Increasing the breadth and potency of response to the seasonal influenza virus vaccine by immune complex immunization

Abstract

The main barrier to reduction of morbidity caused by influenza is the absence of a vaccine that elicits broad protection against different virus strains. Studies in preclinical models of influenza virus infections have shown that antibodies alone are sufficient to provide broad protection against divergent virus strains in vivo. Here, we address the challenge of identifying an immunogen that can elicit potent, broadly protective, antiinfluenza antibodies by demonstrating that immune complexes composed of sialylated antihemagglutinin antibodies and seasonal inactivated flu vaccine (TIV) can elicit broadly protective antihemagglutinin antibodies. Further, we found that an Fc-modified, bispecific monoclonal antibody against conserved epitopes of the hemagglutinin can be combined with TIV to elicit broad protection, thus setting the stage for a universal influenza virus vaccine.

Trial registration: ClinicalTrials.gov NCT01967238.

Keywords: CD23; TIV; immune complex; sialylated Fc; universal influenza vaccine.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

Model for high-affinity antibody production through B-cell CD23 engagement. (A) Signaling through B-cell receptor (BCR) triggers proliferation of low and high-affinity B cells. (B) Coengagement of CD23 with BCR induces increased expression of the inhibitory FcγRIIB, which, in turn, elevates the threshold of BCR affinity that is required for cell activation and results in selection of higher-affinity B cells.

Fig. 2.

Increased breadth of anti-TIV antibodies by administration of sialylated TIV immune complexes. Breadth of the anti-HA response was enhanced by TIV+sIgG (sIC) administration over TIV alone or TIV+aIgG (aIC). (A–E) Binding of serum IgGs was measured by ELISA to homologous antigen (TIV) or the homologous H1 HA component (Cal09 H1) and was not significantly different between groups. Binding to a heterologous H1 (Bris07), H3 (Bris07), or H5 (Hunan02) was enhanced by administration of sIC over aIC or TIV alone. Stars indicate significance level of difference in binding between the serum IgGs of mice in the sIC and aIC (black) or sIC and TIV (red) immunization groups (five mice per group). Significance at each dilution was determined by unpaired T test (Prism 7). *P ≤ 0.05, ****P ≤ 0.00005.

Fig. 3.

Increased potency of anti-TIV antibodies by administration of sialylated TIV immune complexes. Purified serum IgGs (at 7.5 or 75 µg/mL) from mice vaccinated with TIV sIC or aIC were incubated with 5 mLD50 of maNL09 (H1N1) for 30 min at 37 °C. The virus and IgG mixture was then used to challenge BALB/c mice, intranasally (four mice per group). Mice receiving virus incubated with sIC or aIC IgGs at 75 μg/mL IgG were protected from weight loss. At the 7.5 μg/mL dose, only sIC-elicited IgGs protected from weight loss. Negative control mice received PBS.

Fig. 4.

F241A Fc domain and FI6 and 4G05 Fab domains combine to elicit enhanced anti-HA IgGs. (A) F241A mAb+H1 HA (PR8) enhances the affinity of anti-H1 response to a level that is comparable with the sialylated Fc IgG. Thus, the F241A Fc domain was selected for use in subsequent experiments. (B) The anti-H1 stalk titer, a measurement of antibody quantity and affinity, was not increased by every mAb that was tested (Fig. S1), but was increased by immunization with a FI6/4G05 bispecific mAb+H1 HA. (C) A F241A bispecific mAb comprised of 4G05 (28) and FI6 (29) enhances the affinity of the antistalk response when delivered with HA protein. The increase in affinity was not observed in CD23 knockout mice. Affinity was determined by 7M urea ELISA (31). Data are expressed as serum IgG bound to HA following 7M urea treatment/IgG bound without 7M urea treatment. Significance at each dilution was determined by unpaired T test (Prism 7). *P ≤ 0.05, **P ≤ 0.005, ***P ≤ 0.0005, ****P ≤ 0.00005.

Fig. S1.

Mice were immunized with ICs composed of H1 hemagglutinins and the H1-reactive mAbs, (A) C05 (27) (+PR8 HA), (B) 2B06, and (C) 2G02 (28) (+CAL09 HA), that were generated with the F241A Fc domain. Immune sera elicited by these immune complexes were pooled and screened for binding to the H1 HA stalk domain. We did not observe enhanced anti-H1 stalk antibody responses using these ICs. Clear circles, HA/F241A IC; black circles, HA alone.

Fig. S2.

(A) The influenza hemagglutinin can be divided into a globular head domain (blue) and a stalk domain (green). (B) A bispecific mAb comprised of Fab domains from mAbs 19–4G05 (4G05) (28) and FI6 (29) was generated. The 4G05 is a broadly reactive anti-H1 head mAb, while FI6 is a broad, anti-group 1 and group 2 mAb that binds to the HA stalk domain. HA monomer in A and IgG representations in B are not drawn to scale.

Fig. 5.

IgGs elicited by the F241A bispecific mAb/H1 HA ICs protect against an H1 or H5 virus in vivo. (A and B) IgG elicited by PR8 H1 IC immunization protected against lethal challenge with a heterologous H1 virus [mouse adapted A/Netherlands/602/09 (maNL09) or (C and D) lethal H5 subtype virus challenge (reassortant virus expression the HA and NA of A/Viet Nam/1203/2004 (rVN04) (H5N1) with the internal segments of PR8 (H1N1) influenza A virus (50)] (five mice per group).

Fig. 6.

Breadth and potency of the anti-HA TIV response was enhanced by administration with the F241A bispecific mAb. (A–C) IgGs elicited by TIV+F241A bispecific mAb had increased binding activity to homologous H1 HA (Cal09 H1), heterologous H1 (Bris07), or H5 (Hunan02) subtype hemagglutinins. (D) In addition, the F241A/TIV immune complexes elicited more potent antistalk IgGs, measured by in vitro neutralization using a virus that expresses the cH6/1 hemagglutinin (H1 HA stalk domain with an H6 subtype globular head domain) (10). The H1 stalk domain is highly conserved among group 1 influenza virus subtypes, including H5. Endpoint microneutralization titers against cH6/1 virus are shown. (E and F) IgGs elicited by TIV+F214A bispecific mAb protected mice against lethal H5N1 (rVN04) challenge (five mice per group). Significance at each dilution was determined by unpaired T test (Prism 7). *P ≤ 0.05, **P ≤ 0.005.