| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Suppresses IFN-
-STAT1-Dependent Gene Transcription by Enhancing STAT1-PIAS1 Interactions in Epithelia but Not Monocytes/Macrophages1Gastrointestinal Research Group, University of Calgary, Calgary, Alberta, Canada
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
and TGF-
are important regulators of mucosal immunity, typically functioning in opposition to each other. In this study, we assessed whether TGF- could modulate IFN--induced STAT1 signaling. Model epithelial cell lines (HEp-2, HT-29, and T84) or monocytes/macrophages (THP-1 cell line, human blood mononuclear cells) were pretreated with TGF- (1 ng/ml; 5–60 min), followed by IFN- exposure (20 ng/ml; 30 min), and then STAT1 transcriptional activity, DNA-binding activity, phosphorylation, and methylation were assessed. Some epithelia were transfected with an expression plasmid encoding SMAD7 to block TGF--SMAD signaling. Epithelia, but not macrophages, pretreated with TGF- were hyporesponsive to IFN- stimulation as indicated by reduced expression of four STAT1-regulated genes and reduced STAT1 DNA binding on EMSA. However, STAT1 Tyr701-, Ser727 phosphorylation, and nuclear recruitment of STAT1 were not significantly different in IFN- with or without TGF--treated cells, indicating that the effects of TGF- are downstream of IFN-R-JAK-STAT1 interaction. The TGF- effect was not dependent on ERK1/2, p38, or JNK activation but was prevented by overexpression of the inhibitory SMAD7 protein. Additional studies suggest that TGF- blockade of IFN- activity in epithelia is via enhanced sequestering of STAT1 by pre-existing protein inhibitor of activated STAT1. These results demonstrate that TGF- rapidly suppresses IFN--driven STAT1 signaling by reducing DNA binding via promotion of STAT1-protein inhibitor of activated STAT1 interactions and not inhibition of STAT1 activation; an event that may be specific to epithelia and represent a novel mode of action of TGF-. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
IFN- and TGF-1 are key cytokines that often function in opposition to modulate mucosal immunity and inflammation (3). IFN- is a proinflammatory cytokine that is produced primarily by CD4+ T cells (particularly Th 1 cells) and NK cells. Exposure to IFN- can result in the activation of up to 500 genes (4) and, in the case of the enteric epithelium, causes reduced epithelial barrier function, reduced active ion transport, and increased chemokine synthesis (5, 6, 7, 8). In contrast, TGF- is an immunoregulatory cytokine produced by many cell types and, despite its ability to cause fibrosis, is generally considered beneficial by virtue of its anti-inflammatory and immunosuppressive properties (9). For instance, loss of TGF- signaling by expression of a dominant-negative receptor exacerbates colitis, and conversely enhancement of TGF- production significantly ameliorates murine colitis (10, 11). In terms of direct action on the epithelium, TGF- enhances enteric epithelial barrier function, and the in vivo significance of this would be to control the movement of potentially antigenic material from the lumen into the mucosa/submucosa and access to the immune cells therein (7). Moreover, TGF- blocks the loss of barrier function observed in monolayers of human colon-derived epithelial cell lines caused by exposure to bacterial pathogens and immune mediators, such as IFN- (12, 13).
Under normal conditions and during active immune responses or disease, multiple cytokines will be present in the interstitial milieu. Understanding the interplay between these signals on the various target cell types is crucial to fully appreciate cytokine regulation of mucosal immunity and gut function. IFN- antagonism of TGF--induced intracellular signaling is established and, in general, appears to be an IFN--STAT1-dependent event (14): significantly less is known of how TGF- might ablate IFN- signaling. For example, TGF- inhibition of IFN--induced increases in epithelial paracellular permeability (12) could be due to separate and opposing affects on the expression of tight junction proteins that gate the paracellular permeation pathway between adjacent epithelial cells or via direct cross-regulation of intracellular signaling pathways. In support of the latter hypothesis, TGF- treatment reduces NF-
B-dependent IL-8 expression in enteric epithelium (15) and prevents IL-6-induced JAK/STAT3 phosphorylation through increased suppressor of cytokine signaling 1/3 expression in airway epithelium (16).
In this study, we assessed if, and how, TGF- inhibits IFN--STAT1 gene transcription in model epithelia and monocytes/macrophages. The data illustrate that a short duration, low-dose TGF- exposure suppresses IFN--driven STAT1-dependent gene expression and STAT1 DNA binding in epithelia but not in macrophages. TGF- blockade of STAT1 activity in epithelia did not rely on inhibition of Tyr710 or Ser727 phosphorylation or STAT1 methylation; rather, it occurred through enhanced sequestering of STAT1 by existing protein inhibitor of activated STAT1 (PIAS1),3 which may represent a novel mode of action for TGF- regulation of IFN--STAT1 signaling in epithelial cells.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Epithelial cells. Three human-derived epithelial cell lines that are widely used in the analysis of the effects of enteric pathogens and cytokines on epithelial function were used, namely HEp-2 (laryngeal origin), HT-29, and T84 (both of colonic origin) cells (American Type Culture Collection) (7, 17, 18, 19, 20, 21). Each cell line was maintained at 37°C with 5% CO2 in a specific culture medium (Hep2 cells: MEM F11 supplemented with 2% sodium bicarbonate, 2.5% (v/v) penicillin-streptomycin, and 10% FBS; HT-29 cells: DMEM plus 1% penicillin-streptomycin (v/v), 0.5% sodium bicarbonate (v/v), 0.1% L-glutamine, and 5% FBS; T84 cells: a 1/1 mixture of DMEM and Ham’s F-12 medium supplemented with 2% penicillin-streptomycin, 1.5% (v/v) HEPES, and 10% FBS (all from Invitrogen Life Technologies) (18, 19)).
Immune cells. The human THP-1 monocyte cell line (from American Type Culture Collection) was maintained as a suspension in RPMI 1640 culture medium supplemented with 2% (v/v) penicillin-streptomycin, 36 µM HEPES, and 10% FBS. Two million cells were seeded onto 3-cm petri dishes and induced into a macrophage phenotype by adding PMA (10 nM; Sigma-Aldrich). Three days later, cells were serum starved for 16 h, followed by cytokine treatment (22). PBMC from healthy volunteers were isolated as described previously (7). Briefly, venous blood was collected in heparinized vaccu-tubes, diluted 1/2 with sterile PBS (37°C) in a sterile 50-ml tube, and underlain with 10 ml of Ficoll plaque (Amersham Biosciences). Following centrifugation, mononuclear cells were retrieved from the Ficoll-PBS interface, rinsed twice in PBS, and resuspended in RPMI 1640 culture medium at 2 x 106/ml.
Cytokine treatment.
Epithelia cells were seeded at 106 cells/ml into sterile petri dishes or 6-well culture plates and grown to 
70% confluence (determined by phase contrast microscopy), and the culture medium was replaced with serum-free medium for 16 h, followed by IFN- (20 ng/ml) with or without TGF-1 (1 ng/ml, as a 5- to 60-min pretreatment) exposure (R&D Systems). We have previously shown that these doses of cytokine directly affect epithelial function (12, 22).
RT-PCR
IFN--STAT1-dependent gene expression was determined through semiquantitative measurements of IFN--regulated factor-1 (IRF-1), major histocompatibility CIITA, guanylate-binding protein 1 (GBP-1), and inducible NO synthase (iNOS) mRNA (22). Cells were pretreated with TGF-, exposed to IFN- for 30 min, and rinsed in serum-free medium (three times), and RNA was extracted 3.5 h later (i.e., 3 h following IFN-) using RNeasy columns (Qiagen). RNA yield and concentration were determined spectrophotometrically, and 1 µg of RNA was used to make cDNA by reverse transcription using an iSCRIPT kit (Bio-Rad). PCR was conducted on the resulting cDNA with primers designed using Primer3 software (www.genome.wi.mit.edu/cgi-bin/primer/primer3_www.cgi/) from published mRNA sequences (GenBank): IRF-1 forward, 5'-CGA TAC AAA GCA GGG GAA AA-3' (10 pmol), and IRF-1 reverse, 5'-TAG CTG CTG TGG TCA TCA GG-3' (10 pmol); GBP-1 forward, 5'-TGG AAC GTG TGA AAG CTG AG-3' (3 pmol), and GBP-1 reverse, 5'-TGA CAG GAA GGC TCT GGT CT-3' (3 pmol); CIITA forward, 5'-GGG AAA GCT TGT GCA GAC TC-3' (3 pmol), and CIITA reverse, 5'-CAC CCA GGT CAG TGA TGT TG-3' (3 pmol); iNOS forward, 5'-TGT GCT CTT TGC CTG TAT GC-3' (3 pmol), and iNOS reverse, 5'-GGG GAT CTG AAT GTG CTG TT-3' (3 pmol); and -actin forward, 5'-CCA CAG CAA GAG AGG TAT CC-3' (3 pmol), and -actin reverse, 5'-CTG TGG TGG TGA AGC TGT AG-3' (3 pmol). Amplification by PCR was conducted using platinum Taq (Invitrogen Life Technologies) and the following parameters: 95°C for 3 min, followed by 25 cycles of 1 min at 95°C, 60°C annealing for 1 min, 72°C for 1 min, with a final elongation step conducted at 72°C for 10 min. Amplification products were subjected to electrophoresis in a 1% agarose gel in Tris-acetic acid EDTA buffer, and imaged on a Kodak EDAS 290 gel documentation system (Kodak).
EMSA for STAT1
Following cytokine treatment, nuclear protein extracts were obtained, and protein concentrations were determined using the Bradford assay. EMSA were performed as previously described (22, 23): nuclear extracts (5–10 µg protein) in binding buffer were incubated for 30 min with [32P]dCTP (NEN Life Science Products)-labeled oligonucleotide probe high-affinity sis-inducible element (hSIE) containing a high-affinity STAT1 binding site (5'-GTCGACATTTCCCGTAAATC-3' and 5'-TCGACGATTTACGGGAAATG-3' (MOBIX; McMaster University) (24). Samples were electrophoresed through a nondenaturing 6% (40:1 bis/acrylamide) polyacrylamide gel for 2.5 h at 120 V, dried under vacuum at 80°C, and visualized by autoradiography after overnight exposure (–70°C) to Kodak BioMAX MS film. Specificity controls included use of a "cold" unlabeled hSIE dsDNA oligonucleotide for competitive binding and inclusion of STAT1-specific Abs (Santa Cruz Biotechnology) in the reaction mixture.
Immunoblotting and immunoprecipitation
Immunoblotting was used following a standard protocol (12, 22) to assay for changes in whole-cell protein extracts after cytokine treatment using the following Abs: anti-TGF- receptor I, anti-P-Tyr701-STAT1, anti-P-Ser727-STAT1, anti-c-myc, anti-P-ERK1/2, anti-P-Ser63-JNK, anti-P-Ser73-JNK, anti-JNK, anti-p38, anti-phospho-p38 (Cell Signaling Technology), anti-phosphoserine (Abcam), anti-SMAD2/3, anti--actin, anti-STAT1, anti-IRF1, anti-ERK1/2, anti-PIAS1 (Santa Cruz Biotechnology), and anti-FLAG (Sigma-Aldrich) at working dilutions of 1/1000. Secondary Abs were goat anti-rabbit-HRP IgG or goat anti-mouse-HRP IgG (Santa Cruz Biotechnology) and were used at a working dilution of 1/5000.
Immunoprecipitations were conducted with EZview beads (Sigma-Aldrich), according to the manufacturer’s instructions. Briefly, cell lysates were precleared with 40 µl of the EZview bead slurry (1 h, 4°C). Cell lysates were diluted with radioimmunoprecipitation assay (RIPA) buffer to 1 mg/ml and were gently rocked with Ab (2 µg/ml, 2 h, 4°C), and then added to 40 µl of prewashed EZview beads while rocking (1.5 h, 4°C). The beads were centrifuged and washed four times at 5 min to remove unbound proteins. Following the final wash, dissociation from the EZview beads was performed by adding SDS-loading buffer (laemmli) and boiling for 5 min.
In additional experiments, we conducted in vitro binding assays. Protein extracts (0.5 mg/ml) from HEp-2 epithelial cells with or without TGF- treatments (1 ng/ml at 15, 30, and 60 min) were subjected to immunoprecipitation with anti-PIAS1 Ab or an isotype-matched irrelevant Ab (negative control). The captured PIAS1 from control or TGF--treated cells was incubated with 0.5 mg of protein obtained from STAT1-FLAG-transfected cells (see below) that had or had not (i.e., control) been exposed to IFN- (20 ng/ml; 30 min) overnight at 4°C. Following several washes in RIPA buffer, protein complexes were dissociated by boiling in SDS-loading buffer for 5 min, followed by immunoblotting with PIAS1- or FLAG-specific Abs.
Construction of a SMAD7 expression vector
The viral genome was isolated from an adenovirus encoding mouse SMAD7, provided by Dr. P. ten Dijke (Ludwig Institute for Cancer Research, Uppsala, Sweden) (25), through a modified small DNA isolation technique (Hirt protocol) as described previously (26). Briefly, subconfluent monolayers (60–80%) of HEK293 cells were infected at 5 multiplicity of infection for 1 h with gentle agitation. Cytopathic effects were observed 24 h postinfection, at which point cells were dislodged with a cell scrapper and gentle tapping of the culture flask. Detached cells were incubated for 16 h on a plate rocker at 37°C in DNA extraction buffer (50 mM Tris, 2 mM EDTA, 0.5% (w/v) polyoxyethylene sorbitan (Tween 20; Sigma- Aldrich), and 400 µg/ml proteinase K (Roche Diagnostics, Laval, PQ: added immediately before use). DNA was isolated by a 25:24:1 phenol:chloroform:isoamyl alcohol (v/v/v) extraction and precipitated with isopropanol (–20°C overnight). Once precipitated, the DNA was collected by centrifugation (13,000 rpm, 30 min), washed once with 70% ethanol, air-dried, and resuspended in ultrapure water (Invitrogen Life Technologies).
Endonuclease restriction sites for KpnI and BamHI were added to the 5' and 3' end of SMAD7, facilitating insertion into the pcDNA3.1 (Invitrogen Life Technologies) expression vector by PCR. Amplification was conducted using the SMAD7-specific primers: mSMAD7-KpnI, 5'-GGG GTA CCA TGT TCA GGA CCA AAC GAT CTG-3', and mSMAD7-BamHI, 5'-CGG AAT TCC TAC CGG CTG TTG AAG ATG AC-3', with hi-fidelity platinum Taq (Invitrogen Life Technologies) under the following conditions: 95°C 2 min, 30 cycles of 95°C 15 s, 64°C 30 s, 60°C 2 min, and a final 68°C extension for 10 min. Following PCR amplification, vector and SMAD7 insert were sequentially double digested with KpnI, followed by BamHI with 10-fold excess enzyme. Linearlized pcDNA3.1 was treated with shrimp alkaline phosphatase (Fermentas Life Sciences), purified by agarose gel electrophoresis (0.7%), and extracted using the QIAEX II kit (Qiagen). Recovered SMAD7 and lineralized pcDNA3.1 were then ligated overnight at a molar ratio of 3:1, respectively, with T4 DNA ligase (Invitrogen Life Technologies).
Competent DH5
Escherichia coli (Invitrogen Life Technologies) were transformed according to the manufacturer’s instructions and plated on Luria-Bertani-ampicillin (100 µg/ml) agar plates, and 24 h later, ampicillin resistant colonies were isolated for screening by restriction enzyme and PCR analysis (using primers directed against the T7 and BGH sequences flanking the multiple cloning site on the vector). Successfully transformed bacteria were further cultured, and plasmid DNA was purified with a QIA filter Maxiprep kit (Qiagen).
Transient transfection of epithelia
Subconfluent HEp-2 monolayers (60%) were transfected using GenePorter II (GPII) according to the manufacture’s recommendations (Gene Therapy Systems). Briefly, plasmid DNA was incubated in DNA dilution solution for 
5 min and added to the GPII transfection reagent diluted in serum and antibiotic free HEp-2 culture medium. The SMAD7 vector described above was used at a final concentration 4 µg/ml, whereas the SMAD2-myc, SMAD3-myc (provided by Dr. O. Eickelberg, University of Giessen, Giessen, Germany) (27), and STAT1-FLAG (provided by Dr. M. David, University of California, San Diego, CA) (28) plasmids were used at a final concentration of 2 µg/ml. Monolayers were rinsed in serum- and antibiotic-free medium (three times), the DNA transfection reagent added to the cells, and 24 h later, the solution was aspirated and replaced with fresh normal HEp-2 medium (transfection conditions were determined empirically by monitoring protein expression of the plasmid encoded gene over a 48-h period).
Nucleofection of small interfering RNA (siRNA)
Nucleofection was performed according to manufacture’s instructions (Amaxa; EBSE Scientific). Briefly, cells were grown to 70% confluence and dissociated with trypsin EDTA to obtain a single-cell suspension. Cells (1 x 106) were placed into a cuvette with nucleofactor solution V, with an experimentally determined amount of prevalidated "stealth" siRNA directed against PIAS1 (Invitrogen Life Technologies) and nucleofected using program T-027. Verification of the nucleofection regime was ascertained in experiments using a dsRNA FITC-labeled oligo provided in the siRNA kit and verified by confocal microscopy 48 h postnucleofection.
Confocal microscopy
HEp-2 cells were grown in 8-well chamber slides (NUNC Labtek II; VWR Scientific) with or without cytokine treatment and fixed with 10% neutral-buffered formalin. Cells were permeabilized with 0.1% Triton X-100/PBS for 30 min at room temperature, followed by blocking in 5% BSA (Roche Diagnostics) and 5% normal goat serum (Sigma-Aldrich) for 1 h. After washing with 0.1% Triton X-100/PBS, preparations were incubated with the specific primary Abs (i.e., those used in immunoblotting; 1/500) for 1 h at room temperature, rinsed extensively, and incubated with AlexaFluor 488- or AlexaFluor 594-conjugated secondary Abs (1/1000) (Molecular Probes) for 1 h at room temperature. Slides were washed five times in 0.1% Triton X-100/PBS, gel mount was applied, and coverslips were affixed. In experiments in which nuclei were counterstained, propidium iodide was added for 5 min and washed extensively before affixing the coverslip. Preparations were examined on an inverted Zeiss laser-scanning microscope (LSM 510, Axiovert 100 M; Zeiss) equipped with argon (450–514 nm) and helium-neon (543 and 633 nm) lasers. AlexaFluor 488 and AlexaFluor 594 fluorescence were exited using the 488- and 543-nm laser with emissions collected using the standard FITC and rhodamine filter set, respectively. Propidium iodide-labeled nuclei were excited using the 633-nm laser and collected with a Cyto-red filter set (12).
In vitro methylation assay
Semiconfluent HEp-2 monolayers were transferred into serum-free medium lacking amino acids capable of donating methyl groups for 16 h. At this point, the protein synthesis inhibitors cycloheximide (40 µg/ml) and chloramphenicol (100 µg/ml; both from Sigma-Aldrich) were added 1 h before the tritiated methyl donor S-[methyl-3H]-adenosyl-L-methionine (10 µCi/ml; NEN Life Science Products), ensuring that radiolabeling was due to methylation and not incorporation of the radiolabeled amino acid. Monolayers were stimulated with cytokines as indicated above, and whole-cell protein extracts were subjected to immunoprecipitation using anti-STAT1 Abs. Following SDS-PAGE, proteins were transferred onto polyvinylidene difluoride membranes and exposed to Kodak BioMAX MS film in a Kodak LE autoradiography enhancement cassette. Exposure was allowed to proceed for 1 wk at –70°C before development.
Cell surface biotinylation
Biotinylation of the HEp-2 surface proteins was conducted using EZ link Sulfo-NHS-LC biotin (Pierce) according to the manufacturer’s instructions. Briefly, 48 h after transfection, monolayers were washed with ice-cold PBS/Mg2+/Ca2+ (1 mM MgCl2 and 0.1 mM CaCl2 in PBS) and subsequently incubated with biotin (0.75 mg/ml EZ link Sulfo-NHS-LC) for 1 h on ice with gentle agitation. Following removal and quenching (100 mM glycine in PBS/Mg2+/Ca2+) of excess biotin, monolayers were washed three times with ice-cold PBS/Mg2+/Ca2+, and RIPA lysis buffer was added. Immunoprecipitation of the TGFRI was conducted as described above, followed by immunoblotting using an anti-biotin Ab (Promega). Membranes were stripped and reprobed with anti-TGFRI Ab to ensure equal loading (29).
Data presentation
Numerical data are presented as mean ± SEM and were compared by ANOVA, followed by Newman-Keuls statistical comparisons, where p 
0.05 was set as the level of statistically significant difference. Images shown are, in the main, representative of at least three separate experiments, with n values being stipulated in the figure legends.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
(22). However, while direct effects of TGF- on HEp-2, T84 and HT-29 have been shown, neither receptor expression nor SMAD protein mobilization have been presented. In initial experiments, immunoblotting for phosphorylated SMAD2 and immunolocalization of SMAD2 to the nucleus confirmed that HEp-2 cells stimulated with TGF- (1 ng/ml, 15 min) mobilize SMAD2/3 (data not shown).
TGF- pretreatment prevents IFN--induced STAT1-regulated gene expression
To determine whether TGF- would alter IFN--driven STAT1-dependent gene expression, HEp-2 cells were pretreated with TGF- (1 ng/ml: 15, 30, or 60 min) before IFN- stimulation and assayed for IRF-1, CIITA, GBP-1, and iNOS mRNA expression (all STAT1-regulated genes). As shown in (Fig. 1, A–D), pretreatment with TGF- attenuated the IFN--induction of mRNA expression of all four genes in HEp-2 epithelia but not THP-1 macrophages (IFN--induced IRF-1 mRNA expression in T84 epithelia was also reduced by TGF- pretreatment (data not shown)). Corroborating these findings, IFN--induced IRF-1 protein expression was suppressed in TGF--pretreated HEp-2 but not THP-1 cells (Fig. 1E).
|
-induced STAT1 DNA binding is reduced by TGF- pretreatment
To determine whether the reduced IFN--STAT1 driven expression was due to decreased STAT1 DNA binding we used EMSA analysis. Nuclear protein extracts from IFN--treated (20 ng/ml, 30 min) HEp-2 cells showed specific binding to the hSIE radiolabeled dsDNA probe (Fig. 2). Paralleling the reduced IRF-1 expression with TGF- pretreatment, STAT1 DNA binding was reduced in extracts from HEp-2, T84 (Fig. 2), or HT-29 cells (data not shown) pretreated for as little as 15 min with TGF-. Although there was some variability in the degree of inhibition of STAT1 DNA binding, which likely reflects interpassage cell variability (e.g., degree of expression of TGF- or IFN-R), five separate experiments showed that a 15- to 60-min pretreatment with TGF- consistently and significantly reduced IFN--induced STAT1 DNA binding. In stark contrast, inhibition of STAT1 DNA binding by TGF- was not observed in either THP-1 cells or PBMC drawn from three healthy volunteers (Fig. 2, D and E). Reduced STAT1 binding was not due to TGF--elicited SMAD2/3 binding to the hSIE probe as indicated by the complete absence of a signal in epithelial extracts from cells treated with TGF- only. This suggests that the effect TGF- is not at the promoter but is affecting the ability of STAT1 to bind DNA directly. In addition, the data indicate that this TGF- inhibition of STAT1 DNA binding is not a global response and may be specific and aligned with the function of specialized cell types.
|
-induced STAT1 phosphorylation and nuclear localization are not altered by TGF- pretreatment
Inhibition of STAT1 DNA binding in nuclear protein extracts from TGF- plus IFN--treated cells could be due to reduced STAT1 phosphorylation (a critical modification to allow maximal nuclear import and DNA binding (30)) or impaired nuclear import in general. Immunoblot analyses revealed that the levels of total STAT1 and IFN--induced STAT1 Tyr701 phosphorylation in epithelia and THP1 cells were unaffected by a 5- to 60-min pretreatment with TGF- (Fig. 3).
|
2-fold by IFN- (20 ng/ml, 30 min), this was not altered by TGF- pretreatment (n = 3; data not shown).
Confocal microscopy revealed that IFN--induced STAT1 nuclear translocation (Fig. 4, IFN-, arrowheads) was unaffected by pretreatment with TGF- and reciprocally that the SMAD2/3 nuclear localization induced by TGF- (Fig. 4, TGF-, arrows) was not blocked by subsequent IFN- exposure.
|
-induced STAT1 DNA binding
Many TGF- signal transduction events are mediated by the mobilization of proteins unique to the TGF-family of cytokines, designated as SMAD proteins. To determine the contribution of this pathway to the inhibition of STAT1 signaling, HEp-2 cells were transiently transfected with a plasmid (pcSMAD7) encoding the inhibitory SMAD7 protein. Transfection of HEp-2 epithelial cells with 4 µg/ml pcSMAD7 for 48 h resulted in negligible cytotoxicity and increased SMAD7 protein expression (Fig. 5, inset). HEp-2 cells in which SMAD7 was overexpressed were insensitive to TGF- inhibition of IFN--induced STAT1 DNA binding; indeed, SMAD7 overexpression restored IFN--induced STAT1 DNA binding in TGF--pretreated cells (rightmost lane in Fig. 5; n = 3). Increased SMAD7 expression can evoke TGFRI internalization and degradation (32). TGFRI internalization induced by pcSMAD7 was not observed by neither confocal microscopy (analysis of five randomly selected fields of view on three separate epithelial preparations/condition) nor by cell surface biotinylation followed by immunoprecipitation of the TGFRI and antibiotin immunoblotting (n = 3, data not shown).
|
activates the MAPK ERK1/2, and the stress activated kinases p38 and JNK. Overexpression of SMAD7 was not a specific antagonist of SMAD-mediated signaling as evidenced by the prevention of TGF--elicited ERK1/2 phosphorylation (Fig. 6, n = 4).
|
is independent of MAPK/stress-activated protein kinase (SAPK) activity
Phosphorylation of ERK1/2 MAPK was apparent after TGF- exposure (Figs. 6 and 7 inset), allowing for the possibility that ERK1/2 signaling could participate in the TGF--induced inhibition of IFN--STAT1 gene transcription in epithelial cells. To test this hypothesis, the ability of TGF- to inhibit IFN--induced STAT1 DNA binding was assessed in the presence of U0126, a pharmacological inhibitor of MEK1/2, the upstream activating kinase of ERK1/2. Although U0126 (1 h pretreatment; 5 µM (dose determined empirically; Fig. 7 inset)) prevented TGF--induced ERK1/2 activation, this ERK inhibition failed to restore STAT1 DNA binding in TGF- plus IFN--treated cells (Fig. 7). This demonstrates that TGF- inhibition of STAT1-mediated signaling is independent of TGF--mobilized ERK1/2.
|
was not dependent on SAPK activity as stimulation of HEp-2 epithelial cells with TGF- (1 ng/ml; 5–90 min) failed to activate, as gauged by phosphorylation on immunoblots, either p38 or JNK (data not shown; n = 6 and n = 3, respectively).
TGF- mobilized SMAD2/3 does not physically interact with IFN--induced STAT1
The colocalization of SMAD2/3 and STAT1 in the nucleus of TGF- plus IFN--treated cells (Fig. 4) suggested the possibility that the proximity of the transcription factors could allow for a physical interaction and thus reduced STAT1 DNA binding: it has been reported recently that SMAD2/3 and STAT1 may physically interact (33). We assessed whether SMAD2/3 activation would result in sequestering of STAT1, directly preventing DNA binding.
To determine whether sequestering was occurring, coimmunoprecipitation of epitope-tagged STAT1 and SMAD2 or SMAD3 was used. In this manner, cells were transfected with 2 µg/ml full-length STAT1-FLAG and either SMAD2-cmyc or SMAD3-cmyc. Whole-cell protein lysates from IFN- plus TGF--treated cells were divided in two, followed by anti-FLAG or anti-c-myc immunoprecipitation, electrophoresis, and immunoblotting for the reciprocal epitope tag. With this approach, we observed no evidence of a physical interaction between STAT1-FLAG and SMAD2-myc (data not shown) or SMAD3-myc (Fig. 8A). It is possible that the epitope tags may have prevented the STAT1 and SMAD2 or SMAD3 interaction. However, immunoprecipitation of endogenous STAT1, followed by immunoblotting for SMAD2 or SMAD3, or the reciprocal approach (i.e., immunoprecipitation for SMAD 2/3 and immunoblotting for STAT1), revealed no evidence of physical interaction between the endogenous proteins (Fig. 8B).
|
does not affect STAT1 methylation
Methylation of arginine residue 31 of STAT1 may inhibit DNA binding and gene transcription events by enhancing PIAS1-mediated sequestering of STAT1 (28). However, recent publications have questioned this postulate, demonstrating that STAT1 is not methylated in response to IFN- (34, 35). Nevertheless, assessing the possibility that STAT1 methylation could contribute to gene regulation in epithelia, HEp-2 cells were treated with IFN- with or without TGF- in the presence of a 3H-radiolabeled methyl donor. Methylation was determined by immunoprecipitation of STAT1 followed by SDS-PAGE, transfer to polyvinylidene difluoride membranes, and autoradiography. Although an abundance of methylated proteins was detected in whole-cell extracts and in the STAT1 immunoprecipitates, they were also observed in the control precipitates obtained with an irrelevant isotype-matched Ab (data not shown; n = 3). This lack of methylated STAT1 was not a consequence of reduced STAT1 in the extracts because total STAT1 was readily detected by immunoblotting the membranes used for the autoradiography (data not shown). Thus, we have no data to support STAT1 methylation in IFN--treated HEp-2 epithelial cells.
Sequestering of STAT1 by PIAS1 is enhanced by pretreatment with TGF-
Sequestering of phosphorylated STAT1 in the nucleus, thereby preventing DNA binding and resultant gene transcription, can occur through the PIAS1 protein. To assess whether sequestering of STAT1 by PIAS1 was enhanced by TGF-, coimmunoprecipitation of endogenous proteins were conducted for epithelia (HEp-2 cells) and macrophages (THP-1) treated with TGF- with or without IFN-. Whole-cell protein lysates from IFN- with or without TGF--treated cells were divided in two, followed by anti-STAT1 or anti-PIAS1 immunoprecipitation and immunoblotting for the reciprocal protein. Using this approach, and in accordance with other reports (36), we observed a low level of interaction of STAT1-PIAS in IFN--treated epithelial cells and constitutive interaction in THP-1 cells. Moreover, enhancement of STAT1-PIAS interaction by TGF- alone was restricted to HEp-2 cells (Fig. 9A, upper panel), and this association was even more evident when protein extracts from TGF- plus IFN--treated HEp-2 cells were assessed (Fig. 9A, n = 3). TGF- promotion of STAT1-PIAS1 interaction was not apparent in THP-1 cells (Fig. 9, n = 3). Critically, increased STAT1-PIAS1 binding was not due to TGF--induced production of PIAS1 because immunoblotting conducted on whole-cell extracts from the immunoprecipitated experiments revealed equivalent PIAS1 expression in TGF- with or without IFN--treated cells (Fig. 9, bottom panels).
|
. Captured PIAS1 was mixed with equal amounts of protein extracts from STAT1-FLAG-expressing cells with or without IFN- treatment (20 ng/ml, 30 min), followed by immunoblotting for FLAG. As shown in Fig. 10A, PIAS1 retrieved from TGF--treated epithelia showed an enhanced interaction with FLAG (i.e., the tagged STAT1). This further supports our contention that TGF- promotes PIAS1-STAT1 interaction.
|
was the mechanism through which STAT1 DNA binding was reduced, a validated siRNA targeted against PIAS1 was used. Specific knockdown of PIAS1 was accomplished by using an empirically determined dose of prevalidated siRNA directed against PIAS1 (Fig. 11, inset). Nucleofection of epithelia with 100 nM PIAS1 siRNA restored the ability of STAT1 to bind DNA in TGF- plus IFN--treated cells (Fig. 11). These data are not due to a direct response to the siRNA as evidenced by lack of STAT1 DNA binding in nonstimulated nucleofected controls (data not shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
and TGF- are key cytokines: the former mobilizing and perpetuating immune and inflammatory responses, and the latter exerting counterbalancing immunosuppressive and immunoregulatory functions (3). Strategies are being deployed to treat disease by restoring this balance through blocking proinflammatory cytokines or delivering anti-inflammatory cytokines (37). Moreover, cytokine production does not occur in isolation: multiple cytokines are present at any given time, and defining the constituents of the tissue cytokine milieu and how these stimuli interact are significant challenges. Knowledge of the intracellular signaling cascades evoked by cytokines could identify targets to either inhibit or promote cytokine responses in a cell-specific manner (38). Mechanistic insight on how IFN- suppresses TGF- signal transduction has been presented (14), but despite numerous examples of TGF- inhibition of IFN--driven events (13, 39, 40), little is known of TGF- modulation of intracellular signal transduction pathways evoked by IFN-.
Mucosal tissues are unique environments, whose dynamic multifunctional nature is the net effect of the integrated responses of many cell types: epithelium, fibroblasts, immune cells, and nerves. Epithelia and macrophages are important components of the innate immune system, whose activity influence subsequent adaptive immunity. As such, both cell types are important targets for IFN- and TGF-. We assessed if, and then how, TGF- could interfere with IFN--STAT1-driven gene transcription. A short pretreatment with TGF- significantly reduced STAT1 DNA binding and transcriptional activity of four target genes evoked by a 20-fold higher dose of IFN- in model epithelia. This rapid inhibition of STAT1 signaling by TGF- was not observed in macrophages. Divergence in TGF- regulation of IFN--STAT1 signaling in epithelia and macrophages could be due to differences in expression, or affinity, of the cytokine receptors or the efficiency of down-stream signals that transduce receptor-ligand interaction into transcriptional activity, aligning with the function of these cells. Given the role of TGF- in oral tolerance, intestinal cells may be continually exposed to low levels of this cytokine, and this could be coupled to rapid and efficient degradation of SMAD7, as recently suggested by Monteleone et al. (41). Thus, TGF- inhibition of the effects of proinflammatory stimuli, such as IFN-, that lead to loss of epithelial integrity and active participation in mucosal immunity (e.g., IFN- can increase epithelial MHC II expression (39)) would be beneficial. Alternatively, the macrophages’ role is to respond to microbial pathogens, and this function is enhanced by IFN-. The ability of low-dose TGF- (found constitutively in the gut) to override this could be deleterious, preventing local containment of bacteria that could lead to systemic infection, sepsis, and toxic shock.
Focusing on the epithelium and using HEp-2 as a model cell line, we sought to determine the mechanism responsible for inhibition of STAT1-mediated events by TGF-. Signal transduction through STAT1 is initiated by phosphorylation of Tyr701 by JAK2 and is a prerequisite for nuclear localization and DNA binding (42). IFN--evoked STAT1 Tyr701 phosphorylation was not altered by TGF-, indicating that the TGF- effect is downstream of IFN-R-JAK-STAT1 interaction and is not due to enhancement of phosphatase activity. Indeed, data in favor of (43), and refuting (44), TGF--inhibition of IL-12-induced JAK and STAT1 tyrosine phosphorylation in T cells have been reported. In light of normal STAT1 tyrosine phosphorylation in the TGF- plus IFN--treated epithelia, TGF- could prevent nuclear import or additional STAT1 modifications that are believed to maximize transcriptional activity such as serine phosphorylation (45, 46). However, reduced nuclear translocation was not diminished in TGF- plus IFN--treated epithelia, and IFN- induced STAT1 Ser727 phosphorylation that was not altered by TGF- pretreatment. In addition, we found no evidence of STAT1 methylation in our model system, and SUMOlation, that could affect transcriptional, also seems unlikely because a higher m.w. STAT1 should have been observed on immunoblot and/or EMSA analysis.
TGF- activates two major intracellular signal cascades, the unique SMAD2 and 3 proteins that bind with the regulatory SMAD4 protein and move to the nucleus and the ubiquitous MAPK (47, 48). Overexpression of the inhibitory SMAD protein, SMAD7, reduced the capacity of TGF- to inhibit IFN--induced STAT1 DNA binding. Importantly, despite the precedent for SMAD7 causing internalization and degradation of the TGFRI (49, 50), there were no differences in the amount of cell surface TGFRI between SMAD7 overexpressing and control epithelia as determined by confocal microscopy and biotinylation of cell surface proteins: this is in accordance with the recent suggestion that distinct endocytotic pathways control TGFRI turnover (51). Thus, restoration of STAT1 DNA binding in epithelia with enhanced SMAD7 expression is due to disrupted signaling downstream of the TGF- receptor as opposed to simply a lack of the TGFRI.
Inhibition of TGF- signal transduction was not restricted to SMAD-dependant signaling, as SMAD7 overexpression also prevented TGF- activation of ERK1/2 MAPK. Thus, we sought to ascertain the involvement of MAPK/SAPK involvement in TGF- inhibition of IFN--STAT1 signaling. Stimulation with TGF- (1 ng/ml) activated ERK1/2 MAPK but was insufficient to activate p38 or JNK in HEp-2 cells, and subsequent use of a selective MEK1/2 inhibitor, U0126, failed to restore STAT1 DNA binding in IFN- plus TGF--pretreated cells. Taken together, these results demonstrate that the TGF--evoked reductions in STAT1 DNA binding and STAT1-dependent gene transcription are SAPK- and MAPK-independent events.
The colocalization of SMAD2/3 and STAT1 in the nuclei of TGF- plus IFN--treated cells, and a precedent for physical interaction between STAT and SMAD proteins (52), pointed to the possibility of a direct STAT1-SMAD2/3 association blocking STAT1 DNA binding and subsequent transcriptional activity. Testing this hypothesis via coimmunoprecipitation of transfected epitope-tagged STAT1 and SMAD2/3 or endogenous STAT1 and SMAD2/3 proteins, we found no evidence in support of a physical association between STAT1 and SMAD2/3. Although a reduction in STAT1-driven gene transcription could be the result of SMAD2/3 binding with cofactors important in STAT1-mediated gene transcription, such as p300 and CBP (52, 53), this would not account for the reduced STAT1 DNA binding observed on EMSA analysis. Indeed, others have suggested that SMAD inhibition of STAT3 transcriptional events is not a consequence of p300/CBP sequestrating (54, 55).
Rather than a STAT1-SMAD2/3 interaction, the data herein show that TGF- promotes STAT1-PIAS1 binding. Enhanced sequestering by PIAS proteins can occur following increased availability of either PIAS or the target protein. Although TGF- induced activation of p38 increased mRNA production and protein stability of the related PIASx, it is unlikely that such a mechanism is at work here given the inability of 1 ng/ml TGF- to elicit p38 activation in HEp-2 cells, and indeed increased PIAS1 expression in TGF- cells was not observed. Despite the ability of PIAS proteins to interact with SMADs, either enhancing (56) or suppressing their transcriptional activity in response to TGF- (57), increased STAT1-PIAS1 interactions are unlikely to be via a STAT1-PIAS1-SMAD complex since this would have been detected in the coimmunoprecipitation assessment of STAT1 and SMAD2/3 interactions.
Three separate lines of evidence point to TGF- promotion of PIAS-STAT1 interaction as a mechanism for TGF- inhibition of IFN--evoked transcription: 1) coimmunoprecipitations show enhanced detection of STAT1-PIAS1 in TGF- plus IFN--treated cells; 2) in vitro binding assays show that PIAS1 obtained from TGF--treated cells has an enhanced ability to bind STAT1 (in this instance FLAG-tagged STAT1); and, 3) siRNA directed against PIAS1 restored the STAT1 DNA1 binding activity in nuclear extracts from TGF- plus IFN--treated cells. Furthermore, this PIAS1-STAT1 interaction is not due to new protein synthesis (or a technical protein loading issue) and, as discussed above, is unlikely to be a consequence of TGF- modification of STAT1. Thus our data fit best with a model in which TGF-, via a SMAD-dependent pathway, results in alterations to PIAS1 that promotes its interaction with STAT1 and inhibition of STAT1-directed transcription; the modification to PIAS1 needs to be determined.
In summary, TGF- (low dose, short duration exposure) is a potent inhibitor of STAT1-mediated signal transduction in epithelia, but not monocytes/macrophages, that is dependent on enhanced STAT1 sequestering in the nucleus by preexisting PIAS1. We suggest that defining the exact mechanism of TGF- interference with IFN--STAT1 signaling in epithelia will be important in understanding the complexity of cytokine cross-talk in mucosal compartments and as such will be relevant to oral tolerance, maintenance of the epithelial barrier and mucosal immunity in general.
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was supported by Canadian Institutes for Health Research Grant MT-13421 (to D.M.M.). D.M.M. is supported by an Alberta Heritage Foundation for Medical Research Scientist Award and a Canada Research Chair (Tier 1). C.R. is a recipient of Natural Sciences and Engineering Research Council of Canada studentship and had former support from a Premiers Research Excellence Award (to D.M.M.), Canadian Institutes for Health Research, and Ontario Graduate studentships. 
2 Address correspondence and reprint requests to Dr. Derek M. McKay, Gastrointestinal Research Group, Department of Physiology and Biophysics, HS-1877, University of Calgary, 3330 Hospital Drive Northwest, Calgary, Alberta T2N 4N1, Canada. E-mail address: dmckay{at}ucalgary.ca
3 Abbreviations used in this paper: PIAS1, protein inhibitor of activated STAT1; GBP-1, guanylate binding protein 1; hSIE, high-affinity sis-inducible element; iNOS, inducible NO synthase; IRF-1, IFN--regulated factor-1; RIPA, radioimmunoprecipitation assay; SAPK, stress-activated protein kinase; siRNA, small interfering RNA.
Received for publication June 1, 2006. Accepted for publication January 18, 2007.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
and TGF- responses regulate the occurrence of mucosal inflammation. Immunol. Today 18: 61-64. [Medline]. Proc. Natl. Acad. Sci. USA 98: 6674-6679. modulation of epithelial barrier function. J. Immunol. 150: 2356-2363. [Abstract] directly affects barrier function of cultured intestinal epithelial monolayers. J. Clin. Invest. 83: 724-727. [Medline] and tumour necrosis factor : inhibition of increased permeability, but not diminished secretory responses by transforming growth factor 2. J. Immunol. 159: 2382-2390. regulation of immune response. Annu. Rev. Immunol. 24: 99-146. [Medline]1 plasmid: TGF-1-mediated suppression of T helper cell type 1 response occurs by interleukin (IL)-10 induction and IL-12 receptor 2 chain downregulation. J. Exp. Med. 192: 41-52. mediated oral tolerance. J. Exp. Med. 183: 2605-2616. regulation of epithelial tight junction proteins enhances barrier function and blocks EHEC-induced increased permeability. Am. J. Pathol. 167: 1587-1597. 1 preserves epithelial barrier function: identification of receptors, biochemical intermediates, and cytokine antagonists. J. Cell. Physiol. 181: 55-66. [Medline]/SMAD signalling by the interferon /STAT pathway. Nature 397: 710-713. [Medline]1 inhibits nonpathogenic Gram negative bacteria-induced NF-B recruitment to the interleukin-6 gene promoter in intestinal epithelial cells through modulation of histone acetylation. J. Biol. Chem. 278: 23851-23860. 1 is mediated by caspase-dependent and -independent processes. Leukemia 16: 675-682. [Medline]
