvine, Calif.) for 1 h. Labeled cells werewashed three times with ice-cold PBS and lysed with RIPA buffer(10 mM Tris [pH 7.0] with 1% Igepal-630, 0.1% SDS, 0.1% sodiumdeoxycholate, 150 mM NaCl, and 1 mM EDTA; protease inhibitorswere added immediately before use). Cell lysates were placedon ice for 30 min to allow complete lysis and then centrifugedat 9,000 x g for 5 min at 4°C to remove insoluble debris.The ³⁵S incorporated into protein was quantified by precipitationwith 20% trichloroacetic acid, and equal counts per minute perlysate sample were used for protein analysis by SDS-PAGE andautoradiography.

To analyze the change in protein expression in CSM14.1 cellsover time, cells were labeled and washed as described aboveand lysed with ice-cold 1% (wt/vol) CHAPS in IP buffer (proteaseinhibitors were added immediately before use). After completelysis and removal of insoluble cellular debris, we immunoprecipitatedSV proteins from samples to which equal counts per minute perlysate were added using a rabbit polyclonal Ab specific forSV at 4°C overnight. Immunocomplexes were isolated usingprotein G immobilized to agarose beads (Pierce), washed twicewith 1% CHAPS-IP buffer and once with IP buffer, and analyzedby SDS-PAGE and autoradiography.

Statistics. Differences between geometric means of virus replication levelsof undifferentiated and differentiated CSM14.1 cells or betweenthe levels of viability of infected and mock-infected cellswere determined by Student's t test using the RAqua 1.8.1 software(Free Software Foundation, Boston, Mass.).

RESULTS

Top

Results

The immortalized cell line CSM14.1 can differentiate into neurons. CSM14.1 cells in medium with 10% FBS actively proliferated at31°C and then stopped proliferating and differentiated whenthey were shifted to medium with reduced serum (1%) at the nonpermissivetemperature (37 to 39°C) (Fig. 1). During the process ofdifferentiation, a substantial number of cells died. DifferentiatedCSM14.1 cells displayed a neuronal morphology, with large cellbodies and extensive process formation. To determine whetherthis change in morphology correlated with the expression ofneuronal markers, lysates of cells at various times after differentiationwere assessed for the expression of TH, MAP-2, and GFAP by Westernblotting (Fig. 2). Compared to undifferentiated cells, differentiatedCSM14.1 cells expressed increased levels of two neuron-specificcell markers, TH and MAP-2, and reduced levels of an astrocytemarker, GFAP (Fig. 2A and B). Immunofluorescent staining demonstratedthat the majority of the differentiated cells expressed TH,NeuN, NF-68, and NSE (Fig. 2C). For subsequent experiments,cells were differentiated for 3 to 4 weeks.

View larger version (97K):
[in this window]
[in a new window]
FIG. 1. Neural cell line model CSM14.1. CSM14.1 cells were cultured under permissive, proliferating conditions in DMEM-10% FBS at 31°C or differentiated for various periods of time by shifting cells to DMEM-1% FBS and 37 to 39°C. Differentiated CSM14.1 cells gradually assumed a neuronal morphology.

View larger version (78K):
[in this window]
[in a new window]
FIG. 2. Expression of neuronal markers in undifferentiated and differentiated CSM14.1 cells. With differentiation, CSM14.1 cells expressed higher levels of TH, MAP-2, NeuN, NF-68, and NSE and less GFAP than undifferentiated cells. (A) Cell and tissue lysates were analyzed by SDS-PAGE and Western blotting using Ab specific for TH, MAP-2, or GFAP. (B) The level of expression of each cell marker was quantified relative to the amount of actin on individual blots. (C) Monolayers of undifferentiated (proliferating) and differentiated (nonproliferating) cells were analyzed by immunofluorescent staining for expression of TH, NeuN, NF-68, NSE, and GFAP (green). For TH and NeuN staining, nuclei were stained with Hochest dye (blue). Act, actin; AU, arbitrary units; BG, background; 2° Ab, secondary Ab.

SV replication and CPE of undifferentiated and differentiated CSM14.1 cells. After establishing the neuronal characteristics of differentiatedCSM14.1 cells, susceptibility to SV infection and virus-inducedcell death were tested. The production of new infectious viruswas detected in cultures of undifferentiated cells at 7 h andreached peaks of approximately 10^10.5 PFU/ml and 10^4.5 PFU/cellby 24 h after infection (Fig. 3A). Undifferentiated cells developedcytopathic effects (CPE), including cell rounding, detachment,and apoptotic blebs, and essentially all were dead by 4 daysafter infection (Fig. 4A). In contrast, the production of newvirus was first detected in cultures of differentiated cellsat 11 h after infection, and virus production continued to increasethrough 72 h (Fig. 3A). Amounts of virus produced were substantiallylower in differentiated cells than in undifferentiated cells.Peak virus production in undifferentiated cells was 10^7.2 PFU/mland 10^3.1 PFU/cell. Differentiated cells displayed little tono CPE and survived for more than 6 days after infection (Fig.4B).

View larger version (16K):
[in this window]
[in a new window]
FIG. 3. SV replication in undifferentiated and differentiated CSM14.1 cells. (A) Growth curves of SV in undifferentiated and differentiated (3.5 weeks) CSM14.1c cells. (B) Growth curves of SV in undifferentiated CSM14.1 cells cultured under permissive (31°C, 10% FBS) and nonpermissive (39°C, 1%) conditions. Cells were infected by SV 633-eGFP (BHK MOI of 5). Virus titers are expressed as log₁₀ numbers of PFU per cell ± standard errors of the means. P values were determined by Student's t test. *, P < 0.05; **, P < 0.01; ***, P < 0.0001.

View larger version (13K):
[in this window]
[in a new window]
FIG. 4. Viability of undifferentiated and differentiated CSM14.1 cells after infection with SV. The levels of viability of undifferentiated (A) and differentiated (B) CSM14.1c cells infected with SV 633-eGFP (BHK MOI, 5) or mock infected were determined by trypan blue exclusion at various times after infection. P values were determined by Student's t test. *, P < 0.05; **; P < 0.01; ***, P < 0.001.

To determine whether these differences in levels of virus replicationwere due to the culture conditions or to the differentiationstate of the cells, undifferentiated CSM14.1 cells were infectedand cultured at 31°C with 10% FBS or at 39°C with 1%FBS and virus production was assessed over 24 h (Fig. 3B). Thelevels of virus replication were the same, indicating that highertemperature and lower FBS concentration were not the reasonsfor restricted virus replication in differentiated cells.

Viral and cellular protein synthesis in undifferentiated and differentiated CSM14.1 cells. Immunofluorescent staining was used to identify virus-infectedcells. At 24 h after infection, all undifferentiated cells expresseddetectable levels of the E2 glycoprotein, while approximately50% of differentiated cells could be identified to be infectedby this criterion (Fig. 5A). Costaining with Ab to E2 and THat this time indicated that more than 95% of the infected cellsalso expressed TH (data not shown). E2 antigen was detectedin all differentiated cells by day 3 after infection. Similarresults were obtained when cells were assessed using a polyclonalAb to SV (data not shown).

View larger version (50K):
[in this window]
[in a new window]
FIG. 5. Expression of viral structural proteins in SV-infected undifferentiated and differentiated CSM14.1 cells. (A) Infected and mock-infected CSM14.1c cells were analyzed for expression of viral structural proteins by immunofluorescent staining using an anti-E2 MAb, G5 (red). Nuclei were stained with Hoechst stain (blue). Total (B) or E2 (C) viral structural protein expression in infected and mock-infected CSM14.1c cells postinfection was analyzed by Western blotting using an anti-SV polyclonal Ab ( ${alpha}$ -SV) and a picogram quantity detection system. The right blot in panel B shows the results of an analysis of viral protein expression in differentiated cells with a femtogram quantity detection system. Undiff, undifferentiated cells; Diff, differentiated cells; d1 p.i., day 1 postinfection; Uninf Br, uninfected brain.

Examination by Western blotting of SV structural proteins ininfected cells showed detectable levels of capsid, E1/E2, andthe precursor E2 protein, pE2, in undifferentiated cells by24 h after infection (Fig. 5B) and in differentiated cells byday 2 after infection. There was a lower level of expressionof E2 relative to that of pE2 in differentiated cells (E2/pE2ratio, 0.0426), compared to that in undifferentiated cells (E2/pE2ratio, 2.14) (Fig. 5C).

SV infection is characteristically associated with a shutdownof host protein synthesis (47). To examine overall protein synthesis,BHK, undifferentiated CSM14.1, and differentiated CSM14.1 cellswere infected with SV, and newly synthesized proteins were labeledwith ³⁵S for 1 h at various times after infection and analyzedby SDS-PAGE and autoradiography (Fig. 6). Mock-infected cellssynthesized many cellular proteins. By 24 h after infection,the synthesis of SV proteins was predominant in BHK cells andin undifferentiated CSM14.1 cells (Fig. 6A). Viral structuralproteins, E1/E2, pE2, and capsid, were readily identified ininfected BHK cells and in undifferentiated CSM14.1 cells within24 h after infection. On the other hand, viral proteins werenot clearly detected in the differentiated CSM14.1 cells untilday 2 and were maximal on day 3 after infection. DifferentiatedCSM14.1 cells survived SV infection and gradually restored cellularprotein synthesis (Fig. 6A). No newly synthesized viral proteinswere detected in differentiated cells by immunoprecipitationon day 6 after infection (Fig. 6B).

View larger version (68K):
[in this window]
[in a new window]
FIG. 6. Intracellular protein synthesis in SV-infected undifferentiated and differentiated CSM14.1 cells. On various days after infection, SV 633-eGFP-infected and mock-infected cells were deprived of Met and Cys and labeled with Tran³⁵S-label for 1 h. (A) Differentiated (Diff) CSM14.1c cells restored cellular protein synthesis, while undifferentiated (Undiff) CSM14.1c and BHK cells did not. Cell lysates were analyzed by SDS-PAGE and autoradiography. (B) Changes in levels of SV protein synthesis in undifferentiated and differentiated CSM14.1c cells over time. Radioactively labeled proteins were immunoprecipitated with an anti-SV polyclonal Ab and analyzed by SDS-PAGE and autoradiography. d1 p.i., day 1 postinfection.

DISCUSSION

Top

Discussion

An important host factor determining the outcome of alphavirusinfection of neurons is age (21, 29), but there has not beena convenient in vitro model for characterizing this aspect ofinfection. These studies have shown that both undifferentiatedand differentiated CSM14.1 neuronal cells could be infectedwith SV and that the levels of virus replication and CPE weredependent on the maturity of the cells. Undifferentiated CSM14.1cells were efficiently infected, produced large amounts of virus,and died within 3 to 4 days. Differentiated CSM14.1 cells wereless efficiently infected, produced virus more slowly, had 10-to 100-fold-lower peak virus production, and survived for manydays. E2 glycoprotein processing was more efficient in undifferentiatedthan in differentiated cells. Virus infection led to a shutdownof cellular protein synthesis in both undifferentiated and differentiatedCSM14.1 cells, but differentiated cells spontaneously restoredhost protein synthesis. Restoration of cellular protein synthesiswas accompanied by a reduction in viral protein synthesis anda decreased production of infectious virus. These characteristicsof virus replication and neuronal survival are typical of theresponses to SV infection in immature and mature mice (18, 29).

Several experimental approaches have been used to study thesurvival of neurons after SV infection and the response of thesecells to immune mediators. The AT3 rat prostatic adenocarcinomacell line has been transfected with the cellular oncogene Bcl-2,an antiapoptotic protein. AT3-Bcl-2 cells are more resistantthan AT3 cells to virus-induced cell death (35), exhibit Ab-mediateddown-regulation of production of infectious virus, but die 2to 3 days after infection (9, 10). Thus, AT3-Bcl-2 cells havethe advantage of a continuous cell line but do not reflect thesurvival of mature neurons in vivo. Primary neuronal cells,such as rodent cortical and DRG neurons, have also been studied.Cortical neurons and undifferentiated DRG neurons collectedfrom embryonic mice or rats are highly susceptible to virus-inducedcell death (34, 41). DRG neurons differentiated for 2 to 3 weeksin culture remain alive for several days after infection andhave been used successfully to study Ab-mediated clearance ofvirus (34, 54). However, the preparation of DRG neurons is labor-intensiveand the number of cells that can be obtained is limited.

CSM14.1 cells, which can be propagated without losing theirability to differentiate, have been used previously to studyneuronal differentiation (23, 30) and the effects of variousfactors on neurons (12, 13, 38, 57, 58) but have not been usedto study responses to virus infection. Consistent with previousstudies (11, 23), differentiating CSM14.1 cells sequentiallyexpressed general and dopaminergic neuron-specific markers ina manner similar to that observed in developing rodent brain(23). In contrast to what was found in one previous study usingan MAb which detected no GFAP in CSM14.1 cells (23), we foundthat GFAP was detectable in undifferentiated cells by usinga polyclonal Ab and was down-regulated with differentiation.GFAP, an intermediate filament protein, is commonly used asan astrocyte marker but has been found in many nonglial populationsof cells, such as hepatic stellate cells (45), chondrocytesof epiglottic cartilage (4), and fibroblasts (24). The inverselyproportional relationship between GFAP expression and the timeof CSM14.1 cell differentiation is consistent with the conclusionthat CSM14.1 cells assume a more neuronal phenotype as theydifferentiate.

Undifferentiated CSM14.1 cells were very susceptible to SV infectionand produced new virus within 8 h. With differentiation, CSM14.1cells became less susceptible to SV infection and produced newinfectious virus more slowly than undifferentiated cells. Thesedifferences were not due to the culture conditions because undifferentiatedcells infected under the culture conditions for differentiatedcells (39°C and 1% FBS) replicated virus efficiently.

Immunocytochemical staining suggested that all undifferentiatedcells, but not all differentiated cells, were infected initially.In cultures of differentiated cells, the infection spread gradually,with all cells being infected within 3 days. SV structural proteinswere present in neuronal processes that contacted neighboringcells, possibly facilitating the spread of virus from cell tocell rather than the spread by release of infectious virionsinto the media. The delay in the appearance of viral structuralproteins detectable by Western blotting in differentiated CSM14.1cells may reflect the initial infection of fewer cells. However,peak virus production on a per-cell basis was also lower indifferentiated than in undifferentiated cells. The level offully processed E2 relative to that of pE2 was lower in differentiatedthan in undifferentiated cells. Less-efficient posttranslationalprocessing of pE2 into E2 may contribute to the reduced andslower production of SV in differentiated cells. Furthermore,if pE2 is incorporated into virions, production of noninfectiousparticles may also delay the spread of infection in differentiatedCSM14.1 cell cultures.

Shutdown of host protein synthesis was evident by 24 h afterinfection in undifferentiated CSM14.1 cells, and these cellseventually died. In differentiated cells, the ratio of viralto host proteins synthesized gradually increased, peaked byday 3, and decreased thereafter, suggesting that differentiatedneurons controlled virus replication and restored cellular proteinsynthesis. Cell survival and the reappearance of cellular proteinsynthesis in differentiated CSM14.1 cells after infection indicatedan ability of these cells to control virus replication. Controlof virus replication has also been described for interferon-producingcells infected with a strain of SV with mutated nonstructuralP2 (16), suggesting a potential role for interferon in thisprocess.

This pattern of virus growth and the association between neuronalmaturation and reduced susceptibility to virus infection resemblesthe time course of viral replication and outcome of SV infectionin mice (18, 21, 29). Increasing age correlates with reducedvirus production and reduced virus-induced mortality and isnot attributable to maturation of the immune response (18, 29).In newborn mice, peak virus titers are reached within 1 to 2days after infection and neuronal apoptosis is observed (36).In weanling mice, peak virus titers are reached 3 to 4 daysafter infection, cell-to-cell spread is observed, and thereis minimal apoptosis (17, 50). Furthermore, in the absence ofan antiviral immune response, SV-infected neurons of weanlingmice survive, partially control virus replication, and continueto produce virus for many weeks after infection (3, 34).

Age-dependent resistance to virus replication is also a characteristicof a number of other neurotropic viruses, including SemlikiForest virus (SFV) (14, 15, 43), Japanese encephalitis virus(JEV) (32, 42, 56), mouse hepatitis virus (48), measles virus(19, 22), reovirus (39), blue-tongue virus (5), parvoviruses(6, 7), influenza virus (26), and herpesviruses (1, 2, 40, 55).As with the AR339 strain of SV, infection of neonatal mice withthe A7(74) strain of SFV leads to fatal encephalitis, but micethat are at least 2 weeks of age recover uneventfully (44).A7(74) replicates productively, spreads efficiently, and leadsto the death of neurons during axonogenesis and synaptogenesis,whereas replication is restricted in mature neurons (14, 15,43). Type 3 reovirus, induces an acute necrotizing and lethalencephalitis in newborn mice but a nonfatal infection in olderanimals (39). Age-dependent resistance of mice to infectionwith SFV and reovirus is not explained by developmental changesin the adaptive immune response but rather by an intrinsic abilityof mature neurons to restrict viral replication (15, 18, 49).Similarly, JEV preferentially infects immature neurons, andanimals older than 2 weeks survive infection (42, 56).

The susceptibility of immature neurons to JEV may be relatedto a decreased expression of neuronal surface molecules actingas viral receptors as the neurons mature (32), and changes inthe abundance of neuronal receptors may also play a role inage-dependent susceptibility to SV (53), but even after infection,less virus is produced by mature than by immature neurons, indicatingan additional postentry restriction of replication. In additionto receptor expression, potentially important age-dependentdeterminants of resistance that have been identified includedecreased expression of proapoptotic molecules and inflammatoryresponse genes and increased expression of the neuroprotectivechemokine fractalkine and of interferon-inducible genes (25,33, 36).

CSM14.1 cells demonstrate development of resistance as a functionof neuronal maturation and provide a system in which the cellularbasis for the neuronal restriction of SV can be studied. Thisin vitro model system should allow further studies of the molecularmechanisms that determine the age-dependent interactions ofneurons with SV and the mechanisms of host control of virusreplication in neurons.

ACKNOWLEDGMENTS

These studies were supported by research grants RO1 NS18596and NS38932 (D.E.G.) and medical scientist training programgrant T32 GMO7309 (P.S.V.) from the National Institutes of Health.

We appreciate the helpful suggestions of George Oyler, IrinaShatz, and Jeffrey Rothstein; assistance from Marcia Lyons;and support from the Cellular and Molecular Medicine GraduateTraining Program at the Johns Hopkins University School of Medicine.

FOOTNOTES

* Corresponding author. Mailing address: W. Harry Feinstone Department of Molecular Microbiology and Immunology, Johns Hopkins Bloomberg School of Public Health, 615 North Wolfe St., Baltimore, MD 21205. Phone: (410) 955-3459. Fax: (410) 955-0105. E-mail:dgriffin{at}jhsph.edu.

REFERENCES

Top