All animal procedures were conducted at the vivarium of the Institute of Experimental Medicine in full compliance with GOST 33216–2014 “Rules for working with laboratory rodents and rabbits.” The study was approved by the Local Committee on Ethics of Animal Care and Use at the Federal State Budgetary Scientific Institution “Institute of Experimental Medicine,” St. Petersburg (protocol No. 1/21, dated January 28, 2021). The study also complied with the Guide for the Care and Use of Laboratory Animals.

Recombinant influenza virus strains were generated by reverse genetics using the 8-plasmid system as described previously [25]. The backbone virus was based on the attenuation donor strain A/Leningrad/134/17/57 (H2N2) carrying the HA and NA surface antigens from A/South Africa/3626/13 (H1N1)pdm09. The HA gene was modified to express antigenic regions of the Spr1875, PspA, or ScaAB proteins.

The LAIV strain A/17/South Africa/2013/01 (H1N1)pdm09, identical in genome composition but containing an unmodified HA gene, was used as the control. The genome compositions of all vaccine strains are shown in Table 1.

The viruses were obtained from the collection of the A.A. Smorodintsev Virology Department of the Institute of Experimental Medicine, St. Petersburg, Russia, and propagated in embryonated chicken eggs (CE).

Sequencing of the modified HA regions was performed using an automated capillary sequencer ABI Prism 3130xl (Applied Biosystems, USA) and the BigDye Terminator Cycle Sequencing Kit v3.1 according to the manufacturer’s instructions. Viral RNA was isolated using the MagJET™ Viral DNA and RNA Purification Kit (Thermo Scientific, USA).

Amplification of gene fragments prior to sequencing was performed using the SuperScript III One-Step RT-PCR System with Platinum Taq (Invitrogen, USA) and the primers listed in Table 2. Primers were synthesized by (Alcor Bio, Russia).

PCR products were purified from agarose gels using a DNA purification kit (Diaem, Moscow, Russia).

Multiple alignment of nucleotide and amino acid sequences was performed using UGENE software (Unipro, Russia) [26].

Spatial modeling of the HA structure was carried out using AlphaFold Server (AlphaFold3) with homology modeling algorithms. Visualization of the resulting models was performed in Mol* Viewer [27, 28].

Potential N-linked glycosylation sequons were predicted with NetNGlyc, and per-site solvent-accessible surface area (SASA) was evaluated using GetArea [29, 30]. NetNGlyc sites with a jury score ≥ 6/9 were marked as sequence-predicted; sites with strong NetNGlyc support (jury ≥ 8/9) and relative exposure ≥ 60–70% were considered likely to be glycosylated in the modeled conformation.

To determine infectious activity, tenfold serial dilutions of virus stocks were inoculated into 10-day-old embryonated CE. Eggs were incubated at 33°C (optimal), 25°C (suboptimal), or 38–40°C (elevated) for 48 h (33–40°C) or 120 h (25°C). The 50% embryonic infectious dose (EID₅₀) was calculated using the Reed–Muench method and expressed as log₁₀ EID₅₀/0.1 mL [31].

For each dilution, five eggs were used (n = 5 per dilution), and titrations were performed in three independent experiments.

Viruses were classified as temperature-sensitive (ts⁺) if the titer at 39°C was ≤ 5.0 log₁₀ EID₅₀/0.1 mL and lower than the corresponding titer at 33°C.

They were classified as cold-adapted (ca⁺) if the difference between titers at 33°C and 25°C did not exceed 3.0 log₁₀ EID₅₀/0.1 mL.

Virus replication in MDCK II cells (ATCC CCL-34) was determined by endpoint titration in 96-well plates seeded at 5 × 10⁴ cells per well. When monolayers reached approximately 90% confluence, the wells were washed and inoculated with 25 µL of ten-fold virus dilutions (4–6 wells per dilution). After 1 h of virus adsorption, the inoculum was removed, and 150 µL of maintenance medium containing 1 µg/mL TPCK-trypsin was added. Cells were incubated at 33°C with 5% CO₂ for 72 h, after which viral replication was assessed by hemagglutination of the supernatant. Infectious titers were calculated using the Reed–Muench method and expressed as log₁₀ TCID₅₀/mL [31]. MDCK titrations were performed with 4–6 replicate wells per dilution in each independent experiment, and three independent experiments were carried out in total. All titrations were conducted at 33°C to evaluate in vitro replication competence and were not used for determining ts/ca phenotypes.

Female BALB/c mice aged 6–8 weeks (20 animals per group) were immunized intranasally with H1N1, H1-ScaAB, H1-PspA, or H1-Spr strains at a dose of 10⁶ EID₅₀ in a 50 μL volume. Control animals received the same volume of PBS.

Euthanasia was performed according to the Rules for Conducting Work Using Experimental Animals by direct cervical dislocation.

To evaluate viral replication, lungs were collected from five mice per group on day 3 post-immunization and individually homogenized in 1 mL of cold phosphate-buffered saline using a TissueLyser LT homogenizer (QIAGEN, USA). Homogenates were centrifuged at 10,000 rpm to remove debris and stored at −70°C. Viral titers in lung homogenates were determined by titration in 10-day-old embryonated CE and expressed as log₁₀ EID₅₀/0.1 mL. Titers were calculated according to the Reed–Muench method [31].

Serum IgG antibody levels to influenza virus were determined by enzyme-linked immunosorbent assay (ELISA) using high-capacity 96-well plates (Thermo Fisher Scientific, Waltham, USA) pre-coated with 20 agglutinating units (AU) of the wild-type A/South Africa/3626/13 (H1N1)pdm09 virus. To detect antibodies against pneumococcal inserts, plates were coated with recombinant Psp peptide (2 μg/mL) containing the Spr1875 sequence. The Psp protein, composed of three immunodominant fragments of pneumococcal surface virulence factors (PsaA, PspA, and Spr1875), was kindly provided by the Molecular Microbiology Department of the Institute of Experimental Medicine. For assays, plates were coated with Psp in carbonate-bicarbonate buffer (pH 9.6, 100 μL per well), incubated overnight at 4°C, and ELISA was performed as described previously [14]. The antibody titer was defined as the reciprocal of the highest serum dilution yielding an optical density at 490 nm exceeding the mean control value by more than three standard deviations.

The human monocyte–macrophage cell line THP-1 was used to assess early cytokine production in vitro. Cells were maintained in RPMI medium (Roswell Park Memorial Institute, Buffalo, NY, USA) supplemented with 10% fetal calf serum, 100 IU/mL penicillin, and 100 μg/mL streptomycin. THP-1 cells were seeded at 0.6 × 10⁶ cells per well in 24-well plates and incubated at 37°C with 5% CO₂ for 48 h. Viability (> 90%) was confirmed microscopically prior to infection.

Cells were then infected with 1 × 10⁶ EID₅₀/0.1 mL of the A/17/South Africa/2013/01 (H1N1)pdm09 vaccine virus or recombinant vaccine candidates H1-ScaAB, H1-PspA, and H1-Spr. Polyinosinic:polycytidylic acid (Poly I:C; Sigma, St. Louis, USA) at a final concentration of 1 μg/mL served as a positive control, and RPMI medium alone was used as a negative control. Plates were incubated for 24 h, after which supernatants were collected for cytokine analysis. Levels of TNF-α, IL-6, MCP-1, and IFN-α were measured using commercial ELISA kits (Vector-Best, Novosibirsk, Russia) according to the manufacturer’s instructions.

Statistical analyses were conducted using GraphPad Prism 8 (GraphPad Software Inc., USA). Data are presented as mean ± standard deviation (SD) or geometric mean ± SD for titer measurements. Reciprocal viral titers (EID₅₀) and ELISA antibody titers (IgG, sIgA) were log₁₀-transformed prior to analysis. Data normality and variance homogeneity were assessed by the Shapiro–Wilk and Levene tests, respectively. When parametric assumptions were met, group comparisons were performed using one-way ANOVA followed by Tukey’s post-hoc test; otherwise, the Kruskal–Wallis test with Dunn’s multiple-comparison post-hoc test was applied. All tests were two-tailed, and differences were considered statistically significant at p < 0.05. Sample sizes (n), statistical tests used, and multiple-testing corrections are indicated in the respective figure legends.

The Spr1875 insert exhibited an instability index of 34.97, indicating that the protein is relatively stable. Secondary structure predictions show that more than 50% of the sequence remains in unstructured coil regions. A single α-helical segment of three residues was identified, while the rest of the sequence remains as a flexible coil region interspersed with occasional β-fragments. Thus, the overall structure of this fragment appears sparse and flexible, without features of a stable, compact core. Spr1875 lacks distinct linear endoplasmic reticulum-associated degradation (ERAD) motifs, which lowers the likelihood of retention and degradation of newly synthesized HA in the endoplasmic reticulum (ER) [32].

At the same time, neither complete RX-[K/R]-R↓ furin motifs (typical of HA₀) nor other known cleavage sites within the Golgi lumen were identified, suggesting a low probability of unintended proteolysis. Overall, the Spr1875 insert preserves the native HA₀ profile: It is relatively stable and does not contain ERAD-related degradation motifs or protease recognition sites that could compromise HA folding or transport through the ER–Golgi pathway.

The HA-insert PCR product and the amino-acid sequence of the Spr1875 insert are shown in Figure 1A and 1B, respectively. Comparison of the HA model containing the Spr1875 insert with the unmodified A/H1N1 HA structure in PyMOL (Figure 1C) revealed a high degree of structural identity. Superimposition of atomic coordinates yielded a final root mean square deviation (RMSD) of only 0.206 Å, confirming preservation of the overall three-dimensional organization of the protein.

The PspA insert, comprising 103 amino acids (Figure 2A, B), has an estimated isoelectric point (pI) of approximately 5.0 and a very high instability index of 96.28, suggesting a strong tendency toward proteolysis. Its pronounced hydrophilicity (Grand Average of Hydropathicity, GRAVY = –1.32) reflects the predominance of polar and charged residues, which reduces the likelihood of forming a compact hydrophobic core and renders these regions predisposed to an unstructured or flexible conformation.

The PspA insert comprises three short α-helical segments and one β-sheet region, while over 75% of the sequence remains as unstructured coil fragments. Pronounced hydrophilicity and intrinsic instability create a flexible, non-compact region that could be recognized as a misfolded domain and targeted for degradation via the ERAD (HRD1/gp78) pathway. The ELM database identifies RRS (64–66) and RKS (89–91) as potential Nardilysin sites; however, in the context of HA₀, furin specificity (RX-[K/R]-R↓) is more relevant, with cleavage of mature HA occurring on the cell surface by TMPRSS2-like proteases [33]. Therefore, the functional relevance of these ELM predictions within the ER/Golgi remains uncertain.

To assess the impact of the PspA insertion on global HA conformation, structural superposition of the model containing the insertion with the native HA model was performed in PyMOL (Figure 2). The final RMSD between 10,869 atoms was 0.283 Å, indicating overall structural conservation. However, during early alignment cycles up to 267 atoms were excluded (RMSD = 0.42 Å at cycle 2), revealing local conformational distortions.

The ScaAB insertion consists of 202 amino acids (Figure 3A, B).

The high pI (~9.1) of the ScaAB insertion indicates a predominance of basic residues, imparting a positive charge to the HA N-terminus at physiological pH. Combined with an instability index of 31.01 and moderate GRAVY (–0.56), this may perturb signal-peptide recognition and translocation across the ER membrane. PSIPRED predicts a predominance of α-helical structures, accounting for more than half of the sequence, suggesting formation of a partially ordered core. In contrast, β-sheet elements constitute < 15% of the sequence and appear as short, discontinuous fragments within unstructured regions.

The presence of dibasic RX-[KR]-R motifs in ScaAB (positions 81–85 and 136–140) and an additional RRX motif at 154–156 matches furin-like cleavage patterns. However, their activity in the HA context requires experimental confirmation, as the proximity to the N-terminus may prevent proper exposure for proteolytic processing [34].

The large RMSD (18.08 Å) obtained in PyMOL analysis (Figure 3), even after excluding ≈ 600 atoms, together with the alignment warning, indicates substantial global perturbation of the HA-H1-ScaAB structure. Although part of this discrepancy may reflect the size of the inserted domain, the result suggests that ScaAB can markedly alter HA folding and topology.

When assessing potential N-linked glycosylation sites by integrating sequence-based NetNGlyc predictions with per-residue SASA values from GetArea applied to AlphaFold3 models, the following patterns were observed. Despite the N-terminal insertions, canonical HA sequons were retained but renumbered. Notably, stalk-region glycan exposure varied among constructs:

In H1-Spr, the stalk Asn (Asn366) was highly exposed (SASA = 134.51 Ų, 96.2% exposure);

By contrast, principal head sequons remained highly predicted and consistently exposed across all constructs. These findings indicate that the composition and length of N-terminal inserts can locally modulate stalk glycan exposure without affecting canonical head sites.

Based on the results of the evaluation of the replication efficiency of chimeric vaccine candidates H1-Spr 1875, H1-PspA, and H1-ScaAB in chicken embryos (CE) at different incubation temperatures, their phenotypes were characterized (Figure 4).

At the optimal temperature (33°C), the highest titer was observed for the parental H1N1 vaccine strain. The recombinant viruses showed slightly lower levels of infectious activity (Figure 4). At low incubation temperatures, however, replication rates differed significantly (Kruskal–Wallis test, p = 0.017). Pairwise comparisons revealed that vaccine candidates carrying the H1-Spr HA were significantly less efficient in replication at low temperatures compared to the original unmodified H1N1 strain (Figure 4).

All viruses exhibited the ts⁺ phenotype, confirming temperature sensitivity consistent with the requirements for live attenuated vaccines. The H1N1 vaccine strain displayed ts^–/ca^– phenotypes—it did not replicate at 39°C but retained active replication at 25°C (Figure 4). In contrast, the H1-Spr and H1-PspA strains showed markedly reduced replication at 25°C relative to 33°C, indicating the absence of cold adaptation while maintaining temperature sensitivity. The H1-ScaAB virus, however, exhibited a ca⁺ phenotype and replicated more efficiently than the other recombinant candidates at elevated temperature (Kruskal–Wallis test, p = 0.004).

The growth characteristics of the recombinant vector vaccines were analyzed in MDCK cell culture under optimal incubation conditions (33°C) for cold-adapted viruses (Figure 5). Replication of the H1-PspA virus was significantly reduced compared with the parental H1N1 strain, whereas H1-Spr and H1-ScaAB showed slightly lower replication levels, though the differences were not statistically significant.

Replication levels of the studied viruses in mouse lungs were assessed on day 3 after intranasal immunization (Figure 6).

All three recombinant candidates exhibited reduced replication in mouse lungs compared with the parental H1N1 strain on day 3 post-immunization. The reductions were statistically significant for H1-ScaAB and H1-PspA (Figure 6).

The immunogenicity study demonstrated that only the unmodified H1N1 vaccine strain induced a statistically significant increase in serum IgG antibodies to A/South Africa/3626/13 (H1N1)pdm09 three weeks post-immunization (Figure 7A). In contrast, a significant increase in local antiviral IgA was observed only after administration of the recombinant H1-ScaAB strain (Figure 7B).

No statistically significant increase in local IgA was detected following immunization with the other recombinant candidates. However, the highest average serum antibody level to the recombinant pneumococcal peptide was observed in the H1-PspA group (Figure 8A). In addition, the H1-Spr candidate induced a statistically significant increase in antibodies specific to the recombinant pneumococcal peptide Psp (Figure 8B).

In THP-1 monocyte–macrophage cultures, early cytokine production varied among constructs (Figure 9).

Exposure to the unmodified H1N1 strain resulted in the highest IFN-α production, IL-6 levels were elevated following exposure to H1N1 and H1-Spr, while MCP-1 concentrations were highest after H1-ScaAB exposure. Given the limited number of independent experiments, these observations are preliminary and require confirmation in additional biological replicates and more physiologically relevant models.

Table 3 summarizes the main biological characteristics of the recombinant vaccine candidates A/H1N1pdm09 with modified HA, compared to the parental H1N1 strain.

The unmodified H1N1 vaccine strain replicated efficiently in chicken embryos, MDCK cells, and mouse respiratory tissues, and induced a robust systemic immune response to influenza virus. The H1-ScaAB candidate also showed high replication efficiency in chicken embryos and induced a local antibody response to influenza virus. In contrast, the recombinant strains H1-Spr and H1-PspA exhibited lower replication in chicken embryos, MDCK cells, and mouse lungs, resulting in a comparatively weaker immune response to influenza virus. However, these constructs elicited the most pronounced local and systemic antibody responses to the pneumococcal Psp peptide, respectively.