Expansion of IL-23 receptor bearing TNFR2+ T cells is associated with molecular resistance to anti-TNF therapy in Crohn's disease

Abstract

Objective: Anti-tumour necrosis factor (TNF) antibodies are successfully used for treatment of Crohn's disease. Nevertheless, approximately 40% of patients display failure to anti-TNF therapy. Here, we characterised molecular mechanisms that are associated with endoscopic resistance to anti-TNF therapy.

Design: Mucosal and blood cells were isolated from patients with Crohn's disease prior and during anti-TNF therapy. Cytokine profiles, cell surface markers, signalling proteins and cell apoptosis were assessed by microarray, immunohistochemistry, qPCR, ELISA, whole organ cultures and FACS.

Results: Responders to anti-TNF therapy displayed a significantly higher expression of TNF receptor 2 (TNFR2) but not IL23R on T cells than non-responders prior to anti-TNF therapy. During anti-TNF therapy, there was a significant upregulation of mucosal IL-23p19, IL23R and IL-17A in anti-TNF non-responders but not in responders. Apoptosis-resistant TNFR2+IL23R+ T cells were significantly expanded in anti-TNF non-responders compared with responders, expressed the gut tropic integrins α4β7, and exhibited increased expression of IFN-γ, T-bet, IL-17A and RORγt compared with TNFR2+IL23R- cells, indicating a mixed Th1/Th17-like phenotype. Intestinal TNFR2+IL23R+ T cells were activated by IL-23 derived from CD14+ macrophages, which were significantly more present in non-responders prior to anti-TNF treatment. Administration of IL-23 to anti-TNF-treated mucosal organ cultures led to the expansion of CD4+IL23R+TNFR2+ lymphocytes. Functional studies demonstrated that anti-TNF-induced apoptosis in mucosal T cells is abrogated by IL-23.

Conclusions: Expansion of apoptosis-resistant intestinal TNFR2+IL23R+ T cells is associated with resistance to anti-TNF therapy in Crohn's disease. These findings identify IL-23 as a suitable molecular target in patients with Crohn's disease refractory to anti-TNF therapy.

Keywords: apoptosis; crohn’s disease; intestinal T cells; mucosal immunology; tnf.

Figure 1

Mucosal expression level of several cytokines, cytokine receptors and cytokine-inducing transcription factors in patients with Crohn’s disease (CD) prior to the initiation of anti-tumour necrosis factor (TNF) therapy. (A) Intestinal biopsies of the most inflamed area were taken from anti-TNF-naive patients with CD prior to initiation of anti-TNF therapy and qPCR was performed. Patients were stratified as responders (n=8) or non-responders (n=9–11) according to their subsequent clinical (Crohn’s Disease Activity Index reduction of >100 points at week 12) and endoscopic response (Simple Endoscopic Score for Crohn’s Disease <5 on follow-up endoscopy) to anti-TNF therapy with adalimumab. Data were normalised to the housekeeping gene Hprt and compared with a control group without mucosal inflammation. Every symbol represents a single patient. Median value is indicated as a line. **P≤0.01. Responders and non-responders did not show any significant differences in histological inflammatory activity prior to initiation of anti-TNF therapy. (B) Representative immunofluorescence staining for TNFR2, CD3 and isotype controls on intestinal cryosections from anti-TNF responders (n=8) and non-responders (n=7) before initiation of anti-TNF therapy. Scale bars represent 10 µm. Statistical analyses of the staining. 4–6 high power fields (HPFs) of each sample were analysed. Data represent mean values±SEM. *P≤0.05.

Figure 2

Heatmap of genes showing microarray based differential expression of IBD susceptibility and IL23R associated pathway genes during ongoing anti- TNF therapy (A) Gene expression profiles of isolated LPMCs from responders (n=3) and non-responders (n=3) to anti-TNF therapy in CD patients. Shown as a heat map are expression patterns of 40 IBD susceptibility genes and some genes which are additionally involved in IL23R signaling (in red) from microarray data, with the corresponding hierarchical clustering of the experimental conditions (responder or non-responder to anti-TNF therapy in CD patients). Results were corrected for multiple testing with the Benjamini-Hochberg correction. Pearson and Ward were used to compute hierarchical clustering of the experimental conditions. Expression data for each gene/row were normalized to the median expression value of the respective gene across all six samples. (B) Ingenuity Pathway Analysis of CD anti-TNF responders and non-responders during ongoing anti-TNF therapy showing the role of cytokines mediating communication between immune cells. The intensity of the color indicates the level of upregulation (red) or downregulation (green) of the respective cytokines in anti-TNF non-responders compared to anti-TNF responders.

Figure 3

Mucosal expression level of several cytokines, cytokine receptors and cytokine-inducing transcription factors in Crohn’s disease patients during ongoing anti-TNF therapy Mucosal biopsies were taken from the same CD patient prior and during initiated anti- TNF. Response to anti-TNF therapy was defined as a SES-CD <5 during the endoscopic examination where the biopsy was taken under ongoing anti-TNF treatment. Quantitative RT-PCR analysis of the intestinal biopsies from responders (n=7-16) and non-responders (n=6-11) was performed. Data were normalized to the housekeeping-gene Hprt and compared to the respective expression level prior to the initiation of anti-TNF therapy in each single patient. Gene expression is shown as fold induction. Every symbol represents a single patient. Median value is indicated as a line *P≤0.05.; **P≤0.01, ***P≤0.001.

Figure 4

Molecular characterisation of mucosal TNFR2+IL23R+ cells in patients with Crohn’s disease (CD). (A) Flow cytometric analysis of TNFR2 and IL23R cell expression in lamina propria mononuclear cells from intestinal tissue of control patients (CON; n=7) and patients with CD with active mucosal inflammation (CD; n=9–14). (B) Quantitative expression analysis of IL23R, IFN-γ and IL-17A in TNFR2+ and TNFR2− intestinal cells from patients with CD (n=22–33). Isolated mucosal cells were analysed by flow cytometry using a lymphocyte gate. (C) Flow cytometric expression analysis of intestinal cells from patients with CD, which express IFN-γ+IL17-A+ or Tbet+RORγt+ cells within T cells that showed expression of either TNFR2, IL23R, both or none of these receptors (n=9–10). Representative FACS plots of the analyses of the above-described expression profiles of intestinal cells from patients with CD are shown. Data represent mean values±SEM. *p≤0.05; **p≤0.01; ***p≤0.001.

Figure 5

Expression of TNFR2, IL23R, pSTAT3 and Ki-67 in anti-tumour necrosis factor (TNF)-treated patients with Crohn’s disease (CD). (A) Immunofluorescence staining for TNFR2, pSTAT3, Ki-67 and CD3 was performed with cryosections from intestinal biopsies of responders (n=8) and non-responders (n=10–12) to anti-TNF therapy in patients with CD. Scale bars represent 7.5 µm. Positive cells were counted in 4–6 high power fields (HPFs) per slide. Statistical analyses of double-positive cells are shown. (B) Flow cytometric analysis regarding the expression of mucosal TNFR2+IL23R+CD4+ lymphocytes in the intestinal tissue of anti-TNF-naive patients with CD with mucosal inflammation (n=14), as well as responders (n=13) and non-responders (n=16) to anti-TNF therapy patients (upper left panel). Frequencies of TNFR2+IL23R+ T cells before and during anti-TNF treatment in the same patient for anti-TNF responders (n=3) and non-responders (lower left panel). Representative FACS plot images are shown for anti-TNF responders, non-responders and naive patients and the respective isotype control. CD4+ cells were gated and the rate of TNFR2+IL23R+ cells assessed (right panel). Data represent mean values±SEM. *p≤0.05; **p≤0.01.

Figure 6

Significant upregulation of gut-tropic α₄+β₇+CD4+IL23R+ T cells on IL-23 application. (A) CD4+ lamina propria mononuclear cells from patients with Crohn’s disease (CD) (n=10) were isolated with magnetic beads and left untreated or stimulated for 72 hours with TGF-β, IL-6 or IL-23 or a combination of all three cytokines. Expression of the IL23R on CD4+ cells was determined by FACS analysis. (B) CD4+ peripheral blood mononuclear cells from healthy controls (n=6) or patients with CD (n=13) were isolated with magnetic beads and left untreated or stimulated for 72 hours with TGF-β, IL23, a combination of both cytokines, IL12, IL17, IL21 or TNF. Expression of the IL23R on CD4+ cells among α₄−β₇− or α₄+β₇+ cells was determined by FACS analysis. (C) Whole intestinal biopsies from patients with CD were cultivated for 24 hours in an organ culture chamber at 37°C with 95% O₂/5% CO₂ atmosphere. Biopsies were untreated (n=10) or stimulated with 20 ng/mL IL-23 (n=10). Expression of the IL23R on CD4+ cells among α₄−β₇− or α₄+β₇+ cells was determined via flow cytometry. (D) Whole intestinal biopsies from patients with CD were cultivated for 24 hours in an organ culture chamber at 37°C with 95% O₂/5% CO₂ atmosphere. Biopsies were untreated, stimulated with 25 µg/mL anti-TNF antibody or 25 µg/mL anti-TNF antibody plus 20 ng/mL IL-23. Expression of CD4, IL23R and TNFR2 was determined via flow cytometry. Data represent mean values±SEM. *p≤0.05; **p≤0.01; ***p≤0.001.

Figure 7

IL-23 production of blood and intestinal cells of patients with Crohn’s disease (CD). (A) Peripheral blood mononuclear cells from patients with CD (n=16) were untreated or stimulated overnight with LPS, IFN-γ and SAC. IL-23p19 levels were assessed in the supernatant of the cultivated cells. (B) IL-23p19 concentration in the supernatant of cultured lamina propria mononuclear cells (LPMCs) of controls (n=10) and patients with CD (n=7). (C) Quantitative analyses of IL-23p19 expressing CD14+ LPMCs of controls (n=8), anti-tumour necrosis factor (TNF)-naive patients (n=6), anti-TNF responders (n=8) or anti-TNF non-responders (n=10) in patients with CD (upper left panel). Quantitative analyses of IL-23p19 expressing CD16+, CD11c+ and CD15+ LPMCs of controls (n=8), anti-TNF responders (n=8) or anti-TNF non-responders (n=10) in patients with CD (lower left panel). Representative FACS analysis images of intestinal CD14+ macrophages from controls, responders and non-responders that express IL-23p19 (right panel). (D) Representative immunofluorescence staining for CD14 and isotype controls on intestinal paraffin-embedded sections from CD anti-TNF responders (n=5) and non-responders (n=7) before initiation of anti-TNF therapy. 3–6 high power fields (HPFs) of each sample were analysed using ×40 objective magnification. Scale bars represent 5 µm. Statistical analyses of the staining data represent mean values±SEM. *p≤0.05; **p≤0.01; ***p≤0.001.

Figure 8

Influence of anti-tumour necrosis factor (TNF) and IL-23 treatment on the induction of apoptosis in blood and mucosal cells. (A) CD4+ lamina propria mononuclear cells (LPMCs) from patients with CD (n=15) were isolated, stimulated for 72 hours with anti-CD3/28 antibodies and treated with an anti-TNF or anti-IL23p19 antibody or an anti-TNF antibody in combination with recombinant IL-23. Cell apoptosis was analysed via flow cytometry analysis and CD4+ Annexin V-positive/propidium iodide-negative cells were used for analysis of apoptosis cells. Representative FACS images of treated LPMCs are shown. (B) CD4+ blood cells from patients with CD (n=15) were untreated or treated with IL-23 for 24 hours. Cell apoptosis was determined using TUNEL staining. TUNEL-positive cells per 100 cells were quantified and 5 high power fields (HPFs) per sample were analysed. Scale bars represent 10 µm. (C) CD4+ peripheral blood mononuclear cells from patients with CD were isolated (n=6) and the cells were untreated or treated with IL-23 in various concentrations (10–50 ng/mL) for 24 hours. Cell apoptosis was analysed using the cell death ELISA detection kit. (D) Cryosections of intestinal biopsies from patients with CD under ongoing anti-TNF therapy were stained for CD4 and active Caspase and the double-positive cells among CD4+ T cells were determined. Patients were stratified as responders (n=7) or non-responders (n=14) according to their Simple Endoscopic Score for Crohn’s Disease during ongoing anti-TNF therapy. 4–5 HPFs per slide were analysed. Scale bars represent 10 µm. Data represent mean values±SEM. *p≤0.05; **p≤0.01; ***p≤0.001.

Figure 9

Model of IL-23-mediated resistance to apoptosis of mucosal CD4+ T cells in anti-tumour necrosis factor (TNF) refractory Crohn’s disease (CD). In responders to therapy, anti-TNF antibodies bind to mTNF-expressing CD14+ macrophages, thereby inhibiting activation of the TNFR2 on mucosal CD4+ T cells. The resulting blockade of TNFR2-dependent signalling pathways finally leads to the indirect induction of apoptosis in intestinal CD4+ T cells in CD. In anti-TNF refractory patients, TNFR2 bearing gut CD4+ T cells additionally express the IL23R. Heightened production of IL-23 from CD14+ macrophages leads to binding to the IL23R on CD4+TNFR2+ T cells and induction of STAT3 activation, thereby augmenting T-cell resistance to apoptosis that overcomes the induction of apoptosis by anti-TNF antibodies.