key: cord-258286-lodjcj8c
authors: Zhang, Xuming; Hinton, David R.; Cua, Daniel J.; Stohlman, Stephen A.; Lai, Michael M.C.
title: Expression of Interferon-γ by a Coronavirus Defective-Interfering RNA Vector and Its Effect on Viral Replication, Spread, and Pathogenicity
date: 1997-07-07
journal: Virology
DOI: 10.1006/viro.1997.8598
sha: 
doc_id: 258286
cord_uid: lodjcj8c

Abstract A defective-interfering (DI) RNA of the murine coronavirus mouse hepatitis virus (MHV) was developed as a vector for expressing interferon-γ (IFN-γ). The murine IFN-γ gene was cloned into the DI vector under the control of an MHV transcriptional promoter and transfected into MHV-infected cells. IFN-γ was secreted into culture medium as early as 6 hr posttransfection and reached a peak level (up to 180 U/ml) at 12 hr posttransfection. The DI-expressed IFN-γ (DE-IFN-γ) exhibited an antiviral activity comparable to that of recombinant IFN-γ and was blocked by a neutralizing monoclonal antibody against IFN-γ. Treatment of macrophages with DE-IFN-γ selectively induced the expression of the cellular inducible nitric oxide synthase and the IFN-γ-inducing factor (IGIF) but did not affect the amounts of the MHV receptor mRNA. Antiviral activity was detected only when cells were pretreated with IFN-γ for 24 hr prior to infection; no inhibition of virus replication was detected when cells were treated with IFN-γ during or after infection. Furthermore, addition of IFN-γ together with MHV did not prevent infection, but appeared to prevent subsequent viral spread. MHV variants with different degrees of neurovirulence in mice had correspondingly different levels of sensitivities to IFN-γ treatmentin vitro,with the most virulent strain being most resistant to IFN-γ treatment. Infection of susceptible mice with DE-IFN-γ-containing virus caused significantly milder disease, accompanied by more pronounced mononuclear cell infiltrates into the CNS and less virus replication, than that caused by virus containing a control DI vector. This study thus demonstrates the feasibility and usefulness of this MHV DI vector for expressing cytokines and may provide a model for studying the role of cytokines in MHV pathogenesis.

1992). Resistance to IFN-g may lead to incomplete viral clearance and contribute to the establishment of persis-Interferon-g (IFN-g) is a pleiotropic cytokine produced tent infection (Moskophidis et al., 1994) . By contrast, IFNby activated CD4 / and CD8 / T cells and natural killer g is also involved in inflammatory processes. IFN-g incells (Trinchieri and Perussia, 1985; Pestka and Langer, duces the expression of many other inflammatory cyto-1987; Ijzermans and Marquet, 1989) , which exerts both kines, such as interleukin-1 (IL-1) and tumor necrosis antiviral and immunomodulatory effects. These include factor (TNF), and acts synergistically with these cytokines the activation of mononuclear phagocytes, enhancement (Wong and Goeddel, 1986) . The multitude of immunoof the generation of oxygen-free radicals, modulation of modulatory effects of IFN-g makes it a particularly interclass I and II major histocompatability complex (MHC) esting cytokine for studying viral pathogenesis. In the antigen expression, and promotion of differentiation of central nervous system (CNS), no cells constitutively exboth T and B cells (for reviews, see references by Pestka press IFN-g. During encephalomyelitis, for example as and Langer, 1987; Benveniste, 1992) . It plays an ima result of mouse hepatitis virus (MHV) infection, actiportant role in the early phase of many viral infections vated NK cells and T cells which pass through the blood- (Wheelock, 1965; Wong and Goeddel, 1986; Leist et al., brain barrier into the CNS express IFN-g (Bukowski et 1989; Klavinskis et al., 1989; Feducchi and Carrasco, al., 1983; Pearce et al., 1994) . In addition to its effects on 1991; Ramsey et al., 1993; Heise and Virgin IV, 1995;  mononuclear cells, IFN-g acts upon cells of the CNS, Rodriguez et al., 1995) , inhibiting the replication of a varisuch as astrocytes, microglia, and macrophages (Benety of viruses prior to activation of antiviral effector cytoveniste, 1992) . toxic T lymphocyte (CTL) or antibodies. Because of its MHV, a murine coronavirus, causes a variety of disantiviral activity, IFN-g has been implicated in virus cleareases in rodents, such as hepatitis, enteritis, and neuroance and resolution of viral infection (Ramshaw et al., logical diseases, depending on the viral strain (Cheever et al., 1949; Gledhill and Niven, 1955; Ishida et al., 1978) .

lination (Stohlman et al., 1982; Lai and Stohlman, 1992) . may allow studies of the interaction between MHV and the host's immune system by expressing immunoregula-The DL variant derived from the parental JHMV causes an acute, fulminant, necrotizing encephalomyelitis with tory proteins at the foci of viral infection. minimal or no demyelination. By contrast, the neuroattenuated variant 2.2-V-1 derived from DL produces a nonfa-MATERIALS AND METHODS tal encephalomyelitis with extensive demyelination Virus and cells (Fleming et al., 1986 (Fleming et al., , 1987 Wang et al., 1992) . Disease outcome also depends on the genetic background, the

The following virus strains were used in this study: the developmental stage, and the immunological status of neuropathogenic MHV strain JHM isolate (DL), which is the host. Previous studies have shown that immunocoma large plaque variant derived from the parental JHM petent mice infected with MHV exhibited increased exstrain (Stohlman et al., 1982) ; the small plaque variant pression of a number of cytokines, including IL-1, IL-6, DS (Stohlman et al., 1982) ; the neutralization-escape mu-TNF-a, and IFN-g, in the CNS at the time of viral cleartant 2.2-V-1 (Fleming et al., 1987; Wang et al., 1992) , and ance (Pearce et al., 1994) . However, the role of these strain A59, which is both neurotropic and hepatotropic. cytokines in MHV pathogenesis is not fully understood.

The murine astrocytoma cell line (DBT) (Hirano et al., For example, it has been suggested that IFN-g may not 1974) and J774.1 macrophage cell line (obtained from be necessary for induction of the MHC class I molecules the American Type Culture Collection) were used for in on neural cells in vivo (Pearce et al., 1994) , a prerequisite vitro experiments. DBT cells were also used for plaque to CTL-mediated clearance (Stohlman et al., 1995) . Howassay. ever, IFN-g treatment ameliorates MHV-induced disease (Smith et al., 1991) , suggesting that either the antiviral Plasmid construction role or the immunomodulatory role of IFN-g is a critical A previously constructed plasmid p25CAT (Liao and component of MHV infection. Lai, 1994) , which contains the plasmid Bluescript (Pro-MHV contains a single-strand, positive-sense RNA gemega) sequence with a CAT gene inserted behind an IG nome of 31 kb (Lee et al., 1991) . It undergoes rapid recombisequence in the DIssE cDNA (Makino et al., 1988a) , was nation, probably due to its large RNA genome and the used as the basic DI vector. For cloning the murine IFNspecial properties of its RNA-dependent RNA polymerase g gene into the DI vector, a cDNA fragment containing . Similarly, defective interfering (DI) RNAs are the complete IFN-g gene (kindly provided by Dr. J. A. frequently generated in MHV-infected cells. Recently, re-Frelinger, University of Rochester) was generated by combinant DI RNAs have been developed which can replipolymerase chain reaction (PCR) using a pair of primers. cate in the presence of a helper MHV (Makino et al., 1988a, The 5 sense primer (5-TAACTAGTAATCTAATCTAA-1991; Van der Most et al., 1991) . We have modified an MHV ACTTTAAGGAATGAACGCTACACACT-3) contains a re-DI RNA and developed an expression vector. This DI RNA striction enzyme SpeI site (underlined), the coronavirus contains both the 5-and the 3-ends, an internal region of intergenic sequence (in boldface), and the first 16 nucleothe parental MHV genome (Makino et al., 1988b) , and an tides of the IFN-g open reading frame (ORF). The 3 intergenic (IG) sequence, which is a recognition signal for antisense primer (5-TCAGAATTCAATCAGCAGCGAsubgenomic mRNA transcription, followed by an exoge-CTCCT-3) contains the last 15 nucleotides of the IFN-g nous gene. Upon transfection of this DI RNA into MHV-ORF and a restriction enzyme EcoRI site (underlined). infected cells, a subgenomic mRNA is synthesized and the After restriction enzyme digestion of the PCR products inserted gene expressed. This system has been used to with SpeI and EcoRI, a 0.5-kb cDNA fragment was puriexpress the chloramphenicol acetyltransferase (CAT) profied by low-melting-point agarose gel electrophoresis tein and the coronavirus structural protein hemagglutinin/ and directionally cloned into the SpeI and EcoRI sites of esterase (HE) in MHV-infected cells (Liao and Lai, 1994;  p25CAT, resulting in pDE-IFN-g (Fig. 1A) . The resulting Liao et al., 1995) . These proteins are expressed only in construct contains the IFN-g gene placed behind the IG infected cells during virus replication, thus providing some sequence between genes 6 and 7 (IG7) of MHV. degree of targeted gene expression. Furthermore, the expressed HE protein can be incorporated into virus particles, RNA transcription and transfection and the expression can be detected in serial virus passages (Liao et al., 1995) . Thus, this DI RNA expression Plasmid DNA (pDE-IFN-g) was linearized with XbaI, and RNA was transcribed in vitro using T7 RNA polymer-system provides an alternative to an infectious full-length cDNA clone, which is still not available, for studying the ase according to the manufacturer's recommended procedure (Promega). RNA transfection was carried out molecular biology and pathogenesis of coronaviruses.

In the present study, we have used this DI RNA system using the DOTAP method (Boehringer-Mannheim) as described previously (Zhang et al., 1994) . Briefly, mono-to express the murine IFN-g gene. The expressed IFNg exhibited antiviral activity, prevented virus spread in layers of DBT cells grown at approximately 70% confluence in 60-mm petri dishes were infected with MHV at vitro, and altered viral pathogenesis in mice. This system DE-IFN-g RNA. Following centrifugation at 4000 g for 30 min, supernatants were tested for IFN-g using a sand-cells were washed with phosphate-buffered saline (PBS) and covered with 2 ml of prewarmed Eagle's minimum wich ELISA as previously described (Cua et al., 1995) . R4-6A2 (anti-IFN-g) (American Type Culture Collection) essential medium (MEM) containing 1% newborn calf serum (Intragen). Five to ten micrograms of in vitro tran-serum-free hybridoma supernatant was used to coat 96well plates. Biotinylated XMG-1.2 (anti-IFN-g) was ob-scribed RNAs were mixed slowly with 10 ml of DOTAP (Boehringer-Mannheim) in HBS buffer (20 mM HEPES; tained from PharMingen. Avidin-peroxidase and o-phenylenediamine (OPD) were obtained from Sigma Chemical 150 mM NaCl; pH 7.4), and incubated at room temperature for 10 min. The mixture was then added to the cell Co. Recombinant IFN-g (rIFN-g) (Zymogen) was used as ELISA standard, and the concentration of IFN-g is re-culture. The final concentration of DOTAP was 5 mg/ml. ported in international units per milliliter (U/ml). Enzyme-linked immunosorbent assay (ELISA) for IFN-g MHV replication in the presence of IFN-g To quantitate expression of IFN-g, medium was collected at 4, 6, 8, 10, 12, and 24 hr posttransfection from DBT cells were seeded at a concentration of 5 1 10 5 cells per well into 24-well plates and incubated for 24 hr DBT cells infected with JHM or A59 and transfected with at 37Њ in MEM containing 5% newborn calf serum. J774.1 extension. PCR products were analyzed by agarose gel electrophoresis. cells were seeded at a concentration of 5 1 10 4 cells per well into 24-well plates and incubated for 24 hr at 37Њ in Dulbecco's modified MEM (DMEM) containing 10% Dot blot analysis fetal calf serum. Cells were treated with various concen-RT-PCR products were quantitated using the dot blot trations of the DI-expressed IFN-g (DE-IFN-g) or rIFN-g method previously described (Murphy et al., 1993 ; Cua and infected with viruses at an m.o.i. of 1, 0.1, 0. 01, or et al., 1995) . Briefly, PCR-amplified cDNA (10 ml) was 0.001. After virus adsorption for 1 hr, the respective medenatured in 90 ml of denaturing solution (0.4 N NaOH dium with or without IFN-g was added and the cells were and 25 mM EDTA) for 10 min and neutralized by the incubated for the indicated periods of time.

addition of an equal volume of 1 M Tris-HCl, pH 8.0. Samples were transferred to a nylon membrane via a Isolation and detection of intracellular mRNAs

Minifold I Dot Blot apparatus (Schleichel and Schuell), To study the effects of IFN-g treatment on the expresand the wells were washed with 51 SSC (4.38% sodium sion of cellular genes [inducible nitric oxide synthase chloride, 2.2% sodium citrate). Membranes were air-dried (iNOS), interferon-g-inducing factor (IGIF), and MHV reand the cDNA was fixed using a Stratalinker UV oven ceptor (MHVR)], macrophage cells (J774.1) were grown (Stratagene). Following prehybridization [6% 101 SSC, to 90% confluence in 60-mm petri dishes and then treated 0.5% sodium dodecyl sulfate (SDS), 0.1 mg/ml salmon with medium from cells expressing DE-IFN-g or DE-CAT,

sperm DNA] at room temperature for 30 min, 32 P-labeled both of which had been irradiated with UV to inactivate specific probes (Table 1) were added. Following hybridhelper virus. At 24 and 48 hr after treatment, cells were ization at 60Њ, the membranes were washed three times collected and intracellular RNA was isolated as dewith 21 SSC containing 0.1% SDS for 10 min, air dried, scribed previously (Zhang et al., 1994) . To determine the and scanned on an Ambis radioanalytic imaging system effects of MHV infection on the expression of cellular (Ambis Systems). Total counts of each duplicate sample genes, J774.1 cells were infected with MHV-JHM virus at for iNOS, IGIF, and MHVR at each time point were noran m.o.i. of 0.01 at 24 hr after IFN-g treatment. RNA was malized to the control HPRT. The blots were further autoisolated at 24 hr postinfection. The RNA samples were radiographed. used for synthesis of cDNAs by reverse transcription (RT) with random priming hexamers (Boehringer-Mannheim).

Mice To detect individual genes, cDNA pools were subjected to PCR amplification using gene-specific primers (Table  C57BL /6 mice were purchased at 7 weeks of age from The Jackson Laboratory. Mice were infected with 1 1 10 5 1). The gene encoding the housekeeping enzyme hypoxanthine phosphoribosyltransferase (HPRT) was used as PFU of A59 expressing DE-IFN-g or DE-CAT. Preliminary experiments showed no difference in virus replication in an internal control. The PCR was performed for 20 cycles under the following condition: 95Њ for 1 min for denatur-the CNS comparing parental A59 and A59 virus containing the DE-CAT vector. ation, 56Њ for 1 min for annealing, and 72Њ for 2 min for

Virus titers in the CNS were determined by homogenization of half of the brain in PBS followed by plaque assay on monolayers of DBT cells as previously described (Stohlman et al., 1995) . The remaining half of the brains were fixed in Clark's solution (75% ethanol, 25% glacial acetic acid), embedded in paraffin, and stained with hematoxylin and eosin to examine the extent of encephalitis or with the immunoperoxidase method (Vectastain ABC kit; Vector Laboratories, Burlingame, CA) using the anti-nucleocapsid monoclonal antibody J.3.3. (Fleming et al., 1983) to determine the percentage of Cuture medium from DBT cells infected with JHM virus and transfected virus-infected cells.

with either DE-IFN-g or DE-CAT RNA was harvested at various time points posttransfection, and virus titers were determined by plaque assays.

Expression of IFN-g using an MHV DI RNA vector cell metabolism prior to infection or it may be that interferon acts at an early stage of viral replication. The murine IFN-g gene was cloned into the MHV DI To distinguish these possibilities, the culture medium RNA vector (Liao et al., 1995) under the control of the harvested from JHM-infected and DE-IFN-g-transfected MHV IG7 sequence. The resulting RNA, DE-IFN-g RNA, cells late in infection was used to infect DBT cells. This was transfected into MHV-infected cells, and the producmedium contained not only JHM virus but also IFN-g tion of IFN-g in the culture medium was detected by (180 U/ml) (Fig. 1) . Therefore, IFN-g was present through-ELISA. As shown in Fig. 1B , when MHV-JHM was used out the infection, beginning with the initiation of viral as helper virus, IFN-g was secreted into the medium (20 infection. No significant differences in virus titer released U/ml) as early as 6 hr posttransfection and increased from the DE-IFN-g-and DE-CAT-infected cells were dewith time. At 24 hr posttransfection, when cell monotected (both yielded approximately 10 6 PFU/ml) (data not layers were completely lysed, the amount of IFN-g shown). Thus, IFN-g has little antiviral effect even when reached approximately 180 U/ml. When A59 was used present at the initiation of viral infection. as helper virus, the production of IFN-g was detected at

In view of the known mechanisms of action of IFN-a 80 U/ml at 6 hr posttransfection and reached a maximum and -b, whose antiviral activities require preadsorption (approximately 180 U/ml) earlier (at 12 hr posttransfecto cells prior to viral infection (Bianzani and Autonelli, tion) (Fig. 1C) , consistent with the observation that A59 1989), we examined the effects of pretreatment of cells replicates faster than JHM. These results indicated that with IFN-g prior to infection. For this study, the culture MHV DI vector can be used for the production of a semedium from JHM-infected and DE-IFN-g-transfected creted cytokine during MHV infection in vitro.

cells was UV-irradiated to inactivate infectious virus and then used as a source of IFN-g to pretreat DBT cells.

Twenty-four hours later, cells were infected with JHM or replication in vitro A59 virus at m.o.i.'s ranging from 0.1 to 0.001 in the continual presence of DE-IFN-g. Virus titers were deter-IFN-g exerts multiple biological functions both in vitro and in vivo (Trinchieri and Perussia, 1985; Pestka and mined at 24 hr postinfection. As shown in Fig. 3A , DE-IFNg exhibited a slight inhibitory effect on JHM replication Langer, 1987), but its effects on coronavirus infections have not been extensively examined. We first determined (approximately 1 log 10 reducation in virus titer), when an m.o.i. of 0.001 was used; similar results were obtained whether DI-expressed IFN-g had antiviral effects on helper viral replication. Virus titers in the medium of DBT with A59 virus (Fig. 3A) , suggesting that pretreatment of cells with IFN-g prior to viral infection induces an antiviral cells infected with JHM and transfected with DE-IFN-g RNA were determined at various time points after infec-state. This inhibitory effect was less pronounced when higher m.o.i.'s were used (data not shown), suggesting tion and compared to DE-CAT RNA-transfected cells. Figure 2 shows that the virus titers in the presence of DE-that the observed antiviral activity was weak and could be overcome by a higher virus titer. IFN-g were lower by approximately half a log 10 compared to cultures transfected with the DE-CAT RNA. This differ-To further establish that the antiviral effect was due to the specific effects of IFN-g, the UV-inactivated DE-IFN-ence was small but reproducible, suggesting that IFN-g exerts at most a weak antiviral effect. The absence of g preparation was preincubated for 2 hr with a hamster neutralizing monoclonal antibody specific for rIFN-g. significant anti-viral effect of IFN-g in this system could be due to the requirement for interferon to modify host Antiviral effects were completely blocked by this treat- The UV-irradiated supernatants were used either as a source of IFN-g or as a control (CAT) to pretreat cells for 24 hr, and the cells were then infected with either JHM or A59 at an m.o.i. of 0.001. After virus adsorption, cells were incubated with the same supernatants for 24 hr, and the virus titers in culture medium at 24 hr postinfection were determined by a standard plaque assay. (B) Neutralization assay of IFN-g. Both UV-irradiated supernatants (IFN-g and CAT) were incubated with 1 mg/ml of a hamster anti-IFN-g neutralizing monoclonal antibody for 2 hr at room temperature prior to being used for pretreatment of cells. Subsequent procedures were the same as in (A). ment (Fig. 3B ), demonstrating that IFN-g, but not the repli-log 10 , similar to the data obtained with DBT cells. Thus, the absence of strong antiviral effects of IFN-g is not cation of the DI vector itself, was responsible for the antiviral activity. These combined results suggest that due to nonresponsiveness of cells to IFN-g. IFN-g has a weak antiviral effect, which was evident only DI RNA-expressed IFN-g prevents virus spread when cells were pretreated with IFN-g prior to infection.

The relatively weak antiviral effects of IFN-g also could The results described above indicated that antiviral be due to the possibility that DBT cells do not respond effects of IFN-g could be demonstrated only when cells well to IFN-g. Since it is known that macrophages are were pretreated with IFN-g before viral infection and particularly sensitive to IFN-g treatment (Ijzermans and when a low m.o.i. was used. They suggested the possibil-Marquet, 1989), we further determined the inhibitory efity that IFN-g could prevent virus spread, if virus initially fects of IFN-g on MHV replication in an MHV-susceptible infects only a small number of cells. To establish an in macrophage cell line (J774.1). J774.1 cells were previtro model for studying the potential effects of IFN-g in treated with various concentrations of rIFN-g for 24 hr preventing virus spread, UV-irradiated culture medium before and throughout virus infection. As shown in Fig. from DE-IFN-g-transfected cells, which contained IFN-g 4, both A59 and JHM were inhibited by rIFN-g by 1 to 2 at 180 U/ml, was mixed with a very low titer of JHM virus at approximately one infectious particle in each well of a 24-well plate. Cells were observed for cytopathic effects daily for 4 days and the number of fusion plaques was counted. Results of these experiments are presented in Table 2 . The number of plaques increased more slowly when the DE-IFN-g was present (for example, from 1 plaque on Day 1 to 12 plaques on Day 4), as compared to those in the control wells, in which DIexpressed CAT preparation was used (i.e., from 1 plaque on Day 1 to 30 plaques on Day 2 and too numerous to count by Day 3) (Table 2) . Initially, the plaque sizes in the presence of IFN-g were indistinguishable from those of the control wells (data not shown); however, by Day 3 or 4 postinfection, while all plaques in the IFN-g-treated cultures remained of uniform size, plaques in the absence of IFN-g became numerous and heterogeneous It has been suggested that IFN-g induces a number

No. of plaques c on of cellular proteins and enzymes which either act as T cells (Okamura et al., 1995) . MHVR is a member of virus. One milliliter of each culture medium was then mixed with JHM the biliary glycoprotein (BGP)/carcinoembryonic antigen virus and added to the cell monolayers, so that an average of 1 PFU per well was present.

(CEA) family and serves as a receptor for MHV infection b Each sample was quadruplicated in 4 wells of a 24-well plate. (Williams et al., 1991) . Treatment of cells with DI-exc Plaques were counted in the liquid medium using a light micropressed IFN-g for 24 hr increased the expression of iNOS scope.

and IGIF mRNAs. MHV infection did not affect the expresd UC, uncountable due to extensive cytopathic effects and detachment of cells.

due to the rapid spread of progeny virus before IFN-g exhibited its antiviral effect (data not shown). Similar results were obtained when various concentrations of rIFNg (50, 100, and 150 U/ml) were used, suggesting that 50 U/ml rIFN-g is sufficient to prevent virus spread in vitro (data not shown).

Sensitivity of different JHM variants to IFN-g treatment in vitro was assessed in an effort to determine whether the IFN-g sensitivity correlates with the pathogenicity of the virus in vivo. Three JHM variants with different degrees of neurovirulence were used: DL (LD 50 1-5 PFU), DS (LD 50 100-200 PFU), and 2.2-V-1 (LD 50 2000-10,000 PFU) (Stohlman et al., 1982 (Stohlman et al., , 1995 Fleming et al., 1986 Fleming et al., , 1987 . DL causes little demyelination and infects predominantly neurons whereas variant 2.2-V-1 causes extensive demyelination and infects predominantly glial cells with a particular tropism for oligodendrocytes. Variant DS causes less demyelination than variant 2.2-V-1. DBT cells pretreated with IFN-g (180 U/ml) for 24 hr were infected, and the same concentrations of IFN-g were maintained throughout the infection. At 24 hr postinfection, culture medium was collected and virus titer determined by plaque assay. As shown in Fig. 6 , a reduction of approximately 2.5 log 10 in pathogenicity in vivo, groups of C57BL/6 mice were infected with 1 1 10 5 PFU of A59 virus containing either DE-IFN-g small numbers of perivascular and subarachnoid mononuor DE-CAT. Preliminary experiments showed no difference clear cells, the brains of the DE-IFN-g-expressing group in virus replication in CNS between mice infected with pashowed widespread meningomyeloencephalitis with promrental A59 virus and those infected with A59-DE-CAT (data inent perivascular cuffs, infiltration of mononuclear cells not shown). At 6 days postinfection, four mice in each group into the parenchyma, and subarachnoid infiltrates (Fig. 8) . were sacrificed and the brains were examined for MHV This result supports the immunostimulatory effects of IFNtiter and histological changes. The remaining mice in each g. Although this experiment used only a small number of group were monitored daily for survival. Table 3 shows that mice, the data suggest that expression of immunomodulathere was approximately 2.4 log 10 less virus in the CNS of tory molecules from the DI vector can alter the pathogenemice infected with A59 expressing DE-IFN-g vector comsis of MHV-induced disease. pared to the mice infected with A59 expressing DE-CAT vector. Correspondingly, all the mice infected with DE-IFNg-expressing A59 survived the entire 21-day observation 

The molecular basis for the relative IFN resistance of different MHV strains is not yet known. Previous studies This study demonstrates that the MHV DI RNA system have shown that the neutralization-escape mutant 2.2-Vcan be utilized as a vector to express the IFN-g gene 1 of JHM strain has a single nucleotide mutation at posiand that the IFN-g protein is translated and secreted tion 3340 of the S gene, which results in a leucine to from infected cells as a biologically active molecule.

phenylalanine substitution (Wang et al., 1992) . Whether These data represent the first successful attempt to exthis single mutation affects the sensitivity of the virus to press a mammalian cellular gene product using a coro-IFN-g remains unclear. In lymphocytic choriomeningitis navirus DI RNA vector. Thus far, we have demonstrated virus, resistance of various virus strains to IFN-a/b or the feasibility of this DI RNA system for expressing a IFN-g in vitro correlates with their ability to establish prokaryotic bacterial gene CAT (Liao and Lai, 1994) , a persistent infections in adult immunocompetent mice viral structural protein gene HE (Liao et al., 1995) , and (Moskophidis et al., 1994) . One possibility is that IFN the mammalian cellular gene IFN-g (this report). These resistance allows enhanced viral replication and spread, studies showed a broad range of usage of this DI RNA facilitating exhaustion of antiviral CTL, thereby resulting system for expressing various genes of interest.

in virus persistence. Whether MHV utilizes a similar Currently, an infectious, full-length cDNA clone of MHV mechanism to modulate its infection in mice is an inter-RNA is not available; therefore, it is difficult to unequivoesting issue. Correlation between IFN resistance and cally elucidate the mechanism of pathogenesis of MHV viral pathogenicity has also been documented for meaat the molecular level. The development of a DI RNA sles virus, adenovirus, and herpes simplex virus type I expression system thus provides an alternative ap- (Carrigan and Kehl-Knox, 1990; Su et al., 1990 ; Kalvakoproach, allowing the expression of both viral and cellular lanu et al., 1991) . genes to be manipulated. Further, this system allows The in vitro experiments showed that the DI-expressed expression of heterologous gene products at the site of IFN-g had inhibitory effects on virus spread from initially viral replication. This system has an advantage over the infected cells to neighboring uninfected cells. The inhibipassive administration of cytokines for studying viral tory effect was more pronouced at a lower m.o.i., which pathogenesis, since cytokines usually have a short halfapparently allowed sufficient time for IFN-g to activate life, making it difficult to maintain high local concentraan antiviral state in adjacent uninfected cells. Pretreattions at the site of infection. One drawback of the DI ment of cells (astrocytoma and macrophages) with IFNsystem, however, is its limited expression. The DI RNA g is required to induce an antiviral state (Figs. 3 and 4) , cannot be packaged beyond the fourth passage in vitro consistent with previous findings from studies of primary (data not shown). We have attempted to increase retenmouse macrophages (Lucchiari et al., 1991) and other tion of the DI RNA via incorporation of a packaging signal.

target cells (Lewis, 1982) . Expression of both iNOS and However, the expression level of the gene product was IGIF mRNA in macrophages was induced by IFN-g. Howreduced; no significant retention was found (Lin and Lai, ever, whether these molecules mediate the antiviral ef-1993). Nevertheless, our data indicated that, during the fects of IFN-g is not clear. Recently, it was demonstrated first several passages, the expression level of IFN-g was that iNOS expression did not play a significant role in such that a sufficiently high level of IFN-g can be mainthe pathogenesis of the MHV OBLV60 strain (Lane et al., tained locally at the beginning of viral infection. 1997). Nevertheless, we can conclude from our study The virulence of several MHV variants correlates with that the antiviral effects of IFN-g are not mediated by their resistance to IFN-g treatment, suggesting that IFNdown-regulation of MHVR. The precise mechanism of the g may play a role in the pathogenesis of MHV. An earlier antiviral effects of IFN-g will require additional studies, study analyzed the effects of IFN-g during JHM infection as there appears to be discordance between the antiviral using passive transfer of an anti-IFN-g-antibody (Smith effects of NO in vivo and its effects in vitro (Lane et al., et al., 1991) . This treatment significantly enhanced virus 1997). replication and resulted in a higher mortality with de-

The alteration of A59 neuropathogenesis by DE-IFN-g creased survival times. IFN-g treatment of macrophages provides further support for the significance of IFN-g from A/J mice rendered them partially resistant to MHV3 in MHV infection. Inhibition of IFN-g action by passive infection, whereas the macrophages from susceptible transfer of antibody (Smith et al., 1991) enhanced virus BALB/c mice did not respond to IFN-g, suggesting that replication and increased mortality, suggesting that local the resistance of mice to MHV3 infection involves the production of IFN-g by infiltrating leukocytes is a critical sensitivity of macrophages to IFN-g (Lucchiari et al., component of the host response to MHV infection. In 1991; Vassao et al., 1994a,b) . IFN-g was also shown to our experiments, the production of IFN-g by DE-IFN-g be more effective than IFN-a/b in inducing an antiviral resulted in an exaggeration of the host response with state in macrophages infected with MHV (Vassao et al., more prominent encephalitis, improved viral clearance, 1994a). These reports support the notion that IFN-g may and decreased mortality. The increased encephalitis may, in turn, induce local cytokine production and CTL play a role in MHV infection.

The complete 263, 1-16. sequence (22 kilobases) of murine coronavirus gene 1 encoding the

1606-sequence as an upstream cis-acting element for coronavirus sub-1615. genomic mRNA transcription

A murine virus (JHM) causing disseminated encephalomyelitive-interfering RNA as an expression vector: the generation of a tis with extensive destruction of myelin. I. Isolation and biological pseudorecombinant mouse hepatitis virus expressing hemagglutiproperties of the virus

Deletion mapping of a mouse hepatiduced Th2 responses in experimental allergic encephalomyelitis tis virus defective interfering RNA reveals the requirement of an (EAE)-resistant mice: Th2-mediated suppression of autoimmune disinternal and discontiguous sequence for replication

Acquired interferon and tumor necrosis factor exert a synergistic blockade on immunity of A/J mice to mouse hepatitis virus 3 infection: dependence the replication of herpes simplex virus

Defectiveruses: Analysis using monoclonal antibodies to JHM (MHV-4) virus. interfering particles of murine coronavirus: Mechanism of synthesis Virology

Primary structure and translation of a defective-interfering navirus JHM selected with monoclonal antibodies

Experimental demyelination induced by coronavidefective-interfering RNA results from intergenic site insertion

Isolation and characterization of two plaque morphology variants of the JHM neurotropic strain

Mouse hepatitis virus-specific cytotoxic T lymphocytes protect from lethal infection without eliminating virus from the central ingitis virus to alpha/beta interferon and to gamma interferon

Detection of in vivo expression of interleukin-10 using a semi-quantitative polymerherpes simplex virus type I strain is associated with heightened sensitivity to alpha/beta interferon

Immune interferon: A pleiotropic lymphokine with multiple effects

Cloning of a new cytokine that induces IFN-g production by dation of coronavirus defective interfering RNAs

A genetic analysis of macrophage activation and specific antibodies in relation Cytokine induction during T-cell-mediated clearance of mouse hepatitis virus from neurons in vivo

The astrocyte is a target cell in mice persistently infected with mouse hepatitis virus, strain

Interferons and their actions. Annu. of genetic heterogeneous mouse populations to mouse hepatitis virus infection

Sequence analysis of the spike protein gene of murine coronavirus variants: Study of as effector molecules in the resolution of virus infection

Expression of cytokines by recombinant vaccinia viruses: A model leukocytes by phytohemagglutinin

hepatitis virus is a member of the carcinoembryonic antigen family of glycoproteins

Tumour necrosis factor a and b inhibit virus replication and synergize with interferons

Coronavirus leader The role of gamma interferon in infection of susceptible mice with murine coronavirus, MHV-JHM

 Gledhill, A. W., and Niven, J. S. F. (1955) . Latent virus as exemplified by activity. Altogether, these data demonstrated that IFN-g mouse hepatitis virus (MHV). Vet. Rev. Annotat. 1, [82] [83] [84] [85] [86] [87] [88] [89] [90] plays a critical role at least early in A59 infection. The Haller, O. (1981) . Inborn resistance of mice to orthomyxoviruses. Curr.longer-term consequences of ever, cannot be definitively determined from this study Heise, M. T., and Virgin, IV, H. W. (1995) .