top of page

Partial Cross-Protection of Canine Influenza Vaccines

Influenza A viruses, the causative agent of the annual flu, infect birds and some mammals. The viruses are categorized into different subtypes, depending on the types of hemagglutinin (HA) and neuraminidase (NA) proteins located on the surface of the viral envelope. The trivalent flu vaccine contains HA components of the three circulating strains. These viruses undergo significant genetic drift, necessitating an annual update of the flu vaccine. The effectiveness of the vaccine depends on how well the viruses in the vaccine match those in circulation, but antibodies made in response to the vaccine may still provide cross-protection to related viruses.

Dr. Scott Hensley’s lab investigates antigenic drift of influenza viruses and factors that affect vaccine responsiveness. Elinor Willis, a V.M.D.-Ph.D. student in the lab, recently published a paper in Vaccine investigating cross-protection of the influenza antibody response in dogs. In 1999, an interspecies transmission of the equine H3N8 virus was first found in racing greyhounds and later became endemic in the United States. Canine H3N8 (cH3N8) virus causes dog flu that is usually mild, although some may develop a severe and fatal illness. A vaccine for cH3N8 became available in 2009 and has been shown to be effective in preventing severe disease caused by cH3N8 infection. Another subtype, H3N2, was isolated in dogs in South Korea in 2006-2007 and was introduced in the Unites States in 2015, causing a large outbreak in the Midwest. There are currently two conditionally licensed cH3N2 vaccines. Since the canine H3N2 (cH3N2) strain shares the H3 hemagglutanin (HA) with H3N8, Elinor sought to determine whether the cH3N2 and cH3N8 strains are antigenically related, and whether the cH3N8 vaccine provides cross-protection against the cH3N2 strain.


Most vaccines are designed against the HA protein, which consists of a stalk and a globular head domain. A comparison of the amino acid sequences of prototypical cH3N2 and cH3N8 strains showed differences throughout the five antigenic sites in the globular head of H3, while the stalk domain was more conserved. cH3N2 and cH3N8 are also predicted to have different N-linked glycosylation patterns in HA. Using sera from mice vaccinated with either the cH3N8 or the cH3N2 vaccine, she conducted an enzyme-linked immunosorbent assay (ELISA) to assess antibody reactivity and found that the cH3N8 vaccine induced an antibody response against the cH3N2 virus, and vice-versa. Vaccines may also contain internal viral proteins, so Elinor conducted ELISAs using recombinant HA proteins to test whether the antibodies bound to HA. Surprisingly, the antibodies in sera from cH3N8-vaccinated mice did not bind well to the HA recombinant protein from the cH3N2 virus. ELISAs using chimeric H5/H3 HA protein with an H5 globular head and an H3 stalk or an H5/H1 HA protein showed that antibodies from both cH3N8 and cH3N2 vaccines bound the H3 stalk domain but not the H1 stalk domain. Both the cH3N8 and the cH3N2 vaccines elicited antibody responses against internal viral proteins and epitopes on the HA stalk domain of cH3N2 viruses, but only the cH3N2 vaccine elicited an antibody response against the epitopes on the HA head domain of the cH3N2 virus.

This study explored the benefits of the commercially available canine H3N8 influenza vaccine against an avian H3N2 influenza virus that has been circulating in dogs. Willis et al. tested the H3N8 vaccine in a mouse model of influenza infection. They demonstrated that mice previously vaccinated with H3N8 had limited viral replication in the lung during subsequent H3N2 infection. Future studies conducted in dogs may provide evidence on the efficacy of this vaccine against related influenza viruses. The figure above outlines the potential mechanism of this partial protection.

A) Vaccination with the commercially available H3N8 vaccine.

B) Generation of antibodies against conserved viral glycoprotein stalk regions and internal proteins.

C) Cross-reactive antibodies elicited by vaccination can limit infection by H3N8 viruses



Since cH3N2 infections do not cause morbidity or mortality in mice, Elinor compared vaccine efficacy by measuring viral loads in the lungs. Six out of nine mice vaccinated with cH3N2 had no detectable lung viral titers after infection, but seven out of eight mice vaccinated with cH3N8 had viral titers, though lower compared to those in unvaccinated mice. This suggests that the cH3N8 vaccine helps limit cH3N2 replication in the lung, but does not provide sterilizing immunity against cH3N2 infection.


Vaccines take time to generate, and it can be useful to determine whether existing vaccines are effective against new influenza subtypes. While these studies were not performed in dogs, mice have been used as a model system to investigate antibody responses against other influenza viruses. Elinor’s results show that the cH3N8 vaccine does not provide complete protection against the cH3N2 virus, and the conditionally licensed cH3N2 vaccine should be used in areas with both cH3N8 and cH3N2 viruses.


Willis, E., Parkhouse, K., Krammer, F., and Hensley, S. E. Canine H3N8 influenza vaccines partially protect mice against the canine H3N2 strain currently circulating in the United States.Vaccine 34: 5483-5487, Sep 2016.

bottom of page