Epigenetic dysregulation of naive CD4+ T-cell activation genes in childhood food allergy

Abstract

Food allergy poses a significant clinical and public health burden affecting 2-10% of infants. Using integrated DNA methylation and transcriptomic profiling, we found that polyclonal activation of naive CD4+ T cells through the T cell receptor results in poorer lymphoproliferative responses in children with immunoglobulin E (IgE)-mediated food allergy. Reduced expression of cell cycle-related targets of the E2F and MYC transcription factor networks, and remodeling of DNA methylation at metabolic (RPTOR, PIK3D, MAPK1, FOXO1) and inflammatory genes (IL1R, IL18RAP, CD82) underpins this suboptimal response. Infants who fail to resolve food allergy in later childhood exhibit cumulative increases in epigenetic disruption at T cell activation genes and poorer lymphoproliferative responses compared to children who resolved food allergy. Our data indicate epigenetic dysregulation in the early stages of signal transduction through the T cell receptor complex, and likely reflects pathways modified by gene-environment interactions in food allergy.

Fig. 1

Overview of study design and analysis. Clinical assessments occurred at baseline and the allergic subjects were re-assessed several years later at follow-up, at which point 59% had resolved allergy. Bloods were collected at both time points, and frozen for later analysis. Experiments were batched such that naive CD4+ T cells (baseline and follow-up) were flow sorted from bloods prior to cell culture and molecular analyses. NA non-allergic, FA food allergy

Fig. 2

Dynamic remodeling of epigenetic landscape in naive T cells following activation. a Genome-wide view of the –log 10 P value for 4154 differentially expressed genes (red circles) and 558 differentially methylated CpG (blue circles). b Heatmap of average gene expression (rows) for select genes. Cells are colored by the level of expression (blue, low; red, high). c Heatmap shows differentially methylated CpGs by genomic location (y-axis) relative to distance from transcriptional start sites (x-axis). Each panel shows enrichment for specific histone marks. d Relationship between differential methylation and gene expression for selected genes. X-axis shows the difference in expression (activated minus quiescent cells) as percent methylation (10⁻²). Y-axis shows the log 2 fold change. Points in red were significantly differentially methylated and expressed (remodeled genes) at a genome-wide level. Teff: T effector

Fig. 3

Unsupervised clustering analysis of T cell activation genes reveals variation by food allergy status. Principal component clustering analysis of 12-month patient samples based on 4154 T cell activation response genes (a) and 558 CpGs (b). Samples are labeled according to activation status, 0 = quiescent, 1 = activated, and colored according to phenotype, blue = non-allergic, red = food allergy

Fig. 4

T cell hypo-responsiveness in allergic children is underpinned by altered remodeling of metabolic and inflammatory genes. a Proliferative responses and cell viability following T cell activation. Data are expressed as fold change calculated as post–pre-activation cell counts, with bars showing median and interquartile range. Groups were compared using the Mann–Whitney test. ***P < 0.001. NA.con non-allergic quiescent, NA.act non-allergic activated, FA.con food allergy quiescent, FA.act food allergy activated. b Supernatant cytokines, with data expressed as median with interquartile range. Groups were compared using the Mann–Whitney test. c Genome-wide view of the –log 10 P value for 1412 differentially expressed genes (red circles) and 189 differentially methylated CpG (blue circles). d Similarity (Forbes coefficient) between allergy-dmrs and regions marked by activating histone modifications in naive and effector/memory cells, respectively. Similarity calculated by ratio of observed/expected overlap between these regulatory regions and allergy-dmrs. e Relationship between differential methylation and gene expression. X-axis shows delta value expressed as percent methylation (10⁻²) for the comparison of cases–controls. Y-axis shows the log 2 fold change. Points in red were differentially methylated and expressed (remodeled genes) at the genome-wide level. Eff effector, mem memory

Fig. 5

Analysis of allergy candidates at follow-up between persistent and resolved allergics. a The top panels show representative interactions plots for 4 of the 24 remodeled candidates identified at baseline. Children with persistent allergy show cumulative epigenetic perturbation with age. Filled and dashed lines represent the trajectory for the change over time in mean methylation levels (10⁻²). Probe Ids and gene names are shown above, and full statistics are provided in Table 4. b Proliferative responses and cell viability following T cell activation at follow-up are shown as median with interquartile range. c Cytokine production was expressed as median with interquartile range. For b and c groups were compared using the Mann–Whitney test. Exact P values are shown. For cell viability, exact P value was derived from the one-way ANOVA test