Anti-Influenza Hyperimmune Immunoglobulin Enhances Fc-Functional Antibody Immunity During Human Influenza Infection

Abstract

Background: New treatments for severe influenza are needed. Passive transfer of influenza-specific hyperimmune pooled immunoglobulin (Flu-IVIG) boosts neutralizing antibody responses to past strains in influenza-infected subjects. The effect of Flu-IVIG on antibodies with Fc-mediated functions, which may target diverse influenza strains, is unclear.

Methods: We studied the capacity of Flu-IVIG, relative to standard IVIG, to bind to Fcγ receptors and mediate antibody-dependent cellular cytotoxicity in vitro. The effect of Flu-IVIG infusion, compared to placebo infusion, was examined in serial plasma samples from 24 subjects with confirmed influenza infection in the INSIGHT FLU005 pilot study.

Results: Flu-IVIG contains higher concentrations of Fc-functional antibodies than IVIG against a diverse range of influenza hemagglutinins. Following infusion of Flu-IVIG into influenza-infected subjects, a transient increase in Fc-functional antibodies was present for 1-3 days against infecting and noninfecting strains of influenza.

Conclusions: Flu-IVIG contains antibodies with Fc-mediated functions against influenza virus, and passive transfer of Flu-IVIG increases anti-influenza Fc-functional antibodies in the plasma of influenza-infected subjects. Enhancement of Fc-functional antibodies to a diverse range of influenza strains suggests that Flu-IVIG infusion could prove useful in the context of novel influenza virus infections, when there may be minimal or no neutralizing antibodies in the Flu-IVIG preparation.

Figure 1.

Greater Fc gamma receptor (FcγR) cross-linking by anti-hemagglutinin (HA) antibodies in influenza-specific hyperimmune immunoglobulin (Flu-IVIG). Recombinant soluble (rs) FcγR dimer ELISAs were used to compare Flu-IVIG made during 2013 to standard IVIGs made in 2008 (prior to pandemic H1N1 [pH1N1]), 2010, and 2016. Dimeric rsFcγRIIIa (A) and dimeric rsFcγRIIa (B) binding antibodies (optical density [OD]) against pH1N1 recombinant hemagglutinin (rHA) protein (A/California/07/2009 strain) are shown in 4 different IVIG preparations. The (1 / effective concentration 50 [EC₅₀]) × 10³ values for dimeric rsFcγRIIIa (C) and dimeric rsFcγRIIa (D) binding antibodies against diverse rHA proteins were calculated for Flu-IVIG and 3 standard IVIG preparations. Fold increase in the concentration of FcγR cross-linking antibodies in Flu-IVIG relative to standard IVIGs was calculated by dividing the (1 / EC₅₀) × 10³ value for Flu-IVIG by the (1 / EC₅₀) × 10³ value for standard IVIG [eg, Flu-IVIG (1 / EC₅₀) × 10³ = 23.55 (μg/mL)⁻³ and IVIG 2008 (1 / EC₅₀) × 10³ = 4.56 (μg/mL)⁻³, so 23.55 / 4.56 = 5.2-fold increase]. Fold increases in the concentration of FcγR cross-linking antibodies in Flu-IVIG relative to standard IVIGs prepared between 2008 and 2016 are shown. All IVIG preparations were also tested for binding against an irrelevant simian immunodeficiency virus glycoprotein 120; the mean background was calculated and multiplied by 3 to give a threshold of detection represented by the dotted line. Error bars represent standard deviation of the mean of duplicate wells in the FcγR dimer ELISA.

Figure 2.

Enhanced Fc-functional anti-influenza antibodies in influenza-specific hyperimmune immunoglobulin (Flu-IVIG). NK cell activation and ADCC assays were used to study Flu-IVIG compared to standard IVIGs. Antibody-dependent natural killer (NK) cell activation, as measured by the proportion of an NK-like cell line expressing the degranulation marker CD107a by flow cytometry, was assessed to pandemic H1N1 (pH1N1) recombinant hemagglutinin (rHA) protein (A/California/07/2009 strain) (A) and pH1N1-infected A549 cells (A/Auckland/01/2009 strain) (B). (C) ADCC was measured by lactate dehydrogenase (LDH) release from pH1N1-infected A549 cells (A/Auckland/01/2009 strain). NK cell activation (A and B) or target cell killing (C) in the absence of IVIG was also tested in each assay and multiplied by 3 to give a threshold of detection represented by the dotted line. (D) The (1 / effective concentration 50 [EC₅₀]) × 10³ values for ADCC antibodies against diverse rHA proteins and A/Auckland/01/2009(H1N1)-infected target cells were calculated for Flu-IVIG and 3 standard IVIG preparations as described in Figure 1. Fold increases in the concentration of ADCC antibodies in Flu-IVIG relative to standard IVIGs prepared between 2008 and 2016 are shown.

Figure 3.

Influenza-specific hyperimmune immunoglobulin (Flu-IVIG) contains hemagglutinin (HA) stem-specific antibodies. Immunoglobulin G (IgG) enzyme-linked immunosorbent assays (ELISAs), recombinant soluble Fc gamma receptor (rsFcγR) dimer ELISAs, and natural killer (NK) cell activation assays were used to compare Flu-IVIG made during 2013 to standard IVIGs made in 2008 (prior to pandemic H1N1 [pH1N1]), 2010, and 2016. IgG antibodies (optical density [OD]) (A), dimeric rsFcγRIIIa binding antibodies (OD) (B), and NK cell activating antibodies (C) against the pH1N1 HA stem are shown in the 4 different IVIG preparations. All IVIG preparations were also tested for binding against an irrelevant simian immunodeficiency virus glycoprotein 120; the mean background was calculated and multiplied by 3 to give a threshold of detection represented by the dotted line. Error bars represent standard deviation of the mean of duplicate wells in the FcγR dimer ELISA.

Figure 4.

Increased Fc gamma receptor (FcγR) IIIa cross-linking antibodies against hemagglutinin (HA) following infusion of influenza-specific hyperimmune immunoglobulin (Flu-IVIG) into subjects with influenza infection. The recombinant soluble (rs) FcγRIIIa dimer enzyme-linked immunosorbent assays were used to examine serial plasma samples from 24 influenza-infected subjects (15 of whom had pandemic H1N1 [pH1N1] infections) randomized to receive Flu-IVIG or placebo. (A) Means of dimeric rsFcγRIIIa binding antibodies against pH1N1 recombinant HA (rHA) protein (at a 1:40 dilution of plasma) in the 15 pH1N1-infected subjects are shown. Preinfusion antibody levels for each individual patient were subtracted from that patient’s follow-up samples to give change from baseline at every time point, and the adjusted mean changes from baseline (preinfusion) are graphed. Mean optical density (OD) of preinfusion samples was 0.22 for controls and 0.22 for the Flu-IVIG group. (B) Geometric mean titers (GMTs) of dimeric rsFcγRIIIa binding antibodies against pH1N1 rHA in the 15 pH1N1-infected subjects are graphed. GMTs for postinfusion time points are shown adjusted for baseline or preinfusion titer; the mean GMTs for preinfusion samples were 59 for the control group and 62 for the Flu-IVIG group. (C) GMTs of dimeric rsFcγRIIIa binding antibodies against pH1N1 rHA in 24 influenza-infected subjects (15 had pH1N1, 3 had H3N2, and 6 had B infection) are shown. As before, postinfusion GMTs were adjusted for preinfusion titers, which were 80 for the control group and 65 for the Flu-IVIG group. All plasma samples were also tested for dimeric rsFcγRIIIa binding antibodies against an irrelevant simian immunodeficiency virus type 1 glycoprotein 120, and background was subtracted for each plasma sample. Error bars represent standard error of the mean. ***P < .001.

Figure 5.

Enhanced influenza-specific antibody-mediated natural killer (NK) cell activation following influenza-specific hyperimmune immunoglobulin (Flu-IVIG) infusion. NK cell activation assays measuring CD107a expression were used to study serial plasma samples from 15 subjects with pandemic H1N1 (pH1N1) infections randomized to receive either Flu-IVIG or placebo. Mean changes from preinfusion or baseline for NK cell activating antibodies to pH1N1 recombinant hemagglutinin (rHA) protein (A) and geometric mean titers (GMTs) of NK cell activating antibodies to pH1N1-infected A549 cells (B) are shown. Data in A and B are adjusted for preinfusion antibody levels. All plasma samples were also tested for antibody-mediated NK cell activation against an irrelevant simian immunodeficiency virus type 1 glycoprotein 120, and background was subtracted for each individual sample. Error bars represent standard error of the mean. **P < .01, ***P < .001.

Figure 6.

Greater Fc gamma receptor (FcγR) IIa cross-linking antibodies against hemagglutinin (HA) following infusion of influenza-specific hyperimmune immunoglobulin (Flu-IVIG) into subjects with influenza infection. A recombinant soluble (rs) FcγRIIa dimer enzyme-linked immunosorbent assay was used to examine serial plasma samples from 15 subjects with pandemic H1N1 (pH1N1) infections randomized to receive Flu-IVIG or placebo. As in Figure 4A, preinfusion antibody levels for each individual patient were subtracted from that patient’s follow-up samples to give change from baseline at every time point, and the adjusted mean changes from baseline are shown for dimeric rsFcγRIIa binding antibodies (optical density [OD]) against pH1N1 recombinant HA (at a 1:40 plasma dilution). All plasma samples were also tested for dimeric rsFcγRIIa binding antibodies against an irrelevant simian immunodeficiency virus type 1 glycoprotein 120, and background was subtracted for each plasma sample. Error bars represent standard error of the mean. ***P < .001.

Figure 7.

Increased Fc gamma receptor (FcγR) IIIa cross-linking antibodies against hemagglutinins (HAs) from diverse strains of influenza virus following influenza-specific hyperimmune immunoglobulin (Flu-IVIG) infusion. The recombinant soluble (rs) FcγRIIIa dimer enzyme-linked immunosorbent assays were used to study serial plasma samples from 15 subjects with pH1N1 infections randomized to receive either Flu-IVIG or placebo. Mean changes from preinfusion or baseline for dimeric rsFcγRIIIa binding antibodies (optical density [OD]) to recombinant HAs (rHAs) are shown. The rHAs represent the following strains: H2N2 A/Japan/305/1957 and H5N1 A/Vietnam/1194/2004 (A), H3N2 A/Texas/50/2012 and H4N6 A/Swine/Ontario/01911-/1999 (B), and B/Phuket/3073/2013 (C). Error bars represent standard error of the mean. *P < .05, **P < .01, ***P < .001.