Central Role of CD169+ Lymph Node Resident Macrophages in the Adjuvanticity of the QS-21 Component of AS01

Abstract

Saponins represent a promising class of vaccine adjuvant. Together with the TLR4-ligand MPL, QS-21 is part of the Adjuvant System AS01, a key component of the malaria and zoster candidate vaccines that display demonstrated clinical efficacy. However, the mechanism of action of QS-21 in this liposomal formulation is poorly understood. Upon intra-muscular immunisation, we observed that QS-21 rapidly accumulated in CD169⁺ resident macrophages of the draining lymph node where it elicited a local innate immune response. Depletion of these cells abrogated QS-21-mediated innate cell recruitment to the lymph node, dendritic cell (DC) phenotypic maturation as well as the adjuvant effect on T-cell and antibody responses to co-administered antigens. DCs rather than lymph node-resident macrophages were directly involved in T-cell priming by QS-21, as revealed by the decrease in antigen-specific T-cell response in Batf3^-/- mice. Further analysis showed that the adjuvant effect of QS-21 depended on the integration of Caspase-1 and MyD88 pathways, at least in part through the local release of HMGB1. Taken together, this work unravels the key role of lymph node sentinel macrophage in controlling the adjuvant effect of a molecule proven to improve vaccine response in humans.

Figure 1. QS-21 is a potent adjuvant for the induction of cellular and humoral responses.

(A) Experiment timeline. (B) Mice were immunised as in (A) and cytokine production and polyfunctionality of HBs-specific CD4 and CD8 T cells were assessed by flow cytometric cytokine intracellular staining at days 7 and 21. Data is represented as median of 6 (Ag alone groups) or 20 (Ag + QS-21 groups) animals. (C) Percentage of OVA-specific CD8 T cells quantified by H2Kb-SIINFEKL pentamer staining at days 7 and 21. (D) Expression of CD127 and KLRG1 on pentamer-positive CD8 T cells measured by flow cytometry. (F) Anti-HBs IgG1 and IgG2c titres measured by ELISA in the serum of mice immunised with antigens alone (Ag) or antigens formulated with QS-21 on days 7 and 21. Each point represents a single mouse and the horizontal bar represents the geometric mean. Statistical significance was determined by a non-parametric Mann Mann-Whitney test. The data is representative of at least 2 independent experiments.

Figure 2. QS-21 induces rapid changes in gene expression in draining lymph node.

(A,B) Monocyte (CD11b⁺ Ly6C⁺ Ly6G⁻), neutrophil (SSC^hi CD11b⁺ Ly6C^int Ly6G⁺), eosinophil (SSC^hi CD11b⁺ Ly6C⁺ Ly6G⁻ SiglecF⁺) and dendritic cell (CD11c⁺ MHCII⁺) recruitment was assessed by flow cytometry 24 h post immunisation in the muscle (A) and draining lymph node (B). Each point represents a pool of 2 (DLN) or 3 (muscle) mice and the horizontal bar is the geometric mean. (C) Heatmap representation of mRNA expression (Fold change over PBS) of the top upregulated genes (Fold Change over PBS >10 in either the injection site or draining lymph node) measured by microarray analysis. (D,E) Fold change over PBS of the top 50 upregulated genes at 2, 4 or 6 hours in the draining lymph node (D) or injection site (E). The top upregulated genes in the draining lymph node (Cxcl1, il6, il1b) are represented on each chart. (F) InnateDB analysis of pathway over-representation of significantly upregulated genes (p-value < 0.05 and fold change >3) using the hypergeometric algorithm and Benjamini Hochberg correction for p-values.

Figure 3. QS-21-bodipy rapidly accumulates in the draining lymph node where it colocalises with CD11b⁺ CD169⁺ macrophages of the subcapsular sinus.

(A) Mice were injected i.m. with liposomes containing bodipy-labelled QS-21. The draining lymph nodes were recovered 30 min and 3 h post injection and stained with anti-CD11b, anti-CD169 and anti-B220 antibodies and analysed by confocal microscopy. (B) Flow cytometry analysis of subcapsular sinus (CD169⁺ F4/80⁻) and medullary sinus (CD169⁺ F4/80⁺) macrophages in mice 24 h following PBS or QS-21 injection. (C) Quantification of the loss of LN-resident macrophages detected by flow cytometry. Results from two independent experiments. Each point represents one mouse. Statistical significance was determined by a non-parametric Mann-Whitney test.

Figure 4. Clodronate-mediated depletion of CD169⁺ subcapsular sinus macrophages greatly reduces innate and effector responses induced by QS-21.

(A) Monocyte, neutrophil and dendritic cell recruitment to the DLN measured by flow cytometry 24 h post QS-21 injection (Ag: n = 5, QS-21: n = 10). (B) Expression of the co-stimulatory molecules CD80 and CD86 on dendritic cells measured by flow cytometry 24 h post QS-21 injection (Ag: n = 4, QS-21: n = 5) (C) Mice were injected with clodronate or control liposomes 6 days prior to vaccination and immunised as in 1 A. At day 21, median cytokine production (C) and polyfunctionality (D) of HBs-specific CD4 T cells, cytokine production by HBs-specific CD8 T cells (E), and frequency of OVA-specific circulating CD8 T-cells (E) were assessed by flow cytometry (Ag: n = 4, QS-21: n = 8). (F) Anti-HBs IgG1, IgG2c and total IgG titres in the serum of mice injected with QS-21 or Alum were measured by ELISA at day 21 (Ag: n = 8, QS-21: n = 16, Alum: n = 8). Each point represents a single mouse and the horizontal bar represents the geometric mean. Statistical significance was determined by a non-parametric Mann-Whitney test. The data is representative of at least 2 independent experiments.

Figure 5. Batf3-dependent dendritic cells regulate QS-21-mediated T-cell responses.

Mice were immunised as in 1 A and adaptive responses were measured at day 21. (A) Flow cytometry analysis of cytokine production by HBs-specific CD4 T cells in WT and Batf3^−/− mice. (B) polyfunctionality index of CD4 T cells. (C) Flow cytometry analysis of cytokine production by splenic HBs-specific CD8 T cells and of the frequency of circulating OVA-specific CD8 T cells (Ag: n = 1–2, QS-21: n = 8). (D) Anti-HBs IgG1 and IgG2c titres in the serum measured by ELISA (Ag: n = 1–2, QS-21: n = 8). Each point represents one mouse and the horizontal bar represents the geometric mean. Statistical significance was determined by a non-parametric Mann-Whitney test.

Figure 6. QS-21 elicited innate cell recruitment and CD4/CD8 T-cell responses are dependent on Caspase-1.

(A) Bone-marrow derived dendritic cells (BMDCs) from mice lacking NALP3 or Caspase-1 expression were primed overnight with MPL and stimulated with either QS-21 or alum. IL-1β release in the supernatant was determined by ELISA. The data is represented as mean and SEM of 3 independent experiments. (B) WT or caspase-1 KO mice were injected i.m. with either PBS or QS-21 and the draining lymph nodes were recovered at the indicated time points. Whole lymph nodes were lysed and caspase-1 and IL-1β processing were detected by western blot. (C) Innate cell recruitment to the iliac DLN in WT and Caspase-1 KO mice 24 h post QS-21 injection. Absolute numbers of monocytes (Lin-CD11b⁺ Ly6C⁺ Ly6G⁻), neutrophils (SSC^hi Lin⁻ CD11b⁺ Ly6C^int Ly6G⁺) and dendritic cells (Lin⁻ CD11c⁺ MHCII⁺) per DLN were assessed by flow cytometry (PBS: n = 5, QS-21: n = 5–10). (D–F) Mice were immunised as in 1 A. At day 21, cytokine production (D) and polyfunctionality (E) of HBs-specific CD4 were evaluated by flow cytometry. (F) Cytokine production of HBs-specific and frequency of OVA-specifc CD8 T cells were evaluated by flow cytometry (Ag: n = 3, QS-21: n = 9–10). (G) Anti-HBs IgG1 and IgG2c titres in the serum at day 21 were measured by ELISA (Ag: n = 5–6, QS-21: n = 20). Each point represents one mouse and the horizontal bar represents the geometric mean. Statistical significance was determined by a non-parametric Mann-Whitney test. The data is representative of 2 independent experiments.

Figure 7. QS-21-elicited neutrophil and dendritic cell recruitment, cellular CD4/CD8 T-cell responses and antibody production are dependent on MyD88.

(A) Innate cell recruitment in the iliac DLN 24 h post QS-21 injection. Absolute numbers of monocytes (Lin⁻ CD11b⁺ Ly6C⁺ Ly6G⁻), neutrophils (SSC^hi Lin⁻ CD11b⁺ Ly6C^int Ly6G⁺) and dendritic cells (Lin⁻ CD11c⁺ MHCII⁺) per DLN were assessed by flow cytometry (PBS: n = 6, QS-21: n = 7–8). (B–D) WT and MyD88 KO mice were immunised as in 1 A. At day 21, median cytokine production (B) and polyfunctionality (C) of HBs-specific CD4 were evaluated by intracellular staining (Ag: n = 3, QS-21: n = 14). (D) Cytokine production of HBs-specific splenic CD8 T cells and frequency of OVA-specific circulating CD8 T cells assessed by flow cytometry (Ag: n = 3–6, QS-21: n = 14–21). (E) Anti-HBs IgG1 and IgG2c titres in the serum at day 21 were measured by ELISA (Ag: n = 6, QS-21: n = 21). Each point represents one mouse and the horizontal bar represents the geometric mean. Statistical significance was determined by a non-parametric Mann-Whitney test. The data is representative of 2 independent experiments.

Figure 8. QS-21 injection leads to HMGB1 release that is required for optimal CD4 T-cell responses.

(A) Mice were injected i.m. with QS-21 and the draining lymph nodes were recovered 6 h post injection. The whole lymph nodes were cultured for 24 h in complete medium and HMGB1 was detected in the supernatant by ELISA (n = 10). (B) Experiment timeline. Mice received four i.p. peptide (500 μg per mouse – red arrows) injections at day 0 and day 14 (1 h before immunisation and then at 12 h intervals). The FSSE-NH₂ peptide inhibits HMGB1-MD-2 interaction while the scrambled control (SFSE-NH₂) does not. (C–F) FSSE and control peptide-treated mice were immunised as in (B). At day 21, cytokine production (C) and polyfunctionality (D) of HBs-specific CD4 T cells were evaluated after in vitro restimulation by intracellular staining. The data is represented as the median of 6 (Ag) or 20 (QS-21) mice. (E) Cytokine production by HBs-specific splenic CD8 and frequency of antigen (OVA)-specific circulating CD8 T cells assessed by flow cytometry (Ag: n = 6, QS-21: n = 20). (F) Anti-HBs IgG1 and IgG2c titres in the serum at day 21 were measured by ELISA (Ag: n = 3, QS-21: n = 10). Each point represents a single mouse and the horizontal bar represents the geometric mean. Statistical significance was determined by a non-parametric Mann-Whitney test. The data represents a pool of 2 independent experiments.