Insights into the Acquisition of Virulence of Avian Influenza Viruses during a Single Passage in Ferrets

Abstract

Circulating avian influenza viruses pose a significant threat, with human infections occurring infrequently but with potentially severe consequences. To examine the dynamics and locale of the adaptation process of avian influenza viruses when introduced to a mammalian host, we infected ferrets with H5N1 viruses. As expected, all ferrets infected with the human H5N1 isolate A/Vietnam/1203/2004 showed severe disease and virus replication outside the respiratory tract in multiple organs including the brain. In contrast infection of ferrets with the avian H5N1 virus A/Chicken/Laos/Xaythiani26/2006 showed a different collective pattern of infection; many ferrets developed and cleared a mild respiratory infection but a subset (25-50%), showed extended replication in the upper respiratory tract and developed infection in distal sites. Virus from these severely infected ferrets was commonly found in tissues that included liver and small intestine. In most instances the virus had acquired the common virulence substitution PB2 E627K but, in one case, a previously unidentified combination of two amino acid substitutions at PB2 S489P and NP V408I, which enhanced polymerase activity, was found. We noted that virus with high pathogenicity adaptations could be dominant in an extra-respiratory site without being equally represented in the nasal wash. Further ferret passage of these mutated viruses resulted in high pathogenicity in all ferrets. These findings illustrate the remarkable ability of avian influenza viruses that avoid clearance in the respiratory tract, to mutate towards a high pathogenicity phenotype during just a single passage in ferrets and also indicate a window of less than 5 days in which treatment may curtail systemic infection.

Keywords: H5N1 influenza virus; avian influenza; ferret; mammalian adaptation; systemic infection.

Figure 1

Weight loss induced by infection with avian and human H5N1 isolates. (A) Ferrets F1–F10 were infected with the avian H5N1 virus A/Laos (encoding PB2 627E) and (B) ferrets F11–F20 were infected with the human H5N1 isolate A/Viet (encoding PB2 627K). Ferrets were weighed immediately prior to virus inoculation on day 0 and again immediately prior to euthanasia on the indicated day of sampling. Values represent the percentage change from the initial ferret weights. * Indicates that a change in weight was not detected.

Figure 2

Immunostaining of H5N1 virus in liver samples of infected ferrets. Virus (red) was visualised in (A) liver of A/Laos-inoculated ferret F8 on day 5 PI.; (B) liver of A/Laos-inoculated ferret F10 on day 7 PI.; (C) and (D) liver of A/Viet-inoculated ferret F18 on day 5 PI. Boxed area in (C) denotes image shown in (D). Magnifications (A) 20×; (B) 20×; (C) 5×; (D) 20×.

Figure 3

Weight changes following inoculation of ferrets with second passage H5N1 viruses. (A) A/Laos (PB2 627E); (B) A/F8/Liv (PB2 627E); (C) A/F10/Liv (PB2 627K); (D) A/F8/NW (PB2 627E/K); (E) A/F10/NW (PB2 627E)) H5N1 viruses. Two ferrets from each group were sampled on days 5 and 7 PI. * Ferrets were euthanised earlier than scheduled, at the humane endpoint.

Figure 4

The effect of individual mutations at NP 408I or PB2 489P on pathogenicity. Percentage weight change following intranasal infection of ferrets with (A) rgA/L/PB2/627K; (B) rgA/V/PB2/627E; (C) rgA/L/NP/408I; or (D) rgA/L/PB2/489P viruses. * Ferrets were euthanised earlier than scheduled, at the humane endpoint.

Figure 5

The effect of PB2 627K and the PB2 489P, NP 408I combination on viral polymerase activity in mammalian cells. Different combinations of (A) the pHWLaos and pHWViet PB2, PB1, PA and NP genome segment plasmids (designated as LPB2, VPB2, LPB1, VPB1, LPA, VPA, LNP and VNP, respectively); as well as (B) each of the plasmids encoding the single AA changes, pHWLaos PB2 489P (LPB2 489P), pHWLaos PB2 627K (LPB2 627K), pHWLaos NP 408I (LNP 408I) and pHWViet PB2 627E (VPB2 627E), were co-transfected into 293T cells along with plasmids pPOL-NP-LUC and pRL-TK. Following incubation at 37 °C for 24 h in a humidified incubator the relative luciferase activity of each polymerase complex was assessed using the Dual-Luciferase Reporter Assay System (Promega). Values (mean of three replicates ± standard error) are shown relative to the polymerase activity of the luciferase control (pRL-TK + pPOL-NP-LUC only) which represents one unit. Luciferase activity of each polymerase complex genetic constellation was compared to the luciferase activity of the native A/Laos polymerase constellation (LPB2, LPB1, LPA and LNP) using a one-way analysis of variance test. * P ≤ 0.05.