SOCS3 promotor hypermethylation and STAT3-NF-κB interaction downregulate SOCS3 expression in human coronary artery smooth muscle cells

Abstract

Suppressor of cytokine signaling-3 (SOCS3) is an intracellular negative regulator of cytokine signaling pathway. We recently found significant reduction in SOCS3 expression in coronary artery smooth muscle cells (CASMCs) of atherosclerotic swine and also in vitro cultured cells. Here, we investigated the underlying mechanisms of SOCS3 downregulation by IGF-1 and TNF-α in human CASMCs(hCASMCs). We propose that hypermethylation of CpG islands in the SOCS3 promoter is responsible for decrease in SOCS3 expression involving STAT3 and NFkB-p65 interaction. Western blot and qPCR data revealed significant upregulation of SOCS3 (6- to 10-fold) in hCASMC when treated individually with TNF-α (100 ng/ml) or IGF-1 (100 ng/ml). However, a significant decrease (5-fold) was observed by the combined treatment with TNF-α and IGF-1 compared with individual stimulation. IGF-1 phosphorylated STAT3 and TNF-α-activated NF-κB in hCASMCs. In the nuclear extract of hCASMCs stimulated with both TNF-α and IGF-1, there was an interaction between NF-κB-p65 and pSTAT3, as determined by co-immunoprecipitation. Knockdown of STAT3 by small interfering RNA abolished SOCS3 expression in response to IGF-1. Methylation-specific PCR confirmed hypermethylation of SOCS3 promoter in hCASMCs stimulated with both TNF-α and IGF-1, and this was positively associated with elevated levels of DNA methyltransferase-I (9- to 10-fold). Knockdown of DNMT1 increased SOCS3 expression in IGF-1+TNF-α-stimulated cells. Downregulation of SOCS3 in the presence of both TNF-α and IGF-1 in hCASMCs is due to SOCS3 promoter hypermethylation involving STAT3-NFkBp65 interaction. Because TNF-α and IGF-1 are released due to mechanical injury during coronary intervention, hypermethylation of SOCS3 gene could be an underlying mechanism of intimal hyperplasia and restenosis.

inflammatory cytokine, TNF-α (26), and a potent smooth muscle cell growth factor, IGF-1, play a critical role in the development of restenosis after intervention in atherosclerotic arteries (9). TNF-α activates monocytes, macrophages, and endothelial cells through its mitogenic and pro-apoptotic effects, which regulate atherosclerosis (31). Transcription factor NF-κB is known to regulate TNF-α signaling (1). It consists of two subunits: p50 and p65 (Rel A). P65 regulates a cellular protective mechanism against the cytotoxic effects of TNF-α. IGF-1 elicits pleiotropic effects on cells involved in atherosclerosis (35). In addition to PI3-kinase/Akt/PKB pathway, IGF-1 may activate Janus kinase/signal transducer and activator of transcription (JAK/STAT) signaling pathway by upregulating STAT3 activity (5, 17). JAK/STAT signaling is a vital cellular mechanism in the regulation of cytokines and growth factors during inflammation, cellular activation, and proliferation (37, 15, 19, 41, 48). The binding of cytokines and growth factors to their corresponding receptors activates JAK, which then phosphorylates the receptor and STAT protein on specific tyrosine residues. STAT molecules dimerize upon activation and migrate to the nucleus to induce transcription of target genes. Suppressor of cytokines signaling 3 (SOCS3) is a member of eight structurally related intracellular SOCS family proteins (49) that function as tumor suppressor by inhibition of the JAK/STAT signaling pathway (46, 44, 20). SOCS3 decreases the JAK/STAT activity by directly interacting with phosphorylated receptors through the SH2 domain (2, 18, 22, 27, 33) and also by the binding to JAK through its kinase inhibitory region (34, 47).

Recently, we reported a significant reduction in the SOCS3 expression in the neointimal lesion after balloon angioplasty in coronary arteries of atherosclerotic swine (11). These neointimal lesions are also highly positive to both TNF-α and IGF-1. In the in vitro studies, treatment of porcine coronary artery smooth muscle cells with both TNF-α and IGF-1 decreased SOCS3 expression (11). It was indeed interesting to find an upregulation of SOCS3 expression in porcine coronary artery smooth muscle cells when treated individually with either IGF-1 or TNF-α. However, simultaneous treatment of the cells with both mediators significantly diminished SOCS3 expression. In this study, we further examined such phenomenon in human coronary artery smooth muscle cells (hCASMCs) and determined the underlying mechanisms and investigated the role of activated NF-κB and STAT3 proteins in the hypermethylation of SOCS3 gene.

MATERIALS AND METHODS

Reagents.

Recombinant human TNF-α and IGF-1 were purchased from PeproTech (Rocky Hill, NJ). The cell culture media, growth factors, and antibiotics were purchased from Science Cell (Carlsbad, CA). The following antibodies were used: SOCS3 (No. 2923), STAT3 (No. 9132), STAT1 (No. 9172), and DNMT1 (No. 5119) (Cell Signaling Technology, Danvers, MA), pSTAT3(Y705) (ab30646) (Abcam, Cambridge, MA), NF-kBp65(sc-8008) (Santa Cruz Biotechnology, Santa Cruz, CA) and GAPDH (NB300-221) (Novus Biologicals, Littleton, CO). SiRNAs for DNMT1, STAT3, and STAT1 were obtained from Dharmacon (Lafayette, CO).

Smooth muscle cell culture.

hCASMCs were obtained from Cell Applications (San Diego, CA) at passage 2. Cells were grown in smooth muscle cell media (SMCM) supplemented with 10% FBS, smooth muscle cell growth supplement and 1% penicillin-streptomycin. hCASMCs between passages 3 and 5 were used for in vitro experiments. Confluent cells (about 70%) were serum starved for 24 h before stimulation. Cells were stimulated with TNF-α and IGF-1 at 100 ng/ml concentration for 24 h individually and/or in combination. Cells were harvested by trypsinization and were used for isolation of protein, mRNA, and genomic DNA.

Protein isolation and Western blot analysis.

The cells were suspended in a radioimmunoprecipitation assay buffer (Sigma, St Louis, MO) containing 25 mM Tris·HCl (pH 7.6), 150 mM NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, and 0.1% SDS for 5 min in ice with protease inhibitor (Sigma). Suspended cells were sonicated and centrifuged at 10,000 g for 10 min. Supernatant was used for protein analysis. The protein content of the sample was measured using BCA protein assay kit according to the manufacturer's protocol (Sigma). The protein samples (20 μg) were subjected to 10–20% SDS -PAGE (Bio-Rad, Hercules, CA) and then transferred to nitrocellulose membrane (Bio-Rad) for immunoblotting. After the nonspecific protein was blocked with 5% milk, the membrane was washed and incubated for 1 h with targeted antibodies, which were diluted 1:1,000 in nonfat milk in PBS-Tween. The blot was washed again and incubated for another 1 h with horseradish peroxidase-conjugated anti rabbit (Novus Biologicals) diluted at 1:1,000. Finally, immunoblot was developed with ECL chemiluminescence detection reagents (Amersham Pharmacia Biotech) system using UVP Bioimaging system. Results were normalized with housekeeping protein GAPDH.

RNA isolation and quantitative PCR.

After stimulation of cells with TNF-α and IGF-1 for 24 h, total RNA was isolated using the TRIzol reagent (Sigma) method. The yield of RNA was quantified using Nanodrop (Thermo Scientific, Rockford, IL). First-strand cDNA synthesis was done using 1 μg total RNA with oligo dT (1 μg), 5× reaction buffer, MgCl₂, dNTP mix, RNAse inhibitor, and Improm II reverse transcriptase as per Improm II reverse transcription kit (Promega, Madison, WI). After the first strand synthesis, real-time PCR was done using 8 μl cDNA, 10 μl SYBR green PCR master mix (Bio-Rad Laboratories) and forward and reverse primers (10 pM/μl) (Integrated DNA Technologies, San Diego, CA) using a real-time PCR system (CFX96; Bio-Rad Laboratories). The primer sequences used were SOCS3, FP 5-AGCAGCGATGGAATTACCTGGAAC-3′, RP 5-TCCAGCCCAATACCTGACACAGAA-3′; GAPDH, FP 5′-GGGAAGGTGAAGGTCGGAGT-3′, RP 5′-TTGAGGTCAATGAAGGGGTCA-3′. The specificity of the primers was analyzed by running a melting curve. The PCR cycling conditions used were 5 min at 95°C for initial denaturation, 40 cycles of 30 s at 95°C, 30 s at 58°C and 30 s at 72°C. Each real-time PCR was carried out using three individual samples in triplicates, and the threshold cycle values were averaged. Calculations of relative normalized gene expression were done using the Bio-Rad CFX manager software based on the ΔCt method. The results were normalized against housekeeping gene GAPDH. The mRNA expression of DNMT1, DNMT3A, and DNMT3B was examined in a similar manner. The primer sequences used for DNMTs were as follows: DNMT1 F 5′-AAGAACGGCATCCTGTACCGAGTT-3′, DNMT1 R 5′-TGCTGCCTTTGATGTAGTCGGAGT-3′; DNMT3A F 5′-TTTGAGTTCTACCGCCTCCTGCAT-3′, DNMT3A R 5′-GTGCAGCTGACACTTCTTTGGCAT-3′; DNMT3B F 5′-AGTGTGTGAGGAGTCCATTGCTGT-3′, DNMT3BR 5′-GCTTCCGCCAATCACCAAGTCAAA-3′.

Pretreatments.

Before every experiment, hCASMCs were starved overnight in the serum-free smooth muscle cell media. According to the experiment design, cells were pretreated with actinomycin D (0.2 μg/ml for 24 h) or cycloheximide (25 μM for 1 h) in experiments mentioned in Fig. 5 and Tyrphostin AG490 (25 μM for 2 h) in experimental data shown in Fig. 6. Cells were treated with 100 ng/ml IGF-1 and/or 100 ng/ml TNF-α, unless otherwise mentioned for durations as demanded by the experiment. After treatment with IGF-1 and TNF-α, RNA was isolated and subjected to PCR for detection of SOCS3 in case of actinomycin D pretreatment, and protein lysate was subjected to Western blot for detection of SOCS3 in case of pretreatment with cycloheximide.

Fig. 5.

Co-immunoprecipitation of NF-κBp65 with STAT3 in nuclear and cytoplasmic extract upon IGF-1 and TNF-α treatment. A: hCASMCs were serum starved and treated with 100 ng/ml of IGF-1 and TNF-α alone and together for 2 h. Upon completion of treatments, cells were harvested and nuclear and cytoplasmic extracts were isolated. The samples were precleared with protein A/G agarose slurry to remove nonspecific binding. NF-κB p65 antibody was used for immunoprecipitation, and immune complexes were captured with protein A/G agarose slurry. The samples were washed and run on Western blot for detection of association with STAT3 by STAT3 antibody, and the bands were visualized using ECL solution. The picture is representative of 3 individual experiments. B: SOCS3 expression after pretreatment with tyrphostin AG490. hCASMCs were serum starved for 24 h and pretreated with tyrphostin AG490 (25 μM). Cells were then stimulated with IGF-1 or TNF-α (100 ng/ml; 24 h), together or alone. C and E: protein expression of STAT1 and STAT3. Cells were harvested and proteins were isolated after 24 h transfection of STAT1 and STAT3 sall interfering RNAs. D and F: SOCS3 expression after knockdown of STAT1 and STAT3 in presence of TNF-α and IGF-1 alone or in combination in hCASMCs. Protein samples were separated by SDS-PAGE. SOCS3 induction was detected with SOCS3 primary antibody. GAPDH was used as housekeeping gene. The picture is a representative of 3 individual experiments. Data are means ± SE for triplicate experiments. *P < 0.001.

Fig. 6.

The SOCS3 gene structure and methylation-specific PCR analysis. A: a schematic diagram of SOCS3 gene. Bottom line represents genomic sequence of SOCS3 gene containing exon 1, intron 1, and exon 2. The translation start site is +1, and arrow indicates direction of translation. Top line represents whole CpG region including SOCS3 gene. The arrows indicate primers used for MSP method. B: MSP products were analyzed in 2% agarose gel using 100-bp standers. The size of MSP product was 141 bp. C: figure represented the densitometric analysis of PCR products compared with control.

Preparation of cytoplasmic and nuclear extract.

Nuclear and cytoplasmic extraction reagents (NE-PER cat. No: 78833; Pierce Biotechnology, Rockford, IL) were used. Briefly, after treatment, cells were harvested and cell pellets were dissolved in 100 μl CERI. This was vortexed for 15 s and incubated on ice for 10 min. CERII (5.5 μl) was further added, and the mixture was vortexed for 5 s and incubated on ice for 1 min. After another vortex for 5 s, the samples were centrifuged for 5 min at 16,000 g. The supernatant (cytoplasmic extract) was transferred to separate tubes and pellets were resuspended in 50 μl NER. Samples were vortexed for 15 s and incubated on ice for 10 min. This process was repeated for a total of 40 min. Samples were then centrifuged at 16,000 g for 10 min, and supernatant carrying nuclear fraction was transferred to fresh tubes. Samples were stored at −80°C until used.

Immunoprecipitation.

To prepare Protein A/G agarose bead slurry (Santa Cruz) the beads were washed twice in PBS and restored to 50% slurry in PBS. Cell lysate was precleared by adding 100 μl of bead slurry per 1 ml of cell lysate and incubating at 40°C for 10 min on a rocker to reduce nonspecific binding. Agarose was removed by centrifuging lysate at 14,000 g at 40°C for 10 min. Protein concentration of supernatant was determined. Protein (2 mg/ml) was used for immunoprecipitation. Primary antibody was added at a 1:500 dilution and mixture was gently rocked for 2 h at room temperature. The immune complex was captured by adding 100 μl protein A/G agarose bead slurry and gently rocking for 1 h at room temperature. The beads were collected by pulse centrifugation (5 s at 14,000 g). The beads were washed three times with 800 μl ice-cold PBS. The pellet was resuspended in 60 μl sample buffer. The beads were boiled for 5 min to dissociate from immunocomplexes. Beads were collected by centrifugation, and 100 μg of sample per well was run on an SDS-PAGE as described under Western blot section.

Luciferase promoter activity assay.

Effect of TNF-α and IGF-I, individually and in combination, was assessed on NF-κB promoter activity in hCASMCs. Cells were grown in 24-well plate and cotransfected with pNFκB-Luciferase reporter vector and control vector (pRL-TK) (Promega) for 24 h. The cultured cells were replaced in growth medium with 10% serum for a further 24 h before the cytokine stimulation for 4 h. The cells were then collected and lysed. Firefly and Renilla luciferase activities were measured by luminometer using the dual-luciferase reporter assay system. The normalized luciferase activities were calculated by dividing firefly activities by Renilla activities and expressed relative to the control transfected nonstimulated cells.

STAT1, STAT3, and DNMT1 knockdowns.

STAT1, STAT3, and DNMT1 genes were silenced in hCASMCs by transfecting specific siRNAs for 24 h. A nontarget oligo was used as a negative control and the gene silencing was confirmed by Western blot. Transfections were carried out using the transfection reagent siPORTNeoFX (Ambion). Furthermore, the cells were stimulated with TNF-α and IGF-1 individually or together for another 24 h, and SOCS3 expression was analyzed by Western blot.

Methylation-specific PCR.

Genomic DNA was isolated from the treated and untreated hCASMCs using the GeneElute Mammalian Genomic miniprep kit (Sigma) according to the instructions of the manufacturer. The quantity of the DNA was measured by the NanoDrop spectrophotometer (Thermo Scientific). Methylation-specific PCR (MSP) method was used for analysis of the methylation status of the promoter site of SOCS3. Bisulfite reaction was carried out with 1,000 ng of purified DNA using the treatment of Epitect Bisulfite kit from Qiagen (Valencia, CA). Bisulfite-treated DNA was amplified by using either a methylation or unmethylation specific primer sets. All primer sets used for MSP method are shown in Table 1. Methylation-specific PCR reactions were performed with 32 ng of bisulfite-modified DNA in a total volume of 25 μl for 35 cycles using FastStart Taq DNA Polymerase at 1.0U (Roche Diagnostic, Indianapolis, IN), MgCl₂ solution at 3.5 mM, dNTPs at 0.2 mM, and forward primer and reverse primer each at 0.3 μM, with denaturation at 95°C for 30 s, annealing at 60°C for 45 s, and extension at 72°C for 1 min. PCR products were electrophoresed on 1% agarose gel, stained with ethidium bromide, and visualized with UVP Bioimaging system. Densitometric analysis of the PCR product was performed by Labwork software.

Table 1.

Primers used in methylation analysis of suppressor of cytokine signaling-3 promoters

Primer Set	Sense Primer, 5'–3	Antisense Primer, 5'–3'	Size Position	Genomic
Unmethylated 1	GTTGGAGATTTTAGGTTTTG GAATATTT	AAACCCCCAAAACTACCTA AACACCA	150	−529
Methylated 1	GGAGATTTTAGGTTTTCGGA ATATTTC	CCCCCGAAACTACCTAAA CGCCG	141	−525
Umethylated 2	GGAGGGGTTGTTGTTAGGAAT	CAAAACAAAACCAAAAAACA	91	−1266
Methylated 2	GAGGGGTCGTTGTTAGGAA	CACAAAAACCGAAAAAACGC	89	−1265
Unmethylated 3	TGTGGTGGTTGTTTATATATA TTTGTGAGTGTGGTT	CAACCAACAATAACCCACAC TACACCCA	298	−1018
Methylated 3	TATATATTCGCGAGCGCGGTTT	CGCTGCGCCCA GATGTT	268	−1005
Unmethylated 4	AATATTATAAGAAGGTTGG TTGTGTAG	AAATCCAAATCAAACCAC CACAA	108	−68
Methylated 4	AATATTATAAGAAGGTCGGT CGCGTAG	AAATCCAAATCGAAACCG CCGCGA	108	−67

Statistical analysis.

Data are presented as means ± SE. Data were analyzed by using GraphPad Prism. Multiple-group comparisons were performed by Bonferroni's multiple comparison test using one-way ANOVA. Probability (P) value <0.05 was accepted as statistically significant.

RESULTS

IGF-1 and TNF-α induce SOCS3 in a dose- and time-dependent manner.

The protein expression of SOCS3 was examined in hCASMC stimulating with inflammatory cytokine, TNF-α (100 ng/ml), and IGF-1 (100 ng/ml) individually and/or in combination at 12, 24, 36, and 48 h. There was detectable level of SOCS3 in the initial stage without cytokines and mitogen. Time course studies showed maximum SOCS3 expression at 24 h after stimulation with TNF-α or IGF-1alone (Fig. 1, A and B). But significantly decreased SOCS3 expression was observed after 12 h when cells treated with both TNF-α and IGF-1 together (Fig. 1C). There was an increasing expression of SOCS3 in hCASMCs treated individually with increasing concentrations of IGF-1 or TNF-α. Cells were treated with increasing concentrations (10–200 ng/ml) of TNF-α or IGF-1, protein samples were run on SDS-PAGE for Western blot. Maximum expression was seen with a dose of 100 ng/ml after 24 h of incubation. The SOCS3 expression was down with a dose of 200 ng/ml (Fig. 2).

Fig. 1.

Time course of suppressor of cytokine signaling-3 (SOCS3) protein expression with TNF-α and IGF-1. Human coronary artery smooth muscle cells (hCASMCs) were stimulated with TNF-α (100 ng/ml; A), IGF-1 (100 ng/ml; B) alone, or in combination (C) at 12, 24, 36, and 48 h in passage 3. Proteins were isolated from cell lysate and subjected to Western blot using SOCS3 antibody. GAPDH was used as a loading control. N = 3; *P < 0.05, **P < 0.01, ***P < 0.001, compared with control group.

Fig. 2.

SOCS3 protein and mRNA expression in presence of both TNF-α and IGF-1 in hCASMC. A and B: dose-dependent SOCS3 expression in hCASMCs with IGF-1 and TNF-α individually. hCASMC were treated with increasing concentration 10, 20, 50, 100, and 200 ng/ml of IGF-1 (A) or TNF-α (B) for 24 h. C: cells were stimulated with TNF-α (100 ng/ml) and IGF-1 (100 ng/ml) for 24 h alone or in combination. The relative SOCS-3 protein expression was determined using GAPDH as loading control. D: graph represented relative fold change in SOCS-3 mRNA expression compared with control. The real-time PCR and Western blot data represented of three individual data sets. (N = 3; *P < 0.05, **P < 0.01, ***P < 0.001 compared with control group).

SOCS3 expression was then studied at 24 h in the presence of both TNF-α and IGF-1 alone and/or in combination. The SOCS-3 protein expression in hCASMCs significantly increased in the presence of TNF-α (5-fold) or IGF-1 (6-fold) alone. However, there was a fourfold decrease when both TNF-α and IGF-1 were used together compared with individual stimulation (Fig. 2C). The mRNA transcripts of SOCS3 were significantly induced in the presence of TNF-α (10-fold) or IGF-1 (8-fold) alone in hCASMCs. However, a significant decrease (5-fold) in SOCS3 mRNA transcripts was observed in the presence of both TNF-α and IGF-1, compared with that of individual stimulation (Fig. 2D).

STAT3 phosphorylation and NF-κb promoter activity in response to IGF-1 and TNF-α.

hCASMCs were treated with 100 ng/ml dose of IGF-1 and TNF-α for 2 h. Protein was isolated and run on SDS-PAGE for Western blot. With the use of primary antibody specific to pSTAT3, Western blot experiment was performed. Immunoreactive bands were detected using ECL reagent. The pSTAT3 protein bands were detected in response to IGF-1, indicating phosphorylation of STAT3 with IGF-1 (Fig. 3A). No band was detected in untreated cells as well as in the cells treated with TNF-α. The STAT3 phosphorylation, albeit to much lesser extent, was also detected when hCASMCs were treated with IGF-1 and TNF-α together. STAT3 antibody was used as control.

Fig. 3.

pSTAT3 and NF-κB expression in hCASMCs upon treatment with IGF-1 and TNF-α. A: activation of transcription factor STAT3 with IGF-1 treatment. hCASMCs were serum starved for 24 h and treated for 5, 15, and 30 min with TNF-α and IGF-1 (100 ng/ml each). In a separate experiment, cells were treated with different doses of TNF-α and IGF-1 for 30 min. Cells were harvested at the time of completion of treatment, and proteins were isolated. Samples were run for Western blot analysis using pSTAT3 as primary antibody. Bands were visualized with ECL solution. STAT3 was used as a loading control. The figure is a representative of 3 individual experiments. B: effect of TNF-α and IGF-1 NF-κB promoter activity. hCASMCs cells were grown in 24-well plate and cotransfected with pNF-κB-Luciferase reporter vector and control vector (pRL-TK) for 24 h. The cultured cells were replaced in growth medium with 10% serum for a further 24 h before the cytokine and growth factor stimulation for 4 h. The cells were then collected and lysed. Firefly and Renilla luciferase activities were measured by luminometer using the dual-luciferase reporter assay system. The normalized luciferase activities were calculated by dividing firefly activities by Renilla activities and expressed relative to the control transfected nonstimulated cells. P < 0.05 compared with control values (n = 3).

To test if NF-κB, which has a promoter binding site present on SOCS3 response element, is activated in response to TNF-α, luciferase assay was performed. hCASMCs were cotransfected with pNF-κB-Luciferase reporter vector and control vector (pRL-TK) and treated with 100 ng/ml of IGF-1 and TNF-α for 4 h. Samples were subjected to detection of luciferase activity. NF-κB promoter activity was detected only in response to TNF-α, indicating NF-κB activation in response to TNF-α and not IGF-1 (Fig. 3B).

SOCS3 induction by IGF-1 and TNF-α together is blocked upstream of transcription.

We further examined the underlying mechanisms of inhibition of SOCS3 following treatment of hCASMCs with both IGF-1 and TNF-α. To check if this inhibition was occurring at the level of transcription or translation, hCASMCs were pretreated with actinomycin-D or cycloheximide, respectively. After treatment of the cells with actinomycin D (0.2 μg/ml) for 2 h, cells were incubated with 100 ng/ml IGF-1 or TNF-α for 12 h followed by washing of the cells. RNA was extracted using phenol/chloroform method. The RNA sample from SOCS3-overexpressed cells was used as a positive control. When compared with the samples without any treatment, no SOCS3 mRNA could be detected. SOCS3 mRNA expression was detected in actinomycin D-treated cells when incubated with either IGF-1 or TNF-α, and SOCS3 transcription was completely blocked when both IGF-1 and TNF-α were added together. This suggests an inhibition in the pathway before the initiation of transcription. When cells were pretreated with 25 μM cycloheximide for 2 h followed by incubation with both IGF-1 and TNF-α, no SOCS3 could be detected in any case (Fig. 4).

Fig. 4.

Expression of SOCS3 at transcription and translation. A: serum-starved hCASMCs were pretreated with 0.2 μg/ml actinomycin D for 24 h and then treated with 100 ng/ml IGF-1 and TNF-α, alone or together for 24 h. RNA was isolated using phenol/chloroform method, and RT-PCR was run using SOCS3 primers. As control, SOCS3 overexpressed cells were subjected to the same treatment. RNA from SOCS3 overexpressed cells was used as a positive control. Primers for β-actin were used to ensure equal loading in each well. The figure is a representative of 3 individual experiments. B: serum starved hCASMCs were pretreated with 25 μM cycloheximide for 1 h and then treated with 100 ng/ml IGF-1 and TNF-α, alone or together for 24 h. Protein was isolated and resolved on Western blot for detection of SOCS3. Cells were overexpressed with SOCS3, and samples were subjected to same treatment as control. GAPDH was used as housekeeping gene. The figure is a representative of 3 individual experiments.

Transcription factors STAT3 and NF-κBp65 interact upon treatment with IGF-1 and TNF-α.

We examined the interaction between transcription factors NF-κB and STAT3. Cells were treated with 100 ng/ml IGF-1 and TNF-α, alone and together for 3 h. We performed immunoprecipitation of NB-κB p65 from hCASMCs after various treatments. The samples were run on SDS-PAGE for Western blot using pSTAT3 antibody, to check binding with pSTAT3. No bands were detected in untreated cells as well as cells treated with IGF-1 or TNF-α alone. However, in cells incubated with both IGF-1 and TNF-α, a co-immunoprecipitated band was detected, indicating binding of pSTAT3 with NF-κB p65 (Fig. 5A). To further establish this interaction, cells were pretreated with STAT3 inhibitor, AG490, before incubation with IGF-1 and TNF-α, together and alone. The protein isolated from the samples was subjected to Western blotting and SOCS3 protein expression was detected. IGF-1-induced SOCS3 expression was almost completely blocked with the use of STAT3 inhibitor, whereas TNF-α-induced SOCSC3 was still unaffected by this inhibitor. Interestingly, protein expression of SOCS3 was detected in AG490-trested cells when incubated with both IGF-1 and TNF-α (Fig. 5B).

Silencing of STAT3 affects the expression of SOCS3 but not STAT1.

To establish the role of STAT3 and STAT1 in the regulation of SOCS3 expression in the presence of IGF-1 and TNF-α, we performed gene knockdown experiments using small interfering RNA (siRNA) for STAT1 and STAT3 independently. Western blot analysis showed 80–90% knockdown of STAT1 and STAT3 in hCASMCs (Fig. 5, C and E). Silencing of STAT1 in hCASMCs did not affect the increase in SOCS3 expression due to either IGF-1 or TNF-α and also did not affect the decrease in SOCS3 expression when both TNF-α and IGF-1 were used together (Fig. 5D). But, SOCS3 expression was completely diminished after STAT3 silencing in the presence of IGF-1 and in control STAT3-silenced hCASMCs (Fig. 5F). But STAT3 knockdown did not alter SOCS3 expression due to TNF-α alone or in combination with IGF-1 (Fig. 5F).

IGF-1 and TNF-α together trigger hypermethylation in SOCS3 promoter.

We examined the methylation status of CpG islands of SOCS3 promoter in hCASMC in the presence of TNF-α or IGF-1 alone or in combination. The structure of SOCS3 and location of the CpG islands are shown in Fig. 6A. The sequence of the methylation-specific primer and unmethylation-specific primers is shown in Table 1. Four sets of methylated and unmethylated primers were used to study the methylation in different regions in SOCS3 promoter near the recognized STAT binding sites. We were able to detect methylation with first set of primers that recognizes the STAT-binding site at −525 bp upstream of the transcription of SOCS3. We did not find any methylation at other three STAT binding sites on SOCS3 promoter. Methylation was observed only when both TNF-α and IGF-1 were present (Fig. 6B). No methylation was found in normal cells or cells stimulated with either TNF-α or IGF-1 alone. The densitometric analysis data, as shown in Fig. 6C, confirmed these findings.

Combination of IGF-1 and TNF-α increased DNMTs mRNA expression.

We further investigated the mRNA expression level of mammalian DNA methyltransferases, DNMT1, DNMT3A, and DNMT3B, in the presence of either TNF-α or IGF-1 alone or in combination. The mRNA expressions of DNMT1, DNMT3A, and DNMT3B were considerably high when hCASMCs were stimulated with both TNF-α and IGF-1 together. The mRNA expression of DNMT1 was significantly higher (4- to 5-fold) than DNMT3A and DNMT3B in the cells under SOCS3 silencing condition (Fig. 7).

Fig. 7.

DNMTs mRNA expression in presence of TNF-α and IGF-1 in hCASMC. A: mRNA expression of DNMT1. B: DNMT3A mRNA expression. C: DNMT3B mRNA expression. The graph represented relative fold change of mRNA expression compared with control. Real-time PCR analysis represented of 3 individual data sets (N = 3; *P < 0.05, ***P < 0.001 compared with control group).

Silencing of DNMT1 increases the expression of SOCS3.

Because we found that DNMT1 level was significantly high in the presence of IGF-1 and TNF-α together in hCASMCs, we silenced the DNMT1 and analyzed SOCS3 expression. The expression of SOCS3 dramatically increased after knocking down of 80% of DNMT1 in the presence of IGF-1 and TNF-α together (Fig. 8A). The methylation status of SOCS3 promoter was also analyzed after silencing DNMT1 in the presence IGF-1 and TNF-α individually and together. We did not see promoter methylation when IGF-1 and TNF-α present together in hCASMCs after DNMT1 silencing (Fig. 8B).

Fig. 8.

Silencing of DNMT1 in the presence of TNF-α and IGF-1 in hCASMC and methylation analysis. A: protein expression of DNMT1 and SOCS3. hCASMCs were harvested after silencing of DNMT1 followed by simultaneous stimulation with IGF-1 and TNF-α for 24 h. Proteins were isolated and resolved on Western blot. B: analysis of methylation. After silencing of DNMT1, hCASMCs were stimulated with IGF-1 and TNF-α individually or in combination for 24 h. Methylation-specific PCR was performed and analyzed on 2% agarose gel using 100-bp standers.

DISCUSSION

The present study was conducted to understand the cellular and molecular mechanisms of the effect of IGF-1 together with TNF-α on the expression of regulatory protein SOCS3 in hCASMCs. During the development of atherosclerotic plaque and intimal hyperplasia smooth muscle cells are one of the first cells to migrate from the media to the intima in response to an endothelial injury and inflammation. IGF-1, being one of the most potent mitogen, is released not only from smooth muscle cells but also from endothelial cells as well as inflammatory cells, including macrophages. Although inflammatory cells release many cytokines, including IL-18, IL-12, and IFN-γ, TNF-α plays a critical role in bringing about inflammatory response to the site of an atherosclerotic lesion. SOCS3 is mostly known to be induced in response to cytokines, as a regulatory protein.

In this study, we designed the experiments under an environment for hCASMCs similar to those found in an atherosclerotic lesion, where the cells are exposed to cytokines and growth factors simultaneously. During coronary intervention, both cytokines and growth factors are secreted. The increase in the mRNA and protein expression of SOCS3 in response to individual incubation with either TNF-α or IGF-1 is consistent with the findings in the literature in other cells (3–6, 17, 39). However, the most intriguing and novel finding is that the stimulation of hCASMCs simultaneously with TNF-α and IGF-1 significantly decreases the SOCS3 mRNA and protein expression. This supports our recent report in a porcine model (11). However, the underlying molecular mechanisms are unclear.

Two main transcription factor response elements, STAT3 and NF-κB, have been reported on the SOCS3 promoter region (13). We found the activation of NF-κB primarily by TNF-α, whereas IGF-1 prefrentially induced the STAT3 phosphorylation in hCASMCs. However, co-operative or antagonistic action of various proteins may serve as a regulatory function of transcription requiring several proteins. STAT and NF-κB are activated by different pathways and migrate to the nucleus to bring about a transcriptional activity. In mesangial cells, STAT3 inhibits NF-κB activity via direct interaction with NF-κB p65 and thus inhibits the transcription of NF-κB-induced inducible NOS gene (50). Also, STAT5b inhibits NF-κB-mediated activation of target promoters (24). Interestingly, direct interaction between STAT6 and NF-κB may provide a basis for synergistic activation of transcription by IL-4 and activators of NF-κB (36). In a study, JAK-specific inhibitor, tyrphostin AG490, markedly prevented pancreatitis-associated protein1-induced NF-κB inhibition, indicating a cross-talk between JAK/STAT3 and NF-κB signaling pathways (7). In our study, we examined the role of STAT3/NF-κB interaction in the transcription and translation of SOCS3 in the presence of both IGF-1 and TNF-α in hCASMCs. We found a coprecipitated band of pSTAT3- NF-κB in the cells after treatment with IGF-1 and TNF-α exclusively in the nuclear fraction and not in the cytoplasmic fraction. Because SOCS3 gene in its promoter region contains STAT3, STAT5b, and NF-κB response elements, their interaction could interfere with the attachment of the DNA itself and thus decreasing the binding of both STAT3 and NF-κB to their response elements on SOCS3 promoter. This could be supported by the finding in the cells treated with STAT3 inhibitor, AG490, where further treatment of the cells with IGF-1 and TNF-α induced SOCS3 expression, indicating the induction of SOCS3 in a STAT3-independent manner. A similar result was obtained when STAT3 was silenced using siRNA in hCASMCs. Silencing of STAT1 in hCASMCs did not affect the decrease in SOCS3 expression due to combined effect of both IGF-1 and TNF-α. These data suggest that the association of STAT3 (activated by IGF-1) and NF-κB (activated by TNF-α) may not have an overall inhibitory action of protein induced by either of them individually.

Indeed, the activation and physical interaction between STAT3 and NF-κB were found in many human head and neck squamous carcinoma and B lymphocyte cancer cells (10, 12, 38). STAT3 activity is required to maintain NF-κB activity in both cancer cells and tumor-associated hematopoietic cells (21, 45). However, in our study in hCASMCs, although simultaneous activation of both STAT3 and NF-κB inhibits SOCS3 expression, it is possible that such interaction between pSTAT3- NF-κB could upregulate intracellular molecules that are more favorable to uncontrolled growth of vascular smooth muscle cells. However, it is likely that the regulation of genes during the interaction between pSTAT3- NF-κB is controlled by the epigenetic mechanisms including methylation-demethylation process.

Increased expression of DNMT1 has been observed in malignant T lymphocytes when STAT3 binds to promoter and enhancer regions of DNMT1 (51, 52). Many studies have demonstrated the essential role of DNMT1 in gene silencing by promoter hypermethylation in cancer cells (16, 29, 32). Methylation often occurs in the promoter region of SOCS3 in all types of human cancer (8, 14, 23, 25, 28, 30, 40, 42, 43), resulting in the silencing of SOCS3 gene. Vascular smooth muscle cells in neointimal hyperplasia have growth kinetics and proliferation very similar to cancer cells. Therefore, we further studied DNA methylation in SOCS3 promoter using methylation-specific PCR in four different CpG regions of SOCS3 promoter. The CpG island in the human SOCS3 gene spans in the noncoding exon 1, intron 1, and exon 2. Also, out of the four STAT biding sites on SOCS3 promoter, we found an aberrant DNA methylation only in one of the STAT binding sites that lies in the 505 bp upstream to the translational site. No methylation was observed at other STAT binding sites. This suggests that all STAT binding sites are not required to be methylated to silence the SOCS3 promoter. However, it is unclear how STAT binding sites on the SOCS3 promoter influences DNA methylation. But we found high level of DNA methyltransferases, DNMT1, DNMT3A, and DNMT3B, in hCASMCs treated with both TNF-α and IGF-1. Among these DNMTs, the increase in the DNMT1 was the highest (9-fold) compared with the control group. Significant increase in the expression of SOCS3 was observed after silencing of DNMT1 in presence of both IGF-1 and TNF-α in hCASMC. However, no methylation was detected during similar conditions in hCASMCs. High levels of DNMT1 in the presence of both TNF-α and IGF-1 could be responsible for hypermethylation of SOCS3 promoter.

This is the first report demonstrating the methylation of the SOCS3 promoter gene and subsequent reduction of mRNA and protein expression in hCASMCs in the presence of both TNF-α and IGF-1. This could be responsible for the development of intimal hyperplasia following coronary interventional procedures when both inflammatory cytokines and mitogens are released at the site of mechanical injury.

GRANTS

The research reported in this publication was supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health under Award Numbers R01HL090580, R01HL104516, and R01HL112597. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: K.D., K.R., and D.P. performed experiments; K.D., K.R., D.P., and D.K.A. analyzed data; K.D., K.R., D.P., and D.K.A. interpreted results of experiments; K.D. and K.R. prepared Figs.; K.D. drafted manuscript; K.D., K.R., D.P., and D.K.A. approved final version of manuscript; D.K.A. conception and design of research; D.K.A. edited and revised manuscript.