Abstract

Ulcerative colitis (UC) is associated with a high risk of developing colorectal cancer (CRC). The mechanisms by which chronic inflammatory responses in the colon may promote CRC remain only partially understood, but may involve reduced negative regulation of interleukin (IL)-6 signaling towards signal transducer and activator of transcription 3 activation through the loss of SOCS3 expression, unleashing the full carcinogenic potential of this transcription factor. Thus, we analyzed SOCS3 expression in the colon of healthy controls, as well as in a cohort of UC patients with varying degrees of dysplasia. We observe that the loss of epithelial SOCS3 expression delimits the areas subject to dysplasia in UC, suggesting an important tumour-suppressive role of SOCS3 downregulation, early in the transformation process. Importantly, methylation of the SOCS3 promotor appears to constitute an important regulatory mechanism for colonic SOCS3 expression as SOCS3 methylation status in CRC cells correlates with a disability to upregulate SOCS3 upon IL-6 stimulation, whereas forced demethylation using 5-aza-2ʹ-deoxycytidine restores SOCS3 expression and inhibits IL-6-induced p-signal transducer and activator of transcription 3 activation and proliferation. Expression of the DNA methyltransferase gene DMNT1 is prominent in dysplastic cells and correlates with low or absent SOCS3 expression. Thus, induction of DNMT1 expression in the chronically inflamed colon may release IL-6 signaling towards signal transducer and activator of transcription 3 from inhibition through SOCS3 increasing the propensity to malignant transformation. Hence, DNMT1 emerges as a rational target in preventive strategies aimed at counteracting UC–CRC.

Introduction

Ulcerative colitis (UC) is a major manifestation of inflammatory bowel disease (IBD) and is clinically characterized by chronic recurrent colitis and is histopathologically characterized by the presence of a continuous inflammatory infiltrate in the intestinal mucosa ascending from the rectum into the colon ( 1 ). Chronic UC is associated with an increased risk to develop colorectal cancer (CRC) and UC–CRC underlies approximately 10–15% of all mortality in patients with IBD. In these patients, the risk for contracting CRC is approximately 20–100 times higher in the population at large, depending on the absence or presence of Primary sclerosing cholangitis (PSC) as a co-morbidity ( 2 , 3) . As such, UC–CRC represents a special challenge in the prevention and treatment of gastrointestinal cancer.

The pathophysiological mechanisms explaining the unusual high risk of colon cancer in UC remain only partially understood but may involve the chronic proliferation necessary to repair the damage inflicted on the epithelial monolayer by the constant inflammation ( 4 , 5) . In this context, high interleukin (IL)-6, detected in colonic biopsies and serum samples of UC patients, is considered an especially deleterious factor with respect to the development of malignant disease. IL-6 is a pro-inflammatory cytokine that is associated with carcinogenesis at multiple places in the human body. Signaling through the canonical cytokine receptor pathway, IL-6 evokes activation of Janus kinase 2, in turn activating the signal transducer and activator of transcription 3 (STAT3), which can mediate evasion of immunosurveillance, promote cell growth and increase survival signaling ( 6 , 7) . Full manifestation of these pro-oncogenic effects of STAT3, however, requires inactivation of the negative regulators of IL-6 signaling. The most important of the latter mechanisms is expression of the suppressor of cytokine signaling 3 (SOCS3), which is a direct target gene of STAT3 and mediates suppression of IL-6 signaling through the ubiquitination and degradation of signaling intermediates. The question how during UC–CRC carcinogenesis the negative influence of STAT3-dependent induction of SOCS3 is overcome remains one of the most important questions in the field.

We have previously reported on the silencing of SOCS3 expression as a potential mechanism explaining sustained IL-6-dependent STAT3 phosphorylation and activation in the inflammation–dysplasia–carcinoma sequence in UC ( 8 ), but left the underlying molecular mechanism unexplored. Obviously, delineation of these mechanisms might be exceedingly useful for devising rational strategies for the prevention and maybe also for treatment of UC-associated CRC. Interestingly, in cholangiocarcinoma, a disease that shares many characteristics with UC–CRC, IL-6 signaling has been associated with increased expression of DNA methyltransferase-1 (DNMT1) ( 9 ), a CpG island methylator, which has thus been proposed to provide a link between IL-6 levels and inflammation-associated cholangiocarcinoma ( 10 ) and UC-related carcinoma ( 11 ) through DNMT1-mediated tumour-suppressor gene silencing. Causative relationships, however, remain to be established and especially identification of the tumour-suppressor genes involved remains a very open issue.

Here, we explore the possibility that DNMT1 induction by IL-6 in UC mediates silencing of SOCS3 expression and consequently CRC we observe that spatially UC–CRC is restricted to areas with loss of SOCS3 expression, that SOCS3 promoter methylation makes the gene resistant to IL-6-dependent induction of its expression in a DNMT1-dependent fashion and that both in vivo , as well as in patients, IL-6-dependent activation of STAT3 signaling is strictly dependent on DNMT1 enzymatic activity. The dependence of carcinogenic progression of UC–CRC on DNMT1 and the subsequent inhibition of SOCS3 transcription make this methyltransferase a prime candidate target for chemopreventive strategies in this disease.

Materials and methods

Patient biopsies and cell lines

Patients’ characteristics are shown in Table 1 . Paraffin-embedded biopsies from 9 normal controls, 27 active UC, 11 UC–CRC and 14 sporadic CRC were included. Colon biopsy specimens were obtained, after informed consent, during colonoscopy performed in the course of routine clinical care and surveillance. The diagnosis of UC was based on conventional clinical, endoscopic and pathohistological criteria as described by Lennard-Jones ( 12 ). An independent pathologist reconfirmed carcinoma by hematoxylin and eosin staining and histological scoring. Colonic biopsies from patients without abnormalities at colonoscopy and histories of gastrointestinal disease served as controls. Five CRC cell lines were included, which are caco-2, HT29, SW480, SW48 and DLD1. They are cultured in Dulbecco’s modified Eagle’s medium, supplemented with streptomycin 10% (v/v) fetal calf serum, 1% (v/v) penicillin and streptomycin, at 37°C in a humidified atmosphere containing 5% CO 2 . All the in vitro experimentation was repeated at least three times.

Immunohistochemistry staining and scoring

All biopsies were fixed in 10% formalin and embedded in paraffin. The deparaffinization, endogenous peroxidise blocking and staining processes were performed as described previously ( 8 ). Antigen retrieval was performed by boiling in preheated buffer Tris/EDTA pH 9.0 for 15min at 200W in a microwave. Next, slides were transferred to Shandon chambers and blocked by 10% normal human serum, 10% goat serum in phosphate-buffered saline pH 7.4 for 1h at room temperature. Primary antibodies rabbit polyclonal anti-DNMT1 (Abcam, ab19905) and rabbit polyclonal anti-IL-6 (Abcam, ab6672) were added and incubated at 4°C overnight. Envision goat anti-rabbit-horseradish peroxidase (DAKO, Denmark) was used as secondary antibody. Two independent trained scientists scored the slides with SOCS3-positive epithelial cells in a blinded manner. The percentage of epithelial cells that stained positive (immunoreactivity above background) in the area consistent with that used for diagnoses was quantified.

DNA isolation and bisulfate modification

Different cell-line DNA was isolated using trizol reagent. Cell-line DNA samples were bisulfate modified using the EZ DNA methylation TM kit (ZYMO Research). Methylated DNA and unmethylated DNA (millipore) were acquired commercially and used as a positive and negative control in the enzymatic assay.

Methylation-specific PCR

Methylation-specific PCR (MSP)-specific primers were either designed by Methprimer or as described previously by Weber Hengge ( 13 ) and used to amplify methylated and unmethylated DNA sequences as appropriate. A schematic overview of the strategy employed is provided in Supplementary Figure 2 , available at Carcinogenesis Online. MSP was performed on bisulphite-treated genomic DNA. The PCR was performed in a 25 µl volume containing 40ng bisulphite-modified DNA, 5x Gold Taq buffer, 0.2mM deoxynucleoside triphosphates, 10 pmol specific primer mix (forward and reverse primers) and one unit Gold Taq enzyme (Promega). Three regions of SOCS3 promoter were targeted; the PCR conditions and sequences of each primer mix are given in Table 2 . The PCR products were visualized on a 2% agarose gel using ethidium bromide and UV illumination.

Fig. 2.

Different SOCS3 promoter methylation in CRC cell lines. (A) The methylation status was determined in five CRC cell lines using MSP at three different regions of SOCS3 promoter. HT29, SW480, SW48 and DLD1 showed more SOCS3 methylation than caco-2 cells. (B) Induction of SOCS3 expression upon IL-6 stimulation in CRC cell lines. (C) IL-6-induced methylation of SOCS3 gene in caco-2 cells after 48h.

Fig. 2.

Different SOCS3 promoter methylation in CRC cell lines. (A) The methylation status was determined in five CRC cell lines using MSP at three different regions of SOCS3 promoter. HT29, SW480, SW48 and DLD1 showed more SOCS3 methylation than caco-2 cells. (B) Induction of SOCS3 expression upon IL-6 stimulation in CRC cell lines. (C) IL-6-induced methylation of SOCS3 gene in caco-2 cells after 48h.

Normal PCR and pyrosequencing

Bisulphite genomic DNA was amplified region exon 1 to intron before exon 2, on SOCS3 promoter. Product was purified and sequenced by input 15ng per 100-bppurified PCR product per 20 µl reaction. SOCS3 promoter region 5ʹ-untranslated region (UTR) before exon 1 was amplified and prepared using PyroMark™ Vacuum Prep Workstation. The methylation status was quantified by pyrosequencing machine PyroMark™MD, Biotage. The PCR conditions and sequences of each primer mix were shown in Table 2 .

Demethylation treatment

For demethylation, experiments were seeded using 1:10 dilution of a 1×10 5 cells/ml stock cell suspension into 12-well plates. Twenty-four hours later, 5-aza-2ʹ-deoxycytidine (5-Aza-dC; Sigma) was added to a final concentration of 10 μM. Three days (72h) after 5-Aza-dC treatment, the cells were stimulated by IL-6 (100ng/ml) and harvested for quantitative real time polymerase chain reaction (QPCR) and western blot. Independent cells (control and demethylated cells) cultured after 3 and 6 days were viewed by microscope.

RNA isolation and real-time PCR

Total RNA was isolated from cell lines using TRIzol (Sigma). About 1 μg of RNA was reversely transcribed using iScript TM cDNA synthesis kit (Bio-rad). quantitative real time polymerase chain reaction (QPCR) reactions were performed using real-time PCR kit sensimix TM SYBR&Flurescence (Quntace) in a final volume of 25 μl containing 10 pmol primers in IQ5 (Bio-rad). The experimental conditions and primers employed for specific experimentation is provided in Table 2 . Data obtained are expressed as the ratio between SOCS3 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA delta-Ct to yield the fold induction in mRNA expression relative to the appropriate control.

Western blot

Western blotting was performed as described earlier ( 14 ). Immobilon IF transfer membranes (Millipore) were probed with primary antibodies specific for SOCS3 (rabbit polyclonal, abcam, ab16030), p-STAT3 (rabbit monoclonal, #9145, cell signal) and DNMT1 (rabbit polyclonal, Abcam, ab19905). α-tublin (Abcam ab6046) and β-actin (Santa Cruz, E0610) were used to provide the levels of a reference protein to ascertain equal loading. Horseradish peroxidase–labelled goat anti-rabbit and goat anti-mouse were used as secondary antibody (LI-COR Biosciences). Transfer membranes were incubated with antibodies and washed in 50ml sterile light-protecting centrifuge tubes (Greiner bio-one). Protein expression was visualized using fluorescence Odyssey system (LI-COR Biosciences, Lincoln, NE, USA). Quantitative expression data were determined by Odyssey 3.0 software and normalized using β-actin or α-tublin for reference gene expression.

Thiazolyl blue tetrazolium bromide assay

Control cells and demethylated cells were cultured in 96-well plates for in total four days. Subsequently 0.5mg/ml thiazolyl blue tetrazolium bromide was added to the cell medium. As described previously ( 15 ), the absorbance was measured using a microplate reader Model 680XR (BIO-RAD) at 490nm. All experiments were repeated three times. Percentage of cell viability after demethylation was calculated as % Viability = mean OD in sample well/mean OD in control well × 100, where OD is optical density.

Statistical analysis

Statistical analysis was performed using SPSS software version 15.0. The possible statistical differences between were determined using the Mann–Whitney test or, if appropriate, the Kruskal–Wallis test. The correlation between two variables was assessed by determining the Spearman correlation coefficient. A P value < 0.05 was considered statistically significant.

Results

Loss of epithelial SOCS3 expression delimits the neoplastic domain in UC–CRC

We previously proposed that silencing of SOCS3 expression explains sustained IL-6-dependent STAT3 phosphorylation and activation in the inflammation–dysplasia–carcinoma sequence in UC ( 8 ). Similarly downregulation of SOCS3 was not found in CD-dysplasia and CD-CRC patients ( Supplementary Figure 1 , available at Carcinogenesis Online), which highlights the specificity of SOCS3 downregulation in UC-carcinogenesis. A prediction from the hypothesis that SOCS3 downregulation in UC unleashes the full oncogenic potential of IL-6 signaling would be that the cancerous process in UC–CRC should be limited to areas with loss of SOCS3 expression. Hence, we investigated SOCS3 expression in the dysplastic, as well as the adjacent non-dysplastic, epithelial cells ( Figure 1A ). Non-dysplastic epithelial cells were characterized by relatively high levels of SOCS3 expression, whereas the SOCS3 expression was reduced in dysplastic areas. Such a correlation was absent in areas directly adjacent to the dysplastic regions, suggesting that the downregulation of SOCS3 is truly dysplastic tissue specific and not a reflection of general environmental changes in the colon of these patients ( Figure 1A and 1B ). It thus appears that dysplasia per se only occurs in the context of loss of SOCS3 expression and subsequent experiments were initiated to identify the mechanisms underlying the loss of SOCS3 in the inflammation–dysplasia–carcinoma sequence in UC.

Fig. 1.

Specific SOCS3 down regulation in epithelial cells during progression from low-grade dysplasia (LGD) to UC–CRC in the dysplastic and/or cancer region. (A) Immunohistochemistry (IHC) staining shows SOCS3 expression in dysplastic/cancer and adjacent non-dysplastic/cancer area in the same biopsies. (B) Epithelial SOCS3 expression was compared in the same biopsies between dysplastic/cancerous areas and non-dysplastic areas. SOCS3-positive epithelial cells are significantly more numerous in the non-dysplastic/non-cancer area than in the dysplastic/cancer area from the same biopsies ( P < 0.05). To assess SOCS3 expression in different areas of cancer tissues, experienced gastro-pathologist evaluated dysplasia/cancer after hematoxylin and eosin staining and marked dysplastic/cancerous area in every slide. Then two independent trained scientists scored the slides with SOCS3 positive epithelial cells in a blinded manner. The scoring criteria were defined as mild expression <30%, positive expression 30–60%, strong positive expression >60% of SOCS3 positive epithelial cells.

Fig. 1.

Specific SOCS3 down regulation in epithelial cells during progression from low-grade dysplasia (LGD) to UC–CRC in the dysplastic and/or cancer region. (A) Immunohistochemistry (IHC) staining shows SOCS3 expression in dysplastic/cancer and adjacent non-dysplastic/cancer area in the same biopsies. (B) Epithelial SOCS3 expression was compared in the same biopsies between dysplastic/cancerous areas and non-dysplastic areas. SOCS3-positive epithelial cells are significantly more numerous in the non-dysplastic/non-cancer area than in the dysplastic/cancer area from the same biopsies ( P < 0.05). To assess SOCS3 expression in different areas of cancer tissues, experienced gastro-pathologist evaluated dysplasia/cancer after hematoxylin and eosin staining and marked dysplastic/cancerous area in every slide. Then two independent trained scientists scored the slides with SOCS3 positive epithelial cells in a blinded manner. The scoring criteria were defined as mild expression <30%, positive expression 30–60%, strong positive expression >60% of SOCS3 positive epithelial cells.

SOCS3 promoter methylation is a major regulator of SOCS3 expression in the colonic epithelial compartment

We reasoned that the mechanisms mediating inhibition of SOCS3 expression in UC-associated carcinogenesis might be reflected in in vitro models for colon cancer. Thus, a panel of colon cancer cell-line cultures was checked for both unstimulated SOCS3 levels, as well as the potential of IL-6 to provoke SOCS3 gene expression. It appeared that SOCS3 mRNA levels already varied substantially between different cell lines ( Figure 2B ), demonstrating that regulation is different in established colon cancer cell cultures, but that especially the capacity of IL-6 to stimulate transcription of SOCS3 is widely divergent, such induction being very strong in caco-2 cultures but absent in DLD1. A preliminary analysis of IL-6 receptor levels revealed that alternative expression of IL-6 signal transduction elements was unlikely to account for these differences ( Supplementary Figure 3 , available at Carcinogenesis Online). As identification of the mechanisms underlying differences in expression of SOCS3 may hold important clues as to the regulation of the protein in vivo , we checked promoter methylation in this panel as well. The methylation status was determined in five CRC cell lines by MSP at three different regions of SOCS3 gene promoter. It revealed a high negative correlation between promoter methylation and basal SOCS3 levels, but effects were especially prominent when the capacity of IL-6 to induce SOCS3 expression was investigated: IL-6 is almost incapable of doing so when SOCS3 promoter is heavily methylated ( Figure 2A and 2B ). Consistent with these MSP results, pyrosequencing confirmed increased CpG methylation in SOCS3 promoter from colon cell cultures (HT29 17.12%, sw480 18.06%, sw48 18.46%, DLD1 19.33% and caco-2 cells only 10.86%). It thus seems that SOCS3 promoter methylation is an important mechanism controlling IL-6-dependent SOCS3 expression in colon cancer cultures.

Fig. 3.

DAC restores SOCS3 inducibility. (A–B) In DLD1 cells, DAC treatment increases basal levels of SOCS3 expression both at mRNA ( P < 0.05) (A) and protein level ( P < 0.05) (B), and this expression was further increased after IL-6 stimulation. DAC restoration of SOCS3 expression at basal levels reduced the IL-6-induced STAT3 phosphorylation in DLD1 cells ( P < 0.05) (B). (C–D) The restoration of SOCS3 expression by DNMT1 inhibition affects cell growth. DLD1 cells were cultured in medium with or without DAC for 3 days, then for an additional 3 days, with or without IL-6. Cells treated with DAC for the entire 6 days shown a significant reduction in cell growth compared with those cultured without DAC ( P < 0.05). Whereas IL-6 stimulation further increased the growth of cells in the absence of DAC, DAC treatment prevented enhanced cell growth by IL-6.

Fig. 3.

DAC restores SOCS3 inducibility. (A–B) In DLD1 cells, DAC treatment increases basal levels of SOCS3 expression both at mRNA ( P < 0.05) (A) and protein level ( P < 0.05) (B), and this expression was further increased after IL-6 stimulation. DAC restoration of SOCS3 expression at basal levels reduced the IL-6-induced STAT3 phosphorylation in DLD1 cells ( P < 0.05) (B). (C–D) The restoration of SOCS3 expression by DNMT1 inhibition affects cell growth. DLD1 cells were cultured in medium with or without DAC for 3 days, then for an additional 3 days, with or without IL-6. Cells treated with DAC for the entire 6 days shown a significant reduction in cell growth compared with those cultured without DAC ( P < 0.05). Whereas IL-6 stimulation further increased the growth of cells in the absence of DAC, DAC treatment prevented enhanced cell growth by IL-6.

Expression of DNMT1 negatively correlates with capacity for SOCS3 induction

Searching for candidate genes possibly involved in the methylation of the SOCS3 promoter in colonic cells, we decided to study the possible involvement of DNMT1 ( 16 ), because the UC–CRC highly expressed cytokine IL-6 is able to upregulate the expression of DNMT1 ( 10 ). Earlier, we showed that activity of this gene increased chemosensitivity of colorectal cancer cells by inducing epigenetic reprogramming and reducing colorectal cancer-cell stemness ( 17 ), thus the activity of this enzyme constitutes an important epigenetic regulatory event. In this study, we established that ideal time frame for this epigenetic experiment at 48h and thus also in the present study, this temporal window for the epigenetic effects of IL-6 stimulation on nuclear reprogramming was chosen. We observe that IL-6-induced stimulation of colorectal cancer cell cultures provokes DNMT1 expression at both mRNA and protein ( Supplementary Figure 4 , available at Carcinogenesis Online), at this time point, suggesting that expression of DNMT1 methylating activity under the control of inflammatory mediators in the colon. Furthermore, we found that the IL-6-induced DNMT1 expression correlated with enhanced SOCS3 promoter methylation ( Figure 2C ). Hence, DNMT1 appears to be one of the enzymes responsible for increased methyltransferase activity leading to methylation of the SOCS3 promoter in turn negatively influencing feedback on IL-6-dependent signal transduction.

Fig. 4.

IL-6 and DNMT1 expression and their correlation in patient biopsies. (A) Immunohistochemistry (IHC) staining was used to visualize DNMT1 expression in the biopsies from controls, UC–CRC and active UC. (B) Analysis of the number of IL-6 positive epithelial cells. IL-6 expression is significantly higher in UC–CRC ( P = 0.0002), sporadic CRC ( P = 0.0017) and active UC ( P < 0.0001) than in normal controls. Biopsies from UC-CRC had even higher IL-6 expression than sporadic CRC ( P = 0.006). (C) Analysis of DNMT1 positive epithelial cells. DNMT1 expression is significantly higher in UC–CRC ( P = 0.020) and active UC ( P = 0.004) compared with controls. (D) Overall the expression levels of IL-6 and DNMT1 in all the biopsies show positive correlation (Spearman’s correlation coefficient r = 0.371, P = 0.002).

Fig. 4.

IL-6 and DNMT1 expression and their correlation in patient biopsies. (A) Immunohistochemistry (IHC) staining was used to visualize DNMT1 expression in the biopsies from controls, UC–CRC and active UC. (B) Analysis of the number of IL-6 positive epithelial cells. IL-6 expression is significantly higher in UC–CRC ( P = 0.0002), sporadic CRC ( P = 0.0017) and active UC ( P < 0.0001) than in normal controls. Biopsies from UC-CRC had even higher IL-6 expression than sporadic CRC ( P = 0.006). (C) Analysis of DNMT1 positive epithelial cells. DNMT1 expression is significantly higher in UC–CRC ( P = 0.020) and active UC ( P = 0.004) compared with controls. (D) Overall the expression levels of IL-6 and DNMT1 in all the biopsies show positive correlation (Spearman’s correlation coefficient r = 0.371, P = 0.002).

DNMT1 regulates the SOCS3 expression and function in colon cancer cell cultures

Our data suggest that IL-6-dependent DNMT1 induction mediates SOCS3 downregulation, in turn releasing IL-6 signaling towards STAT3 from its negative regulation. A prediction from this hypothesis is that inhibition of DNMT1 should restore IL-6-dependent induction of SOCS3 and counteract methylation of SOCS3 promoter. Thus, we tested the effect of the methylation inhibitor 5-aza-2ʹ-deoxycytidine (DAC) on DLD1 cells. Indeed, DAC treatment increased SOCS3 mRNA and protein levels in otherwise unchallenged cells ( Figure 3A and 3B ). Furthermore, following the treatment with the methylation inhibitor, cells became susceptible to IL-6-dependent SOCS3 induction. Finally, concomitantly IL-6-dependent induction of STAT3 was impaired ( Figure 3B ), as well as IL-6-dependent cell proliferation was impaired ( Figure 3C and 3D ). To illustrate the effect of methylation inhibitor on cell growth, DLD1 cells were cultured in medium with or without DAC for 3 days, then for an additional 3 days, with or without IL-6. Cells treated with DAC for the entire 6 days showed a significant reduction in cell growth compared with those cultured without DAC ( P < 0.05). Whereas IL-6 stimulation further increased the growth of cells in the absence of DAC, DAC treatment prevented enhanced proliferation by IL-6. Thus, these in vitro experiments provide support for the notion that an IL-6-dependent downmodulation of methyltransferases reduces regulation of IL-6 signaling towards STAT3 and is permissive for uncontrolled colonic cell proliferation and subsequently we investigated whether support for this notion could be obtained in patient samples of UC–CRC.

IL-6 is associated with DNMT1 over-expression in CRC

To investigate the possible importance of DNMT1 expression in the pathogenesis of UC–CRC, we investigated biopsies from controls, UC, UC–CRC and sporadic CRC for the expression of IL-6 (showed previously ( 8 )) and DNMT1 ( Figure 4A ). As expected, IL-6 expression was significantly higher in UC–CRC ( P = 0.0002), sporadic CRC ( P = 0.0017) and active UC ( P < 0.0001) than in normal controls, in agreement with a role of IL-6 in tumor progression in the colon. Furthermore, biopsies from UC–CRC have higher IL-6 expression compared with sporadic CRC ( P = 0.006; Figure 4B ), in agreement with an important role of IL-6 expression for specifically IBD-associated cancer. Importantly, also DNMT1 expression is significantly higher in UC–CRC ( P = 0.020) and active UC ( P = 0.004) compared with controls ( Figure 4C ). There seems to be some correlation between the expression of IL-6 and DNMT1 in the epithelial cells, irrespective of disease phenotype or activity (Spearman r = 0.371, P = 0.002). Taken together, these data support the notion that induction of DNMT1 is an important factor in permitting UC–CRC to develop.

Discussion

The highly increased incidence of CRC associated with chronic inflammatory disorders lumped together under the common denominator IBD is of complex and poorly understood etiology, but a large body of evidence in the corpus of biomedical literature implicates increased IL-6 signaling through STAT3 in this phenomenon ( 18–20 ). We previously reported substantial silencing of SOCS3 expression in the inflammation–dysplasia–carcinoma sequence in UC ( 8 ), whereas the same situation was not found in CD-dysplasia and CD-CRC patients ( Supplementary Figure 1 , available at Carcinogenesis Online). Apart from multi-effective immune inducers on SOCS3 expression, the location difference in large and small intestine of these two disease types would also lead the specificity of SOCS3 downregulation exerting the sustained IL-6-dependent STAT3 activation in UC-carcinogenesis. However, limited data could not lead to more clear conclusions. Earlier, it was shown that SOCS3 is an important negative regulator of STAT3 during intestinal inflammation, suggesting that loss of SOCS3 would provoke activation of this pro-oncogenic transcription factor ( 21 ). Indeed, we reported that concomitant with the loss of SOCS3, STAT3 activation increased, providing a rational explanation as to why SOCS3 loss is related to an increased propensity to develop UC–CRC. Thus, we proposed that the loss of SOCS3 expression represents a critical event, permissive for the development of UC–CRC, but the molecular mechanism as to why chronic inflammation would provoke the loss of SOCS3 expression remained obscure at most.

In the present study, we demonstrate that SOCS3 downregulation delimits the extent of the dysplastic and cancerous domain in patients with UC-dyspalsia and UC–CRC respectively, further supporting SOCS3 downregulation is a critical event here. Furthermore, we provide evidence that in the colonic compartment, the SOCS3 promoter is subject to regulation through methylation. Finally, we show that IL-6 signaling can provoke an increase in methylation of SOCS3 by stimulating increased DNMT1 expression and that results in increased signaling through STAT3. Finally, we provide evidence that DNMT1 is upregulated in vivo in UC and its expression correlates to both the areas with high IL-6 signaling and those which cancer is developed. Hence we propose that under the influence of the prolonged signaling through IL-6, the SOCS3 promoter is methylated releasing signaling towards STAT3 from inhibition and allowing the cancerous process to progress. Thus, DNMT1 emerges as an attractive target for rational strategies aimed at preventing CRC in patients IBD.

From our data, it emerges that SOCS3 methylation involves large areas of its promoter (see also Supplementary Figure 2 , available at Carcinogenesis Online), as MSP showed strong methylation in three different regions of this promoter. Importantly, one of areas positively identified as being methylated includes the STAT3 binding element, which is considered essential for induction of SOCS3 in response to IL-6 and seen as the main element responsible for limiting the extent of SOCS3 induction in vivo . The relevance of this finding for the cancerous process is illustrated by our observation that absence of SOCS3 expression delimits the dysplastic domain in UC. Thus, it seems that methylation of this particular element might constitute a rate-limiting step in the generation UC–CRC.

In our patient cohort, no specific significant differences between the genders emerged. It has been reported, though, that DNA methylation patterns change with increasing age and these changed methylation pattern contribute to ageing. Specifically, in a study investigating cytosine methylation in the promoter of 80 genes, the degree of methylation showed significant positive correlation with the male sex and with age ( 22 ). Also in different neuroendocrine tumours, gene methylation is positively associated with older age of patients and more aggressive manifestation of cancer ( 23 ). In our patients, we did detect more methyltransferase expression and SOCS3 gene methylation in older patients, suggesting that age is a factor for the mechanisms described in the present study as well.

IL-6/p-STAT3 signaling in vitro has been shown to stimulate both colonic epithelial cell proliferation and to induce resistance to apoptosis ( 24 ). In apparent agreement, in animals genetically devoid of SOCS3 in the colonic mucosa, enhanced sensitivity to repeated dextran sulfate sodium (DSS) challenge is noted with respect to the occurrence of colon carcinoma ( 25 ). The current data add to this concept as they provide a mechanistic link to the previous murine data to actual patients. However, the role of STAT3 in colonic physiology is not clear cut. Recent data indicate that STAT3 expression can also inhibit inflammation in the colon through multiple but mainly tolerogenic mechanisms ( 26 ). Hence, strategies at increasing SOCS3 expression may be accompanied exacerbation of disease. Disregarding this complication in the clinical management of UC per se, our data do show that preventing SOCS3 promoter methylation may be the way forward for novel strategies aimed at avoiding the development of UC–CRC. Various DNMT inhibitors are currently in clinical testing and might be interesting to study the effect of these in, for instance, UC patients with PSC co-morbidity as such this way would be a novel rational strategy for dealing with this complication of UC.

Table I.

Patients’ characteristics

  Healthy control ( N = 9)   Active UC ( N = 27)   UC–CRC ( N = 11)   Sporadic CRC ( N = 14)  
Gender c 
   Male 15 
   Female 12 
Mean age, year ± standard deviation 39.6±17.8 42.0±15.3  (56.1±16.5) a  (68.3±12.2) a 
Mean duration of disease, year ± standard deviation – 11.9±9.6  (22.3±5.8) b NA 
Extent of disease 
   Extensive colitis – 
   Left-side colitis – 18 
   Proctitis – 
Therapy 
   Aminosalicylates – 14 – 
   Corticosteroids – 13 – 
   Immunosuppressives – 11 – 
   Biologicals – – 
   None – – 
  Healthy control ( N = 9)   Active UC ( N = 27)   UC–CRC ( N = 11)   Sporadic CRC ( N = 14)  
Gender c 
   Male 15 
   Female 12 
Mean age, year ± standard deviation 39.6±17.8 42.0±15.3  (56.1±16.5) a  (68.3±12.2) a 
Mean duration of disease, year ± standard deviation – 11.9±9.6  (22.3±5.8) b NA 
Extent of disease 
   Extensive colitis – 
   Left-side colitis – 18 
   Proctitis – 
Therapy 
   Aminosalicylates – 14 – 
   Corticosteroids – 13 – 
   Immunosuppressives – 11 – 
   Biologicals – – 
   None – – 

a The mean age of patients in sporadic CRC group (68.3±12.2) and UC–CRC group (56.1±16.5) was significantly larger than those of healthy control group and active UC group.

b The mean age duration of disease of patients in UC–CRC patients (22.3±5.8) was significantly longer than that of patients in other groups ( P < 0.01).

c No gender difference in each group.

Table II.

PCR conditions and primers

 Primer sequnece Annealing temperature (cycle) SOCS3 gene amplified Product size (bp) 
SOCS3 MSP 
M1F GTTTCGGTTTCGTACGTAGTTAGTC 55 (× 45, 20s) Exon 1 M = 180  
M1R TAAAATCCACAAAAAAACCTTCG    
U1F TTTTGGTTTTGTATGTAGTTAGTTG 56.5 (× 40, 30s)  U = 179  
U1R TAAAATCCACAAAAAAACCTTCAC    
M2F GGAGATTTTAGGTTTTCGGAATATTTC 56.5 (× 40, 30s) Intron before exon 2, STAT3 binding site M = 142  
M2R CCCCCGAAACTACCTAAACGCCG    
U2F GTTGGAGATTTTAGGTTTTTGGAATATTTT 56.5 (× 40, 30s)  U = 151  
U2R AAACCCCCAAAACTACCTAAACACCA    
M3F GTTTGATTCGTAGTTGGGTTTTC 55 (× 40, 25s) 5'-UTR before exon 1 M = 104  
M3R GACCTAAAAAACCTCCCGAT    
U3F GTTTGATTTGTAGTTGGGTTTTTG 55 (× 40, 25s)  U = 105  
U3R CAACCTAAAAAACCTCCCAAT    
SOCS3 pyrosequencing 
GTTTGGTTGTGGGGTAGTTTTATT 63 (× 45, 45s) 5'-UTR before exon 1 259 
AAGTGTGAATGAGAAGTTGGGG    
Q-PCR GAPDH 
GCATTGCCCTCAACGACCAC 60 (× 40, 15s)   
CCACCACCCTGTTGCTGTAG    
SOCS3      
TCTGTCGGAAGACCGTCAAC 60 (× 40, 15s)   
GGTCCAGGAACTCCCGAATG    
IL-6R      
TGTGCGTCGCCAGTAGTGTC 60 (× 40, 15s)   
CGGCAGTGACTGTGATGTTG    
DNMT1      
AGCCCGTAGAGTGGGAATGG 60 (× 40, 15s)   
TGAGGCAGGAGGGTCTCTTG    
 Primer sequnece Annealing temperature (cycle) SOCS3 gene amplified Product size (bp) 
SOCS3 MSP 
M1F GTTTCGGTTTCGTACGTAGTTAGTC 55 (× 45, 20s) Exon 1 M = 180  
M1R TAAAATCCACAAAAAAACCTTCG    
U1F TTTTGGTTTTGTATGTAGTTAGTTG 56.5 (× 40, 30s)  U = 179  
U1R TAAAATCCACAAAAAAACCTTCAC    
M2F GGAGATTTTAGGTTTTCGGAATATTTC 56.5 (× 40, 30s) Intron before exon 2, STAT3 binding site M = 142  
M2R CCCCCGAAACTACCTAAACGCCG    
U2F GTTGGAGATTTTAGGTTTTTGGAATATTTT 56.5 (× 40, 30s)  U = 151  
U2R AAACCCCCAAAACTACCTAAACACCA    
M3F GTTTGATTCGTAGTTGGGTTTTC 55 (× 40, 25s) 5'-UTR before exon 1 M = 104  
M3R GACCTAAAAAACCTCCCGAT    
U3F GTTTGATTTGTAGTTGGGTTTTTG 55 (× 40, 25s)  U = 105  
U3R CAACCTAAAAAACCTCCCAAT    
SOCS3 pyrosequencing 
GTTTGGTTGTGGGGTAGTTTTATT 63 (× 45, 45s) 5'-UTR before exon 1 259 
AAGTGTGAATGAGAAGTTGGGG    
Q-PCR GAPDH 
GCATTGCCCTCAACGACCAC 60 (× 40, 15s)   
CCACCACCCTGTTGCTGTAG    
SOCS3      
TCTGTCGGAAGACCGTCAAC 60 (× 40, 15s)   
GGTCCAGGAACTCCCGAATG    
IL-6R      
TGTGCGTCGCCAGTAGTGTC 60 (× 40, 15s)   
CGGCAGTGACTGTGATGTTG    
DNMT1      
AGCCCGTAGAGTGGGAATGG 60 (× 40, 15s)   
TGAGGCAGGAGGGTCTCTTG    

Supplementary material

Supplementary Figures S1S4 can be found at http://carcin.oxfordjournals.org/

Funding

Part of this research is supported by the State Scholarship Fund from the Chinese Scholarship Council (No 2007101714).

Acknowledgements

Conflict of Interest Statement : The authors report no conflict of interest.

Ethics approval: This study was conducted with the approval of the relevant institutional ethics board.

References

1.
Cho
J.H.
2008
The genetics and immunopathogenesis of inflammatory bowel disease
.
Nat. Rev. Immunol.
  ,
8
,
458
––
466
.
2.
Eaden
J.A.
et al
2001
The risk of colorectal cancer in ulcerative colitis: a meta-analysis
.
Gut
  ,
48
,
526
––
535
.
3.
Eaden
J.
2004
Review article: colorectal carcinoma and inflammatory bowel disease
.
Aliment. Pharmacol. Ther.
  ,
20
(Suppl. 4),
24
––
30
.
4.
Sartor
R.B.
2006
Mechanisms of disease: pathogenesis of Crohn's disease and ulcerative colitis
.
Nat. Clin. Pract. Gastroenterol. Hepatol.
  ,
3
,
390
––
407
.
5.
Schottelius
A.J.
et al
2006
Cytokines, NF-kappaB, microenvironment, intestinal inflammation and cancer
.
Cancer Treat. Res.
  ,
130
,
67
––
87
.
6.
Grivennikov
S.
et al
2009
IL-6 and stat3 are required for survival of intestinal epithelial cells and development of colitis-associated cancer
.
Cancer Cell
  ,
15
,
103
––
113
.
7.
Bollrath
J.
et al
2009
gp130-mediated Stat3 activation in enterocytes regulates cell survival and cell-cycle progression during colitis-associated tumorigenesis
.
Cancer Cell
  ,
15
,
91
––
102
.
8.
Li
Y.
et al
2010
Disease-related expression of the IL6/STAT3/SOCS3 signalling pathway in ulcerative colitis and ulcerative colitis-related carcinogenesis
.
Gut
  ,
59
,
227
––
235
.
9.
Wehbe
H.
et al
2006
Interleukin-6 contributes to growth in cholangiocarcinoma cells by aberrant promoter methylation and gene expression
.
Cancer Res
  .,
66
,
10517
––
10524
.
10.
Braconi
C.
et al
2010
MicroRNA-dependent regulation of DNA methyltransferase-1 and tumor suppressor gene expression by interleukin-6 in human malignant cholangiocytes
.
Hepatology
  ,
51
,
881
––
890
.
11.
Foran
E.
et al
2010
Upregulation of DNA methyltransferase-mediated gene silencing, anchorage-independent growth, and migration of colon cancer cells by interleukin-6
.
Mol. Cancer Res.
  ,
8
,
471
––
481
.
12.
Lennard-Jones
J.E.
1989
Classification of inflammatory bowel disease
.
Scand. J. Gastroenterol. Suppl.
  ,
170
,
2
––
6
; discussion 16-19.
13.
Weber
A.
et al
2005
SOCS-3 is frequently methylated in head and neck squamous cell carcinoma and its precursor lesions and causes growth inhibition
.
Oncogene
  ,
24
,
6699
––
6708
.
14.
Wang
M.
et al
2007
Deficiency of TNFR1 protects myocardium through SOCS3 and IL-6 but not p38 MAPK or IL-1beta
.
Am. J. Physiol. Heart Circ. Physiol.
  ,
292
,
H1694
––
1699
.
15.
Yin
Y.
et al
2010
Blocking effects of siRNA on VEGF expression in human colorectal cancer cells
.
World J. Gastroenterol.
  ,
16
,
1086
––
1092
.
16.
Zhu
Y.M.
et al
2007
Expression of human DNA methyltransferase 1 in colorectal cancer tissues and their corresponding distant normal tissues
.
Int. J. Colorectal Dis.
  ,
22
,
661
––
666
.
17.
Kodach
L.L.
et al
2010
Statins augment the chemosensitivity of colorectal cancer cells inducing epigenetic reprogramming and reducing colorectal cancer cell ‘stemness’ via the bone morphogenetic protein pathway
.
Gut
  ,
60
,
1544
––
1553
.
18.
Kim
S.
et al
2009
Carcinoma-produced factors activate myeloid cells through TLR2 to stimulate metastasis
.
Nature
  ,
457
,
102
––
106
.
19.
Neurath
M.F.
et al
2010
IL-6 signaling in autoimmunity, chronic inflammation and inflammation-associated cancer
.
Cytokine Growth Factor Rev
  .,
22
,
83
––
89
.
20.
Fantini
M.C.
et al
2008
Cytokines: from gut inflammation to colorectal cancer
.
Curr. Drug Targets
  ,
9
,
375
––
380
.
21.
Suzuki
A.
et al
2001
CIS3/SOCS3/SSI3 plays a negative regulatory role in STAT3 activation and intestinal inflammation
.
J. Exp. Med.
  ,
193
,
471
––
481
.
22.
Bocklandt
S.
et al
2011
Epigenetic predictor of age
.
PLoS One
  ,
6
,
e14821
.
23.
Liu
L.
et al
2005
Epigenetic alterations in neuroendocrine tumors: methylation of RAS-association domain family 1, isoform A and p16 genes are associated with metastasis
.
Mod. Pathol.
  ,
18
,
1632
––
1640
.
24.
Schneider
M.R.
et al
2000
Interleukin-6 stimulates clonogenic growth of primary and metastatic human colon carcinoma cells
.
Cancer Lett
  .,
151
,
31
––
38
.
25.
Rigby
R.J.
et al
2007
Suppressor of cytokine signaling 3 (SOCS3) limits damage-induced crypt hyper-proliferation and inflammation-associated tumorigenesis in the colon
.
Oncogene
  ,
26
,
4833
––
4841
.
26.
Li
Y.
et al
2011
New insights into the role of STAT3 in IBD
.
Inflamm. Bowel Dis.
  , 18, 1177–1183.

Abbreviations

    Abbreviations
  • CRC

    colorectal cancer

  • DNMT1

    DNA methyltransferase- 1interleukin

  • IBD

    inflammatory bowel disease

  • STAT3

    signal transducer and activator of transcription 3

  • SOCS3

    suppressor of cytokine signaling 3

  • MSP

    methylation-specific PCR

  • UC

    ulcerative colitis.