Characterization of Effect of Enterovirus D68 in 129/Sv Mice Deficient in IFN-α/β and/or IFN-γ Receptors

Abstract

Enterovirus D68 (EV-D68), a respiratory RNA virus in the family Picornaviridae, is implicated as a potential etiological agent for acute flaccid myelitis in preteen adolescents. The absence of a specific therapeutic intervention necessitates the development of an effective animal model for EV-D68. The AG129 mouse strain, characterized by the double knockout of IFN-α/β and IFN-γ receptors on the 129 genetic background, has been proposed as a suitable model for EV-D68. The goals of this study were to assess the effect of a nonmouse-adapted EV-D68 strain (US/MO/14-18947, NR-49129) in AG129 (IFN-α/β and IFN-γ receptors null), A129 (IFN-α/β receptor null), G129 (IFN-γ receptor null), and the 129 background strain (129S2/SvPasCrl) when infected intraperitoneally at 10 d of age. Both AG129 and A129 strains demonstrated similar clinical signs (paralysis, paresis, lethargy, dyspnea [characterized by prominent abdominal respiration], and morbidity requiring euthanasia) induced by EV-D68. While G129 and 129S2 strains also exhibited susceptibility to EV-D68, the severity of clinical signs was less than in AG129 and A129 strains, and many survived to the experimental endpoint. Histopathological and immunohistochemical data confirmed EV-D68 tropism for the skeletal muscle and spinal cord and suggest that the dyspnea observed in infected mice could be attributed, in part, to lesions in the diaphragmatic skeletal muscles. These findings contribute valuable insights into the pathogenesis of EV-D68 infection in this mouse model and provide investigators with key information on virus dose and mouse strain selection when using this mouse model to evaluate candidate EV-D68 therapeutics.

Introduction

Enterovirus D68 (EV-D68), initially identified in 1962,¹ is a positive-sense nonenveloped ssRNA virus in the virus family Picornaviridae. Enteroviruses infecting humans are classified into four species, namely Enterovirus A, Enterovirus B, Enterovirus C, and Enterovirus D, with each comprising multiple types.²² EV-D68, a type within Enterovirus D, is spread primarily by the respiratory route. It can cause severe respiratory disease but is also reported increasingly to be associated with acute flaccid myelitis (AFM), a rapidly progressing neurologic condition that causes paralysis, mostly in preteen children.^11,12

Various EV-D68 strains have been isolated, with some strains exhibiting nonvirulence in mouse models, while others, including CA/14-4231, US/MO/14-18947, and US/IL-14-18952, demonstrate virulence.^6,32 The 7.4-kb RNA genome of EV-D68 encodes a polyprotein that undergoes self-digestion to yield 4 structural proteins (VP1, VP2, VP3, and VP4) and 7 nonstructural proteins (2A, 2B, 2C, 3A, 3B, 3C, and an RNA-dependent RNA polymerase). Notably, position 88 of VP3 has been identified as a virulence determinant, with a change from isoleucine to valine attenuating virulence.³²

AFM symptoms include acute onset of flaccid limb weakness or paralysis, cranial nerve dysfunction, and distinctive lesions in the brainstem and central gray matter of the spinal cord identified by MRI.^10,24 Following an EV-D68 respiratory disease outbreak in the United States^11,12,16 and Canada in 2014, there was subsequently a large number of young patients diagnosed with AFM, suggesting a causal relationship. However, the mechanism that causes EV-D68 to trigger AFM is not clear. As monitored by the Centers for Disease Control and Prevention, most of the EV-D68 infection cases in the United States have occurred between August and November in biennial cycles in 2014, 2016, and 2018. Although a peak of EV-D68 infections was not observed in 2020 and 2022, future EV-D68 outbreaks are being closely monitored.³¹

Unfortunately, no approved vaccine or antiviral treatments have been developed so far, and treatments are mostly supportive or symptomatic.⁷ Although some proposed treatments have shown anti–EV-D68 activity in vitro^3,5,20,23,26 and in neonatal mouse models,^4,25 there remains no approved effective treatment for humans. A well-characterized animal model can contribute to the development of antiviral therapies to prevent and manage new EV-D68 outbreaks.

Several mouse strains have been used to study EV-D68 pathogenesis (reviewed in Vermillion and colleagues²⁹). Adult mice are, in general, refractory to EV-D68 infection, often showing minimal clinical signs.^3,8 This can be overcome using neonatal mice, infection with a mouse-adapted virus, or the use of immune-deficient strains.²⁹ The double knockout, immune-deficient strain AG129 is null for receptors of the type I IFN-α/β (A) and type II IFN-γ (G) on the 129 genetic background. Compared with immune-competent mice, AG129 mice display an increased susceptibility to several viruses,^1,27,28 including EV-D68.^3,13 Although others have characterized the pathogenesis of EV-D68 in AG129,^3,8,13 its effect on the single knockout strains, A129 (null for IFN-α/β receptor only) and G129 (null for IFN-γ receptor only), and the background 129 strain have not been reported.

In this study, we compared morbidity and mortality in AG129, A129, G129, and the wild-type 129 genetic background strain using a nonmouse adapted EV-D68 under identical experimental conditions. We hypothesized that the AG129 strain would be the most susceptible to EV-D68, A129 and G129 strains would be intermediate in susceptibility, and the 129 strain would be resistant. In addition, we performed histologic analysis of infected AG129 mice to characterize virus-induced pathology in the spinal cord and muscle. A major goal of the study was to provide investigators data on dose and strain selection when using the AG129 mouse model to evaluate candidate EV-D68 therapeutics.

Materials and Methods

Ethics statement.

All procedures were approved by the Stanford Institutional Animal Care and Use Committee (Administrative Panel for Laboratory Animal Care), and were in compliance with the Guide for the Care and Use of Laboratory Animals.¹⁵ Stanford University’s animal care program is accredited by AAALAC International. Prior to the study we consulted the ARRIVE 2.0 guidelines.¹⁸

Animals.

The AG129 strain^14,27,28 was obtained from Dr. Harry Greenberg (Stanford University). The 129S2/SvPasCrl strain was purchased from Charles River Laboratories (Hollister, CA). A129 and G129 strains were generated by crossing AG129 with 129S2/SvPasCrl and selecting for homozygotes of each knockout. Mice were housed under barrier conditions on a 12-h dark/12-h light cycle at 21.1 °C to 23.3 °C (70 °F to 74 °F) and 30% to 70% relative humidity. Animals were housed in irradiated individually ventilated cages, prebedded with irradiated ⅛-in. corncob or AlphaDri bedding (Innocage^®, Innovive, San Diego, CA), and provided with irradiated Enviro-Dri^® crinkled paper strands for enrichment (Shepherd Specialty Papers, Milford, NJ). Mice were fed an irradiated commercial diet (Teklad 2918, Envigo) and were provided with bottled acidified (pH 2.5 to 3.0) water (Aquavive^®, Innovive, San Diego, CA). Mouse colonies were monitored for adventitious viral, bacterial, and parasitic pathogens by dirty bedding sentinels that were tested quarterly. The sentinels were shown to be free of mouse hepatitis virus, minute virus of mice, murine norovirus, mouse parvovirus, mouse rotavirus (epizootic diarrhea of infant mice), Theiler murine encephalomyelitis virus (GDVII strain), lymphocytic choriomeningitis virus, murine respirovirus (Sendai virus), mouse adenovirus 1 and 2, mousepox virus (ectromelia virus), pneumonia virus of mice, reovirus 3, Mycoplasma pulmonis, Citrobacter rodentium, Corynebacterium kutscheri, Rodentibacter spp., Salmonella spp., Streptococcus pneumoniae, Helicobacter spp., mites, lice, pinworms, Spironucleus muris, and Giardia muris.

Viruses and infection.

EV-D68 viral stock US/MO/14-18947, NR-49129 (Missouri 2014 outbreak strain) was obtained from BEI Resources, National Institute of Allergy and Infectious Diseases, NIH, and propagated and titered by plaque assay on human rhabdomyosarcoma cells (CCL-136, ATCC). Virus stock was diluted in rhabdomyosarcoma tissue culture medium supplemented with 20 mM HEPES. The number of litters and sample sizes of each strain were as follows: AG129, 10 litters, n = 44; A129, 6 litters, n = 20; G129, 8 litters, n = 34; and 129S2, 7 litters, n = 43.

Mice were infected intraperitoneally on the lower right abdominal quadrant on postnatal day 10 (P10) with 10³ to 10⁶ pfu in a volume of 0.1 mL.⁸ Body weights were taken daily and animals were monitored for clinical signs. A clinical score was assigned according to the following criteria: 1, no clinical signs; 2 (mild), gait abnormalities and paresis; 3 (moderate), splayed legs or paralysis but still able to actively move; 4 (severe), lethargy, labored breathing (dyspnea). Clinical signs that prompted early euthanasia included lethargy associated with hypothermia, failure to move when stimulated, or weight loss for 2 consecutive days associated with lethargy and labored breathing, with or without paresis/paralysis. Moribund mice were euthanized with carbon dioxide followed by decapitation according to AVMA Guidelines for the Euthanasia of Animals.⁹

Necropsy and histopathology.

AG129 pups infected with 10⁵ pfu of EV-D68 (n = 11) were euthanized via CO₂ asphyxiation on 4 d postinfection (dpi) (n = 6) or 5 dpi (n = 5) when they showed severe clinical signs. Age-matched, uninfected, control pups (n = 5) were euthanized in parallel. Tissues were collected and immersion-fixed in 10% neutral buffered formalin for at least 72 h. Selected bones (vertebral column, pelvic limbs) were decalcified in Rapid Cal Immuno (BBC Biochemical) for 2 h. Selected tissues (vertebral column, including spinal cord and paravertebral muscles; pelvic limbs, including bones and associated pelvic limb muscles; diaphragm; lungs) were submitted for histology. The vertebral column was transversely sectioned at 2- to 3-mm intervals generating 10 to 14 total evaluable sections of cervical, thoracic, and lumbosacral regions. Additional step sections at 200-µm intervals of the vertebral column were obtained from selected mice (3 total per mouse). Pelvic limbs were disarticulated at the level of the coxofemoral joint and longitudinally bisected. Formalin-fixed tissues were submitted to the Stanford University Animal Histology Service and embedded in paraffin, sectioned at 5 µm, and stained with hematoxylin and eosin. Slides were examined by a board-certified veterinary pathologist (K.M.C.). An ordinal histopathologic grading scale (0 to 4) was designed to assess the distribution of skeletal muscle degeneration and necrosis within the paravertebral muscles, pelvic limbs, and diaphragm. Briefly, a score of 0 indicated no lesions, a score of 1 indicated lesions present in less than 25% of the skeletal muscle, a score of 2 indicated lesions in 25% to 50% of the skeletal muscle, a score of 3 indicated lesions in 50% to 75% of the skeletal muscle, and a score of 4 indicated lesions in more than 75% of the skeletal muscle.

Immunohistochemistry.

Sections (5 µm) of vertebral column, pelvic limb, diaphragm, and lungs were obtained from the formalin-fixed paraffin blocks of selected mice. Slides were deparaffinized and subjected to enzyme-induced antigen retrieval (0.6 U/mL proteinase K in Tris/EDTA buffer, pH 8.0, 5 min, 21 °C). Protein blocking was performed using pH 7.4 TBS (50 mM Tris-Cl, 150 mM NaCl) supplemented with 0.5% Tween 20 and 3% goat serum. EV-D68 VP1 polyclonal antibody (GeneTex, GTX132313) was used at a dilution of 1:400 for formalin-fixed tissues and at 1:200 for formalin-fixed tissues that had undergone decalcification. Tissues were incubated in primary antibody at room temperature for 90 min and rinsed with TBS. Prior to the secondary antibody, endogenous peroxidase was blocked by immersion in 3% hydrogen peroxide in phosphate-buffered saline (pH 7.4) for 10 min, then rinsed 3 × 3 min with TBS-Tween (0.05% Tween 20, pH 7.4). Slides were incubated with secondary antibody (goat anti-rabbit IgG (H+L), HRP conjugated, Columbia BioSci, HRP-114) in a humid chamber for 45 min at room temperature, then rinsed 3 × 3 min in TBS-Tween. The immunoreaction was visualized with the Abcam DAB substrate kit (Abcam, ab64238) and counterstained with Harris hematoxylin (VWR, 10143-608). Slides were dehydrated in ethanol and xylene, then cover-slipped with a nonpolar mountant (Thermo Scientific Cytoseal^TM XYL, 8312-4).

Statistical analysis.

Prism 9.0 software (GraphPad Software, San Diego, CA) was used to generate Kaplan–Meier survival curves of EV-D68–infected mice. Data were analyzed for statistical significance using t tests or 2-way ANOVAs with post hoc t tests. Statistical significance was set at P < 0.05.

Results

EV-D68 infection induces clinical signs and mortality in AG129, A129, G129, and 129S2 mice.

We first investigated the effect of varying doses of EV-D68 on the severity of viral-induced clinical signs in the AG129 strain. Lower doses increased survivability (Figure 1). No pup survived doses given intraperitoneally at 10⁶ pfu/mouse (n = 14) or 10⁵ pfu/mouse (n = 10), with all attaining early euthanasia criteria by 4 or 5 dpi (Figure 1). At 10⁴ pfu/mouse, 10/11 pups (91%) developed mild to moderate clinical signs such as splayed pelvic limbs, gait abnormalities, and pelvic limb paralysis or attained early euthanasia criteria while 1/11 pups (9%) survived until the 14 dpi endpoint. Similarly, at 10³ pfu/mouse, 3/9 (33%) developed mild to moderate clinical signs and 6/9 (67%) survived until the 14 dpi endpoint. Paresis or paralysis was observed more often in the right pelvic limb possibly due to the site of injection (lower right abdominal quadrant).¹³ Of the infected pups that survived through the study period, 71% (5/7) had improvement of clinical signs, for example, limb paralysis improving to paresis/gait abnormalities.

Figure 1.

Kaplan–Meier survival curves of AG129-infected intraperitoneally at P10 with different doses of EV-D68 (US/MO/14-18947) (10³, 10⁴, 10⁵, and 10⁶ pfu/mouse) demonstrating that severe clinical signs leading to euthanasia are dose-dependent. Mice infected with a higher dosage of virus had earlier onsets of severe clinical signs and lower probability of survival. Comparisons of all survival curves were significantly different: 10³ pfu compared with 10⁴ pfu: P = 0.0032; 10⁴ pfu compared with 10⁵ pfu: P = 0.0005; 10⁵ pfu compared with 10⁶ pfu: P = 0.0104.

A129 P10 pups infected intraperitoneally with 10⁶ pfu of EV-D68 had a clinical course similar to AG129, showing a rapid onset of clinical signs by 3 dpi and attaining early euthanasia criteria by 4 dpi (Figure 2A). The AG129 and A129 survival curves at 10⁶ pfu were not statistically different (P = 0.5260). AG129 and A129 survival curves were statistically different at 10⁵ pfu/mouse (Figure 2B, P = 0.0268) in part due to AG129 reaching an earlier euthanasia time point (4 dpi). Additional studies are necessary to determine whether this statistical difference is biologically meaningful. AG129 and A129 survival curves were not significantly different from each other at 10⁴ pfu/mouse (Figure 2C, P = 0.3893).

Figure 2.

Kaplan–Meier survival curves of AG129, A129, G129, and 129S2 mice infected intraperitoneally at P10 with EV-D68 (US/MO/14-18947): (A) 10⁶ pfu/mouse, (B) 10⁵ pfu/mouse, and (C) 10⁴ pfu/mouse. AG129 and A129 mice were highly susceptible to EV-D68 infection, all reaching early euthanasia criteria by 4 dpi at a dose of 10⁶ pfu/mouse (A) and 5 to 6 dpi at 10⁵ pfu/mouse (B). Although AG129 and A129 survival curves were similar at 10⁵ pfu/mouse, they were statistically different (P = 0.0146) due to AG129 reaching an earlier euthanasia endpoint (AG129 MST = 4 dpi, A129 MST = 5 dpi). AG129 and A129 survival curves were not statistically different at 10⁶ (A, P = 0.5260) and 10⁴ (C, P = 0.8259) pfu/mouse. G129 and 129S2 survival curves were not significantly different from each other at all EV-D68 doses tested: (A) 10⁶ pfu P = 0.2226; (B) 10⁵ pfu P = 0.1063; (C) 10⁴ pfu P = 0.3923, and a large percentage of pups of both strains survived until the end of the experiment (14 dpi). AG129 survival curves were statistically significant from G129 at all doses tested: (A) 10⁶ pfu, P < 0.0001; (B) 10⁵ pfu, P < 0.0001; (C) 10⁴ pfu, P = 0.0005.

G129 P10 pups infected intraperitoneally with 10⁶ pfu of EV-D68 started to show signs of lethargy, pelvic limb paresis or paralysis, gait abnormalities, and dyspnea characterized by abdominal breathing on 4 dpi, and 29% (5/17) became moribund on 5 dpi and were euthanized. Similar to AG129 and A129 mice, the right pelvic limb was affected more often than the left. However, despite initially showing clinical signs (paresis, single limb paralysis, gait abnormalities, weight loss), 9/17 (53%) regained weight and limb mobility and survived to 14 dpi, the end of the experiment (Figure 2).

Surprisingly, the 129S2 background strain was also susceptible to EV-D68 at P10. 45% (18/40) of 129S2 mice showed virus-induced symptoms similar to G129 and attained early euthanasia criteria. The G129 and 129S2 survival curves were not significantly different from each other (Figure 2, 10⁶ pfu, P = 0.2226; 10⁵ pfu, P = 0.1063; 10⁴ pfu, P = 0.1063), and 47% (7/15) of G129 and 63% (12/19) of 129S2 at 10⁶ pfu survived the infection until 14 dpi.

EV-D68 exhibits tropism for skeletal muscle in AG129 mice.

To gain insights into what may be the etiology for pelvic limb paralysis and dyspnea, we performed a histologic and immunohistochemical examination of selected tissues of infected AG129 mice. AG129 pups (n = 11) infected with 10⁵ pfu of EV-D68 were euthanized when severe symptoms were observed. Infected pups exhibited polyphasic myodegeneration and necrosis within skeletal muscle of the diaphragm (Figure 3), pelvic limbs (Figure 4), and paravertebral muscles (data not shown). In all locations, skeletal muscles exhibited varying degrees of degeneration (swollen and vacuolated sarcoplasm), necrosis (loss of cross striations, hypereosinophilia, shrunken sarcoplasm, karyorrhectic to pyknotic nuclei), and regeneration (increased basophilia, nuclear rowing, prominent nucleoli, occasional mitotic figures). Moderate numbers of neutrophils and fewer histiocytes were present within regions of myocyte loss and within the endomysial connective tissue.

Figure 3.

Diaphragmatic myodegeneration and necrosis in EV-D68 (US/MO/14-18947)–infected AG129 pups. (A) Degenerate and necrotic myofibers (arrowheads) are surrounded by moderate numbers of neutrophils and fewer histiocytes (asterisk). Hematoxylin and eosin. (B) Normal diaphragmatic skeletal muscle from control pup. Hematoxylin and eosin. (C) Areas of myocyte regeneration were characterized by increased sarcoplasmic basophilia and nuclear rowing (arrowheads). Hematoxylin and eosin. (D) EV-D68 immunohistochemistry. Strong punctate to granular immunoreactivity to EV-D68 was present in degenerate and necrotic myocytes. In all figures, scale bar = 20 μm.

Figure 4.

Pelvic limb myodegeneration and necrosis in EV-D68 (US/MO/14-18947)–infected AG129 pups. (A) EV-D68–infected pups exhibited dropout of skeletal muscle fibers and expansion of the endomysial space (arrowheads). Hematoxylin and eosin, scale bar = 1 μm. (B) Normal pelvic limb skeletal muscle from control pup. Hematoxylin and eosin, scale bar = 1 μm. (C) Higher magnification of boxed region in (A) demonstrating widespread myodegeneration and necrosis of pelvic limb skeletal muscle. Hematoxylin and eosin, scale bar = 50 μm. (D) EV-D68 immunohistochemistry. Strong punctate to granular immunoreactivity to EV-D68 was present in degenerate and necrotic myocytes, scale bar = 50 μm.

Within the paravertebral muscles there were increased levels of muscular damage in the cranial to caudal direction, with lumbosacral paravertebral muscles exhibiting more widespread and severe damage than cervical or thoracic paravertebral muscles (Figure 5). In 8/11 mice, the right pelvic limb exhibited more widespread muscular damage than the left pelvic limb, correlating to the observed clinical predilection for right-sided paresis/paralysis. In all EV-D68–infected mice, more than 50% of the diaphragm exhibited muscular degeneration and necrosis, while 2 mice exhibited damage in more than 75% of the diaphragm. No histologic lesions were noted in any of the control pups.

Figure 5.

Paravertebral muscles in the lumbosacral region are more severely affected compared with the cervical or thoracic regions in EV-D68 (US/MO/14-18947)–infected AG129 mice (n = 11). No lesions were present in control mice (n = 5). Average lesion scores were calculated based on the percentage of affected muscle. A score of 0 indicated no lesions, a score of 1 indicated lesions in less than 25% of the skeletal muscle, a score of 2 indicated lesions in 25% to 50% of the skeletal muscle, a score of 3 indicated lesions in 50% to 75% of the skeletal muscle, and a score of 4 indicated lesions in more than 75% of the skeletal muscle. Cervical: median, 0.67; mean, 0.85; thoracic: median, 1.33; mean, 1.34; lumbosacral: median, 2.33; mean, 2.19. Means for each region were statistically different from each other (P < 0.001, Brown–Forsythe ANOVA).

Immunohistochemistry for EV-D68 confirmed viral presence within skeletal muscle fibers of the diaphragm (Figure 3D), pelvic limbs (Figure 4D), and paravertebral muscles (data not shown). Punctate to granular, dark brown immunoreactivity was present within the sarcoplasm of degenerate to necrotic skeletal muscle fibers. Unaffected skeletal muscle fibers in EV-D68–infected pups lacked immunoreactivity.

EV-D68 infection results in neutrophilic myelitis of the ventral gray matter in AG129 mice.

A total of 9/11 pups infected with EV-D68 exhibited myelitis characterized by low numbers of neutrophils within the ventral gray matter of the spinal cord, occasional karyorrhectic debris, gliosis, and neuropil rarefaction (Figure 6). When present, myelitis was seen sporadically along the length of the spinal cord. Myelitis was most common within the thoracic spinal cord (9/9) and occasionally seen within the lumbosacral (2/9) or cervical spinal cord (1/9). Immunohistochemistry for EV-D68 confirmed viral presence within the neuropil and karyorrhectic debris of affected ventral gray matter horns (Figure 6C).

Figure 6.

Neutrophilic myelitis of the gray matter in EV-D68 (US/MO/14-18947)–infected pups. (A) Transverse section of thoracic spinal cord. Lesions were limited to the ventral gray horns. Hematoxylin and eosin, scale bar = 200 μm. (B) Higher magnification of the boxed region shown in (A) demonstrates neutrophilic inflammation (asterisks), neuropil rarefaction (arrowheads), and scattered karyorrhectic debris. Hematoxylin and eosin, scale bar = 20 μm. (C) EV-D68 immunohistochemistry. Immunoreactivity to EV-D68 (arrowheads) correlated with regions of neutrophilic myelitis, scale bar = 20 μm.

Discussion

A major goal of the study was to provide investigators data on dose and mouse strain selection when using the AG129 mouse model to evaluate candidate EV-D68 therapeutics.

This study demonstrates differential susceptibility and resultant pathologic outcomes to a nonmouse adapted EV-D68 infection in IFN-α and/or IFN-γ receptor null mice on the same 129 genetic background strain, thereby advancing our understanding of the antiviral innate immune response mechanisms. To our knowledge, this is the first study comparing the single and double knockout using a standardized experimental condition. Despite our hypothesis that AG129 would be most susceptible to EV-D68, we found that 10-d-old AG129 and A129 mice were fairly similar in their response at the doses tested, and both were highly susceptible to infection by nonmouse-adapted EV-D68, resulting in severe clinical signs and a high mortality rate. In contrast, G129 and the 129S2 parental wild-type mice were more refractory to EV-D68 infection, resulting in milder symptoms and less mortality. These observations collectively highlight the pivotal role of type I interferon signaling as a dominant antiviral defense mechanism against EV-D68 in murine models. Histopathological and immunohistochemical analyses in the AG129 mouse model show the presence of EV-D68 in skeletal muscle where it causes myodegeneration and necrosis. Overall, AG129 and A129 emerge as reliable models for screening candidate antiviral drugs to anticipate and manage future EV-D68 outbreaks and AFM.

An EV-D68 animal model ideally should recapitulate the genetics and molecular pathogenesis of the disease, should mimic molecular and physiologic clinical features of human infection, and should provide a platform to determine the effectiveness of new therapeutics on a clinical population. Several animal models of EV-D68 have been developed by previous studies: intranasal inoculation of the cotton rat has been proposed as a model for EV-D68, but no clinical signs were reported and its response to challenge were transient and limited to the lungs.¹⁷ In a ferret model, intranasally infected ferrets showed minimal clinical signs of respiratory illness, and lung pathology was observed only in ferrets receiving high viral loads.³⁴ Mouse models are beneficial due to their small size, which allows a larger sample size, numerous genetic mutations, and lower dosages for often limited experimental therapeutics. Several immunocompetent strains (BALB/c, C57BL/6J, FVB/NJ, ICR, KM, NIH, SJL/J, and Swiss Webster)²⁹ and a C57BL/6 strain transgenic for Tg21 and the IFN-α/β receptor knockout³² have been tested as models for EV-D68. Because they were infected by different routes, at different ages, and using different strains of EV-D68, direct comparisons are difficult. The double knockout strain AG129 has been identified as a reliable model for EV-D68 study.^3,8 However, the susceptibility of single knockout strains A129 and G129 to EV-D68 has not been previously explored. Our investigation focused on bridging this gap, providing a comprehensive analysis of the susceptibility profiles of both A129 and G129 strains in comparison to the studied AG129 model. Through this comparative examination, our study contributes valuable insights into the distinct host–virus interactions exhibited by each knockout strain, enhancing the overall understanding of EV-D68 pathogenesis in this mouse model.

We rejected our hypothesis that the immunocompetent 129S2 strain would be resistant to EV-D68 when infected intraperitoneally at P10. To our knowledge, this study is the first to report that inbred 129S2 P10 pups are susceptible, albeit weakly, to EV-D68, aligning it with the outbred ICR strain.³³ At all doses tested, a small but not insignificant percentage of infected 129S2 mice reached early euthanasia criteria.

We have demonstrated dose-dependent different mortality rates of P10, intraperitoneally infected pups for AG129, A129, G129, and 129S2 mice (Figure 2B). The variability in virulence observed among different mouse strains has elucidated the differential roles of interferon receptors in the EV-D68 host defense mechanism. Our results suggest that the IFN-γ receptor knockout is not sufficient to increase EV-D68 susceptibility in 129S2 mice, suggesting a less important role of IFN-γ in EV-D68 virus defense. IFN-γ is a cytokine critical to both innate and adaptive immunity. Functions of IFN-γ include activating macrophages and stimulating natural killer cells and neutrophils.^2,21 Type III interferons (IFN-λ) can also induce antiviral activity.³⁰ Further studies aimed at understanding the role of type II and III interferons in the context of EV-D68 infection would benefit possible antiviral development.

In keeping with previous reports, we confirmed EV-D68 tropism for skeletal muscle^3,13,25 and spinal cord^6,33 as demonstrated by histopathologic and immunohistochemical evidence of degeneration and necrosis in the skeletal muscles of the pelvic limbs, paravertebral muscles, and diaphragm and neutrophilic myelitis within the spinal cord of infected AG129 mice (Figures 3, 4, 5, and 6).

We propose that the primary cause of the gait abnormalities (splayed legs, paresis) and paralysis in EV-D68–infected mice is due to EV-D68–induced lesions in the pelvic limb and paravertebral skeletal muscles. As suggested by Morrey and colleagues,¹³ we infer this is a result of the intraperitoneal infection site, which was performed in the lower right abdomen. In keeping with this, we found that paralysis was commonly observed in the right pelvic limb, that the caudal paravertebral muscles were more affected than the cranial paravertebral muscles (Figure 5), and that the right pelvic limb muscles were more affected than the left muscles (data not shown). We cannot rule out the contribution of the neutrophilic myelitis identified in the spinal cord, although lesions were more commonly identified in the thoracic spinal cord (9/9) and only occasionally in the lumbosacral spinal cord (2/9).

Dyspnea, characterized by prominent abdominal breathing, was a clinical sign in this mouse model. Tachypnea was also reported in EV-D68–infected BALB/c and ICR neonatal mice.²⁵ Although EV-D68 is a respiratory virus, dyspnea has not been reported in most mouse models even if infected intranasally^3,6,8,13,19 and/or using a mouse lung-adapted virus.^3,8,13 An increase in enhanced pause (Penh) measured by plethysmography has been reported^3,8 but the authors did not report any signs of dyspnea. One explanation for absence of dyspnea could be the age at infection (4 to 10 wk),^3,8,19 mouse strain,⁶ or virus isolate. Although interstitial pneumonia was noted in tachypneic BALB/c mice,²⁵ we did not observe lesions in the lungs of dyspneic AG129 pups. On the other hand, we identified lesions in the muscles of the diaphragm (Figure 6). We propose that one possible cause of the dyspnea in the AG129 mouse model is the observed diaphragmatic skeletal muscle fiber lesions compromising its function. However, it cannot be ruled out that lesions in the associated spinal cord or nerves innervating the diaphragm may also be contributory.

Our data suggest that the paralysis and dyspnea observed in the AG129 mouse model does not faithfully replicate the human AFM syndrome. However, this does not lessen the importance of the AG129/A129 mouse model for the development of EV-D68 therapeutics given that the clinical signs in this model are highly predictable and severe. If there is concern of the use of an interferon-deficient strain, our data also show that the wild-type 129S2 strain is also susceptible to EV-D68 under the same experimental conditions, although the results are less predictable. We propose that any compound that ameliorates the clinical course in this model would be a strong candidate to be examined further.

Limitations of this study are that only knockouts of IFN-α/β and IFN-γ receptors on the 129 genetic background were evaluated and that we focused on only the US/MO/14-18947 strain of EV-D68. Given that many mutations are on the C57BL/6 genetic background, a similar study examining EV-D68 in C57BL/6 IFN-α/β and/or IFN-γ receptor knockouts would be informative.

In conclusion, we show that AG129 and A129 are 2 suitable strains to further assess nonmouse adapted EV-D68–induced clinical signs and can be used to screen and evaluate candidate antivirals to prevent and treat future EV-D68 outbreaks. To our knowledge, this is the first report that the single knockout, A129, but not G129, shows similar susceptibility to EV-D68. We also report, to our knowledge for the first time, that 129S2 is susceptible to high doses of EV-D68 under the conditions tested. Finally, we show that EV-D68–infected mice demonstrate dyspnea that we propose is due to lesions in the diaphragm’s skeletal muscle. Future studies should aim to elucidate the mechanisms of viral entry within the CNS and the factors determining regional susceptibility within the spinal cord. Understanding these aspects is crucial for developing targeted therapeutic and preventive strategies against EV-D68 infections and associated neurologic complications.

Conflict of Interest

The authors have no conflicts of interest to declare.

Funding

This research was supported by the Discovery Innovation Fund (to C.M.N. and K.M.C.), NIH grants R01 AI130123 (to J.E.C.), 1 R01 AI153169 (to J.E.C.), and 75N93020C00054 (to J.S.G.), and by the Department of Comparative Medicine.