Mice Lacking Alpha/Beta and Gamma Interferon Receptors Are Susceptible to Junin Virus Infection

Abstract

Junin virus (JUNV) causes a highly lethal human disease, Argentine hemorrhagic fever. Previous work has demonstrated the requirement for human transferrin receptor 1 for virus entry, and the absence of the receptor was proposed to be a major cause for the resistance of laboratory mice to JUNV infection. In this study, we present for the first time in vivo evidence that the disruption of interferon signaling is sufficient to generate a disease-susceptible mouse model for JUNV infection. After peripheral inoculation with virulent JUNV, adult mice lacking alpha/beta and gamma interferon receptors developed disseminated infection and severe disease.

Argentine hemorrhagic fever (AHF) is a highly lethal disease that is endemic in central Argentina (16) and characterized by hematological, immunological, neurological, and hemorrhagic manifestations. The etiological agent of AHF—a member of clade B New World human pathogenic arenaviruses—Junin virus (JUNV) enters cells using the human transferrin receptor 1 (hTfR1) (22). The absence of an efficient receptor required for the virus to enter cells has been previously proposed as an explanation for the resistance of laboratory mice to JUNV infection (22, 23). For that reason, generation of transgenic mice expressing human or humanized TfR1 was offered as a solution for the production of a JUNV-susceptible murine model (22, 23).

In contrast, JUNV replication in rat and mouse astrocytes and in neurons has been well documented (1-3, 8, 10-14, 21). Consistent with those findings, a recent study demonstrated the ability of New World clade B pathogenic arenaviruses to utilize a TfR1-independent entry pathway, although less efficiently (7). On the basis of previously published studies reporting inhibition of type I interferon induction mediated by nucleocapsid protein (NP) and Z proteins of arenaviruses (6, 18, 19), we hypothesized that counteracting the interferon response is critical for JUNV to establish productive infections in mice.

Increased susceptibility of mice lacking functional interferon signaling to JUNV infection.

We then performed animal experiments to determine whether the natural resistance of laboratory mice to JUNV infection is based on the absence of a specific cellular receptor for JUNV or on the functional interferon response. For that reason, 4- to 8-week-old male and female wild-type hybrid mice (strain 129 were backcrossed twice with C57BL/6 mice) (a total of 15 mice in trial 1 and a total of 8 mice in trial 2) and mice deficient for alpha/beta and gamma interferon receptor (IFN-α/βγR^−/−) (a total of 5 mice in trial 1 and a total of 8 mice in trial 2) (24) on the same genetic background were inoculated intraperitoneally on day 0 with wild-type JUNV Romero strain (20) using 1 × 10⁴ PFU/animal. All animal studies were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Texas Medical Branch (UTMB) and housed at biosafety level 4 (BSL-4).

Animals were humanely euthanized upon development of severe body weight loss exceeding 15% (trial 1) or 20% (trial 2). Statistical analysis comparing survival time and changes in body weight and temperature was performed by using GraphPad Prism (San Diego, CA) using Kaplan-Meier survival analysis (log-rank test) and two-way repeated-measure analysis of variance, respectively, followed by Holm-Sidak posttest for pairwise multiple comparison.

Similar disease development was observed in both trials. Wild-type mice had gained 18.9 and 19.5% in weight by the end of the first and second studies or trials, respectively. In contrast, IFN-α/βγR^−/− mice have lost 17% (trial 1) and 24% (trial 2) of their body weight by day 13 postinfection (p.i.) (Fig. 1A). In the first trial, on day 13 p.i., one mouse from the IFN-α/βγR^−/− group died (Fig. 1B). A decrease (although not statistically significant) in body temperature at day 12 p.i. preceded the first deaths (Fig. 1C). Four out of 13 IFN-α/βγR^−/− mice developed a scruffy coat starting as early as day 8 p.i. One JUNV-infected IFN-α/βγR^−/− mouse died, and all mice were morbid by day 14 p.i. (Fig. 1A and B) with mean survival times of 13.4 and 13.5 day p.i. for the first and second trials, respectively. We want to emphasize that in our experimental setting, the true lethality was masked by the fact that we euthanized animals on the basis of body weight loss.

FIG. 1.

Increased susceptibility of IFN-α/βγR^−/− mice to JUNV infection. The susceptibility of IFN-α/βγR^−/− mice to JUNV compared to weight- and age-matched wild-type mice was tested in two studies, trial 1 and trial 2. The mice are shown as follows in the figure: 1-Hybrid/IFNa/βγR−/−, IFN-α/βγR^−/− mice in trial 1; 1-Hybrid, wild-type mice in trial 1; 2-Hybrid/IFN-a/βγR−/−, hybrid IFN-α/βγR^−/− mice in trial 2; 2-Hybrid, wild-type mice in trial 2. (A) Percent change in body weight. Each data point represents the average percent change in the body weight of mice relative to the baseline weight (day before infection [day −1]) ± standard error of the mean (SEM) (error bar). P values with the corresponding critical levels for statistically significant comparisons of the body weight changes of the IFN-α/βγR^−/− mice compared to the wild-type control mice are provided in the figure. (B) Survival analysis. P values with the corresponding critical levels for statistically significant comparisons of survival comparing survival time for the IFN-α/βγR^−/− mice with the wild-type control mice are provided in the figure. (C) Percent change in body temperature. Each data point represents the average percent change in body temperature of mice relative to the baseline temperature (day −1 for trial 1 and day 0 for trial 2) ± SEM.

More importantly, disseminated infection was detected in moribund or dead IFN-α/βγR^−/− mice, while the organs from wild-type mice contained no infectious virus. In both studies, infectious virus levels, which are shown as log(mean PFU/g tissue ± SD), were highest in the kidneys (5.1 ± 1.1) (trial 1) and 4.7 ± 0.7 (trial 2) but were also relatively high in the brain, liver, spleen, and heart, with average values of 3.3 ± 0.7/3.0 ± 1.1, 3.3 ± 0.8/4.0 ± 1.1, 3.1 ± 0.7/3.7 ± 1.0, and 3.1 ± 0.3/2.9 ± 1.4, respectively (the value for the first trial is shown before the slash, and the value for the second trial is shown after the slash) (Fig. 2A). To confirm the absence of JUNV dissemination in wild-type mice earlier in infection, we similarly infected 7 wild-type mice with JUNV and sacrificed them on days 2 (2 mice), 4 (2 mice), and 6 (3 mice) p.i. No measurable viral load was detected in all organs and serum samples tested, suggesting an absence of JUNV dissemination in wild-type adult mice. The titer of infectious virus was determined as previously described (25).

FIG. 2.

JUNV virus replication in adult mice or in cell lines of murine origin. (A) Disseminated infection in JUNV-infected IFN-α/βγR^−/− mice. The levels of infectious JUNV in the brains and peripheral organs of IFN-α/βγR^−/− mice in trial 1 at the following times are shown: 13 days p.i. (mouse 35 [M35] to M37), 14 days p.i. (M38 and M39), or upon death (M35) or development of severe disease (weight loss exceeding 15%). Data from trial 2 are from 14 days p.i. (M04, M05, and M11), 10 days p.i. (M01 to M03, M09, and M10), or upon scheduled sacrifice (M02, M03, M09, and M10) or body weight loss exceeding 20%. The total number of animals tested is indicated for each organ. The limit of detection of 50 PFU/g tissue is indicated by the dashed line. (B) JUNV production in MEF from wild-type mice and wild-type IFN-α/βγR^−/− mice. Data represent the average ± SEM of three replicate samples of MEF in tissue culture. The limit of detection of 10 PFU/ml is indicated by the dashed line.

Despite undetectable viral loads, the levels of JUNV-specific neutralizing antibodies were measurable in the wild-type mice 31 days p.i. using a plaque reduction neutralization test (PRNT) (25) with 80% reduction titers in the range of 20 to 160 (Table 1) indicating productive infection with JUNV. By the time of euthanasia, 4 out of 6 tested IFN-α/βγR^−/− mice had PRNT titers ranging from 20 to 80 (Table 1). The low titers of neutralizing antibodies correlated with the functionality of the knockout phenotype.

TABLE 1.

Titers of JUNV-neutralizing antibodies in IFN-α/βγR^−/− and wild-type hybrid mice^a

Trial and mouse type	Day p.i.	No. positive/no. testedb	No. positive/80% reduction titer by PRNTc
Trial 1
IFN-α/βγR−/−	13	1/1	1/1:20
Wild-type	31	6/7	4/1:40, 2/1:20
Trial 2
IFN-α/βγR−/−	6	0/4
	10	2/4	1/>1:80, 1/1:40
	14	1/1	1/>1:40
Wild-type	6	0/2
	10	0/1
	31	5/6	2/1:20, 1/1:80, 2/1:160

^aSeroconversion was assessed by 80% reduction by PRNT. Samples that were below the limit of detection (80% reduction in PRNT < 1:20) were considered negative.

^bNo. of mice positive for antibodies to JUNV/number of mice tested.

^cThe number of mice positive for antibodies to JUNV is shown before the slash, and the 80% reduction value by PRNT is shown after the slash (e.g., 1/1:20 means that one mouse had a 80% reduction titer by PRNT of 1:20).

Interferon signaling deficiency does not affect JUNV production in vitro.

We infected three replicate samples of 5 × 10⁴ mouse embryonic fibroblasts (MEF) derived from IFN-α/βγR^−/− mice or hybrid mice with JUNV at a multiplicity of infection (MOI) of 0.1 PFU/cell. Supernatants were collected daily, and virus production was determined by plaque assay (Fig. 2B). JUNV production rose exponentially after 24 h p.i., peaking 72 h p.i. The levels of JUNV were undistinguishable in the cell lines from different mice at all times p.i. These data most likely indicate that in vivo interferon does not directly restrict JUNV multiplication but activates innate and adaptive host defense responses (15).

Histopathologic changes in visceral organs and brains of JUNV-infected IFN-α/βγR^−/− mice.

To evaluate the relevance of pathological changes accompanying JUNV infection of IFN-α/βγR^−/− mice to the pathogenesis of AHF in humans, comparative histologic examinations were performed by a pathologist who did not know the genotype of the animals.

The most prominent histopathological changes were seen in the hearts and brains of IFN-α/βγR^−/− mice by day 13 or 14 p.i., whereas animals examined at day 10 p.i. showed fewer lesions. By day 13 or 14 p.i., all IFN-α/βγR^−/− animals (n = 6) demonstrated multifocal myocarditis that ranged from expansion of the myocardial interstitium by mononuclear inflammatory cells without overt myocytic necrosis to confluent areas of mixed inflammatory infiltrates associated with infarct-like zones of myocardial necrosis (Fig. 3R). It is striking that myocarditis has been reported for humans with AHF (5). The brains of all six IFN-α/βγR^−/− mice examined at day 13 or 14 p.i. showed a moderate to intense leptomeningitis, with perivascular cuffing of neutrophils mixed with lymphocytes (Fig. 3B). The leptomeningitis or meningoencephalitis observed in the brains of JUNV-infected IFN-α/βγR^−/− mice was described in about one quarter of autopsied human cases (5). Wild-type and mock-infected IFN-α/βγR^−/− mice showed no histologic evidence of myocarditis (Fig. 3Q, S, and T) or inflammatory pathology in the brain (Fig. 3A, C, and D).

FIG. 3.

Comparative histopathology of JUNV-infected mice. Representative histopathologic changes of alpha/beta and gamma interferon receptor-deficient mice (Hybrid/IFN-α/βγR^−/−) or wild-type (Hybrid) mice infected with JUNV in comparison with the corresponding noninfected animals. Mice were challenged with a 1 × 10⁴ PFU dose of JUNV. Animals were humanely killed upon development of severe disease such as weight loss exceeding 20%. Representative hematoxylin-and-eosin-stained sections are shown. Original magnifications, ×10 (brain, liver, spleen, and kidney sections) and ×20 (heart sections). Bars, 500 μm (A to P) and 100 μm (Q to T). (A to D) Brain sections showing leptomeningitis in JUNV-infected, interferon receptor null mice. (E to H) Liver sections, showing large periportal and perivenular inflammatory lesions in JUNV-infected, interferon receptor null mice. (I to L) Spleen sections, showing effacement of normal splenic architecture in JUNV-infected, interferon receptor null mice 14 days p.i. (M to P) Kidney sections showing patchy interstitial inflammation in the cortex in JUNV-infected, interferon receptor null mice. (Q to T) Heart sections showing marked myocarditis in interferon receptor null mice.

The spleens of IFN-α/βγR^−/− JUNV-infected mice showed prominent alterations in microarchitecture, including an increase in white pulp volume with expansion of the periarteriolar lymphoid sheath (Fig. 3J). However, no frank necrosis or white pulp lymphoid depletion has been seen in IFN-α/βγR^−/− JUNV-infected mice as occurs in humans with AHF (5, 9, 17). The spleens of wild-type and mock-infected IFN-α/βγR^−/− mice showed distinct splenic architectures (Fig. 3I, K, and L).

The livers of infected wild-type mice showed rare intrasinusoidal aggregates of inflammatory cells but no hepatocyte damage and no portal inflammation. Infected IFN-α/βγR^−/− mice showed profound inflammatory changes, with large periportal and perivenular mixed inflammatory infiltrates with associated hepatocellular necrosis/apoptosis (Fig. 3F). Liver lesions are a common finding in humans with AHF.

The kidneys of IFN-α/βγR^−/− mice showed small, scattered foci of lymphohistiocytic infiltration on day 10 p.i. and confluent lesions expanding the cortical interstitium by day 13 or 14 p.i. (Fig. 3N). Lesions consisted of granulocytes and nuclear debris, which, together with effacement of the normal tubulointerstitial architecture, suggested necrosis. These lesions were absent in the kidneys of wild-type animals and mock-infected IFN-α/βγR^−/− mice. The kidney lesions described in humans did not resemble the patchy interstitial mononuclear cell infiltrates seen in the cortices of JUNV-infected IFN-α/βγR^−/− mice. However, viral particles and antigen were seen in tubular epithelial cells of kidneys of autopsied human patients (4). Hemorrhagic lesions were not seen in any murine organ.

On balance, the results of our histologic studies suggest that JUNV organ and tissue tropism may be similar in humans and IFN-α/βγR^−/− mice but that the inflammatory reaction is more prominent in the murine model reported here. Moreover, we were able to show for the first time the presence of disseminated infection with the development of severe disease after inoculation of JUNV in mice lacking a functional interferon system. These animals can provide a safer and more cost-effective alternative for testing antiviral agents against JUNV in the BSL-4 laboratory.