A molecular atlas of innate immunity to adjuvanted and live attenuated vaccines, in mice

Abstract

Adjuvants hold great potential in enhancing vaccine efficacy, making the understanding and improving of adjuvants critical goals in vaccinology. The TLR7/8 agonist, 3M-052, induces long-lived humoral immunity in non-human primates and is currently being evaluated in human clinical trials. However, the innate mechanisms of 3M-052 have not been fully characterized. Here, we perform flow cytometry, single cell RNA-seq and ATAC-seq to profile the kinetics, transcriptomics and epigenomics of innate immune cells in murine draining lymph nodes following 3M-052-Alum/Ovalbumin immunization. We find that 3M-052-Alum/OVA induces a robust antiviral and interferon gene program, similar to the yellow fever vaccine, which is known to confer long-lasting protection. Activation of myeloid cells in dLNs persists through day 28 and single cell analysis reveals putative TF-gene regulatory programs in distinct myeloid cells and heterogeneity of monocytes. This study provides a comprehensive characterization of the transcriptomics and epigenomics of innate populations in the dLNs after vaccination.

Fig. 1. 3M-052-Alum/OVA induces changes in activation and cell frequency of innate immune cells in the draining LN on Day 1.

a Schematic of flow cytometry phenotyping, CITE-seq and scATAC-seq in this study. b Mean $\pm$SEM of CD86 median fluorescence intensity (MFI) in innate immune cell subsets from the dLNs at each timepoint (n = 5 for each group; naïve group pooled from three independent experiments, n = 15). Two-way ANOVA with Dunnett’s multiple comparisons test in comparison to Day 0 (*, adjusted p-value < 0.05; **, padj < 0.01; ****, padj < 0.0001). c Fold-change of mean % live cells $\pm$SEM across innate immune cell subsets in the dLNs at day 1 post-immunization compared to naïve control (n = 5 for each group; naïve group pooled from three independent experiments, n = 15). Two-way ANOVA with Dunnett’s multiple comparisons test in comparison to Day 0 (*, adjusted p-value < 0.05; **, padj < 0.01; ****, padj < 0.0001). Source data are provided as a Source Data file.

Fig. 2. 3M-052-Alum/OVA induces a global anti-viral innate immunity program in draining LN 1 day post-immunization.

a UMAP embedding of 21,664 cells classified by cell type. b Scaled expression of cell-type-specific genes used to classify cells by type. c Number of upregulated (red) and downregulated (blue) genes in each cell type, FDR < 0.05 and absolute LogFC > 0.25. d Log fold change of genes from (c) across all cell types. Genes and cell types were k-means and hierarchically clustered. Functional groups of relevant gene clusters are highlighted. e BTMs significantly enriched (FDR < 0.001) after 1- day post-immunization with 3M-052-Alum/OVA. Overrepresentation analysis of the genes from (c) was used to determine significance. Adjusted P-values were measured using hypergeometric distribution with Benjamini Hochberg correction.

Fig. 3. 3M-052-Alum/OVA induces a transcriptional signature on Day 1 similar to YF-17D.

a UMAP embedding of 15,330 cells classified by cell type. b Number of upregulated (yellow) and downregulated (blue) genes in each celltype, FDR < 0.05 and absolute LogFC > 0.25. c Correlation of the logFC of 3M-052-Alum/OVA and YF across monocytes. R = Pearson’s correlation; p represents two-tailed P values. d Absolute logFC of each gene in monocytes in 3M (red) and YF (yellow) (n = 20967). Statistical test performed was two-sided t-test. Center line of boxplot corresponds to the median, the bounds are the 75th and 25th percentiles (interquartile range; IQR), and the whiskers are the largest or smallest value no further than 1.5*IQR from the bounds. e Log fold change of genes from (b) across all cell types. f Correlation between the logFC of each gene in human samples 7 days after immunization with the logFC of the corresponding homolog in mouse myeloid cells 1 day after YF-17D immunization. R = Pearson’s correlation; p represents two-tailed P values. g Genes previously identified as significantly changing in humans after YF-17D immunization. Left, human transcriptional changes after YF-17D vaccination; Right, corresponding changes in mouse monocytes after immunization.

Fig. 4. scATAC-seq reveals differential TF motif accessibility between 3M-052-Alum/OVA and YF-17D.

a UMAP embedding of 44327 cells from innate immune cells immunized with 3M-052-Alum/OVA or YF-17D. b Differentially accessible genes, FDR < 0.05 and absolute logFC > 0.1. c Correlation between logFC in monocytes in scATAC and scRNA-seq. R is Pearson’s correlation. d Volcano plots of gene accessibility logFC in monocytes in 3M-052-Alum/OVA and YF-17D. e Representative tracks of an Irf inducible site on Chromosome 16, spanning genes Mx1 and Mx2. Peaks are shown in red below the track. f TF logFC 1 day after 3M-052-Alum/OVA immunization. g Top, differentially accessible TF motifs (FDR < 0.05) 1 day after immunization. Bottom, the percentage of 3M-052-Alum/OVA DA TF motifs that change in the same direction after YF-17D immunization. h The accessibility correlation network of genes and transcription factors in day 1 monocytes. Each node is a differentially accessible gene or TF; edges represent the correlation between nodes.

Fig. 5. 3M-052-Alum/OVA induces long-lasting changes in innate immune cells that persisted up to Day 28.

a Number of significantly expressed (RNA) genes that remain up or downregulated compared to baseline (FDR < 0.05 absolute logFC > 0.1) and the total number of significantly accessible (ATAC) motifs at day 28. b BTMs significantly enriched (FDR < 0.001) after 1- and 28-days post-immunization with 3M-052-Alum/OVA across dLN innate immune cells. Overrepresentation analysis of the differentially expressed genes was used to determine significance. c Temporal expression pattern of all genes within modules M111.1 and M111.0 (viral sensing & immunity; IRF2 targets network) in both RNA (left panels) and ATAC (right panels, red line, and square represent the mean fold change). d Heatmap of TF-ISG correlation in each cell subset. Pearson correlation is shown here (left). Rank plots of TF motifs that correlate with Ifit3 expression in each cell type (right). e Top differentially accessibile TF motifs at day 28 in YF-17D compared to 3M-052-Alum/OVA (FDR < 0.05 absolute logFC > 0.1).

Fig. 6. Differentially expressed and accessible genes in 3M-052-Alum/OVA are driven by different sub-clusters of monocytes on Day 28.

a Re-embedding of monocytes present before (day 0) and 1 and 28 days after immunization with 3M-052-Alum/OVA in scRNA-seq. b Sub-clustering and re-embedding of monocytes at baseline and 28 days after 3M-052-Alum/OVA immunization in scRNA-seq (top) and scATAC-seq (bottom). c Gene expression and accessibility score of antiviral BTM module highlighted in Fig. 5b, across day 0 and day 28 monocyte subclusters. d Re-clustering of Ly6Chi monocyte subcluster identified in (c) (left); heatmap of Irf and AP-1 gene accessibility within Ly6Chi monocyte subclusters (right), performed on scATAC-seq data. e Re-clustering of Ly6Chi monocyte subcluster identified in (c) (left); heatmap of Irf and AP-1 gene accessibility within Ly6Chi monocyte subclusters (right), performed on scRNA-seq data.