Retrovirus-Based Virus-Like Particle Immunogenicity and Its Modulation by Toll-Like Receptor Activation

Abstract

Retrovirus-derived virus-like particles (VLPs) are particularly interesting vaccine platforms, as they trigger efficient humoral and cellular immune responses and can be used to display heterologous antigens. In this study, we characterized the intrinsic immunogenicity of VLPs and investigated their possible adjuvantization by incorporation of Toll-like receptor (TLR) ligands. We designed a noncoding single-stranded RNA (ncRNA) that could be encapsidated by VLPs and induce TLR7/8 signaling. We found that VLPs efficiently induce in vitro dendritic cell activation, which can be improved by ncRNA encapsidation (_ncRNAVLPs). Transcriptome studies of dendritic cells harvested from the spleens of immunized mice identified antigen presentation and immune activation as the main gene expression signatures induced by VLPs, while TLR signaling and Th1 signatures characterize _ncRNAVLPs. In vivo and compared with standard VLPs, _ncRNAVLPs promoted Th1 responses and improved CD8⁺ T cell proliferation in a MyD88-dependent manner. In an HIV vaccine mouse model, HIV-pseudotyped _ncRNAVLPs elicited stronger antigen-specific cellular and humoral responses than VLPs. Altogether, our findings provide molecular evidence for a strong vaccine potential of retrovirus-derived VLPs that can be further improved by harnessing TLR-mediated immune activation.IMPORTANCE We previously reported that DNA vaccines encoding antigens displayed in/on retroviral VLPs are more efficient than standard DNA vaccines at inducing cellular and humoral immune responses. We aimed to decipher the mechanisms and investigated the VLPs' immunogenicity independently of DNA vaccination. We show that VLPs have the ability to activate antigen-presenting cells directly, thus confirming their intrinsic immunostimulatory properties and their potential to be used as an antigenic platform. Notably, this immunogenicity can be further improved and/or oriented by the incorporation into VLPs of ncRNA, which provides further TLR-mediated activation and Th1-type CD4⁺ and CD8⁺ T cell response orientation. Our results highlight the versatility of retrovirus-derived VLP design and the value of using ncRNA as an intrinsic vaccine adjuvant.

Keywords: RNA; TLR ligand; human immunodeficiency virus; immunogenicity; mice; retroviruses; vaccines; virus-like particle.

FIG 1

ncRNA system and validation of _ncRNAVLPs. (A) Structure of the ncRNA-encoding plasmid. The box sizes indicate the relative lengths of the genetic sequences according to the scale provided. PBS, primer-binding site; CMV, cytomegalovirus immediate-early promoter; psi, retroviral encapsidation sequence; U3, R, and U5, MLV LTR sequences. (B) Schematic representation of MLV-derived standard VLPs (left) and _ncRNAVLPs (right). Packaged cellular RNAs and ncRNA (two copies of single-stranded virus-derived RNA) are illustrated. (C) Validation by RT-qPCR of the presence of ncRNA in pseudoparticles. Gray curves, VLPs; black curves, _ncRNAVLPs. Untransfected cell lysate was used as a negative control (dashed curves). Duplicates of one experiment representative of three are shown. (D) ncRNA-specific qPCR conducted on VLPs or _ncRNAVLPs with (RT+) or without (RT−) a reverse transcription step. (E) Evaluation of endotoxin levels in three different production batches of VLPs and _ncRNAVLPs using a LAL assay. A commercial OVA protein batch was used as a control.

FIG 2

In vitro effect of VLPs carrying or not carrying ncRNA on bone marrow-derived dendritic cell activation. Immature BMDCs from C57BL/6 (A to F) or MyD88^−/− (G to L) mice were incubated for 24 h in the presence of 1, 5, or 10 μg/ml VSV-G-pseudotyped MLV-Gag VLPs or _ncRNAVLPs. CD80 and CD86 expression levels were analyzed by flow cytometry. (A, B, G, and H) Representative histograms of CD80 (A and G) and CD86 (B and H) on C57BL/6 BMDCs (A and B) or MyD88^−/− BMDCs (G and H) cultured with 5 μg/ml VLPs (thin lines), _ncRNAVLPs (thick lines), or medium alone (shaded histogram) are shown. (C to F and I to L) Related percentages of CD80⁺ (C and I) and CD86⁺ (D and J). Medium alone, LPS (100 ng/ml), and R848 (1 μg/ml) were used as a negative control and two positive controls, respectively. The results represent the means + standard deviations (SD) of duplicates for each dose of VLPs from one experiment representative of two (C and D, I and J) and the means ± standard errors of the mean (SEM) of two independent experiments with a dose of 5 μg/ml are represented for CD80 (E and K) and CD86 (F and L) for C57BL/6 (E and F) and MyD88^−/− (K and L) BMDCs. *, P ≤ 0.05; ns, not significant; Mann-Whitney test.

FIG 3

Transcriptome analysis of splenic dendritic cells after in vivo injection of VLPs or _ncRNAVLPs. Dendritic cell-specific gene set enrichment analysis using Gene Ontology Database signatures allowed us to identify molecular signatures that were differentially enriched in the VLP and _ncRNAVLP groups compared with PBS. (A) Results were mapped using Cytoscape software as a network of signatures (nodes) related by similarity (edges). The node size is proportional to the total number of genes in each set. Groups of functionally related signatures are circled and labeled (modules). The gray nodes represent signatures shared between VLP and _ncRNAVLP groups; the red nodes represent signatures that are specific for the _ncRNAVLP group. FDR q value = 0.05; P value = 0.005. (B) Heat maps showing three ncRNA-specific signatures. Samples were clustered using distance-based hierarchical clustering regarding gene expression. The heat map colors represent the gene expression (red for high, black for middle, and green for low expression). The Gene Ontology exact names of signatures are as follows: DC differentiation, dendritic cell differentiation; TLR signaling, positive regulation of Toll-like receptor signaling pathways; and Th1 response, positive regulation of T helper type 1 immune response. (C) Validation by RT-qPCR of the relative quantities of eight different genes from the three selected signatures.

FIG 4

In vitro and in vivo effects of VLPs carrying or not carrying ncRNA on T cell proliferation and polarization. (A and B) In vitro proliferation of antigen-specific CD8⁺ T cells. CFSE-stained OVA-specific OT-I splenic lymphocytes were cultured for 3 days with antigen-presenting cells from C57BL/6 (A) or MyD88^−/− (B) mice in the presence of 5 μg/ml of VSV-G-pseudotyped Gag-OVA VLPs (gray bars) or _ncRNAVLPs (black bars). The percentages of divided cells were evaluated by flow cytometry analysis of CFSE-low cells among CD8⁺ Vα2⁺ live cells. Medium alone and OVA-I peptide were used as negative and positive controls, respectively. Means of triplicates from three independent experiments are shown. (C to E) In vivo proliferation of antigen-specific CD8⁺ T cells. C57BL/6 or MyD88^−/− mice (n = 5 per group) were injected i.v. with 1 μg of VSV-G-pseudotyped Gag-OVA VLPs or _ncRNAVLPs or PBS in the control group. Six hours later, the mice received 1.5 × 10⁶ CFSE⁺ OVA-specific CD8⁺ T cells from OT-I mice. After 3 days, spleens were collected, and the proliferation of OVA-specific CD8⁺ T cells was evaluated by flow cytometry. (C) One representative dot plot of the CFSE profile from each group is depicted. The percentages of divided OVA-specific CD8⁺ T cells for each dose in C57BL/6 (D) and MyD88^−/− (E) mice are shown. The results represent the mean values + SEM. (F and G) In vivo proliferation and differentiation of antigen-specific CD4⁺ T cells. The same experiment as for panel D was conducted with OT-II cells instead of OT-I cells. (F) Means of the percentages of divided OVA-specific CD4⁺ T cells and SEM. (G) Geometric means (mean fluorescence intensity [MFI]) of T-bet/GATA3 expression ratios among divided CD4⁺ T cells. *, P ≤ 0.05; **, P ≤ 0.01; ns, not significant; Mann-Whitney test.

FIG 5

T and B cell immune responses in mice vaccinated with HIV-pseudotyped VLPs carrying or not carrying ncRNA. (A) Schematic representation of the vaccination protocol. C57BL/6 or MyD88^−/− mice (n = 5 per group) were immunized 3 times at 2-week intervals with 25 μg of HIV-pseudotyped Gag–gp33-41 VLPs or _ncRNAVLPs. Sera were collected from blood samples at weeks 6 and 12. (B and C) Cellular responses were evaluated at week 12 by standard IFN-γ ELISPOT assay after specific restimulation with either gp33-41 (B) or HIV-gp140 (C). The results represent individual values and group means. (D and E) Specific anti-gp140 antibody concentrations were evaluated in sera of immunized or naive C57BL/6 mice by anti-GP120 ELISA at weeks 6 (D) and 12 (E). The results represent the mean values + SEM of measured concentrations. *, P ≤ 0.05; **, P ≤ 0.01; ns, not significant; Mann-Whitney test.