The TNF family member TL1A induces IL‐22 secretion in committed human T_h17 cells via IL‐9 induction

Short abstract

Mechanism identified by which TL1A differentially regulates the expression of TH17 effector cytokines IL‐17 and IL‐22, in human TH17 cells.

Abstract

TL1A contributes to the pathogenesis of several chronic inflammatory diseases, including those of the bowel by enhancing T_H1, T_H17, and T_H2 responses. TL1A mediates a strong costimulation of these T_H subsets, particularly of mucosal CCR9⁺ T cells. However, the signaling pathways that TL1A induces in different T_H subsets are incompletely understood. We investigated the function of TL1A on human T_H17 cells. TL1A, together with TGF‐β, IL‐6, and IL‐23, enhanced the secretion of IL‐17 and IFN‐γ from human CD4⁺ memory T cells. TL1A induced expression of the transcription factors BATF and T‐bet that correlated with the secretion of IL‐17 and IFN‐γ. In contrast, TL1A alone induced high levels of IL‐22 in memory CD4⁺ T cells and committed T_H17 cells. However, TL1A did not enhance expression of IL‐17A in T_H17 cells. Expression of the transcription factor aryl hydrocarbon receptor, which regulates the expression of IL‐22 was not affected by TL1A. Transcriptome analysis of T_H17 cells revealed increased expression of IL‐9 in response to TL1A. Blocking IL‐9 receptor antibodies abrogated TL1A‐induced IL‐22 secretion. Furthermore, TL1A increased IL‐9 production by peripheral T_H17 cells isolated from patients with Crohn’s disease. These data suggest that TL1A differentially induces expression of T_H17 effector cytokines IL‐17, ‐9, and ‐22 and provides a potential target for therapeutic intervention in T_H17‐driven chronic inflammatory diseases.

Abbreviations

AhR

aryl hydrocarbon receptor

BATF

basic leucine zipper transcription factor, ATF‐like

CD

Crohn’s disease

DR3

death receptor 3

IBD

inflammatory bowel disease

ILC

innate lymphoid cell

PMA

phorbol myristate acetate

RORγt

RAR‐related orphan receptor γ

UC

ulcerative colitis

Introduction

TL1A, a member of the TNF superfamily, and its receptor DR3 are up‐regulated during chronic intestinal inflammation in UC and CD and in mouse models of chronic intestinal inflammation [1, 2, 3, 4–5]. Its expression is mainly confined to inflamed tissues of the colon and small bowel. Both TL1A and its receptor DR3 are particularly highly expressed on gut‐homing CD4⁺CCR9⁺ mucosal T cells [1, 6]. Its expression is also markedly enhanced in lamina propria macrophages from patients with CD [5].

TNFSF15, the gene encoding TL1A, has been identified as a susceptibility and severity gene in IBD by genome‐wide association studies [7]. We have demonstrated that patients with CD who carry a TNFSF15 risk haplotype have higher TL1A protein expression by peripheral blood monocytes and develop significantly more intestinal strictures and worse inflammation in the ileocolonic region [8, 9]. We, and others, have shown that transgenic mice that overexpress TL1A in the myeloid or lymphoid compartment develop small‐intestine inflammation and fibrosis that are characterized by overexpression of IFN‐γ, IL‐17, and IL‐13 [8, 10, 11]. On the contrary, administration of neutralizing TL1A antibodies attenuates chronic colitis by affecting both T_H1 and T_H17 responses, which suggests that it is a major regulator of intestinal inflammation during colitis [12]. It promotes proliferation of T cells and augments IFN‐γ production by T_H1 cells and IL‐17 production by T_H17 cells [6, 12, 13, 14, 15–16]. It exerts its biologic effects by binding to its receptor DR3, which is primarily expressed on activated T cells, resulting in the activation of the NF‐κB and MAPK signaling pathways [17, 18].

Although compelling in vivo data endorse a role for TL1A in promoting T_H17 differentiation and effector function, its role in the differentiation of T_H17 cells in vitro is still controversial [19]. It is undisputed that it enhances the secretion of IL‐17 in effector T cells, but it has been demonstrated that it can inhibit human and murine T_H17 differentiation in vitro by mechanisms that are independent of STAT1 and IL‐2 signaling [20]. The exact inhibitory signaling mechanisms of TL1A during the differentiation of naïve T cells into T_H17 cells and the stimulatory signaling pathways engaged by it in memory T cells and committed T_H17 cells remain to be elucidated.

In this study, we investigated the effect of TL1A on cytokine secretion by committed peripheral human T_H17 and memory CD4⁺ T cells. TL1A, in the presence of TGF‐β, IL‐6, and IL‐23, enhanced the secretion of IL‐17 and IFN‐γ from human memory CD4⁺ T cells by enhancing the gene expression of the transcription factors BATF and T‐bet. Furthermore, its expression induced a cell population of IL‐17/IFN‐γ double‐positive cells. TL1A alone induced IL‐22 in memory CD4⁺ T cells. Also, stimulation with TL1A alone led to the maximum induction of IFN‐γ and IL‐22 in committed CD45RO⁺CCR6⁺ human T_H17 cells, whereas IL‐17 secretion was not affected in these cells. It did not affect transcriptional levels of the transcription factor AhR, which regulates the expression of IL‐22, suggesting a signaling pathway of TL1A‐induced IL‐22 that is independent of AhR. We performed transcriptome analysis of TL1A‐stimulated T_H17 cells and identified IL‐22 and IL9 as genes with the highest induction by TL1A. It significantly induced IL‐9 secretion in T_H17 cells in a dose‐dependent manner. Blocking IL‐9 receptor antibodies completely abrogated the TL1A‐induced IL‐22 secretion in T_H17 cells. Furthermore, TL1A significantly induced the secretion of IL‐9 but not of IL‐17 in peripheral T_H17 cells isolated from patients with CD. Our data suggest that TL1A is an important immune modulator in human T_H17 cells and leads to the induction of IFN‐γ, IL‐9, and IL‐22 in these cells, increasing their proinflammatory propensity. TL1A therefore represents an attractive interventional target in chronic T_H17‐driven inflammatory diseases.

MATERIAL AND METHODS

Isolation of CD4⁺ T cells from PBMCs and cell sorting

Blood was obtained from healthy volunteers or patients with CD after informed consent, in accordance with procedures established by the Cedars‐Sinai Institutional Review Board (IRB 3358, 2673, 6522). All the patients had received a diagnosis of CD, according to standard clinical, endoscopic, radiologic, and histologic findings. CD4⁺ T cells were isolated from PBMCs by negative selection, using the human CD4⁺ T cell Enrichment Kit (STEMCELLS Technologies, Vancouver, BC, Canada). CD4⁺ T cells were stained with anti‐CD45RO, anti‐CD45RA, anti‐CD25, and anti‐CCR6 antibodies (eBioscience, San Diego, CA, USA). CD45RA⁻CD45RO⁺CD25⁻ or CD45RA⁻CD45RO⁺CD25⁻CCR6⁺ cells were collected with the FACSAria III Cell Sorter (BD Biosciences, San Jose, CA, USA).

T cell stimulation

T cells were stimulated with immobilized anti‐CD3 (5 μg/ml; BD Biosciences) and anti‐CD28 (2 μg/ml; Bristol‐Myers Squibb, New York, NY, USA) in the presence of TGF‐β1 (3 ng/ml; PeproTech, Rocky Hill, NJ, USA), IL‐6 (50 ng/ml; PeproTech), IL‐23 (20 ng/ml; R & D Systems, Minneapolis, MN, USA), TL1A (100 ng/ml; Fitzgerald Industries International, Acton, MA, USA), and neutralizing antibodies to IL‐4 (2 μg/ml) and IFN‐γ (3 μg/ml; BioLegend, San Diego, CA, USA) for 3 days, or as indicated. In selected experiments, blocking antibodies to IL‐9 receptor (BioLegend) or isotype controls were added at the time of stimulation.

Quantitative real‐time PCR analyses

T‐bet, Batf, RORA (Integrated DNA Technologies, Coralville, IA, USA), IL‐9, AhR, and RORC (ThermoFisher Scientific, Waltham MA, USA) transcripts were amplified by quantitative real‐time RT‐PCR with TaqMan probes [21]. Batf amplification was performed with RT² SYBR Green ROX qPCR Mastermix (Qiagen, Valencia, CA, USA). Replicate C _t values were normalized to replicate reference gene (β‐actin) C _t values (^Δ C _t), and relative expression was calculated with respect to the indicated reference sample (^ΔΔ C _t). Primer/probe sequences used were T‐bet (probe: TCC GCC GTC CCT GCT TGG TGA TGA, forward: CCA AGT TTA ATC AGC ACC AGA CAG, reverse: GCC ACA GTA AAT GAC AGG AAT GG); RORA (probe: TTG ATG GGA AGT ATG CCA GC, forward: CGG TGC CTT TGA CTC TCA GAA CAA CAC CG, reverse: TCT TTC CAA ATT CAA ACA CAA AGC); and Batf (forward: CGA GTT GCT GCT CAG AGA AGT CGG, reverse: TCC TCA TGG AGC TTG TCA GCC TTC).

ELISA

Human IL‐17A, IFN‐γ, IL‐9 (all eBioscience), and IL‐22 (R&D Systems) were quantified with ELISA kits.

Flow cytometry

For the detection of intracellular cytokines, cultured cells were stimulated with PMA (50 nM), ionomycin (1 μg/ml), and monensin (eBioscience) for 4 h at 37°, or monensin was added for the last 4 h of stimulation. Cells were harvested and stained for CD4 and intracellular IL‐17 (Alexa Fluor 647‐IL‐17), IFN‐γ (FITC‐IFN‐γ), IL‐9 (eFluor 660‐IL‐9) (all eBioscience), IL‐22 (PE‐IL‐22; R&D Systems), or isotype controls, with the Fixation and Permeabilization kit (eBioscience). For selected experiments cells were stained with DR3, CCR6 (both eBioscience), CCR9, CXCR3 (both R&D Systems), CCR3 (BD Bioscience), and IL‐9. For the analysis of cell proliferation, cells were stained for CD4, IL‐9, and Ki67 (eFluor450‐Ki67, eBioscience). For the analysis of apoptosis, cells were harvested according to the manufacturer’s instructions and stained with annexin‐V (APC‐annexin‐V) and Fixable Viability dye (eFluor 780; both eBioscience) and acquired within 2 h. Cells were acquired on a CyAnTM ADP flow cytometer (Dako, Carpinteria, CA, USA) and analyzed with FlowJo software (TreeStar Inc., Ashland, OR, USA).

Gene expression analysis

RNA was isolated from stimulated T cells at the indicated time points with the RNeasy Micro kit (Qiagen). cRNA amplification and biotin labeling were performed with 500 ng total RNA using the Illumina TotalPrep RNA Amplification Kit (ThermoFisher Scientific). cRNA yield was determined with a NanoDrop spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA), and 1.5 mg biotin‐labeled cRNA was hybridized onto HumanHT‐12_V4 Expression Beadchips (Illumina, San Diego, CA, USA) and scanned with the Illumina iScan according to the manufacturer’s protocol. All samples from 5 independent experiments were run together at the same time under identical conditions. Data visualization and quality control were performed with the Gene Expression module v1.9.0 of GenomeStudio v2011.1 software (Illumina). Data were normalized and analyzed using BRB‐Array Tools (Version 4.4.0; http://brb.nci.nih.gov/BRB‐ArrayTools/; NIH, National Cancer Institute, Frederick, MD, USA). Hierarchical clustering of total array data and class comparison between control and TL1A‐treated cells were performed with the BRB‐Array Tools. Detection at P < 0.01 was tagged as significant expression and used for analysis. Ingenuity software (Qiagen) was used for gene pathway and function analysis. Microarray expression data for GSE89133 are publicly available at the Gene Expression Omnibus (GEO) repository (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE89133).

Statistics

Statistical significance was determined by Student’s t test. P < 0.05 indicated statistical significance.

RESULTS

TL1A enhances IL‐17 production by human CD45RO⁺ memory T cells

We and others have demonstrated that TL1A is a strong costimulator of T_H1 responses [4, 6, 15]. However, its effect on human memory CD4⁺ T cells and committed human T_H17 cells and the signaling pathways involved are incompletely understood. To establish the role of TL1A on IL‐17 secretion from memory CD45RO⁺ T cells, we used high‐speed flow cytometry for sorting CD45RO⁺ T cells from enriched CD4⁺ T cells isolated from PBMCs from healthy volunteers. The highly enriched CD4⁺CD45RO⁺ T cells were stimulated with anti‐CD3/CD28 in the presence of TGF‐β, IL‐6, and IL‐23 and in the presence or absence of TL1A. A combination of TGF‐β, IL‐6, IL‐23, and TL1A significantly enhanced the secretion of IL‐17 compared to TGF‐β, IL‐6, and IL‐23 ( Fig. 1A). TL1A alone significantly increased the secretion of IL‐17 from CD4⁺CD45RO⁺ memory T cells. Furthermore, stimulation of cells with TGF‐β, IL‐6, IL‐23, and TL1A significantly enhanced IFN‐γ secretion, whereas stimulation with TGF‐β, IL‐6, and IL‐23 had no effect on it (Fig. 1B). When we analyzed the secretion of IL‐22, we observed that TL1A alone led to maximum secretion of IL‐22 in CD4⁺CD45RO⁺ memory T cells (Fig. 1C). In contrast, TGF‐β suppressed IL‐6‐ and TL1A‐induced IL‐22 secretion by CD4⁺CD45RO⁺ cells, consistent with previously reported data [22]. To understand the molecular mechanisms by which TL1A enhances IL‐17 and IFN‐γ production by human CD4⁺CD45RO⁺ T cells, we analyzed the expression of the T_H17 transcription factors BATF and the T_H1 transcription factor T‐bet. BATF has been described as a pioneer transcription factor for the differentiation of T_H17 cells [23, 24]. Analysis of the expression of BATF revealed that a combination of TGF‐β, IL‐6, IL‐23, and TL1A led to maximum induction of BATF mRNA (Fig. 1D). In contrast, TL1A alone significantly increased the expression level of the T_H17 transcription factor RORγt (encoded by the human gene RORC), but did not further increase RORγt expression when stimulated with TGF‐β, IL‐6, and IL‐23 (Supplemental Fig. 1A). Similarly, it significantly increased the expression of RORA, a transcription factor that is essential for T_H17 differentiation, but did not further increase RORA expression when stimulated with TGF‐β, IL‐6, and IL‐23 [25]. Stimulation with TL1A in combination with TGF‐β, IL‐6, and IL‐23 also led to maximum induction of T‐bet mRNA (Fig. 1E), confirming our data of maximum induction of IFN‐γ in CD4⁺CD45RO⁺ T cells stimulated with TGF‐β, IL‐6, IL‐23, and TL1A. Furthermore, intracellular cytokine staining showed that TL1A, in combination with TGF‐β, IL‐6, and IL‐23, enhanced the percentage of IL‐17⁺ cells (Fig. 1F). TL1A also enhances the IFN‐γ⁺ cell population and induces a cell population of IL‐17/IFN‐γ double‐positive cells. These data confirm concomitant secretion of IL‐17 and IFN‐γ in CD4⁺CD45RO⁺ memory T cells by stimulation with TL1A. To determine whether the enhanced induction of IL‐17 and IFN‐γ in CD4⁺CD45RO⁺ T cells by TL1A was dependent on the expression level of the TL1A receptor DR3, we stimulated CD4⁺CD45RO⁺ T cells with TGF‐β, IL‐6, and IL‐23, in the presence or absence of TL1A, and analyzed DR3 expression. Consistent with previous publications, we observed very low expression levels of DR3 in freshly isolated CD4⁺CD45RO⁺ T cells (Supplemental Fig. 1B). We observed an up‐regulation of DR3 expression in the presence of TGF‐β and IL‐6 compared to the level in unstimulated cells (Supplemental Fig. 1C). However, neither IL‐23 nor TL1A could further increase the expression of DR3, suggesting that the increase in IL‐17 and IFN‐γ production in the presence of TL1A is not due to higher DR3 expression.

Figure 1.

TL1A enhances IL‐17, IFN‐γ, and IL‐22 secretion from memory CD4⁺CD45RO⁺ T cells. Human CD4⁺CD45RO⁺ T cells were stimulated with plate‐bound anti‐CD3 and anti‐CD28 antibodies, with the addition of the indicated cytokines. Culture supernatants were analyzed for IL‐17 (A), IFN‐γ (B), and IL‐22 (C) after 72 h by ELISA. Human CD4⁺CD45RO⁺ T cells were stimulated for 48 h, and expression of BATF (D) and T‐bet (E) was analyzed by quantitative real‐time PCR. Data represent the mean of duplicates ± sd. A representative experiment of 2 (D, E) or of at least 4 (A–C) independent experiments with similar results is shown. (F) Human CD4⁺CD45RO⁺ T cells were stimulated as described. After 72 h incubation, the cells were restimulated with PMA/ionomycin for 6 h and the percentages of IFN‐γ‐ and IL‐17‐producing cells were measured by intracellular cytokine staining using flow cytometry. Data are representative of 4 independent experiments with similar results. *P < 0.05, **P < 0.01, ***P < 0.005.

TL1A increases IL‐22 production in committed T_H17 cells

It has been said that human peripheral CD45RO⁺CCR6⁺ cells are committed T_H17 cells [26, 27]. To determine whether TL1A could enhance the production of IL‐17 from committed T_H17 cells, we isolated human CD45RO⁺CCR6⁺ cells from peripheral blood and stimulated them with TL1A for 3 d. Stimulation of T_H17 cells with TL1A did not significantly enhance the secretion of IL‐17 compared to that in unstimulated cells ( Fig. 2A ). We also did not observe an effect of the stimulation on the expression level of RORC (Fig. 2B), supporting our observation that TL1A does not induce IL‐17 expression in committed T_H17 cells. However, we observed an enhancement of IFN‐γ production by TL1A alone in committed T_H17 cells, but it did not reach significance (Fig. 2C). In contrast, it significantly increased the secretion of IL‐22 (Fig. 2D). Similar to our observation in memory T cells, TL1A did not overcome the suppression of IL‐22 secretion induced by TGF‐β. To determine the time‐course of IFN‐γ and IL‐22 secretion in response to TL1A, we stimulated committed T_H17 cells for different time points in a larger cohort of human volunteers. TL1A alone significantly increased the secretion of IFN‐γ at 24 and 48 h of stimulation by 1.6‐ and 1.5‐fold, respectively (Fig. 2E). Furthermore, TL1A alone significantly increased the secretion of IL‐22 at different time points, starting as early as 6 h post stimulation up to 72 h of stimulation (4.3‐, 2.1‐, 2.0‐, and 1.8‐fold, for 6, 24, 48, and 72 h, respectively) (Fig. 2F). The expression of IL‐22 is regulated by the transcription factor AhR [28]. To elucidate the mechanism of TL1A‐induced expression of IL‐22 in T_H17 cells, we analyzed the expression of AhR at different time points. We did not observe significant changes in expression of AhR mRNA in response to TL1A stimulation at 24 or 48 h, suggesting that TL1A induces IL‐22 via AhR‐independent pathways (Fig. 2G).

Figure 2.

TL1A increases IL‐22 and IFN‐γ secretion from committed CD4⁺CD45RO⁺CCR6⁺ T_H17 cells. Human CD4⁺CD45RO⁺CCR6⁺ T_H17 cells were stimulated with plate‐bound anti‐CD3 and anti‐CD28 antibodies with the addition of the indicated cytokines. Supernatants were analyzed for IL‐17 (A), IFN‐γ (C), and IL‐22 (D) after 72 h by ELISA. (B) Expression of RORC was analyzed by quantitative real‐time PCR. Data represent the mean of duplicates ± sd. A representative experiment of 3 with similar results is shown. Box‐and‐whisker plot showing the IFN‐γ (E) or IL‐22 (F) concentrations in culture supernatants for control and TL1A‐treated cells. The boxes represent the median and 25th and 75th percentiles, and the whiskers represent the minimum and maximum values of the distribution (IFN‐γ, n = 6‐7; IL‐22, n = 5 for 6 h, and at least 20 for 24, 48, 72 h). (G) Expression of AhR was analyzed by quantitative real‐time PCR. Data represent mean results of 3 to 4 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.005.

Increased IL‐22 production in committed T_H17 cells by TL1A stimulation is driven by IL‐9 secretion

To determine the mechanism of TL1A‐induced IL‐22 production in T_H17 cells, we performed global gene expression analysis of TL1A‐stimulated T_H17 cells ( Fig. 3A, B ; Supplemental Fig. 2A, Supplemental Table 1). Unsupervised clustering analysis did not show significant differences between controls and TL1A‐treated cells but showed clustering at the different time points. A heat map visualization at the different time points revealed genes with expression that differed by at least 1.5‐fold in T_H17 cells stimulated with TL1A vs. controls (Fig. 3A). Among the genes that demonstrated the highest up‐regulation in TL1A‐stimulated T_H17 cells at all time points analyzed (24, 48, and 72 h) were IL‐22 and IL‐9 (Fig. 3B). We also confirmed the induction of IFN‐γ (24 h time point) at a lower fold increase than IL‐9 and IL‐22. Ingenuity Pathway Analysis (Qiagen) confirmed the IL‐9/IL‐22 pathways as the highest hit in TL1A‐stimulated T_H17 cells (Supplemental Fig. 2B, and data not shown). Because IL‐9 has been described as a T_H17‐associated cytokine, we hypothesized that the induction of IL‐9 by TL1A would lead to the up‐regulation of IL‐22. First, we confirmed our gene array data by quantitative real‐time PCR. IL‐9 mRNA expression was induced early on in TL1A‐stimulated T_H17 cells and remained up‐regulated for up to 72 h ( Fig. 4A ). Next, we confirmed TL1A‐induced IL‐9 secretion in T_H17 cells. We observed a dose‐dependent increase in IL‐9 secretion after stimulation with TL1A at 24 and 48 h (Fig. 4B). Similar to the induction of IL‐22 by TL1A in T_H17 cells, we observed that TL1A alone is a strong inducer of IL‐9, particularly at early time points (Fig. 4C). Furthermore, we observed synergistic induction of IL‐9 under T_H17 and TL1A stimulation (Fig. 4C). Concomitant with the dose‐dependent increase of IL‐9 secretion by TL1A stimulation we observed an increase in IL‐9⁺, IL‐22⁺, and IL‐9⁺IL‐22⁺ cells, as detected by intracellular cytokine staining ( Fig. 5A ). We observed a significant increase in the percentage of IL‐9⁺ cells with a concentration of 100 ng/ml of TL1A at 24 and 48 h of stimulation (Fig. 5B). On a per‐cell basis, IL‐9 output (as measured by mean fluorescence intensity) was higher in TL1A‐stimulated cells that reached significance for the highest TL1A concentration used (Supplemental Fig. 3A), suggesting that TL1A induces IL‐9 expression and enhances the IL‐9 output by IL‐9‐producing cells. We confirmed the time course of IL‐9 secretion by T_H17 cells in response to TL1A in a large cohort of human healthy volunteers. TL1A alone significantly increased the secretion of IL‐9 at 24 and 48 h after stimulation by 3.2‐, and 2.6‐fold, respectively (Fig. 5C). To assess whether the increase in the percentage of IL‐9⁺ cells after TL1A stimulation is caused by the preferential proliferation of this subpopulation, we analyzed the degree of proliferation after stimulation with TL1A in IL‐9⁺ and IL‐9⁻ T_H17 cells. The total percentage of proliferating CD4⁺ T cells was similar in control and TL1A‐stimulated cells at 48 and 72 h after stimulation (54% vs. 63% and 77% vs. 80% of CD4⁺Ki67⁺ cells, respectively). Although there was a trend toward a higher proportion of IL‐9⁺Ki67⁺ cells in TL1A conditions, it did not reach statistical significance (Supplemental Fig. 3B). We observed a population of Ki67⁻IL‐9‐producing cells in control and TL1A conditions, suggesting that proliferation was not necessary for IL‐9 production (data not shown). Next, we analyzed the degree of apoptosis in T_H17 cells treated with TL1A. We did not observe an overall increase in annexin‐V⁺ cells in TL1A‐treated T_H17 cells (Supplemental Fig. 3C). However, because of the incompatibility of annexin‐V and intracellular IL‐9 staining, we were unable to distinguish whether apoptotic cells belong to the IL‐9⁺ or IL‐9⁻ subpopulations. These data suggest that TL1A induces the expression of IL‐9, rather than leading to a preferential proliferation of IL‐9‐producing cells.

Figure 3.

Transcriptional profiling of committed CD4⁺CD45RO⁺CCR6⁺ T_H17 cells stimulated with TL1A. Human CD4⁺CD45RO⁺CCR6⁺ T_H17 cells were stimulated with plate‐bound anti‐CD3 and anti‐CD28 antibodies, with or without TL1A for the indicated time points. (A) Transcriptional profiling of human CD4⁺CD45RO⁺CCR6⁺ T_H17 cells, treated with or without TL1A by whole‐genome gene expression arrays. Heat map of transcripts with significant differences in expression between control and TL1A‐treated T_H17 cells at 24, 48, and 72 h of stimulation (P < 0.01). (B) Fold change of selected genes that were significantly up‐/down‐regulated at 24, 48, and 72 h of TL1A stimulation (P < 0.01). Data are results of 5 independent experiments.

Figure 4.

TL1A enhances IL‐9 mRNA expression and secretion from committed CD4⁺CD45RO⁺CCR6⁺ T_H17 cells. Human CD4⁺CD45RO⁺CCR6⁺ T_H17 cells were stimulated with plate‐bound anti‐CD3 and anti‐CD28 antibodies, with or without TL1A for the indicated time points. (A) Expression of IL‐9 mRNA was analyzed by quantitative real‐time PCR. Each circle represents one donor. Means are shown. (B) Supernatants were analyzed for IL‐9 at different time points by ELISA. Dose–response curve for TL1A stimulation at 24 and 48 h. (C) T_H17 cells were stimulated with the addition of the indicated cytokines for 24, 48, or 72 h. Data represent the mean of duplicate experiments ± sd. One representative experiment of 3 (B, C) with similar results is shown. *P < 0.05, **P < 0.01, ***P < 0.005.

Figure 5.

TL1A‐induced IL‐22 secretion from committed CD4⁺CD45RO⁺CCR6⁺ T_H17 cells is dependent on IL‐9 secretion. Human CD4⁺CD45RO⁺CCR6⁺ T_H17 cells were stimulated with plate‐bound anti‐CD3 and anti‐CD28 antibodies with the addition of TL1A. (A) Cells were stimulated for 24 h with the indicated concentration of TL1A. Monensin was added for the last 4 h of stimulation, and the percentages of IL‐9‐ and IL‐22‐producing cells were measured by intracellular cytokine staining with flow cytometry. Representative flow cytometry plots are shown. (B) Percentage of IL‐9⁺ cells after 24 and 48 h of stimulation with the indicated concentration of TL1A. Each circle represents one donor. Means are shown as lines. (C) Box‐and‐whisker plot showing the IL‐9 concentrations in culture supernatants for control and TL1A treated cells. The boxes represent the median and 25th and 75th percentiles, and the whiskers represent the minimum and maximum values of the distribution (n = 6 for 6 h, and 19–28 for 24, 48, and 72 h). (D) Human CD4⁺CD45RO⁺CCR6⁺ T_H17 cells were stimulated with anti‐CD3, anti‐CD28, and TL1A in the presence of blocking IL‐9 receptor antibodies or isotype controls for 2 days. Supernatants were analyzed for IL‐22 by ELISA. The secretion of IL‐22 in anti‐IL‐9 receptor antibodies or isotype control‐treated cells was compared to TL1A‐stimulated cells (set at 100% of IL‐22 secretion). Data represent the means of duplicates ± sd. One representative experiment of 3 (A, D) with similar results is shown. *P < 0.05, **P < 0.01.

To prove our hypothesis that TL1A‐induced secretion of IL‐22 is dependent on IL‐9, we used anti‐IL‐9R blocking antibodies. In agreement with our hypothesis, we observed a dose‐dependent inhibition of TL1A‐induced IL‐22 (Fig. 5D). We observed similar results with neutralizing anti‐IL‐9 antibodies (data not shown). These data demonstrate that TL1A induces IL‐22 in T_H17 cells via the early induction of IL‐9.

We performed intracellular cytokine staining to further characterize the phenotype of IL‐9‐producing cells. Only 8.6% of all IL‐9⁺ cells were also IL‐17⁺ cells, and 39.5% of all IL‐9⁺ cells were also IFN‐γ⁺ cells ( Fig. 6A ). TL1A increased the percentage of IFN‐γ⁺IL‐9⁺ cells, whereas the percentage of IFN‐γ⁺ cells did not change in the presence of TL1A. Next, we analyzed the chemokine receptor profile of IL‐9 producing T_H17 cells. We did not observe expression of CCR6, CCR9, CXCR3, or CCR3 on IL‐9⁺ cells (Supplemental Fig. 4A–D).

Figure 6.

TL1A induces IL‐9 in CD4⁺CD45RO⁺CCR6⁺ T_H17 cells from patients with CD. (A) Human CD4⁺CD45RO CCR6⁺ T_H17 cells were stimulated with plate‐bound anti‐CD3 and anti‐CD28 antibodies with the addition of TL1A. Cells were stimulated for 48 h and then restimulated with PMA, ionomycin, and monensin for 4 h, and the percentages of IL‐9‐ and ‐17‐producing cells and IL‐9‐ and IFN‐γ‐producing cells were measured by intracellular cytokine staining using flow cytometry. Representative flow cytometry plots are shown. (B, D) CD4⁺CD45RO⁺CCR6⁺ T_H17 cells were isolated from patients with CD. Cells were stimulated with indicated cytokines for 48 or 72 h. (B) Supernatants were analyzed for IL‐9, ‐17, and ‐22 at 72 h by ELISA. Each circle represents one patient. Means are shown (red lines). (C) Cells were stimulated with indicated cytokines for 48 h and then restimulated with PMA, ionomycin, and monensin for 4 h and the percentages of IL‐9‐ and IL‐22‐producing cells were measured by intracellular cytokine staining. Representative flow cytometry plots are shown. One representative experiment of 3 (A) or 2 (C) with similar results is shown. **P < 0.01.

TL1A increases IL‐9 production in committed T_H17 cells isolated from patients with CD

The expression of TL1A is increased in patients with CD and has been implicated in disease pathogenesis. To assess whether it increases the expression of IL‐9 in T_H17 cells from patients with CD, we isolated T_H17 cells from peripheral blood of affected patients. Consistent with our data from healthy donors, TL1A significantly up‐regulated IL‐9 secretion in T_H17 cells isolated from patients with CD (2.7‐ and 3‐fold increase for 48 and 72 h, respectively) (Fig. 6B; Supplemental Fig. 4E). In contrast, it did not significantly increase the secretion of IL‐17. We also observed an increase in IL‐22 (average: 1.7‐fold increase) and IFN‐γ (average: 2.0‐fold increase) secretion upon stimulation with TL1A. Intracellular staining confirmed an increase in the percentage of IL‐9⁺ cells upon stimulation (Fig. 6C). Taken together, these data show that TL1A increases the secretion of IL‐9 but not of IL‐17 in peripheral T_H17 cells from patients with CD.

DISCUSSION

TL1A, a member of the TNF superfamily, has been implicated in the pathogenesis of IBD. Genome‐wide association studies have identified TNFSF15 (the gene encoding TL1A) as an IBD susceptibility and severity gene [29]. Patients with CD carrying a TNFSF15 risk haplotype overexpress TL1A and are more prone to intestinal fibrosis [7, 30]. TL1A promotes T_H1 responses and has also been implicated in T_H17 responses, but the data demonstrating the role of TL1A in the differentiation of T_H17 cells have been conflicting [19, 20, 31], and the signaling pathways engaged by TL1A during the differentiation of T_H cells have not been clearly defined.

In the current study, TL1A greatly increased the secretion of the T_H17 cytokines IL‐17A, IL‐22, and the T_H1 cytokine IFN‐γ from memory CD4⁺ T cells. TL1A also induced IL‐22, IL‐9, and IFN‐γ from committed human T_H17 cells isolated from healthy controls and patients with CD. The T_H17 pathway has been linked to several inflammatory diseases, including colitis, psoriasis, arthritis, and experimental autoimmune encephalomyelitis [32]. It has been demonstrated that TL1A plays an important role in T_H17‐mediated diseases, including autoimmune encephalomyelitis and chronic colitis [12, 16]. We analyzed the expression of T_H17/T_H1‐driving transcription factors and observed that TL1A induces high expression of BATF and T‐bet, correlating with high expression of IL‐17 and IFN‐γ in these conditions. BATF and IRF4 have been proposed as “pioneer factors” in the differentiation of T_H17 cells by contributing to the initial chromatin accessibility and facilitating the access of other transcription factors [33]. We observed that TL1A synergized with TGF‐β, IL‐6, and IL‐23 to induce maximum BATF expression in memory CD4⁺ T cells. TL1A alone increased the expression of the T_H17 transcription factors RORA and RORC, but did not further enhance their expression in response to TGF‐β, IL‐6, and IL‐23 stimulation. Our data suggest that the induction of T‐bet and BATF by TL1A may lead to the secretion of IFN‐γ and IL‐17A in memory T cells, respectively.

TL1A induces a cell type of IL‐17/IFN‐γ coproducing T_H17 cells a phenotype that has been defined as a pathogenic T_H17 subtype [34]. Recent data derived from TL1A^−/− mice further suggest that TL1A plays an essential role in the optimal differentiation as well as effector function of murine T_H17 cells [16]. We observed strong increases of IL‐22 in memory CD4⁺ T cells and committed human T_H17 cells induced directly by TL1A without the need for costimulation by additional cytokines. In contrast to memory CD4⁺ T cells, TL1A had no effect on IL‐17 production in committed T_H17 cells isolated from healthy donors or patients with CD, suggesting a differential regulation by TL1A of IL‐22, IL‐9, and IFN‐γ on the one hand and IL‐17 on the other. Our data are consistent with recent findings of differential regulation of IL‐22 and IL‐17 in human T cells [35]. Although IL‐6 alone induces IL‐22, the combination of TGF‐β and IL‐6 results in the production of IL‐17 but not IL‐22 in T cells. We observed that TL1A alone was a stronger inducer of IL‐22 than IL‐6 in memory CD4⁺ T cells and in committed T_H17 cells. Recently, a population of memory T cells has been described that produce IL‐22 but neither IL‐17 nor IFN‐γ [36, 37]. TL1A stimulation did not affect the transcriptional levels of AhR, a transcription factor that regulates IL‐22 production in CD4⁺ T cells, suggesting that TL1A induces IL‐22, independent of AhR in T_H17 cells [38, 39]. Indeed, transcriptional profiling of human T_H17 cells stimulated with TL1A identified IL‐9 as a cytokine highly induced by TL1A in these cells. Mechanistically, we demonstrated that blocking antibodies to IL‐9 receptor‐attenuated TL1A induced secretion of IL‐22 in T_H17 cells. Although TL1A has been shown to increase proliferation in T cells, we did not observe any significant changes in proliferation in IL‐9⁺ or total T_H17 cells, suggesting that it induces the expression of IL‐9 rather than leading to a preferential proliferation of IL‐9 producing cells. Similar to the induction of IL‐17/IFN‐γ coproducing cells by TL1A, we observed that it induced IL‐22/IL‐9‐ and IFN‐γ/IL‐9‐coproducing cells. IL‐9 and T_H9 cells have been associated with the pathogenesis of IBD, particularly in UC. IL‐9⁺CD4⁺ T cells were enriched in the intestinal mucosa of patients with UC [40, 41], and mice with IL‐9‐deficiency were protected from intestinal inflammation in a mouse model of UC [40]. In the current study, peripheral T_H17 cells from patients with CD were also producing IL‐9, and TL1A significantly increased IL‐9 secretion of these cells. Recently, it has been demonstrated that TL1A enhances IL‐9 secretion of murine T_H9 cells but TL1A‐induced secretion of IL‐9 in other T_H subsets was not reported in that publication [42]. Our data suggest that TL1A exerts its proinflammatory effects in vivo on multiple levels by inducing not only IL‐17, but also IFN‐γ, IL‐9, and IL‐22 in memory T cells and IL‐22, IL‐9, and IFN‐γ in committed T_H17 cells. Although we did not observe an increase in IL‐17 by TL1A in T_H17 cells from healthy donors or patients with CD, the induction of IL‐9, IFN‐γ, and IL‐22 by TL1A may render these cells more pathogenic in vivo.

Our data provide important insights into the effector function of human T_H17 cells and have implications for the understanding and potential therapeutic manipulation of T_H1/T_H17‐associated diseases. Clinical trials using anti‐IL‐17 neutralizing antibodies in patients with CD have failed to show a beneficial effect [43, 44]. On the contrary, some patients receiving anti‐IL‐17 antibodies developed more severe inflammation. These patients were carriers of a TL1A (TNFSF15) risk polymorphism, and their adverse response to anti‐IL‐17 antibodies may have been related to higher mucosal TL1A expression, suggesting an important TL1A‐IL‐17 network in patients with CD [43].

Although IL‐17 and ‐22 have been initially described to be produced by T_H17 cells, both cytokines exert distinct in vivo functions. IL‐22 has been implicated in epithelial cell homeostasis, tissue repair, and wound healing, and a different regulation between IL‐17 and ‐22 has been suggested [35]. Our data confirm and extend a different regulation between IL‐17 and ‐22 production, depending on the T cell subtype (memory T cell and T_H17 cells) and the cytokines used for in vitro stimulation. Although TL1A alone induces IL‐22 secretion in memory T cells and T_H17 cells, TL1A requires the presence of TGF‐β, IL‐6, and IL‐23 for enhanced secretion of IL‐17 by memory T cells and is unable to enhance IL‐17 secretion by T_H17 cells. IL‐22 has been shown to be protective in mouse models of IBD by inducing antimicrobial peptides and promoting tissue repair [45]. Conversely, deregulated IL‐22 signaling contributes to the pathogenesis of IBD and colon cancer, and although IL‐22 exerts a protective effect, elevated levels have been seen in the mucosa of patients with IBD and colon cancer [46, 47–48]. IL‐22 and ‐9 have seemingly opposing direct effects on intestinal epithelial cells. IL‐9 directly impairs intestinal barrier function and prevents mucosal wound healing in mice by regulating the expression of tight junction proteins and epithelial cell proliferation [40]. However, in colorectal cancer, IL‐22⁺ cells promote cancer stemness and correlate negatively with patient outcome [47].

The expression of the TL1A receptor DR3 is high on activated T cells and on committed human T_H17 cells and also on ILCs [20, 48, 49]. Indeed, murine and human ILC3s respond to stimulation with IL‐23 and TL1A or IL‐23, IL‐1β, and TL1A with increased expression of IL‐22 [48, 50]. However, although we observed a direct effect of TL1A on IL‐22 secretion in memory CD4⁺ T cells and committed T_H17 cells, in ILCs, TL1A alone did not induce IL‐22 and needed IL‐23 for its induction, suggesting different signaling pathways in CD4⁺ T cells and ILCs [48, 50]. The lack of TCR signaling in ILCs may explain the different requirements for IL‐22 secretion by TL1A. A recent publication demonstrating IL‐22 secretion by TL1A requiring the costimulation of IL‐15 under TCR‐independent conditions supports this notion [51]. These published findings and our current data suggest that TL1A is a strong inducer of IL‐22 in T cells and ILCs and that IL‐22 derived from both cell populations contribute to the pathologies in IBD.

Prior publications have demonstrated that TL1A suppresses the differentiation of naïve CD4⁺ T cells into T_H17 cells by enhancing the production of IL‐2, which acts as a negative regulator of T_H17 differentiation [16, 20]. In contrast, TL1A enhanced IL‐17 expression in activated murine CD4⁺ T cells by mechanisms independent of its ability to enhance proliferation [20]. Our data confirm and extend these findings, demonstrating that TL1A induces IL‐17 expression in human effector/memory CD4⁺ T cells. Furthermore, it induced multiple T_H17‐associated cytokines, including IL‐22, IL‐9, and IFN‐γ in T_H17 cells isolated from healthy controls and from patients with CD. Collectively, these results support the view that TL1A’s activities depend on the activation status of the T cell population and that it differentially regulates different cytokines.

AUTHORSHIP

K.S.M., L.S.T., M.T., Q.T.Y., B.C.S., T.H., and E.M. performed experiments and analyzed data. K.S.M., S.R.T., and D.P.B.M. designed the experiments, and wrote the manuscript.

DISCLOSURES

K.S.M. and S.R.T. have a patent application (U.S. Application No. 61/856,491; International application No. PCT/US2014/047326) related to TL1A research. The authors declare no other financial conflicts of interests.

Supporting information

Supplementary Material Files