The authors have declared that no competing interests exist.
Conceived and designed the experiments: BB VK CT MVB TK RV RHEF JG. Performed the experiments: BB VK CT HJWMK JJO. Analyzed the data: BB VK CT MVB HJWMK JJO JJ AMO. Contributed reagents/materials/analysis tools: JG JJO JJ AMO. Wrote the paper: BB WK RV RHEF.
Human monoclonal antibodies have been identified which neutralize broad spectra of influenza A or B viruses. Here, we dissect the mechanisms by which such antibodies interfere with infectivity. We distinguish four mechanisms that link the conserved hemagglutinin (HA) epitopes of broadly neutralizing antibodies to critical processes in the viral life cycle. HA-stem binding antibodies can act intracellularly by blocking fusion between the viral and endosomal membranes and extracellularly by preventing the proteolytic activation of HA. HA-head binding antibodies prevent viral attachment and release. These insights into newly identified ways by which the human immune system can interfere with influenza virus infection may aid the development of novel universal vaccines and antivirals.
Influenza viruses continue to be a major cause of morbidity and mortality due to shortcomings of currently available vaccines and antivirals. Despite the well-established role of neutralizing antibodies in the defense against influenza virus infection
Neutralizing antibodies mainly target the hemagglutinin (HA) protein, the major envelope glycoprotein of influenza viruses. The HA protein is synthesized as a single precursor protein (HA0) and requires cleavage by host serine proteases into two disulfide-linked subunits, HA1 and HA2, for the virus to be infectious
The vast majority of neutralizing antibodies in infected or vaccinated individuals interferes with attachment of the virus to cellular receptors by binding to exposed, highly variable loops that surround the receptor binding site. Antibodies binding to these regions are typically strain-specific and immunity following natural exposure or vaccination is mostly restricted to closely related strains. However, in the last five years, several human antibodies with remarkably broad neutralizing activity against influenza virus have been generated and characterized. Most of these broadly neutralizing antibodies (bnAbs), such as CR6261, F10, CR8020, FI6, and CR9114, were shown to bind to epitopes in the HA stem which are highly conserved among various influenza virus subtypes and have heterosubtypic neutralizing activity
Stem-binding neutralizing antibodies have been postulated to inhibit the fusion process based on their interaction with the HA2 subunit and lack of activity in hemagglutination-inhibition (HAI) assays, which specifically detect antibodies that interfere with attachment of the virus to sialic acid receptors. Indirect evidence supporting this notion comes from biochemical studies showing that such antibodies can block the conformational changes of recombinant HA required for membrane fusion
(
The finding that stem-binding bnAbs reach late endosomes in complex with the virus is congruent with the assumption that such antibodies can prevent infection by blocking fusion of the viral and endosomal membranes. To directly observe the interference of viral fusion by bnAbs, a single particle fusion assay was applied (
(
The inhibition of the fusion between the virus and the endosome is a mechanism shared by all neutralizing stem binding bnAbs described to date. Inhibiting the cleavage of HA0 into HA1 and HA2 fragments removes the fusogenic potential of HA and is a second mechanism adding to the potency of some of the stem binding Abs. Stem-binding bnAbs CR8020 and FI6 recognize epitopes which partially overlap with the fusion peptide and bind close to the cleavage site of HA
(
Head-binding neutralizing antibodies are well-documented to prevent viral attachment to the receptor. However, we have recently described two bnAbs, CR8033 and CR8071, which bind to the globular head of influenza B HA and are able to inhibit viral egress
(
Confirmation that head-binding antibodies inhibit egress comes from Scanning EM (SEM) images showing that whereas separate budding particles are present at the surface of infected cells in the presence of stem-binding antibody CR6261, large aggregates of particles are visible in the presence of each of the head-binding antibodies (
We hypothesized that HA head-binding antibodies inhibit egress by cross-linking of newly formed virions to each other and to HA on the cell membrane. In line with this hypothesis, the presence of the monovalent Fab fragments of CR8057, CR8033 and CR8071 had no effect on the amount of HA in the supernatant of cells infected with H3N2 and influenza B virus, respectively (
The
(
Suspension PER.C6® (sPER.C6®) cells
Purified wild type influenza viruses A/Puerto Rico/8/1934 (H1N1) and A/Aichi/1968-X31 (6∶2 reassortant of A/Puerto Rico/8/1934 with the HA and NA segments of A/Aichi/1968 (H3N2)) propagated in eggs were purchased from Charles River Laboratories and used for live cell imaging. Stock samples were certified to contain 2 mg of protein per mL and stored at −80°C. A/New Caledonia/20/1999 (H1N1), A/NYMC/X-181 (6∶2 reassortant of A/Puerto Rico/8/1934 with the HA and NA segments of A/California/07/2009 (H1N1)), A/Puerto Rico/8/1934 (H1N1), A/New Caledonia/20/1999 (H1N1), A/Brisbane/59/2007 (H1N1), A/Wisconsin/67/2005 (H3N2), A/Aichi/2/1968-X31 (H3N2), A/NYMC/X-161B (A/Puerto Rico/8/1934 with the HA and NA segments of A/Wisconsin/67/2005 (H3N2)), and B/Florida/04/2006 were grown by infecting sPER.C6® cells with virus at MOI 1×10−4 in infection medium (AEM and VP-FSM (2∶1), supplemented with 2.6 mM L-glutamine and 3 µg/mL trypsin (all reagents from Invitrogen)). After 72 h of incubation, virus containing cell culture supernatant was harvested by centrifugation at 4000 g for 10 min. Virus aliquots were stored at −80°C. For colocalisation and entry studies in live cells, A/New Caledonia/20/1999 and A/NYMC/X-161B were purified by ultracentrifugation at 27,000 rpm for 2 h at 4°C through a 25% sucrose cushion. The virus pellet was resuspended in NTE buffer (150 mM NaCl, 10 mM Tris, 1 mM EDTA), pH 7.4 overnight at 4°C before aliquotting and storage at −80°C.
Uncleaved viruses were produced by infecting sPER.C6® cells with cleaved virus at MOI 2 for 2 h in infection medium without trypsin. Cells were subsequently washed extensively with 10% FBS in PBS and incubated in infection medium in the absence of trypsin. Virus supernatant was harvested by centrifugation at 4000 g for 10 min. All incubations were done at 35°C, 10% CO2, on a shaking platform. Uncleaved status of HA was confirmed by Western blotting after probing with H1-HA or H3-HA specific polyclonal serum and infection assays to confirm the absence of infection without prior treatment with 5 µg/mL trypsin for 30 min at 37°C.
All viruses were specifically titrated to reach >90% infection in each of the experimental conditions. Controls confirmed the successful infection in every experiment.
Fully human IgG1 antibodies CR6261, CR8020, CR8033, CR8057, CR9020, CR11054, and CR11055 were constructed and expressed as described previously (Ekiert et al., 2011). Fab fragments were obtained by IdeS digestion of antibodies, followed by purification via protein G (GE Healthcare), cation exchange (MonoS, GE Healthcare), and gel filtration (Superdex200, GE Healthcare). All antibodies and Fab fragments were more than 97% pure and monomeric. Influenza A nucleoprotein (NP) specific monoclonal mouse antibody was obtained from Abbiotec (clone 5D8) and for influenza B NP from Santa Cruz (sc-52027). Goat F(ab′)2 anti-mouse- or anti-human Alexa Fluor conjugated secondary antibodies (Invitrogen) were used for fluorescent imaging at 2 µg/mL. HA specific rabbit polyclonal serum for Immunoblot analysis was obtained from Protein Sciences. Secondary HRP-coupled anti-rabbit F(ab′)2-fragment were purchased from Jackson Immuno Research Laboratories (111-036-047). Polyclonal sheep sera directed against B/Florida/4/2006 (07/356, sheep 478 and 479) were obtained from the NIBSC and derived from sheep immunized with the respective purified HA. HA-specific polyclonal serum was derived from mice immunized intramuscularly three times at a 3-week interval with 50 µg plasmid DNA encoding full-length A/Brisbane/59/07 HA, codon-optimized for mammalian expression, mixed with 50 µg plasmid DNA encoding murine Granulocyte Macrophage-Colony Stimulating Factor (GM-CSF).
In all experiments antibodies were either used at a range of concentrations or at sufficiently high concentration to neutralize the virus under the given experimental settings. This was confirmed by neutralization controls in every experiment. Imaging also confirmed that the used antibody concentrations were sufficient to binding nearly 100% of viral particles including infectious and potentially non-infectious particles (
Purified and concentrated viruses were diluted in HNE buffer (5 mM Hepes, 140 mM NaCl, 0.2 mM EDTA, pH 7.4) for labeling. The lipophilic fluorescent dye, Octadecyl Rhodamine B chloride (R18, Molecular Probes) dissolved in DMSO or DMSO alone as a mock labeled control was added to the samples to a final dye concentration of 1–2 µM and 0.4–0.5% DMSO. The samples were mixed for 2–3 h at room temperature, protected from light. Unincorporated dye was removed by passing the virus-dye solution through a PD-10 desalting column (GE Healthcare). Fractions containing labeled virus were pooled and labeling verified by fluorescence microscopy.
To confirm that labeling did not affect the infectivity of viruses, labeled- and mock-labeled virus samples were compared in imaged based infection assays (
For imaging studies, HA-specific monoclonal antibodies were fluorescently labeled according to manufacturer's guidance with the amine reactive dyes (Molecular Probes) Alexa Fluor 488 (AF488) or Alexa Fluor 647 (AF647). Briefly, dye dissolved in DMSO was added to antibodies diluted in sodium bicarbonate buffer to a basic pH. For each antibody different dye concentrations where tested to avoid over-labeling. Contents were mixed and incubated for ∼2 h protected from light. Free dye was removed from the sample by desalting and buffer exchange using PD-10 sephadex G-25 columns (GE Healthcare). Antibodies were labeled with 3–8 dyes per IgG molecule.
The biological activity of all labeled antibodies was confirmed and compared to unlabeled antibodies in viral neutralization assays before they were used in imaging experiments. Only batches of labeled antibodies showing less than 2 fold differences in titer where used.
MDCK cells were seeded on the day of experiment at 40,000 cells/well into 96-well flat bottom plates. Antibodies were serially diluted, mixed with an equal volume of viral inoculum and incubated for 2 h at 37°C in medium (DMEM supplemented with 2 mM L-glutamine and 3 µg/mL trypsin-EDTA). The mixture (∼100 TCID50/well) was then added to confluent MDCK monolayers in quadruplicate. Cells were cultured for 72 h before supernatant was added to an equal volume of 1% Turkey red blood cells and incubated for 1 h at room temperature in a 96-well V-bottom plate. The absence of hemagglutination was defined as protection. Titers were determined using the Spearman-Kärber formula.
Virus was diluted to 8 HA units/50 µL and 25 µL was combined in quadruplicate wells with an equal volume of antibody serially diluted in PBS. Plates were incubated for 1 h at 37°C in 96-well V-bottom plates. 50 µL of 1% Turkey red blood cells was then added to each well and incubated for 1 h at room temperature. Button formation was scored as evidence of hemagglutination inhibition. Titers were determined using the Spearman-Kärber formula.
All experiments were performed using black flat bottom 96-well imaging plates (BD Falcon) which were sealed with oxygen permeable film (Sigma Aldrich) before imaging. Images were taken after laser-based auto-focusing using a Pathway 855 high content imager (Becton Dickinson) equipped with different objectives (Olympus: 4X 0.16 NA, 20X 0.75 NA, and 40X 0.90 NA). Movies were taken with the 40X objective at4 frames/s while alternating between two channels over the duration of 3–5 min. For the overnight tracking of cells images were automatically taken at pre-defined positions over the duration of ∼15 h at ∼30 min intervals. Confirming infection (NP expression) after fixation and staining of the cells was carried out at the same pre-defined positions with the 40X objective, and also throughout the well with a 20X objective to determine the percentage of infected cells. Images of individual channels where overlaid and movies were compiled using ImageJ software
Cells were infected with an MOI of 3 for at least 15 h and then rinsed twice with PBS followed by fixation with 80% ice cold acetone for 10 min. After removing the acetone and drying the wells the plates were washed 3 times with 300 µL per well wash buffer (PBS, 0.05% Tween-20) then incubated for 1 h with mouse anti-influenza NP antibody (1 µg/mL) in antibody dilution buffer (1% BSA, 0.1% Tween-20 in PBS) at room temperature. After washing three times with 300 µL wash buffer the wells were incubated with 2 µg/mL goat-anti mouse AF488 labeled secondary antibody and 1 µg/mL 4′,6-diamidino-2-phenylindole (DAPI) for 1 h. After three wash steps buffer was replaced with 100 µL PBS containing 0.25 mM Sodium Azide, plates sealed, and imaged.
All staining steps described below were performed for 1 h at room temperature in the dark.
An immunofluorescence entry assay was designed to assess the ability of HA head-binding antibodies to prevent viral internalization into cells. R18-labelled H1N1 or H3N2 (MOI 3) was pre-incubated with Alexa-Fluor 647 labeled HA-specific antibodies to a final concentration of 30 µg/mL (200 nM) for 1 h at 37°C before being added to MDCK cells seeded in 96-well black-sided imaging plates (Becton Dickinson). MDCK cells stably expressing a GFP cell marker (OriGene, Rockville, USA) were incubated with virus for 15 min at 37°C followed by treatment of the cells with 0.05 U/well neuraminidase (Sigma) for 5 min at 37°C to remove non-internalized viruses. Cells were washed twice with PBS before imaging live in CO2-independent medium supplemented with 2 mM L-glutamine.
An immunofluorescence internalization assay was designed to assess the ability of HA stem-binding antibodies to be internalized into cells in complex with infectious virus particles. A pre-determined amount of R18-labelled H1N1 or H3N2 virus giving rise to 90–100% infection under the following experimental conditions was pre-incubated with Alexa-Fluor 647 labeled HA-specific bnAbs to a final concentration of 30 µg/mL (200 nM) for 1 h at 37°C. MDCK cells were treated for 5 min at 37°C with the cell permeant nuclear counterstain Hoechst 33342 (10 µg/mL, Invitrogen), followed by treatment with 3 µM tubulin tracker green reagent (Molecular Probes) for 30 min at 37°C to stain the microtubules. Cells were then incubated for 15 min at 37°C with the prepared virus-antibody mixture, followed by treatment with 0.05 U/well neuraminidase (Sigma Aldrich) for 5 min at 37°C. All reagents were diluted in CO2-independent medium supplemented with 2 mM L-glutamine. Cells were washed four times in medium before imaging live in CO2-independent medium supplemented with 2 mM L-glutamine and the glucose oxidase/catalase oxygen scavenging system (GODCAT, 1% glucose, 0.5 mg/mL glucose oxidase, 40 µg/mL catalase; all reagents from Sigma) to prevent photobleaching
Mock-labelled H1N1 or H3N2 virus was pre-incubated with Alexa-Fluor 647 labeled anti-HA bnAbs to a final concentration of 30 µg/mL (200 nM) for 1 h at 37°C. Immediately prior to infection, MDCK cells were treated for 5 min at 37°C with the cell permeant nuclear dye Hoechst 33342 (Invitrogen) at 10 µg/mL. Cells were then infected with the virus-mAb mixture (MOI 3) mixed 1∶1 with 100 nM Lysotracker Red reagent (Molecular Probes) for 15 min at 37°C, followed by treatment with 0.05 U/well neuraminidase (Sigma Aldrich) for 5 min at 37°C. All reagents were diluted in CO2-independent medium supplemented with 2 mM L-glutamine. Cells were washed four times with medium before imaging live in CO2-independent medium supplemented with 2 mM L-glutamine and the glucose oxidase/catalase oxygen scavenging system.
To determine percentage colocalization between R18-labelled virus and AF647-labelled antibodies, images were analyzed using ImageJ software with the particle analysis plugin – 3D Object Counter
A pre-determined amount of R18-labelled H1N1 or H3N2 virus giving rise to 90–100% infection under the following experimental conditions, was pre-incubated with Alexa-Fluor 647 labeled anti-HA bnAbs to a final concentration of 30 µg/mL (200 nM) for 1 h at 37°C. MDCK cells stably expressing a GFP cell marker, were seeded into 96-well black-sided imaging plates and subsequently infected with the virus-mAb mixture for 15 min at 37°C, followed by treatment with 0.05 U/well neuraminidase for 5 min at 37°C. All reagents were diluted in CO2-independent medium supplemented with 2 mM L-glutamine. Cells were washed extensively before imaging live for 15 h in CO2-independent medium supplemented with 2 mM L-glutamine and 1% FBS. The following day, cells were fixed with ice cold 80% acetone for 10 min and stained for influenza A NP expression as previously described to confirm infection inhibition in the presence of neutralizing antibody.
Fusion experiments were executed as described in the supporting information (
To study the additive effect of HA-cleavage inhibition, uncleaved A/Wisconsin/67/05 (H3N2, MOI 3) was either first incubated with trypsin (Gibco) at 1.5 µg/mL, followed by incubation with antibodies serially diluted from 0–10 µg/mL (0–67 nM), or, first incubated with antibodies serially diluted from 0–10 µg/mL, followed by incubation with trypsin at 1.5 µg/mL. FBS was added to a final concentration of 10% after trypsin treatment to inhibit trypsin activity and all incubation steps were carried out for 45 min at 37°C. Virus-antibody mixtures were then added to confluent MDCK monolayers and allowed to incubate overnight. HA cleavage status was verified by Western blot analysis with a portion of the treated samples (data not shown). Cells were fixed with ice cold 80% acetone for 10 min and stained for influenza A NP expression as described above. Calu3 cells were infected with cleaved A/Wisconsin/67/05 (H3N2) or A/New Caledonia/20/99 (H1N1) with an MOI 3 in DMEM supplemented with 2 mM L-Glutamine. Three hours post infection cells were washed twice with PBS and incubated overnight with a concentration range (0–100 µg/mL) of test or control antibody in 50 µL medium and incubated overnight. The following day, the medium of three replicate wells was pooled and spun down for 10 min at 200× g to remove cell debris. One well from each triplicate was used to obtain cell lysate by resuspending the cell layer in 150 µL lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, pH 7.5). As a positive control for HA cleavage, supernatant from cells infected with virus in the absence of antibody was used and treated with 5 µg/mL trypsin for 30 min at 37°C for complete HA cleavage. Samples were then subject to Western blot analysis. To confirm viral infection, plates were also fixed and stained with ice cold 80% acetone for 10 min and stained for Influenza A NP expression.
Single-particle fusion, and cleavage inhibition data were analyzed using a 4-parameter logistic model in which for variance stabilization the ‘
To be able to determine the effect of trypsin on the potency of CR8020, the model was modified to include an indicator variable in the estimation of the inflection point that takes the value 0 for data before trypsin and a 1 for data after trypsin (C+ID*Δc). The difference, in location, between the dose-response curves is then indicated by Δc and represents a difference in potency of the mAb under these conditions. Statistical analysis was performed using IBM SPSS statistics (version 20).
Relative amounts and cleavage status of hemagglutinin in the samples were determined by Western blotting. First 2 µL reducing agent (Invitrogen) and 5 µL 4× loading buffer (Ivitrogen) was added to 13 µL sample followed by 10 min incubation at 90°C. Proteins in each sample were resolved by 4–12% Bis-Tris SDS-PAGE (NUPAGE, Invitrogen) followed by trans-blotting onto a PVDF membrane (0.45 µm, P-Immobilon, Millipore, Massachusetts) in transfer buffer (NUPAGE, Invitrogen) containing 5% methanol at 30 V for 60 min. The membrane was blocked by incubation in blocking solution containing 4% non-fat dry milk (Bio-Rad) in TBST (20 mM Tris-HCl, 150 mM NaCl, 0.2% Tween 20) overnight at 4°C. The blocked membrane was incubated with rabbit anti-HA1 or -HA3 polyclonal serum, 60 ng/mL for 1 h at room temperature and washed 3 times with TBST. Subsequently, the membrane was incubated with goat anti-rabbit peroxidase conjugated F(ab′)
Four hours prior to the experiment, 40,000 MDCK cells per well were seeded in DMEM/glutamine into flat bottom 96 well imaging plates (BD Falcon). The amount of virus needed to achieve 90–100% infection was titrated in a separate experiment. The required amount of virus was added to the cells washed twice with PBS and incubated at 37°C, 5% CO2. After three hours, the supernatants were removed and cells were washed twice with PBS to remove non-internalized virus particles. Cells were replenished with 50 µL infection medium containing serial diluted antibodies. After incubation for 16–18 h at 37°C, 5% CO2, the supernatants were harvested, spun down to remove debris (200× g for 10 min). The remaining cells were lysed (Tris HCl pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% (v/v) Triton-X). Lysate and supernatant samples were treated with loading buffer and reducing agent, incubated for 10 min at 90°C, and analyzed by SDS-PAGE and Western blot to determine the amount of virions produced and released into the supernatant. As a control for infection, replicate identically-treated wells were fixed with 80% acetone and the number of infected cells was assessed using the imaged based infection assay (data not shown).
MDCK cells seeded on coverslips (sterile 15 mm thermanox plastic, Thermo Scientific, #174969) were infected with a pre-determined amount of virus (separate experiment) to yield 90–100% infected cells 18 h post infection. Three hours after the initial infection, the supernatants were removed; cells were washed thrice with PBS, before media containing the indicated concentration of antibodies were added. After an additional 15 h, the cell culture medium was removed and cells were fixed in phosphate buffered 2.5% glutaraldehyde buffer pH 7.4 for 1–2 h and stored at 4°C until further analysis. The coverslips were rinsed in PBS followed by distilled water and then dehydrated in 70%, 95%, anhydrous ethanol and finally in acetone and subjected to critical point drying in acetone and liquid CO2. Finally, the cells were mounted on alumina stubs and coated with a thin layer of carbon and examined in a Zeiss Ultra 55 SEM field emission microscope using an accelerating voltage of 3 keV and InLens detection at Vironova, Sweden.
MDCK cells seeded on coverslips were infected with a pre-determined amount of virus (separate experiment) to yield 90–100% infected cells 18 h post infection. Three hours after the initial infection, the supernatants were removed; cells were washed thrice with PBS, before media containing the indicated concentration of antibodies were added. After an additional 15 h, the cell culture medium was removed and cells were fixed in phosphate buffered 2.5% glutaraldehyde buffer for 1–2 h and stored at 4°C until further analysis. The samples were subsequently scraped and pelleted using a table top centrifuge, before being washed twice in 0.1 M phosphate buffer and chemically post-fixed with 2% osmium tetroxid (OsO4) in 0.1 M phosphate buffer for 2 h at 4°C followed by stepwise dehydration with ethanol, followed by LX 112-embedding by stepwise infiltration and polymerization at 60°C. Microtome sections of ∼60 nm were prepared and applied to one-slot formvar nickel grids. The sections were finally post-stained with uranyl acetate and Reynold's lead citrate before being imaged with a FEI Tecnai 10 electron microscope run at 100 kV accelerating voltage using a 2k x 2k Veleta CCD camera (Olympus Soft Imaging Systems) at Vironova, Sweden.
(TIF)
(TIF)
(TIF)
(TIF)
(TIF)
(TIF)
(AVI)
(AVI)
(AVI)
(AVI)
(AVI)
(AVI)
(DOC)
(DOC)
(DOC)
We thank N. Coppens, M. Jongeneelen, E. Geelen, C. Ophorst, T. Nahar, and O. Sahin for excellent technical support; M. Koldijk for statistical analyses; E. Brinkman, K. Hegmans, N. Hafkemeier, G. Perdok, A. Apetri and their teams for producing the mAbs and Fabs; H. Kuipers for mouse serum directed against the HA of A/Brisbane/59/2007 and D. Zuijdgeest, R. Roozendaal and A. Apetri for valuable comments. B.B., W.K., J.G., V.K., C.T., M.V.B., H.J.W.M.K., T.K., J.J., R.V., and R.H.E.F. are employees of Crucell.