Cell-by-Cell Deciphering of T cells in Allergic Inflammation

Abstract

Technical advances in single-cell RNA sequencing (scRNA Seq) render it possible to examine the transcriptome of single cells in allergic inflammation with high resolution in the context of their specific microenvironment, specific treatment, and disease status. Using a recently published scRNA Seq study of tissue T cells as an example, we introduce the major pipeline steps, illustrate the options of scRNA Seq platforms, summarize the new knowledge gained from this study, and provide directions for future research. The presented scRNA Seq study elucidated the T cell heterogeneity present in an allergic inflammatory tissue, focused on eosinophilic esophagitis (EoE), a prototypic, chronic, allergic disease that provided a unique opportunity to probe the pathogenesis of allergic inflammation on the tissue level through readily available, endoscopically procured biopsies. The scRNA Seq analysis identified 8 populations of CD3+ T cells and defined 2 disease-specific populations of CD3+CD4+ T cells, including a markedly activated type 2 cytokine-producing pathogenic cell population distinguished by expression of the short-chain fatty acid receptor FFAR3, and a population of T regulatory-like cells. In addition to presenting and interpreting the new findings within the prior literature, we postulate about future single-cell next-generation sequencing platforms in this burgeoning field.

Capsule summary:

“What Do We Know?”

• There is clinically relevant heterogeneity in tissue T cells

• Despite a CD8+ T cell dominance in the esophageal tissue, Th2 cytokines are produced by CD4+ T cells

• There is an eosinophilic esophagitis (EoE)-specific “super activation” of the CD4+ tissue T cell compartment

• CD4+ T cells in allergic inflamed tissue are notably enriched (~30%) in Th2 cytokine production compared to CD4+ T cells in the circulation (typically < 0.1%)

• An unexpectedly high prevalence of “T regulatory cell–like” cells co-exist in the allergic tissue, which likely contributes to the chronicity and phenotype of the disease

• Key surface markers and transcription factors expressed by tissue Th2 cells are notably enriched in an active lipid sensory pathway that includes the receptor for short-chain fatty acids (FFAR3)

• Butyrate and other short-chain fatty acids may contribute to the disease pathogenesis. The data presented argue for a positive contributory role

“What Is Still Enigmatic?”

• Is there an allergen-specific T cell receptor (TCR) repertoire beyond oligoclonality? — prompting large-scale, high-throughput TCR clonotyping

• Can the plasticity of Treg cells, if they exist in EoE tissue in situ, be modulated for the treatment of EoE?

• What is the relative contribution of Th2 cytokines from non-T cells, such as innate lymphoid cells (ILC2s) and mast cells in the epithelium and lamina propria?

• Will single-cell RNA sequencing (scRNA Seq) improve the understanding and eventual effectiveness of EoE treatments?

• Do the observations and conclusions from this scRNA Seq study apply to other tissues with allergic inflammation?

A typical human cell consists of a diploid genome composed of 2 copies of approximately 3 billion base pairs of DNA and over hundreds of millions of bases of mRNA differentially expressed by a myriad of cell types in the body. Advances in Next-Generation Sequencing (NGS) have allowed profiling of the collection of mRNA species (the transcriptome) expressed in specific organs, tissues, and cells, but these advances have relied on analysis of bulk populations of cells, typically a mixture of millions of cells from isolated tissue or cell culture. With the advances in single-cell capture and the automated cDNA library generation pipeline, it is now possible to examine the transcriptome of single cells, a process referred to as single-cell RNA sequencing (scRNA Seq). This breakthrough technology allows higher resolution of cellular differences and a better comprehension of the function of individual cells in the context of their specific microenvironment, specific treatment, and/or disease contexts. Conceivably, the scRNA Seq platform can achieve many unique objectives beyond conventional methodology, including identification of rare cell populations, defining disease subtypes, discovery of novel cellular markers, characterization of cellular heterogeneity and subsets, elucidation of disease mechanisms, and opportunity for precision and personalized medicine. The basic process of scRNA Seq involves isolation of single cells, nucleic acid extraction, RNA reverse transcription and amplification, cDNA library preparation (including NGS barcoding), NGS, and bioinformatic data analyses (typical pipeline depicted in Figure 1).

Figure 1. Schematic work flow of single-cell RNA sequencing with tissue cells from biopsy tissue.

A visual summary of the major steps involved in the single-cell RNA sequencing platform that was applied to studying tissue cells isolated from biopsies is shown. The major steps are listed sequentially, consisting of 4 indispensable modules: single-cell acquisition, single-cell barcoding (to ensure each single cell is specifically represented by a unique molecular DNA sequence), cDNA library generation, and next-generation sequencing (NGS).

To date, scRNA Seq has begun to be employed in exploring circulating and tissue-residing cells in select diseases with a focus on cancer. Its usefulness has only recently been applied to studying allergic diseases. Eosinophilic esophagitis (EoE) provides a unique opportunity to probe the molecular and cellular mechanisms of human allergic inflammation as tissue is readily available by routine endoscopy, which can provide multiple research biopsies from each patient. EoE is a prototypic severe allergic disorder mediated by gene-environment interactions (1) involving the interplay of the innate (mucosal epithelium and eosinophils) and adaptive (T cell) immune systems and driven by type 2 cytokines (Th2 cytokines), especially IL-13. EoE provides an unprecedented opportunity to scrutinize the tissue-residing cells, including the T cells, which co-migrate into esophageal mucosa with eosinophils. Driven by the goal to uncover mechanisms of type 2 immunity in EoE tissue, our team built a comprehensive platform by FACS sorting the single-cell suspension isolated from enzyme-digested human esophageal biopsies, eventually obtaining 1088 single CD3+ tissue T cells from the biopsies of patients with active EoE, disease remission, or no history of disease (controls) (2). Guided by pilot experiments in mice, our team opted for the C1 Fluidigm™ System as the platform for the scRNA Seq study (3, 4) due to its ability to directly and physically examine the single-cell chamber (thereby removing any chambers with more than one cell and/or with suboptimal morphology), quality control of the cDNA library before the NGS, lack of the 3’ bias in the synthesized cDNAs, and relatively deep number of NGS reads per cell. In the study, we aimed to answer the following questions: 1) what is the heterogeneity of tissue T cells? 2) how different are tissue CD3+ T cells from their circulating counterparts? and 3) do any EoE–specific T cell subpopulations exist and how do they causally contribute to pathogenesis? Though prior studies have identified increase type 2 cytokines in allergic tissue (5, 6), it remained unknown which cells produce these cytokines and the relative level of different cytokine-producing cells, particularly in allergic inflammation in situ. In the allergy field, controversy exists on the cellular sources of cytokines, particularly those involved in EoE. Conventional wisdom focuses on CD4+ T cells (13) and more recently on type 2 innate lymphoid cells (ILC2s) (14, 15) as the chief sources of Th2 cytokines, but there is an emerging body of literature that supports Th2 cytokine production by a variety of other cell populations, including CD8+ T cells (16, 17), invariant natural killer T cells (iNKTs) (18), specialized Th2A cells (7) , mast cells, and reprogrammed T regulatory cells (Tregs) (19, 20). Accordingly, the scRNA Seq study determined the following key findings in the allergic inflammation in EoE tissue: 1) conventional CD3+ T cells produce large amounts of type 2 cytokines in the epithelial mucosa; 2) a relatively large fraction of tissue CD4+ T cells constitutively produce type 2 cytokines without ex vivo T cell stimulation that is typically required to detect Th2 cytokine production ex vivo (e.g., allergen, phorbol 12-myristate 13-acetate [PMA], anti-CD3/CD28); and 3) though highly abundant and producing large amounts of type II interferons, CD8+ T cells do not substantially and directly contribute to type 2 cytokine production. As a caveat, all these findings are in the context of allergic inflammation in the esophageal epithelium.

The classical theory has proposed that the type 2 inflammation observed in EoE tissue originates as a result of local, food allergen–stimulated Th2 cells. The analysis of human Th2 cells in the tissue is rather limited, with only one study showing augmented IL-5+ T helper cells in the blood of subjects with active EoE (8). For atopic diseases other than EoE, nearly all of the published studies focused on deciphering type 2 cytokines are based on tissue immunostaining or flow cytometric analysis. Furthermore, detection of Th2 cytokine production in isolated cells has been primarily confined to cells that are polyclonally stimulated in vitro. In contrast, the scRNA Seq analysis did not require ex vivo stimulation of the T cells, perhaps because 1) we were probing the specific allergic inflammatory site and 2) the high sensitivity of the employed scRNA Seq technique. We found that the type 2 cytokines were produced by ~10% of the freshly isolated/unmanipulated tissue T cells in EoE tissue, a phenomenon verified at mRNA and protein levels. In the T helper compartment, Th2 cytokine-capable T cells reached the surprisingly high percentage of > 30% of CD4+ T cells in EoE. This enrichment contrasts with the frequency of these cells in the blood—typically < 5% of CD4+ T cells even in individuals with allergy (9–11). The enrichment of bona fide Th2 cells in EoE tissue suggests that there is an active in situ differentiation process in addition to the recruitment of circulating Th2 cells. Although the scRNA Seq data indicated that the pathogenic effector Th2 cells and the Treg-like cells are largely CD103 (integrin αE – the reported intraepithelial lymphocyte marker) negative, it is still unclear which component contributes dominantly, mucosal resident T cells or newly recruited T cells.

It is likely that local food allergen uptake by antigen-presenting cells in the esophagus may be involved in stimulating the production of cytokines by the pathogenic tissue effector Th2 cells. Consistent with prior literature, the type 2 cytokine-producing effector Th2 cells (e.g., those producing IL-4, IL-5, and IL-13) also expressed mRNAs that encode for CRTH2 (a PGD2 receptor), ST2L (the IL-33 receptor), IL-17RB (the IL-25 receptor), and HPGDS (a rate-limiting enzyme involved in synthesizing prostaglandins in Th2 cells) (7, 9). In addition, we identified expression of the mRNA for FFAR3 (2), which encodes a short-chain fatty acid (SCFA) receptor (12). Focusing on the transcriptome specifically expressed in these tissue Th2 cells, we used gene ontology analysis and identified a lipid-sensitive pathway that was centered on the FFAR3-SCFA axis and specific to allergic tissue. We subsequently substantiated these findings by demonstrating that FFAR3 ligation (engaging butyrate and other SCFAs) enhanced type 2 cytokine production (in vitro and in a murine model in vivo) in mouse and human systems, and, interestingly, that IL-4 induced FFAR3 expression during Th2 cell development. These data also suggested that butyrate, which can be derived from microbiota (13) and/or influenced by diet (14), may be a key positive regulator of type 2 immunity in EoE tissue (Figure 2). Indeed, butyrate-producing bacteria are present in the human esophagus, and esophageal dysbiosis occurs in EoE (15, 16). These findings add controversy to the primary viewpoint that butyrate and other SCFA have anti-inflammatory effects perhaps mediated by the alternative SCFA receptors on other cell types (e.g., FFAR2 on Tregs) (17) in the lower gastrointestinal tissue compartment. The proposed interactions of esophageal microbiota and butyrate-receiving pathogenic Th2 cells is illustrated in Figure 2.

Figure 2. The proposed short-chain fatty acid (butyrate)-mediated interactome of microbiota, EoE diet, and effector Th2 cells.

The FFAR3–short-chain fatty acid (SCFA) axis represents a novel and EoE-specific pathway identified by single-cell RNA sequencing (scRNA Seq) on tissue T cells. This finding could have significant impact on upper gastrointestinal Th2 mechanisms in EoE, as the mucosal immunocyte-luminal content interactome has been little known. The gastrointestinal luminal microbiota provides the source of butyrate as a major metabolite of the butyrate-capable flora. To a certain degree, the diet of a person with eosinophilic esophagitis (EoE) would also contribute to variations in SCFA concentration due to dietary fiber differences. These two factors would also interact with the commensal flora dysbiosis observed in EoE and collectively contribute to the Th2 pathogenesis centered on tissue effector Th2 cells. DC, dendritic cells; peTh2, pathogenic effector Th2.

The C1 platform and the associated NGS read depth provided strong sequence coverage, allowing the collaborative team to examine for the presence of T cell receptor (TCR) clonotype enrichment in EoE. In contrast to the expected allergen-specific oligoclonality in EoE tissue, abundant TCR polyclonality in the EoE T cell compartment in patients with active EoE was shown as assessed by analysis of Shannon entropy. Likewise, patients with active disease had the most diversified TCR clonality. One potential interpretation of the observed EoE TCR polyclonality is 2 fold: 1) 500 T cells are not a robust number for statistical power in this type of analysis and/or 2) even though the TCR sequences can be reliably extracted, current bioinformatic tools may not readily allow precise prediction of allergen reactivity (18). Allergen-reactive TCRs are highly diversified and dynamic (19), meaning that multiple TCR clonotypes could react with the same food allergen across a myriad of epitopes that evade the oligoclonality screening. Recently, a new chemistry of the 10x Genomics was developed for TCR compatibility (20), providing larger scale and depth for TCR analyses.

The scRNA Seq of esophageal mucosa identified two disease-associated T cell subpopulations, a population of Treg-like CD4+ T cells (enriched in EoE) and a population of pathogenic effector Th2 cells (specific to EoE). Intriguingly, even though the Treg-like cells also increase in EoE and outnumber the pathogenic Th2 cells, the influx of Treg-like cells do not successfully suppress the allergic inflammation in active disease. Our findings demonstrate that these two cell types share some gene expression profiles, such as GATA3 and CD25/IL2RA (but not type 2 cytokine production), and do not yet substantiate that these two populations may transition between each other in the tissue (21, 22). The lack of the optimal Treg inhibition could be due to 1) insufficient cell numbers that are overwhelmed by the bona fide Th2 cells; 2) inhibitory signal from the microenvironment and antigen-presenting cells; 3) temporal and spatial issues that are not addressed by single-timepoint studies; 4) lack of co-stimulation signals and induction agents (e.g., retinoic acid and hyaluronic acid) (22–24); and/or 5) lack of sufficient TGF-β expression or signaling essential for induced Treg function in the peripheral gastrointestinal mucosal tissue (25, 26). Notably, EoE is not the only disease with ineffective Treg-like cells accumulating in the gastrointestinal mucosa, as this paradox is also observed in inflammatory bowel disease (27, 28). Future studies are expected to elucidate the bona fide suppressive capacity of esophageal regulatory T cells and ways to enhance their cellular and molecular effectiveness. The identified tissue Th2 cells in the esophagus resemble the recently published “pathogenic effector Th2 cells (peTh2)” (9) and allergy-inducing “Th2A” cells (7). However, translating the observations from the esophageal allergic inflammation may not equally apply to allergic inflammation in other tissues.

EoE biopsy tissue has proven to be an invaluable tool to understand the cellular immunology of this allergic disorder, as a plethora of leukocytes and other relevant cells (e.g., epithelial and endothelial cells) are readily retrievable from this inflamed tissue (29). The C1 Fluidigm™ platform is limited by the number of cells that are analyzed and its incapacity to use a barcoded antibody for labeling the surface protein of interest (as can be done with scCITE sequencing [Cellular Indexing of Transcriptomes and Epitopes by Sequencing] (30)). Going forward, an emerging new scRNA Seq platform (10x Genomics) provides an opportunity for a higher throughput means to expand scRNA Seq to by at least 10-fold higher cell numbers; this will likely uncover an even more precise understanding of the distinct functions of each cell type and the interactive cellular mechanisms in allergic inflammation. For example, even though EoE is named and centered around eosinophil-associated inflammation, we know relatively little about the molecular and cellular function of eosinophils directly in this disease—a subject with direct therapeutic implications, given the recently emerging new classes of biological agents that now target eosinophils (31). ScRNA Seq of infiltrating eosinophils in EoE tissue will likely be a key step in understanding this cell type, its probable heterogeneity, and its role in EoE.

ScRNA Seq analysis is part of a dawn of a new era of fine-resolution technologies that are revolutionizing translational research. The field of single-cell sequencing is rapidly evolving; for example, single-cell analyses that focus on epigenomics (e.g., DNA methylome sequencing and single-cell assay for transposase-accessible chromatin with sequencing [scATAC Seq] (32)) and protein expression (e.g., scCITE Seq (30)) are now becoming readily available. From an anatomical perspective, spatial transcriptomics (33) and “paired-cell scRNA Seq” (34) provide opportunities to elucidate diseased processes cell by cell in the context of in vivo tissue structure.

This may aid in understanding the interaction of disease-specific Treg and Th2 cells in EoE. Capitalizing on these advances will be key in pursuing a better understanding of molecular etiologies, identifying novel pharmaceutical targets, and eventually treating allergic diseases by modulating the immune system.

Abbreviations:

ScRNA Seq

Single-cell RNA Sequencing

Th2

T helper type 2

NGS

Next-Generation Sequencing

EoE

Eosinophilic Esophagitis