AS04 drives superior cross-protective antibody response by increased NOTCH signaling of dendritic cells and proliferation of memory B cells

Abstract

Introduction: The Gardasil-4^® vaccine targets HPV types 6, 11, 16 and 18 and is formulated with amorphous alum. Cervarix^® targets HPV types 16 and 18 using AS04 (Al(OH)3 + TLR4 agonist MPL) to enhance immune response. Cervarix elicits higher cross-protection against other high-risk HPV types, likely mediated by AS04.

Methods: To investigate mechanisms of cross-neutralizing potential, six monozygotic twins (12 females aged 9-13 years) were vaccinated with either Cervarix or Gardasil-4 (2 doses, 6 months apart). Serum neutralizing antibody titers against HPV 6,16,18,31,33,45,52, and 58 were assessed pre-vaccination and 7 days post-second dose. Multi-omic single cell RNA and ATAC sequencing of PBMCs was performed at the latter timepoint.

Results: Cervarix generated higher cross-neutralizing antibody titers than Gardasil-4. Higher frequencies of dendritic cells and memory B cells were observed. Gene Set Enrichment Analysis (GSEA) indicated enhanced pathways related to NOTCH2 signaling in DCs and cell cycling/RNA translation in B cells, correlating positively with cross-neutralizing antibody titers. Increased chromatin accessability in genes related to NOTCH signaling in cDC1 was also observed. Cervarix-vaccinated subjects showed increased DC-to-memory B signaling, through upregulation of NOTCH ligands. Engagement of NOTCH was associated to BCL2 expression in memory B cells, supporting an anti-apoptotic state.

Conclusion: Increased DC signaling, including NOTCH, through AS04 in Cervarix supports cell survival and sustained RNA translation in memory B cells, 7 days post-vaccination. This may enhance adaptive immune cell maturation, providing a mechanism that can lead to improved cross-reactivity.

Keywords: HPV vaccine; adjuvant; dendritic cells; immune response; memory B cell.

Figure 1

Cervarix induces higher neutralizing antibody titers against closely related HPV types and enhances DC and Memory cell frequencies in peripheral blood. (A) Study design: Six female homozygotic twins (n = 12) were vaccinated with either Cervarix, containing AS04, or Gardasil-4, containing alum. All participants were HPV-naive and received the vaccine at the recommended age of 9-13 years in Belgium. Each regimen consisted of two doses administered six months apart (day 0 and day 180). Serum samples were collected on day 0 (pre-vaccination) and seven days after the second dose (day 187), while blood for PBMC isolation was collected on day 187 only. (B) Neutralizing antibody titers: neutralizing antibody titers against HPV6, 16, 18, 31, 33, 45, 52 and 58 were measured using a pseudovirion-based neutralizing assay in all participants at both timepoints. Geometric mean titers (GMT) were calculated for each HPV type and vaccine. The line graph presents log-transformed GMT on day 0 and day 187 for each vaccine separately, with colored line representing specific HPV types. (C) Neutralizing antibody levels on day 187: Boxplots display the median (IQR) log-transformed neutralizing antibody titers on day 187 for each HPV type, grouped by vaccine (orange: Cervarix, green: Gardasil-4). (D) Single cell RNA sequencing analysis: A total of 79,817 cells were analyzed using single-cell RNA sequencing. The UMAP plot visualizes all single cells after dimensionality reduction, clustering, and annotation. (E) A subset of all DCs and B cells were selected for further investigation. The UMAP plot visualizes the subset of single cells after dimensionality reduction and re-clustering. (F) Cell frequencies: Boxplots show the median (IQR) of relative cell frequencies per vaccine. (G) Cytokine concentrations: Boxplots show median (IQR) for each measured cytokine. Mono, monocytes; NK, Natural Killer cells; cDC1/cDC2, conventional dendritic cells type 1 or type 2; pDC, plasmacytoid dendritic cell; CD4, CD4+ T cells; CD8, CD8+ T cells; GMT, geometric mean titer; HPV, human papilloma virus; significance was assessed by the actual p-values by Wilcoxon rank test.

Figure 2

Cervarix induces transcriptional responses in DCs and memory B cells that correlate with antibody titers. (A) Clustered heatmap of all genes involved in identified enriched pathways in cDC1, pDC, and B memory cells. Color represents column-wise z-scores. (B) Heatmap of normalized enriched scores (NES) for significantly enriched reactome pathways in cDC1, pDC, and memory B cells. All pathways that were enriched in any of DC subsets are shown. Additional pathways enriched in B cell subsets are included in Supplementary Table 3 . (C) Boxplot showing the median (IQR) of average gene expression of NOTCH2 signaling pathway genes in cDC1, grouped by vaccine. (D) Scatter plots showing correlation between the module score of NOTCH2 signaling pathway per participant with neutralizing antibody titers against all measured HPV types. cDC1, conventional dendritic cells type 1; cDC2, conventional dendritic cells type 2; pDC, plasmacytoid dendritic cell; NES, normalized enrichment score.

Figure 3

Cervarix enhances chromatin accessibility of NOTCH genes in cDC1. (A) Heatmap showing normalized chromatin accessibility at the top 200 DARs in cDC1 for each subject. Regions were classified as follows: promoter −2,000 bp to +500 bp; distal −10 kbp to +10 kbp – promoter; trans< −10 kbp or > +10 kbp. (B) Heatmap of normalized accessibility of NOTCH-related DARs in cDC1 for each subject. (C) The top enriched motifs identified from NOTCH-related DARs in cDC1. (D) Chromatin accessibility at the locus of NOTCH ligands (NOTCH1, NOTCH2, NOTCH3, NOTCH4, DLL4, JAG1) in cDC1 grouped per vaccine. The coverage tracks represent the aggregate signal of transposase-accessible regions, with peaks indicating open chromatin regions. The plot extends 50 bp upstream and 10 kb downstream of the gene to capture potential regulatory elements. cDC1, conventional dendritic cells type 1; cDC2, conventional dendritic cells type 2; pDC, plasmacytoid dendritic cell.

Figure 4

Enhanced signaling by DCs via NOTCH after Cervarix vaccination. (A) Bar plot showing the total strength of interactions per vaccine. (B) Heatmap comparing the differential strength of interactions between Cervarix and Gardasil-4. Rows represent sending cells (y-axis), while columns indicate receiving cells (x-axis). Red denotes stronger signaling in Cervarix, while blue indicates stronger signaling in Gardasil-4. The heatmap is clustered based on the differential strength of interactions. Bar plots along the rows and columns indicate the total strength of signals sent or received, respectively.C-D. Heatmaps showing outgoing signals for each cell type in Cervarix (C) and Gardasil-4 (D). “Outgoing” refers to signals sent by the specified cell types. Signals are ranked by ‘importance’ and the extent to which cells utilize them. Bar plots on the columns represent the total number of signals sent by each cell type, while the bar plots on the rows indicate the total number of signals sent overall. Green highlights the relative strength of the outgoing signal. E-F NicheNet analysis of DC signaling to memory B cells. (E) Ligand activity plot highlighting the most important ligands sent by cDC1 and pDC to memory B cells. (F) Target gene heatmap showing the influence on expression of BCL2, JUNB and CD40 in memory B cells (columns) by the top ligands sent by DCs (rows). cDC1/cDC2, conventional dendritic cells type 1 or type 2; pDC, plasmacytoid dendritic cell; AUPR, area under the precision recall curve.

Figure 5

T help via NOTCH signaling to B memory cells. (A) UMAP projecting displaying all CD4+ T cells after dimensionality reduction and clustering and manual annotation of T helper subsets. (B) Boxplots representing the median (IQR) relative frequencies of CD4+ T cell subsets, stratified by vaccine (orange: Cervarix, green: Gardasil-4). (C) Dotplot showing significantly enriched pathways related to NOTCH signaling in Th1 and Th2 cells. Positive NES indicates a pathway enriched in Cervarix. (D) Heatmap showing the differential strength of interactions between cell types. Rows represent sender cells (y-axis), and columns represent receiver cells (x-axis). Blue indicates stronger signaling in Cervarix group, and red indicates stronger signaling in Gardasil-4 group. The heatmap is clustered by the differential interaction strength, with bar plots on the rows and columns indicating the weight of signals sent or received, respectively. (E) NicheNet analysis of Th2 signaling to memory B cells. Target gene heatmap showing the expression of genes in memory B cells (columns) influenced by several cytokines sent by Th2 cells (rows). (F) Boxplots showing the median (IQR) of average expression of Immunoglobulin-related genes in B memory cells, stratified by vaccine. Tfh, follicular T helper cell; Th, helper T cell; Treg, regulatory T cell.

Figure 6

NOTCH signaling and cell cycling and survival gene signatures are associated with antibody breadth in other vaccines. (A) The Gene Expression Omnibus (GEO) database was searched for publicly available gene expression data that included antibody titers across multiple strains or pathogen types. Two relevant studies were identified. The first examined a conjugated polysaccharide vaccine against meningococcus, with serum bactericidal antibody titers available for MenA and MenC. The second study analyzed gene signatures of an mRNA COVID-19 vaccine and their relationship with neutralization antibody titers against two SARS-CoV-2 strains. For both studies, titers across strains were summed, and subjects were categorized into high and low breadth groups based on the median value. (B-C). Bar charts displaying the top 20 enriched pathways in high- vs. low-breadth subjects for GSE52245 (B) and GSE169159 (C). (D) Heatmap showing the NES of pathways commonly enriched in both vaccines, clustered per vaccine. NES, normalized enrichment score; GEO, gene expression omnibus; nAB, neutralizing antibodies; SBA, serum bactericidal assay.