This phase 1, open-label, dose escalation clinical trial (ClinicalTrials.gov, NCT04579250) was designed to evaluate the safety, tolerability, and immunogenicity of H10ssF in healthy adults. The primary objectives were evaluation of safety and tolerability of the vaccine after a single 20 mcg dose or two 60 mcg doses administered 16 weeks apart. The secondary objectives were evaluation of the antibody responses at two weeks after each vaccination with 20 mcg or 60 mcg H10ssF. The doses were chosen based on prior clinical experience with the ferritin nanoparticle vaccine platform26. Group sizes were selected to ensure the ability to detect severe adverse events and to allow for meaningful immunogenicity studies. For the 20 mcg dose participants (n = 5), there was a 90% chance to observe at least one SAE if the true rate was at least 0.369 and a more than 90% chance to observe no SAE if the true rate was less than 0.021. For the 60 mcg dose participants, within each age cohort (n = 15), there was a greater than 90% chance to observe at least one SAE if the true rate was at least 0.145 and a more than 90% chance to observe no SAE if the true rate was no more than 0.0069. To assess the primary safety and tolerability endpoints, participants recorded solicited local and systemic reactogenicity for seven days following vaccination, and clinical and laboratory assessments were conducted at protocol-specified visits for 40 weeks. Unsolicited AEs were recorded for 28 days after each vaccination, graded according to a modified FDA Toxicity Grading Scale for Healthy Adults and Adolescent Volunteers Enrolled in Preventative Vaccine Clinical Trials (see Protocol). Influenza-like illness (ILI), serious adverse events, and new chronic medical conditions were recorded for the duration of the clinical trial, which lasted 40 weeks after the first vaccination. Nasopharyngeal swabs were collected for ILIs for laboratory confirmation of influenza infection by PCR. Secondary objectives included the evaluation of H10 stem-specific antibody responses at two weeks after each vaccination.

The H10ssF (VRC-FLUNPF0103-00-VP) vaccine is a ferritin nanoparticle displaying eight stabilized-stem HA trimers from A/Jiangxi-Donghu/346/2013 (H10N8, equivalent to A/Jiangxi-Donghu/IPB13/2013) influenza. H. pylori non-heme ferritin spontaneously assembles into stable nanoparticles, allowing influenza HA protomers expressed on each ferritin monomer to associate and properly fold into HA trimers44. The H10ss antigen was developed by replacing the H10 head with a glycine-rich linker, followed by a series of structure-based mutations to improve expression and immunogenicity24,25. The H10ssF vaccine was developed at the Vaccine Production Program Laboratory (VPPL) and produced at the VRC pilot plant (operated under contract by the Vaccine Clinical Materials Program, Leidos Biomedical Research, Frederick, MD) under Good Manufacturing Practices. The vaccine was vialed at 0.7 ± 0.1 mL in 3 mL glass vials at a concentration of 180 ± 36 mcg/mL. The formulation buffer consisted of 20 mM sodium phosphate, pH 7.2, 100 mM sodium chloride, 5% w/v sucrose, and 0.01% w/v Pluronic F-68 (Poloxamer 188). The vaccine was stored frozen at −35 °C to −15 °C and thawed immediately prior to use. To administer the 20 mcg dose, the vaccine was diluted in sterile PBS, whereas no dilution was needed to administer the 60 mcg dose.

H10ssF was administered by needle and syringe into the deltoid muscle. Participants were observed for a minimum of 30 min following each injection, including recording vital signs and local reactogenicity. Safety was first established at 20 mcg. Once the protocol safety review team (PSRT) found no safety concerns by two weeks after the first three participants received 20 mcg, a dose-escalation ensued with enrollment of 18–50-year olds to receive the 60 mcg dose. An additional safety review per protocol occurred following two weeks of safety observations for the first three recipients of 60 mcg H10ssF before enrollment was opened for administration of second doses of 60 mcg H10ssF and for enrollment of the older cohort of individuals, aged 55–70 years. Participants aged 51–54 at the start of the trial (birth years 1965 to 1970) were excluded to establish two distinct cohorts based on likely first influenza exposure.

The trial was conducted at the NIH Clinical Center by the VRC Clinical Trials Program, NIAID, NIH in Bethesda, MD. Participants were recruited from the greater Washington, DC area with NIH institutional review board (IRB)-approved written and electronic media, advertisement campaigns, and community outreach strategies. Inclusion criteria specified the selection of adults aged 18–70 (exclusive of individuals born between January 1, 1965, and December 31, 1970), in general good health, and who had received at least one licensed influenza vaccine from 2015 to the time of enrollment. Exclusion criteria included receipt of a previous investigational H10 influenza vaccine and receipt of a licensed influenza vaccine within 6 weeks prior to enrollment. Full exclusion and inclusion criteria are included in the study protocol (https://storage.googleapis.com/ctgov2-large-docs/50/NCT04579250/Prot_SAP_001.pdf). Written informed consent was obtained from all participants prior to enrollment. The study followed guidelines for conducting clinical human subjects research in accordance with 45 CFR Part 46 from the US Department of Health and Human Services, US Food and Drug Administration regulations for investigational products, and principles expressed in the Declaration of Helsinki. The clinical trial protocol was reviewed and approved by the NIH IRB.

Participants’ serum samples were collected for immunogenicity analyses between baseline and 40 weeks after vaccination. HA binding antibodies were assayed using the Meso Scale Diagnostics (MSD) electrochemiluminescence (ECLIA) 10-Assay U-PLEX Development Pack (Cat# K15235N). Protein ligands were biotinylated at an AviTag site located proximal to the C-terminus from the trimer foldon, coupled to a unique streptavidin U-PLEX linker at 10 mcg/mL, and combined to form a multiplexed coating solution according to the manufacturer’s instruction. The stabilized stem proteins included: H1ss: A/Michigan/45/2015, H3ss: A/Finland/486/2004, H5ss: A/Indonesia/05/2005, H7ss: A/Anhui/1/2013, and H10ss: A/Jiangxi-Donghu/346/2013. The full-length proteins were H7FL: A/Shanghai/02/2013 and H10FL: A/Jiangxi-Donghu/346/2013. 96-well U-PLEX plates were coated for 1 h shaking at room temperature (RT). Following coating, plates were washed and incubated with serially diluted test samples and reference standard for 1 h shaking at RT. 315-53-1F12 (group 1 and 2 binding) antibody was used as a reference standard28. The binding of 50 ng/mL of 315-53-1F12 to H10ss was assigned a value of 100 arbitrary units per milliliter (AU/mL)26. AU/mL values for 315-53-1F12 binding to the heterologous HAs were assigned relative to H10ss binding. After sample incubation, plates were washed and 0.25 mcg/mL SULFO-TAG (MSD, cat #R91AO-1)-labeled anti-human IgG, IgM, and IgA (H + L) secondary antibody (ThermoFisher Cat#31128, RRID AB_228255, lot #0031128) was added for 1 h, shaking at RT. After the detection antibody incubation, plates were washed and a 1X MSD Read Buffer was applied, and plates were analyzed using the MSD Sector Imager S600. All samples were tested in duplicate. Samples with a replicate coefficient of variation > 30% were retested. Serial dilutions of samples within the dynamic range of the standard curve were interpolated to determine a sample’s AU/mL. Results were plotted and analyzed using Prism version 8 or newer (GraphPad, San Diego, CA).

Anti-ferritin binding antibodies were measured by MSD ECLIA. MSD 96-well bare plates (Cat# L15XA-3) were coated with respective ferritin proteins for 16–24 h at 4 °C. Ferritin antigens used were those previously described26,27. Following the overnight incubation, plates were washed and blocked in 5% bovine serum albumin for 1 h, shaking at RT. Following blocking, plates were washed and incubated with serially diluted test samples (1:100 followed by seven 3-fold serial dilutions) and control for 2 h shaking at RT. Plates were then washed and 5 mcg/mL SULFO-TAG (MSD, cat #R91AO-1)-labeled anti-human IgG, IgM, and IgA (H + L) secondary antibody (ThermoFisher Cat#31128, RRID AB_228255, lot #0031128) was added for 1 h, shaking at RT. After the detection antibody incubation, plates were washed, 1X MSD Read Buffer was applied, and plates were analyzed using the MSD Sector Imager S600. All samples were tested in duplicate. Serial dilutions of the sample were used to calculate and assign an area under the curve (AUC) value as the primary readout. Results were plotted and analyzed using Prism version 8 or newer (GraphPad, San Diego, CA). Samples with a replicate coefficient of variation >30% were retested.

Generation of the replication-restricted reporter (R3ΔPB1) virus subtypes H1N1, H3N2 and H10N8, as well as “rewired” replication-restricted reporter (R4ΔPB1) virus H7N9 is described elsewhere28. Briefly, to produce the R3ΔPB1 viruses the viral genomic RNA encoding functional PB1 was replaced with a gene encoding the fluorescent protein (TdKatushka2), whereas R4ΔPB1 H7N9 virus has the PB1 and H1 viral segments modified (“rewired”) to prevent HA reassortment: the HA coding region was inserted between PB1 genomic packaging signals and the fluorescent protein TdKatushka2 between the HA genomic packaging signals. The R3/R4ΔPB1 viruses were rescued by reverse genetics and propagated in the complementary cell line which expresses PB1 constitutively. Each virus stock was titrated by determining the fluorescent units per mL (FU/mL) prior to use in the experiments. For virus titration, serial dilutions of virus stock in OptiMEM were mixed with pre-washed MDCK-SIAT1-PB1 cells (8 × 105 cells/ml) and incubated in a 384-well plate in quadruplicate (25 mcL/well). Plates were incubated for 18–26 h at 37 °C with 5% CO2 humidified atmosphere. After incubation, fluorescent cells were imaged and counted by using a Celigo Image Cytometer (Nexcelom) with a customized red filter for detecting TdKatushka2 fluorescence.

Neutralization activity of serum antibodies was determined using a reporter virus microneutralization assay as previously described26,28. Briefly, serial dilutions of Receptor Destroying Enzyme-treated serum were mixed with an equal volume of R3/R4ΔPB1 virus. The reporter viruses used were H10N8: A/Jiangxi-Donghu/346/2013, H7N9: A/Anhui/1/2013, H3N2v: A/Indiana/10/2011, H3N2: A/Singapore/INFIMH-16-0019/2016, and H1N1: A/Michigan/45/2015. After incubation, pre-washed MDCK-SIAT1-PB1 cells were added to the serum-virus mixtures and transferred to 384-well plates in quadruplicate (25 mcL/well). Plates were incubated and imaged as described above. The percent neutralization was calculated for each well by constraining the virus control (virus plus cells) as 0% neutralization and the cell-only control (no virus) as 100% neutralization. A 7-point neutralization curve was plotted against serum dilution for each sample, and a five-parameter logistic curve fit generated using LabKey was used to estimate the 50% (IC50, dilution−1) or 80% (IC80, dilution−1) inhibitory concentrations. Of note, serum dilutions take into account the sera concentrations after mixing with virus at 1:1 ratio. The limit of detection (LOD) of the assay is 40 dilution−1; one-half the LOD (20 dilution−1) was used to impute values below the LOD. For the calculation of GMTs, the results using the one-half LOD imputation were reported. Because most participants had an H10 neutralizing antibody titer below the LOD at baseline, statistical comparisons between time points for the H10 assay were subjected to a sensitivity analysis. Substitution of zero instead of half LOD did not change the conclusions of statistical testing. When different imputation methods changed the p value, the significance was reported as p ≤ x, where x is the least significant p value resulting from the two test methods.

The capacity of anti-HA antibodies in participant sera to bind human FcγRIIIa was measured using a recombinant soluble human FcγRIIIa (rsFcγRIIIa) dimer ELISA as previously described45. Briefly, 50 ng of recombinant H10 HA protein for the H10N8: A/Jiangxi-Donghu/346/2013 virus or recombinant H7 HA from the H7N9: A/Shanghai/02/2013 virus was coated in the wells of 96-well NUNC Maxisorp plates (Thermo Fisher Scientific, Waltham, MA, USA) overnight at 4 °C. The wells of the plates were blocked with 1% bovine serum albumin (BSA; Sigma-Aldrich, St. Louis, MO, USA) and 1 mM EDTA (Sigma-Aldrich, St. Louis, MO, USA) in phosphate-buffered saline for 1 h at 37 °C. Two-fold serial dilutions of participant sera, starting at a 1:10 dilution, were added to the wells of the plate and incubated for 1 h at 37 °C. Plates were washed then incubated with 50 mcL of 0.1 mcg/mL biotinylated human rsFcγRIIIa dimer for 1 h at 37 °C. After washing, plates were incubated with a 1:10,000 dilution of Pierce High Sensitivity HRP-Streptavidin (Thermo Fisher Scientific, Waltham, MA, USA) for 1 h at 37 °C then washed. The color was developed by adding 3,3,5,5-tetramethylbenzidine (TMB; Sigma-Aldrich, St. Louis, MO, USA) then stopped with 1 M hydrochloric acid (HCl), and absorbance read at 450 nm on a SPECTROstar Nano microplate reader (BMG Labtech).

Binding and neutralizing antibody titers were log10-transformed before analysis, after which data was assumed to be normally distributed. Binding antibody AU/mL GM or neutralizing antibody GMT and corresponding 95% CIs were calculated for each antigen at each time point. Comparisons of antibody AU/mL or neutralization titers between weeks were made using paired t tests. Fold changes and corresponding 95% CIs were calculated. Between group comparisons were performed pairwise at each available time point using two-sample t tests unless noted otherwise and are exploratory because of small group sample sizes. For the microneutralization assay, values at the LOD were imputed to half the LOD. GMT and corresponding 95% CIs were calculated for each antigen at each time point. To determine whether the H10ssF vaccine increases ferritin binding antibody responses, two-sided Wilcoxon Sign-Rank tests were performed stratified by group. For the FcγRIIIa analysis, absorbance readouts over the entire dilution series were used to calculate an area under the curve (AUC) for each sample. The AUC calculation was performed using the trapezoid method in R using the package DescTools46. AUCs were log10-transformed before analysis. AUC GMs and corresponding 95% CIs were calculated at each time point. Fold-changes between visits with corresponding 95% CIs were calculated, and comparisons between weeks were made using paired t tests. Between group comparisons were performed pairwise at each available time point using two-sample t tests. All analyses were two sided and results were deemed significant if the p value < 0.05. No multiple comparison adjustments were employed per protocol. R version 4.3.0 was used to analyze the microneutralization and ADCC data. R version 4.2.1 was used to analyze the anti-ferritin and ECLIA data.