CXCL13 promotes broad immune responses induced by circular RNA vaccines

Abstract

Antibody responses induced by current vaccines for influenza and SARS-CoV-2 often lack robust cross-reactivity. As hubs where diverse immune cells converge and interact, the alterations in the immune microenvironment within lymph nodes (LNs) are intricately linked to immune responses. Herein, we designed a lipid nanoparticle (LNP) loaded with circular RNA (circRNA) and targeted to LNs, in which CXCL13 was directly integrated into antigen-encoding circRNA strands. We demonstrated that CXCL13 alters the transcriptomic profiles of LNs, especially the upregulation of IL-21 and IL-4. Meanwhile, CXCL13 promotes the formation of germinal center and elicits robust antigen-specific T cell responses. With the codelivery of CXCL13 and the antigen, CXCL13 enhances cross-reactive antibodies against influenza virus and SARS-CoV-2, achieving protection against both homologous and heterologous influenza virus challenges in a mouse model. Notably, the targeted modification of LNP surfaces with antibodies helps address some of the challenges associated with lyophilized LNP vaccines, which is crucial for the long-term storage of LNP-circRNA vaccines. Overall, the circRNA-based antigen-CXCL13 coexpression system developed herein provides a simple and robust platform that enhances the magnitude and breadth of antibody responses against multiple viral glycoproteins, highlighting the potential utility of CXCL13 in inducing broad immune responses.

Keywords: CXCL13; SARS-CoV-2; broadly cross-reactive antibodies; circRNA vaccine; influenza virus.

Fig. 1.

Design and characterization of HA-CXCL13-circRNA tLNPs vaccines targeted to LNs. (A) Schematic diagram of circRNA circularization by group I intron autocatalysis. When circRNA enters the cell, the internal ribosome entry site (IRES) sequence recognizes the ribosome, facilitating the direct completion of the protein expression process. (B) Schematic diagram of the construction of LN targeted-LNPs (tLNPs). The targeting antibodies (anti-DEC-205) are chemically conjugated with the LNPs to the maleimide group in the lipid DSPE-PEG-maleimide. (C) Representative transmission electron microscopy image of LNPs or tLNPs. (Scale bar, 500 nm.) (D) CircRNA was transfected into HEK-293T cells, and the protein concentration in HEK-293T cells was measured by ELISA 48 h after transfection. (E–G) Encapsulation efficiency (E), particle size (F), and polydispersity index (PDI) (G) of LNPs and tLNPs. (H) The antigen distribution of HA-circRNA LNPs and HA-circRNA tLNPs in various tissues. The pie chart represents the average values from three independent experiments. Each segment of the pie chart reflects the proportionate data calculated based on these replicates. (I-K) Schematic illustration of the experimental design (I). BALB/c mice (n = 5) were immunized twice with 10 µg of HA-CXCL13-circRNA LNPs or HA-CXCL13-circRNA tLNPs. Serum samples were collected at 14 and 35 d after the primary immunization for IgG titer (J) and hemagglutination inhibition (HAI) titer (K) detection. Data are represented as the mean ± SD. Unpaired two-tailed Student’s t test was performed for comparison, as indicated in the figures; **P < 0.01; ***P < 0.001; ns, not significant.

Fig. 2.

Transcriptome analysis of LNs revealed distinct immune response between CXCL13 expression and control mice. C57BL/6 mice (n = 5) were immunized via intramuscular injection with 10 µg of HA-CXCL13-circRNA tLNPs or HA-circRNA tLNPs. Two weeks after the initial immunization, a booster dose of the same amount was administered. Inguinal dLNs were collected 2 d after the booster immunization for RNA sequencing. (A) The PCA model plot of transcriptomic profiles from the inguinal dLNs of mice inoculated with HA-CXCL13-circRNA tLNPs and HA-circRNA tLNPs. PCA Plot showing the global differences between CXCL13 expression and control groups. (B) Volcano plots of differentially expressed genes between different groups. The red dots represent genes that are up-regulated, the blue dots represent genes that are down-regulated, and the gray dots represent genes that are not different. FDR < 0.01, |log₂ (fold change)| > 1. (C) KEGG enrichment analysis of differentially expressed genes. The data are represented as circles, where the size indicates the gene count for that particular process, and the color represents the –Log₁₀ P-value calculated with one-sided Fisher’s exact test with Benjamini–Hochberg correction. (D) Heatmap visualization of scaled gene expression levels for selected pathways of interest. (E) The expression of IL-21 and IL-4 in mouse LNs inoculated with HA-circRNA tLNPs or HA-CXCL13-circRNA tLNPs. Data are represented as the mean ± SD. Unpaired two-tailed Student’s t test was performed for comparison, as indicated in the figures; ***P < 0.001. PCA: principal component analysis; KEGG: Kyoto Encyclopedia of Genes and Genomes; FDR: false discovery rate.

Fig. 3.

HA-CXCL13-circRNA enhanced both humoral and cellular immune responses. C57BL/6 mice were immunized with HA-CXCL13-circRNA tLNPs or HA-circRNA tLNPs on days 0 and 14, and inguinal dLNs were harvested for FCM analysis. (A) Representative flow cytometry plots of GC B cells (gated on CD3⁻B220⁺CD95⁺GL7⁺ cells) in dLNs from days 0, 7, 14, 21, 28, 35, and 42. See SI Appendix, Fig. S4 for the full gating strategy. (B) Representative flow cytometry plots of GC Tfh cells (gated on B220⁻CD3⁺CD8⁻CD4⁺CXCR5^hiPD-1^hi cells), mantle Tfh (mTfh, gated on B220⁻CD3⁺CD8⁻CD4⁺CXCR5^intPD-1^int), and non-Tfh (gated on B220⁻CD3⁺CD8⁻CD4⁺CXCR5⁻) in dLNs from days 0, 7, 14, 21, 28, 35, and 42. See SI Appendix, Fig. S4 for the full gating strategy. (C) Numbers of GC B cells over time (n = 5 mice per time point). (D) Numbers of GC Tfh cells over time (n = 5 mice per time point). (E) Total number of Tfr cells (gated on B220⁻CD3⁺CD8⁻CD4⁺PD-1⁺BCL6^hiFoxP3⁺) in dLNs from days 14, 21, and 28 (n = 5 mice per time point). (F and G) Statistical results of CD4⁺ TEM (gated on B220⁻CD3⁺CD8⁻CD4⁺CD44⁺CD62L⁻) and TCM (gated on B220⁻CD3⁺CD8⁻CD4⁺CD44⁺CD62L⁺) in the spleen at day 21 (n = 5 mice per group). (H and I) Statistical results of CD8⁺ TEM (gated on B220⁻CD3⁺CD4⁻CD8⁺CD44⁺CD62L⁻) and TCM (gated on B220⁻CD3⁺CD4⁻CD8⁺CD44⁺CD62L⁺) in the spleen at day 21 (n = 5 mice per group). (J and K) Number of cytokine-producing CD4⁺ T cells (J) and CD8⁺ T cells (K) in the spleen at day 21 postimmunization (n = 7 mice per group). Pie charts representing the proportions of cytokine-producing cells positive for one, two, or three cytokines. See SI Appendix, Fig.S5 for the full gating strategy. Data are represented as the mean ± SD. Statistical significance was determined by one-way or two-way ANOVA with Tukey’s multiple comparisons; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant.

Fig. 4.

CXCL13-circRNA-adjuvanted HA-circRNA immunization elicits cross-reactive antibody responses. (A) BALB/c mice (n = 5) were intramuscularly injected with PBS, HA-CXCL13-circRNA tLNPs, or HA-circRNA tLNPs. After 2 or 7 d, the inguinal dLNs were collected and analyzed for weight. The results are presented as the fold of the LN weight from the PBS-injected contralateral side. (B) The GC response in the inguinal dLNs (day 10) was analyzed by immunofluorescence. LN sections were stained for B220 (blue) and GL7 (red). Sizes of GCs in representative immunofluorescent images (Right panel). (C–E) BALB/c mice (n = 8) were immunized by intramuscular injection with 10 µg of HA-CXCL13-circRNA tLNPs or HA-circRNA tLNPs and boosted with the same dose after 2 wk. Serum samples were collected at 14 and 35 d after primary immunization and measured by ELISA for antibody levels. (C) Serum IgG responses were measured against HA (PR8, CF09, HK03, and SH13) proteins after immunization with HA-CXCL13-circRNA or HA-circRNA (PR8) vaccines. (D) Serum IgG responses were measured against HA (V11, P09, HK08, and SH13) proteins after immunization with HA-CXCL13-circRNA or HA-circRNA (V11) vaccines. (E) Serum IgG responses were measured against stalk region of HA (SH13) proteins after immunization with HA-CXCL13-circRNA or HA-circRNA (PR8, V11, HK03, and SH13) vaccines. (F) BALB/c mice (n = 8) were immunized with HA-CXCL13-circRNA or HA-circRNA vaccines constructed from HA sequences of different influenza strains (PR8, V11, HK03, and SH13). After immunization, the HAI titers against H1N1 (PR8) in serum were measured. Data are represented as the mean ± SD. Unpaired two-tailed Student’s t test was performed for comparison, as indicated in the figures; *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

Fig. 5.

CXCL13 enhances protection against homologous and heterologous influenza virus challenge. (A) Schematic illustration of the experimental design. (B) Mice were immunized twice with HA-CXCL13-circRNA tLNPs or HA-circRNA tLNPs constructed using PR8 (H1N1). On day 49, mice were intranasally infected with PR8 (H1N1), and body weight changes (C), viral load (D), and survival (E) were recorded for 12 days. N = 10 (C, E) or 5 (D) mice per group. (F) Hematoxylin and eosin (H&E) staining and pathological scoring of the lung tissues from different groups of mice at 5 days postchallenge. (G) Immunostaining of lung tissues with an influenza virus HA-specific monoclonal antibody (mAb) and quantification of the mean fluorescence intensity (MFI). (H) Mice were immunized twice with HA-CXCL13-circRNA tLNPs or HA-circRNA tLNPs constructed using H3N2 (V11). Body weight changes (I), viral load (J), and survival (K) were recorded for 12 days after infection with PR8 (H1N1). N = 10 (I, K) or 5 (J) mice per group. (L) H&E staining and pathology scoring at 5 days postchallenge. (M) Immunostaining and MFI quantification. The data are presented as representative images from three independent visual fields randomly collected from each of the three mice per group. The pathology score for each mouse is the average score from three independent visual fields. The MFI for each mouse is the average value from three independent visual fields. Data are represented as the mean ± SD. Unpaired two-tailed Student’s t test was performed for comparison, as indicated in the figures; *P < 0.05; **P < 0.01; ***P < 0.001.

Fig. 6.

CXCL13-circRNA-adjuvanted SARS-CoV-2 RBD-circRNA immunization induces cross-reactive antibodies against multiple viral variants. (A) Schematic diagram of the circRNA, encoding the SARS-CoV-2 trimeric RBD and CXCL13. The arrow indicates the expected open-reading frame. IRES, internal ribosome entry site. (B) Mice were intramuscularly injected with PBS, RBD-CXCL13-circRNA tLNPs, or RBD-circRNA tLNPs. After 2 or 7 d, the inguinal dLNs were collected and analyzed for weight. The results are presented as the fold of the LN weight from the PBS-injected contralateral side. N = 5 mice per group. (C) The GC response in the inguinal dLNs (day 10) was analyzed by immunofluorescence. LN sections were stained for B220 (blue) and GL7 (red). Sizes of GCs in representative immunofluorescence images (Right panel). (D) Schematic illustration of the experimental design. Groups of BALB/c mice were immunized by intramuscular injection with 10 µg of RBD-CXCL13-circRNA tLNPs or RBD-circRNA tLNPs, both based on the Wuhan-Hu-1 RBD sequence, and boosted with the same dose after 2 wk. Serum samples were collected at 14 and 35 d after the prime immunization for antibody detection. (E–G) Antibodies binding to spike RBD, spike S1 and spike (Wuhan-Hu-1) (E), and RBD from relevant variants (F and G) were analyzed by ELISA. N = 8 mice per group. (H) Mice were immunized with RBD-CXCL13-circRNA tLNPs or RBD-circRNA tLNPs from different SARS-CoV-2 variants, and then the anti-wild-type SARS-CoV-2 neutralizing antibodies in serum were detected by pseudovirus neutralization assay. NT₅₀, 50% neutralization titers. N = 8 mice per group. Data are represented as the mean ± SD. Unpaired two-tailed Student’s t test was performed for comparison, as indicated in the figures; *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.