Cell-Based Systems Biology Analysis of Human AS03-Adjuvanted H5N1 Avian Influenza Vaccine Responses: A Phase I Randomized Controlled Trial

Abstract

Background: Vaccine development for influenza A/H5N1 is an important public health priority, but H5N1 vaccines are less immunogenic than seasonal influenza vaccines. Adjuvant System 03 (AS03) markedly enhances immune responses to H5N1 vaccine antigens, but the underlying molecular mechanisms are incompletely understood.

Objective and methods: We compared the safety (primary endpoint), immunogenicity (secondary), gene expression (tertiary) and cytokine responses (exploratory) between AS03-adjuvanted and unadjuvanted inactivated split-virus H5N1 influenza vaccines. In a double-blinded clinical trial, we randomized twenty adults aged 18-49 to receive two doses of either AS03-adjuvanted (n = 10) or unadjuvanted (n = 10) H5N1 vaccine 28 days apart. We used a systems biology approach to characterize and correlate changes in serum cytokines, antibody titers, and gene expression levels in six immune cell types at 1, 3, 7, and 28 days after the first vaccination.

Results: Both vaccines were well-tolerated. Nine of 10 subjects in the adjuvanted group and 0/10 in the unadjuvanted group exhibited seroprotection (hemagglutination inhibition antibody titer > 1:40) at day 56. Within 24 hours of AS03-adjuvanted vaccination, increased serum levels of IL-6 and IP-10 were noted. Interferon signaling and antigen processing and presentation-related gene responses were induced in dendritic cells, monocytes, and neutrophils. Upregulation of MHC class II antigen presentation-related genes was seen in neutrophils. Three days after AS03-adjuvanted vaccine, upregulation of genes involved in cell cycle and division was detected in NK cells and correlated with serum levels of IP-10. Early upregulation of interferon signaling-related genes was also found to predict seroprotection 56 days after first vaccination.

Conclusions: Using this cell-based systems approach, novel mechanisms of action for AS03-adjuvanted pandemic influenza vaccination were observed.

Trial registration: ClinicalTrials.gov NCT01573312.

Conflict of interest statement

KME has received grant funding for group B streptococcal vaccine research from Novartis and has participated as a member on a Novartis Data Safety Monitoring Board on influenza vaccines. CBC has received institutional grant support for staphylococcal vaccine research from Pfizer and Novartis Vaccines. JBG, TJL, and HH are employed by The Emmes Corporation. The Emmes Corporation is an independent contract research organization NIH-NIAID-DMID contracted with for collecting and analyzing data for clinical trials and studies to evaluate the safety and efficacy of vaccines, devices, and therapeutics for infectious diseases. The Emmes Corporation does not have any patents, products in development or marketed products to declare. These affiliations do not alter our adherence to PLOS ONE policies on sharing data and materials; however, we have not released raw sequencing data to maintain accordance with subject confidentiality requirements as outlined in the study’s Informed Consent Document, as approved by the Vanderbilt University Medical Center Institutional Review Board. The other authors declare that they have no conflicts of interest.

Figures

Fig 1. CONSORT diagram outlining study group enrollment and randomization.

Fig 2. AS03-adjuvanted H5N1 vaccination induced early IP-10 and IL-6 serum cytokine and later protective immune responses relative to non-adjuvanted vaccine.

Serum Antibody and cytokine responses by vaccine group (red: SV-AS03 (n = 10), blue: SV-PBS (n = 10)). (A) HAI and Nt antibody GMT and 95% CI at each time point by vaccine group; p-values are based on two-sided t-test on the log scale adjusting for unequal variances if necessary. No multiple-testing adjustment was carried out. (B) Hemagglutination inhibition (HAI) and neutralizing antibody (Nt) titers in individual subjects at days 28 and 56 by vaccine group; titer is represented by bar height (left y-axis) while fold change from baseline is shown as a connected black line (right y-axis), cut offs are indicated by grey lines (solid: 1:40 titer, dashed: 4-fold change). (C) Median fold change and 95% bootstrap CI for serum cytokines/chemokines with significantly different responses at each post-vaccination time point by vaccine group; individual subject fold changes are shown in lighter colors; p-values are based on two-sided non-parametric exact Wilcoxon rank-sum test. No multiple-testing adjustment was carried out. (D) Radar chart of cytokines/chemokine median fold changes at each time point by vaccine group.

Fig 3. AS03-adjuvanted H5N1 vaccination primarily modulated early innate immune cell gene expression responses.

Heatmaps of baseline log2 fold changes of DE genes for each cell type at each post-vaccination time point (A: monocytes, B: neutrophils, C: DC, D: NK cells, E: B cells, F: T cells). Dendograms were obtained using complete linkage clustering of uncentered pairwise Pearson correlation distances between log2 fold changes. Co-expressed gene cluster membership based on multiscale bootstrapping is highlighted along the gene dendogram on the left side. Cells are color-coded by log2 fold change (in red: upregulated from baseline; in green: downregulated from baseline.

Fig 4. Response time trends of clusters of co-expressed differential genes by vaccine group.

Co-expressed gene cluster time trends of baseline log2 fold change in LCPM of DE genes by vaccine group. Header indicates cell type, primary functional classification, and time point. Mean log2 fold change across cluster genes is drawn in bold. Individual mean gene log2 fold changes are plotted in lighter colors. Multiscale bootstrapping was used to determine co-expressed gene clusters. (A) Interferon-inducible clusters with increased responses for the AS03-SV group identified in monocytes, B cells and T cells at day 1. (B) Interferon-inducible clusters with similar gene composition but different responses for the AS03-SV group in monocytes and neutrophils at day 1. (C) Additional interferon-inducible cluster modulated by AS03-SV identified in neutrophils at day 3. (D) Cell proliferation-related clusters with increased responses for the AS03-SV group identified in NK cells at day 3. (E) MHC class-II-related cluster with increased responses for the PBS-SV group identified in NK cells at day 28.

Fig 5. AS03-adjuvanted H5N1 vaccination triggered immune system pathways including interferon signaling, cell proliferation and antigen processing and presentation.

Heatmap of MSigDB Reactome pathways enriched in DE genes. Gene sets significantly enriched in at least two conditions (time point or cell type) are shown. Heatmap cell counts represent the number of DE genes in a pathway. Cells with numbers in brackets indicate significantly enriched sets. Cells are color-coded by Jaccard similarity index (in dark-red: high level of similarity, in light-yellow: low level of similarity). The index was calculated by comparing DE genes that had any Reactome annotation with genes in a certain Reactome pathway. Pathways were clustered based on the Jaccard distance between binary enrichment patterns across conditions.

Fig 6. AS03-adjuvanted H5N1 vaccination resulted in increased cell cycle responses in NK-cells relative to non-adjuvanted vaccine 3 days post-vaccination.

The pathway map is based on the KEGG Cell cycle pathway. Pathway node color gradient encodes log2 fold change difference (LFCD) between vaccine groups (for multi-gene nodes the median LFCD is used). In red: up-regulated for the SV-AS03 group compared to the SV-PBS group, in green: down-regulated for the SV-AS03 group, in black: fold change close to 1, in dark grey: genes filtered out due to low overall expression, light grey: gene missing database mapping, white: non-human gene. DE genes are highlighted using red and green label colors.

Fig 7. Antigen processing and presentation responses for Neutrophils and NK-cells showed increased MHC-II sub-pathway activity that differed between vaccine groups.

(A) Neutrophil responses. (B) NK-cell responses. To the right: Heatmaps of subject-specific baseline log2 fold changes for KEGG Antigen processing and presentation genes across post-vaccination days by vaccine group. Colored in red: up regulated from baseline; colored in green: down regulated from baseline. To the left: KEGG pathway maps for post-vaccination days at which this pathway was significantly perturbed (day 1 for neutrophils, day 28 for NK-cells). Pathway node color gradient encodes log2 fold change difference (LFCD) between vaccine groups (for multi-gene nodes the median LFCD was used). In red: increased log2 fold change response for the SV-AS03 group relative to the SV-PBS group, in green: decreased log2 fold change response for the SV-AS03 group relative to the SV-AS03 group, and vice versa, in black: fold change close to 1, in dark grey: genes filtered out due to low overall expression, light grey: gene missing database mapping, white: non-human gene. DE genes are highlighted using red and green label colors.

Fig 8. Differential gene overlap between immune cell types.

(A) Numbers of overlapping DE genes for each cell type and post-vaccination time point combination. Cells are color-coded by number of overlapping genes (dark-red: high overlap, light-yellow: low overlap). (B) Venn diagram summarizing DE gene overlap between neutrophils, monocytes, and dendritic cells at day 1.

Fig 9. Network analysis of proteins encoded by the 80 core AS03-responsive genes shared between neutrophils, monocytes, and dendritic cells 1 day post-vaccination.

Nodes are labeled by corresponding gene name. Edges represent experimentally determined (reported in at least two publications) binary protein-protein interactions or protein complex membership. Color and size of individual nodes indicates the degree of mean log2 fold change difference (LFCD) in corresponding gene expression between vaccine groups. Further details are provided as part of the legend within the figure.

Fig 10. AS03-adjuvanted vaccine gene responses in neutrophils, monocytes and dendritic cells correlated with serum IP-10 and MIP−1α serum cytokine responses 1 day post-vaccination.

Canonical correlation plots summarizing key correlation patterns between changes in serum cytokine/antibody responses and gene responses at day 1 in (A,B) neutrophils (SV-AS03 and SV-PBS, respectively), (C) monocytes (SV-AS03), and (D) dendritic cells (SV-AS03). Red: upregulated genes; Green: downregulated genes; +/- gene name prefix: DE gene (yes/no); colored-dots: KEGG pathway class (BRITE level 2); grey lines: co-expressed gene clusters. The serum-based variable set (in black) included baseline log fold changes for 12 serum cytokines as well as baseline log fold changes of antibody titers (HAI: hemagglutination inhibition, Nt: microneutralization inhibition). The strength of the correlation is encoded by the distance from the center of the circle as well as by the angle between variables when viewed as vectors originating from the center. An acute angle (<90°) between variables represents a positive correlation, an obtuse angle (>90°) indicates a negative correlation, while a right angle is observed for zero correlation. Thus, maximum correlation is achieved when variables are closely placed together in the outer circle (or directly opposed on the outer circle.

Fig 11. AS03-adjuvanted vaccine NK-cell gene responses correlated with serum IP-10 cytokine responses 3 days post-vaccination.

Canonical correlation plots summarizing key correlation patterns between changes in serum cytokine/antibody responses and gene responses at day 3 in NK cells; (A) SV-AS03, (B) SV-PBS. See Fig 10legend for additional information.

Fig 12. Key molecular immune events observed in peripheral blood cells after AS03-adjuvanted H5N1 split-virus vaccination.

(A) Intramuscular (IM) administration of AS03-adjuvanted H5N1 vaccine results in interferon (IFN) signaling. (B) Activation of IFN-inducible genes, in particular for monocytes (Mo), dendritic cells (DC), and neutrophils (Neut). (C) Significant up-regulation of CXCL10, the gene encoding IP-10, in peripheral blood Mo and DC implies that these cells contribute to increased serum IP-10 levels (in addition to local immune cells). (D) MHC class I- related antigen presentation genes are upregulated in monocytes Mo, Neut, and DC. (E) MHC class II- related antigen presentation genes are upregulated in Neut and, to a lesser extent, in Mo. At day 3, (F) IP-10 stimulates expansion of NK cells. (G) IFN-stimulated genes remain upregulated in Neut. (H) Upregulation of MHC class I and class II molecules stimulate cytotoxic T cells (CD8+) and T helper cells (CD4+), respectively, to initiate transition to the adaptive immune response. (I) Generation of protective H5-specific antibodies after a 2-dose vaccine regimen. Solid lines indicate that direct evidence for these events was seen in our serologic or transcriptomic data, while dashed lines indicate no direct evidence was observed.