Incorporation of Membrane-Anchored Flagellin into Influenza Virus-Like Particles Enhances the Breadth of Immune Responses^▿

Cell lines and viruses.

Spodoptera frugiperda Sf9 cells were maintained as suspension cultures in flasks with serum-free SF900 II medium (Gibco-BRL) at 27°C with stirring at a speed of 80 rpm. Madin-Darby canine kidney (MDCK) cells were cultured in Dulbecco's modification of Eagle's medium (DMEM) (Cellgro) plus 10% fetal bovine serum. Influenza A/PR8 (H₁N₁) virus was grown in and purified from hen egg embryonic fluid as described previously (41). Mouse-adapted PR8 and A/Philippines/2/82/X-79 (H₃N₂) (A/Philippines) viruses were prepared as described previously (42).

Construction of a membrane-anchored flagellin gene.

A full-length membrane-anchored flagellin encoding gene was generated by fusing a signal peptide (SP) from honeybee mellitin and the transmembrane (TM) and cytoplasmic tail (CT) regions from the influenza A virus PR8 hemagglutinin (HA) to the 5′ and 3′ termini of the flagellin gene in frame, respectively. The mellitin SP-encoding fragment was PCR amplified from the plasmid M-TM.CT_MMTV (60) by use of primers 5′-GGTTCTAGAATGAAATTCTTAGTC-3′ and 5′-GTGGGATCCTTTCATGTTGATCGG-3′ (XbaI and BamHI sites are underlined) and cloned into cloning vector pBluescript (−) with XbaI/BamHI sites, resulting in plasmid pBluescript-SP. The Salmonella enterica serovar Typhimurium flagellin gene (fliC; GenBank accession no. D13689) was amplified from plasmid pEM045 pEF6 FliC stop (a kind gift from Alan Aderem, Institute of Systems Biology, Seattle, WA) by using primers 5′-GCAGGATCCATGGCACAAGTCAT-3′ and 5′-CGCGAATTCACGCAGTAAAGAGAG-3′ (underlined sites are BamHI and EcoRI, respectively) and inserted into pBluescript-SP, resulting in pBluescript-SP-FliC. HA TM-CT was amplified from plasmid pc/pS1 containing the full-length HA gene by using primers 5′-GCTAGAATTCCAGATTCTGGCGATC-3′ and 5′-GCTAGGGCCCTTATCAGATGCATATTCT-3′ (underlined sites are EcoRI and ApaI, respectively) and cloned into pBluescript-SP-FliC to produce pBluescript-SP-FliC-HA tail. The full-length membrane-anchored flagellin gene was amplified from pBluescript-SP-FliC-HA tail by using primers 5′-GCTCGTCGACATGAAATTCTTAG-3′ and 5′-GCTACTCGAGTTATCAGATGCATATTC-3′ (SalI and XhoI sites, respectively, are underlined) and cloned into pFastBac 1 under the control of the polyhedrin promoter. The sequence of the membrane-anchored flagellin gene was verified by DNA sequencing.

Generation of rBVs.

A recombinant baculovirus (rBV) expressing membrane-anchored flagellin was derived from the transfer plasmid pFastBac1 constructed as described above by using a Bac-to-Bac expression system (Invitrogen) according to the manufacturer's instructions. rBVs expressing PR8 HA and M1 were described previously (42).

Determination of the cell surface expression of membrane-anchored flagellin.

The presence of the membrane-anchored flagellin on cell surfaces was determined by a cell surface expression assay. Sf9 cells were seeded in six-well plates at 10⁶ cells/well. The infection with the flagellin-expressing rBV and isotopic labeling were performed as described previously (60). The biotinylation of cell surface proteins and immunoprecipitation were carried out (62). Final samples were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Gels were dried and then used for autoradiography.

Production and characterization of cVLPs.

Flagellin-containing influenza VLPs were produced by infection of Sf9 cells with rBVs expressing PR8 HA, M1, and membrane-anchored flagellin. Standard VLPs containing HA/M1 and control VLPs containing M1 only were produced by infecting Sf9 cell with rBVs expressing PR8 HA and M1 as described previously (42). Cell culture supernatants were collected on day 3 postinfection and cleared by a brief centrifugation (4,000 × g for 20 min at 4°C). VLPs were pelleted by ultracentrifugation at 100,000 × g for 1 h at 4°C. The pellets were resuspended in phosphate-buffered saline (PBS) at 4°C overnight. VLPs were further purified through a 20%-35%-60% discontinuous sucrose gradient at 100,000 × g for 1 h at 4°C. The VLP band between 35% and 60% was collected and then diluted with PBS and pelleted at 100,000 × g for 1 h at 4°C. VLPs were resuspended in PBS overnight at 4°C. The resulting VLPs were characterized by Western blot analysis, hemagglutination activity analysis, and electric microscopic observation. For Western blot analysis, HA and M1 bands were probed by mouse anti-HA or M1 polyclonal antibodies. Membrane-anchored flagellin was detected by rabbit antiflagellin polyclonal antibodies (Provided by Alan Aderem). The flagellin content in cVLPs was estimated by comparison with a standard purified soluble standard flagellin in Western blotting. The hemagglutination activity of VLPs was determined by the capacity to hemagglutinate chicken red blood cells (42). For electron microscopy, VLP samples (5 to 10 μl; 0.1 mg/ml protein) were examined as described previously (60).

Treatment with glycosidases.

Peptides N-glycosidase F (PNGase F) and endoglycosidase H (endo-H) (New England Biolabs) were used to determine the glycosylation of membrane-anchored flagellin by following the manufacturer's instructions. Flagellin-containing VLPs (10 μg in a volume of 10 μl) were mixed with 1 μl of denaturing buffer and heated at 100°C for 10 min. After being cooled in an ice bath for 2 min, samples were mixed with reaction buffer and PNGase F or endo-H and incubated at 37°C for 1 h. The reaction was terminated by adding sodium dodecyl sulfate-polyacrylamide gel electrophoresis loading buffer, and the mixture was heated at 100°C for 5 min. Samples were subjected to Western blotting analysis.

TLR-5-specific bioactivity assay.

A RAW264.7 cell-based assay (31) with modifications was used to determine the bioactivity of membrane-anchored flagellin in VLPs. The RAW264.7 cell line is a mouse macrophage cell line which expresses TLR-2 and -4 but not TLR-5 (64). In brief, 80%-confluent RAW264.7 (TLR-5-negative) cells in a 75-cm² T flask were transfected with 10 μg of plasmid pUNO-hTLR5 expressing the human TLR-5 (InvivoGen) by use of the transfection reagent Lipofectamine (Invitrogen) by following the manufacturer's instructions. Six hours posttransfection, cells were removed from the T flask with a cell scraper and seeded into 96-well plates by using 5 × 10⁴ cells/well in 100 μl of fresh medium. Nontransfected RAW264.7 cells (TLR-5 negative) were also seeded into 96-well plates for comparison. The TLR-5-positive and -negative cells were incubated with 100 μl of serially diluted purified soluble flagellin, flagellin-containing VLPs, or standard HA/M1 VLPs in DMEM, and supernatants were collected after 24 h. Levels of tumor necrosis factor alpha (TNF-α) production stimulated by soluble flagellin, flagellin-containing VLPs, or standard HA/M1 VLPs in both TLR-5-positive and TLR-5-negative cell cultures were determined by enzyme-linked immunosorbent assay (ELISA). TLR-5 bioactivity was expressed as the level of TNF-α production of TLR-5-positive cells from which was subtracted that of TLR-5-negative cells stimulated by flagellin, flagellin-containing VLPs, or standard HA/M1 VLPs.

Immunization and challenge.

Inbred female BALB/c mice were obtained from Charles River Laboratory. Mouse groups (six mice per group) were immunized twice with 10 μg/mouse of VLPs at 4-week intervals (weeks 1 and 4). For virus challenge, mice were anesthetized with isoflurane and infected with 40 times the 50% lethal dose (40×LD₅₀) (LD₅₀ = 50 PFU/mouse) of mouse-adapted A/PR8 virus (2,000 PFU) or 40×LD₅₀ (LD₅₀ = 25 PFU/mouse) of mouse-adapted A/Philippines virus (1,000 PFU) in 50 μl of PBS per mouse 4 weeks after the boosting immunization. For the determination of lung virus titers, six mice from each group were sacrificed on day 4 postchallenge. Blood samples were collected on weeks 0, 3, and 7 by retro-orbital plexus puncture. After clotting and a brief centrifugation, serum samples were collected and stored at −80°C prior to use for assays.

Antibody titration.

The influenza virus-specific serum antibody endpoint titers, including those for immunoglobulin G (IgG) and subtypes (IgG1, IgG2, and IgG2b), were determined by ELISA as described previously (42). In brief, 96-well microtiter plates (Nunc Life Technologies) were coated with 100 μl/well of inactivated PR8 virus (5 μg/ml) in PBS overnight at 4°C. For serum IgG titers against the heterologous A/Philippines virus (H₃N₂), plates were coated with 100 μl/well of inactivated A/Philippines virus (5 μg/ml). The serum samples were serially diluted in twofold steps. After being washed and blocked with 1.5% bovine serum albumin, plates were used to bind antibody with the diluted sera. The detection color was developed by binding horseradish peroxidase-labeled goat anti-mouse IgG, IgG1, IgG2a, or IgG2b (Southern Biotechnology) at 37°C for 1 h. After extensive washing, the substrate TMB (Zymed, Invitrogen) was applied. The optical density at 450 nm (OD₄₅₀) was read using an ELISA reader (model 680; Bio-Rad). The highest dilution which gave an OD₄₅₀ twice that of the naive group without dilution was designated as the antibody endpoint titer.

Virus neutralization and HI assays.

A neutralization assay was performed using MDCK cells as described previously (42). Hemagglutination inhibition (HI) assays were carried out as described previously (7) with modifications. Briefly, 8 hemagglutination units of PR8 or A/Philippines virus were mixed with serially diluted receptor-destroying enzyme-pretreated serum samples in a total volume of 50 μl and incubated at 37°C for 1 h. An equal volume of chicken blood cells (0.5%) was mixed with the virus-serum mixture and incubated at 25°C for 30 min. HI titers were recorded.

Cytokine assays.

Interleukin-2 (IL-2), gamma interferon (IFN-γ), TNF-α, and IL-4 levels were determined by ELISA (eBioscience, San Diego, CA) according to the manufacturer's instructions. Splenocytes (1 × 10⁶ cells) isolated from immunized mice were seeded into 96-well issue culture plates and stimulated with a mixture of two major histocompatibility complex class I (MHC-I) HA peptides (IYSTVASSL and LYEKVKSQL) or a pool of five MHC-II HA peptides (SFERFEIFPKE, HNTNGVTAACSH, CPKYVRSAKLRM, KLKNSYVNKKGK, and NAYVSVVTSKYN RRF) at a concentration of 10 μg/ml. The plates were incubated at 37°C for 2 days, and cell culture supernatants were collected for cytokine assays.

Construction of the membrane-anchored flagellin coding sequence.

To incorporate flagellin into VLPs as a molecular adjuvant, the gene must be modified to enable membrane translocation, transport, and cell surface expression. As schematized in Fig. 1A, a membrane-anchored flagellin-encoding gene was constructed by merging the coding sequence for the SP from the honeybee protein mellitin at the N terminus and a TM-CT from influenza HA in frame at the C terminus. As a heterologous SP, mellitin SP is known to improve glycoprotein cell surface expression in an insect cell system (60). The TM-CT sequence from HA provides a membrane anchor sequence which would be expected to engage the modified flagellin to assemble into influenza virus matrix protein (M1)-derived VLPs (1). rBV was generated using the resulting membrane-anchored flagellin coding sequence. As shown in Fig. 1B, left, flagellin fused with the HA TM-CT region was expressed well in rBV-infected Sf9 cells. The modified protein showed two major bands in cell lysates by the autoradiography (Fig. 1B, left, lane 1). The lowest band is around 55 kDa, which corresponds to the molecular mass estimated according to its amino acid composition in a nonglycosylated form. The top band is about 6.5 kDa, indicating that glycosylation of this modified protein occurs in the insect cell protein expression system. There are other bands between the two main bands, suggesting differences in glycosylation of the membrane-anchored flagellin. As shown in Fig. 1B, right panel, the surface-expressed flagellin corresponds to the top band with a molecular mass of 65 kDa, suggesting that only glycosylated flagellin was transported to cell surfaces.

Production of cVLPs containing flagellin.

As described previously, VLPs were produced in an rBV-derived protein expression system in Sf9 insect cells and purified by gradient centrifugation (42). To optimize the production of cVLPs, rBVs expressing HA, M1, and flagellin at various of multiplicities of infection (MOI) shown in Fig. 1C, bottom, were compared. The results in Fig. 1C, top, demonstrate that standard influenza VLPs with a high HA content resulted from coinfection of HA- and M1-expressing rBVs at MOI of 4 and 2, respectively, and the VLPs (total protein concentration, 1 mg/ml) have an HA titer as high as 2,048 U, as titrated with chicken blood cells. The cVLPs containing flagellin were produced by coinfection of rBVs expressing HA, M1, and flagellin at MOI of 6, 2, and 6, respectively. These cVLPs have HA and M1 contents comparable to those of standard HA and M1 VLPs, as shown by the Western blot results in Fig. 1C, top. The HA titer of the cVLPs (protein concentration, 1 mg/ml) was 2,048 U, the same as that for standard VLPs. When purified recombinant flagellin was used as a standard, a Western blot comparison (not shown) showed that the cVLPs have a flagellin content of about 8 μg/100 μg VLPs. To further confirm the morphology and integrity of these cVLPs, the cVLP samples were examined by electron microscopy after negative staining. As shown in Fig. 1D, enveloped VLPs with projections on the surface were observed, with diameters of about 80 to 100 nm. In addition, the cVLPs have morphological characteristics similar to those of standard HA/M1 VLPs, as described previously (42). These results indicate that membrane-anchored flagellin, together with HA, is incorporated into M1-derived VLPs with a morphology and size similar to those of standard HA/M1 VLPs and influenza virions.

Characterization of the membrane-anchored flagellin in cVLPs.

There are six potential N-linked glycosylation sites in the full-length flagellin sequence with an NXT/S motif (Asn19, 101, 200, 346, 446, and 465, respectively), and four of them are located in the TLR-5-recognizing region (Asn19, 101, 446, and 465). To further characterize the possible glycosylation of the membrane-anchored flagellin in cVLPs, VLPs containing flagellin were treated with PNGase F or endo-H. PNGase F is an amidase which can remove N-linked oligosaccharides from glycoproteins, whereas endo-H cleaves the chitobiose core of high-mannose and hybrid oligosaccharides from N-linked glycoproteins (29). As shown in Fig. 2A, most of the modified flagellin in cell lysates is seen as two main portions on the blot with molecular masses of 55 and 65 kDa (lane 1). The flagellin incorporated into cVLPs corresponds to the upper band (65 kDa) (lane 2). After treatment of VLPs with PNGase F, flagellin bands of faster mobility at 55 kDa were observed (Fig. 2A, lane 3), demonstrating that the membrane-anchored flagellin is glycosylated by N-linked oligosaccharides. When the glycosylated flagellin in VLPs was treated by endo-H, an intermediate band (around 60 kDa) was observed (Fig. 2A, lane 4), revealing the partial sensitivity of the flagellin in VLPs to endo-H. These results indicate that at least some of the oligosaccharides are of the high-mannose type. In conclusion, these results indicate that flagellin in VLPs is glycosylated and that the oligosaccharides are linked to the flagellin peptide backbone by N-type glycosidic linkages.

The adjuvant effect of flagellin is based on its TLR-5-activating activity. To evaluate the ability of the membrane-anchored flagellin in VLPs to function as a TLR-5 ligand, flagellin-containing VLPs were analyzed by a mouse macrophage cell line RAW264.7-based assay (31), and results were compared to those from purified soluble flagellin. As shown in Fig. 2B, flagellin-containing cVLPs stimulated TLR-5-positive RAW264.7 cells to produce TNF-α over a broad concentration spectrum, similar to what was seen with soluble flagellin. The 50% effective concentration (concentration which produces 50% of maximal activity) of flagellin-containing VLPs was around 8 ng/ml, whereas the 50% effective concentration of soluble flagellin was 1 ng/ml. Because the flagellin content in cVLPs is about 8%, the results indicate that the TLR-5 agonist activity of membrane-anchored flagellin in cVLPs is comparable to that of soluble flagellin.

cVLPs containing flagellin induce enhanced humoral immune responses.

It is well recognized that flagellin in full-length, truncated, or fusion protein forms enhances antigen-specific antibody responses (8, 18, 31). To evaluate the ability of membrane-bound flagellin in influenza cVLPs to function as an adjuvant, the humoral immune response against influenza viral antigen was determined for mice immunized with standard HA/M1 VLPs or flagellin/HA/M1 cVLPs. As shown in Fig. 3A, high levels of serum antigen-specific IgG were promoted by priming or priming plus boosting for mice immunized with flagellin-containing cVLPs. A 2,500-fold higher IgG titer was achieved by the flagellin-containing VLP group after only the priming immunization compared with that of the standard HA/M1 VLP group, and this IgG level was comparable to that of the standard VLP group after two immunizations, demonstrating a significant enhancement of responses promoted by the incorporated flagellin. After two immunizations, the IgG level of the cVLP group remained two times higher than that of the standard VLP group (P < 0.05). In contrast, when mice were immunized with mixtures of HA/M1 VLPs plus soluble recombinant flagellin, no significant difference in antibody response was detected compared to what was seen for HA/M1 VLPs alone. These results indicate that the incorporation of the membrane-anchored flagellin into VLPs is important for its adjuvant effect.

Previously, we determined that influenza VLP vaccines induce mixed Th1/Th2-type immune responses (42). To further evaluate the serum antibody response induced by flagellin, production levels of IgG subtypes IgG1, IgG2a, and IgG2b were determined. As shown in Fig. 3B, C, and D, both standard VLPs and cVLPs promoted the production of all three IgG subtypes compared to what was seen for the control (M1-only) VLP group, demonstrating that both Th1 and Th2 immune responses were induced by VLP vaccines; this was consistent with our previous observation (42). However, the flagellin-containing VLPs elicited a level of IgG2a (IgG1/IgG2a ratio, 0.5) significantly higher than that seen for standard VLPs (IgG1/IgG2a ratio, 1.5; P < 0.05), but this was not the case for IgG1, demonstrating that Th1-biased type-mixed responses and IgG2a-dominant class switching were effectively promoted by the incorporation of flagellin compared to standard VLPs.

Flagellin stimulates enhanced virus neutralization and HI activity.

Virus neutralization activity is the most important serological assay to reflect the functional antibodies providing protective immunity. To determine the effects of flagellin on conferring protective humoral responses, sera from mouse groups immunized with HA/M1 VLPs or flagellin-containing HA/M1 VLPs were evaluated for neutralization activities against PR8 virus. As shown in Fig. 4A, sera from standard VLP-immunized mice 3 weeks after the boost immunization showed a neutralization titer (50% plaque reduction) of 1,280. In contrast, the flagellin-containing VLP group showed a virus neutralization titer of 4,000, more than threefold higher, revealing the effectiveness of flagellin incorporated into VLPs as an adjuvant. The enhanced responses were also demonstrated by the HI titers, which are based on blocking the ability of influenza HA to agglutinate erythrocytes by specific antibodies. As shown in Fig. 4B, the flagellin-containing VLP group achieved an HI titer of 1,080, threefold higher than that of the standard VLP group (P < 0.05), which had a mean HI titer of 360. The neutralization activity and HI titers were found to be highly consistent, demonstrating that functional antibodies elicited by influenza VLPs are directed against the HA. Similar to what was found with the serum IgG titers, immune sera from the group immunized with a mixture of soluble flagellin plus HA/M1 VLPs achieved levels of neutralization and HI titers similar to those of the standard HA/M1 VLP group.

A concern for using a protein component as an adjuvant is the antigenicity of the protein itself, and preexisting immunity against flagellin might block its further function as an adjuvant. To evaluate the effects of preexisting antiflagellin antibody, mice were preimmunized intramuscularly twice with 10 μg of recombinant flagellin. Subsequently, the same group was immunized twice with 10 μg of cVLPs at 4-week intervals. As shown in Fig. 4C, this resulted in a significant mean antiflagellin IgG titer of 2.7 ×10⁵, and this titer was stable for 8 weeks. Interestingly, the PR8-specific IgG titers of flagellin-preimmunized mice rose to levels similar to those of the cVLP control group without flagellin preimmunization, as shown in Fig. 4C. In conclusion, flagellin is an effective adjuvant to promote antigen-specific humoral responses when incorporated into VLPs, and the presence of preexisting flagellin immunity was not found to decrease its adjuvant function.

Because cVLPs induced significantly higher humoral responses as described above, we evaluated whether these immune sera confer a cross-reaction with a heterosubtypic virus. Therefore we determined the serum IgG titer and HI titer against a heterosubtypic A/Philippines virus (H₃N₂). We found that though the humoral responses against A/Philippines were at low levels compared to those against A/PR8, flagellin-containing cVLPs induced significantly high IgG (2,800) and HI (60) titers against A/Philippines compared to those induced by standard VLPs (all P values were <0.05) (Fig. 5A and B). These results demonstrated the effects of the membrane-anchored flagellin on broadening the spectrum of immunity to confer heterosubtypic immune responses.

Flagellin promotes antigen-specific T-cell responses.

Several microbial products, such as lipopolysaccharide and CpG DNA, have recently been reported to induce dendritic cell (DC) maturation by binding to TLR family molecules, including TLR-5 (16, 20, 32, 52). DCs are found in physical contact with naive T cells in vivo (12, 63). Consequently, DCs recognize microbial products and activate antigen-specific T-cell clonal expansion. To test the effects of membrane-anchored flagellin on enhancing the production of cytokines, cytokine secretions from splenocytes of mice immunized with flagellin-containing VLPs were determined and compared to those of the standard VLP-immunized mice. The results in Fig. 6 show that spleen T cells from mice immunized with flagellin-containing VLP secreted high levels of IL-2, IFN-γ, TNF-α, and IL-4 when stimulated by MHC-I or -II HA peptides or inactivated PR8 virus compared to those for the standard VLP group (all P values were <0.05). MHC-II-recognized HA peptides induced relatively higher levels of secretion of cytokines, demonstrating a CD4-dominant memory T-cell population; both Th1 (IL-2 and IFN-γ)- and Th2 (TNF-α and IL-4)-type cytokine production was observed. Also, we did not observe detectable TNF-α secretion from splenocytes from the standard VLP group upon stimulation with either MHC-I or -II HA peptides, but splenocytes from flagellin-containing VLP-immunized mice produced high levels of TNF-α.

Flagellin-containing VLPs promote protective immunity against lethal virus challenge.

Providing protective immunity to laboratory animals against lethal virus challenge is the most important goal for preclinical vaccine studies. To determine whether the enhanced antibody and T-cell responses observed in our studies conferred protection, immunized mice were challenged intranasally with mouse-adapted PR8 viruses at 40×LD₅₀. As shown in Fig. 7A and B, mice immunized with the standard VLPs, flagellin-incorporating VLPs, or a mixture of standard VLPs plus soluble flagellin retained their healthy status, as measured by body weight. These groups showed 100% protective immunity to lethal PR8 virus challenge. By contrast, no protection was observed using M1 VLPs, though mice immunized with M1 VLPs showed a minor improvement clinical score, as represented by body weight loss and a 1- to 2-day delay in reaching the endpoint upon lethal virus challenge compared to what was seen for the negative control group.

Because the flagellin-containing VLP group showed much higher antibody levels and enhanced T cellular responses, we tested whether their enhanced immunity provides cross-protection to a heterologous virus challenge. Thus, immunized mice were challenged with an A/Philippines/82 H₃N₂ virus at 40×LD₅₀ per mouse. Data shown in Fig. 7C demonstrate that all VLP groups lost weight, as did the naive group, but the flagellin-containing VLP-immunized group lost less weight. Though there was a 1-day delay compared to the control group, all mice in the standard VLP group or in the group immunized with the standard VLPs plus soluble flagellin, as well as those of the M1 VLP group, reached the endpoint by day 8 or 9. Notably, in the flagellin-containing VLP-immunized group, mice began to regain body weight from day 9 (Fig. 7C), and 67% of mice survived the challenge with the heterologous A/Philippines virus, as shown in Fig. 7D.

Rapid clearance of virus from the body and low virus loads after infection are important signs of decreased morbidity and mortality after infection. To further evaluate the protective immunity induced by flagellin-containing VLPs, immunized mice were intranasally infected with A/PR8 or A/Philippines virus at the same dose used in the challenge experiment. Mice were sacrificed on day 4, and the virus loads in lungs were determined by plaque assay on MDCK cells. As shown in Fig. 8, virus titers were not detected for either the standard or the flagellin-containing VLP groups in the PR8-infected mice. In contrast, naive mice and mice of the M1 VLP group were carrying high virus loads of 1.5 × 10⁹ and 8.5 × 10⁸ PFU/ml of lung extract, respectively. For A/Philippines virus-infected mice, the standard VLP group showed a lung virus titer of 4.2 × 10⁶ PFU/ml of lung extract, whereas the flagellin-containing VLP group showed 1.6 × 10⁴ PFU/ml of lung extract, a relatively low titer compared to that of standard VLP group (P < 0.05). The naive and M1 VLP groups showed titers of 1.5 × 10⁹ and 8.7 × 10⁸ PFU/ml of lung extract, respectively. These results demonstrate that by incorporating flagellin into VLPs as an adjuvant, influenza VLPs induce enhanced immunity, providing complete protection against a homologous virus challenge and significant cross-protection against a heterosubtypic virus challenge.