Article Text

Download PDFPDF

Paper
Protein tyrosine phosphatase N2 regulates TNFα-induced signalling and cytokine secretion in human intestinal epithelial cells
  1. Michael Scharl1,
  2. Declan F McCole2,
  3. Achim Weber3,
  4. Stephan R Vavricka1,
  5. Pascal Frei1,
  6. Silvia Kellermeier1,
  7. Theresa Pesch1,
  8. Michael Fried1,
  9. Gerhard Rogler1
  1. 1Division of Gastroenterology and Hepatology, Zurich Center for Integrative Human Physiology, University Hospital Zurich, Zurich, Switzerland
  2. 2Division of Gastroenterology, University of California, San Diego, School of Medicine, La Jolla, California, USA
  3. 3Department of Pathology, Institute of Surgical Pathology, University Hospital Zurich, Zurich, Switzerland
  1. Correspondence to Professor Dr Gerhard Rogler, Department of Internal Medicine, Clinic for Gastroenterology and Hepatology, University Hospital Zurich, Rämistrasse 100, 8091 Zurich, Switzerland; gerhard.rogler{at}usz.ch

Abstract

Objective The Crohn's disease (CD) susceptibility gene, protein tyrosine phosphatase N2 (PTPN2), regulates interferon γ (IFNγ)-induced signalling and epithelial barrier function in T84 intestinal epithelial cells (IECs). The aim of this study was to investigate whether PTPN2 is also regulated by tumour necrosis factor α (TNFα) and if PTPN2 controls TNFα-induced signalling and effects in IECs.

Methods T84 IECs were used for all cell studies. Protein levels were assessed by western blotting, mRNA levels by reverse transcription–PCR (RT–PCR) and cytokine levels by ELISA. PTPN2 knock-down was induced by small interfering RNA (siRNA). Imaging was performed by immunohistochemistry or immunofluorescence.

Results TNFα treatment elevated PTPN2 mRNA as well as nuclear and cytoplasmic protein levels and caused cytoplasmic accumulation of PTPN2. Biopsy specimens from patients with active CD showed strong immunohistochemical PTPN2 staining in the epithelium, whereas samples from patients with CD in remission featured PTPN2 levels similar to controls without inflammatory bowel disease (IBD). Though samples from patients with active ulcerative colitis (UC) revealed more PTPN2 protein than non-IBD patients and patients with UC in remission, their PTPN2 expression was lower than in active CD. Samples from patients with CD in remission and responding to anti-TNF treatment also showed PTPN2 levels that were similar to those in control patients. Pharmacological inhibition of nuclear factor-κB (NF-κB) by BMS-345541 prevented the TNFα-induced rise in PTPN2 protein, independent of apoptotic events. PTPN2 knock-down revealed that the phosphatase regulates TNFα-induced extracellular signal-regulated kinase 1/2 (ERK1/2) and p38 phosphorylation, without affecting c-Jun N-terminal kinase (JNK), inhibitor of κB (IκB) or NF-κB phosphorylation. Loss of PTPN2 potentiated TNFα-induced secretion of interleukin 6 (IL-6) and IL-8. In TNFα- and IFNγ-co-treated cells, loss of PTPN2 enhanced protein expression of inducible nitric oxide synthase (iNOS).

Conclusions TNFα induces PTPN2 expression in IECs. Loss of PTPN2 promotes TNFα-induced mitogen-activated protein kinase signalling and the induction of inflammatory mediators. These data indicate that PTPN2 activity could play a crucial role in the establishment of chronic inflammatory conditions in the intestine, such as CD.

  • PTPN2
  • TNFalpha
  • inflammatory bowel disease
  • Crohn's disease
  • cell biology
View Full Text

Statistics from Altmetric.com

Significance of this study

What is already known about this subject?

  • A mutation within the gene locus encoding PTPN2 is associated with IBD.

  • PTPN2 regulates IFNγ-induced STAT signalling in intestinal epithelial cells.

  • PTPN2 restricts the IFNγ-induced barrier defect in intestinal epithelial cells.

What are the new findings?

  • PTPN2 expression is increased by TNFα in intestinal epithelial cells and elevated in active CD.

  • PTPN2 regulates TNFα-induced MAPK signalling.

  • PTPN2 regulates TNFα-induced cytokine secretion.

How might it impact on clinical practice in the foreseeable future?

  • These data show that PTPN2 is activated by TNFα in intestinal epithelial cells and regulates TNFα-induced signalling and cytokine secretion. These findings suggest a possible mechanism to explain how dysfunction of PTPN2 could contribute to the onset of chronic intestinal inflammation, such as Crohn's disease.

Introduction

Crohn's disease (CD) and ulcerative colitis (UC) represent two subtypes of chronic intestinal inflammation and are collectively referred to as inflammatory bowel disease (IBD). The pathogenesis of IBD includes genetic, immunological and bacterial factors. There is evidence that a genetic predisposition favours a prominent role for an epithelial barrier defect in the onset of disease.1 As a result of increased access to the lamina propria, commensal bacteria cause an overwhelming activation of the immune system, which is—due to genetic mutations in regulatory genes—out of control and thus amplifies the inflammatory state.2

Intestinal epithelial cells (IECs) form a primary barrier against luminal antigens and toxins, and contribute to the regulation of essential intestinal immune responses, such as antigen presentation to immune cells or cytokine production. Especially in CD, an increased rate of IEC apoptosis can be observed that may exacerbate the barrier defect and increase the passage of antigens through the intestinal epithelium.3 4

During IBD, the levels of proinflammatory cytokines are highly elevated, whereas the secretion or activity of anti-inflammatory mediators is strongly impaired.5 Both UC and CD are characterised by their specific cytokine profile, and elevated levels of tumour necrosis factor α (TNFα) are well established as critical events in IBD, most prominently in CD.6

TNFα can cause proliferation, differentiation, inflammation or cell death depending on the specific cell type and the cellular signalling milieu.7 The key regulatory factor dictating the specific cellular responses to TNFα is the balance in the activation of TNFα-induced signalling pathways.8 TNFα binding to its ubiquitously expressed type 1 TNF receptor (TNFR1) can result in the recruitment of inhibitor of κB (IκB) kinase (IKK).9 Recruited IKK targets IκB for degradation, and the released transcription factor, nuclear factor-κB (NF-κB), can translocate into the nucleus to initiate gene transcription.7 On the other hand, TNFα signalling can also lead to the activation of mitogen-activated protein kinase (MAPK) isoforms p38, extracellular signal-regulated kinases 1 and 2 (ERK1/2) and c-Jun N-terminal kinase (JNK), resulting mainly in the activation of the transcription factor, activator protein 1 (AP-1).7 10 Among the TNFα-inducible genes are important proinflammatory mediators such as the interleukins IL-6 and IL-8, as well as the inducible nitric oxide synthase (iNOS). Additionally, TNFα can trigger the activation of certain caspases, especially caspase-3 and caspase-7, that are crucial for the induction of apoptosis.11 TNFα antibodies have been shown to reduce inflammation and IEC apoptosis in patients with active CD by in vivo administration12 13 and are a mainstay of treatment for refractory CD14 and UC.15

Important new avenues of investigation into the pathogenesis of IBD have been directed by genome-wide association studies that identified a number of disease-associated candidate genes. One of those genes, protein tyrosine phosphatase non-receptor type 2 (PTPN2), has recently been associated with CD and UC.16 17 PTPN2, also known as T cell protein tyrosine phosphatase (TC-PTP), exists in a nuclear 45 kDa and a cytoplasmic 48 kDa isoform. Following its activation, the catalytically more important nuclear isoform can exit the nucleus and accumulate in the cytoplasm.18 In various cell types, PTPN2 inactivates important immune mediators—that is, signal transducer and activator of transcription 1 (STAT1)19 and STAT320—as well as p38 and ERK1/2,21 and has also been implicated in the regulation of apoptosis in pancreatic β cells.22 Of special interest with respect to CD is that PTPN2 regulates intestinal epithelial barrier function, thus identifying a novel link between a disease-associated gene and a key pathophysiological event in CD.23

A role for PTPN2 in regulating TNFα-induced signalling and function in IECs has not yet been studied. We demonstrate that PTPN2 is activated by TNFα via NF-κB and regulates TNFα-induced MAPK signalling. From a functional perspective, loss of PTPN2 results in increased levels of proinflammatory mediators and correlates with elevated numbers of apoptotic events in IECs. These data identify PTPN2 as a key regulator of TNFα effects in IECs and further emphasise a potential role for PTPN2 in the pathogenesis of CD.

Materials and methods

Material

Human TNFα (Promokine, Heidelberg, Germany), human interferon γ (IFNγ; Promokine), BMS-345541 (Sigma, St. Louis, Missouri, USA), rabbit antilamin A/C (BD Biosciences, San Jose, California, USA), mouse anti-β-actin (Sigma), mouse anti-PTPN2 (Calbiochem, San Diego, California, USA) and rabbit anti-ERK1/2 antibodies (Santa Cruz, Santa Cruz, California, USA) were obtained from the sources noted. Rabbit antiphospho-ERK1/2 (Thr202/Tyr204), rabbit anti-p38, mouse antiphospho-p38 (Thr180/Tyr182), rabbit anti-JNK, rabbit anti-phospho-JNK (Thr183/Tyr185), rabbit anti-p65-NF-κB, rabbit antiphospho-p65-NFκB (Ser536), rabbit anti-IκBα, rabbit antiphospho-IκBα (Ser32) and rabbit anti-iNOS antibodies were obtained from Cell Signalling Technologies (Danvers, Massachusetts, USA). Rabbit anticaspase-3 antibody (Cell Signaling) detected both full-length and cleaved protein variants. All other reagents were of analytical grade and were acquired commercially.

Cell culture

Human colonic crypt T84 epithelial cells were cultured in a humidified atmosphere with 10% CO2 in 4.5% high glucose Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, California, USA) supplemented with 10% newborn calf serum. Cells were separated by trypsinisation and 1×106 cells were seeded onto either 12 mm or 30 mm Millicell-HA semi-permeable filter supports (Millipore, Bedford, Massachusetts, USA). Cells used for experiments were between passage 21 and 32. Before treatment, cells were cultured for 8–14 days, and transepithelial electrical resistance was measured to ensure intact monolayers. According to their receptor localisation, TNFα (100 ng/ml) and IFNγ (100 ng/ml) were added basolaterally, while BMS-345541 (25 μM) was added bilaterally.

Patient samples

Tissue specimens were prospectively collected from the colon of male and female non-IBD control patients who were asymptomatic and presented for colon cancer screening (n=11; average age: 56±4 years), from male and female patients with clinically and macroscopically active CD (n=16; average age: 45±3 years), from male and female patients with CD clinically and macroscopically in remission (n=10; average age: 42±4 years), from male and female patients with clinically and macroscopically active UC (n=11; average age: 37±5 years), from male and female patients with UC clinically and macroscopically in remission (n=10; average age: 42±4 years) and from male and female patients with clinically and macroscopically inactive CD (n=4) receiving anti-TNF treatment (average age: 56±4 years). Histological assessment of active or inactive disease was performed by a pathologist. Tissue samples were immediately transferred into 4% formalin and stored at 4°C until further analysis. Written informed consent was obtained before specimen collection, and studies were approved by the Cantonal Ethics Committee of the Canton of Zurich.

Histological PTPN2 score

Protein levels of PTPN2 in the investigated biopsy tissue specimens were analysed by a pathologist in a blinded fashion. The following scoring system was used for evaluating PTPN2 levels: 0=PTPN2 protein not detectable; 1=PTPN2 protein poorly detectable; 2=PTPN2 protein obviously detectable; 3=PTPN2 protein heavily expressed. All of the studied samples were evaluated.

RNA isolation and real-time PCR

RNA analysis, primer sequences and PCR conditions are described in the supplementary methods.

Preparation of whole-cell lysates

Whole-cell lysates were generated using M-Per Mammalian protein extraction reagent (Pierce Biotechnology, Rockford, Illinois, USA) according to the manufacturer's instructions and as described in the supplementary methods.

Preparation of cytoplasmic and nuclear lysates

Separate cytoplasmic and nuclear lysates were obtained using a Nuclear Extract Kit (Active Motif, San Diego, California, USA) according to the manufacturer's instructions and as described in the supplementary methods.

Western blotting

Experimental procedures are described in detail in the supplementary methods.

Immunofluorecence microscopy

A total of 5×105 T84 cells were seeded onto 12 mm Millicell-HA filters for 2 (small interfering RNA (siRNA)-transfected cells) to 6 days before stimulation. Experimental conditions and staining protocols are described in detail in the supplementary methods.

Immunhistochemistry

Immunhistochemical studies were performed on tissue specimens using a peroxidase-based method with diaminobenzidine (DAB) chromogen (DAKO, Glostrup, Denmark) as described previously.24 Experimental conditions and staining protocols are described in detail in the supplementary methods.

siRNA transfection

Experimental conditions and oligonucleotide sequences are described in the supplementary methods.

ELISA

Experimental procedures are described in the supplementary methods.

Statistical analysis

Data are presented as means±SEM for a series of n experiments. Data are expressed as a percentage of the respective control. Statistical analysis was performed by analysis of variance (ANOVA) followed by Student–Newman–Keuls post hoc test. p Values <0.05 were considered significant.

Results

TNFα increases PTPN2 expression in T84 IECs

To identify the maximally effective dose of TNFα capable of inducing PTPN2 protein expression in T84 IECs, we stimulated these cells with 1, 10 or 100 ng/ml TNFα for 24 h. As shown by representative western blots and densitometric analysis, 10 ng/ml TNFα significantly induced PTPN2 protein expression, and this effect was further enhanced by 100 ng/ml TNFα (figure 1A). Based on this observation, we continued with a concentration of 100 ng/ml TNFα in all of our subsequent experiments. We next studied whether TNFα could elevate PTPN2 mRNA levels. We treated T84 cells for 2, 8, 16, 24, 48 and 72 h, and TNFα increased PTPN2 mRNA levels significantly after 8 and 48 h incubations (figure 1B).

Figure 1

Tumour necrosis factor α (TNFα) increases expression levels and affects cellular localisation of protein tyrosine phosphatase N2 (PTPN2) in T84 cells. (A) PTPN2 protein in whole-cell lysates derived from TNFα-treated (24 h) T84 cells shown by representative western blots of PTPN2 and the loading control, β-actin. The relative protein level was assessed by densitometric analysis (n=4). (B) PTPN2 mRNA expression in response to TNFα, normalised to glyceraldehyde phosphate dehydrogenase (GAPDH). Four similar experiments were performed, each in triplicate. (C) Nuclear PTPN2 protein in TNFα-treated T84 cells shown by representative western blots. Lamin A/C was used as a loading control (n=4). (D) Cytoplasmic PTPN2 protein in TNFα-treated T84 cells is demonstrated by representative western blots and densitometric analysis (n=4). β-Actin was used as a loading control. Asterisks denote significant differences from the respective control (*p<0.05; **p<0.01; ***p<0.001). #p<0.05; ##p<0.01; ###p<0.001 vs 72 h TNFα treatment.

We then tested whether TNFα could affect PTPN2 protein expression at different time points. We collected separate cytoplasmic and nuclear protein lysates from T84 cells after TNFα treatment for 24, 48 and 72 h. In the nuclear compartment, TNFα increased PTPN2 protein by 24 h. Although PTPN2 protein was elevated after 48 h treatment, these levels were not statistically significantly above control values. After 72 h of TNFα stimulation, nuclear PTPN2 protein levels had reached controls levels again (figure 1C). In contrast, TNFα did not affect cytoplasmic PTPN2 protein by treatment for 24 and 48 h. However, after 72 h TNFα stimulation, PTPN2 protein was significantly elevated in the cytoplasmic compartment (figure 1D). These data demonstrate that TNFα induces PTPN2 mRNA as well as protein expression in T84 IECs.

TNFα causes cytoplasmic accumulation of PTPN2 in T84 IECs

The different pattern of PTPN2 protein expression in the different cell compartments in response to TNFα suggests that TNFα could affect the intracellular distribution of the phosphatase in T84 cells. Therefore, we treated the cells with TNFα for 72 h and performed immunofluorescence studies. PTPN2 was equally distributed in the cytoplasmic and the nuclear compartment of untreated control cells (figure 2A). While PTPN2 in the nuclei was still equally visible compared with control cells after 72 h TNFα treatment, the amount of PTPN2 in the cytoplasm had clearly increased (figure 2B). These findings correlate with our protein data and support the hypothesis that TNFα treatment not only increases PTPN2 protein levels, but also causes a cytoplasmic accumulation of PTPN2 by 72 h treatment.

Figure 2

Tumour necrosis factor α (TNFα) treatment causes cytoplasmic accumulation of protein tyrosine phosphatase N2 (PTPN2) in T84 cells. Immunofluorescence microscopy of T84 cells shows PTPN2 staining in red and nuclear staining in blue. Cells were (A) either left untreated or (B) incubated with TNFα (100 ng/ml) for 72 h. Each panel shows representative images from three similar experiments per condition. Green arrows point to PTPN2 in the nucleus, white arrows in the cytoplasm. Orange line indicates 10 μM. Magnification: ×40.

PTPN2 protein is elevated in active CD

To demonstrate the relevance of these findings to human disease, we performed immunohistochemical analysis of PTPN2 in human colonic biopsy specimens from non-IBD patients, from patients with clinically and macroscopically active CD or CD in remission, from patients with clinically and macroscopically active UC or UC in remission, as well as from patients with CD responding to anti-TNF treatment. In non-IBD patients, only low levels of PTPN2 protein were detectable (figure 3A). In contrast, the epithelial layer of tissue specimens from patients with active CD featured strong PTPN2 staining, indicating that the protein is strongly expressed in IECs from patients with active CD (figure 3B). Intestinal biopsy samples from patients with CD in remission revealed similar PTPN2 protein levels to those of control patients (figure 3C). Interestingly, while tissue specimens from patients with active UC showed more PTPN2 staining than those from non-IBD control patients, PTPN2 protein levels in these samples were clearly less than in active CD (figure 3D). Samples from patients with UC in remission revealed levels of PTPN2 protein similar to control patients or patients with CD in remission (figure 3E). Samples from patients with CD in remission receiving anti-TNF treatment revealed a staining intensity similar to samples from non-IBD one control patients or from CD as well as UC patients in remission (figure 3F). PTPN2 protein levels in the tissue samples were then scored by a pathologist in a blinded manner. With this approach we could fully confirm the visual observations, since PTPN2 levels in samples from patients with active disease reached a considerably higher score than in all of the other groups. The score that was obtained by analysing samples from CD patients in remission (with or without anti-TNF treatment) was comparable with those obtained in samples from patients with active UC or with UC in remission and slightly higher than in non-control patients (figure 3G). These data correlate with our previous PTPN2 mRNA findings in patient samples23 and underline the relevance of our experiments for human disease.

Figure 3

Protein tyrosine phosphatase N2 (PTPN2) protein is strongly expressed in intestinal biopsies from patients with active Crohn's disease (CD). Representative images of PTPN2 staining in intestinal biopsies derived from (A) non-inflammatory bowel disease (IBD) control patients (n=11), (B) patients with active CD (n=16), (C) patients with CD in remission (n=10), (D) patients with active ulcerative colitis (UC) (n=11), (E) patients with UC in remission (n=10) and (F) patients with CD in remission responding to antitumour necrosis factor α (TNFα) treatment (n=4) are demonstrated by immunohistochemistry. Black arrows point to nuclear and white arrows to cytoplasmic PTPN2 protein in intestinal epitelial cells. Magnification: ×10 or ×40 (enlargements), respectively. For A–E, tissue specimens from two different, representative patients are shown. (G) Histogram represents the statistical analysis of PTPN2 protein levels in the immunohistochemically stained tissue samples as scored by a blinded pathologist. Asterisks indicate significant differences compared with the score obtained by analysing the samples from non-IBD controls (***p<0.001).

TNFα increases PTPN2 protein via NF-κB activation

We next studied the mechanisms responsible for the effects of TNFα on PTPN2. To investigate whether TNFα could mediate its effects on the PTPN2 protein level via NF-κB, we stimulated T84 cells with TNFα and/or the highly selective pharmacological IKK inhibitor, BMS-345541 (25 μM), for 24 h. As shown by representative western blots and densitometric analysis, TNFα induced phosphorylation of IκBα (figure 4A) as well as of the p65 subunit of NF-κB (figure 4B), while BMS-345541 alone or in combination with TNFα suppressed IκBα (figure 4A) and p65-NF-κB phosphorylation levels even below control values (figure 4B). The same membranes as used in figure 4B were then stripped and re-probed for PTPN2. TNFα elevated PTPN2 protein by 24 h of treatment (figure 4C). However, co-treatment with BMS-345541 completely prevented the cytokine-induced rise in PTPN2 protein (figure 4C). Representative western blots of total IκBα (figure 4A) and of the loading control, β-actin (figure 4C), show that the overall protein level in cells (co-)treated with BMS-345541 was not altered. Since the amount of cleaved caspase-3, a marker of apoptotic cell death, was not altered under any conditions (figure 4D), these findings demonstrate that the massive reduction in PTPN2 and p65-NF-κB protein (figure 4B,C) was probably not due to cell death.

Figure 4

Tumour necrosis factor α (TNFα) induces protein tyrosine phosphatase N2 (PTPN2) protein via nuclear factor-κB (NF-κB) in T84 cells. (A) Representative western blots show phosphorylated inhibitor of κBα (IκBα) (Ser32) (above) and total IκBα (below) after TNFα (100 ng/ml) and/or BMS-345541 (25 μM) treatment for 24 h, followed by densitometric analysis of three similar experiments. (B) Phosphorylation of the p65 subunit of NF-κB (Ser536) (above) and protein level of total p65-NF-κB (below) after TNFα and/or BMS-345541 treatment (24 h) are shown by representative western blots of three similar experiments and followed by densitometric analysis. (C) Membranes used in B were stripped and re-probed. PTPN2 (above) and β-actin protein levels (below) in TNFα- and/or BMS-345541-treated cells (24 h) are shown by representative western blots, and densitometric analysis represents the protein level of the respective proteins (n=3). (D) Representative western blots of three similar experiments show protein levels of full-length and cleaved caspase-3 in TNFα- and/or BMS-345541-treated cells (24 h). Asterisks indicate significant differences compared with the respective control (*p<0.05; **p<0.01; ***p<0.001). #p<0.05; ###p<0.001 vs 24 h TNFα treatment.

PTPN2 regulates TNFα-induced MAPK, but not NF-κB activation

To study whether PTPN2 could modify TNFα-induced signalling in IECs, we performed PTPN2 knock-down studies using siRNA. We transfected T84 cells with either non-specific siRNA or PTPN2-specific siRNA constructs, and subsequently treated these cells with TNFα for 24 h. TNFα treatment elevated PTPN2 protein in T84 cells transfected with control siRNA while PTPN2-specific siRNA caused a maximal decrease in PTPN2 protein of ∼75.5±5.1% (figure 5A). A representative western blot of the loading control, β-actin, demonstrated that the transfection procedure itself did not affect the overall protein level in our cells (figure 5A). We then tested whether PTPN2 knock-down could affect the TNFα-induced activation of the three MAPK isoforms, namely ERK1/2, p38 and JNK, as assessed by their phosphorylation status. TNFα treatment increased the phosphorylation of ERK1/2 in control siRNA cells, and loss of PTPN2 potentiated this effect (figure 5B). Although TNFα alone was not sufficient to induce the phosphorylation of p38 in cells transfected with control siRNA, loss of PTPN2 permitted TNFα to induce phosphorylation of p38 (figure 5C). In contrast, neither TNFα treatment nor PTPN2 knock-down affected the phosphorylation of JNK (figure 5D). Interestingly, while TNFα induced phosphorylation of IκBα (figure 5E) as well as of p65-NF-κB (figure 5F), these effects were not altered by PTPN2 knock-down (figure 5E,F). These findings demonstrate that PTPN2 regulates the activation of the MAPK isoforms, ERK1/2 and p38, but not IκBα or p65-NF-κB in response to TNFα in IECs.

Figure 5

Protein tyrosine phosphatase N2 (PTPN2) knock-down enhances tumour necrosis factor α (TNFα)-stimulated extracellular signal-regulated kinase 1/2 (ERK1/2) and p38 phosphorylation. T84 cells were transfected with either non-specific small interfering RNA (siRNA) or PTPN2-specific siRNA and treated with TNFα (100 ng/ml) for 24 h. (A) Representative western blots for PTPN2 and β-actin followed by densitometric analysis of four similar experiments. (B) Representative western blots and densitometry showing phosphorylation (Thr202/Tyr204) and protein expression of ERK1/2, (C) phosphorylation (Thr180/Tyr182) and protein expression of p38, (D) phosphorylation (Thr183/Tyr185) and protein expression of c-Jun N-terminal kinase (JNK), (E) phosphorylation (Ser32) and protein expression of inhibitor of κBα (IκBα) and (F) phosphorylation (Ser536) and protein expression of the p65 subunit of nuclear factor-κB (NF-κB). Three similar experiments were performed for each protein. Asterisks indicate significant differences compared with the respective control (*p<0.05; **p<0.01; ***p<0.001). ##p<0.01 vs 24 h TNFα treatment of control siRNA-transfected cells.

Knock-down of PTPN2 potentiates the expression of proinflammatory mediators

We next investigated whether PTPN2 might affect the expression and/or secretion of proinflammatory mediators in response to TNFα treatment of T84 IECs. While TNFα induced the secretion of the pleiotropic cytokine IL-6 into the supernatant of T84 monolayers, this effect was further enhanced by PTPN2 knock-down (figure 6A). The concentration range of IL-6 in the cell supernatant was between 0.130 and 3.615 pg/ml. Similarly, loss of PTPN2 also potentiated the TNFα-induced secretion of the proinflammatory IL-8 in our studies (figure 6B). The concentration range was between 192 and 1,702 pg/ml. A further important mediator of inflammation is iNOS, and excessively elevated levels of iNOS have been found in the intestine of patients with IBD.25 As expected, TNFα treatment alone had no effect on iNOS protein (data not shown). Since cytokine co-treatment is required to stimulate iNOS expression,26 we co-treated either control siRNA- or PTPN2 siRNA-transfected cells with TNFα (100 ng/ml) and IFNγ (100 ng/ml) for 24 h. PTPN2-specific siRNA still caused a significant knock-down of the phosphatase under these conditions (data not shown). Although iNOS protein was not clearly elevated in PTPN2-competent cells that had been co-treated with the cytokine mix, loss of PTPN2 resulted in a significant increase in iNOS protein in response to TNFα and IFNγ co-treatment (figure 6C). These data demonstrate that PTPN2 restricts the TNFα-induced increase in the levels of proinflammatory mediators and therefore regulates inflammatory responses in IECs.

Figure 6

Protein tyrosine phosphatase N2 (PTPN2) regulates the level of inflammatory mediators in response to tumour necrosis factor α (TNFα) and/or interferon γ (IFNγ) in T84 cells. T84 cells were transfected with either non-specific small interfering RNA (siRNA) or PTPN2 siRNA. (A) ELISA analysis of supernatants from TNFα-treated T84 cells (100 ng/ml, 24 h) were analysed for interleukin 6 (IL-6). Five similar experiments were performed and the concentration range of IL-6 in the cell supernatant was between 0.130 and 3.615 pg/ml. Measurements were performed in duplicate. (B) IL-8 levels in the supernatant of TNFα-treated T84 cells (100 ng/ml, 24 h) were assessed by ELISA. Three similar experiments were performed and the concentration range of IL-8 in the cell supernatant was between 192 and 1702 pg/ml. Measurements were performed in duplicate. (C) T84 cells were transfected with either non-specific siRNA or PTPN2-specific siRNA and treated with IFNγ (100 ng/ml), TNFα (100 ng/ml) or a combination of both cytokines. Representative western blots and densitometric analysis of inducible nitric oxide synthase (iNOS) and β-actin after treatment with TNFα and IFNγ for 24 h. Three similar experiments were performed. Asterisks indicate significant differences compared with the respective control (*p<0.05; **p<0.01; ***p<0.001). #p<0.05; ##p<0.01; ###p<0.001 vs 24 h TNFα treatment of control siRNA-transfected cells.

Discussion

We have previously shown that PTPN2 is activated by IFNγ and regulates intestinal epithelial barrier function.23 Here, we demonstrate that PTPN2 is also upregulated by TNFα and controls TNFα-induced signalling, cytokine secretion and apoptosis. Both of these cytokines play important roles in the pathogenesis of IBD, and of CD in particular. An impaired epithelial barrier, dysregulated cytokine profiles and increased levels of IEC apoptosis are well described pathological findings in these diseases. Our studies revealed that all of these parameters are critically regulated by the activity of PTPN2. These observations are aligned with studies of homozygous PTPN2-deficient mice that feature elevated serum levels of IFNγ and TNFα as well as increased production of nitric oxide. These mice die within 5 weeks of birth, suffering from a progressive systemic inflammatory disease.27 28

With respect to TNFα induction of PTPN2 expression, we did observe that PTPN2 mRNA expression showed a biphasic increase in response to TNFα by 6 and 48 h. The first peak is likely to be responsible for the increase in nuclear PTPN2 protein after 24 h cytokine treatment, possibly causing dephosphorylation of nuclear signal transduction molecules. The latter peak may account for the increased cytoplasmic PTPN2 protein in response to 72 h TNFα. Since a certain amount of the nuclear PTPN2 has translocated into the cytoplasm the second expression peak could serve to maintain a certain basal level of PTPN2 in the nucleus.

Of special interest with respect to human disease was the finding that PTPN2 protein is strongly expressed in the intestinal epithelium of patients with active CD, while PTPN2 levels in tissue specimens from patients with CD in remission were clearly lower than in samples from patients with active disease and comparable with PTPN2 levels in non-IBD control patients. These data suggest that PTPN2 expression might be critically dependent on CD activity and correlate with the observation that patients with active CD have higher serum levels of TNFα than non-IBD patients,6 29 as well as with our in vitro data showing that PTPN2 expression can be induced by TNFα. As shown in figure 3D, intestinal biopsies from patients with active UC had PTPN2 protein levels that were slightly higher than in samples from non-IBD control patients, but lower than in samples from patients with active CD. This finding suggests that the upregulation of PTPN2 in CD might be somewhat disease specific, since none of the samples from patients with active UC showed PTPN2 levels that were comparable with those observed in patients with active CD. One possible explanation for this might be that patients with active CD present with higher serum TNFα levels than patients with active UC.29

An NF-κB-binding motif has been described within the PTPN2 promoter region.30 In our experiments NF-κB activity regulates PTPN2 protein expression in response to TNFα, while PTPN2 knock-down affected neither TNFα-induced IκBα nor p65-NF-κB activation. The latter finding was somewhat expected since NF-κB activation, and also IκBα activation, occurs after phosphorylation of selective serine residues, while PTPN2 only targets tyrosine residues. PTPN2 regulated the TNFα-induced phosphorylation of the MAPK isoforms ERK1/2 and p38, both of which are activated, at least in part, by tyrosine phosphorylation. These data are in good agreement with a previous study from van Vliet et al, showing that PTPN2 regulates TNFα-induced MAPK signalling in monkey and mouse fibroblasts.21 Taken together, these observations suggest that PTPN2 plays an important role in maintaining the balance of TNFα-induced signalling pathways. In particular, TNFα-mediated NF-κB activity appears to restrict TNFα-induced MAPK activity via the upregulation of PTPN2.

A large number of proinflammatory mediators are, at least in part, regulated via the activity of MAPKs. Therefore, we speculated that loss of PTPN2 could alter the secretion of IL-6 and IL-8, or the expression of iNOS, in T84 IECs in response to TNFα. Our data demonstrate that TNFα induces the secretion of both of these interleukins, and loss of PTPN2 potentiates this effect. IL-6 is critically involved in the switch from innate to adapted immune responses, since it regulates the recruitment of leucocytes to sites of inflammation as well as the pattern of infiltrating leucocytes into the acutely inflamed tissue.31 In colonic biopsies from patients with active IBD, IL-6 levels correlate with disease activity.32 Van Vliet et al demonstrated increased secretion of IL-6 in TNFα-treated, PTPN2-deficient mouse fibroblasts via elevated MAPK activity.21 These findings are in good agreement with our data in IECs, showing that PTPN2 knock-down results in increased p38 and ERK1/2 phosphorylation and, subsequently, in elevated IL-6 secretion in response to TNFα.

It is well established that TNFα induces the secretion of IL-8 from IECs. Our data demonstrated that PTPN2 knock-down further promotes the TNFα-induced secretion of IL-8. This finding seems plausible since loss of PTPN2 results in increased TNFα-mediated activation of ERK1/2 and p38. Both of these MAPKs have been implicated in the secretion of IL-8 from IECs in response to TNFα.33 Further, these data correlate with observations seen in intestinal biopsies from patients with IBD demonstrating elevated IL-8 levels in the intestinal mucosa.34

Excessive levels of nitric oxide as well as iNOS have been demonstrated in the inflamed mucosa of patients with IBD.25 While low levels of nitric oxide seem to exert a protective effect on the intestinal epithelium and intestinal barrier function, high levels are clearly associated with progressive cell damage and inflammation.35 These findings are consistent with our studies, showing that TNFα and IFNγ co-treatment caused a small, but noteworthy, increase in iNOS protein, correlating with a possible protective effect of moderate levels of iNOS during inflammatory conditions. However, loss of PTPN2 permits the cytokine mix to induce the expression of iNOS excessively, a finding that can also be observed in IBD and that is associated with more severe inflammation. These data, limited though they are by extrapolation of in vitro cell studies to a complex human disease, are at least consistent with our hypothesis that PTPN2 protects the intestinal epithelium and loss of PTPN2 results in cell damage and higher levels of inflammation.

In summary, we have demonstrated that PTPN2 is activated by TNFα and regulates TNFα-induced MAPK signalling. On a functional level, loss of PTPN2 is associated with increased expression and secretion of proinflammatory mediators. These proinflammatory events can be observed in the intestinal epithelium of patients with IBD. Moreover, a mutation in the PTPN2 gene has been associated with IBD, while we have recently shown that PTPN2 expression is elevated in active CD. We can speculate that if the IBD-associated PTPN2 mutation causes a dysfunction of the phosphatase in human colonic epithelial cells, then expression of a dysfunctional PTPN2 mutant could result in elevated activity of proinflammatory mediators and pathways in the intestinal epithelium. Therefore, our data support a preliminary hypothesis that dysfunction of PTPN2 could make an important contribution to the pathogenesis of IBD.

References

View Abstract

Footnotes

  • Funding This work was supported by an educational grant from Essex Chemie, Switzerland, to MS, a research grant from the University of Zurich to MS, a research grant from the European Crohn's and Colitis Organisation (ECCO) to MS, a grant from the Swiss National Science Foundation (grant no. 310030-120312) to GR, a research grant from UCB to GR, a Senior Research Award from the Crohn's and Colitis Foundation of America (CCFA) to DFM and by the Swiss IBD cohort (SIBDC).

  • Competing interests MS discloses grant support from Essex. GR discloses grant support from Abbot, Ardeypharm, Essex, FALK, Flamentera, Novartis, Tillots, UCB and Zeller.

  • Ethics approval This study was conducted with the approval of the Cantonal Ethics Committee of the Canton Zurich, Switzerland.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Digest
    Emad El-Omar Severine Vermeire Alexander Gerbes