H5N1 influenza vaccine induces a less robust neutralizing antibody response than seasonal trivalent and H7N9 influenza vaccines

Abstract

Conventional inactivated avian influenza vaccines have performed poorly in past vaccine trials, leading to the hypothesis that they are less immunogenic than seasonal influenza vaccines. We tested this hypothesis by comparing the immunogenicity of the H5N1 and H7N9 vaccines (avian influenza vaccines) to a seasonal trivalent inactivated influenza vaccine in naïve ferrets, administered with or without the adjuvants MF59 or AS03. Vaccine immunogenicity was assessed by measuring neutralizing antibody titers against hemagglutinin and neuraminidase and by hemagglutinin -specific IgG levels. Two doses of unadjuvanted vaccines induced low or no HA-specific IgG responses and hemagglutination-inhibiting titers. Adjuvanted vaccines induced comparable IgG-titers, but poorer neutralizing antibody titers for the H5 vaccine. All adjuvanted vaccines elicited detectable anti- neuraminidase -antibodies with the exception of the H5N1 vaccine, likely due to the low amounts of neuraminidase in the vaccine. Overall, the H5N1 vaccine had poorer capacity to induce neutralizing antibodies, but not HA-specific IgG, compared to H7N9 or trivalent inactivated influenza vaccine.

Fig. 1

Immunization schedule and vaccine groups. Each ferret received 7.5 µg of hemagglutinin (HA) protein per vaccine strain per dose. Graphics depicted in this figure were generated in-house

Fig. 2

Immunogenicity profile in ferrets vaccinated with unadjuvanted, MF59 or AS03-adjuvanted TIV, H7N9 or H5N1 vaccines. Sera was collected after the first dose given 22 days post-vaccination (DPV) and the second dose (B + 21 dpv) and tested by a hemagglutination-inhibition (HAI), b ELISA for antigen-specific IgG and c microneutralization (MN) assay. Antibody titers are expressed as mean log₁₀-antibody titer, with the error bars representing the standard deviation. Dashed line indicates the limit of assay detection; Y = 1 for HAI and MN assays and Y = 2 for ELISA. Statistically significant differences were tested by ANOVA of log-transformed antibody titers for all antigens and vaccine groups and post hoc tests with Bonferroni adjustments for multiple comparisons. Asterisk denotes p-value < 0.05, **p < 0.01, ***p < 0.001. The complete results are provided in Supplemental Fig. 1

Fig. 3

Antibody titers against NA in ferrets vaccinated with unadjuvanted, MF59 or AS03-adjuvanted TIV, H7N9 or H5N1 vaccines. Sera was collected after the first dose (22 dpv) and second dose (B + 21 dpv) and tested by enzyme-linked lectin assay (ELLA) against homologous antigens. Antibody titers are expressed as mean log₁₀-antibody titer, with the error bars representing the standard deviation. Dashed line (at Y = 1) indicates the limit of assay detection. Statistically significant differences were tested by ANOVA of log-transformed antibody titers for all antigens and vaccine groups and post hoc tests with Bonferroni adjustments for multiple comparisons. Asterisk denotes p-value < 0.05, **p < 0.01, ***p < 0.001

Fig. 4

Neuraminidase (NA) contents in the H5N1, TIV and H7N9 vaccines. a Western blots of vaccine preparations. Native (N) and denatured (D) vaccine preparations were probed with polyclonal serum raised against the specified NA subtype. Arrowheads indicate the expected size of NA in the vaccines (red) and controls (white). The N1-controls (lane 3,4,5) used were a concentrated and purified reverse-genetics (rg) reassortant virus containing A/Viet Nam/1203/2004 (H5N1) HA and NA, the reassortant H6N1 (A/Viet Nam/1203/2004, NA) used in the ELLA assay, and a purified N1 protein without the ectodomain derived from A/California/04/2009 (H1N1). The N2-control used was a concentrated and purified A/Brisbane/10/2007 (H3N2) virus and the N9-control used was a concentrated and purified rg-reassortant virus containing A/Anhui/1/2013 (H7N9) HA and NA. Blots were overexposed to capture faint NA-bands in the H5N1 vaccine. Samples in lanes 1–7, 8–10 and 11–13 were processed under the same conditions respectively and run as three separate blots. No quantitative comparison is made amongst the blots. b Neuraminidase activity in TIV, H7N9 and H5N1 vaccines as detected by enzyme-linked-lectin assay (ELLA)

Fig. 5

Cross-reactive antibodies against the N1 of the H5 strain elicited by TIV vaccination. Sera from the TIV vaccinated ferrets were tested for cross-reactive antibodies against N1 from the H5 virus. Antibody titers are expressed as mean log₁₀-antibody titer, with the error bars representing the standard deviation. Dashed line (at Y = 1) indicates the limit of assay detection. Statistically significant differences were tested by ANOVA of log-transformed antibody titers for all antigens and vaccine groups and post hoc tests with Bonferroni adjustments for multiple comparisons. Asterisk denotes p-value < 0.05, **p < 0.01, ***p < 0.001

Fig. 6

Ratio of neutralizing antibodies as measured by a hemagglutination-inhibition (HAI) and b microneutralization (MN) assays per influenza-specific IgG titer. Proportion was determined based on area-under-the-curve (AUC) of the respective antibody titer induced over time, with the assumption of a linear antibody increase between the sampling periods. Statistical significance was determined by ANOVA, with Bonferroni’s correction applied for multiple pairwise comparisons. Asterisk denotes p-value < 0.05, **p < 0.01, ***p < 0.001