Self-amplifying mRNA Expressing COBRA Hemagglutinin Elicits Long-lasting Broadly Reactive Antibodies Against Seasonal Influenza A Viruses

Abstract

Development of universal or broadly-reactive influenza virus vaccines is critical for addressing emerging pandemic strains, as well as improving the effectiveness and longevity of annual, seasonal influenza virus vaccines. The next generation of influenza vaccines need to address expanding the breadth of vaccine induced immune response to neutralize drifted variants, enhance the longevity of elicited immunity, and preferably use single-shot platforms that will reduce the number of vaccinations and expand the number of doses available. In this report, influenza hemagglutinin sequences, developed using computationally optimized broadly-reactive antigen (COBRA) methodology, were expressed from a self-amplifying mRNA (samRNA) vector to elicit broadly-reactive, protective immunity following a single vaccination of mice or ferrets. Three COBRA HA antigens representing an H1 HA (Y2) or two H3 HA (J4 or NG2) were expressed from individual samRNA vectors and administered individually or mixed as H1/H3 HA samRNA vaccines. In addition, two HA antigens were expressed from the same vector (Y2-J4 or Y2-NG2) as a dual expressing samRNA vaccine. Mice or ferrets vaccinated with these samRNA vaccines had long-lasting antibodies with hemagglutination-inhibition activity against a panel of H1N1 and H3N2 influenza strains representing past, current, and future drifted influenza virus variants. In addition, samRNA expressed COBRA HA antigens elicited H1 and H3 specific T cell resposnes against HA head and stem regions. Animals challenged with H1N1 or H3N2 influenza viruses had little weight loss or signs of morbidity and little to no virus detected in the lungs or nasal washes following challenge. Overall, samRNA vectors, expressing COBRA HA antigens, efficiency elicited broadly-reactive and protective immune responses following a single vaccination.

INTRODUCTION

Influenza virus induced respiratory disease continues to be a major public health issue and results in ~200,000 hospitalizations and 30,000-50,000 deaths annually in U.S. [1]. Management of disease caused by influenza viruses results in annual medical costs of $11.2 billion in the U.S. and up to ~€14 billion in the E.U. [2] Vaccination is the most effective way to prevent severe disease and control virus transmission in people. Nevertheless, fewer than 50% of eligible adults get the annual influenza vaccine [3].

To combat seasonal influenza virus infection, multiple commercial influenza vaccines are available [4]. The effectiveness of these vaccines varies from season to season based upon the match of the recommended hemagglutinin (HA) and neuraminidase (NA) in each vaccine to the circulating influenza viruses during each influenza season. Influenza viruses have a high mutation rate that results in a continuous need to update the vaccine each season and allows the circulating viruses to frequently evade host immune responses [5]. One approach to improve influenza vaccines is the use of a computationally optimized broadly reactive antigen (COBRA) methodology to generate HA immunogens that elicit immune responses that recognize a larger number of influenza strains with and across subtypes, such as H1, H2, H3, H5, H7, IBV, as well as neuramindase N1 and N2 [6–11]. COBRA HA antigens can be administered as recombinant proteins or expressed on the surface of virus-like particles (VLPs) to elicit protective antibody responses [6, 8]. New technologies associated with genetic based vaccine platforms, including messenger RNA (mRNA), which express a gene of interest in vivo to stimulate high titer antibodies, have been successfully developed to protect against the SARS-CoV-2 during the recent pandemic [12].

In this report, an alternative mRNA vaccine technology, self-amplifying mRNA (samRNA) was used to express COBRA HA proteins and the durability of the immune response was assessed for a single prime vaccination regimen and compared to a prime and boost regimen. samRNA vaccines are typically based on an alphavirus genomes that retain the genes coding for RNA replication, but are deleted in the structural genes necessary for infection. In place of the structural genes, vaccine antigen sequences can be inserted and the replication machinery can drive replication of both the full length RNA [13, 14], as well as the preferential replication from the internal subgenomic promoter, to drive higher levels of RNA expressing the vaccine antigen resulting in greater protein expression compared to conventional mRNA [15]. samRNA vaccines retain the advantages of mRNA vaccines, such as rapid development, cell-free synthesis and enhanced safety. These vaccines can administered with lower, and fewer, doses due to their replicative ability and more durable antigen expression [16]. This system results in a potent vaccine platform that drives robust T-cell and B cell responses at low doses and has effectively been utilized to protect non-human primates against SARS-CoV-2 challenge [17]. The samRNA platform elicits potent immune responses as therapeutic vaccines in oncology (Phase 1 and 2 studies) [18] and as a propylactic vaccine against SARS-CoV-2 [19]. In this study, the same mRNA platform was used to express next generation, COBRA-based HA vaccines to elicit high titer, long-lasting protective immune responses against H1N1 and H3N2 influenza viruses following vaccination in mice and ferrets.

MATERIALS AND METHODS

COBRA hemagglutinin (HA) H1 and H3 sequences

COBRA hemagglutinin (HA) proteins corresponding to H1N1 and H3N2 seasonal influenza viruses (IAVs) were design based on the next-generation COBRA methodology as previously described [24], [30]. Briefly, Y2 (H1) COBRA HA was derived from 6232 full length wild-type influenza A(H1N1) HA protein amino acid sequences, residues 1-566 (starting with Methionine as the first amino acid), from human H1N1 virus infections collected from January 1, 2014 to December 31, 2016 were downloaded from the EpiFlu online database and organized by their date of collection [31].

H3 COBRA HA, NG2, was originated by extracting full length HA sequences pertaining to 22,144 human wild-type influenza A(H3N2) viruses. H3N2 sequences of infections collected from human isolates ranging between January 1, 2016 to December 31, 2018 were used to download the HA residues 1-566 (Methionine as the first amino acid) from online database and organized in order of collection [26].

H3 COBRA HA, J4, was originated by extracting full length HA sequences pertaining to 3536 human wild-type influenza A(H3N2) viruses. H3N2 sequences of infections collected from human isolates ranging between May 1, 2013 to April 30, 2016 were used to download the HA residues 1-566 (Methionine as the first amino acid) from online database and organized in order of collection [6].

Self-amplifying mRNA (samRNA) Vector Vaccine.

HA nucleotide sequences were codon optimized using the COOL algorithm [20] and syntheized as dsDNA fragments (IDT, Coralville, IA) for cloning into PacI/BstBI sites of a pUC02-VEE vector. For dual antigen expressing vectors, a second wild-type alphavirus promoter driving a second HA was placed after the first open expression cassette by cloning into the BstB1 restriction site. To ensure that both H3 COBRA antigens are expressed efficiently, a 6X-HIS tag was added to dual constructs for detection (Figure 1). Capped samRNA was synthesized in vitro using HiScribe T7 Quick High Yield RNA Synthesis Kit (New England Biolabs, Ipswich, MA) and purified using a RNeasy Maxi Kit (Qiagen) according to the manufacturer’s protocol. samRNA was subsequently encapsulated in a lipid nanoparticle (LNP) using a self-assembly process in which an aqueous solution of RNA is rapidly mixed with a lipid mixture in ethanol. RNA encapsulation efficiency was measured using Ribogreen RNA quantitation reagent (Thermo Fisher) and confirmed to be >95% in all batches analyzed. samRNA-LNP was formulated into a buffer containing 5 mM Tris (pH 8.0), 10% sucrose, 10% maltose.

Figure 1. Schematic and protein expression of samRNA Y2, NG2 and J4 COBRA HA samRNA constructs.

(A) Schematic of monovalent COBRA constructs with an arrow indicating the position of samRNA sub-genomic promoter (B) Schematic of bivalent COBRA constructs with each HA antigen being driven by a separate sub-genomic promoters, indicated by arrows (C) Western blots against lysates from BHK-21 cells transfected with samRNA constructs and probed with anti-H1, H3, 6XHis or actin antibodies. MK-RBD-T4-His6 is a positive non-HA samRNA control for Histag expression and the negative control is a non-transfected cell lysate.

BHK-21 cells were transfected with 0.5 μg of samRNA/LNP and incubated for 24h before harvesting by centrifugation washing in 1xPBS and resuspension in 1XSDS/PAGE buffer with 2.5% β-mercaptoethanol, prior to incubation at >90 °C for 5 min. Samples (10 μL) were separated on a 4–20% SDS-PAGE gel and blotted onto a PVDF membrane using a Trans-BlotTurbo transfer system (BioRad). Membranes were blocked for 1 h in 5% skim milk in TBST (Tris-buffered saline plus 0.1% Tween-20) and then probed with either an anti-H1 mouse monoclonal antibody (Sino Biologicals, Beijing, China) at a 1:1000 dilution, an anti-H3 antibody (Sino Biologicals) at a 1:2000 dilution or an anti-Hist6 antibody (R&D Bio, Minneapolis, MN) at a 1:2000 dilution for 2 h at room temperature. Membranes were washed (0.05% Tween 20 in 1X TBS) and then probed with a Rabbit anti-Mouse HRP antibody at a 1:10,000 dilution (Bethyl Labs) for 1 h beforewashing and detection with a SuperSignal West Femto Maximum Sensitivity Substrate (Pierce). A parallel blot using a mouse anti-actin antibody (Bethyl Labs) at a 1:1000 dilution was used to ensure equivalent protein amounts per well (Figure 1C).

On the day of administration, samRNA-LNP vaccines were thawed at RT for ~15 min prior to dilution and vaccine formulation. Each samRNA-LNP component was individually diluted to a concentration of 0.1μg/μl in phosphate buffered saline (PBS), mixing via inversion. When samRNA-COBRA vaccines were combined the individual vaccine components were diluted together such that each component was 0.1 μg /μl (Suppl. Fig. 1) in PBS.

Vaccination

Fitch ferrets (Mustela putorius furo, female, 6-12 months of age), negative for antibodies to circulating influenza A (H1N1 and H3N2) and influenza B viruses, were de-scented and purchased from Triple F Farms. Ferrets (n=36) were infected with A/Singapore/6/1986 (SG/86; H1N1) and A/Panama/2007/199 (PN/99; H3N2) viruses simultaneously (5x10⁵ PFU/per virus/animal) and monitored for adverse reactions and clinical signs, including weight loss, loss of activity, nasal discharge, and diarrhea (Fig. 2). One group of ferrets (N=6) was not infected (Mock). At day 65 and day 91 post-infection, ferrets (N=6 per group) were vaccinated intramuscularly with different vaccine formulations consisting of, Y2 (H1) and NG2 (H3) COBRA HA expressing samRNA-LNPs, Y2 (H1) and NG2 H3 and J4 (H3) COBRA HA samRNA-LNPs, and 2 mock vaccine formulations of PBS only (Fig. 2). Blood was harvested from anesthetized ferrets via the anterior vena cava at days 29, 84, and 105 post-vaccination. Serum samples were centrifuged at 6,000 rpm for 10 min and clarified sera were stored at −20 ± 5°C.

Figure 2. The experimental outlines.

(A) Schematic of pre-immunized ferret study. Fitch ferrets (n=30) were infected (i.n.) with both live influenza H1N1 Sing/86 and H3N2 Pan/99 viruses at a dose of 5x10^5 PFU/per virus/animal. Ferrets (n=30) were allowed to recover for 65 days, leaving 6 ferrets not pre-immunized as naive mock controls. All ferrets were then vaccinated intramuscularly with samRNA-COBRA vaccines on days 65 and 91 as follows: 10μg each of samRNA-Y2 and samRNA-NG2 (n=6), 10μg each of samRNA-Y2, NG2, and J4 (n=6), 10μg each of samRNA-Mich/15 and Sing/16 (n=6), 10μg each of samRNA-Mich/15, HK/14, and Sing/16 (n=6), or PBS (n=12). From each ferret, less than 3mL of blood was collected on days 29, 84, and 105 of the experiment. At day 130, all ferrets (n=36) were infected (i.n.) with 4 × 10^6 PFU/animal of BR/18 H1N1 influenza virus. Nasal washes were collected on days 3 and 5 post-infection to evaluate viral titer in infected animals. For 14 days following infection, ferrets were tracked for clinical signs and body weight. (B) Acute study: Immunologically naïve DBA/2J mice (n=36) were randomly divided into 4 groups (n=9/group) and vaccinated intramuscularly at day 0 and boosted at day 28 with 10 μg of either samRNA-Y2, samRNA-J4, samRNA-NG2, or PBS (mock). Blood samples were collected at days 14 and 42. Mice were challenged with either BS/18 (3.6*10^6PFU/mouse/50μL) or KS/17 (3.1*10^8PFU/mouse/50μL) at day 56. Long-term study (bottom): Immunologically naïve DBA/2J mice (n=126) were randomly divided into 7 groups (n=18/group) and vaccinated intramuscularly with 5 μg of either samRNA-Y2+NG2 (1 dose), samRNA-Y2+NG2 (2 doses), samRNA-Y2+J4 (1 dose), samRNA-Y2+J4 (2 doses), samRNA-Y2+NG2+J4 (1 dose), samRNA-Y2+NG2+J4 (2 dose), or PBS (mock). All mice were vaccinated at day 0 and boosted at day 28. For those groups vaccinated with only one dose, only PBS was administered at day 0. Blood samples were collected at days 14, 42, 56, 84, 112, 140, and 196. Mice (n=9/group/virus) were challenged with BS/18 (3.6*10^6PFU/mouse/50μL) or SW/13 (3.1*10^8PFU/mouse/50μL) at day 196. Lungs (n=3/group/virus) were collected 3 days post-infection. Body weight and clinical signs were closely monitored for 14 days post-infection (d.p.i). (C). Immunologically naïve DBA/2J mice were randomly divided into 6 groups (n=9/group) and vaccinated intramuscularly at day 0 and boosted at day 28 with 10 μg of either samRNA-Y2, samRNA-J4, samRNA-NG2, samRNA-Y2-NG2, samRNA-NG2-Y2, or samRNA-J4-NG2. Blood samples were collected at day 14 and 42/49. Spleen were harvested from mice (n=4) at day 35. Mice were challenged with either BS/18 (3.6*10^6PFU/mouse/50μL) or SW/13 (3.1*10^8PFU/mouse/50μL) at day 56. Lungs (n=3/group/virus) were collected 3 days post-infection (day 59). Body weight and clinical signs were closely monitored for 14 days post-infection (d.p.i) and the experiment was terminated at day 70.

DBA/2J mice (females, 6 to 8 weeks old) were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). The mice were housed in microisolator units and allowed free access to food and water. All animals were cared for under the USDA guidelines for laboratory animals, and all procedures were approved by the University of Georgia Institutional Animal Care and Use Committee (IACUC) (approval no. A2021-06-016-A9).

For the acute study, the 36 immunologically naïve DBA/2J mice were randomly divided into 4 groups and vaccinated intramuscularly with either 10μg of samRNA-Y2, 10μg of samRNA-J4, 10μg of samRNA-NG2, or PBS. Mice were vaccinated in their hind legs on day 0 and day 28. Blood samples were collected 14 days after each vaccination (Fig. 2). Blood samples were collected and sera was clarified following centrifugation at 2,500 rpm for 10 min and then stored in the −20 °C freezer. A second set of mice were vaccinated with samRNA vectors expressing Y2 and NG2 HA proteins from the same construct or samRNA vectors expressing NG2 and J4 HA proteins from the same construct. These mice were vaccinated, blood collected, and challenged using the same regimen as the mice in the acute study, however, on day 35 spleens were collected from 2 mice per group for T cell analyses (Fig. 2).

For the longevity study, the 126 immunologically naïve DBA/2J mice were randomly divided into 7 groups and vaccinated intramuscularly with either samRNA-Y2 mixed with samRNA-NG2 (1 dose), samRNA-Y2 mixed with samRNA-NG2 (2 doses), samRNA-Y2 mixed with samRNA-J4 (1 dose), samRNA-Y2 mixed with samRNA-J4 (2 doses), samRNA-Y2 mixed with samRNA-NG2 mixed samRNA-J4 (1 dose), samRNA-Y2 mixed with samRNA-NG2 mixed samRNA-J4 (2 doses), or PBS. Mice were vaccinated (5μg/samRNA) by their hind legs once or twice with the vaccines listed above. Blood samples were collected at day 14, 28, 42 and 49 after the vaccination. Serum was isolated from the blood by centrifugation at 2,500 rpm for 10 min and then stored in the −20°C freezer.

Viral challenge

For ferrets, at day 130, animals were infected with egg derived A/Brisbane/02/2018 (BR/18; H1N1) virus (4x10⁶ PFU/mL) via intranasal inoculation with 500μl per nare. Ferrets were then monitored for 14 days for adverse reactions and clinical signs, including weight loss, loss of activity, nasal discharge, and diarrhea for 14 days. Nasal washes were collected at days 3 and 5 post-infection. Ferrets were then euthanized at 14 days post infection.

Mice vaccinated with samRNA-Y2 and half of the mice in the PBS group were challenged with a H1N1 strain A/Brisbane/02/2028 (BR/18) at the final concentration of 3.6 x10⁶ PFU/mouse/50μL in PBS on day 56 by intranasal innoculation. Mice vaccinated with samRNA-J4/NG2 and half of the mice in the PBS group were challenged with a H3N2 strain A/Kansas/14/2017 (KS/17) at the final concentration of 3.1 x10⁸ PFU/mouse/50μL in PBS on day 56 by intranasal innoculation. In a separate study, mice vaccinated with the same single HA expressing samRNA vectors or a dual expressing HA samRNA vectors were challenge with BR/18 H1N1 influenza virus at a dose of 3.6 x10⁶ PFU/mouse/50μL or a mouse adapted A/Switzerland/9715293/2013 (SW/13) H3N2 influenza virus at a final concentration of 3.36x10⁵ PFU/mouse/50uL at day 56 post-vaccination.

All mice were monitored daily for recording the body weight loss and clinical signs for 14 days after the challenge. On day 59, 3 mice in each vaccinated group, 3 BS/18-challenged mice in the mock group, and 3 KS/17-challenged mice in the mock group were sacrificed for the lung tissue. Lungs were frozen on dry ice, and stored at −80 ± 5°C until viral plaque assays were performed.

For the longevity study, 50% of the mice in each group were challenged with the H1N1 BS/18 at the final concentration of 3.6 x10⁶ PFU/mouse/50μL in PBS, and the other half were challenged with the H3N2 SW/13-mouse adapted at the final concentration of 3.36 x10⁵ PFU/mouse/50μL in PBS 24 weeks after the vaccination by intranasal innoculation All mice were monitored daily for recording the body weight loss and clinical signs for 14 days after the challenge. On day 3 post-infecion, 6 mice in each group (3 for the BS/18 challenge and 3 for the SW/13 challenge) were sacrificed for the lung tissue. Lungs were frozen on dry ice, and stored at −80 ± 5°C until viral plaque assays were performed.

Hemagglutination Inhibition (HAI) Assay

The HAI assay was used in this study to evaluate antibodies binding to HA protein with the function to inhibit red blood cell agglutination. Collected serum samples were tested for HAI activity [21] against a panel of 3 H1N1 viruses, A/Michigan/45/2015 (MI/15), A/Brisbane/02/2018 (BS/18), and A/Victoria/2570/2019 (VC/19), and 6 H3N2 viruses, A/Hong Kong/4801/2014 (HK/14), A/Singapore/INFIMH-16-0019/2016 (SG/16), A/Kansas/14/2017 (KS/17), A/Hong Kong/45/2019 (HK/19), A/Tasmania/503/2020 (TS/20), and A/Darwin/6/2021 (DR/21). One part sera were treated with three parts receptor-destroying enzyme (RDE) (Denka Seiken, Co., Tokyo, Japan) and incubated overnight at 37 °C. The RDE-treated sera incubated at 56°C for 45 min to inactivated the RDE in the sera. Six parts PBS was added to make the final dilution of 1:10 of the sera samples.

Samples plates were prepared by adding 50 μl of RDE-treated sera to column 1 and 25 μL and 50 μl of PBS to columns 2-11 and column 12 of a 96-well V-bottom plate. Samples were serially diluted (1:1). An equal volume of H1N1 or H3N2 virus measuring 8 hemagglutination units (HAU)/50 μL was added into each well and the plates were incubated at RT for 30 min. A solution of 0.8% turkey erythrocytes diluted in PBS (50 μL) were added to each well and then the plate was mixed by agitation and incubated at RT for additional 30 min. The HAI titer was determined as the reciprocal dilution of the last well that contained non-agglutinated RBCs. A serum sample measuring a HAI titer greater than 1:40 was defined as seroprotective in accordance with the WHO and European Committee for Medicinal Products guidelines to evaluate influenza vaccines [22].

Enzyme-linked Immunoabsorbant assay (ELISA)

96-well QuickPlex plates (Meso Scale Discovery, Rockville, MD) were coated with 50 μL of 1μg/mL of recombinant Y2, J4, NG2, A/California/4/09, A/Victoria/2570/2019 or A/Darwin/6/2021 HA proteins expressed by transient transfection in HEK293F cells (Thermo Fisher, Waltham, MA, USA) diluted in DPBS (Corning, Corning, NY, USA), and incubated at 4 °C overnight. Wells were washed three times with agitation using 250 μL of PBS + 0.05% Tween-20 (Teknova, Hollister, CA, USA) and plates blocked with 150 μL Superblock PBS (Thermo Fisher Scientific, Waltham, MA) for 1 h at room temperature Test sera was diluted in 10% species-matched serum (Innovative Research, Novi, MI, USA) and tested in single wells on each plate. Wells were washed and 50 μL of the diluted samples were added to wells and incubated for 1 h at room temperature on an orbital shaker. Wells were washed and incubated with 25 μL of 1μg/mL SULFO-TAG labeled anti-species antibody (MSD), diluted in DPBS + 1% BSA (Sigma-Aldrich, St. Louis, MO, USA), for 1 h at room temperature on an orbital shaker. Wells were washed and 150 μL Read Buffer T (MSD) added. Plates were read immediately using the QPlex SQ 120 (MSD) ECL plate reader with Methodical Mind software v1.0.37. For each sample, the endpoint titer was calculated as the reciprocal serum dilution at 2-fold the average background value for the plate, interpolated by a linear regression using the two highest dilutions that gave values >2-fold the average background. Data analysis was performed using the R programming language and graphed using GraphPad Prism 9.

Enzyme-linked cytokine immunospot assay.

T-cell enzyme-linked cytokine immunospot (ELISpot) assays were performed using an ELISpot Kit from Cellular Technology Limited (CTL) (Shaker Heights, OH, USA - Catalog No. mu-IFN-γ M/10). Briefly, 96-well polyvinylidene fluoride (PVDF) membrane high-protein-binding filter plates were coated with purified anti-murine IFN-γ (Biotin) (CTL, Shaker heights, OH, USA) and incubated overnight at 4°C. Spleens were harvested in complete Roswell Park Memorial Institute (RPMI) 1640 media + 1% fresh L-glutamine (Thermo Fisher, Waltham, MA, USA) under sterile conditions and processed using GentleMACS OctoDissociator (Miltenyi Biotec, Germany). The next day, plates were washed with PBS and 100mL of peptide pools diluted in CTL-Test Medium was added to the designated wells at 2x concentration, resulting in a final concentration of 1 μg/mL. Next, 100mL containing 3x10⁵ splenocytes diluted in CTL-Test Medium were added to each well containing the peptide pools. HA specific peptides (14 and 15-mers) from A/New York/18/2009 (H1N1)pdm09 (Catalog No. NR-19245) and A/Perth/16/2009 (H3N2) (Catalog No. NR-19266) were obtained from the BEI Resources Repository (NIH). Each 70-peptide array with 7 amino acid overlaps span each HA protein. The peptides in the H1N1 and H3N2 peptide arrays were pooled into 20 different pools, each containing 7 peptides and resuspended in 100% DMSO and individual peptides were used in pools at a final concentration of 1 μg/mL. For each group, an unstimulated control (cells only, without peptides) and a positive control using PMA (500 ng/mL) and ionomycin (50 ng/mL) were included alongside the peptide-stimulated wells. Following 24 h incubation at 37°C and 5% CO₂, plates were washed twice (PBS + 0.05% Tween-20 (Thermo Fisher, Waltham, MA, USA) and incubated with anti-murine IFN-γ (Biotin) detection antibody (Cellular Technology Limited, Shaker Heights, OH, USA) and incubated with Strepavidin-Alkaline Phosphatase (1:1000) diluted in Strepavidin-Alkaline Phosphatase diluent (CTL, Shaker Heights, OH, USA) as previously described [23, 24]. Spots were developed using Blue Developer Substrate Solution (CTL, Shaker Heights, OH, USA) to detect cytokine-producing cells and plates were scanned and counted using a CTL ImmunoSpot^® reader and ImmunoSpot^® Software version 7.0.38.2 Professional SC was used for data analysis.

Viral Plaque assay

Madin-Darby Canine Kidney (MDCK) cells (1 × 10⁶ cells) were seeded in each well of a six-well plate in 2mL of Dulbecco’s Modified Eagle Medium (DMEM) Growth media was added one day prior to performing the plaque assay. Collected ferret nasal washes were serially diluted 10-fold in DMEM + 1% penicillin-streptomycin. The seeded MDCK cells were washed 3X with DMEM + 1% penicillin-streptomycin to remove residual fetal bovine serum (FBS) and mixed with 100 μL of each dilution of nasal wash. After 1 h incubation, cells were washed twice with DMEM + 1% penicillin-streptomycin. Finally, 2× MEM (2 mL) and 1.6% agarose overlay (Cambrex, East Rutherford, NJ, USA) was added into each well and the plates were incubated at 37 °C with 5% CO2 for another 48-72 hours, monitoring for overcrowding of plaques. The overlay was removed from each well and then cells then were fixed with 10% buffered formalin for 10 min and stained with 1% crystal violet (Fisher Science Education, Waltham, MA, USA) for 10 min at RT. All plates were then rinsed using tap water to remove excess crystal violet. The plaques were counted and the lung viral titers were calculated and presented as PFU/mL of nasal filtrate.

RESULTS

samRNA vaccines express COBRA HA proteins in monovalent & bivalent single samRNA constructs

To confirm the potency of the samRNA vaccines, BHK-21 cells were transfected with samRNA/LNP formulated vaccines and expression of HA was confirmed in cellular lysates by Western blot. All antigen designs (Fig. 1A–B), both monovalent and bivalent, expressed HA of the expected ~65 kDa size (Fig. 1C). For the bivalent constructs, expression was superior from the second promoter compared to the first promoter position. As the anti-H3 antibody cannot distinguish between the two H3 COBRA HA proteins, separate vaccine constructs were generated with a 6X HiIS-tag on either the first or second protein. Probing with the anti-HIS antibody confirmed expression of both proteins with a slight bias in increased expression from the second promoter (Fig. 1C).

samRNA vaccines expressing influenza HA proteins elicit antibodies with HAI activity in ferrets against both H1N1 and H3N2 viruses

To examine the vaccine effectiveness in an animal model with pre-existing influenza immunity, ferrets were infected with a historical H1N1 and H3N2 influenza strain and then vaccinated with samRNA vectors expressing either wild-type or COBRA HA proteins. Prior to vaccination, all ferrets seroconverted to both viruses used for the pre-immune infection. Ferrets vaccinated with a mixture of samRNA-Y2 and samRNA-J4 had antisera with high HAI activity against MI/15 and BS/18 H1N1 viruses after one vaccination and lower, albeit 1:40 titers on average, against VC/19 (Fig. 3A–C). Serum HAI activity rose following a booster vaccination with HAI titers ranging from 1:160 to 1:2048 (Fig. 3A–C). The same sera had HAI activity against the panel of H3N2 influenza viruses that were, on average, between 1:40-1:80 after one vaccination, but rose 2 to 4-fold following a booster vaccination (Fig. 3D–I). Ferrets vaccinated with three samRNA vectors, expressing one H1 COBRA HA (Y2) and two H3 COBRA HA proteins (J4 and NG2) had, on average, HAI titers >1:40 against all the viruses tested after the second vaccination, with all but one ferret seroconverting against all 9 strains tested (Fig. 3). Mock vaccinated ferrets, regardless if they were naïve or had pre-existing anti-influenza immunity, had no detectable HAI titers.

Figure 3. Serum HAI antibody titers in vaccinated ferrets.

Prime-boost vaccination with multi-valent COBRA-SAM elicit HAI titers against a panel of H1N1 influenza viruses, Mich/15 (A), BR/18 (B), and Vic/19 (C), and H3N2 influenza viruses, HK/14 (D), Sing/16 (E), KS/17 (F), HK/19 (G), TS/20 (H), and Dar/21 (I). Serum samples collected at days 84 and 105 were assayed (x-axis). HAI titers depicted as absolute mean +/− S.E.M. (y-axis). Ferrets were vaccinated with either a mixture of SAM vaccines expressing Y2 and J4 (blue) or Y2, J4 and NG2 (red). Unvaccinated mice with pre-existing immune responses to influenza viruses (orange) and unvaccinated naïve mice (black). The lower dotted line represent a 1:40 HAI titer and the upper dotted line represent a 1:80 HAI titer. A p value of less than 0.05 was defined as statistically significant (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001).

Vaccinated ferrets were protected against H1N1 influenza virus challenge.

At 130 days, all ferrets were challenged with BS/18 H1N1 influenza virus (Fig. 4). Unvaccinated naïve ferrets lost, on average, ~10% of their original body weight by day 5 post-infection (Fig. 4A). Unvaccinated ferrets with pre-existing immunity lost ~5% of their original body weight by day 4 post-infection and began to recover. In contrast, all vaccinated, pre-immune ferrets did not lose any weight over the 14 days of observation. Two of the mock vaccinated naive ferrets had signs of lethargy at days 5 and 6 post-infection with high viral nasal wash titers observed at 3 days post-infection (Fig. 4B). Vaccinated ferrets had no detectable viral nasal wash titers with no detectable clinical signs. All ferrets survived infection, regardless if they were vaccinated.

Figure 4. Challenge of pre-immune vaccinated ferrets.

Ferrets (n=6 per group) were infected intranasally with the H1N1 BR/18 (4×10^6 PFU/mL). (A). Percentage of orginal weight was recorded for 14 days following infection. The X-axis represents days post infection is depicted on the x-axis and percent of original weight is depicted on the y-axis. The dotted line represents 25% of original weight. (B). Ferret nares were washed with PBS at 3 days post infection and influenza virus titers were assessed in the collected fluid. Each vaccine group is depicted on the X-axis and viral nasal wash titers are shown as plaque forming units (PFU) per mL of nasal wash. A p value of less than 0.05 was defined as statistically significant (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001).

Monovalent COBRA-SAM vaccines can elicit broad protection in naïve mice

To test the effectiveness of the samRNA-COBRA vaccines in more depth, immunologically naïve mice were vaccinated twice with one of the three samRNAs expressing a single COBRA HA vaccine (Fig. 1A). At day 14, sera collected from mice vaccinated with one samRNA-Y2 HA expressing vaccine had HAI activity against a panel of H1N1 viruses from 2009-2019 (Fig. 5A). Mice vaccinated with samRNA vaccines expressing either H3 COBRA HA (J4 or NG2) had HAI activity against a panel of H3N2 viruses isolated between 2013 to 2019 (Fig 5B–C). The H1N1-specific HAI activity was 16-fold higher following the second vaccination with HAI titers greater than 1:1280 (Fig. 5D). Serum HAI activity increased by against all of these H3N2 influenza viruses following the second vaccination of either samRNA-J4 or samRNA-NG2 (Fig. 5E–F).

Figure 5. Hemagglutination-inhibition (HAI) titers.

Monovalent samRNA-COBRA vaccines elicited broadly reactive antibodies in naïve mice model. Pooled sera sample was used to measure the HAI titers in each group. A-C: HAI titers against H1N1 and H3N2 panels for sera collected at week 2, representing 2 weeks post prime of samRNA-Y2 (A), samRNA-J4 (B), or samRNA-NG2 (C). D-F: HAI titers against H1N1 and H3N2 panels for sera collected at week 6, representing 2 weeks post boost of samRNA-Y2 (D), samRNA-J4 (E), or samRNA-NG2 (F). Virus strains used in the HAI assay are listed on the x-axis. HAI titers are shown on the y-axis. The lower dotted line indicates the HAI titer of 1:40, and the upper dotted line indicates 1:80.

All vaccinated mice were protected from an BS/18 H1N1 viral challenge with little weight loss (Fig. 6A) and no detectable viral lung titers at 3 days post-infection (Fig. 6B). In contrast, all unvaccinated (mock) mice lost >25% of their original weight and succumbed to infection by day 6 post-infection with high viral lung titers. Vaccinated mice were infected with the H3N2 influenza virus, KS/17, however, there was little or no detectable weight loss in any of the mice, including the unvaccinated mice (Fig. 6C). Whereas, there was detectable, albeit low, viral lung titers in unvaccinated mock mice following KS/17 challenge, there was no detectable viral lung titers in the samRNA-J4 or samRNA-NG2 vaccinated mice (Fig. 6D).

Figure 6. Challenge of mice with influenza virus.

Mice were infected with one of two influenza viruses BS/18 (H1N1) at the dose of 3.6*10^6 PFU/animal or KS/17 (H3N2) at a dose of 6.7*10^6 PFU/animal. (Panels A and C) Body weight (n=6/group/virus) was recorded daily for 14 days after the influenza virus challenge. Days following influenza virus challenge (D1 to D14) are depicted on the x-axis and percentage of their original body weight depicted on the y-axis. (Panels B and D) Lung viral titers (n=3/group/virus) were detected in lung tissues collected 3 days after the challenge. Viral titers are recorded absolute mean values ± S.E.M. of PFU/g lung tissue.A p value of <0.05 was defined as statistically significant (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001).

Multivalent samRNA-COBRA vaccines elicit broadly reactive antibodies

Next, mice were vaccinated either once with a single vaccination or twice in a prime-boost regimen with a mixture of samRNA vaccines expressing different combinations of the H1 and H3 COBRA HA proteins (Fig. 2B). Collected sera was tested against a panel of influenza viruses to evaluate antibody responses against a broad number of influenza viruses. Following a single vaccination, sera collected at 2 weeks post-vaccination from mice vaccinated with samRNA-Y2 and samRNA-J4 had antibodies with HAI activity against both H1N1 influenza viruses tested, as well as 6 of the 8 H3N2 influenza viruses (Fig. 7A). HAI activity was, on average, less than 1:40 against both KS/17 or DR/21. Antibodies with HAI activity was maintained over the 24-week period of observation. Mice vaccinated twice with samRNA-Y2 and samRNA-J4 had antibodies with high HAI activity against 6 of the H3N2 viruses and the titers against KS/17 and Dar/21 were both above 1:40 at day 42 (2 weeks post-boost) (Fig. 7A). At 24 weeks post-vaccination, HAI activity was reduced 2 to 4-fold from the day 42 peak against the panel of influenza viruses, but in general were maintained at high levels. Similar results were observed in mice vaccinated with samRNA-Y2+NG2 (Fig. 7B) or samRNA-Y2+NG2+J4 (Fig. 7C). These titers were maintained for 24 weeks against the two strains used for challenge, BS/18 and SW/13. There was a significant increase in serum HAI activity in mice administered one vaccination between week 2 and week 4 post-vaccination. These titers were statistically similar to HAI titers two weeks post-boost in mice vaccinated two times (Suppl Fig. 1).

Figure 7. Longevity of elicited antibody titers following vaccination.

Multi-valent samRNA-COBRA vaccines broadened and maintained the HAI activities over 6 months. HAI titers against H1N1 and H3N2 panels for sera collected at week 6 or week 28, representing 2 or 24 weeks post prime or boost of samRNA-Y2+ samRNA-NG2 blend (A), samRNA-Y2+ samRNA-J4 blend (B), or samRNA-Y2+ samRNA-NG2+ samRNA-J4 blend (C). The x-axis represents different virus strains. The y-axis represents the Log2 HAI titers with absolute mean values ± SEM. The legend indicates the different time points. The lower dotted line indicates the HAI titer of 1:40, and the upper dotted line indicates at titer of 1:80. A p value of <0.05 was defined as statistically significant (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001).

One dose of multivalent samRNA-COBRA vaccine elicits long-term protective antibodies

At 24 weeks post-vaccination, mice were challenged with either BS/18 (H1N1) or SW/13 (H3N2) (Fig. 8). There was no significant weight loss in any vaccinated mice, however, mock vaccinated mice lost more than 20% of the original body weight between day 5-7 post-infection (Fig 8A–B). In addition, there was no detectable viral lung titers in vaccinated mice in lungs collected day 3 post-infection, however, all mock vaccinated mice had viral lung titers >1x10^6 pfu/g at day 3 post-infection (Fig 8B–D).

Figure 8. Influenza virus challenge of vaccinated mice.

Mice were infected with one of two influenza viruses BS/18 (H1N1) at a dose of 3.6*10^6 PFU/animal or SW/13 (H3N2) at a dose of 3.36*10^5 PFU/animal. (Panels A and C) Body weight (n=6/group/virus) was recorded daily for 14 days after the influenza virus challenge. Days following influenza virus challenge (D1 to D14) are depicted on the x-axis and percentage of their original body weight depicted on the y-axis. (Panels B and D) Lung viral titers (n=3/group/virus) were detected in lung tissues collected 3 days after the challenge. Viral titers are recorded absolute mean values ± S.E.M. of PFU/g lung tissue.

Vaccination with samRNA expressing two HA proteins from one vector.

In order to reduce the number of vaccine constructs in the mixture, a dual expression HA samRNA vectors were developed to express two HA antigens from two different promoters on the same vector (Fig. 1B). One dual expressing vector expressed Y2 (H1) and NG2 (H3) HA proteins and the second dual expressing vector expressed J4 and NG2 HA, both H3 HA antigens. Mice were vaccinated with each dual expressing samRNA vector and the elicited immune responses were compared to mice vaccinated with samRNA vectors expressing each of the 3 HA proteins on individual constructs. All mice vaccinated with the samRNA-Y2 vaccine had high anti-HA antibody titers against the COBRA Y2 HA and wild-type HA antigens from CA/09 and VC/19 (Fig. 9). Similarly mice vaccinated with the samRNA-J4 or samRNA-NG2 vaccine had binding antibodies against the two COBRA HA proteins, as well as the wild-type HA protein isolated from the DR/21 H3N2 influenza virus (Fig. 9). Mice vaccinated with samRNA vectors expressing both Y2 and NG2 had high anti-HA antibodies against all the H1 and H3 HA antigens tested. Similarly, mice vaccinated with samRNA vectors expressing the two H3 HA antigens, J4 and NG2, had high titers anti-HA antibodies against only the H3 HA proteins.

Figure 9. Anti-HA IgG antibody binding titers.

Balb/C mice were vaccinated with 5 αg of samRNA vaccines (n=6 per group) expressing either single HA antigens (Y2, NG2 or J4) or combinations of two Cobra HA antigens (Y2 & NG2, NG2 & Y2, or J4 & NG2). All animals were boosted 28 days later with the same dose. Sera was harvested at d56 and anti-HA IgG titers were determined by a MSD endpoint dilution assay. Anti-HA IgG endpoint titers were determined against H1 and H3 COBRA antigens, and wild-type HA proteins (CA4/09, Vic/19 & Dr/21). Geometric mean (annotated), and geometric SD. LOD=50, samples below LOD (10,000) set to half the LOD (5,000 are shown.

Collected serum samples were tested for HAI activity against a panel of H1N1 and H3N2 influenza viruses (Fig. 10). Mice vaccinated with the samRNA-Y2 vaccine had, on average, HAI activity >1:80 against all 5 H1N1 influenza viruses isolated between 2009 and 2019 at day 14 post-vaccination (Fig. 10A). In contrast, mice vaccinated with samRNA-Y2-NG2 dual HA expressing vaccines had, on average, HAI activity >1:40 against all H1N1 influenza viruses tested except for CA/09. Following the second vaccination, mice vaccinated with samRNA-Y2 or the samRNA-Y2-NG2 vaccine had high HAI titers that ranged from 1:640 to 1:1280 against all 5 H1N1 viruses tested (Fig. 10C).

Figure 10. Assesement of HAI titers following vaccination of mice with dual-HA expressing samRNA vaccines.

Mice were vaccinated with samRNA vaccines expressing two HA antigens from a single vector. Sera was collected from mice (n=8/group) at 14 days (Panels A and B) or day 42 post vaccination (Panels C and D). HAI activity was assessed against a panel of H1N1 viruses (CA/09, MI/15, BR/18, GD/19, and VC/19) (Panels A and C) , as well as H3N2 viruses (SW/13, HK/14, SG/16, KS/17, HK/19, SA/19, TS/20, and DA/21) (Panels B and D). Mice were vaccinated with samRNA vaccines expressing Y2, NG2 and Y2, NG2, J4, or J4 and NG2 HA proteins. HAI titers are depicted as absolute mean values +/− S.E.M. on the y-axis. The lower dotted line represent a 1:40 HAI titer and the upper dotted line represent a 1:80 HAI titer. HAI titers were statistically analyzed using a two-way ANOVA Tukey’s multiple comparison test by Prism 10 software. A p value of less than 0.05 was defined as statistically significant (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001).

Similarly, mice vaccinated with the samRNA-NG2 vaccine had, on average, HAI activity >1:40 against 6 of the 8 H3N2 influenza viruses isolated between 2013 and 2021 at day 14 following vaccination (Fig. 10B). Mice vaccinated with samRNA-J4 or with the dual vaccine expressing both H3 COBRA HA proteins (NG2 and J4) had, on average, HAI activity >1:40 against 7 of the H3N2 influenza virus tested (Fig. 10B). In contrast, sera collected at day 14 from mice vaccinated with samRNA-Y2-NG2 dual HA expressing vaccines had little to no HAI activity against any of the H3N2 viruses (Fig. 10B). In contrast, mice vaccinated with samRNA-NG2, samRNA-J4, or samRNA-J4-NG2 dual HA expressing vaccines had average HAI activity >1:40 against all H3N2 influenza viruses in the panel at day 42 with high HAI titers against SW/13, HK/14, SG/16, HK/19, SA/19, and TS/20 (Fig. 10D). In addition, mice vaccinated with samRNA-Y2-NG2 dual HA expressing vaccine had HAI activity ≥1:40 against 6 or the 8 viruses at day 42 post-vaccination.

At day 35 post-vaccination (7 days post-boost), spleens were collected from vaccinated mice and assessed for T cell responses (Fig. 11). Mice vaccinated with samRNA-Y2 or samRNA-Y2-NG2 had detectable interferon (INF) gamma (γ) secreting splenocytes against a set of H1N1 HA-specific peptides (Fig. 11). On average, mice with ≥200/million INF-γ secreting splenocytes detected peptide pools 3, 4, 9, 10, 13, 15, 16, and 20 (Fig. 11A), which identified 8 specific peptides regions in the Y2 amino acid sequence (Suppl Fig. 2A). Fewer number of INF-γ secreting splenocytes were detected in mice vaccinated with samRNA-Y2-NG2 vaccine (Fig. 11A). Overall, 4100 The number of INF-γ secreting spots were detected from collected spleens against the H1N1 peptides in mice vaccinated with samRNA-Y2 and 1300 spots detected in from splenocytes collected from mice vaccinated with samRNA-Y2-NG2 (FIg. 11B). Mice vaccinated with samRNA-NG2 vaccine had low numbers of INF-γ secreting splenocytes detected, however, dual samRNA vaccines expressing NG2 and Y2 or NG and J4 had 6-7 fold higher number of INF-γ secreting spots detected ranging from 300-350 spots (Fig. 11C–D). In contrast, Mice vaccinated with samRNA-J4 vaccine had even higher numbers of INF-γ secreting splenocytes (470 spots) detected (Fig. 11E–F). Splenocytes reacted to two peptide pools in the H3N2 panel that identified one dominate peptide region (peptide 129) in the stem region of HA (Fig. 11E and Suppl Fig. 2D). For the 10 H1N1 peptides identified, 8 of them tracked to the stem region of HA and two were located in the HA head region (Fig. Suppl 2C).

Figure 11. Assessment of HA-specific T cell responses elicited by samRNA-COBRA HA expressing vaccines.

Naïve DBA2/J mice were vaccinated with samRNA-COBRA vaccines: samRNA-Y2, samRNA-NG2-Y2, samRNA-J4, samRNA-J4-NG2, samRNA-NG2, or samRNA-empty control. Activated T cells specific for HA-H1N1 (A) or H3N2 (C, E) were quantified by ELISPOT. T cells isolated from the spleen of vaccinated mice were tested for recognition of overlapping H1 or H3 70 peptides pools (14-15-mer length of peptides and overlapping by 7 aa) and assessed for IFN-γ secretion by ELISPOT. The frequency of reactive T cells per million splenocytes by H1N1 peptide pools (A) or H3N2 peptide pools (C-D). The cumulative number of counted IFN-γ spots spots for all 20 H1N1 peptides pools (B) or 20 H3N2 peptide pools (D). The Specific peptides (p) were identified as stimulating the highest IFN-γ secretion highlighted with p number above the peptide bar (A, C, E). The average number of IFN-γ secreting splenocyte spots with the S.E.M is depicted. Number of IFN-γ count spots were statistically analyzed using two-way ANOVA Tukey’s multiple comparison test by Graphpad Prism 10 software.

DISCUSSION

Innovative new strategies are needed for developing broadly-reactive, next generation influenza vaccines. Both new antigen designs as vaccine targets, as well as effective, one-shot delivery platforms are game changers for combating emerging influenza viral variants. In this report, hemagglutinin sequences, derived using a computationally optimized broadly reactive antigen (COBRA) methodology, were combined with a self-amplifying mRNA vaccine platform to elicit high titer, long-lasting, protective antibodies against a broad panel of H1N1 and H3N2 influenza drifted variants. These COBRA HA sequences have been expressed and purified as recombinant protein [25, 26], on the surface of virus-like particles [27, 28], as inactivated split viruses [29], or as mRNA LNPs [30]. In general, two or more vaccinations using these vaccine platforms are necessary to elicit high titer neutralizing antibodies. However, these vaccines were effective following a single vaccination with some of the highest HAI titers observed using a COBRA HA antigen, particularly in animals with pre-existing anti-influenza immunity. Historically, COBRA HA antigens effectively elicit broadly-reactive HAI titers against contemporary and drifted influenza virus variants (REFS). Combining the samRNA platform with COBRA antigens combines the advantages of the platform and the antigen and could lead to improvents in influenza vaccines that are needed particular in populations that are most at risk.

Most people are exposed to influenza viruses in the first 5 years of life [31]. Initial influenza virus infections often imprint on the host immune system and can influence future immune responses following vaccination or subsequent infections [32, 33]. In healthy adults, a single influenza vaccination is sufficient to recall memory B and T cells, but in children, the elderly, or immunocompromised people, a single vaccination may not be sufficient to elicit high protective immune responses. In children 6 months to 8 years of age that are often immunologically naïve to influenza viruses, the U.S. Center for Disease Control and Prevention recommends administration of 2 doses of the seasonal influenza vaccine [34].

Next-generation influenza virus vaccines not only need to elicit broadly-reactive immune responses against viral variants, but these responses should be durable and last for multiple seasons. Immunologically naïve mice that were administered a single dose of the samRNA-COBRA vaccine were protected against influenza virus infection 24 weeks after vaccination (Fig. 7). In humans, immune responses elicited by influenza virus vaccination that maintain protective immunity for multiple seasons would be an improvement over currently commercial influenza virus vaccines. In most people, protective HAI titers wane over a 12 month period particularly in the elderly [35]. A vaccine formulation that elicits long-lasting immune responses against ever drifting viral variants has many advantages for vaccine manufacturing by allowing for the production of influenza virus vaccines year round using the same sequences and formulations. A single shot vaccine also increases the number of doses available for the population and may result in larger numbers of people receiving the vaccine over multiple influenza virus seasons.

Almost all adults have pre-existing immune memory against past infections and vaccinations [36]. Upon vaccination or infection with a variant strain, subsets of memory B and T cells are often recalled in response and provide a burst of antibodies and cellular responses [32, 36]. To mimic the human condition, ferrets were infected with historical strains of influenza to establish a “pre-immune state” against subtypes of influenza viruses. These ferrets vaccinated with samRNA-COBRA vaccines had antibodies with HAI activity against both H1N1 and H3N2 influenza viruses following one vaccination, which was boosted after the second vaccination. All ferrets had HAI activity (>1:40) that is considered protective for people vaccinated with seasonal influenza vaccines [37, 38]. The samRNA-COBRA vaccines were quite effective at generating neutralizing antibodies against future drifted viral variants and did not appear effected by the previously imprinted immunity, also referred to as original antigenic sin [39]. This vaccine platform is self-amplifying and expresses the antigen sequences over a longer period of time (up to 7 weeks) [40] compared to protein-based vaccines [41, 42], providing antigen presentation and immune stimulation more effectively.

While the H1N1 component of the samRNA-COBRA vaccine was quite effective following a single vaccination, it took two vaccinations to elicit high titer antibodies against a subset of H3N2 influenza viruses. A second dose of the H3N2 component was necessary in previous bivalent COBRA HA protein vaccine studies to elicit high HAI titers against H3N2 drifted variants [6]. Antigenic H3N2 influenza virus variants evolve more frequently than H1N1 influenza viruses or influenza B viruses [43]. Therefore, H3N2 viruses are often the most challenging seasonal influenza subtype to design effective vaccines against. In this study, two different COBRA H3 HA sequences were tested representing different time eras. The J4 HA sequence was developed using input sequences from 2013-2016 and input sequences from 2016-2018 were used to design the NG2 HA sequence [44, 45]. The samRNA-COBRA vaccines were assessed with one samRNA-H3 HA vaccine (J4) coupled with a samRNA-H1 HA vaccine (Y2) or a second vaccine was designed with two samRNA-H3 HA components (J4 and NG2) with a samRNA-H1 HA vaccine (Y2). HAI activity elicited by the vaccines were effective against many viruses with one H1 and one H3 HA component, however, when HAI activity was low (<1:40) against a particular H3 strain, the additional of a second H3 HA (NG2) often statistically enhanced HAI titers without a decline in the HAI activity against the H1N1 viruses (Fig. 4). The addition of the NG2 component may elicit HAI activity against more recent H3N2 strains. Alternatively, the two H3 COBRA HA components may elicit complementary antibodies that recognize a broader number of HA epitopes on drifted H3N2 viruses allowing for broader neutralizing activity. For example, the Y2 and NG2 combination elicited antibodies with lower HAI activity against KS/17 with a single vaccination. However, when the J4 component was added to the vaccine, 5 out 6 vaccinated ferrets had protective antibody titers against KS/17 following a single vaccination (Fig. 4 and Suppl. Fig. 1). In addition, samRNA expressing J4 had enhanced HAI activity against SW/13, KS/17, TX/17, or HK/19 following vaccination compared to the same antigen delivered as recombinant HA [42]. However, we did note differences in the CD4+ T cell responses when HA antigens were administer alone or in combination with other HA antigens with lower number overall CD4+ T cell responses noted with two HA proteins in the vaccine. The immune response has more eptiopes to process with differing regions and may reduce the overall targeting of specific regions, whether there were two H3 HA proteins or and H3 and H1 protein administered together. There were more dominate HA stem peptides identified using the H1 peptide pools and more HA head peptides identified using the H3 peptide pool. Simulating CD4+ T cells specific to the antigen in the vaccine does help to enhance antibody titers and affinity maturation and antigen presentation {Juno, 2022 #1762}. However, one of the limitations of the T cell analyses is the use of mismatched peptides to the COBRA HA sequences in the vaccine and therefore, more critical peptides could have been identified using a matched peptide set.

Overall, inclusion of more than one H3 HA component in a next generation vaccine may be advantageous. Current commercial influenza vaccines contain up to 4 components representing one H1N1 and one H3N2 influenza A viruses and one or two influenza B viruses. For the 2024-2025 influenza season, a trivalent version of the influenza vaccine was recommended to manufacturers [46]. There may be value in adding a second H3N2 component to the influenza virus vaccine to increase breadth of elicited protective immune responses against this subtype. A multivalent COBRA rHA-based vaccine containing 5 or 8 antigen components is still a highly effective influenza vaccine [47]. The samRNA vaccine approach is ideal for developing a multivalent influenza vaccines either by including individual samRNAs co-expressing single antigens or by utilizing samRNAs that express multiple antigens on a single samRNA vaccine vector. In this study, COBRA samRNA expressing one two COBRA sHA were as effective as mixing two samRNA vectors expressing each vaccine component independently. Overall, the samRNA-COBRA vaccine platform elicits durable immune responses following a single dose in both immunologically naïve and pre-immune hosts.

Supplementary Material

1

Supplemental Figure 1. HAI titers detected for 24 weeks following vaccination. A single vaccination of samRNA-COBRA HA expressing vaccines can elicit comparable HAI titers as a prime/boost vaccination regimen over a prolonged period. HAI titers against BS/18 was detected for sera collected from 2 to 24 weeks post prime (A) or boost (B). HAI titers against SW/13 was detected for sera collected from 2 to 24 weeks post prime (C) or boost (D). Weeks post-vaccination are depicted on the x-axis. HAI titers are depicted as absolute mean values ± S.E.M. on the Y-axis. The lower dotted line indicates the HAI titer of 1:40, and the upper dotted line indicates 1:80. A p value of less than 0.05 was defined as statistically significant (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001).

Supplemental Figure 2. Identification of specific immunostimulatory peptides and location of epitopes on HA. Specific peptides were identified that stimulate splenocytes to secrete IFN-γ following incubation in ELISPOT. (A) A set of 139 peptides specific for the HA protein from A/New York/18/2009 (H1N1)pdm09 were used to assess collected spleens following vaccination with samRNA vaccines expressing Y2 COBRA HA and (B) a set of 139 peptides array specific for the HA protein from A/Perth/16/2009 (H3N2) were used to assess collected spleens following vaccination with samRNA vaccines expressing J4 COBRA HA. Full length COBRA HA amino acid sequences were aligned with the amino acid sequences. (C) Red boxes represent the sequences of H1N1 peptides candidates in the sequence alignment that elicited HA-specific T cell responses. Clusters of H1N1-specific HA specific T cell epitopes were identified between amino acids 457 and 543. Sequences alignment was performed using ClustalOmega EMBL-EBI Job Dispatcher sequence analysis tools [48]. (D) Red boxes represent the sequences of H3N2 peptides candidates in the sequence alignment that elicited HA-specific T cell responses. Clusters of H3N2-specific T cell epitopes in HA were identified between amino acids 97 and 119.