Human MxA Protein Protects Mice Lacking a Functional Alpha/Beta Interferon System against La Crosse Virus and Other Lethal Viral Infections

Abstract

The human MxA protein is part of the antiviral state induced by alpha/beta interferon (IFN-α/β). MxA inhibits the multiplication of several RNA viruses in cell culture. However, its antiviral potential in vivo has not yet been fully explored. We have generated MxA-transgenic mice that lack a functional IFN system by crossing MxA-transgenic mice constitutively expressing MxA with genetically targeted (knockout) mice lacking the β subunit of the IFN-α/β receptor (IFNAR-1^−/− mice). These mice are an ideal animal model to investigate the unique antiviral activity of human MxA in vivo, because they are unable to express other IFN-induced proteins. Here, we show that MxA confers resistance to Thogoto virus, La Crosse virus, and Semliki Forest virus. No Thogoto virus progeny was detectable in MxA-transgenic mice, indicating an efficient block of virus replication at the primary site of infection. In the case of La Crosse virus, MxA restricted invasion of the central nervous system. In contrast, Semliki Forest virus multiplication in the brain was detectable in both MxA-expressing and nonexpressing IFNAR-1^−/− mice. However, viral titers were clearly reduced in MxA-transgenic mice. Our results demonstrate that MxA does not need the help of other IFN-induced proteins for activity but is a powerful antiviral agent on its own. Moreover, the results suggest that MxA may protect humans from potential fatal infections by La Crosse virus and other viral pathogens.

La Crosse virus (LACV) and closely related viruses of the California serogroup of bunyaviruses (family, Bunyaviridae) infect humans in many countries of the northern hemisphere (4, 13). LACV is the most important arboviral cause of pediatric encephalitis in the United States. From 1996 to 1997, a total of 252 cases of LACV encephalitis have been reported (5). It has been estimated that there may be as many as 300,000 LACV infections annually in the midwestern United States alone (4). However, the vast majority of infections is clinically inapparent or associated with mild symptoms, suggesting that humans have a powerful defense against LACV infections.

It is well known that the interferon (IFN) system plays a pivotal role in the first line of defense against viruses. Many cell types produce and secrete alpha and beta IFN (IFN-α/β) in response to viral infections in a paracrine fashion, thereby signalling the presence of an invading virus to neighboring cells. The binding of IFN-α/β to their specific cell surface receptors triggers the intracellular Jak/STAT pathway, leading to the activation or enhanced expression of more than 50 genes (34, 37, 38). Their combined activities generate a so-called antiviral state. The proper functioning of the IFN-α/β system is essential for the survival of certain viral infections. Blocking IFN-α/β activity in mice by the injection of antibodies directed against IFN-α and IFN-β leads to a dramatically increased sensitivity to many viruses (14, 15, 17). Furthermore, genetically targeted (knockout) mice lacking the β subunit of the IFN-α/β receptor (IFNAR-1^−/− mice) are unable to establish an antiviral state and, as a consequence, are highly susceptible to many viral infections, despite the presence of an otherwise intact immune system (6, 28). However, the contribution of an individual IFN-induced protein to the generation of the antiviral state is difficult to assess, because various effector proteins appear to have overlapping antiviral activities (38).

Mx proteins are among the few effector proteins of the IFN-α/β system with known antiviral activity. They are highly conserved large GTPases with homology to dynamin and have been found in all vertebrate species investigated so far, including mammals, birds, and fish (reviewed in references 3 and 41). The human MxA protein is a cytoplasmic protein (1, 40) which is rapidly induced in response to acute viral infections (33). Transfected cells, expressing MxA under the control of a constitutive promoter, are resistant to infections with viruses of several RNA virus families, namely, Orthomyxoviridae (10, 11, 30, 31), Paramyxoviridae (35, 36, 44), Rhabdoviridae (31), Bunyaviridae (9, 25), and Togaviridae (27). A first indication for the role of MxA in vivo came from transgenic mice which constitutively express human MxA but lack functional mouse Mx proteins (29). These MxA-transgenic mice were completely resistant to infections with Thogoto virus (THOV), a tick-borne orthomyxovirus, and they proved to be less sensitive to infections with influenza A virus and vesicular stomatitis virus (29).

Here, we demonstrate that the function of a single IFN-induced effector protein can be studied in vivo without interference from activities of other IFN-induced proteins. To that end, we crossed MxA-transgenic and IFNAR-1^−/− mice resulting in MxA^+/+ IFNAR-1^−/− mice. We show that MxA expression is sufficient to protect IFNAR-1^−/− mice against a lethal challenge dose of THOV. Furthermore, enhanced resistance was observed against LACV and Semliki Forest virus (SFV), a neurotropic virus of the family Togaviridae.

MATERIALS AND METHODS

Mice.

The generation of the MxA-transgenic mouse lines L and G, as well as the generation of IFNAR-1^−/− knockout mice, was described previously (28, 29). MxA-expressing IFNAR-1^−/− knockout mice that originated from the MxA-transgenic L line were generated as follows. Mice homozygous for the MxA transgene (MxA^+/+) were mated with IFNAR-1^−/− mice. Resulting F₁ offspring (MxA^+/− IFNAR-1^+/−) were interbred. The IFNAR-1 genotype of the F₂ generation was analyzed by PCR as described previously (28). To test whether F₂ animals were homozygous for the MxA transgene, they were backcrossed with BALB/c mice, and the MxA genotypes of the progeny were analyzed by PCR as described previously (28). Mice homozygous for both MxA and the IFN-α/β receptor defiency (MxA^+/+ IFNAR-1^−/−) were used for further breeding. A second MxA-transgenic IFNAR-1^−/− mouse line was generated with the MxA-transgenic G line. For unknown reasons, breeding of MxA-transgenic mice of line G never yielded homozygous females (29). Therefore, male mice homozygous for MxA of line G were first crossed with female IFNAR-1^−/− mice. F₁ animals thereof were interbred, and the resulting F₂ progeny were tested for their MxA and IFNAR-1 genotypes as described above. F₂ males homozygous for the MxA transgene as well as the IFN-α/β receptor deficiency (MxA^+/+ IFNAR-1^−/−) were selected and backcrossed with IFNAR-1^−/− mice. The resulting offspring (MxA^+/− IFNAR-1^−/−) were used for virus challenge experiments. All mouse lines described in this paper have mutations in the endogenous mouse Mx genes Mx1 and Mx2 (39, 42). As a consequence, functional Mx1 and Mx2 proteins are not expressed in these mice.

Analysis of MxA expression in transgenic mice.

Animals were anesthetized and exsanguinated, and a variety of organs and tissue samples were removed, snap frozen in liquid nitrogen, and stored at −70°C. The frozen samples were homogenized in a buffer containing 50 mM Tris-HCl (pH 8.3), 50 mM KCl, 5 mM MgCl₂, 1 mM EDTA, and 0.1% Triton X-100. Subsequently, the cells were lysed by sonication. The lysates were cleared by centrifugation at 10,000 × g for 10 min and mixed with sodium dodecyl sulfate (SDS)-gel sample buffer (26). Protein samples (20 μg per lane) were separated by SDS–10% polyacrylamide gel electrophoresis. Transfer to nitrocellullose membranes (Millipore, Bedford, Mass.) and Western blot analysis were carried out essentially as previously described (1), with a monoclonal antibody specific for MxA (21) and a chemiluminescence detection kit (Pierce, Rockford, Ill.).

Virus stocks.

The Sicilian (SiAr126) isolate of THOV (2) was grown in BALB/c mice as previously described (19). Stock virus prepared from liver homogenates contained 7 × 10⁶ PFU per ml as titrated on Swiss mouse 3T3 cells. The original strain of LACV (43) was grown on baby hamster kidney (BHK-21) cells yielding a titer of 1.2 × 10⁸ 50% tissue culture infective doses (TCID₅₀) per ml as determined on Vero cells. The SFV prototype strain was grown on Swiss mouse 3T3 cells yielding a titer of 6.8 × 10⁹ TCID₅₀ per ml as determined on the same cell type.

Experimental viral infections.

For each set of experiments mice were age matched. Five- to eight-week-old mice were anesthetized and intraperitoneally infected with 300 PFU of THOV, 10⁵ TCID₅₀ of LACV, or 10² TCID₅₀ of SFV. The animals were monitored for clinical symptoms at least once a day.

Detection of virus yields.

Mice were anesthetized and exsanguinated, and organs and tissue samples were removed, snap frozen in liquid nitrogen, and stored at −70°C. The frozen samples were weighed and transferred to a vial containing 9 volumes of phosphate-buffered saline (PBS) solution per weight of tissue sample. The organs were ground with quartz sand, and the resulting suspensions were cleared by centrifugation and again frozen at −70°C. Virus yields were determined by the TCID₅₀ method with Swiss mouse 3T3 cells for THOV and SFV and Vero cells for LACV.

Immunohistochemical analysis.

Mouse brains were fixed in PBS containing 4% formaldehyde for 48 h and subsequently washed in PBS. Coronal and sagital slices of approximately 3 mm were dehydrated through graded alcohols and embedded in paraffin. Sections of 3-μm nominal thickness were stained with hematoxylin and eosin or stained for cellular and viral proteins. Immunostaining for the glial fibrillary acidic protein (GFAP) was carried out with a rabbit antiserum specific for GFAP (DAKO, Copenhagen, Denmark) and a biotinylated swine anti-rabbit immunoglobulin serum (dilution, 1:300 and 1:250, respectively). Visualization was achieved by using avidin-peroxidase and diaminobenzidine. For the immunostaining of MxA a mouse monoclonal antibody specific for MxA and polyclonal rabbit anti-mouse immunoglobulin serum was used (dilution, 1:50 and 1:20, respectively). Visualization was carried out by using calf intestinal alkaline phosphatase complexed with a mouse monoclonal anti-alkaline phosphatase antibody (dilution, 1:50). Immunostaining of viral antigens was performed with a polyclonal rabbit anti-C protein of SFV (dilution, 1:50) and a polyclonal rabbit antiserum specific for the nucleocapsid protein of LACV (dilution, 1:50) (kindly provided by Raju Ramasamy, Meharry Medical College, Nashville, Tenn.). Mouse monoclonal anti-rabbit immunoglobulins (dilution, 1:25) and a rabbit polyclonal anti-mouse immunoglobulin (dilution, 1:25) were used as bridging antibodies. The remaining steps of the procedure were the same as those used for the immunostaining of MxA. The secondary and tertiary antibodies were purchased from DAKO. All immunostained sections were counterstained with hematoxylin.

RESULTS

Generation of MxA-transgenic mice lacking a functional IFN-α/β receptor.

We have previously generated two MxA-transgenic mouse lines, designated G and L (29). The expression of the transgene is controlled by the 3-hydroxy-3-methylglutaryl coenzyme A reductase gene promoter (12), resulting in an IFN-independent constitutive expression of MxA in various organs (29).

To generate transgenic mice which express MxA but lack a functional IFN-α/β receptor, we crossed transgenic L mice homozygous for MxA (MxA^+/+) and knockout mice homozygous for IFN-α/β receptor deficiency (IFNAR-1^−/−) (28). Animals of the resulting F₁ generation were interbred, and the genotype of the F₂ generation was analysed. Subsequently, MxA^+/+ IFNAR-1^−/− mice of the F₂ generation were used to breed the F₃ generation. If not indicated otherwise, homozygous MxA^+/+ IFNAR-1^−/− mice derived from line L were used for all challenge experiments described in this study.

The expression of MxA was verified by Western blot analysis of protein extracts from various organs of an MxA-transgenic IFNAR-1^−/− mouse with a monoclonal antibody specific for the MxA protein (21). Figure 1 shows that MxA is detectable in all tissues but muscles. The highest expression levels were found in the brain and spleen, as previously noted in mice of the parental L line (29).

FIG. 1.

Detection of MxA protein by Western blot analysis. Various organs and tissue samples from an adult MxA^+/+ IFNAR-1^−/− mouse (originating from the MxA-transgenic L line) were collected, cell extracts were prepared, and samples of 20 μg of protein per lane were separated by SDS-polyacrylamide gel electrophoresis. MxA was detected with an MxA-specific monoclonal antibody (21).

Human MxA confers complete resistance to THOV in IFNAR-1^−/− knockout mice.

To test whether MxA is able to reverse the virus-sensitive phenotype of mice lacking a functional IFN system, MxA-transgenic mice and appropriate control animals were infected with THOV, a tick-borne orthomyxovirus that causes lethal infections in mice lacking functional mouse Mx alleles (19). However, THOV is extremely sensitive to the antiviral action of MxA in cell cultures (10) and in vivo (29). Mice of the genotypes MxA^+/+ IFNAR-1^+/+, MxA^+/+ IFNAR-1^−/−, MxA^−/− IFNAR-1^−/−, and MxA^−/− IFNAR-1^+/+ were infected intraperitoneally with 300 PFU of THOV. All MxA-transgenic mice survived the challenge, irrespective of whether the animals were deficient for the IFN-α/β receptor or not (Fig. 2A). In contrast, mice lacking the MxA transgene succumbed to infection within 3 days (IFNAR-1^−/− mice) or 5 days (IFNAR-1^+/+ mice) (Fig. 2A). In parallel, we monitored the multiplication of THOV in two additional mice per group. These animals were sacrificed 36 h after infection, and their livers were removed. Upon intraperitoneal inoculation, THOV is known to infect the liver and to replicate to high titers in this organ (7, 19). As expected, liver homogenates of animals without MxA exhibited high viral yields exceeding 10⁶ TCID₅₀ per g of tissue. In contrast, no virus was detectable in MxA-transgenic mice (Fig. 2B). Taken together, the data indicate that the expression of MxA in IFNAR-1^−/− mice is sufficient to confer complete protection against THOV infection.

FIG. 2.

MxA-transgenic IFNAR-1^−/− mice are resistant to THOV infection. (A) Adult mice of four different genotypes were challenged with THOV and monitored for survival. MxA^+/+ IFNAR-1^−/− (■), IFNAR-1^−/− (□), MxA^+/+ IFNAR-1^+/+ (●), and IFNAR-1^+/+ mice (○) were infected with 300 PFU of THOV per ml by the intraperitoneal route. (B) Virus growth in livers of susceptible and resistant mice. Two animals per genotype were infected as described above and sacrificed 36 h after infection. Their livers were removed and assayed for infectivity. Each column represents the mean virus titer for two animals.

Human MxA restricts multiplication of LACV in IFNAR-1^−/− knockout mice.

Mice show a strong age-dependent susceptibility to LACV, whereby 6-week-old mice acquire complete resistance (22, 23). However, we have recently observed that adult IFNAR-1^−/− knockout mice are highly susceptible to experimental infections with LACV (20). LACV multiplication has been shown to be inhibited by the antiviral activity of MxA in Vero cells (9). To assess the activity of MxA on LACV multiplication in vivo, we challenged adult MxA-transgenic IFNAR-1^−/− and IFNAR-1^−/− control mice with 10⁵ TCID₅₀ of LACV by the intraperitoneal route. As expected, the infection of 6-week-old IFNAR-1^−/− mice led to the development of severe neurological symptoms and to the death of 17 of 18 animals with a mean survival time of 6.5 days (Fig. 3A). In contrast, 7 of 16 MxA-transgenic IFNAR-1^−/− mice survived infection. The nine MxA-transgenic animals that died showed an increased mean survival time of 8 days (Fig. 3A). To corroborate these findings, we infected animals of another MxA-transgenic IFNAR-1^−/− mouse line which was generated by using MxA-transgenic G mice. Again, all of the IFNAR-1^−/− control mice died after infection with LACV, whereas four of seven MxA^+/− IFNAR-1^−/− mice survived, although they were heterozygous for the MxA transgene (data not shown). All experimental infections with LACV were carried out with approximately 10 50% lethal doses to assure 100% killing of the IFNAR-1^−/− mice. At lower virus doses, we observed higher survival rates for both the MxA^+/+ IFNAR-1^−/− mice and IFNAR-1^−/− control mice (data not shown).

FIG. 3.

Enhanced resistance of MxA-transgenic IFNAR-1^−/− mice to LACV infection. (A) Adult MxA^+/+ IFNAR-1^−/− (■) and MxA^−/− IFNAR-1^−/− (□) mice were infected with 10⁵ TCID₅₀ of LACV by the intraperitoneal route and monitored for survival. (B) Virus growth in brains of susceptible and resistant mice. Two MxA^+/+ IFNAR-1^−/− mice (animals 1 and 2) and three MxA^−/− IFNAR-1^−/− mice (animals 3, 4, and 5) were infected as described above and sacrificed 6 days after infection. Tissue samples were removed and assayed for infectivity.

To investigate MxA-mediated inhibition of LACV in more detail, three additional IFNAR-1^−/− and two MxA-transgenic IFNAR-1^−/− mice were infected intraperitoneally with 10⁵ TCID₅₀ of LACV and sacrificed 6 days later. By that time, two IFNAR-1^−/− mice showed severe clinical symptoms including complete paralysis of the hind legs. In the diseased IFNAR-1^−/− mice, high titers (≥10⁸ TCID₅₀ per g of tissue) of LACV were found in the brain (Fig. 3B, animals 3 and 4). The third animal showed no clinical symptoms. Nevertheless, the brain still contained 10³ TCID₅₀ of LACV per g of brain tissue (Fig. 3B, animal 5), indicating that the virus had started to multiply in the central nervous system (CNS). In contrast, no virus was detectable in the two MxA-transgenic IFNAR-1^−/− mice (Fig. 3B, animals 1 and 2).

Brain tissues from the same mice were further analyzed by histological and immunohistological methods. LACV multiplication in the brains of IFNAR-1^−/− mice lacking MxA was confirmed by immunostaining the viral nucleocapsid protein with specific antibodies (Fig. 4B). Furthermore, hematoxylin and eosin staining (Fig. 4C) or immunostaining for GFAP (Fig. 4D) revealed infiltration of mononuclear inflammatory cells and astrocytes characteristic of meningitis and pronounced astrogliosis. In brains of healthy MxA-transgenic IFNAR-1^−/− mice expressing MxA in the majority of cells (Fig. 4E), no signs of virus infection and no mononuclear cell infiltrates were detectable (Fig. 4F to H).

FIG. 4.

Histology and immunostaining of brain sections from an MxA^+/+ IFNAR-1^−/− and an MxA^−/− IFNAR-1^−/− mouse. Both types of animals were infected with 10⁵ TCID₅₀ of LACV by the intraperitoneal route. Mice were sacrificed 6 days after infection, and brain sections were prepared. Micrographs show the brain of an IFNAR-1^−/− mouse (panels A to D) and that of an MxA^+/+ IFNAR-1^−/− mouse (panels E to H) immunostained for MxA protein (panels A and E), the nucleocapsid protein of LACV (panels B and F), or GFAP (panels D and H) or stained with hematoxylin and eosin (HE) (panels C and G).

These results demonstrate that (i) IFNAR-1^−/− mice are a suitable animal model for studies on LACV-mediated pathogenesis and (ii) human MxA is able to inhibit the multiplication of LACV in vivo. To investigate the cause of death of the MxA-transgenic IFNAR-1^−/− mice that developed clinical symptoms and finally succumbed, we assessed virus replication and pathology in the brain of one diseased animal of that group. The animal clearly died from acute meningoencephalitis. The viral titer was determined to be 10⁸ TCID₅₀ per g of brain tissue, which is comparable to the titers observed in brains of IFNAR-1^−/− mice (Fig. 3B, animals 3 and 4). Moreover, immunohistochemical analysis revealed the accumulation of LACV nucleocapsid protein, as well as infiltrates of lymphocytes and astrocytes (data not shown).

Human MxA restricts multiplication of SFV in IFNAR-1^−/− knockout mice.

Experimental infections of mice with SFV lead to a wide range of pathologies depending on the virus strain and the age of the host. For example, 129Sv mice survive high doses of the SFV prototype strain without apparent clinical symptoms (8). In contrast, adult IFNAR-1^−/− mice are extremely sensitive to the prototype strain of SFV and are killed upon inoculation with as few as 10 infectious particles (28). We infected MxA-transgenic IFNAR-1^−/− mice and IFNAR-1^−/− control mice with 100 PFU of the prototype strain of SFV by the intraperitoneal route. All 22 IFNAR-1^−/− mice rapidly exhibited severe neurological symptoms and died within 6 days, as expected (Fig. 5A). In contrast, only a fraction (13 of 22) of the MxA-transgenic animals developed disease and succumbed to the infection (Fig. 5A). The surviving animals did not show any clinical symptoms during an observation period of 30 days. Furthermore, the mean survival time of diseased MxA-transgenic IFNAR-1^−/− mice was increased compared to that of IFNAR-1^−/− mice (6.8 and 4.5 days, respectively). In an additional experiment, we infected MxA-transgenic IFNAR-1^−/− mice derived from the G line. Six of fifteen MxA^+/− IFNAR-1^−/− mice and 0 of 14 IFNAR-1^−/− control mice survived the infection with 100 PFU of SFV (data not shown). All SFV infections were carried out with approximately 10 50% lethal doses to assure 100% killing of the IFNAR-1^−/− control mice. With less inoculum, we observed higher survival rates of both MxA^+/+ IFNAR-1^−/− mice and control mice (data not shown).

FIG. 5.

Enhanced resistance of MxA-transgenic IFNAR-1^−/− mice to SFV infection. (A) Adult MxA^+/+ IFNAR-1^−/− (■) and MxA^−/− IFNAR-1^−/− (□) mice were infected with 100 PFU of SFV by the intraperitoneal route and monitored for survival. (B) Virus growth in various organs of susceptible and resistant mice. Two MxA^+/+ IFNAR-1^−/− and two MxA^−/− IFNAR-1^−/− mice were infected as described above and sacrificed 4 days after infection. Tissue samples were removed and assayed for infectivity. Each column represents the mean virus titer for two animals.

To examine SFV multiplication and pathology, two MxA- transgenic IFNAR-1^−/− mice and two IFNAR-1^−/− control mice were infected with SFV and sacrificed 4 days later (Fig. 5B). In IFNAR-1^−/− mice virus replication occurred in all organs of the mice tested. For example, up to 3 × 10¹⁰ TCID₅₀ per g of tissue was observed in the brain (Fig. 5B). Organs of MxA-transgenic IFNAR-1^−/− mice contained drastically reduced viral titers ranging from 0.05% (brain) to 2% (lung) of the corresponding titers in IFNAR-1^−/− mice lacking MxA (Fig. 5B). In parallel, brain sections were subjected to immunohistological analysis (Fig. 6). Large amounts of the C protein of SFV were found in the brains of IFNAR-1^−/− mice (Fig. 6B). Staining with hematoxylin and eosin (Fig. 6C) and immunostaining for GFAP (Fig. 6D) revealed an acute encephalomyelitis in IFNAR-1^−/− mice. In contrast, viral antigens were not detectable in the brains of healthy MxA-transgenic IFNAR-1^−/− mice (Fig. 6F). Furthermore, no signs of encephalitis were found in the brains of these mice (Fig. 6G and H). However, diseased MxA-transgenic IFNAR-1^−/− mice accumulated large amounts of SFV antigens in their brains and showed the typical immunohistological picture of acute encephalitis usually observed in SFV-infected IFNAR-1^−/− control mice (data not shown). Evidently, SFV is able to overrun the protective effect of MxA in some of the MxA-transgenic IFNAR-1^−/− mice.

FIG. 6.

Histology and immunostaining of brain sections from an MxA^+/+ IFNAR-1^−/− mouse and an MxA^−/− IFNAR-1^−/− mouse. Both were infected with 100 PFU of SFV per ml by the intraperitoneal route. Mice were sacrificed 4 days after infection, and brain sections were prepared. Micrographs show the brain of an MxA^+/+ IFNAR-1^−/− mouse (panels A to D) and that of an MxA^−/− IFNAR-1^−/− mouse (panels E to H) immunostained for MxA protein (panels A and E), the C protein of SFV (panels B and F), or GFAP (panels D and H) or stained with hematoxylin and eosin (HE) (panels C and G).

These results demonstrate that MxA inhibits the multiplication of SFV in vivo. However, the death of some MxA-transgenic IFNAR-1^−/− mice indicates that MxA expression was not sufficient to block SFV replication completely.

DISCUSSION

Here, we show that human MxA protein protects MxA-transgenic mice from lethal virus infection independent of other IFN-induced proteins. We generated MxA-transgenic IFNAR-1^−/− mice because they express MxA constitutively in various organs but are unable to mount an endogenous IFN-α/β response. Challenge experiments revealed that these mice were highly resistant to infection with THOV, a tick-borne orthomyxovirus. We have previously shown that THOV is inhibited by MxA in transgenic mice (29). The previous experiments were performed with inbred mouse strains that lack functional mouse Mx proteins but possess an otherwise intact IFN-α/β system. It was therefore possible that other IFN-induced proteins would act in conjunction with ectopic MxA to yield complete protection. The results presented here clearly demonstrate that this is not the case. MxA alone is able to block the multiplication of THOV without the help of other IFN-induced proteins. Since antibodies against THOV have been detected in the sera of various species including man (7), one has to assume that human infections can occur but that the virus may be completely inhibited by the action of human MxA protein. Accordingly, only a few clinical cases due to THOV virus infection in humans have been reported.

In cell cultures, MxA has the potential to inhibit a wide range of RNA viruses, including members of the families Orthomyxoviridae, Paramyxoviridae, Rhabdoviridae, Bunyaviridae, and Togaviridae (18). However, the crucial question about the importance of MxA for the antiviral defense in vivo has remained unanswered. LACV and closely related viruses frequently infect humans (4, 13). In spite of the high number of infections, clinical cases are rare. We and others have shown previously that the multiplication of LACV and other members of the family Bunyaviridae is inhibited by MxA in stably transfected Vero cells (9, 25). These results suggested that MxA is part of the antiviral defense mechanism against bunyaviruses in humans. However, experiments elucidating the effect of MxA in vivo were not possible, because a suitable animal model was missing. The laboratory mouse Mus musculus shows a strong age-dependent susceptibility to experimental LACV infections. Suckling mice die from infection irrespective of the site of virus inoculation, whereas adult mice develop a lethal encephalitis only when infected by the intracerebral route (22, 23). The known sensitivity of IFNAR-1^−/− knockout mice to many viral infections (6, 28) prompted us to test these mice for their susceptibility to LACV (20). Our results demonstrate that adult IFNAR-1^−/− mice indeed represent a suitable animal model for studies on LACV-mediated pathogenesis. Interestingly, the neurological symptoms and the immunohistological findings observed in LACV-infected IFNAR-1^−/− mice resemble those described in rare cases of acute LACV encephalitis in humans (24).

Here, we show that about 40% of the MxA-transgenic IFNAR-1^−/− mice survived the experimental infections without apparent clinical symptoms while 100% of IFNAR-1^−/− control mice died. Furthermore, MxA-transgenic animals that succumbed to infection showed a delayed onset of disease. It is presently not clear why only a subset of MxA-transgenic animals survived. The extent to which virus growth was blocked by MxA in muscle cells, the major extraneuronal replication site (16), may be critical to the course of the disease. For unknown reasons, the level of MxA expression in muscle cells is low in both MxA-transgenic founder lines, G and L (29), as well as in MxA-transgenic IFNAR-1^−/− mice. As shown in Fig. 4E and 6E MxA expression in the brain is not uniform and there are cells expressing higher levels of MxA than others. Most likely, the inhibition of virus replication is not complete in the periphery and, depending on the MxA expression levels at the site of CNS entry, LACV might gradually overcome MxA-mediated inhibition. In humans, most LACV infections follow a subclinical course, which may be the result of the induction of MxA expression, but nevertheless in a few cases acute encephalitis is produced (4, 5). In view of the present findings it is conceivable that the degree of clinical manifestations may depend on the extent of IFN production and hence MxA expression during infection. In the few cases where acute illness is observed, the inefficient induction of antiviral effector proteins like MxA might allow uncontrolled LACV replication at the site of primary infection followed by virus spread to the brain. Furthermore, humans with genetic defects in IFN signalling or the MxA gene may be predisposed towards LACV encephalitis. It would be interesting to determine the proper function of IFN signalling and MxA in severe cases of acute LACV encephalitis.

SFV, a member of the family Togaviridae, is so far the only positive-stranded RNA virus which is affected by the antiviral action of MxA (27). Here, we show that MxA is also able to inhibit the multiplication of SFV in vivo. One hundred percent of IFNAR-1^−/− mice died upon infection, whereas only 60% of MxA-transgenic IFNAR-1^−/− mice succumbed. Furthermore, viral titers were lower in MxA-transgenic IFNAR-1^−/− mice than in IFNAR-1^−/− control mice. It should be emphasized that SFV was detectable in the brains of all MxA-transgenic animals analyzed. Obviously, MxA is not able to prevent initial infection of the CNS. Rather, MxA appears to reduce virus replication within the brain. The reason why some animals became progressively diseased after a few days and finally succumbed remains elusive. The simplest explanation is that SFV replication occurred initially in brain cells expressing low levels of MxA. This may have led to greater virus loads in the CNS, and gradually the MxA-mediated block of virus multiplication may have been overcome in an increasing number of cells.

Our results demonstrate that MxA is able to protect transgenic animals against a number of viruses and conclusively establish this protein as an important intracellular mediator of the antiviral effects of IFN-α/β.

Progress in genetic engineering should allow us in the near future to introduce genes like MxA into farm animals in order to improve their disease resistance. A weak spot in the live cycle of arthropod-borne pathogens is their transmission by vectors. Therefore, some viruses could be controlled by MxA even before humans or livestock get infected. In an attempt to establish pathogen-derived resistance in arthropod vectors, the multiplication of LACV was shown to be inhibited in mosquitoes that express genetic elements of the LACV genome (32). An alternative strategy might be to use MxA for the generation of LACV-resistant mosquitoes.