Human sera

Serum was obtained from a vaccine safety and immunogenicity trial (INF-V-A017, EudraCT #2012-001693-28)35, in which subjects received Inflexal V, a trivalent seasonal influenza vaccine of the 2011/2012 composition (H1N1 A/California/07/2009, H3N2 A/Victoria/210/2009 and B/Brisbane/60/2008). Blood was collected on the day of vaccination and four weeks after vaccination. The study was approved by the ‘Universitair Ziekenhuis Antwerpen Comité voor Medische Ethiek’ ethical review board. Written informed consent was obtained from all subjects prior to enrolment.

Vaccines and adjuvants

DNA encoding tag-free H1 stem (H1 stem v1, design and alignment with full-length HA amino acid sequence as described by Impagliazzo et al.7) and H1 stem v2, optimized for expression and manufacturability, were synthesized and cloned into pcDNA3 DNA plasmid with a modified cytomegalovirus promotor (Genscript). All H1 stem proteins were based on the HA sequence from H1N1pdm A/California/07/2009, except for the H1 stem v1 used in the murine challenge experiment which was based on H1N1 A/Brisbane/59/2007 HA sequence and also known as ‘#4900”7. Recombinant protein was transiently expressed in ExpiCHO cells.

H1 stem v2 has 31% amino acid identiy similarity with full length H1N1 A/Brisbane/59/2007 HA, while the H1N1 A/Brisbane/59/2007 HA stem domain has 67% amino acid identiy similarity with H1 stem v2. H1 stem v2 has 28% amino acid identiy similarity with full length H5N1 A/Hong Kong/156/1997 HA, while the H5N1 A/Hong Kong/156/1997 HA stem domain has 61% amino acid identiy similarity with H1 stem v2.

Expression titers and trimer content of recombinant H1 stem protein

Expi293F culture supernatants were harvested at day 3 post transfection, clarified by 0.22 µm filtration. Analytical SEC at day of harvest was performed on an ultra-high-performance liquid chromatography system (Vanquish, Thermo Fisher Scientific) and µDAWN light Scattering detector (Wyatt) coupled to an Optilab µT-rEX Refractive Index Detector (Wyatt). The cleared cell culture supernatants were applied to a Unix-C SEC-300 15 cm column (Sepax) with the corresponding guard column (Sepax) equilibrated in running buffer (150 mM sodium phosphate, 50 mm NaCl, pH 7.0) at 0.3 ml/min. Analytical SEC data was analyzed using the Chromeleon software package (Thermo Fisher Scientific).The trimer content was calculated as the percentage of total protein based on peak heights (in mAU) of the monomeric and trimeric species.

Purification of recombinant H1 stem protein

H1 stem v1 and v2 were purified from the clarified ExpiCHO culture supernatant, harvested at day 7, by a two-step protocol. First H1 stem protein was captured by affinity chromatography using a resin consisting of an H1 stem protein binding single domain immobilized to POROS beads (Thermo Fisher Scientific). Following elution by application of a step gradient of elution buffer (0.1 M TRIS, 2 M MgCl2, 40% propylene glycol, pH 7.4), the antigen containing fractions were pooled and further purified by size exclusion chromatography (Superdex 200 pg 2.6 × 60 cm) using 20 mM TRIS, 150 mM NaCl, pH 7.8 as mobile phase.

In vitro H1 stem protein characterization

Thermal stability of purified H1 stem protein was assessed by Differential Scanning Fluorimetry (DSF). Melting temperatures were determined by monitoring the fluorescent emission of Sypro Orange Dye (Thermo Fisher Scientific) added to 6 µg of protein in solution. Starting at 25 °C, the temperature was increased at a rate of 54 °C per hour to a final temperature of 95 °C and melting curves were measured using a ViiA7 real-time PCR machine (Applied Biosystems), and Tm50 values were derived from the negative first derivative. Protein antigenicity was evaluated by enzyme-linked immunosorbent assay (ELISA). H1 stem protein was directly coated in half area 96-well plated (0.25 µg/mL) and incubated with mAb CR911428, CT14929 (10 µg/mL starting concentration), and Fc fused SD3830 (4 µg/mL starting concentration) in a 3-fold dilution series following incubation with an anti-human Fc horseradish peroxidase secondary antibody (mouse anti-human IgG, Jackson ImmunoResearch, cat 109-035-098, 3750-fold diluted). The luminescent signal of bound antibodies was determined following substrate addition (BM Chemiluminescence ELISA Substrate (POD), Roche) and the EC50 values of the S-curves were determined as an average of 3 independent assays.

Adjuvants

Aluminum phosphate [AlPO4; Adju-Phos™] and aluminum hydroxide [Al(OH)3: Alhydrogel™ 2%] adjuvants were manufactured and supplied by Croda Denmark, formerly known as Brenntag Biosector, Federikssund, Denmark. The amount of adsorbed protein was indirectly defined by measuring the residual protein in the supernatant after centrifugation of the protein-adjuvant mixture57. The liposomal adjuvants, SLA-LSQ, were manufactured and supplied by Access to Advanced Health Institute (AAHI, Seattle, USA), formerly known as Infectious Diseases Research Institute (IDRI), using QS-21 purified from Quillaja saponaria bark extract as previously reported58. The AS01B adjuvant (GlaxoSmithKline Biologicals, Rixensart, Belgium) was obtained from the adjuvant vial from the Shingrix vaccine (i.e., without mixing in lyophilized glycoprotein E from the separate antigen vial).

In vitro characterization of SLA-LSQ

SLA-LSQ stock solutions were mixed 1:1 with the formulation buffer (10 mM Tris, 2.5% sucrose, 150 mM NaCl), either with or without H1 stem protein, and characterized by absorbance (turbidity at 350 nm) and fluorescence emission spectroscopy (excitation 280 nm, emission 340 nm) using a Biotek Neo plate reader. In addition, dynamic light scattering was evaluated using a Wyatt Dynapro II plate reader. After characterization, Nile Red was added and mixed into the formulations and the fluorescence emission at 635 nm was measured after excitation at 550 nm. The SLA-LSQ formulations were also characterized by fluorescence microscopy, using 10 µl in a MultiCount 10™ microscopy slide, according to the staining method described by Demeule et al.59. Zeta-potential (mV) was determined using a Zetasizer Nano (Malvern). Nanoparticle tracking analysis (NTA) was measured using a Nanosight NS300 (Malvern) with a flowcell. To visualize the individual particles in NTA, formulations required a dilution of 300-times in 10 mM Tris, 2.5% sucrose and 150 mM NaCl. Video capture was performed during 120 s.

Mouse studies

Mouse experiments were approved by the Dutch Central Authority for Scientific Procedures on Animals (Centrale Commissie Dierproeven) and conducted in accordance with the European guidelines (EU directive on animal testing 2010/63/EU and ETS 123) and local Dutch legislation on animal experiments. The in vivo phase of mouse studies was performed at Triskelion, Zeist, The Netherlands. Female BALB/cAnNCrl mice aged 6–8 weeks at the start of study were purchased from Charles River Laboratories (Germany). Mice were vaccinated intramuscularly with 100 μl (divided over the two hind legs) H1 stem v2 vaccine adjuvanted with either 50 μg 2% AlPO4, Al(OH)3, Al(OH)3 + PBS or Al(OH)3 + NaCl. The characteristics of the aluminum salt-based adjuvant formulations are listed in Table 1. Blood was collected via the tail vein. Immunizations and blood collections were performed on conscious, restrained animals. Mice were intranasally inoculated with 12.5xLD50 of H1N1 A/Brisbane/59/2007 or H5N1 A/Hong Kong/156/1997 in a volume of 50 µl (25 µl per nostril) under general anesthesia with ketamine/xylazine applied intraperitoneally. Health monitoring was performed twice daily during the 21-day follow-up after virus inoculation. Animals were considered moribund and euthanized for ethical reasons if lethargy is noted in four subsequent observations. The surviving animals were sacrificed on day 21 by O2/CO2 anesthesia followed by CO2 asphyxiation.

NHP study

Housing and handling of NHPs was performed in accordance with the standards of the AAALAC International’s reference resource: the 8th edition of the Guide for the Care and Use of Laboratory animals, Animal Welfare Act as amended, and the 2015 reprint of the Public Health Service (PHS) Policy on Human Care and Use of Laboratory Animals. Handling of samples and animals occurred in compliance with the Biosafety in Microbiological and Biomedical Laboratories (BMBL), 5th edition (Centers for Disease Control). This study was performed under IACUC-approved protocol no. 18-19.

Fifty-seven adult (28 male and 29 female) Cynomolgus monkeys (Macaca fascicularis) were purchased from Beijing Prima Biotech (China) and the in vivo phase of the study was performed at Alpha Genesis Inc, Yemassee, South Carolina. Prior to study start, all animals were subjected to a health assessment and they were confirmed negative for tuberculosis, simian immunodeficiency virus (SIV), simian retrovirus (SRV), simian T-lymphotropic virus (STLV), Herpes B virus, measles and seronegative for influenza viruses as measured by a haemagglutinin inhibition assay (HI assay) for influenza strains known to have been recently circulating in the human population prior to study start. The animals were assigned to 7 study groups using randomized stratification based on body weight and sex.

Groups of 8–9 influenza-naïve cynomolgus macaques were intramuscularly immunized with 50 µg H1 stem v2 protein. The antigen was either unadjuvanted (Group 1) or mixed with SLA-LSQ adjuvant formulations that differ in the final amount of SLA and QS21: 50 µg SLA/50 µg QS-21 (Group 2), 25 µg SLA/25 µg QS-21 (Group 3), 10 µg SLA/ 25 µg QS-21 (Group 4), 5 µg SLA/ 2 µg QS-21 (Group 5). Two groups were immunized with either H1 stem v2 adjuvanted with 750 µg aluminum hydroxide adjuvant [Al(OH)3 + NaCl, Group 6] or human dose of AS01B (50 µg MPL®:50 µg QS-21; Group 7). Animals were immunized through the intramuscular route in week 0, week 8 and week 34 in an injection volume of 0.75 mL. Blood for serum was obtained at regular intervals. One animal from group 7 (AS01B) was lost to follow-up as it did not recover from anesthesia after biotechnical procedures in week 12. After reviewing of all data, it was decided that the death was deemed unrelated to vaccine or adjuvant. For all procedures, the animals were anesthetized intramuscularly with ketamine hydrochloride (10–20 mg/kg) or tiletaminezolazepam (5–8 mg/kg). The injection site was marked with indelible ink and scored according to the Draize Grading scale for 3 days post immunization, where edema was graded as none (no swelling), minimal (slight swelling), mild (defined swelling—distinct), moderate (defined swelling— raised), severe (pronounced swelling). Erythema was graded as none (normal color), minimal (light pink), mild (bright pink/ pale red), moderate (bright red), severe (dark red).

Mouse ELISA

HA-specific binding IgG antibody levels were measured by ELISA against a H1N1 A/California/07/2009 full-length HA. Briefly, 96-well ½ area plates were coated for 2 h with 0.5 µg/ml H1 A/California/07/2009 HA protein (Protein Sciences) at 37 °C in a humidified incubator. Plates were washed with PBS with 0.05% Tween (PBS-T) and subsequently blocked with block buffer (PBS + 2% BSA). After incubation for 1 hour, the plates were washed with PBS-T and serum/ control samples were added to the plates in duplicate, serially diluted in block buffer. Following a wash with PBS-T, a 1:20.000 dilution of anti-mouse IgG-HRP (KPL cat 474-1802) was added, incubated for 1 h, washed with PBS-T and Enhanced ChemiLuminescence (ECL, Bio-Rad) detection substrate was added. The luminescence signal was determined after 10 min, expressed as relative luminescence units (RLU). RLU signals were normalized per plate, and the log10 transformed normalized values were fitted using a four-parameter logistic model. Titers were calculated as a dilution ratio between a sample and the reference standard at 50% of the fitted signal (log10 relative titer).

NHP and human ELISA

HA-specific binding IgG antibody levels were measured by ELISA against C-terminally biotinylated H1N1 A/California/07/2009 full-length HA and a C-terminally biotinylated H1N1 A/California/07/2009 HA stem with different engineered stabilization domains60 to exclude irrelevant cross-reactive responses to non-HA derived epitopes. Briefly, 96-well ½-area plates were incubated with 5 μg/mL streptavidin for 2 h at 37 °C in a humidified incubator. Plates were washed with PBS with 0.05% Tween (PBS-T) and subsequently blocked with block buffer (PBS + 1% casein). Plates were incubated with 1 µg/ml C-terminally biotinylated H1N1 A/California/07/2009 full-length HA or HA stem. After incubation for 1 h, serum samples or 3 µg/ml reference standard (mAb CR911428) were added to the plates in duplicate, serially diluted in block buffer and incubated for 1 h. Following a wash with PBS-T, a 1:3750 dilution of anti-human IgG-HRP (Jackson ImmunoResearch, cat 109-035-098) was added, incubated for 1 h, washed with PBS-T and Enhanced ChemiLuminescence (ECL, Bio-Rad) detection substrate added. The luminescence signal was determined after 10 min, expressed as RLU. RLU signals were normalized per plate, and log10 transformed normalized values fitted using a four-parameter logistic model. Titers were calculated as a dilution ratio between a sample and the reference standard at 10% of the fitted signal (log10 relative titer).

ADCC hFcγRIIIa reporter assay

To measure HA-specific antibody-dependent cellular cytotoxicity (ADCC) Fc effector functions elicited by H1 stem vaccination, serum samples were analyzed using an ADCC hFcγRIIIa reporter assay (Promega). Briefly, human lung carcinoma-derived A549 epithelial cells were maintained in Dulbecco’s Modified Eagle Medium supplemented with fetal calf serum. Two days before the experiment, cells were transfected with plasmid DNA encoding H1N1 A/California/07/2009 full-length HA. One day before the assay, transfected cells were harvested and seeded in 96-well plates. After 24 h, serum samples were diluted in assay buffer and heat-inactivated, followed by serial dilution in assay buffer. The cells were replenished with fresh assay buffer and ADCC Bioassay effector cells (a stable Jurkat cell line expressing human hFcγRIIIa [V158 variant], human CD3γ, and an NFAT-response element driving expression of a luciferase reporter gene) were added and incubated. Bio-Glo Luciferase Assay System substrate was added, and the luminescence signal was read. Data are expressed as endpoint titers, which are defined as the reciprocal dilution where a pre-defined, arbitrarily threshold (5 × 104 RLU) is reached.

Pseudovirus entry inhibition assay

Neutralizing antibody titers were measured with a pseudo virus particle entry inhibition assay (ppVNA), using a panel of strains expressing influenza A group 1 HA and neuraminidase (NA) at Monogram (San Francisco, USA): H1N1 A/Fort Monmouth/1/1947, H1N1 A/South Carolina/1/1918, H1N1 A/USSR/90/1977, H1N1 A/New Caledonia/20/1999, H1N1 A/Brisbane/59/2007, H1N1 A/California/07/2009, H5N1 A/Hong Kong/156/1997, H5N1 A/Indonesia/05/2005, H6N2 A/Turkey/Massachusetts/3740/1965, H8N4 A/duck/Yangzhou/02/2005, H9N2 A/Hong Kong/1073/1999.

Influenza virus neutralization was measured using a recombinant pseudo-influenza virus assay that involves a single round of infection. A replication-defective retroviral vector (RTV1.F-lucP.CNDO delta U3) containing a firefly luciferase gene was co-transfected into human embryonic kidney (HEK) 293 cell cultures along with HA and NA expression vectors. A fourth expression vector containing a human airway serine protease used in processing the HA protein, TMPRSS2, was also transfected. The resultant pseudoviruses were harvested from culture supernatants, filtered, titrated and stored at −80 °C. Murine leukemia virus envelope (aMLV) pseudotyped virus was used as a control. The pseudovirus stocks, at a concentration giving approximately 30,000-300,000 RLU per well, were incubated at 37 °C for one hour with 3- or 4-fold serial diluted serum samples in a 96-well plate starting at a 1:10 dilution. HEK293 cells were then added to each well and incubated at 37 °C in 5% CO2. After 3 days, luciferase substrate and cell lysing reagents were added to the plates which were read on a luminometer. Inhibition curves were fitted by a four-parameter sigmoidal function using nonlinear least-squares and bootstrapping. Resulting curves were used to calculate the serum dilution/antibody concentration required to inhibit virus infection by 50% (IC50). Neutralization titers were expressed as the reciprocal of the IC50 serum dilution (ID50).

Neutralizing-breadth potency plots were generated according to Joyce et al.61. Briefly, the percentage of human-infecting influenza HA subtypes neutralized by serum (defined as percentage coverage) was calculated per animal as the minimum branch length connecting all neutralized viruses (at the specified log10 reciprocal ID50) divided by total branch length of the phylogenetic tree generated from all tested viral strains. The sequence alignment and the phylogenetic tree were generated using the Neighbor Joining method and Jukes-Cantor distance measure using CLC Main workbench 8.1 (Qiagen Aarhus, Denmark). The area under the curve (AUC) of breadth potency plots was calculated using the trapezoidal rule to quantify the breadth of neutralization. Breadth-potency plots were generated and AUC calculations performed using R.4.1.1 and libraries ape 5.6.262 and tidyverse 1.3.163.

Intracellular cytokine staining (ICS)

Antigen-specific cytokine secretion was measured by ICS at Texas Biomedical Research Institute (San Antonio, Texas). 1–2 × 106 PBMC were stimulated overnight with peptide pools of 15-mer peptides with 11-amino acid overlap (covering the entire length of H1N1 A/California/07/2009 HA or H1 stem), negative control (medium) or positive control (SEB), in the presence of Brefeldin A at 37 °C (5% CO2 and 95% humidity). Samples were stained by Live/Dead Aqua dye (Invitrogen) for dead cell discrimination. Cells were surface stained with anti-CD3-PerCP-Cy5.5 (cloneSP34.2, BD cat. 552852, 1:3400 dilution), anti-CD4-APC-H7 (clone L200, BD cat. 560837 1:3400 dilution), and anti-CD8-PE-Cy7 (clone SK1, Biolegend cat. 344712, 1:5667 dilution) for 30 minutes at 4 °C. The cells were then permeabilized with Cytofix/Cytoperm and subsequently stained intracellularly using anti-IFN-γ-APC (clone B27, Biolegend cat. 506510, 1:6000 dilution), anti-TNF-α-FITC (Mab11, BD cat. 552889, 1:2000 dilution), anti-IL-2-BV421 (MQ1-17H12, BD cat. 564164, 1:6000 dilution), anti-CD69-PE (clone FN50, Biolegend cat. 310906, 1:6000 dilution) antibodies for 30 min at room temperature. T cells were identified by consecutive gating on size (lymphocytes; FSC-A versus side scatter-A), single cells (forward scatter [FSC]-H versus FSC-A), live cells, CD3+, CD4+, or CD8+ cells, and CD69+ plus cytokine-positive (Supplemental Fig. 10). Boolean gating was used to measure single, double or triple IFN-γ, TNF-α or IL-2 positive cells. The frequency of cytokine-positive cells is presented after subtraction of the background response detected in the corresponding medium-stimulated sample of each individual animal.

Luminex

Antigen-specific cytokine secretions were measured with a 14-plex cytokine secretion assay at the Texas Biomedical Research Institute. Briefly, 4 × 105 PBMC were stimulated overnight in the presence of peptide pools of 15-mer peptides with 11-amino acid overlap, covering the entire length of H1N1 A/California/07/2009 HA or H1 stem, negative control (medium) or positive control (SEB). 24 h after stimulation an aliquot of the supernatant was removed to measure concentrations of: IL-22, IP-10, IL-4, Perforin, IFN-y, IL-17, I-TAC, TNF-a, sCD40L, MIG, IL-9, IL-6, MCP-1 and IL-2 by Luminex, according to manufacturer’s instructions. The full-length H1 HA and H1 stem group means have been centered and scaled with the root-mean-squared-error from an analysis-of-variance, with group as factor. Hierarchical clustering has been applied in the resulting heatmap to both the rows and columns. The rows have been sorted within the trees by the mean absolute deviation of the standardized group means, and the columns have been sorted by the mean of the standardized group means. In this way, the more statistically significant assays are at the top and the groups with higher responses are on the right.

Statistical analysisMouse study comparing HA stem v1 vs v2

Comparisons of the survival proportion were performed using a two-sided Fisher’s exact test.

Mouse study comparing aluminum-based adjuvants

Comparisons of the survival proportion with the mock-group were performed using a two-sided Fisher’s exact test with a 4-fold Bonferroni-correction and a step-wise approach starting with the highest dose of H1 stem per alum salt formulation. Across dose comparisons of the survival proportion were performed using a two-sided Fisher’s exact test with a 4-fold Bonferroni-correction.

Comparisons of the ELISA titers with the mock-group were performed using an ANOVA with treatment as a group factor with a 4-fold Bonferroni-correction and a step-wise approach starting with the highest dose of H1 stem per alum salt formulation. Across dose comparisons of the ELISA titers were performed using an ANOVA with treatment as a group factor with a 4-fold Bonferroni-correction.

NHP study

Comparisons between the SLA-LSQ groups and the AS01B group were performed using the Wilcoxon rank sum test for all assays. No multiple comparison adjustment was applied.

Human sera

Comparisons were performed using a paired nonparametric Wilcoxon matched-pairs signed rank test.

For all statistical tests, the significance level was 5%. All statistical calculations were done in SAS 9.4 (SAS Institute Inc).