Lipid nanoparticles as adjuvant of norovirus VLP vaccine augment cellular and humoral immune responses in a TLR9- and type I IFN-dependent pathway

It has been reported that LNPs may have intrinsic adjuvant activity (30–33). To further verify that LNPs might be repurposed as a standalone adjuvant for protein vaccine platforms, we explored the adjuvant activity of LNPs for norovirus bivalent VLP vaccines with the genotype GI.1/GII.4. VLPs were prepared as described previously (7) and detected using SDS-PAGE and transmission electron microscopy (TEM) (Fig. S1a through c). LNPs were prepared at a ratio of four lipids (Dlin-MC3-DMA:DSPC:cholesterol:DMPE-PEG2000) of 50:10:38.5:1.5. TEM and dynamic light scattering (DLS) analyses showed that the average particle size of the LNPs was 219 nm and the polydispersity index was 0.059 in phosphate-buffered saline (PBS) (Fig. S1d and e). Zeta potential measurements showed a potential of −5.94 mV in PBS (Fig. S1f). Mice were then immunized intramuscularly (i.m.) with GI.1 and GII.4 VLPs combined with low, medium, and high doses of LNP individually at weeks 0 and 4, as depicted in Fig. 1a. Ionized lipids removed from LNP (IRL) at a medium dose and 250 µg Alum at an optimized dose were selected as controls. The optimal dose of Alum adjuvant for GI.1/GII.4 VLPs was determined in a dose-dependent manner (Fig. S1g through i).

Both anti-GI.1 and GII.4 VLP IgG titers showed an LNP dose-dependent trend after the initial immunization and increased significantly at week 2 (Fig. 1b and c). IgG antibodies were significantly higher in the medium or high LNP + GI.1/GII.4 group than in the Alum + GI.1/GII.4 group at week 2 (P < 0.05). These differences were not significant after the booster immunization. The IRL + GI.1/GII.4 induced low levels of IgG antibody titers as GI.1/GII.4 VLP without adjuvant addition, which is consistent with previous studies showing that ionized lipids (Dlin-MC3-DMA) are a vital component of LNP adjuvant activity (32), despite the different ionizable lipids used.

Next, we assessed the ability of the antisera to block the interaction of VLPs with histo-blood group antigens (HBGAs), a binding receptor for NoVs, which pig gastric mucin-type III was used as the source of HBGAs. Blockade antibody levels are correlated with protection against clinical NoV gastroenteritis (34). Both GI.1- and GII.4-specific blocking antibodies showed an LNP dose-dependent trend after the initial immunization, which obviously increased at week 2. Specifically, compared with the Alum + GI.1/GII.4 group, the low, medium, and high LNP + G1.1/GII.4 groups showed much higher GI.1- and GII.4-specific blocking antibody titers at weeks 2, 4, and 6. However, at week 12, the blocking antibody titers were similar between high LNP + G1.1/GII.4 group and Alum + GI.1/GII.4 group. The IRL + GI.1/GII.4 group induced comparable GI.1- and GII.4-specific blocking antibody titers in the G1.1/GII.4 group (Fig. 1h and i). The results from the blocking antibodies resembled those from the binding IgG antibodies. In addition, the neutralizing ability was tested by VLP-HiBiT entry assay. Consistent with HBGA blocking results, LNP-adjuvanted groups produced earlier neutralizing antibodies at week 2. And the neutralizing antibody titer in the Alum-adjuvanted group was similar to that in the high LNP + GI.1/GII.4 group at week 12 (Fig. 1h and i).

We further detected the magnitude of the IgG subclass profile, including IgG1, IgG2a, IgG2b, and IgG3, against GI.1 or GII.4 VLPs at week 6 post-priming. The low, medium, and high LNP + GI.1/GII.4 groups induced much higher GI.1 or GII.4 VLP-specific Th1 subclass antibody IgG2a, IgG2b, and IgG3 titers, while inducing comparable Th2 subclass antibody IgG1 titers compared with the Alum + GI.1/GII.4 group (Fig. 1d and e). The IgG2a/IgG1 ratios in the LNP-adjuvanted group were much higher than those in the Alum group, indicating that the LNP adjuvant improved the Th1 subclass antibody response compared with Alum (Fig. 1f and g). To further verify this phenomenon, we detected IgG subclass dynamics. All LNP groups induced much higher Th1 subclass antibody IgG2a, IgG2b, and IgG3 titers than Alum at most time points (Fig. S2a through d). The Alum-adjuvanted group induced the highest Th2 subclass antibody IgG1 at weeks 8 and 12. The IgG2a/IgG1 dynamic ratio showed that the LNP adjuvant greatly enhanced the IgG2a/IgG1 ratio at all time points compared with the Alum-adjuvanted vaccine (Fig. S2e and f), which indicates a switch from a Th2 biased response to a more balanced Th1/Th2 response than Alum.

Because germinal centers (GCs) are essential for the development of high-quality antibody responses (35), antigen-specific GC B cells from mouse draining lymph nodes (dLNs), including the popliteal, inguinal, and iliac LNs, were detected on day 10 after immunization. The high LNP + GI.1/GII.4 group showed significantly higher total GC B cell numbers (P < 0.01) and GI.1 (P < 0.01) or GII.4 (P < 0.05) VLP-specific GC B cell numbers than the Alum + GI.1/GII.4 group (Fig. 3f through h). The stronger GC B cell responses may be related to the earlier humoral immune responses induced by LNPs, as shown in Fig. 1.

These experiments demonstrate that the LNP formulation used in this study has intrinsic adjuvant activity in assisting bivalent NoV VLP vaccines, which relies on ionizable lipid components. In addition, compared with the conventional Alum adjuvant, the LNP adjuvant could induce stronger and earlier antibody responses and a more balanced Th1/Th2 response than Alum. Despite early advantages in the LNP-adjuvanted group, the antibody responses in Alum-adjuvanted group was still able to "catch up" over time.

To detect the memory T cell immune responses, mouse splenocytes were harvested and re-stimulated in vitro with GI.1 or GII.4 VLPs and then detected using enzyme-linked immunospot (ELISpot) and intracellular cytokine staining assays at week 12. It showed that the number of Th1-type cytokine-producing splenocytes, such as IFN-γ or IL-2, significantly increased in medium LNP + GI.1/GII.4 and high LNP + GI.1/GII.4 groups compared with the Alum + GI.1/GII.4 group and showed a dose effect in three LNP-adjuvanted groups (Fig. 2a and b). Splenocytes in these three LNP-adjuvanted groups also produced much higher levels of IL-4 than Alum; however, the IL-4 level was lower than that of IL-2 and IFN-γ. The ELISpot results were similar under both GI.1 or GII.4 VLP stimulation.

Intracellular cytokine staining assays showed that the frequencies of CD4⁺ T cells with IL-2 production were much higher in the three LNP-adjuvanted groups than that in the Alum-adjuvanted group in a dose-dependent manner. The frequencies of IFN-γ-producing CD8⁺ T cells were obviously higher in the three LNP-adjuvanted groups than that in the Alum-adjuvanted group under GII.4 stimulation; however, no discernible differences were observed under GI.1 stimulation (Fig. 2c through f). To further ascertain the cellular immune response, effector T cells were detected 10 days after a single injection (Fig. 3a through e). Consistent with the memory T cell response, the frequencies of CD4⁺ T cells with IL-2 and IFN-γ production were significantly higher in low- and high-dose LNP-adjuvanted groups than that in the Alum-adjuvanted group. The frequencies of IFN-γ-producing CD8⁺ T cells were significantly higher in high LNP + GI.1/GII.4 groups than that in the Alum + GI.1/GII.4 group under GII.4 stimulation; however, no discernible differences were observed under GI.1 stimulation.

Overall, the LNP adjuvant significantly enhanced the frequency of both Th1-type cytokine-producing T cell responses specific to GI.1/GII.4 VLP compared with Alum.

Having established the intrinsic adjuvant effect of LNPs on NoV VLP vaccines, we further explored the underlying mechanism of LNPs in initiating an innate immune response, which is required for generating an effective adaptive immune response. DLNs were collected 24 h after immunization and subjected to RNA-seq. RNA-seq revealed that a large number of DEGs compared with naïve mice were induced in the high LNP, high LNP + GI.1/GII.4, and low LNP + GI.1/GII.4 groups, while lower numbers of DEGs were induced in the Alum, Alum + GI.1/GII.4, and GI.1/GII.4 groups. Accordingly, the number and expression levels of DEGs were relatively lower in the low LNP + GI.1/GII.4 than in the high LNP and high LNP + GI.1/GII.4 groups (Fig. 4a and b).

Further analyzing the immune function-related genes in each group, we found that LNPs upregulated numerous I-IFN pathway-related genes, such as IFN-α, IFN-β, C-X-C motif chemokine ligand (CXCL) 9, CXCL10, CXCL11, and interferon regulatory factor 7 (IRF7), in a dose-dependent manner. Genes for TLRs (TLR3, TLR7, and TLR9) and RIG-I-like receptors [ddx58 (RIG-1) and ifih1 (MDA-5)] were also upregulated, indicating that LNPs may enter the cell and activate receptors in the endosomal compartment and the cytoplasm. In addition, cytokine-related genes, such as IL-1β, IL-6, IL-10, and IL-1rn, and antigen processing and presentation genes, such as histocompatibility 2, T region locus 24 (H2-T24), antigen peptide transporter 2 (TAP2), and Fc gamma receptor 1 (Fcgr1), as well as the costimulatory molecules CD80, CD86, and CD40, were upregulated obviously in the LNP-adjuvanted groups (Fig. 4c). DEGs between the high LNP and naïve groups were used to perform pathway enrichment analysis using the reactome and Kyoto Encyclopedia of Genes and Genomes (KEGG). I-IFN signaling, class I MHC-mediated antigen processing and presentation, and TLR signaling pathways were significantly enriched (Fig. 4d and e).

The relative protein levels of inflammatory cytokines or chemokines in dLNs 24 h after immunization were further analyzed using a mouse cytokine array. Chemokines and cytokines such as CXCL9, CXCL10, CXCL11, chemokine (C-C motif) ligand (CCL) 2, CCL12, CCL3, IL-1a, IL-1ra, and tissue inhibitor of metalloprotease-1 (TIMP-1) were upregulated in the high LNP, low LNP + GI.1/GII.4, and high LNP + GI.1/GII.4 groups in a dose-dependent manner compared with the naïve, Alum, Alum + GI.1/GII.4, or GI.1/GII.4 groups (Fig. 4f and g). This result was consistent with the RNA-seq results, which indicated that LNPs could induce the pro-inflammatory microenvironment and activate the innate immune response, especially through the I-IFN signaling pathway.

To elucidate whether the pro-inflammatory environment induced by LNPs translates into activation and changes in the composition of immune cells, we analyzed the number and activation status of leukocytes in dLNs (Fig. 5a through d). Compared with PBS, low and high LNP significantly increased the number of monocytes (CD11b⁺ Ly6C⁺) at 24–72 h, the number of neutrophils (CD11b⁺ Ly6G⁺) at 4–72 h after injection (P < 0.05), and the number of B cells at 72 h (P < 0.01). However, the numbers of cDC (CD11c⁺ MHCII⁺), pDC (CD11b⁻ CD11c^int PDCA-1⁺), and macrophages (CD11b⁺ F4/80⁺) changed slightly over time. Furthermore, changes in cellularity in the dLNs were accompanied by strong activation of innate immune cells, as demonstrated by CD69 expression on neutrophils, macrophages, CD4⁺ T cells, CD8⁺ T cells, and B cells, peaking at 24 h and falling at 72 h post-injection (Fig. 5c). Although the number of cDC and pDC changed slightly and even decreased at 24 h, vigorous activation of these two cell types was observed by CD86 expression at 24 h and was sustained until 72 h, indicating their vital role in initiating the immune response by LNPs (Fig. 5d).

Because the I-IFN signaling pathway was significantly enriched by RNA-seq, its impact on LNP-activated innate immune responses was further explored. Twenty-four hours after injection with high LNP, the innate immune responses were detected in lymphocytes from dLNs in wild-type (WT) and IFNαR1^−/− mice by flow cytometry. The data showed that, compared with WT mice, the activation of innate immune cells, measured by CD69 or CD86 expression, was significantly downregulated in many cell types, including macrophages, CD4⁺ T cells, CD8⁺ T cells, B cells, NK cells, monocytes, cDCs, and pDCs (Fig. 5e and f). Moreover, I-IFN-related or interferon-stimulated genes (ISGs) in LNs were detected using qPCR. The data showed the expression of IRF7, CXCL9, CXCL10, IFN-γ, CCL2, and IFIT3 were significantly downregulated in IFNαR1^−/− mice at 24 h (Fig. 5g). These results indicate that i.m. injection of the LNP adjuvant induces an extensive but transient local immunostimulatory microenvironment and that the I-IFN signal is vital for the activation of the innate immune response by LNPs.

Because LNPs induced vigorous and sustained dendritic cell (DC) activation in vivo in mice, we further tested the effect of LNPs on Bone marrow-derived dendritic cell (BMDC) activation in vitro. BMDCs were obtained by culturing isolated mouse bone marrow cells in the presence of granulocyte-macrophage colony-stimulating factor (GM-CSF), and the purity of the BMDCs reached >90% using flow cytometry (Fig. S3d). LNPs significantly upregulated the expression of costimulatory molecules such as CD80, CD86, MHC I (H-2Kb/H-2Db), and MHC II (I-A/I-E) in BMDCs in a dose-dependent manner (Fig. 6a), indicating BMDC maturation and activation. To further exclude the effect of endotoxin contamination, BMDCs were stimulated with polymyxin B (PMB) or PMB with LNPs for 24 h. PMB, which blocks the biological effects of lipopolysaccharide by binding to lipid A, did not block the activation of BMDCs by LNPs (Fig. 6a). This confirmed that LNP-activating BMDCs were not due to lipopolysaccharide contamination.

Since RNA-seq showed that LNPs induced extensive upregulation of I-IFN-related genes, the expression of these genes in BMDCs stimulated with LNPs for 8 h was detected using qPCR. The results showed that LNPs significantly induced the expression of IFN-α1, IFN-α4, IFN-β1, and CXCL10 in a dose-dependent manner (Fig. 6b). Therefore, LNPs can induce the I-IFN signaling pathway in BMDCs.

Since TLR9, TLR7, and TLR3 were significantly upregulated by LNPs and have been reported to be closely related to I-IFN signaling, we further determined whether LNP-induced BMDC activation depends on the TLR signaling pathway. BMDCs were pretreated with small-molecule inhibitors of TLR4 (TLR4i, TAK-242), TLR3 (TLR3i, CU CPT4a), TLR7 (TLR7i, eupatorine), or TLR9 (TLR9i, ODN2088) separately, followed by the addition of LNP for 8 h. Then, the expression of IFN-β1 and CXCL10 was analyzed using qPCR (Fig. 6c). It showed that TLR9 inhibitor significantly reduced CXCL10 (P < 0.05) and IFN-β1 (P < 0.01) expression in LNP100-treated BMDCs and that TLR3 and TLR7 inhibitors partially reduced IFN-β1 and CXCL10 expression in LNP100-treated BMDCs, with statistical significance just for IFN-β1. However, the TLR4 inhibitor only partially inhibited the production of both cytokines, with no statistical significance. Consistent with the effects of TLR9 inhibitors in vitro, mice injected with TLR9 inhibitors significantly inhibited LNP-stimulated I-IFN signaling-related genes, IRF7, CXCL9, CXCL10, CCL2, and IFIT3 (Fig. 6d).

In summary, LNPs activate I-IFN signaling partially by activating the TLR9 pathway, thereby promoting DC activation for efficient antigen presentation.

I-IFN is a cytokine family with broad functions, including the activation of the innate immune response and the shaping of the adaptive immune response. I-IFN can aid in the activation of DCs and lead to the activation and type 1 immune polarization of T cells (26). Therefore, we explored the effect of I-IFN signaling on the adaptive immune response induced by LNP-adjuvanted NoV VLP vaccines. WT or IFNαR1^−/− mice were i.m. injected with high LNP + GI.1/GII.4 VLP. Effector T cells on day 10 and memory T cells at week 6 from mouse spleens were detected using flow cytometry or ELISpot after restimulation with GI.1 or GII.4 VLPs in vitro (Fig. 7a).

The flow cytometry results showed that on day 10, IFN-γ-producing CD4⁺ and CD8⁺ T cell ratios were significantly lower in IFNαR1^−/− mice compared with those in WT mice. There is a significant decrease in IL-2-producing CD4⁺ T cell ratios under GI.1 VLPs stimulation (P < 0.05), and there was no statistical significance for decrease in IL-2-producing CD4⁺ T cell ratios under GII.4 VLPs stimulation(Fig. 7b and c). At week 6, memory T cells showed a similar decreasing trend in IFNαR1^−/− mice compared with WT mice; however, there was no statistical significance (Fig. 7d and e).

The results from the ELISpot assay showed that a significantly lower level of IFN-γ was produced by splenocytes from IFNαR1^−/− mice compared with that from WT mice on day 10 or at week 6 after immunization (Fig. 7f and g). Similar results were observed for IL-2 and IL-4 on day 10 and week 6, although the differences were seldom statistically significant.

In addition, GI.1 and GII.4 VLP-specific IgG, IgG1, IgG2b, IgG2c, and IgG3 and blocking antibody and neutralizing antibodies were detected. Compared with WT mice, GI.1 VLP-specific IgG, IgG2b, IgG2c, and IgG3 and blocking and neutralizing antibodies were significantly reduced in IFNαR1^−/− mice (Fig. 7h) and GII.4 VLP-specific IgG, IgG2c, and IgG3 and blocking and neutralizing antibodies showed similar decreased trends in IFNαR1^−/− mice, with statistically significant results just for IgG2c. In addition, Th2 subclass IgG1 antibody titers were not obviously affected specific for GI.1 VLPs and even significantly higher specific for GII.4 VLPs in IFNαR1^−/− mice than WT mice (Fig. 7i). In contrast, the immune response of Alum-adjuvanted VLPs were not apparently affected in IFNαR1^−/− mice compared to those in WT mice as shown in Fig. S4.

These data show that the LNP-mediated I-IFN signaling pathway is critical for the induction of Th1-type cytokine-expressing CD4⁺ and CD8⁺ T cell immune responses as well as for blocking antibody and Th1 subclass antibody production.

Female BALB/c and C57BL/6J mice (6–8 weeks old) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. IFNαR1 knockout mice were purchased from Shanghai Model Organisms Center, Inc.

GI.1 VLPs (GenBank ID: NP_056821.2↗) and GII.4 VLPs (GenBank ID: KC631827.1↗) were produced as described in a previous study (7). Briefly, GI.1 and GII.4 VLP expression strains were induced to express VLP proteins according to the Pichia pastoris Expression Manual (Invitrogen). Yeast cells, after induced expression, were collected and resuspended in 0.15 M PBS containing 1 mM phenylmethylsulfonyl fluoride, homogenized at 4°C with 1,500 bar, and then the supernatant was collected after centrifugation at 12,000 rpm for 30 min. Then, 7% PEG6000 and 0.2 M NaCl were added to concentrate the supernatant. And the pellet was resuspended in 0.15 M PBS and dissolved overnight at 4°C. The next day, after centrifugation at 12,000 rpm for 30 min, the supernatant was retained and subjected to sucrose cushion and sucrose gradient ultracentrifugation as previously described (7). Fractions containing the VLPs were collected and diluted with PBS. The concentrated VLPs were quantified using SDS-PAGE and Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, USA).

The LNPs were prepared using a method described before (69, 70). Briefly, lipids were dissolved in ethanol and mixed at a molar ratio of 50:10:38.5:1.5 (Dlin-MC3-DMA:DSPC:cholesterol:DMPE-PEG2000). The lipid mixture was mixed with 50 mM citrate buffer (pH 4.0) at a ratio of 1:3, then dialyzed with PBS and concentrated by 10 kd Amicon ultracentrifuge filter (EMD Millipore, Billerica, MA). The concentrated preparation was passed through a 0.45 µm filter and stored at 4°C until use. And the particle size was detected by DLS using a Zetasizer Nano ZS90 (Malvern Instruments) at room temperature. Zeta potential measurements were conducted on the same instrument as DLS.

Electron microscopy was performed as described previously (7). VLPs and LNPs were negatively stained with 2% phosphotungstic acid, and TEM was performed using a Hitachi HT7700 transmission electron microscope (Hitachi, Japan) operated at 80 kV to analyze the shape and integrity of the VLPs and LNPs.

The different immunization regimens were presented as follows.

VLP-specific total IgG and IgG1/IgG2a/IgG2b/IgG3/IgG2c subclass antibodies were detected using ELISA, as described in detail elsewhere (7). Briefly, sera were diluted twofold at an initial dilution of 1:500 and incubated with NoV VLP-coated plates. The plates were then incubated with a 1:50,000 dilution of horseradish peroxidase (HRP)-conjugated anti-mouse IgG (Abcam, UK) or a 1:5,000 dilution of HRP-conjugated anti-mouse IgG1/IgG2a/IgG2b/IgG3/IgG2c antibody (Southern Biotech, USA). After washing, 3,3',5,5'-tetramethylbenzidine (TMB) substrate (NCM Biotech, China) was added and then were stopped by adding 2 M sulphuric acid. Optical density was measured at 450 nm using a microplate spectrophotometer (Agilent, USA). The average endpoint titer for each serum sample was determined as the reciprocal of the highest serum dilution with an optical density above the set cutoff value.

The ability of antisera to block the interaction of VLPs with HBGAs was evaluated using a blocking assay, as described in detail elsewhere (7). Briefly, 10 µg/mL pig gastric mucin-type III was coated on an ELISA plate. Individual sera were subjected to twofold gradient dilutions with an initial 100-fold dilution, and equal volumes of mixtures of NoV VLPs and serially diluted serum samples were incubated at room temperature for 1 h and 15 min before being added to the plates. Bound VLPs were detected using NoV VLP type-specific antibodies, followed by incubation with HRP-conjugated goat anti-rabbit IgG (Abcam, UK) and TMB substrate. The OD450 was measured as described above. Maximum binding was determined using VLPs without mouse serum. The blocking titer 50 (BT50) was expressed as the reciprocal of the highest serum dilution that blocked 50% of the maximal VLP binding. An arbitrary titer (BT50 = 50) was assigned to samples with a blocking index of <50% at the lowest serum dilution of 1:100.

NoV VLP-HiBiT-based live-cell entry assay system (a gift from professor Huang and Zhang) was adopted to detect the neutralizing antibodies. According to the previous report (41), 293T-FUT2-LgBiT cells were cultured in poly-L-Lys-coated 96-well plates at a concentration of 4 × 10⁴ cells per well in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS), sodium pyruvate, 100 U/mL penicillin, and 100 µg/mL streptomycin at 37°C for 24 h. Then, the cells were chilled at 4°C for 1 h. Meanwhile, an equal-volume mixture of VLP-HiBiT with 1.5 µg/mL and serially diluted serum samples were incubated for 1 h at room temperature and then chilled at 4°C for 10 min. After removal of the culture supernatant, cooled VLP/serum mixtures were added to prechilled 293T-FUT2-LgBiT cells and incubated at 4°C for 1 h. After washing with cold PBS, Nano-Glo Live Cell Substrate (Promega) diluted in Opti-MEM was added to the wells and then the plates were incubated at room temperature for 1 h in the dark, followed by measurement of cellular luciferase signal.

ELISpot assays were performed on isolated murine splenocytes as previously described (7). Multiscreen HTS-IP filter plates (Millipore, USA) were coated with anti-mouse IFN-γ, IL-2, and IL-4 antibodies according to the manufacturer's instructions (Mabtech, Sweden). The resuspended splenocytes were plated at a density of 5 × 10⁵ cells/well. VLPs were diluted in RPMI 1640 medium to a final concentration of 20 µg/mL and added to the plate. Negative (RPMI 1640 medium) and positive controls (Cell Stimulation Mix; Thermo Fisher Scientific) were added to each assay. Plates were incubated for 48 h at 37°C and 5% CO₂ and then were incubated by adding biotinylated anti-mouse IFN-γ, IL-2, and IL-4 antibodies, HRP-conjugated streptavidin, and TMB substrate. Spots were counted using an Immuno-spot analyzer (Cellular Technology Ltd., USA). The results are expressed as the number of spot-forming cells in 5 × 10⁵ cells.

Mouse dLNs from all groups were used for cytokine profiling, and samples from two mice in each group were mixed. The dLNs were homogenized in PBS containing protease inhibitors. Triton X-100 was then added at a final concentration of 1%. Samples were frozen at −80°C, thawed, and centrifuged at 10,000 × g for 5 min to remove cellular debris. The relative levels of different cytokines were assessed using the mouse cytokine array Panel A (Cat#ARY006, R&D Systems), according to the manufacturer's instructions. Pixel density was evaluated using ImageJ Software.

RNA-seq and differential expression analysis were performed at Novogene Corporation, which included read count normalization, model-dependent P-value estimation, and false discovery rate value estimation based on multiple hypothesis testing. This was preceded by raw read filtering, mapping clean reads to a reference genome using HISAT2, and determining the fragments per kilobase per million mapped fragments values for all samples. Differentially expressed genes were evaluated based on their log2 fold change and adjusted P-values. Those with |log2 (foldchange)| ≥ 1 and adjusted P-values ≤0.05 were considered to be differentially expressed and significant. Pathway analysis of significantly differentially expressed genes identified from the RNA-seq results was performed using reactome and KEGG in R Studio v.3.0.3. Pathways and biological processes with adjusted P-value ≤0.05 were considered significant.

Mouse spleens were homogenized with a syringe plunger and passed through 40 µm cell strainers to produce single cells. Erythrocytes were lysed in lysis buffer (10 mM KHCO₃, 150 mM NH₄Cl, and 10 mM EDTA, pH 7.4). DLNs (popliteal, inguinal, and iliac LNs) were harvested after immunization and digested with a combination of 1 mg/mL type IV collagenase (Cat#c5138, Sigma-Alidrich) and 100 µg/mL DNAase I (Cat#D8071, Solarbio) at 37°C for 20 min, homogenized with a syringe plunger, and filtered through a 40 µm cell strainer to make a single-cell suspension. Cell viability was calculated to be 95%–100%, as determined by trypan blue exclusion.

All staining steps were performed at 4°C in PBS buffer (containing 0.5% bovine serum albumin and 50 mM EDTA-2Na). Single-cell suspensions were blocked with an anti-CD16/CD32 monoclonal antibody (Thermo Fisher Scientific) and stained using the LIVE/DEAD staining kit (Invitrogen). After staining for surface markers or intracellular cytokines, all samples were fixed with 1% paraformaldehyde for 30 min, collected on a BD LSRFortessa (BD Biosciences), and analyzed using FlowJo v.10.

BMDCs were derived from 6- to 8-week-old female BALB/c mice and cultured as described previously (66). Bone marrow cells were isolated from mouse leg bones under sterile conditions. Then, 2 × 10⁶ bone marrow cells were seeded in a 90 mm diameter culture dish in RPMI 1640 medium containing 10% fetal bovine serum, 50 µM β-mercaptoethanol, and 20 ng mL⁻¹ of GM-CSF. On the 3rd day, 10 mL of medium (RPMI 1640 medium containing 10% FBS, 50 µM β-mercaptoethanol) containing GM-CSF (20 ng mL⁻¹) was added. On day 6, the cells were centrifuged and resuspended in 10 mL of culture medium to continue the culture. The BMDCs were collected on day 9 for subsequent experiments.

BMDCs (1 × 10⁶ cells/well) were seeded on a 24-well plate and stimulated with 25, 50, and 100 µg/mL LNP, as well as with PMB (10 µg/mL), PMB + 100 µg/mL LNP for 24 h, and the same volume of media was added in cells as a medium control. Cells were collected for flow cytometry to detect the expression of CD80, CD86, MHC I (H-2Kb/H-2Db), and MHC II (I-A/I-E) in the CD11c⁺ population. For cytokine expression analysis, BMDCs were stimulated with 25, 50, and 100 µg/mL LNP for 8 h, and cellular RNA was extracted to measure the expression levels of cytokines using real-time PCR.

BMDCs were seeded in a 24-well plate at a density of 1 × 10⁶ cells/well, and pretreated with TLR3 inhibitor CU CPT 4a (25 µM), TLR4 inhibitor TKA-242 (100 nM), TLR7 inhibitor (eupatorine, 10 µM), or TLR9 inhibitor ODN2088 (10 µM) for 1 h. Subsequently, the cells were stimulated with 100 µg/mL LNP for 8 h, and RNA was extracted to measure the expression levels of cytokines using real-time PCR.

RNA was extracted from cells using an RNA isolation kit (Cat#AC0205, Sparkjade). RNA was reverse transcribed using the Hifair III 1st Strand cDNA Synthesis SuperMix (Yeasen). Real-time PCR was performed using the SYBR Premix kit (Yeasen) and analyzed using a Roche 480 Real-Time PCR System (Eppendorf, Germany). The primers used in this study are listed in Table S7.

The Prism 9.5.1 software (GraphPad, USA) was used to perform statistical analyses. Data are shown as mean ± SEM. For multiple-group comparisons, one-way or two-way analysis of variance was used, followed by Tukey's post hoc multiple comparison test. Statistical differences were considered significant when P-values ≤0.05.