Pathogenicity testing of influenza candidate vaccine viruses in the ferret model

Abstract

The development of influenza candidate vaccine viruses (CVVs) for pre-pandemic vaccine production represents a critical step in pandemic preparedness. The multiple subtypes and clades of avian or swine origin influenza viruses circulating world-wide at any one time necessitates the continuous generation of CVVs to provide an advanced starting point should a novel zoonotic virus cross the species barrier and cause a pandemic. Furthermore, the evolution and diversity of novel influenza viruses that cause zoonotic infections requires ongoing monitoring and surveillance, and, when a lack of antigenic match between circulating viruses and available CVVs is identified, the production of new CVVs. Pandemic guidelines developed by the WHO Global Influenza Program govern the design and preparation of reverse genetics-derived CVVs, which must undergo numerous safety and quality tests prior to human use. Confirmation of reassortant CVV attenuation of virulence in ferrets relative to wild-type virus represents one of these critical steps, yet there is a paucity of information available regarding the relative degree of attenuation achieved by WHO-recommended CVVs developed against novel viruses with pandemic potential. To better understand the degree of CVV attenuation in the ferret model, we examined the relative virulence of six A/Puerto Rico/8/1934-based CVVs encompassing five different influenza A subtypes (H2N3, H5N1, H5N2, H5N8, and H7N9) compared with the respective wild-type virus in ferrets. Despite varied virulence of wild-type viruses in the ferret, all CVVs examined showed reductions in morbidity and viral shedding in upper respiratory tract tissues. Furthermore, unlike the wild-type counterparts, none of the CVVs spread to extrapulmonary tissues during the acute phase of infection. While the magnitude of virus attenuation varied between virus subtypes, collectively we show the reliable and reproducible attenuation of CVVs that have the A/Puerto Rico/9/1934 backbone in a mammalian model.

Keywords: Ferret; Influenza; Pathogenicity; Vaccine viruses.

Fig. 1. Generation, validation, and safety testing of candidate vaccine viruses

Influenza virus reassortant candidate vaccine viruses (CVVs) were created according to WHO guidance for development of vaccine reference viruses (WHO, 2005b). Methodology is described in detail in Sections 2 and 3.

Fig. 2. Mean nasal wash viral titer of wild-type and CVVs in ferrets

Three ferrets per group were inoculated with 10⁶ PFU or EID₅₀ of virus (grey bars, wild-type virus; black bars, CVV). Nasal washes were collected on alternate days p.i. Mean titers for each group are shown as log₁₀ EID₅₀/ml + standard deviation (SD) with the exception of H7N9 viruses which are shown as log₁₀ PFU/ml. The limit of detection was 1.5 log₁₀ EID₅₀/ml or 1 log₁₀ PFU/ml. *, p < 0.05 between wild-type and CVV groups.

Fig. 3. Systemic replication of wild-type and CVVs in ferrets

Three ferrets per group were inoculated with 10⁶ PFU or EID₅₀ of virus (grey bars, wild-type virus; black bars, CVV), and tissues were collected 3 days p.i. for virus titration. Mean titers for each group are shown as log₁₀ EID₅₀/g or ml + SD with the exception of H7N9 viruses which are shown as log₁₀ PFU/g or ml. All titers are reffective of virus detection in 3/3 ferrets unless specified otherwise. NT, nasal turbinates; Tr, trachea; Lg, lung; BnOB, olfactory bulb; BnA, anterior brain; BnP, posterior brain; Int, intestine (pooled). The limit of detection was 1.5 log₁₀ EID₅₀/g or ml or 1 log₁₀ PFU/g or ml (all samples are expressed per g with the exception of NT, which is expressed per ml).