Age-dependent resistance to lethal alphavirus encephalitis in mice: analysis of gene expression in the central nervous system and identification of a novel interferon-inducible protective gene,mouse ISG12

Several different mammalian neurotropic viruses produce an age-dependent encephalitis characterized by more severe disease in younger hosts. To elucidate potential factors that contribute to age-dependent resistance to lethal viral encephalitis, we compared central nervous system (CNS) gene expression in neonatal and weanling mice that were either mock infected or infected intracerebrally with a recombinant strain, dsTE12Q, of the prototype alphavirus Sindbis virus. In 1-day-old mice, infection with dsTE12Q resulted in rapidly fatal disease associated with high CNS viral titers and extensive CNS apoptosis, whereas in 4-week-old mice, dsTE12Q infection resulted in asymptomatic infection with lower CNS virus titers and undetectable CNS apoptosis. GeneChip expression comparisons of mock-infected neonatal and weanling mouse brains revealed developmental regulation of the mRNA expression of numerous genes, including some apoptosis regulatory genes, such as the proapoptotic molecules caspase-3 and TRAF4, which are downregulated during development, and the neuroprotective chemokine, fractalkine, which is upregulated during postnatal development. In parallel with increased neurovirulence and increased viral replication, Sindbis virus infection in 1-day-old mice resulted in both a greater number of host inflammatory genes with altered expression and greater changes in levels of host inflammatory gene expression than infection in 4-week-old mice. Only one inflammatory response gene, an expressed sequence tag similar to human ISG12, increased by a greater magnitude in infected 4-week-old mouse brains than in infected 1-day-old mouse brains. Furthermore, we found that enforced neuronal ISG12 expression results in a significant delay in Sindbis virus-induced death in neonatal mice. Together, our data identify genes that are developmentally regulated in the CNS and genes that are differentially regulated in the brains of different aged mice in response to Sindbis virus infection.     

BCL-2 and BAX protect adult mice from lethal Sindbis virus infection but do not protect spinal cord motor neurons or prevent paralysis

Cellular proteins that regulate apoptotic cell death can modulate the outcome of Sindbis virus (SV) encephalitis in mice. Both endogenous and overexpressed BCL-2 and BAX proteins protect newborn mice from fatal SV infection by blocking apoptosis in infected neurons. To determine the effects of these cellular factors on the course of infection in older animals, a more neurovirulent SV vector (dsNSV) was constructed from a viral strain that causes both prominent spinal cord infection with hind-limb paralysis and death in weanling mice. This vector has allowed assessment of the effects of BCL-2 and BAX on both mortality and paralysis in these hosts. Similar to newborn hosts, weanling mice infected with dsNSV encoding BCL-2 or BAX survived better than animals infected with control viruses. This finding indicates that BCL-2 and BAX both protect neurons that mediate host survival. Neither cellular factor, however, could suppress the development of hind-limb paralysis or prevent the degeneration of motor neurons in the lumbar spinal cord. Infection of BAX knockout mice with dsNSV demonstrated that endogenous BAX also enhances the survival of animals but has no effect on paralysis. These findings for the spinal cord are consistent with earlier data showing that dying lumbar motor neurons do not exhibit an apoptotic morphology. Thus, divergent cell death pathways are activated in different target populations of neurons during neurovirulent SV infection of weanling mice.     

Protection against Fatal Sindbis Virus Encephalitis by Beclin, a Novel Bcl-2-Interacting Protein

bcl-2, the prototypic cellular antiapoptotic gene, decreases Sindbis virus replication and Sindbis virus-induced apoptosis in mouse brains, resulting in protection against lethal encephalitis. To investigate potential mechanisms by which Bcl-2 protects against central nervous system Sindbis virus infection, we performed a yeast two-hybrid screen to identify Bcl-2-interacting gene products in an adult mouse brain library. We identified a novel 60-kDa coiled-coil protein, Beclin, which we confirmed interacts with Bcl-2 in mammalian cells, using fluorescence resonance energy transfer microscopy. To examine the role of Beclin in Sindbis virus pathogenesis, we constructed recombinant Sindbis virus chimeras that express full-length human Beclin (SIN/beclin), Beclin lacking the putative Bcl-2-binding domain (SIN/beclinΔBcl-2BD), or Beclin containing a premature stop codon near the 5′ terminus (SIN/beclinstop). The survival of mice infected with SIN/beclin was significantly higher (71%) than the survival of mice infected with SIN/beclinΔBcl-2BD (9%) or SIN/beclinstop (7%) (P < 0.001). The brains of mice infected with SIN/beclin had fewer Sindbis virus RNA-positive cells, fewer apoptotic cells, and lower viral titers than the brains of mice infected with SIN/beclinΔBcl-2BD or SIN/beclinstop. These findings demonstrate that Beclin is a novel Bcl-2-interacting cellular protein that may play a role in antiviral host defense.     

Dominant inhibitory Ras delays Sindbis virus-induced apoptosis in neuronal cells.

Mature neurons are more resistant than dividing cells or differentiating neurons to Sindbis virus-induced apoptotic death. Therefore, we hypothesized that mitogenic signal transduction pathways may influence susceptibility to Sindbis virus-induced apoptosis. Since Ras, a 21-kDa GTP-binding protein, plays an important role in cellular proliferation and neuronal differentiation, we investigated the effect of an inducible dominant inhibitory Ras on Sindbis virus-induced death of a rat pheochromocytoma cell line, PC12 cells. Dexamethasone induction of dominant inhibitory Ras (Ha Ras(Asn17)) expression in transfected PC12 cell lines (MMTV-M17-21 and GSrasDN6 cells) resulted in a marked delay in Sindbis virus-induced apoptosis, compared with infected, uninduced cells. The delay in death after Sindbis virus infection in induced versus uninduced PC12 cells was not associated with differences in viral titers or viral infectivity. No delay in Sindbis virus-induced apoptosis was observed in Ha Ras(Asn17)-transfected PC12 cells if dexamethasone induction was initiated less than 12 h before Sindbis virus infection or in wild-type PC12 cells infected with a chimeric Sindbis virus construct that expresses Ha Ras(Asn17). The delay in Sindbis virus-induced apoptosis in induced Ha Ras(Asn17)-transfected PC12 cells was associated with a decrease in cellular DNA synthesis as measured by 5'-bromo-2'-deoxyuridine incorporation. Thus, in PC12 cells, inducible dominant inhibitory Ras inhibits cellular proliferation and delays Sindbis virus-induced apoptosis. These findings suggest that a Ras-dependent signaling pathway is a determinant of neuronal susceptibility to Sindbis virus-induced apoptosis.     

Inhibition of virus-induced neuronal apoptosis by Bax.

The Bax protein is widely known as a pro-apoptotic Bcl-2 family member that when overexpressed can trigger apoptosis in multiple cell types and is important for the developmental cell death of neurons. However, Bax was found here to be a potent inhibitor of neuronal cell death in mice infected with Sindbis virus. Newborn mice, which are highly susceptible to a fatal infection with neurotropic Sindbis virus, were significantly protected from neuronal apoptosis and fatal disease when infected with a recombinant Sindbis virus encoding Bax. Deletion of the N terminus of Bax, which mimics cleaved Bax, converted Bax into a pro-apoptotic factor in vivo. As mice mature during the first week after birth, they acquire resistance to a fatal Sindbis virus infection. However, Bax-deficient mice remained very sensitive to fatal disease compared with their control littermates, indicating that endogenous Bax functions as a survival factor and contributes to age-dependent resistance to Sindbis virus-induced mortality. The protective effects of Bax were reproduced in cultured hippocampal neurons but not in cultured dorsal root ganglia neurons. These findings indicate that cell-specific factors determine the anti-apoptotic versus pro-apoptotic function of Bax.     

Radiation-induced apoptosis in the neonatal and adult rat spinal cord

This study was designed to characterize radiation-induced apoptosis in the spinal cord of the neonatal and young adult rat. Spinal cords (C2-T2) of 1-, 2- and 10-week-old rats were irradiated with a single dose of 8, 18 or 22 Gy. Apoptosis was assessed histologically according to its specific morphological features or by using the TUNEL assay. Cell proliferation was assessed immunohistochemically using BrdU. Identities of cell types undergoing apoptosis were assessed using immunohistochemistry or in situ hybridization using markers for neurons, glial progenitor cells, microglia, oligodendrocytes and astrocytes. The time course of radiation-induced apoptosis in 1- or 2-week-old rat spinal cord was similar to that in the young adult rat spinal cord. A peak response was observed at about 8 h after irradiation, and the apoptosis index returned to the levels in nonirradiated spinal cords at 24 h. The neonatal rat spinal cord demonstrated increased apoptosis compared to the adult. Values for total yield of apoptosis over 24 h induced by 8 Gy in the neonatal rat spinal cord were significantly greater than that in the adult. Immunohistochemistry studies using Leu7, galactocerebroside, Rip and adenomatous polyposis coli tumor suppressor protein indicated that most apoptotic cells were cells of the oligodendroglial lineage regardless of the age of the animal. No evidence of Gfap or factor VIII-related antigen-positive apoptotic cells was observed, and there was a small number of apoptotic microglial cells (lectin-Rca1 positive) in the neonatal and adult rat spinal cord. In the neonatal but not adult rat spinal cord, about 10% of the apoptotic cells appeared to be neurons and were immunoreactive for synaptophysin. Labeling indices (LI) for BrdU in nonirradiated 1- and 2-week-old rat spinal cord were 20.0 and 16.3%, respectively, significantly greater than the LI of 1.0% in the 10-week-old rat spinal cord. At 8 h after a single dose of 8 Gy, 13.4% of the apoptotic cells were BrdU-positive in 10-week-old rat spinal cord, whereas 62.4 and 44.1% of the apoptotic cells showed BrdU incorporation in 1- and 2-week-old rat spinal cord, respectively. Regardless of the age of the animal, the apoptosis indices in BrdU-positive cells were greater than those in BrdU-negative cells. We conclude that the neonatal spinal cord demonstrates a greater level of apoptosis after exposure to ionizing radiation than the young adult spinal cord. This increase in apoptosis may be associated in part with the greater percentage of proliferating cells in the neonatal spinal cord, which demonstrate a greater level of radiation-induced apoptosis than nonproliferating cells.     

Role of viral persistence in retaining CD8(+) T cells within the central nervous system

The continued presence of virus-specific CD8(+) T cells within the central nervous system (CNS) following resolution of acute viral encephalomyelitis implicates organ-specific retention. The role of viral persistence in locally maintaining T cells was investigated by infecting mice with either a demyelinating, paralytic (V-1) or nonpathogenic (V-2) variant of a neurotropic mouse hepatitis virus, which differ in the ability to persist within the CNS. Class I tetramer technology revealed more infiltrating virus-specific CD8(+) T cells during acute V-1 compared to V-2 infection. However, both total and virus-specific CD8(+) T cells accumulated at similar peak levels in spinal cords by day 10 postinfection (p.i.). Decreasing viral RNA levels in both brains and spinal cords following initial virus clearance coincided with an overall progressive loss of both total and virus-specific CD8(+) T cells. By 9 weeks p.i., T cells had largely disappeared from brains of both infected groups, consistent with the decline of viral RNA. T cells also completely disappeared from V-2-infected spinal cords coincident with the absence of viral RNA. By contrast, a significant number of CD8(+) T cells which contained detectable viral RNA were recovered from spinal cords of V-1-infected mice. The data indicate that residual virus from a primary CNS infection is a vital component in mediating local retention of both CD8(+) and CD4(+) T cells and that once minimal thresholds of stimuli are lost, T cells within the CNS cannot survive in an autonomous fashion.     

TGF-beta 2 reduces demyelination, virus antigen expression, and macrophage recruitment in a viral model of multiple sclerosis

TGF-beta 2 is a potent immunoregulatory mediator that influences B cell, T cell, and macrophage function. To test whether this cytokine alters pathology in a model of virus-induced demyelinating disease, we treated SJL/J mice with TGF-beta 2 either before or after infection with Theiler's murine encephalomyelitis virus. Treatment continued three times weekly through day 35 postinfection. TGF-beta 2 administration resulted in significantly smaller lesions and fewer virus Ag-positive cells in the spinal cords of infected SJL/J mice. Mice treated with TGF-beta 2 had similar levels of virus-specific IgG as infected, control-treated mice. TGF-beta 2 administration significantly increased the level of non-virus-specific activated CTLs, but had no effect on virus-specific CTLs. TUNEL revealed a decrease in the number of apoptotic nuclei in the spinal cord white matter of mice treated in vivo with TGF-beta 2. Immunostaining with an Ab to F4/80 revealed that TGF-beta 2-treated mice had significantly fewer F4/80-positive cells in the white matter of the spinal cord as compared with infected control-treated mice. These data suggest that TGF-beta 2 may control virus-induced demyelination via an immunomodulatory mechanism that reduces macrophage infiltration.     

Identification of genes involved in the host response to neurovirulent alphavirus infection

Single-amino-acid mutations in Sindbis virus proteins can convert clinically silent encephalitis into uniformly lethal disease. However, little is known about the host gene response during avirulent and virulent central nervous system (CNS) infections. To identify candidate host genes that modulate alphavirus neurovirulence, we utilized GeneChip Expression analysis to compare CNS gene expression in mice infected with two strains of Sindbis virus that differ by one amino acid in the E2 envelope glycoprotein. Infection with Sindbis virus, dsTE12H (E2-55 HIS), resulted in 100% mortality in 10-day-old mice, whereas no disease was observed in mice infected with dsTE12Q (E2-55 GLN). dsTE12H, compared with dsTE12Q, replicated to higher titers in mouse brain and induced more CNS apoptosis. Infection with the neurovirulent dsTE12H strain was associated with both a greater number of host genes with increased expression and greater changes in levels of host gene expression than was infection with the nonvirulent dsTE12Q strain. In particular, dsTE12H infection resulted in greater increases in the levels of mRNAs encoding chemokines, proteins involved in antigen presentation and protein degradation, complement proteins, interferon-regulated proteins, and mitochondrial proteins. At least some of these increases may be beneficial for the host, as evidenced by the demonstration that enforced expression of the antiapoptotic mitochondrial protein peripheral benzodiazepine receptor (PBR) protects neonatal mice against lethal Sindbis virus infection. Thus, our findings identify specific host genes that may play a role in the host protective or pathologic response to neurovirulent Sindbis virus infection.     

Apoptosis in Alphavirus Encephalitis

Sindbis virus causes acute encephalitis in mice and serves as a useful model for encephalitic alphaviruses that infect humans. The outcome of infection is determined by whether infected neurons are resistant to virus-induced programmed cell death or activate their apoptotic pathway. The host immune response may also cause death of infected neurons. Determinants of neuronal apoptosis include the maturity of the neuron, the virulence of the infecting virus and the cellular immune response to infection. In many situations viral and cellular factors that decrease virus replication also decrease apoptosis. Antiviral antibody can downregulate virus replication in surviving neurons without affecting cell viability. Other innate and induced host immune responses can alter the outcome of infection without a change in virus production. Failure to induce apoptosis in infected neurons leads to long-term persistence of small amounts of viral RNA in the nervous system of infected mice despite the clearance of infectious virus. The molecular mechanisms that govern these pathogenesis factors are beginning to be elucidated.     

Depletion of Blood-Borne Macrophages Does Not Reduce Demyelination in Mice Infected with a Neurotropic Coronavirus

Mice infected with the neurotropic coronavirus mouse hepatitis virus strain JHM (MHV-JHM) develop a chronic demyelinating disease with symptoms of hindlimb paralysis. Histological examination of the brains and spinal cords of these animals reveals the presence of large numbers of activated macrophages/microglia. In two other experimental models of demyelination, experimental allergic encephalomyelitis and Theiler's murine encephalomyelitis virus-induced demyelination, depletion of hematogenous macrophages abrogates the demyelinating process. In both of these diseases, early events in the demyelinating process are inhibited by macrophage depletion. From these studies, it was not possible to determine whether infiltrating macrophages were required for late steps in the process, such as myelin removal. In this study, we show that when macrophages are depleted with either unmodified or mannosylated liposomes encapsulating dichloromethylene diphosphate, the amount of demyelination detected in MHV-infected mice is not affected. At a time when these cells were completely depleted from the liver, approximately equivalent numbers of macrophages were present in the spinal cords of control and drug-treated animals. These results suggest that blood-borne macrophages are not required for MHV-induced demyelination and also suggest that other cells, such as perivascular macrophages or microglia, perform the function of these cells in the presence of drug.     

High Numbers of Viral RNA Copies in the Central Nervous System of Mice during Persistent Infection with Theiler's Virus

The low-neurovirulence Theiler's murine encephalomyelitis viruses (TMEV), such as BeAn virus, cause a persistent infection of the central nervous system (CNS) in susceptible mouse strains that results in inflammatory demyelination. The ability of TMEV to persist in the mouse CNS has traditionally been demonstrated by recovering infectious virus from the spinal cord. Results of infectivity assays led to the notion that TMEV persists at low levels. In the present study, we analyzed the copy number of TMEV genomes, plus- to minus-strand ratios, and full-length species in the spinal cords of infected mice and infected tissue culture cells by using Northern hybridization. Considering the low levels of infectious virus in the spinal cord, a surprisingly large number of viral genomes (mean of 3.0 x 10(9)) was detected in persistently infected mice. In the transition from the acute (approximately postinfection [p.i.] day 7) to the persistent (beginning on p.i. day 28) phase of infection, viral RNA copy numbers steadily increased, indicating that TMEV persistence involves active viral RNA replication. Further, BeAn viral genomes were full-length in size; i.e., no subgenomic species were detected and the ratio of BeAn virus plus- to minus-strand RNA indicated that viral RNA replication is unperturbed in the mouse spinal cord. Analysis of cultured macrophages and oligodendrocytes suggests that either of these cell types can potentially synthesize high numbers of viral RNA copies if infected in the spinal cord and therefore account for the heavy viral load. A scheme is presented for the direct isolation of both cell types directly from infected spinal cords for further viral analyses.     

Molecular analysis of neurovirulent strains of Sindbis virus that evolve during persistent infection of scid mice.

To understand the role of tissue-specific adaptation and antibody-induced selectional pressures in the evolution of neurovirulent viruses, we analyzed three strains of Sindbis virus isolated from the brains of persistently infected scid mice and four strains of Sindbis virus isolated from the brains of scid mice with viral reactivation following immune serum treatment. For each viral isolate, we tested neurovirulence in weanling BALB/c mice and sequenced regions of the E2 and E1 envelope glycoprotein genes that are known to contain important determinants of Sindbis virus neurovirulence. One strain isolated from a persistently infected scid mouse and two strains isolated from scid mice with viral reactivation were neurovirulent, resulting in mortality in 80 to 100% of weanling BALB/c mice. All three neurovirulent strains contained an A-->U change at nucleotide 8795, which predicts a Gln-->His substitution at E2 amino acid position 55. No nucleotide changes were detected in the other sequenced regions of the E2 and E1 envelope glycoprotein genes or in the avirulent isolates. Our findings indicate that tissue-specific adaptations, rather than antibody-induced selectional pressures, are a critical determinant of the evolution of neurovirulent strains of Sindbis virus and provide evidence that E2 His-55 is an important neuroadaptive mutation that confers neurovirulence properties on Sindbis virus.     

Activation of divergent neuronal cell death pathways in different target cell populations during neuroadapted sindbis virus infection of mice

Infection of adult mice with neuroadapted Sindbis virus (NSV) results in a severe encephalomyelitis accompanied by prominent hindlimb paralysis. We find that the onset of paralysis parallels morphologic changes in motor neuron cell bodies in the lumbar spinal cord and in motor neuron axons in ventral nerve roots, many of which are eventually lost over time. However, unlike NSV-induced neuronal cell death found in the brain of infected animals, the loss of motor neurons does not appear to be apoptotic, as judged by morphologic and biochemical criteria. This may be explained in part by the lack of detectable caspase-3 expression in these cells.     

Sindbis virus-induced neuronal death is both necrotic and apoptotic and is ameliorated by N-methyl-D-aspartate receptor antagonists.

Virus infection of neurons leads to different outcomes ranging from latent and noncytolytic infection to cell death. Viruses kill neurons directly by inducing either apoptosis or necrosis or indirectly as a result of the host immune response. Sindbis virus (SV) is an alphavirus that induces apoptotic cell death both in vitro and in vivo. However, apoptotic changes are not always evident in neurons induced to die by alphavirus infection. Time lapse imaging revealed that SV-infected primary cortical neurons exhibited both apoptotic and necrotic morphological features and that uninfected neurons in the cultures also died. Antagonists of the N-methyl-D-aspartate (NMDA) subtype of glutamate receptors protected neurons from SV-induced death without affecting virus replication or SV-induced apoptotic cell death. These results provide evidence that SV infection activates neurotoxic pathways that result in aberrant NMDA receptor stimulation and damage to infected and uninfected neurons.     

Induction of apoptosis by Sindbis virus occurs at cell entry and does not require virus replication

Sindbis virus (SV) is an alphavirus that causes encephalitis in mice and can lead to the apoptotic death of infected cells. To determine the step in virus replication during which apoptosis is triggered, we used UV-inactivated SV, chemicals that block virus fusion or protein synthesis, and cells that do and do not express heparan sulfate, the initial binding molecule for SV infection of many cells. In initial experiments, UV-inactivated neuroadapted SV (NSV) induced apoptosis in Chinese hamster ovary (CHO) cells lacking heparan sulfate in the presence of cycloheximide. When fusion of prebound UV-inactivated NSV was rapidly induced at the plasma membrane by exposure to acidic pH, apoptosis was induced in CHO cells with or without heparan sulfate in the presence or absence of cycloheximide in a virus dose-dependent manner. In N18 neuroblastoma cells, the relative virulence of the virus strain was an important determinant of apoptosis induced by UV-inactivated SV. Treatment of N18 cells with monensin to prevent endosomal acidification an hour before, but not 2 h after, exposure to live NSV blocked the induction of cell death, as did treatment with NH(4)Cl or bafilomycin A1. These studies indicate that SV can induce apoptosis at the time of fusion with the cell membrane and that virus replication is not required.     

Virus infection induces neuronal apoptosis: A comparison with trophic factor withdrawal

Multicellular organisms can employ a number of defences to combat viral replication, the most dramatic being implementation of a cell autonomous apoptotic process. The overall cost to the viability of an organism of losing infected cells by apoptosis may be small if the dying cells can be substituted. In contrast, suicide of irreplaceable cells such as highly specialised neurons may have a more dramatic, even fatal consequence. Previous in vitro approaches to understanding whether neurotropic viruses cause neurons to apoptose have utilised transformed cell lines. These are not in the appropriate state of differentiation to provide an accurate indication of events in vivo. We have chosen to characterise the ability of a model CNS disease-causing virus, Semliki Forest virus (SFV), to infect and trigger apoptosis in primary cultures of nerve growth factor (NGF)-dependent sensory neurons. These cells are known to die when deprived of NGF and constitute a useful indicator of apoptosis. We observe that infection causes cell death which bears the morphological hallmarks of apoptosis, this occurs even in the present of survival promoting NGF and is concomitant with new virus production. Using the TUNEL (transferase dUTP nick end labelling) technique we show that SFV-induced apoptosis involves DNA fragmentation and requires caspase (CED-3/ICE cysteine protease) activation, as does apoptosis induced by NGF-deprivation. Extensive areas of apoptosis, as defined using a combination of ultrastructural analysis and TUNEL occur in infected neonatal mouse brains. The novel evidence that infection of primary neurons with SFV induces apoptosis with activation of one or more caspases defines a system for the further anlaysis of apoptosis regulation in physiologically relevant neurons.     

Bcl-2 inhibits apoptosis and extends recombinant protein production in cells infected with Sindbis viral vectors

Viruses carrying foreign genes are often used for the production of recombinant proteins in mammalian cells and other eukaryotic expression systems. Though high levels of gene expression are possible using viral vectors, the host cell generally responds to the infection by inducing apoptotic cell death within several days, abruptly ending protein production. It has recently been demonstrated, however, that apoptosis can be suppressed in virally infected cells using anti-apoptotic genes, such as bcl-2. In this study, stably transfected rat carcinomal cell lines, AT3-bcl2 and AT3-neo, were infected with a Sindbis virus carrying the gene for chloramphenicol acetyltransferase (CAT) in an effort to determine the effect of bcl-2 on cell viability and recombinant protein production. Infected AT3-bcl2 cells consistently maintained viabilities close to 100% and a growth rate equivalent to that of uninfected cells (0.040 h^-1). In contrast, the Sindbis viral vector induced apoptosis in the AT3-neo cells, which were all dead by three days post-infection. Though infected AT3-neo cells generated higher levels of heterologous protein, over 1000 mUnits per well, CAT activity fell to zero by two days post-infection. In contrast, chloramphenicol acetyltransferase was present in AT3-bcl2 cells for almost a week, reaching a maximum level of 580 mUnits per well. In addition, recombinant protein production in AT3-bcl2 cells was extended and amplified by the regular addition of virus to the culture medium, a process which resulted in expression for the duration of the cell culture process.Abbreviations BHK Baby Hamster Kidney - CAT chloramphenicol acetyltransferase - dsSV-CAT double subgenomic Sindbis viral vector containing the gene for CAT - MOI multiplicity of infection