Suppressor of cytokine signaling 3 (SOCS3) limits damage ... - Nature

Oncogene (2007) 26, 4833–4841

& 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

Suppressor of cytokine signaling 3 (SOCS3) limits damage-induced crypt hyper-proliferation and inflammation-associated tumorigenesis in the colon RJ Rigby1, JG Simmons1, CJ Greenhalgh2, WS Alexander3 and PK Lund1 1 Department of Cell and Molecular Physiology, University of North Carolina, Chapel Hill, NC, USA; 2Cancer and Haematology Division, Walter and Eliza Hall Institute of Medical Research, Victoria, Australia and 3Division of Autoimmunity and Transplantation, Walter and Eliza Hall Institute of Medical Research, Victoria, Australia

Intestinal injury or chronic inflammation induce cytokines that promote crypt regeneration and mucosal repair. If excessive or prolonged, such mechanisms may increase colon cancer risk. Factors that terminate or limit cytokine action in intestinal epithelial cells (IEC) may protect against crypt hyperplasia and neoplasia. We hypothesized that suppressor of cytokine signaling-3 (SOCS3) is such a factor. Mice with Vilin-promoter/Cre-recombinase (VC)mediated IEC-specific SOCS3 gene disruption (VC/HO), WT/HO littermates with floxed but intact SOCS3 genes and VC/WT mice were studied. Colon was examined after acute dextran sodium sulfate (DSS)-induced mucosal injury or after azoxymethane (AOM) and chronic DSS. Signaling pathways were examined in colon, cultured IEC or colon cancer cell lines. VC/HO mice showed no basal phenotype. After acute DSS, VC/HO exhibited enhanced crypt proliferation and crypt hyperplasia and reduced transforming growth factor (TGF) b expression in colon. Inflammation and mucosal damage were similar across genotypes. Following AOM/DSS, VC/HO mice had increased size, number and load of colonic tumors and increased STAT3 and nuclear factor-kappa B (NF-jB) activation in colon. In vitro, SOCS3 overexpression reduced proliferation, IL-6-mediated STAT3 activation and tumor necrosis factor (TNF) a-mediated NF-jB activation. We conclude that cytokine induction of SOCS3 normally provides an intrinsic mechanism to limit injury-induced crypt hyperproliferation and inflammationassociated colon cancer by regulating both STAT3 and NF-jB pathways. Oncogene (2007) 26, 4833–4841; doi:10.1038/sj.onc.1210286; published online 12 February 2007 Keywords: SOCS3; colon; cancer; mouse model

Correspondence: Dr RJ Rigby, Department of Cell and Molecular Physiology, University of North Carolina at Chapel Hill, Rm 6336, MBRB, CB.7545, 103, Mason Farm Road, Chapel Hill, NC 27599, USA. E-mail: [email protected] Received 8 September 2006; revised 7 December 2006; accepted 7 December 2006; published online 12 February 2007

Introduction Ulcerative colitis increases risk of colon cancer and mortality with colitis-associated cancer (CAC) and is higher than in sporadic colorectal cancer (Itzkowitz, 2003; Munkholm, 2003). Considerable evidence implicates cytokines, particularly interleukin-6 (IL-6) in CAC. IL-6 directly stimulates proliferation (Schneider et al., 2000) of some colon cancer cell lines. Plasma levels of IL-6 correlate with tumor size and prognosis in colon cancer patients (Chung and Chang, 2003). A mutation in the IL-6 receptor is linked to increased colon cancer risk in humans (Landi et al., 2003). Signaling pathways linked to IL-6 also play a role in CAC. Typically, IL-6 binds to a membrane complex containing the ligand binding subunit, IL-6Ra and a signal-transducing subunit gp130. Ligand binding causes gp130 dimerization and activation of Janus kinases (JAK), JAK1, JAK2 and Tyk2, of which JAK1 appears to play a major role (Heinrich et al., 2003). JAK activation of gp130 then mediates downstream responses by activating signal transducers and activators of transcription (STAT), particularly STAT3, as well as Src homology domain 2 containing protein tyrosine phosphatase (SHP2), a docking protein that activates mitogenactivated protein kinase/extracellular signal-regulated kinase (ERK) (Tebbutt et al., 2002; Heinrich et al., 2003). Mice with mutated STAT3-binding sites on gp130 showed impaired mucosal healing after acute injury with DSS, associated with enhanced ERK activation (Tebbutt et al., 2002). Mice with a mutated SHP2 binding site on gp130 developed spontaneous gastric tumors associated with increased STAT3 and decreased ERK activation (Judd et al., 2004). Mice with interferon gamma-inducible disruption of STAT3 genes in both macrophages and intestinal epithelial cells (IEC) develop enterocolitis (Alonzi et al., 2004). However, STAT3 phosphorylation is constitutively activated in some colon cancer cell lines and in these model systems, STAT3 inhibitors reduce growth and promote apoptosis (Nakamura et al., 2004; Rivat et al., 2004). Recent studies in the AOM/DSS model of colon carcinogenesis indicate a key role of IL-6 and STAT3 in promoting tumor formation and growth (Becker et al., 2004). The apparently dual roles of IL-6 and STAT3 in protecting against colitis, but promoting colon tumors, support a

SOCS3 limits tumor growth RJ Rigby et al

4834

Results SOCS3 overexpression limits proliferation of IEC and colon cancer cell lines IEC-6 and Caco-2 intestinal epithelial cell lines were used to test the effects of SOCS3 overexpression on spontaneous cell proliferation. Cells were transfected with empty vector or SOCS3 expression plasmid and cell number was evaluated over a 3-day culture period, in serum. The numbers of empty vector-transfected cells increased over the 3-day culture period, whereas cells transfected with SOCS3 expression plasmid showed no increase in cell number (Figure 1). Generation of IEC-specific SOCS3 knockout mice Mice homozygous for pLox-modified SOCS3 alleles (WT/HO) (Croker et al., 2003) were mated with mice hemizygous for Villin-promoter/Cre-recombinase transgene (VC). To demonstrate IEC-specific SOCS3 deletion, genomic DNA was extracted from isolated IEC from small intestine and colon, or liver and kidney of Oncogene

Fold increase in cell #

a

IEC-6

Empty vector

4

SOCS3

3

p≤0.05

2 1 0 day 1

b Fold increase in cell #

concept that factors which limit the magnitude or duration of STAT3 activation in IEC may favor normalization of crypt homeostasis after injury and protect against CAC. We hypothesize that SOCS3 is such a factor. SOCS3 belongs to a family of structurally related proteins discovered as negative feedback inhibitors of cytokine action (Greenhalgh et al., 2002). IL-6 potently induces SOCS3 in macrophages, T cells and hepatocytes and SOCS3 limits IL-6-mediated STAT3 activation and induction of acute-phase response genes in these cell types (Campbell et al., 2001; Croker et al., 2003; Chen et al., 2006). In humans, dysregulated JAK-STAT signaling and methylation-mediated silencing of SOCS1, SOCS2 or SOCS3 genes occurs in tumors of multiple organs (He et al., 2003; Oshimo et al., 2004; Sutherland et al., 2004). In human lung cancer cells, restoring SOCS3 expression reduced STAT3 activation, induced apoptosis and decreased tumor growth (He et al., 2003). The role of SOCS3 in growth or apoptosis of IEC or CAC has not yet been defined. In Caco2 cells, IL-6 induced SOCS3, and SOCS3 overexpression decreased IL-6 induced nuclear factor-kappa B (NF-kB) activation (Wang et al., 2003). SOCS3 is known to be upregulated in animal models of inflammatory bowel disease (IBD) and in human IBD patients, and SOCS3 expression has been reported in epithelial cells, macrophages and T cells (Miller et al., 2000; Suzuki et al., 2001; Mizoguchi et al., 2003). Expression of SOCS3 in multiple cell types during intestinal inflammation, together with the complex roles of STAT3 in preserving mucosal integrity yet promoting neoplasia, underscores the importance of defining cell-type specific roles of SOCS3. Here, we develop mice with IEC-specific SOCS3 gene deletion and demonstrate that SOCS3 plays an important role in limiting inflammation-induced crypt hyperproliferation and tumor growth in the colon.

day 2

day 3

Caco2

2 1.5 1 0.5 0 day 1

day 2

day 3

Figure 1 Fold increase in cell number in IEC-6 and Caco2 cells grown in serum at 1–3 days after transfection of SOCS3 expression plasmid or empty vector. SOCS3 overexpression reduced cell proliferation (% ¼ Po0.05 vs empty vector).

VC/HO and WT/HO littermates. Following polymerase chain reaction (PCR) with primers flanking the pLoxmodified region of the SOCS3 gene a 740-bp fragment corresponding to a pLox-modified, but otherwise intact SOCS3 gene, was detected in liver and kidney of VC/HO mice and in IEC, liver and kidney of WT/HO mice. Small intestine and colon IEC from VC/HO mice showed only a 368-bp fragment corresponding in size to the Cre-recombinase-modified SOCS3 allele lacking SOCS3 coding sequence (Figure 2a). Figure 2b shows reverse transcription–PCR (RT–PCR) quantification of SOCS3 messenger ribonucleic acid (mRNA) in total RNA prepared from isolated colonic or small intestine IEC, kidney and liver of VC/HO or WT/HO mice. SOCS3 mRNA was undetectable in IEC from colon or small intestine of VC/HO mice but was detected in IEC from WT/HO mice and was similarly expressed in kidney and liver of VC/HO and WT/HO mice. Hydroxymethylbilane synthase (HMBS) mRNA was similar in all samples. IEC-specific SOCS3 deletion enhanced crypt proliferation and hyperplasia following injury by acute DSS Histology was performed on colon of VC/HO and WT/ HO littermates given water, or at 3 days after DSS treatment (DSS þ 3). We focused on DSS þ 3 because we have previously established that this is a time of early mucosal repair and maximal SOCS3 expression (Miller et al., 2000; Williams et al., 2001). VC/HO and WT/HO mice given water alone showed no significant difference


4835

a

WT/HO C

SI

L K

VC/HO C SI

L K

floxed SOCS3 740bp excised 368bp

Socs3 mRNA

b

2 WT/HO VC/HO 1

0 COLON IEC

SMALL INTESTINE IEC

KIDNEY

LIVER

COLON IEC

SMALL INTESTINE IEC

KIDNEY

LIVER

HMBS mRNA

2

1

0

Figure 2 (a) PCR for floxed but intact SOCS3 gene (740 bp) or Cre-excised gene (368 bp) on DNA from IEC prepared from colon (C) or small intestine (SI), liver (L) or kidney (K). (b) SOCS3 mRNA and housekeeping gene (HMBS) mRNA abundance in isolated from IEC from colon and small intestine and kidney or liver of WT/HO and VC/HO animals. SOCS3 mRNA is ablated or dramatically reduced in IEC, but not kidney or liver of VC/HO animals compared to WT/HO controls.

in histology, crypt depth or 5-bromodeoxyuridine (BrdU) incorporation in the distal colon (Figure 3). However, after DSS, VC/HO showed pronounced crypt hyperplasia and dramatically increased BrdU incorporation into crypts in distal colon relative to WT/HO mice (Po0.001; Figure 3b). A significant increase in crypt depth was seen in both genotypes following DSS treatment (Po0.001) but the increase in crypt depth was more dramatic in VC/HO mice (390723 mm) than WT/HO (27371 mm, Po0.001). These data indicate that SOCS3 expressed in IEC normally limits crypt hyperproliferation and hyperplasia after mucosal injury. Total colitis scores did not differ significantly in water- or DSS-treated WT/HO and VC/HO colon (Figure 3d), indicating that epithelial SOCS3 deletion specifically affects epithelial repair after injury and not susceptibility to initial damage. This is consistent with only low level expression of SOCS3 in uninjured or noninflamed intestine (Miller et al., 2000). The more variable colitis score in VC/HO water controls may indicate increased susceptibility to mild inflammation or mucosal damage in a subset of mice. Counting apoptotic cells in hematoxylin and eosin (H&E) stained sections of colon of DSS-treated WT/HO and VC/HO mice

revealed a non-significant trend (P ¼ 0.2) towards reduced crypt apoptosis in the colon of DSS-treated VC/HO compared with WT/HO mice (Figure 3e). Northern blot hybridization revealed reduced TGFb mRNA (0.5670.1-fold, Pp0.03) in colon from DSStreated VC/HO compared with WT/HO mice (Figure 3f). Increased tumor size and number in VC/HO compared with WT/HO and VC/WT following AOM/DSS treatment We analysed colon tumors in VC/HO, WT/HO and VC/ WT mice after AOM/DSS treatment. Tumor size and tumor number were both significantly increased in VC/ HO mice vs other genotypes (Figure 4), which translated into a 2.370.4-fold increase in total tumor burden vs WT/HO (Pp0.03), and a 4.070.7-fold increase in tumor burden compared with VC/WT (Pp0.01). The more pronounced difference between VC/HO and VC/ WT compared with WT/HO is consistent with previous reports that flox modification of SOCS3 alleles leads to a small reduction in SOCS3 expression vs mice with unmodified SOCS3 genes (Croker et al., 2003). Increased STAT3, NF-kB and STAT1 activation in VC/ HO vs VC/WT and WT/HO mucosa following AOM/ DSS treatment STAT3, NF-kB and STAT1 activation were assessed by electromobility shift assay (EMSA) on nuclear extracts from colonic mucosa at 80 days after AOM/DSS or no treatment. Low levels of basal STAT3 and NF-kB DNA-binding activity were detected in colon of all genotypes. AOM/DSS-treated VC/HO mice showed increased STAT3, NF-kB and STAT1 DNA-binding activity in colonic mucosa compared with other genotypes. Immunohistochemistry for phosphorylated STAT3 (pSTAT3) revealed more nuclear pSTAT3 in hyperplastic or dysplastic IEC from AOM/DSS-treated VC/HO, compared with VC/WT and HO/WT colon (Figure 5b). Increased STAT3 activation in VC/HO colon following AOM/DSS treatment was confirmed by Western immunoblotting for pSTAT3 (Figure 5c). In accordance with increased NF-kB-binding activity, VC/ HO colon exhibited increased levels of IKKb, but no difference in phosphorylated or total ERK across genotypes. No significant difference in TGFb mRNA was detected in colon of AOM/DSS-treated VC/HO and other genotypes (Figure 5d). SOCS3 overexpression inhibits IL-6/STAT3 and TNFa/ NF-kB pathways in cancer cell lines IL-6- and TNFa-induced activation of STAT3 or NF-kB were examined in SW480 cells, a colon cancer cell line that we have found to express little or no endogenous SOCS3. Cells were infected with empty or SOCS3-expressing adenovirus (Ad-empty or AdSOCS3). Western blots showed IL-6-dependent tyrosine phosphorylation of STAT3 in Ad-empty infected cells and this was inhibited in Ad-SOCS3 infected cells (Figure 6a). Treatment of SW480 cells with a molar Oncogene


4836

a

c 25

500

WT/HO

20

400

VC/HO

crypt depth (µm)

15 10 5

100 0

30

DSS+3

p=0.7

water

20 15 10 5

DSS+3

e % apoptotic cells

25 Colitis score

200

0 water

d

p≤0.05

300

f 0.6

1.2

0.5

1

0.4

p=0.2

0.3 0.2 0.1

water

DSS+3

0.8 0.6 0.4 0.2

0

0

TGFβ mRNA

# BrdU positive cells/crypt

b

0 WT/HO

VC/HO

DSS+3

Figure 3 (a) Photomicrographs of H&E and BrdU-stained sections of distal colon from WT/HO and VC/HO mice at DSS þ 3 or water controls. Histograms compare WT/HO (black) and VC/HO (white) at 3 days after DSS or water controls for (b) number of BrdU-positive cells per crypt; (c) mean crypt depth; (d) total colitis score; (e) apoptotic cells per crypt; (f) TGFb mRNA.

excess of sIL-6R, pre-mixed with IL-6, to elicit IL-6 trans-signaling, induced STAT3 phosphorylation and Ad-SOCS3 infected cells showed a diminished response (Figure 6b). TNFa, but not IL-6, induced phosphorylation of NF-kB p65 in empty vector infected cells and this effect was inhibited in SOCS3 infected cells (Figure 6c).

Discussion Previous studies have shown that SOCS3 exerts negative feedback effects on IL-6 signaling in macrophages, hepatocytes and Caco2 colon cancer cells (Croker et al., Oncogene

2003; Wang et al., 2003; Yasukawa et al., 2003; Ogata et al., 2006b). We demonstrate here that SOCS3 overexpression reduces basal proliferation of IEC-6, and Caco2 cells, providing new evidence that SOCS3 can exert anti-proliferative effects in non-transformed IEC and colon cancer cell lines. SOCS3 is upregulated in multiple cell types in addition to IEC during mucosal injury or inflammation (Miller et al., 2000; Suzuki et al., 2001). We therefore developed mice with IEC-specific SOCS3 gene deletion (VC/HO) to define specifically the functional significance of IEC-SOCS3 in vivo. Consistent with observations that SOCS3 expression is low in the colon in the basal state, lack of SOCS3 in IEC did not result in discernible phenotypic changes in the colon or


4837

a VC/WT

VC/HO

WT/HO

p

0.05

25

60

VC/WT

2

20

50

VC/HO

1.5 1 0.5

Tumor burden

2.5 Tumor counts

Tumor score

b

15 10 5

SIZE

WT/HO

30 20 10 0

0

0

40

NUMBER

LOAD

Figure 4 (a) Representative photographs of tumors in distal colon of VC/WT, VC/HO and WT/HO following AOM/DSS treatment. Note larger tumors in VC/HO colon compared with WT controls. (b) Histograms show average tumor size, number and load in VC/ WT (n ¼ 5), VC/HO (n ¼ 7) and WT/HO (n ¼ 7) colon following AOM/DSS treatment. Note the significant increase in tumor size, number and load in VC/HO compared with VC/WT or WT/HO colon.

in basal crypt proliferation. Dramatically enhanced crypt proliferation and crypt hyperplasia in VC/HO during recovery from acute DSS-induced mucosal injury provide evidence that SOCS3 normally limits injuryinduced crypt proliferation and hyperplasia in colon. Furthermore, these anti-proliferative effects of SOCS3 appear to protect against development and growth of colonic tumors during chronic intestinal inflammation, because VC/HO mice show enhanced tumor size, number and burden after AOM/DSS. Our findings in the AOM/DSS model of CAC demonstrate a tumor suppressor role of SOCS3 during chronic colitis and are consistent with a recent report that hepatocyte-specific SOCS3 deletion promotes hepatitis C associated hepatocellular carcinoma (Ogata et al., 2006b). Promoter hypermethylation and silencing of SOCS3 genes has been reported in liver (Niwa et al., 2005), lung (He et al., 2003) and squamous cell carcinoma of the head and neck (Weber et al., 2005). Our observations suggest that it will be of interest to assess if epigenetic silencing of SOCS3 is associated with susceptibility to colitisassociated crypt hyperplasia, dysplasia or neoplasia in humans. Phenotypic effects of IEC-SOCS3 deletion on colon tumorigenesis were associated with enhanced activation of STAT3, NF-kB and STAT1, indicating that during chronic inflammation, SOCS3 normally limits activation of multiple signaling pathways implicated in risk of CAC. Effects of SOCS3 deletion on STAT3 activation are consistent with data in other cell types, suggesting that the major role of SOCS3 is to limit hyperactivation

of STAT3 by proinflammatory cytokines, particularly IL-6 (Croker et al., 2003; Ogata et al., 2006a). Findings that SOCS3 effectively inhibits STAT3 activation by IL-6 trans-signaling, as well as by IL-6 alone is relevant to evidence for a major role of IL-6 trans-signaling in CAC (Becker et al., 2004). Mice with mutations in the SHP2 binding site on gp130 exhibit constitutively active STAT3 and develop spontaneous gastric, but not colonic tumors (Judd et al., 2004). In contrast, IECselective SOCS3 deletion led to STAT3 hyperactivation and enhanced colon tumorigenesis only following injury and inflammation and not at baseline. The differences between gp130 and SOCS3 mutants may reflect differences in basal STAT3 activation in stomach and colon, a role of macrophage STAT3 hyperactivation in gp130 mutants (Judd et al., 2006) and not in IECselective SOCS3 mutants and potentially, an effect of STAT3 hyperactivation on the trefoil factor tumor suppressors in stomach (Judd et al., 2004) and not colon epithelium. Future evaluation of the effects of macrophage-specific deletion of SOCS3 on CAC, as well as stomach phenotype or trefoil expression in stomach or colon of our existing IEC-selective SOCS3 deletion mutants could be important in elucidating different mechanisms of inflammation-associated gastric and colon cancer. Enhanced NF-kB activation in colon of AOM/DSS-treated VC/HO mice strengthens a small, but growing, body of evidence that SOCS3 can limit NF-kB pathways as well as STAT pathways more traditionally linked to SOCS. Increased IKKb in VC/HO colon suggests that SOCS3 normally limits Oncogene


4838 VC/WT

VC/HO

WT/HO

VC/WT

VC/HO

AOM/DSS

WATER

WT/HO

a

Cold oligo

STAT3

NF-kB

STAT1

b

VC/WT

c

WT/HO

VC/WT WT/HO VC/HO

d

pY-ERK

WT/HO TGFβ mRNA

IKKB

VC/WT

1.5

pY-STAT3 Total STAT3

VC/HO

VC/HO

1

0.5

Total ERK 0 AOM/DSS Figure 5 (a) EMSA on pooled nuclear extracts from VC/WT (n ¼ 3), WT/HO (n ¼ 3) and VC/HO (n ¼ 3) colonic mucosa. Note the increase in STAT3, NF-kB and STAT1 activation, particularly in VC/HO colon following AOM/DSS treatment. Data are representative of two individual experiments. In AOM/DSS-treated VC/HO mice STAT3 DNA binding was 1.570.1-fold higher than in other genotypes, NF-kB (1.570.1-fold higher and STAT1 1.770.4-fold higher). Addition of excess unlabeled (cold) oligomer blocked DNA binding, indicating specificity of DNA:protein interaction. (b) Immunostaining of pSTAT3 in areas of hyperplasia and dysplasia in VC/WT, WT/HO and VC/HO colon following AOM/DSS treatment. Note the more pronounced nuclear staining of epithelial cells in VC/HO colon compared with VC/WT and WT/HO littermates. (c) Western immunoblotting on mucosal extracts from VC/WT, WT/HO and VC/HO colon following AOM/DSS treatment. VC/HO colon showed increased pY-STAT3 and not total STAT3 (1.570.1-fold vs VC/WT) and increased IKKb (1.7-fold vs VC/WT), but no difference in phosphorylated or total ERK. (d) TGFb mRNA abundance in colon of VC/WT, WT/HO and VC/HO colon following AOM/DSS treatment. No significant difference in TGFb expression was seen between genotypes.

expression of a kinase that phosphorylates IkB and targets associated NF-kB for ubiquitin-dependant degradation. Findings that SOCS3 overexpression limits the rapid activation of NF-kB by TNFa in colon cancer cells provide novel evidence that SOCS3 may directly impact on TNFa action as well as that of IL-6 and cytokines signaling through gp130-linked receptors. This possibility is supported by two reports that TNFa can activate JAKs and that a related SOCS, SOCS1, can limit JAK activation (Miscia et al., 2002; Kimura et al., 2004). Further studies of the interactions between TNFa/JAK and SOCS3 are warranted. Enhanced STAT1, as well as STAT3, activation in VC/HO colon Oncogene

during chronic inflammation adds to the evidence that SOCS3 can limit STAT1 in vivo (Croker et al., 2003) and is consistent with recent studies implicating hyperactivation of STAT1 in spontaneous colon tumors (Hanada et al., 2006). A model of global hepatic deletion of SOCS3, using adenovirus-Cre, linked enhanced liver fibrosis to increased STAT3 activation and enhanced TGFb expression in regions of cirrhosis (Ogata et al., 2006a). In contrast, no significant upregulation of TGFb was seen in colon of VC/HO despite the enhanced STAT3 activation in the AOM/DSS model. Thus, effects of STAT3 on TGFb may therefore be cell type or context-


4839

a

Empty -

IL-6:

SOCS3 +

+


b

sIL-6R + IL-6 (ng/ul):

Empty 0

5

10

0

SOCS-3 5 10


c

Empty Treatment:

No Tx

IL-6 TNF

SOCS3

studied at 3 days after recovery on water (DSS þ 3). Mice given water alone were used as controls. For AOM/DSS treatment, sex-matched WT/HO and VC/HO littermate pairs and VC/ WT mice were given a single intraperitoneal (i.p.) injection of 10 mg/kg AOM. Seven days later, mice were administered 2.5% DSS for 5 days, followed by 14 days recovery on water and this process was repeated three times. Mice were studied 80 days following AOM injection. Sample collection Entire colon was collected and tissue was either fixed in formalin for histology and/or tumor counting or frozen. Tumors were counted and scored according to size (score of 1p1, 2 ¼ 1–2, 3 ¼ 2–3, 4X3 mm). Nuclear extracts or RNA were prepared from whole tissue, mucosal scrapes or isolated IEC. IEC were isolated using dithiothreitol/ethylenediaminetetraacetic acid methods described previously (Miller et al., 2004).

No Tx IL-6 TNF

pY-p65 Total p65 Figure 6 Western blots for tyrosine-phosphorylated and total STAT3 or NF-kB p65 in SW480 cells infected with either empty or SOCS3 adenovirus. SOCS3 inhibited both classical (a) and transIL-6 signaling-induced (b) STAT3 phosphorylation and also TNFa-induced p65 phosphorylation (c).

specific. In conclusion, our results demonstrate a role of intestinal epithelial SOCS3 in limiting injury and inflammation associated crypt hyperproliferation and tumorigenesis by limiting multiple signaling pathways linked to intestinal cancer.

Materials and methods Animals Mice homozygous for pLox modification of SOCS3 genes (WT/HO) were derived on the C57BL6 background as described previously (Croker et al., 2003) and mated with C57BL6 mice hemizygous for VC transgene (Jackson Labs, Bar Harbor, ME, USA). VC and WT mice heterozygous for the floxed SOCS3 allele were then crossed to derive study animals. Sex-matched, 6–8-week old, WT/HO or VC/HO littermates were studied. VC/WT, were used as additional controls in selected experiments. Genotyping used the following primers. WT or floxed SOCS3 alleles: 50 -GAGTTTTCTCTGGGCGTCCTCCTA-30 and 50 -TGG TACTCGCTTTTGGAGCTGAA-30 ; VC transgene: 50 -GTG TGGGACAGAGAACAAACCG-30 and 50 -TGCGAACCTC ATCACTCGTTGC-30 ; floxed or Cre-excised SOCS3 alleles: 50 -ACGTCTGTGATGCTTTGCTG-30 and 50 -TCTTGTGTC TCTCCCCATCC-30 , yield a 740- or a 368-bp fragment, respectively. DSS and AOM/DSS treatment Sex-matched WT/HO and VC/HO littermate pairs were given 3% DSS (TDB Consultancy, Sweden) for 5 days. Mice were

Cell lines IEC-6, Caco-2 (ATCC, MD, USA) and SW480 cells were propagated in Dulbecco’s modified Eagle’s medium supplemented with 10% heat inactivated fetal bovine serum, 50 U/ml penicillin, 50 mg/ml streptomycin and 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES). Plasmid or adenoviral vectors expressing SOCS3 or empty vector controls were used for SOCS3 overexpression. Plasmids were used for proliferation studies in IEC-6 and Caco-2 cells to avoid the problem that adenoviral vectors alone can impact on cell proliferation. Plasmid pBIG2i expressing human SOCS-3 and empty vector were provided by Dr Richard Furlanetto and 2–5 mg transfected into 1 106 subconfluent cells using nucleofection (Amaxa Biosystems, Germany). Total cell number was determined using the CyQUANT cell proliferation assay kit (Molecular Probes, Carlsbad, CA, USA). For signaling experiments in SW480 cells adenovirus expressing the hSOCS3 sequence excised from pBIG2i or empty adenovirus were used. Infection was performed at 100 multiplicity of infection and added to cells cultured in serumfree medium at 24–48 h before experiments. Cells were treated with rhIL-6 or rhTNFa (Peprotech, NJ, USA) at concentrations ranging from 0 to 20 ng/ml for 30 min. For transsignaling experiments, SW480 cells were treated with molar excess (100 mg/ml) of sIL-6R (Peprotech, NJ, USA) premixed with IL-6. Transfection or infection efficiency was monitored by cotransfecting with green fluorescent protein and by level of SOCS3 mRNA expression (data not shown).

Real-time quantitative PCR Real-time qPCR was performed on the Light Cycler (Roche, Switzerland) using primers SOCS3 F: 50 -GCACAAGC ACAAAAATCCAGC-30 , R: 50 -AGAAGCCAATCTGCCCC TG-30 and FAM-labeled probes 50 -FAM-CCAACGGTCGG TAGCTCCCAGTGA-TAMRA-30 designed by Primer Express software (Applied Biosystems, Foster City, CA, USA). Results were normalized to the housekeeping gene HMBS using primers 50 -TGTGTTGCACGATCCTGAAAC-30 and 50 -CTCCTTCCAGGTGCCTCAGAA-30 and 50 -FAM-TTCG CTGCATTGCTGAAAGGG-TAMRA-30 probe. Cycling conditions were 501C for 2 min, 951C for 2 min, 45 941C, 5 s and 601C, 20 s. Cycle threshold values for SOCS3 and HMBS in the same sample were analysed using RelQuant software (Roche Diagnostics, Switzerland). Oncogene


4840 Histological analyses Coded H&E-stained sections from three different regions of colon (proximal, mid and distal) were scored for crypt damage, inflammation severity and extent as detailed in Williams et al. (2001). The depth of well-oriented crypts (at least 20 per section) was measured in H&E-stained sections to quantify crypt hyperplasia (Ulshen et al., 1993). BrdU (200 mg/kg) was administered i.p. 90 min before anesthesia and BrdU antibody (Zymed, San Fransisco, CA, USA) used to stain paraffinembedded sections. The number of BrdU-labeled cells per crypt was evaluated in 10–20 well-oriented crypts per section. The number of apoptotic bodies were counted on H&E stained sections, and expressed as a percentage of total cells per crypt. Electro mobility shift assay EMSA was performed on nuclear extracts using consensus double-stranded oligomers for STAT3 (50 -GATCCTTCTGG GAATTCCTAGATC-30 ), NF-kB (50 -AGTTGAGGCGACT TTCCCAGGC-30 ) and STAT-1 (50 -CATGTTATGCATAT TGGAGTAAGTG-30 ) all from Santa Cruz (Santa Cruz, CA, USA). Immunohistochemistry IHC for activated STAT3 was carried out on paraffinembedded sections using anti-phospho-STAT3 (Cell signaling #9134) and secondary antibodies conjugated to horseradish peroxidase. Northern blots TGFb mRNA abundance was assessed by standard northern hybridization conditions on 30 mg total RNA using 32P-labeled

riboprobe generated from primers F: 30 -AGCCCGAAGCG GACTACTAT-50 , R: 50 -AGCCCTGTATTCCGTCTCCT-30 and T7: TAATACGACTCACTATAGGGAGCCCTGTAT TCCGTCTCCT-30 . GAPDH mRNA was used as loading control. Western blot analysis Cells or tissues were solubilized in sodium dodecyl sulfate (SDS) buffer and equivalent amounts of protein size-fractionated SDS-polyacrylamide gels. Immunoblotting used primary antibodies from Santa Cruz: anti-pY-STAT3 (SC-8059), antitotal STAT3 (SC-482); anti-total ERK (SC-93), anti-pERK (SC-7383), b-actin (SC-47778) or IKKb (Cell signaling #2684). Blots were probed with IRDye-800 and AlexaFluor-680conjugated secondary antibodies to allow detection and quantification by the Odyssey imaging system (LICOR, NE, USA). Acknowledgements We thank Kirk McNaughton for assistance with histology and immunohistochemistry, Dr Daniel Meechan for help with realtime PCR, Victoria Newton for assistance with BrdU cell counting and Eileen Hoyt for guidance with EMSA assays. This work was supported by a senior research award and postdoctoral fellowship from the Crohns’ and Colitis foundation of America. Gene therapy and imaging core facilities of the Center for Gastrointestinal Biology and Disease (#P30 DK34987) assisted this work. We thank Dr Richard Furlanetto and Dr Robert Coffrey for provision of plasmids and cells.

References Alonzi T, Newton IP, Bryce PJ, Di Carlo E, Lattanzio G, Tripodi M et al. (2004). Induced somatic inactivation of STAT3 in mice triggers the development of a fulminant form of enterocolitis. Cytokine 26: 45–56. Becker C, Fantini MC, Schramm C, Lehr HA, Wirtz S, Nikolaev A et al. (2004). TGF-beta suppresses tumor progression in colon cancer by inhibition of IL-6 transsignaling. Immunity 21: 491–501. Campbell JS, Prichard L, Schaper F, Schmitz J, StephensonFamy A, Rosenfeld ME et al. (2001). Expression of suppressors of cytokine signaling during liver regeneration. J Clin Invest 107: 1285–1292. Chen Z, Laurence A, Kanno Y, Pacher-Zavisin M, Zhu BM, Tato C et al. (2006). Selective regulatory function of Socs3 in the formation of IL-17-secreting T cells. Proc Natl Acad Sci USA 103: 8137–8142. Chung YC, Chang YF. (2003). Serum interleukin-6 levels reflect the disease status of colorectal cancer. J Surg Oncol 83: 222–226. Croker BA, Krebs DL, Zhang JG, Wormald S, Willson TA, Stanley EG et al. (2003). SOCS3 negatively regulates IL-6 signaling in vivo. Nat Immunol 4: 540–545. Greenhalgh CJ, Miller ME, Hilton DJ, Lund PK. (2002). Suppressors of cytokine signaling: relevance to gastrointestinal function and disease. Gastroenterology 123: 2064–2081. Hanada T, Kobayashi T, Chinen T, Saeki K, Takaki H, Koga K et al. (2006). IFNgamma-dependent, spontaneous development of colorectal carcinomas in SOCS1-deficient mice. J Exp Med 203: 1391–1397. He B, You L, Uematsu K, Zang K, Xu Z, Lee AY et al. (2003). SOCS-3 is frequently silenced by hypermethylation and Oncogene

suppresses cell growth in human lung cancer. Proc Natl Acad Sci USA 100: 14133–14138. Heinrich PC, Behrmann I, Haan S, Hermanns HM, MullerNewen G, Schaper F. (2003). Principles of interleukin (IL)-6type cytokine signalling and its regulation. Biochem J 374: 1–20. Itzkowitz S. (2003). Colon carcinogenesis in inflammatory bowel disease: applying molecular genetics to clinical practice. J Clin Gastroenterol 36: 70–74. Judd LM, Alderman BM, Howlett M, Shulkes A, Dow C, Moverley J et al. (2004). Gastric cancer development in mice lacking the SHP2 binding site on the IL-6 family co-receptor gp130. Gastroenterology 126: 196–207. Judd LM, Bredin K, Kalantzis A, Jenkins BJ, Ernst M, Giraud AS. (2006). STAT3 activation regulates growth, inflammation, and vascularization in a mouse model of gastric tumorigenesis. Gastroenterology 131: 1073–1085. Kimura A, Naka T, Nagata S, Kawase I, Kishimoto T. (2004). SOCS-1 suppresses TNF-alpha-induced apoptosis through the regulation of Jak activation. Int Immunol 16: 991–999. Landi S, Moreno V, Gioia-Patricola L, Guino E, Navarro M, de Oca J et al. (2003). Association of common polymorphisms in inflammatory genes interleukin (IL)6, IL8, tumor necrosis factor alpha, NFKB1, and peroxisome proliferatoractivated receptor gamma with colorectal cancer. Cancer Res 63: 3560–3566. Miller ME, Goebel HN, Williams KL, Keku TO, Pucilowska JB, Sartor RB et al. (2000). Suppressor of cytokine signalling-3 (SOCS-3) mRNA is preferentially induced during experimental enterocolitis, ulcerative colitis and Crohn’s disease. Regulatory Peptides 94: 67–68.


4841 Miller ME, Michaylira CZ, Simmons JG, Ney DM, Dahly EM, Heath JK et al. (2004). Suppressor of cytokine signaling-2: a growth hormone-inducible inhibitor of intestinal epithelial cell proliferation. Gastroenterology 127: 570–581. Miscia S, Marchisio M, Grilli A, Di VV, Centurione L, Sabatino G et al. (2002). Tumor necrosis factor alpha (TNF-alpha) activates Jak1/Stat3-Stat5B signaling through TNFR-1 in human B cells. Cell Growth Differ 13: 13–18. Mizoguchi E, Xavier RJ, Reinecker HC, Uchino H, Bhan AK, Podolsky DK et al. (2003). Colonic epithelial functional phenotype varies with type and phase of experimental colitis. Gastroenterology 125: 148–161. Munkholm P. (2003). Review article: the incidence and prevalence of colorectal cancer in inflammatory bowel disease. Aliment Pharmacol Ther 18(Suppl 2): 1–5. Nakamura Y, Chang CC, Mori T, Sato K, Ohtsuki K, Upham BL et al. (2004). Augmentation of differentiation and gap junction function by kaempferol in partially-differentiated colon cancer cells. Carcinogenesis 26: 665–671. Niwa Y, Kanda H, Shikauchi Y, Saiura A, Matsubara K, Kitagawa T et al. (2005). Methylation silencing of SOCS-3 promotes cell growth and migration by enhancing JAK/ STAT and FAK signalings in human hepatocellular carcinoma. Oncogene 24: 6406–6417. Ogata H, Chinen T, Yoshida T, Kinjyo I, Takaesu G, Shiraishi H et al. (2006a). Loss of SOCS3 in the liver promotes fibrosis by enhancing STAT3-mediated TGF-beta1 production. Oncogene 25: 2520–2530. Ogata H, Kobayashi T, Chinen T, Takaki H, Sanada T, Minoda Y et al. (2006b). Deletion of the SOCS3 gene in liver parenchymal cells promotes hepatitis-induced hepatocarcinogenesis. Gastroenterology 131: 179–193. Oshimo Y, Kuraoka K, Nakayama H, Kitadai Y, Yoshida K, Chayama K et al. (2004). Epigenetic inactivation of SOCS-1 by CpG island hypermethylation in human gastric carcinoma. Int J Cancer 112: 1003–1009. Rivat C, Wever OD, Bruyneel E, Mareel M, Gespach C, Attoub S. (2004). Disruption of STAT3 signaling leads to

tumor cell invasion through alterations of homotypic cellcell adhesion complexes. Oncogene 23: 3317–3327. Schneider MR, Hoeflich A, Fischer JR, Wolf E, Sordat B, Lahm H. (2000). Interleukin-6 stimulates clonogenic growth of primary and metastatic human colon carcinoma cells. Cancer Lett 151: 31–38. Sutherland KD, Lindeman GJ, Choong DY, Wittlin S, Brentzell L, Phillips W et al. (2004). Differential hypermethylation of SOCS genes in ovarian and breast carcinomas. Oncogene 23: 7726–7733. Suzuki A, Hanada T, Mitsuyama K, Yoshida T, Kamizono S, Hoshino T et al. (2001). CIS3/SOCS3/SSI3 plays a negative regulatory role in STAT3 activation and intestinal inflammation. J Exp Med 193: 471–481. Tebbutt NC, Giraud AS, Inglese M, Jenkins B, Waring P, Clay FJ et al. (2002). Reciprocal regulation of gastrointestinal homeostasis by SHP2 and STAT-mediated trefoil gene activation in gp130 mutant mice. Nat Med 8: 1089–1097. Ulshen MH, Dowling RH, Fuller CR, Zimmermann EM, Lund PK. (1993). Enhanced growth of small bowel in transgenic mice overexpressing bovine growth hormone. Gastroenterology 104: 973–980. Wang L, Walia B, Evans J, Gewirtz AT, Merlin D, Sitaraman SV. (2003). IL-6 induces NF-kappa B activation in the intestinal epithelia. J Immunol 171: 3194–3201. Weber A, Hengge UR, Bardenheuer W, Tischoff I, Sommerer F, Markwarth 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. Williams KL, Fuller CR, Dieleman LA, DaCosta CM, Haldeman KM, Sartor RB et al. (2001). Enhanced survival and mucosal repair after dextran sodium sulfate-induced colitis in transgenic mice that overexpress growth hormone. Gastroenterology 120: 925–937. Yasukawa H, Ohishi M, Mori H, Murakami M, Chinen T, Aki D et al. (2003). IL-6 induces an anti-inflammatory response in the absence of SOCS3 in macrophages. Nat Immunol 4: 551–556.

Oncogene