Prior upregulation of interferon pathways in the nasopharynx impacts viral shedding following live attenuated influenza vaccine challenge in children

Abstract

In children lacking influenza-specific adaptive immunity, upper respiratory tract innate immune responses may influence viral replication and disease outcome. We use trivalent live attenuated influenza vaccine (LAIV) as a surrogate challenge model in children aged 24-59 months to identify pre-infection mucosal transcriptomic signatures associated with subsequent viral shedding. Upregulation of interferon signaling pathways prior to LAIV is significantly associated with lower strain-specific viral loads (VLs) at days 2 and 7. Several interferon-stimulated genes are differentially expressed in children with pre-LAIV asymptomatic respiratory viral infections and negatively correlated with LAIV VLs. Upregulation of genes enriched in macrophages, neutrophils, and eosinophils is associated with lower VLs and found more commonly in children with asymptomatic viral infections. Variability in pre-infection mucosal interferon gene expression in children may impact the course of subsequent influenza infections. This variability may be due to frequent respiratory viral infections, demonstrating the potential importance of mucosal virus-virus interactions in children.

Keywords: LAIV; asymptomatic respiratory viral infection; influenza; interferon-stimulated genes; mucosal; transcriptome.

Graphical abstract

Figure 1

Study design, nasopharyngeal LAIV viral loads, and presence of asymptomatic respiratory viruses (A) Study design and sampling. Influenza vaccine naive children (n = 82) given a single dose of the northern hemisphere 2017-18 LAIV (Nasovac-S Serum Institute of India) were included in the study. NPS, nasopharyngeal swab; HAI, hemagglutinin inhibition titer. (B) Heatmap showing nasopharyngeal viral shedding at days 2 and 7 following LAIV (log₁₀ 50% egg infectious dose equivalent [EID₅₀]/mL) for 2009 pandemic H1N1 (pH1N1), H3N2, and influenza B viruses in children seronegative (HAI titer < 1:10) and seropositive for each corresponding strain. Detection of asymptomatic respiratory viruses at baseline are displayed, with key for each virus as per Figure 1D (X indicates samples not available for testing). HPIV-1, human parainfluenza 1; seasonal CoVs, seasonal coronaviruses (229E, OC43, NL63). X denotes children where no result was available due to lack of sample availability. (C) Comparison of day 2 and day 7 nasopharyngeal viral loads in children who are seronegative and seropositive to each influenza strain. Red lines denote median value. The p values are from Mann-Whitney U test. (D) Prevalence of asymptomatic respiratory virus in 33/79 (41.8%) children in nasopharyngeal swabs taken prior to vaccination.

Figure 2

Baseline transcriptional profiles associated with LAIV strain shedding at day 2 and day 7 in children seronegative to each influenza strain prior to vaccination (A) Selected pathways from gene set enrichment analysis (GSEA) using the Spearman correlation coefficients between normalized gene expression (rlog) at baseline and strain-specific viral loads (log₁₀ EID₅₀/mL) at day 2 and day 7 as rank, and Reactome pathways set. NES, normalized enrichment score. The size of circles is proportional to the adjusted −log₁₀ p value from GSEA, while the intensity of the circle color denotes the NES for each pathway/strain-specific viral load combination (deeper color indicates higher NES). Blue circles denote enrichment of genes in nasopharynx prior to LAIV challenge that negatively correlate with viral load at day 2 and day 7, and red circles denote enrichment of genes that positively correlate with viral loads. Significant pathways (adjusted p value < 0.1) are highlighted with a black outline. The number of leading edge (LE) genes is provided for each pathway, i.e., genes contributing to the enrichment signal. (B) Volcano plot of differentially expressed genes (DEGs) at baseline in children with (n = 33) and without (n = 46) asymptomatic viral infections, defined by a log₂ fold-change of ±0.5 and adjusted p value of 0.001. The 28 genes from interferon (IFN) signaling pathways in (A) also found in upregulated DEGs are highlighted. (C) Protein-protein interaction network of overlapping genes (n = 27, shown in orange) from IFN signaling pathways shown in (A) and upregulated DEGs in children with asymptomatic respiratory viruses shown in (B). Network was constructed using Network Analyst, the InnateDB interaction database, and the minimum network option. Proteins in orange scale are colored by log₂ fold-change of corresponding gene in the DEG analysis. Additional connecting protein nodes not found as DEGs, but are key within the network are shown in gray. Of these, CREBBP, IRF1, RELA, STAT1, and UBC are found within the IFN signaling pathway LE genes. (D) Examples of inverse correlation between baseline IFN gene expression and LAIV strain viral load (day 2, H3N2). Rho represents Spearman correlation coefficient. Values from children with asymptomatic respiratory viruses detected at baseline are colored purple. (E) Antibody response (HAI) and hemagglutinin (HA)-specific CD4+ T cell response to LAIV strains in children with and without evidence of asymptomatic respiratory viral infections at baseline. Antibody response is expressed as the fold-rise in geometric mean titer after vaccination and CD4+ T cell response as the fold-change after vaccination in the percentage of CD4+ T cells expressing IFN gamma (IFNg) following stimulation with strain-specific HA peptides. The p values are from Mann-Whitney U test. Shown are median and interquartile range for each plot.

Figure 3

Cell-type-specific gene expression signatures at baseline associated with presence of asymptomatic respiratory viruses and LAIV shedding in children seronegative to each influenza strain prior to vaccination (A) Fold-difference in cell-type-specific gene expression scores for different cell types between children with and without asymptomatic respiratory viruses pre-LAIV challenge. Cell-type-specific expression scores from bulk RNA sequencing (RNA-seq) data were generated using the sum of the normalized expression values from the top 50 genes defining each cell type derived from a previously described nasal wash single-cell RNA-seq data in influenza-infected individuals. Only cell types with a significant difference between groups (Mann-Whitney U test; Figure S1) are displayed. The fold-difference was calculated using the median values for children positive and negative for asymptomatic respiratory viruses. (B) GSEA using the Spearman correlation coefficients between rlog prior to LAIV challenge and strain-specific viral loads (log₁₀ EID₅₀/mL) at days 2 and 7 as rank and cell-type-specific gene set as in Cao et al. (2020). The size of circles is proportional to the adjusted −log₁₀ p value from GSEA, while the intensity of the circle color denotes the NES for each cell-type/strain-specific viral load combination (deeper color represents higher NES). Blue circles denote enrichment of genes in nasopharynx prior to LAIV challenge that negatively correlate with viral load at day 2 and day 7, and red circles denote enrichment of genes that positively correlate with viral loads. Significant pathways (adjusted p value < 0.1) are highlighted with a black outline. (C) Correlation between cell-type-specific gene expression scores and NESs for the IFN signaling pathway from single sample GSEA using the Reactome pathway set. Cell-type-specific scores were calculated following exclusion of any genes also present in the IFN signaling pathway gene set to avoid correlation due to these overlapping genes (classical dendritic cells [DCs]: HLA-DP1, HLA-DPB1, HLA-DQA1, HLA-DQB1, HLA-DQB2, HLA-DRB5, IRF8; no overlapping genes for macrophage, eosinophil, or goblet cell gene sets). R, Spearman correlation coefficient.