VP1 is the primary determinant of neuropathogenesis in a mouse model of enterovirus D68 acute flaccid myelitis

Abstract

Enterovirus D68 (EV-D68) is an emerging pathogen that can cause severe respiratory and neurologic disease [acute flaccid myelitis (AFM)]. Intramuscular (IM) injection of neonatal Swiss Webster (SW) mice with US/IL/14-18952 (IL52), a clinical isolate from the 2014 EV-D68 epidemic, results in many of the pathogenic features of human AFM, including viral infection of the spinal cord, death of motor neurons, and resultant progressive paralysis. In distinction, CA/14-4231 (CA4231), another clinical isolate from the 2014 EV-D68 outbreak, does not cause paralysis in mice, does not grow in the spinal cord, and does not cause motor neuron loss following IM injection. A panel of chimeric viruses containing sequences from IL52 and CA4231 was used to demonstrate that VP1 is the main determinant of EV-D68 neurovirulence following IM injection of neonatal SW mice. VP1 contains four amino acid differences between IL52 and CA4231. Mutations resulting in substituting these four amino acids (CA4231 residues into the IL52 polyprotein) completely abolished neurovirulence. Conversely, mutations resulting in substituting VP1 IL52 amino acid residues into the CA4231 polyprotein created a virus that induced paralysis to the same degree as IL52. Neurovirulence following infection of neonatal SW mice with parental and chimeric viruses was associated with viral growth in the spinal cord.

Importance: Emerging viruses allow us to investigate mutations leading to increased disease severity. Enterovirus D68 (EV-D68), once the cause of rare cases of respiratory illness, recently acquired the ability to cause severe respiratory and neurologic disease. Chimeric viruses were used to demonstrate that viral structural protein VP1 determines growth in the spinal cord, motor neuron loss, and paralysis following intramuscular (IM) injection of neonatal Swiss Webster (SW) mice with EV-D68. These results have relevance for predicting the clinical outcome of future EV-D68 epidemics as well as targeting retrograde transport as a potential strategy for treating virus-induced neurologic disease.

Keywords: AFM; EV-D68; VP1; neurovirulence.

Fig 1

Following IM injection IL52, but not CA4231, induces paralysis in MAVS ko mice. Two-day-old MAVS ko mice were infected with 1000 TCID50 EV-D68 strains IL52 (N = 3) and CA4231 (N = 4) by IM inoculation. The graph shows the average paralysis score of infected mice at DPI 4 and 8. The paralysis score of individual mice was based on a previously established scoring system where each limb is given a score of 0 (normal function)–3 (complete paralysis). Error bars are SEM. Statistical analysis was performed using PRISM (unpaired t-test).

Fig 2

IL52 induces paralysis in mice that is associated with viral growth in the spinal cord and loss of motor neurons. Two-day-old SW mice were infected with 1000 TCID50 EV-D68 strains IL52 (N = 17) and CA4231 (N = 15) by IM inoculation. (A) The graph shows the average paralysis over time of mice infected with IL52 (black circles) and CA4231 (gray circles). The paralysis score of individual mice was based on a previously established scoring system where each limb is given a score of 0 (normal function)–3 (complete paralysis). Error bars are SEM. (B) The scatter plot shows the paralysis scores for individual IL52 (black circles, N = 17) and CA4231 (gray circles, N = 15)-infected mice at DPI 10. The mean paralysis score is shown. Error bars are SEM. Statistical analysis was performed using PRISM (unpaired t-test). (C) In separate experiments, the injected muscle and spinal cords were harvested from IL52 (N = 7) and CA4231 (N = 6)-infected mice at DPI 6. The scatter plot shows titers for individual mice as well as the average viral load per gram of tissue for each group. Error bars are SEM. Statistical analysis was performed using PRISM (unpaired t-test), ns = not significant. (D and E) Spinal cords were harvested from mock-, CA4231-, and IL52-infected mice (N = 6). Immunohistochemical analysis of spinal cord sections demonstrated that the spinal cord motor neurons (large staining cells in white box) are present in mock- and CA4231-infected animals but are completely absent from the anterior horns of mice infected with IL52. The images (D) are representative sections. Graph (E) shows the motor neuron counts for individual mice as well as the average (for each group) number of motor neurons in the right (R) and left (L) lumbar enlargements (N = 6). Error bars are SEM. Statistical analysis was performed comparing the total number of motor neurons (both enlargements) using PRISM (unpaired t-test).

Fig 3

The non-structural proteins do not determine neurovirulence following infection of mice with IL52. This figure contains data from parental viruses (Fig. 2) that is presented for comparison. (A) IL52 and CA4231 have 13 nucleotide differences in the 5’UTR and 1 nucleotide difference in the 3’UTR. (B) The IL52 and CA4231 polyproteins have 12 amino acid differences: 2 in VP3; 4 in VP1; 1 in 2A, 2B, and 3A; and 3 in 3D. (C) Diagram of parental strains (IL52 and CA4231), engineered chimeras [4231(4241)IL52 and 4231(3220)IL52], and viruses with individual CA4231 amino acid residues (2A, 2B, 2A and 2B, and 3A) in an IL52 background. Individual amino acid differences between IL52 and CA4231 are shown by black arrows. Proteins and UTR sequences from IL52 are shown in black. Proteins and UTR sequences from CA4231 are shown in gray. Individual mutations where CA4231 amino acid residues replace IL52 residues are marked with a gray line in an otherwise black segment. The white regions have no amino acid differences between IL52 and CA4231 (VP4, VP2, 2C, 3B, and 3C). Two-day-old SW mice were infected by IM injection with 1000 TCID50 EV-D68 strains IL52 (N = 17), CA4231 (N = 15), 4231(4241)IL52 (N = 16), 4231(3220)IL52 (N = 24), 2A/IL52 (N = 8), 2B/IL52 (N = 8), 2A2B/IL52 (N = 11), and 3A/IL52 (N = 8). The average paralysis score (DPI 10) and SEM are shown to the right of the genome diagrams. (D) The scatter plot shows the individual paralysis scores of each mouse as well as the average paralysis for each group. IL52-infected mice are represented by black circles, CA4231-infected mice are represented by gray circles, and chimera/mutant-infected mice are represented by black diamonds. Error bars are SEM. Statistical analysis was performed using PRISM (unpaired t-test). Paralysis scores for 2A/IL52, 2B/IL52, 2A2B/IL52, and 3A/IL52 were not significantly different for those of IL52.

Fig 4

VP1 is the main determinant of neurovirulence following infection of mice with EV-D68. This figure contains data from parental viruses (Fig. 1) that is presented for comparison. (A) Diagram of engineered and parental (same data as shown in Fig. 1 but included as a comparison) viruses. Engineered viruses include viruses with individual CA4231 amino acid residues in VP1 (VP1/IL52 contains 4 CA4231 amino acid residues) and VP3 (VP3/IL52 contains 2 CA4231 residues), or the CA4231 5’UTR (UTR/IL52) in an IL52 background. Individual amino acid differences between IL52 and CA4231 are shown by black arrows. Proteins and UTR sequences from IL52 are shown in black. Proteins and UTR sequences from CA4231 are shown in gray. Individual mutations where CA4231 amino acid residues replace IL52 residues are marked with a gray line in an otherwise black segment. The white regions have no amino acid differences between IL52 and CA4231 (VP4, VP2, 2C, 3B, and 3C). Two-day-old SW mice were infected by IM injection with 1000 TCID50 EV-D68 strains IL52 (N = 17), CA4231 (N = 15), VP1/IL52 (N = 16), VP3/IL52 (N = 16), and UTR/IL5 (N = 24). The average paralysis score at DPI 10 and SEM are shown to the right of the genome diagrams. (B) The scatter plot shows the paralysis score of individual mice at DPI 10 as well as the average paralysis score. IL52-infected mice are represented by black circles, CA4231-infected mice are represented by gray circles, and chimera/mutant-infected mice are represented by black diamonds. Error bars are SEM. Statistical analysis was performed using InStat (ANOVA). Paralysis scores for VP3/IL52 and UTR/IL52 were not significantly different from those of IL52. (C and D) In a separate experiment spinal cord and the injected muscle from infected mice were harvested at DPI 6. The scatter plots show viral loads for individual mice as well as the average virus loads per gram of tissue in the spinal cord (C) and muscle (D) for IL52, CA4231, VP1/IL52, VP3/IL52, and UTR/IL5 (N = 8 for all viruses). IL52-infected mice are represented by black circles, CA4231-infected mice are represented by gray circles, and chimera/mutant-infected mice are represented by black diamonds. Error bars are SEM. Statistical analysis was performed using InStat (ANOVA). Spinal cord titers for VP3/IL52 and UTR/IL52 were not significantly different for those of IL52. Muscle titers for VP1/IL52 and UTR/IL52 were not significantly different for those of IL52.

Fig 5

IL52 sequences confer neurovirulence to CA4231. This figure contains data from parental viruses (Fig. 1) for comparison. (A) Diagram of parental viruses and engineered chimeras and viruses with specific IL52 nucleotide and amino acid residues in a CA4231 background. Individual amino acid differences between IL52 and CA4231 are shown by black arrows. Proteins and UTR sequences from IL52 are shown in black. Proteins and UTR sequences from CA4231 are shown in gray. Individual mutations where CA4231 amino acid residues replace IL52 residues are marked with a black line in an otherwise gray segment. The white regions have no amino acid differences between IL52 and CA4231 (VP4, VP2, 2C, 3B, and 3C). Two-day-old SW mice were infected by IM injection with 1000 TCID50 EV-D68 strains IL52 (N = 17), CA4231 (N = 15), IL52(3220)4231 (N = 11), UTRVP3VP1/4231 (N = 24), and VP1/4231 (N = 17). The average paralysis score at DPI 10 and SEM are shown to the right of the genome diagrams. (B) The scatter plot shows the paralysis scores for each mouse as well as the average paralysis score for each group at DPI 10. IL52-infected mice are represented by black circles, CA4231-infected mice are represented by gray circles, and chimera/mutant-infected mice are represented by black diamonds. Error bars are SEM. Statistical analysis was performed using InStat (ANOVA). Paralysis values for UTRVP3VP1/4231 and VP1/IL52 were not significantly different from those of IL52.

Fig 6

Increased neurovirulence in IL52(3220)4231 reflects higher paralysis scores at DPI 3 and higher spinal cord titers at DPI 6. This figure contains data from parental viruses (Fig. 1) as a comparison. Two-day-old SW mice were infected with 1000 TCID50 EV-D68 strains IL52 (N = 17) and IL52(3220)4231 (N = 11) by IM inoculation. (A) The graph shows the average paralysis of mice infected with IL52 (black circles) and IL52(3220)4231 (gray circles). The paralysis score of individual mice was based on a previously established scoring system where each limb is given a score of 0 (normal function)–3 (complete paralysis). Error bars are SEM. (B) Scatter plot showing individual paralysis scores for IL53 (black circles) and IL52(3220)4231 (black diamonds) as well as average scores for the group at DPI 3. Error bars are SEM. Statistical analysis was performed using PRISM (unpaired t-test). (C) Scatter plot showing individual viral spinal cord titers for IL52 (black circles) and IL52(3220)4231 (black diamonds), as well as the average titer for each group. Error bars are SEM. Statistical analysis was performed using PRISM (unpaired t-test).