Identification of cow milk epitopes to characterize and quantify disease-specific T cells in allergic children

Abstract

Background: Cow milk (CM) allergy is the most prevalent food allergy in young children in the United States and Great Britain. Current diagnostic tests are either unreliable (IgE test and skin prick test) or resource-intensive with risks (food challenges).

Objective: We sought to determine whether allergen-specific T cells in CM-allergic (CMA) patients have a distinct quality and/or quantity that could potentially be used as a diagnostic marker.

Methods: Using PBMCs from 147 food-allergic pediatric subjects, we mapped T-cell responses to a set of reactive epitopes in CM that we compiled in a peptide pool. This pool induced cytokine responses in in vitro cultured cells distinguishing subjects with CMA from subjects without CMA. We further used the pool to isolate and characterize antigen-specific CD4 memory T cells using flow cytometry and single-cell RNA/TCR sequencing assays.

Results: We detected significant changes in the transcriptional program and clonality of CM antigen-specific (CM+) T cells elicited by the pool in subjects with CMA versus subjects without CMA ex vivo. CM+ T cells from subjects with CMA had increased percentages of FOXP3⁺ cells over FOXP3^- cells. FOXP3⁺ cells are often equated with regulatory T cells that have suppressive activity, but CM+ FOXP3⁺ cells from subjects with CMA showed significant expression of interferon-responsive genes and dysregulated chemokine receptor expression compared with subjects without CMA, suggesting that these are not conventional regulatory T cells. The CM+ FOXP3⁺ cells were also more clonally expanded than the FOXP3^- population. We were further able to use surface markers (CD25, CD127, and CCR7) in combination with our peptide pool stimulation to quantify these CM+ FOXP3⁺ cells by a simple flow-cytometry assay. We show increased percentages of CM+ CD127^-CD25⁺ cells from subjects with CMA in an independent cohort, which could be used for diagnostic purposes. Looking specifically for T_H2 cells normally associated with allergic diseases, we found a small population of clonally expanded CM+ cells that were significantly increased in subjects with CMA and that had high expression of T_H2 cytokines and pathogenic T_H2/T follicular helper markers.

Conclusions: Overall, these findings suggest that there are several differences in the phenotypes of CM+ T cells with CM allergy and that the increase in CM+ FOXP3⁺ cells is a potential diagnostic marker of an allergic state. Such markers have promising applications in monitoring natural disease outgrowth and/or the efficacy of immunotherapy that will need to be validated in future studies.

Keywords: Food allergy; T-cell epitopes; TCR repertoires; antigen-specific T cells; cow milk; cow milk allergy; food allergy diagnostics; pediatric food allergy; single-cell RNA-seq.

Figure 1:. Discovery and assessment of a T cell epitope megapool from CM extract.

A) 2D immunoblot of CM extract showing protein spots (green) reactive with IgG(pink), IgE(blue) or both(white). 96 circled spots were selected for mass spectrometry analysis. B) Table showing the known allergen or novel antigen associated with each spot on the gel as determined by mass spectrometry. C) Peptide pools were evaluated by 14-day culture followed by Fluorospot assay. Bar graph showing the percentage of IL-5 responders across all pools for each antigen. D) Screened peptides were assessed for homology to human peptides using PepMatch. The pie charts show percentage of indicated reactive or non-reactive peptides where the top human homolog hit had the indicated number of mis-matched amino acids. E) Table describing the contents and peptide inclusion criteria for each of the three megapools. F,G) Bar graphs showing percentage of responders in (F) allergic (CMA-BR and CMA-BT) versus non-CMA controls or (G) CMA-BR versus CMA-BT subjects expressing IL-5, IFNγ, and IL-10 (n= 11 controls, 3 CMA-BT, 5 CMA-BR) to the megapools, CM extract (CME) or PHA.

Figure 2:. scRNA-Sequencing of M111 epitope megapool reactive memory CD4 T cells.

A) Experimental design for the study. PBMC from Control, CMA-BT, and CMA-BR subjects were stimulated with the M111 megapool for 6 hours and antigen specific cells (CM+) were sorted using flow markers CD137 and CD154 for 10X scRNA-Seq. B) Dot plot showing percentage CM+ cells in each donor. Median value is indicated. C) UMAP of 98,883 cells colored by CM+ and CM− sort fractions. D) UMAP colored by identified Seurat cluster. E) Dot plot showing expression of TNFRSF9 (CD137) and CD40LG (CD154) in each cluster where size represents percent of cells expressing the marker and color represents expression level. F) Stacked bar graphs representing the distribution of clusters within CM+ and CM− sort fractions to determine true antigen specific cells. Statistics were performed using a Yates’ continuity corrected chi-square test. G) CM− and CM+ cells were downsampled to the same number of cells per donor and any donor with fewer than 30 cells was removed. Shannon and Chao diversity indexes were calculated and statical analysis performed using paired Wilcoxon tests. *=p<0.05, **=p<0.01, ***=p<0.001, and ****=p<0.0001.

Figure 3:. Increase in clonally expanded CM+ FOXP3+ cells with CMA-BR by scRNA-Seq and flow cytometry.

CM+ cells from clusters 3, 7, and 15 were subset from the total dataset. A) Heatmap showing expression of genes differentiating clusters 3, 15, and 7 where each row is a gene and each column a cell. Selected genes with high expression in each cluster are boxed to the left and clusters are identified as FOXP3+ or FOXP3−. B) Dot plots showing the proportions of cells in each of the three clusters per donor and separated by allergic group. C) Plot showing the ratio of CM+ FOXP3−/FOXP3+ cells per donor where the CMA group is shown both as a total CMA and split into severity groups. D) Stacked bar graphs showing clone sizes within each cluster as a proportion of cells in the cluster. E) Dot plot showing expression of genes differentiating the FOXP3+ and FOXP3− clusters where size represents percent of cells expressing the marker and color represents expression level. F) Example flow plots showing CD127(IL7R)/CD25(IL2RA) expression in CM+ cells. G) Quantification of the gate drawn in (F) compared across allergy groups. H) Example flow plots showing CD127(IL7R)/CCR7 expression in CM+ cells. I,J) Quantification of the gates drawn in (H) compared across allergy groups. Median value is indicated. Statistics were performed by Mann-Whitney test with correction for multiple comparisons where appropriate. *=p<0.05, **=p<0.01. Trending values are included.

Figure 4:. Strong interferon response signature in CM+ FOXP3+ cells increased with CMA-BR.

CM+ FOXP3+ cells were subset and re-clustered. A) UMAP colored by Seurat cluster and annotated by distinguishing gene expression signatures. B) Pie charts showing representation of cells from each allergy group within each cluster. Number in paratheses is the number of cells in that cluster. C) Pie charts showing relative clone sizes within each cluster. D) Volcano plot showing differentially expressed genes in total CM+ FOXP3+ cells between CMA-BR and Control groups. Colored dots are statistically significant (adjusted p value < 0.05; control=blue, CMA-BR=red) and genes involved in indicated GO processes are labeled. E) Dot plot showing expression of selected genes in each allergy group where size represents percent of cells expressing the marker and color represents expression level. F) CM+ FOXP3+ cells were merged with CM− FOXP3+ cells and re-clustered. UMAP is colored by indicated cluster. G) Violin plots showing expression levels of interferon responsive genes associated with CM+ FOXP3+ cluster C2.

Figure 5:. Pathogenic, clonally expanded Th2 cells increased with CMA.

CM+ FOXP3− cells were subset and re-clustered. A-C) UMAP of 6,568 re-clustered CM+ FOXP3− cells colored by (A) Seurat cluster, (B) Th2 Marker Module Score and (C) clone size. D) Violin plot showing expression of the Th2 Marker Module Score across all CM+ FOXP3− cells with cut-off line drawn at 0.25. E) Dot plot of the abundance of Th2 Marker high cells per donor within the CM+ FOXP3− cells. Median value is shown. Statistics were calculated by Mann-Whitney test where *=p<0.05. F) Stacked bar graphs showing relative clone sizes in the Th2 marker high subset vs. the rest of the CM+ FOXP3− Th cells. Cell number in each group is indicated. G) Dot plot showing expression of selected genes in the Th2 Marker high cells where size represents percent of cells expressing the marker and color represents expression level. H) Bar plot showing top functional enrichment pathways of upregulated genes in CMA-BR subjects determined from total CM+ FOXP3− cells.