Transcriptomic analysis of human IL-7 receptor alpha low and high effector memory CD8+ T cells reveals an age-associated signature linked to influenza vaccine response in older adults

Abstract

Here, we investigated the relationship of the age-associated expansion of IL-7 receptor alpha low (IL-7Rα^low ) effector memory (EM) CD8⁺ T cells with the global transcriptomic profile of peripheral blood cells in humans. We found 231 aging signature genes of IL-7Rα^low EM CD8⁺ T cells that corresponded to 15% of the age-associated genes (231/1,497) reported by a meta-analysis study on human peripheral whole blood from approximately 15,000 individuals, having high correlation with chronological age. These aging signature genes were the target genes of several transcription factors including MYC, SATB1, and BATF, which also belonged to the 231 genes, supporting the upstream regulatory role of these transcription factors in altering the gene expression profile of peripheral blood cells with aging. We validated the differential expression of these transcription factors between IL-7Rα^low and ^high EM CD8⁺ T cells as well as in peripheral blood mononuclear cells (PBMCs) of young and older adults. Finally, we found a significant association with influenza vaccine responses in older adults, suggesting the possible biological significance of the aging signature genes of IL-7Rα^low EM CD8⁺ T cells. The results of our study support the relationship of the expansion of IL-7Rα^low EM CD8⁺ T cells with the age-associated changes in the gene expression profile of peripheral blood cells and its possible biological implications.

Keywords: IL-7 receptor alpha; age; gene expression; human; memory CD8+ T cells; vaccine response.

Figure 1

IL‐7Rα^low and ^high effector memory CD8⁺ T cells have distinct global gene expression profiles. Gene expression microarray analysis was done on IL‐7Rα^low and ^high effector memory (EM) CD8⁺ T cells purified from PBMCs of healthy donors (n = 3). (a) Heatmap displays of the differential expression of 774 genes. Red and blue indicate up‐ and downregulations, respectively. (b) Enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways by 380 upregulated and 394 downregulated genes in IL‐7Rα^low cells compared to IL‐7Rα^high cells. (c) Bar plot shows expression levels of the genes encoding the molecules known to be up or downregulated in IL‐7Rα^low cells compared to IL‐7Rα^high cells

Figure 2

IL‐7Rα^low effector memory CD8⁺ T cells express a set of aging signature genes that are associating with human chronological age. (a) Venn diagram depicts overlapping genes between the differentially expressed genes (DEGs) of IL‐7Rα^low versus ^high effector memory (EM) CD8⁺ T cells and age‐associated genes identified from a meta‐analysis of whole‐blood global gene expression microarrays on approximately 15,000 human subjects. p value indicates significance of the number of overlapping genes based on a random permutation strategy (see Methods). (b) Scatter plot of the overlapping genes shows a correlative relationship between log2‐fold changes of IL‐7Rα^low versus ^high EM CD8⁺ T cells and age‐associated z scores from Figure 1 and the meta‐analysis in (a), respectively. Dots on the plot indicate individual genes. Genes with the top 10 highest age‐associated z scores are highlighted with labels and arrows. (c) Bar plot shows enriched biological processes of up and downregulated aging signature genes in IL‐7Rα^low EM CD 8⁺ T cells in comparison with IL‐7Rα^high EM CD8⁺ T cells

Figure 3

A network model describes interactions between key transcription factors and the aging signature genes regulated by these transcription factors. (a,b) Bar plots display the numbers and −Log10 p values of the target genes for the individual transcription factors. p values were computed using Fisher's exact test. (c) A network model describes regulatory relationships between transcription factors and functional modules. The relationships of the age‐associated z scores of individual genes with the expressional fold changes of these same genes between IL‐7Rα^low and ^high cells are shown as the nodes and border colors, respectively. Gray solid and pink arrowed edges represent protein–protein and transcription–target interactions respectively

Figure 4

IL‐7Rα^low effector memory CD8⁺ T cells have differential expression of the transcription factors MYC, SATB1, and BATF that change in peripheral blood mononuclear cells with age. (a) qPCR analysis of the indicated genes in IL‐7Rα^high and ^low effector memory CD8⁺ T cells purified from PBMCs of 5–10 mixed‐age human subjects. (b) Western blot analysis of MYC, SATB1, and BATF in IL‐7Rα^high (H) and ^low (L) effector memory CD8⁺ T cells purified from PBMCs of 6–9 mixed‐age human subjects. Representative blots and the relative density of each molecule compared to β‐actin are shown. (c) qPCR analysis of the indicated genes in peripheral blood of young (n = 30) and older (n = 24) adults. Bars and error bars indicate the means ± and SEM (a,b). Bars indicate the means (c). p values were obtained by the paired (a,b) and unpaired (c) t tests

Figure 5

IL‐7Rα^low effector memory CD8⁺ T cells have differential expression of FGFBP2, GZMH, and CX3CR1 that change in peripheral blood mononuclear cells with age. (a) qPCR analysis of the indicated genes in IL‐7Rα^high and ^low effector memory CD8⁺ T cells purified from PBMCs of 5–10 mixed‐age human subjects. Bars and error bars indicate means ± and SEM. (b) Flow cytometric analysis of fibroblast growth factor binding protein 2 (FGFBP2), granzyme H (GZMH), granzyme B (GZMB), and CX3CR1 in IL‐7Rα^high and ^low effector memory CD8⁺ T cells of PBMCs. Representative data from 5 to 10 mixed‐age human subjects. (c) qPCR analysis of indicated genes in peripheral blood of young (n = 35–43) and older (n = 28–33) adults. Bars indicate the means. p values were obtained by the paired (a) and unpaired (c) t tests

Figure 6

Association of the aging signature genes of IL‐7Rα^low effector memory CD8⁺ T cells with vaccine responses in young and older adults. (a,b) Bar plots display normalized enrichment scores (NES) of the target genes of the indicated transcription factors in the Yale 1 and Yale 2 cohorts. Gene Set Enrichment Analysis (GSEA) was performed in older (a, age ≥ 65) and young (b, age ≤ 35) adults, separately. NES indicate the distribution of genes in the network model across a list of genes ranked by signal‐to‐noise ratio between two groups. Higher NES indicates a shift of genes belonging to the network toward either end of the ranked list, representing up (positive NES) or down (negative NES) regulation (see Figure S1). (c,d) Boxplots display top 10 gene scores in vaccine responders and nonresponders of older and young adults from the Yale 1 (c) and Yale 2 (d) cohorts. (e) Receiver operating characteristic (ROC) curves show the performance of a logistic regression model examining the relationship between the top 10 gene score and vaccine responses in older adults of the Yale 1 and Yale 2 cohorts. (f) Boxplot depicts top 10 gene scores in vaccine responders and nonresponders of older adults (age ≥ 65) from the Mayo cohort. (g) ROC curve shows performance of the regression model in older adults (age ≥ 65) of the Mayo cohort. (H) Boxplots display the differential expression of indicated genes associated with immunosenescence (B3GAT2/CD57) and cytotoxicity (GZMH, PRF, and FGFBP2) of T cells between vaccine responders and nonresponders in older adults of the Mayo cohort. Vaccine responses were determined by hemagglutination inhibition (HAI) antibody responses (Yale 1 and Yale 2 cohorts) or a combination of HAI antibody responses and B‐cell Enzyme Linked ImmunoSpot (ELISPOT) count values (Mayo cohort). The boxes show the 25th‐75th percentile range and the center line is the median. Whiskers show 1.5 times IQR from the 25th or 75th percentile values. Data points beyond the whiskers are displayed using dots. p‐values were obtained by the Wilcoxon rank‐sum statistics (c,d,f,h)