Different immunity elicited by recombinant H5N1 hemagglutinin proteins containing pauci-mannose, high-mannose, or complex type N-glycans

Abstract

Highly pathogenic avian influenza H5N1 viruses can result in poultry and occasionally in human mortality. A safe and effective H5N1 vaccine is urgently needed to reduce the pandemic potential. Hemagglutinin (HA), a major envelope protein accounting for approximately 80% of spikes in influenza virus, is often used as a major antigen for subunit vaccine development. In this study, we conducted a systematic study of the immune response against influenza virus infection following immunization with recombinant HA proteins expressed in insect (Sf9) cells, insect cells that contain exogenous genes for elaborating N-linked glycans (Mimic) and mammalian cells (CHO). While the antibody titers are higher with the insect cell derived HA proteins, the neutralization and HA inhibition titers are much higher with the mammalian cell produced HA proteins. Recombinant HA proteins containing tri- or tetra-antennary complex, terminally sialylated and asialyated-galactose type N-glycans induced better protective immunity in mice to lethal challenge. The results are highly relevant to issues that should be considered in the production of fragment vaccines.

Figure 1. 3D structure model.

(A) Insect cell expressed HA attached with pauci-mannose N-glycans and (B) mammalian cell expressed HA attached with complex-type N-glycans were created by the crystal structure of HA (A/Vietnam/1194/04, PDB ID: 2IBX) and Glyprot.

Figure 2. Recombinant HA protein expression and purification

. Soluble HA proteins were constructed using the HA cDNA sequences of H5N1 A/Thailand/1(KAN-1)/2004. The PQRERRRKKRG multibasic protease cleavage site between HA₁ and HA₂ was mutated to PQRETRG to retain the uncleaved protein. The C-terminal sequence was fused with a GCN4pII leucine zipper sequence and ended with a His-tag to facilitate purification. Recombinant H5HA proteins were purified from the culture supernatants of Sf9, Mimic, and CHO cells and analyzed by SDS-PAGE with (A) Coomassie Blue staining and (B) western blotting with anti-H5 antibodies. Proteins were also analyzed following treatment with (C) Endo H and (D) PNGase F.

Figure 3. HILIC-HPLC glycan patterns.

Purified recombinant HA proteins were placed on SDS-PAGE gels and deglycosylated using PNGase F. Glycans were isolated from protein backbones and analyzed using HILIC-HPLC to identify individual glycan structures. A dextran ladder was created with HILIC-HPLC and used as a standard to provide glucose unit (GU) values for individual peaks recognized in glycan samples from recombinant HA proteins. Shown are HA N-glycan profiles of (A) Sf9-rHA, (B) Mimic-rHA, and (C) CHO-rHA. Red inverted triangles, fucose; blue squares, N -acetylglucosamine; green circles, mannose; yellow circles, galactose; pink diamonds, sialic acid.

Figure 4. Antibody responses elicited by recombinant HA proteins in immunized mice.

Female BALB/c mice (6-8 wk old, five mice per group) were intramuscularly immunized with 20 µg of recombinant HA proteins over a 3-week interval; blood was collected 2 wk after the second inoculation. Total anti-HA IgG antibody titers were measured using ELISAs with (A) Sf9-rHA or (B) CHO-rHA. Total IgG antibody avidities were measured after treatment with 6 M urea followed by ELISA binding to (C) Sf9-rHA or (D) CHO-rHA. Data represent geometric mean ± standard error; results were analyzed using one-way ANOVAs with Tukey's tests, with statistical significance at p<0.05 as asterisks indicated.

Figure 5. Antibody isotypes/subtypes and antibody-secreting B cells.

The antibody isotypes of IgM and IgA and the IgG1, IgG2a, IgG2b and IgG3 subtypes in sera were analyzed using ELISA plates coated with (A) Sf9-rHA or (B) CHO-rHA. Mouse splenocytes were collected 3 wk following second inoculations, and then reacted with (C) Sf9-rHA or (D) CHO-rHA to measure the numbers of antibody-secreting B cells in spleens; ELISPOT assays were performed using HRP-conjugated anti-mouse IgG antibodies. Data represent mean ± standard deviation; results were analyzed using one-way ANOVAs with Tukey's tests, with statistical significance at p<0.05 as asterisks indicated.

Figure 6. T cell responses in splenocytes and T-cell stimulation by antigen-presenting dendritic cells.

Splenocytes were added to each well in 96-well plates (5×10⁵ cells/well) and stimulated with 1 µg/ml pooled peptides (15-mer overlapping 8 amino acids) spanning the HA₁ of H5HA (A/Viet Nam/1203/2004) to determine (A) IFN-γ- and (B) IL-4-secreting T cells using ELISPOT assays. Bone marrow-derived dendritic cells (DCs) were treated with LPS and recombinant HA proteins and then fixed with paraformaldehyde followed by quenching with glycine. Pre-treated DCs were co-incubated with splenocytes from mice immunized with Sf9-rHA for 2 d. Antigen presentation was determined by measuring (C) IFN-γ- and (D) IL-2-secreting T cells using ELISPOT assays. DCs pretreated with LPS and pulsed with PBS were used as a negative control. Data represent mean ± standard deviation. Results were analyzed using two-tailed Student’s t tests with statistical significance at p<0.05 as asterisks indicated.

Figure 7. Neutralizing antibodies against homologous and heterologous clades of H5N1 viruses.

(A) Neutralizing antibodies were measured as the reduced luciferase activity of H5HA-pseudotyped particles (H5pp) following the incubation of sera with homologous (KAN-1; clade 1) H5-pseudotyped particles. (B) HI antibody titers against the KAN-1 virus were determined by reacting influenza VLPs with turkey red blood cells. Cross-clade neutralizing antibodies against (C) Cclade 2.3.4 of A/Anhui/1/2005 and (D) clade 2.3.2 of A/Hubei/1/2010 were also determined using H5pp assays. Data represent mean ± standard deviation. Results were analyzed using one-way ANOVAs with Tukey's tests, with statistical significance at p<0.05 as asterisks indicated.

Figure 8. Protective immunity in mice.

Female BALB/c mice (6−8 wk old, five mice per group) were intramuscularly immunized with 20 µg of recombinant HA proteins over a 3-week interval. Immunized mice were intranasally challenged with 10× MLD₅₀ H5N1 virus (NIBRG-14) 3 wk after their second inoculations. Survival rates (A) and body weights (B) were recorded for 14 d. Body weight loss over 25% was used as an end-point. Body weight loss is presented as mean ± standard deviation.