Platelet response to influenza vaccination reflects effects of aging

Abstract

Platelets are uniquely positioned as mediators of not only hemostasis but also innate immunity. However, how age and geriatric conditions such as frailty influence platelet function during an immune response remains unclear. We assessed the platelet transcriptome at baseline and following influenza vaccination in Younger (age 21-35) and Older (age ≥65) adults (including community-dwelling individuals who were largely non-frail and skilled nursing facility (SNF)-resident adults who nearly all met criteria for frailty). Prior to vaccination, we observed an age-associated increase in the expression of platelet activation and mitochondrial RNAs and decrease in RNAs encoding proteins mediating translation. Age-associated differences were also identified in post-vaccination response trajectories over 28 days. Using tensor decomposition analysis, we found increasing RNA expression of genes in platelet activation pathways in young participants, but decreasing levels in (SNF)-resident adults. Translation RNA trajectories were inversely correlated with these activation pathways. Enhanced platelet activation was found in community-dwelling older adults at the protein level, compared to young individuals both prior to and post-vaccination; whereas SNF residents showed decreased platelet activation compared to community-dwelling older adults that could reflect the influence of decreased translation RNA expression. Our results reveal alterations in the platelet transcriptome and activation responses that may contribute to age-associated chronic inflammation and the increased incidence of thrombotic and pro-inflammatory diseases in older adults.

Keywords: RNASeq; age-specific immunity; flow cytometry; frailty; immunosenescence; platelets; tensor decomposition; vaccination.

FIGURE 1

Analysis of RNASeq pre‐vaccination data, (a) PCA using the 500 most variable genes, (b) correlation of PC1 with platelet activation marker expression, (c) expression heatmap of most variable RNAs (RNAs starting with ‘MT‐’ are mitochondrial genes)

FIGURE 2

Schematic of non‐negative CP tensor decomposition (NCPD) for platelet vaccine response data. The RNA‐by‐participant‐by‐day tensor is decomposed into components that represent a time‐course pattern of RNA expression for a subset of RNAs across a subset of participants. Such components can be correlated with activation or deactivation of pathways in specific groups

FIGURE 3

Sample component scores from non‐negative CP tensor decomposition (NCPD) of time‐course platelet transcriptomic response, (a) hierarchical clustering and (b) PCA of sample component scores. (c) Association of component scores with age (Spearman correlation) and group (Kruskall‐Wallis test), biological sex, and vaccine response (Mann‐Whitney U test). Bolded values indicate significance of association (p < 0.10)

FIGURE 4

Tensor components related to platelet activation and age group. (a,c,e) Participant and Day scores for Components 1, 2, and 5, respectively, (b,d,f) expression levels for the top 5 scoring participants and top 2 scoring RNAs in each component, (g) Venn diagram of overlapping RNAs in Components 1, 2, and 5, (h) overrepresented Reactome pathways shared by the three components

FIGURE 5

Evaluation of protein and RNA expression of the platelet activation markers p‐selectin (CD62p, encoded by the SELP gene), CD63, and CD40L. (a,c) Scatter plots depicting log‐normalized gene expression for SELP and percent cells positive for SELP (CD62p) as measured by flow cytometry, respectively; (b,d) Log2 fold change at day 2 and 7 post‐vaccine relative to prevaccination levels of SELP gene expression (a) or SELP (CD62p) surface expression (c) in Young (RNASeq, n = 28; flow cytometry n = 14), Older (Comm) (RNASeq, n = 20; flow cytometry, n = 17) and Older (SNF) (RNASeq n = 17, flow cytometry n = 17) adults; (e,f) Generalized linear mixed effect models for log normalized expression (RNASeq) and percent cells positive (flow cytometry) of the three markers in (e) RNASeq and (f) flow cytometry at prevaccination, and days 2 and 7 post‐vaccination. Significance for (b), (d): ***, p < 0.001; **, p < 0.01; *, p < 0.05; open bracket, p < 0.10; significance for (f): orange asterisks indicate the Older (Comm) group was significantly different from Young and Older (SNF) at all days and all time points at adj. p value at least <0.01; purple asterisks indicate that Young and Older (SNF) were significantly different at day 0 at adj. p value at least <0.05

FIGURE 6

Young adult vaccine high responders show different expression trajectories compared to low responders. (a) Trajectories for high‐ and low responders of top 5% scoring Component 2 RNAs in Young adults. The fraction of RNAs that are significantly differentially expressed from this set listed on top. Inset: Component 2 score for Young high‐ vs. low responders. (b) Over‐represented Reactome pathways for top Component 2 genes, (c) Mean trajectories for Young high‐ and low‐responders for genes PF4 and PPBP