reduction of inflammatory bowel diseaseinduced tumor ...molecular carcinogenesis 52:726–738 (2013)...

MOLECULAR CARCINOGENESIS 52:726–738 (2013)

Reduction of Inflammatory Bowel Disease-InducedTumor Development in IL-10 Knockout Mice WithSoluble Epoxide Hydrolase Gene Deficiency

Wanying Zhang,1 Jie Liao,1 Haonan Li,1 Hua Dong,2 Han Bai,1 Allison Yang,1 Bruce D. Hammock,2

and Guang-Yu Yang1*1Department of Pathology, Northwestern University, Feinberg School of Medicine, Chicago, Illinois2Department of Entomology, University of California, Davis, California

Soluble epoxide hydrolase (sEH) quickly inactivates anti-inflammatory epoxyeicosatrienoic acids (EETs) by convertingthem to dihydroxyeicosatrienoic acids (DHETs). Inhibition of sEH has shown effects against inflammation, but little isstudied about the role of sEH in inflammatory bowel disease (IBD) and its induced carcinogenesis. In the present study,

the effect of sEH gene deficiency on the development of IBD-induced tumor development was determined in IL-10knockout mice combined with sEH gene deficiency. Tumor development in the bowel was examined at the age of25 wk for male mice and 35 wk for female mice. Compared to IL-10(�/�) mice, sEH (�/�)/IL-10(�/�) mice exhibiteda significant decrease of tumor multiplicity (2 � 0.9 tumors/mouse vs. 1 � 0.3 tumors/mouse) and tumor size

(344.55 � 71.73 mm3 vs. 126.94 � 23.18 mm3), as well as a marked decrease of precancerous dysplasia. The signifi-cantly lower inflammatory scores were further observed in the bowel in sEH(�/�)/IL-10(�/�) mice as compared to IL-10(�/�) mice, including parameters of inflammation-involved area (0.70 � 0.16 vs. 1.4 � 0.18), inflammation cell

infiltration (1.55 � 0.35 vs. 2.15 � 0.18), and epithelial hyperplasia (0.95 � 0.21 vs. 1.45 � 0.18), as well as largerulcer formation. qPCR and Western blotting assays demonstrated a significant downregulation of cytokines/chemo-kines (TNF-a, MCP-1, and IL-12, 17, and 23) and NF-kB signals. Eicosanoid acid metabolic profiling revealed a signifi-

cant increase of ratios of EETs to DHETs and EpOMEs to DiOMEs. These results indicate that sEH plays an importantrole in IBD and its-induced carcinogenesis and could serve as a highly potential target of chemoprevention and treat-ment for IBD. � 2012 Wiley Periodicals, Inc.

Key words: inflammatory bowel disease; carcinogenesis; soluble epoxide hydrolase; IL-10; eicosanoid acid metabolic

profiling

INTRODUCTION

Inflammatory bowel disease (IBD), including ul-cerative colitis and Crohn’s disease, is a longstand-ing chronic active inflammatory process with anincreased risk for the development of cancer inthe intestinal tract. Aberrant metabolites of arachi-donic acid in the process of long-term inflammationplay a crucial role in inflammation-inducedcarcinogenesis [1,2]. Three key metabolic pathwaysare involved in arachidonic acid metabolism, in-cluding cyclooxygenase (COX)-mediated prostaglan-dins, lipoxygenase (LOX)-mediated leukotrienes andhydroxyeicosatetraenoic acids (HETEs), and cyto-chrome P450-mediated epoxyeicosatrienoic acids(EETs) [1]. There are numerous studies that indicatethe COX- and LOX-mediated products promotingcarcinogenesis [2]. But the third pathway of P450-mediated EETs is under study.

The role of P450-EETs-soluble epoxide hydrolase(sEH) pathway in inflammation has been assessed inrodent models [3–9]. The first finding for anti-inflammatory activity by EETs was reported in 1999and showed that physiological concentrationsof EETs or overexpression of p450-2J2 (that trans-formed arachidonic acid to EETs) inhibits inflamma-tion through decreasing tumor necrosis factor-a

(TNF-a)-induced vascular cell adhesion moleculeexpression (VCAM-1) and reducing leukocyte adhe-sion and infiltration to the vascular walland inflamed tissues, and suppresses transcriptionfactor NF-kB and IKK kinase [10]. Under physiologiccondition, EETs are quickly inactivated by sEH thatconverts them to their corresponding dihydroxyei-cosatrienoic acids (DHETs) [7]. It has been demon-strated that sEH inhibition or gene knockoutstabilizes EETs and increases EETs to DHETs ratio[11,12], resulting in dramatically anti-inflammatory

Abbreviations: IBD, inflammatory bowel disease; COX, cyclooxy-genase; LOX, lipoxygenase; HETEs, hydroxyeicosatetraenoic acids;EETs, epoxyeicosatrienoic acids; sEH, soluble epoxide hydrolase; TNF-a, tumor necrosis factor-a; VCAM-1, vascular cell adhesionmolecule; DHETs, dihydroxyeicosatrienoic acids; LC/MS-MS, liquidchromatography/mass spectrometry; MPO, myeloperoxidase; IFN,interferon; MCP-1, monocyte chemoattractant protein-1; PGE2, pros-taglandin E2; TXB2, thromboxane B2; LTB4, leukotriene B4; 5-HETE,5-hydroxyeicosatetraenoic acid; MCP-1, monocyte chemoattractantprotein-1.

*Correspondence to: Department of Pathology, NorthwesternUniversity, Feinberg School of Medicine, 303 E. Chicago Avenue,Chicago, IL 60611.

Received 7 January 2012; Revised 26 March 2012; Accepted 3April 2012

DOI 10.1002/mc.21918

Published online 19 April 2012 in Wiley Online Library(wileyonlinelibrary.com).

� 2012 WILEY PERIODICALS, INC.

function [3–9,13]. Using LPS-induced systematic in-flammation in mice, treatment with sEH inhibitorreduces the production of nitric oxide, cytokines,and proinflammatory lipid mediators and signifi-cantly improves survival [7,14]. In the smoking in-duced bronchitis model in rats, sEH inhibitor showsa significant decrease of total bronchoalveolar in-flammatory cells by 37% in tobacco smoke-exposedrats [8]. A combination of sEH inhibitor and EETsexhibits more significant effect on reducing tobaccosmoking-induced inflammation [8]. sEH knock-outmice (disruption of Epx2 gene which encodes sEH)are phenotypically normal with only minimally de-creased body weight in males, but have an alteredarachidonic acid metabolism [15]. sEH knockout inmice results in a significant shift of the epoxy-fattyacid to diol ratio, led to protection against myocar-dial ischemia-reperfusion injury, modulation ofthe inflammatory response to cerebral ischemia,and improvement of glucose homeostasis [16–18].However, there is no study on investigating the roleof sEH gene deficiency in long-term inflammation-induced carcinogenesis.A spontaneous IBD and its induced carcinogenesis

in IL-10 knockout [IL-10(�/�)] mice serves as a rep-resentative model of IBD [2]. Typically, IL-10(�/�)mice spontaneously develop chronic enterocolitiswith the transmural inflammation pattern in thesmall and large bowel which is phenotypically mim-icking human Crohn’s disease [19]. But the occur-rence of IBD greatly varies from 3 to 12 months.Synchronizing the development of IBD has beenachieved through the short-period administration ofpiroxicam (1-wk treatment) [20]. Development ofcancer in intestinal tract, mainly in the cecum andproximal small bowel, has been observed in IL-10(�/�) mice with long-term IBD [21].Our short-term animal experiment has showed

that sEH gene deficiency in IL-10(�/�) mice resultsin a significant attenuation of inflammatory activityin the gastrointestinal tract (published separately).In the present study, development of tumors anddysplasia in intestinal tracts was analyzed histopath-ologically and immunohistochemically in the long-term survival IL-10(�/�) mice and double knockoutIL-10(�/�)/sEH(�/�) mice. Inflammation activityand its correlation with tumor development wereinvestigated. qPCR and Western blotting assays wereperformed to analyze key inflammatory cytokines/chemokines and NF-kB signals. The eicosanoid met-abolic profile was analyzed using a liquid chroma-tography/mass spectrometry (LC/MS-MS) method.

MATERIALS AND METHODS

Animal Breeding and Genotyping

All animal experiments were approved by the In-stitutional Animal Care and Use Committee atNorthwestern University. IL-10(�/�) mice in a

C57BL/6J background were purchased from JacksonLaboratory (Bar Harbor, ME), and sEH(�/�) micein a C57BL/6J background were provided by Dr.Hammock at the University of California in Davis,CA. To produce homozygous double knockoutsEH(�/�)/IL-10(�/�) mice, female IL-10(�/�) micewere first crossed with male sEH (�/�) mice to gener-ate IL-10(þ/�)/sEH(þ/�) double heterozygous litters(F1 colonies). Then, sEH(�/�)/IL-10(�/�) mice wereproduced by further cross-breeding IL-10(þ/�)/sEH(þ/�) mice (F2 colonies). Genotyping was per-formed based on the protocol from Jackson Laborato-ry. All mice were housed in micro-isolator cages (3–5 mice/cage) in a room illuminated with 12-h/12-hlight–dark cycle and with free access to waterand chow in the animal facilities at the Center forComparative Medicine at Northwestern University(Chicago, IL).

Animal Experiments and Tissue Processing

IL-10(�/�), sEH(�/�)/IL-10(�/�), sEH(�/�), andC57BL/6J wild-type mice (equal number of each gen-der per group) were used. Since the occurrence ofIBD varies greatly in IL-10(�/�) mice, all mice at theage of 4 wk were treated with 200 ppm piroxicam inAIN 93M diet for a week. At age 5 wk, AIN 93M dietwas administered to the mice until the end of experi-ment. Food and water consumption were monitoreddaily, and body weights were recorded every week.For examination of the tumor development, malemice were scarified at age 25 wk due to the exhibi-tion of more than 20% loss of body weight; and fe-male animals were sacrificed at age 38 wk.

Blood and plasma were collected via heartacupuncture and stored in a �808C freezer untilanalysis. All of the key organs including entire gas-trointestinal tract, liver, spleen, kidney, etc were col-lected. Briefly, small and large intestines wereinflated by in situ intra-luminal perfusion withchilled normal saline solution and opened longitu-dinally. The number, size and distribution of intesti-nal lesions (ulcer and tumor) were determined.Tumor volume was determined using the equationV ¼ 4/3pr3, where r was the average tumor radiusobtained from the two diameter measurements.Mucosa tissue from half of entire gastrointestinaltract was collected freshly in RNA later for biochem-ical assay including protein extraction for Westernblot and RNA extraction for qPCR. The other halfof the entire gastrointestinal tract (including stom-ach, small bowel and colon) were fixed in a 10%buffered formalin, prepared as ‘‘Swiss’’ rolls, andprocessed for paraffin sections for histopathologicaland immunohistochemical analyses.

Histopathological Analysis

Intestinal tumor was analyzed based on the estab-lished criteria [22,23]. Adenocarcinoma was diag-nosed by the invasion of neoplastic cells through

ROLE OF sEH IN IBD-INDUCED CARCINOGENESIS 727

Molecular Carcinogenesis

the basement membrane into laminar propria andsubmucosa, and further classified as polypoid well,moderate, or poorly differentiated pattern based onformation of glandular structure and mucinous car-cinoma with mucin production in more than 75%tumor. Dysplasia was characterized by the partialloss of cell polarity and maturation, nuclear atypia,and an increase in mitotic figures [24].

The grading of IBD was based on the establishedcriteria, including the parameters of: (1) the portionof area involved by inflammation, (2) intensityof inflammatory cell infiltration, (3) reactive/regenerative epithelial hyperplasia, and (4) ulcer for-mation [22]. Briefly, the inflammation involved areawas scored from 0 to 3. 0: no inflammation in intes-tine 1: <25% of intestine involved by inflammation;2: 25–50%; and 3: more than 50%. The inflammato-ry cell infiltration was scored from 0 to 3: 0, mini-mal inflammatory cell infiltration in the mucosa ofintestinal tract, mainly are plasma cells and lympho-cytes; 1: focal intense inflammatory cell infiltration(<30% inflammatory cells in total cells countedin the inflamed areas) with occasional neutrophils(<10% of inflammatory cells are neutrophilshighlighted by myeloperoxidase immunostaining);2: focal highly intense inflammatory cells (30–60%inflammatory cells in the inflamed areas with 10–20% myeloperoxidase (MPO)-positive-stained neu-trophils) and epithelial injury and mucosal erosion,and 3: ulcer formation with intense and more than60% inflammatory cells infiltration (>20% MPO-positive-stained neutrophils) in the inflamed areas.Epithelial hyperplasia was scored as 0–3 accordingto the ratio of the mucosal thickness to the entireintestinal wall, 0: the ratio is less than 1/3; 1: 1/3–1/2; 2: 1/2–3/4; and 3: more than 3/4. The score forulcer was 0: no ulcer and 1: with ulcer formation.The total inflammation index will be the sum of allparameters.

Immunohistochemistry

Immunohistochemical staining was performedusing avidin–biotin–peroxidase complex method onparaffin-embedded intestinal sections as previouslydescribed [22]. The primary antibodies are myelo-peroxidase antibody (Rabbit pAb, 60 mg/mL, Abcam,Cambridge, MA) and Ki-67 antibody (Rabbit mAb,1:100 diluted, Vector Laboratories, Inc., Burlingame,CA). Biotinyted anti-rabbit IgG secondary antibody(1:200) and ABC complex (avidin–biotin complex)were purchased from Vector Laboratories, Inc.Diaminobenzidine (Sigma-Aldrich Co., St. Louis,MO) was used as the chromogen. Slides were washedwith 1� TBST buffer between incubations and coun-terstained with Mayer’s hemotoxylin for 1½ min.The number of myeloperoxidase-positive-stained in-flammatory cells per high power (40� objectivelens) was counted for each specimen. Cell prolifera-tion was analyzed by Ki-67-labeled cells and

proliferation index was the percentage of Ki-67-positive cells in the total number of cells counted.

Quantitative Real-Time PCR

Total RNA was extracted from the colonic mucosausing the RNeasy kit (Qiagen, Inc., Valencia, CA).The concentration was determined using a Smart-Spec Plus Spectrophotometer (BioRad, Hercules,CA). First-strand cDNA was synthesized using 1 mgof total RNA in a 20 mL reverse transcriptase reac-tion mixture using SuperScript III Platinum Two-Step q-RT-PCR Kit with SYBR Green (Invitrogen,Carlsbad, CA) according to the manufacturer’sinstructions. The region of IFN-g, TNF-a, and mono-cyte chemoattractant protein-1 (MCP-1) mRNA wereamplified using primers: IFN-g (Forward 50-TCA-AGTGGCATAGATGTGGAAGAA; Reverse 50-TGGCT-CTGCAGGATTTTCATG), TNF-a (Forward 50-CCAACGGCATGGATCTCAAAGACA; Reverse 50-TGAGA-TAGCAAATCGGCTGACGGT), MCP-1 (Forward 50-ACTGAAGCCAGC TCTCTCTTCCTC; Reverse 50-TTCCTTCTTGGG GTCAGCACAGAC) as describedin literatures [25,26]. The region of IL-12, IL-17, andIL-23 mRNA were amplified using primers IL-12(Forward 50-CAGATAGCCCATCACCCTGT; Reverse50-ACGGCCAGAGAAAAACTGAA), IL-17 (Forward50-TCTCTGATGCTGTTGCTGCT; Reverse 50-CGTG-GAACGGTTGAGGTAGT), and IL-23 (Forward 50-AATAATGTGC CCCGTATCCA, Reverse 50-CTGGAG-GAGTTGGCTGAGTC). GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was used as internal con-trol (Forward 50-GCACAGTCAAGGCCGAGAAT; Re-verse 50-GCCTTCTCCATGGTGGTGAA). All real-time PCR reactions were performed in a 25 mL mix-ture containing 1/10 volume of cDNA preparation(2.5 mL), 12.5 mL SYBR Green buffer (Invitrogen),0.5 mM of each primer, 8.5 mL DEPC (diethylprocar-bonate) (EMD Chemicals, Gibbstown, NJ). Real-timequantitation was performed using the MiniOpticonReal-Time PCR System (BioRad). PCR conditionswere: 508C for 2 min, 958C for 2 min, followed by40 cycles of 958C, 15 s; 588C, 3 s; 508C, 1 s. Data ofeach mRNA expression were shown as the relativefold of change normalized by that of GAPDH.

Western Blotting

One hundred fifty microliters of whole tissue/celllysate of intestinal mucosa (n ¼ 3 mice per group)was separated by 10% SDS–PAGE and proteinexpression was analyzed using the followingantibodies: rabbit anti-NF-kB mAb P65 and anti-phosphorylated NF-kB p65 (S276) mAb (1:1,000,Cell Signaling Technology, Danvers, MA), rabbitanti-IKK-a (1:1,000, Cell Signaling Technology),goat anti-VCAM-1 IgG Ab (1:750, R&D System,Minneapolis, MN), and mouse anti-b-actin mAb(1:2000, Sigma-Aldrich Co.). Signals were visualizedwith SuperSignal West Pico Chemiluminescent Sub-strate (Thermo Scientific, Rockford, IL).

728 ZHANG ET AL.


Analysis of Arachidonic Acid Metabolic Profile Using

LC/MS-MS Method

Serum specimens were spiked with 10 mL of50 nM internal standard I (d4-6-keto-PGF1a, d4-PGE2, d4-TXB2, d4-LTB4, d11-14,15-DiHETrE, d4-9-HODE, d8-5-HETE, d11-11(12)-EpETrE) and wereextracted by solid phase extraction using Oasis HLBcartridges (3 cm3 60 mg, Waters, Milford, MA). TheHLB cartridges were first washed with 2 mL ethylacetate, 2 mL methanol twice, and 2 mL 95:5 (v/v)water/methanol with 0.1% acetic acid. Then,200 mL serum samples were loaded onto thecartridges, washed with 6 mL 95:5 (v/v) water/methanol with 0.1% acetic acid, and dried for20 min with low vacuum. The target analytes wereeluted with 0.5 mL methanol, followed by 2 mL ofethyl acetate into the tubes with 6 mL 30% glycerolin methanol as the trap solution. The volatile sol-vents were evaporated using vacuum centrifugation(Speed-Vac) until 2 mL trap solution remained in thetube. The residues were dissolved in 50 mL of metha-nol containing 200 nM internal standards II (1-cyclohexyl-dodecanoic acid urea, CUDA). The sam-ples were mixed with a vortex mixer for 2 min, cen-trifuged at 14 000g for 5 min and then transferredto auto sampler vials with 150 mL inserts for LC/MS-MS analysis.Liquid chromatography/tandem mass spectrome-

try (LC/MS-MS) analysis of oxylipids was performedusing a modified method based on the previouspublication [27]. An Agilent 1200 SL liquid chroma-tography series (Agilent Corporation, Palo Alto,CA) with an Agilent Eclipse Plus C18 2.1 mm �150 mm, 1.8 mm column were used for the oxyli-pins separation. The mobile phase A was water with0.1% acetic acid while the mobile phase B was com-posed of acetonitrile/methanol (80/15, v/v) and0.1% acetic acid. Gradient elution was performed ata flow rate of 250 mL/min and the gradient used isdescribed in the attached table. The injection vol-ume was 10 mL and the samples were kept at 48C inthe auto sampler. Analytes were detected by nega-tive MRM mode using a 4000 QTrap tandemmass spectrometer (Applied Biosystems InstrumentCorporation, Foster City, CA) equipped withan electrospray ionization source (Turbo V). TheQTrap was set as follows: CUR ¼ 20 psi, TEM ¼5008C, GS1 ¼ 50 psi, GS2 ¼ 30 psi, CAD ¼ High,IS ¼ �4500 V, DP ¼ �60 V, EP ¼ �10 V. Calibra-tion curves were generated by 10 mL injections ofseven standards containing each analyte, internalstandard I, and internal standard II for quantifica-tion purpose.

Statistical Analysis

Each analyzed parameter was expressed asmean � SD, unless otherwise stated, with at leastthree independent measurements. Continuous

variables were compared with the Student’s t-test,whereas categorical variables were compared withChi-square test. All statistical tests were two-sided,statistical significance was taken as P < 0.05.

RESULT

Effect of sEH Gene Deficiency on IBD-InducedCarcinogenesis in the Bowel

All of the mice including wild-type, sEH(�/�), IL-10(�/�), and sEH(�/�)/IL-10(�/�) mice were moni-tored daily. In order to synchronize the develop-ment of IBD, the mice were fed an AIN 93M dietcontaining 200 ppm of piroxicam for 1 wk (at age4 wk) [28]. Early mucosal injury in the bowel wasobserved 10 d after piroxicam use; IL-10(�/�) mice,sEH(�/�)/IL-10(�/�) mice, sEH(�/�) mice, andwild-type control animals all exhibited mild muco-sal ulcer formation in the cecum and proximal smallbowl. sEH(�/�)/IL-10(�/�) and sEH(�/�) miceshowed much milder ulcerations as compared tothe IL-10(�/�) mice. While the sEH(�/�) and wild-type control mice showed a full recovery from theerosion and ulceration 4 wk after piroxicam use, thesEH(�/�)/IL-10(�/�) and IL-10(�/�) mice contin-ued to display chronic active IBD (n ¼ 3 mice eachstrain each gender mice). Thus, the IBD process wassynchronically induced in the sEH(�/�)/IL-10(�/�)and IL-10(�/�) mice at 4 wk postpiroxicamtreatment.

At age 17 wk, male IL-10(�/�) and sEH(�/�)/IL-10(�/�) mice started to display loss of body weightand progression worsened. At age 25 wk, more than50% male IL-10(�/�) mice and sEH(�/�)/IL-10(�/�) mice showed >20% decrease of body weight ascompared to male wild-type mice. Male IL-10(�/�)mice displayed more severe diarrhea and bowel in-flammation than male sEH(�/�)/IL-10(�/�) mice.But male sEH(�/�)/IL-10(�/�) mice exhibitedsmaller prostate and testicles, indicating deductionof body weight probably due to alteration of testos-terone [29]. Female sEH (�/�)/IL-10(�/�) mice didnot exhibited any gross abnormalities until age38 wk.

Extensive pathological analysis was performed forthe entire gastrointestinal tract of all mice. In wild-type and sEH(�/�) mice, no tumor or inflammationwas observed. In IL-10(�/�) mice, 85% mice (17/20)developed polypoid tumor masses (Figure 1A), 53%tumors located in duodenal/proximal small boweland 47% in cecum/right colon with average2 � 0.9 tumors/mouse (tumor multiplicity) and av-erage tumor volume 344.55 � 71.73 mm3. Histopa-thology revealed that most tumors were polypoidwell to moderately differentiated adenocarcinomawith invasion into laminar propria or submucosa(Figure 1B), and only one mouse developed an inva-sive mucinous carcinoma (Figure 1C and D). Com-pared to IL-10(�/�) mice, a significant decrease of



tumor multiplicity and volume was observed insEH(�/�)/IL-10(�/�) mice that had average 1 �0.3 tumor/mouse (Figure 1E) and average tumor vol-ume 126 � 23.18 mm3 (Figure 1F, P ¼ 0.028). InsEH(�/�)/IL-10(�/�) mice, 33% tumors occurredin the duodenum/proximal small bowel and 67% inthe cecum and right colon; but no statistical differ-ence was observed for tumor location as well as forgenders, as compared to IL-10(�/�) mice. Histopa-thology revealed that all of the intestinal tumors insEH(�/�)/IL-10(�/�) mice were polypoid well differ-entiated adenocarcinoma, and no invasive mucin-ous carcinoma was seen.

IBD-induced dysplasia in the bowel was analyzedhistopathologically and further characterized withKi-67-labeld proliferative cells. Morphologically,IBD-induced dysplasia was characterized as the par-tial loss of nuclear polarity and cell maturation withnuclear atypia, and an increase in mitotic figures, asseen in Figure 2A and B. Ki-67-labeled proliferativecells were observed in the crypt epithelium and ex-tended to the luminal surface in the dysplastic le-sion, as shown in Figure 2C and D. In nondysplasticmucosa, Ki-67-labeld proliferative cells were onlyobserved in the crypt epithelium. As summarized inhistogram in Figure 2E, 90% (18/20) of IL-10(�/�)mice had average 3 � 1.19 dysplasia/mouse; andonly 60% of sEH(�/�)/IL-10(�/�) mice carried

dysplasia with average 2 � 1.15 dysplasia/mouse.Ki-67 labeled cell proliferation was further semi-quantitatively analyzed and showed a significant de-crease of cell proliferation in inflammation-inducedepithelial hyperplasia, dysplasia and adenocarcino-ma, as seen in Figure 2F.

Effect of sEH Gene Deficiency on Intestinal ChronicActive Inflammation

Inflammatory activity in the intestinal tract wassemi-quantitatively analyzed histopathologically, in-cluding the parameters of inflammation-involvedarea, inflammatory cell infiltration, inflammation-caused reactive epithelial hyperplasia, and ulcerformation according to our established criteria [22].In IL-10(�/�) mice, 90% (18/20) displayedextensive and intense inflammation with ulcer for-mation in the bowel (Figure 3A), and the averageinflammatory scores were 1.40 � 0.18 for inflamma-tion-involved area, 2.15 � 0.18 inflammation cellinfiltration, and 1.45 � 0.18 for inflammation-caused reactive epithelial hyperplasia, furthermore35% IL-10(�/�) mice had large active ulcer(Figure 3A) with extensive myeloperoxidase-labeledneutrophil infiltration (Figure 3C). Compared to IL-10(�/�) mice, only 45% (9/20) of sEH(�/�)/IL-10(�/�) mice showed focal mild inflammation(Figure 3B), and the average inflammatory scores

Figure 1. Morphology of colitis-induced adenocarcinoma and his-tograms of tumor number and size in IL-10(�/�) and sEH(�/�)/IL-10(�/�) mice. (A) A polypoid well-differentiated adenocarcinomawas observed in the intestinal mucosa under low-power field(125�), and (B) high magnification (500�) of square area of (A),showing adenocarcinoma invading into laminar propria in the colon-ic mucosa in IL-10(�/�) mice. (C) Low-power view of transmuralinvasion of mucinous carcinoma in the colon in IL-10(�/�) mice(125�), and (D) high magnification (250�) of square area of (C),

showing well-differentiated mucinous carcinoma invading into mus-cularis propria. (E) Histogram of tumor multiplicity (number oftumors/mouse): showing a significant decrease of tumor multiplicityin sEH(�/�)/IL-10(�/�) mice (n ¼ 20) as compared to IL-10(�/�)mice (n ¼ 19, P ¼ 0.01). (F) Distribution of tumor size: the tumorsize in sEH(�/�)/IL-10(�/�) mice (n ¼ 20) was significantly smallerthan that in IL-10(�/�) mice (n ¼ 19). Results were reported asmean � SD. The statistical difference between sEH(�/�)/IL-10(�/�)and IL-10(�/�) animals was calculated and labeled in the figure.

730 ZHANG ET AL.


were 0.70 � 0.16 for inflammation-involvedarea (P ¼ 0.007), 1.55 � 0.35 for inflammationcell infiltration (P ¼ 0.032), and 0.95 � 0.21 for in-flammation-caused reactive epithelial hyperplasia

(P ¼ 0.044), as summarized in histogram inFigure 3E. Only 10% sEH(�/�)/IL-10(�/�) mice de-veloped small active ulcers (10% vs. 35%, P < 0.05).MPO immunohistochemical staining was performed

Figure 2. Dysplastic lesions and cell proliferation labeled withKi-67 immunohistochemistry. (A) Representative dysplastic lesion inIL-10(�/�) mice. (B) Representative dysplastic lesion in sEH(�/�)/IL-10(�/�) mice. (C) Dysplastic lesion, showing Ki-67-labeled prolifer-ative cells diffusely distributed throughout the crypts to the luminalsurface epithelium in the colon in IL-10(�/�) mice. (D) The prolifer-ative cells labeled with Ki-67 were fewer and predominantly limitedwithin the crypts and focally extend to the luminal surface of colonin sEH(�/�)/IL-10(�/�) mice. (E) Histogram of the number of dyspla-sia (counted throughout the entire colon): showing a marked de-crease of the number of dysplasia in sEH(�/�)/IL-10(�/�) mice

(n ¼ 19) as compared to that in IL-10(�/�) (n ¼ 20); but notreached statistical significance (P > 0.05). (F) Histogram of the aver-age percentage of Ki-67-labeld proliferative cells: showing asignificant decrease of Ki-67-labeld proliferative cells in sEH(�/�)/IL-10(�/�) mice (n ¼ 10) as compared to IL-10(�/�) mice (n ¼ 7) inlesions of inflammation (P ¼ 0.05), dysplasia (P ¼ 0.03), and adeno-carcinoma (P ¼ 0.05). Results represented as mean � SD. Percent-age of positive-stained Ki-67 cells in total cells counted in differentlesions in the colon was analyzed by three randomly chosen areas ofeach lesion under 40� HPF.



to analyze active neutrophil infiltration in theinflamed areas. Compared to that of IL-10(�/�)mice (Figure 3C), a much significant decreaseof MPO-positive neutrophils in the inflamed areaswas observed in sEH(�/�)/IL-10(�/�) mice

(Figure 3D). Overall inflammatory index (sum ofeach parameter) was significantly decreased insEH(�/�)/IL-10(�/�) mice compared to that of IL-10(�/�) mice (3.3 � 0.74 vs. 5.35 � 0.46, P ¼ 0.003,Figure 3E).

Figure 3. Morphological analysis and semi-quantification of in-flammatory activity and ulcer formation in IL-10(�/�) and sEH(�/�)/IL-10(�/�) mice. (A) The colon of IL-10(�/�) mouse exhibits alarge ulcer with adjacent hyperplastic epithelial changes andextensive inflammatory cell infiltration. (B) The colon of sEH(�/�)/IL-10(�/�) mouse showed a small healed ulcer with mild hyperplasticepithelium and less inflammatory cell infiltration. (C) Immunohis-tochemistry of myeloperoxidase (MPO), a square area in photo Ashows intense MPO-positive neutrophil infiltration; and (D) MPOimmunohistochemistry, a square area in photo B displays much

significant less MPO-positive neutrophil infiltration. (E) Histogram ofinflammatory severity in IBD: showing a semi-quantification of in-flammatory severity, ulcer, hyperplasia, and lesion area involved insEH(�/�)/IL-10(�/�) mice (n ¼ 20) and in IL-10(�/�) mice (n ¼ 19).The statistically significant difference was observed between sEH(�/�)/IL-10(�/�) and IL-10(�/�) mice in inflammatory severity(P ¼ 0.03), hyperplasia (P ¼ 0.04), extent of that involved(P ¼ 0.007), and total inflammatory index (P ¼ 0.003). Results werereported as mean � SD.

732 ZHANG ET AL.


Analysis of Key Inflammatory Cytokines and Chemokines

As Well As NF-kB and VCAM-1 Signals With QuantitativeReal-Time-PCR and Western Blot Approaches

Transcriptional expressions of inflammatory cyto-kines and chemokines were analyzed quantitativelyusing real-time-PCR approach for freshly collectedcolonic mucosa (n ¼ 6 mice/group). As seen inFigure 4, all of wild-type mice showed minimal ex-pression of IFN-g, TNF-a, MCP-1, IL-12, IL-17, andIL-23. By setting the value of each cytokine expres-sion in wild-type mice as 1, a minimal increase ofMCP-1, IL-17, and IL-23 was seen in sEH(�/�),but did not reach a statistical significance. A signifi-cant increase of the mRNA expressions in thesecytokines/chemokines was found in IL-10(�/�)mice as compared to wild-type animals, including13-fold increase in TNF-a, 5-fold increase in IFN-g,10-fold increase in MCP-1, 5-fold increase in IL-12,and 4-fold increase in both IL-17 and IL-23.Compared to IL-10(�/�) mice, a highly significantdecrease of the mRNA expression in these inflamma-tory cytokines and chemokines was observed insEH(�/�)/IL-10(�/�) mice and showed 7-fold de-crease in TNF-a, 3-fold decrease in IFN-g, 8-fold de-crease in MCP-1, 4-fold decrease in IL-12, 1-folddecrease in IL-17, and 2-fold decrease in IL-23 ascompared to IL-10(�/�) mice, as seen in Figure 4(P < 0.01).Expressions of total and phosphorylated NF-kB, as

well Ikka and VCAM-1 were measured using the

Western blotting approach and densitometryanalysis for whole tissue/cell lysates of freshly col-lected colonic mucosa (n ¼ 6/group). As shown inFigure 5A and B, all of these inflammatory signalswere significantly decreased in sEH(�/�)/IL-10(�/�)mice as compared to those in IL-10(�/�) mice.

Plasma Levels of Eicosanoid Profile Analyzed With a

Liquid Chromatography-Tandem Mass Spectrometry(LC/MS-MS) Assay

Quantitative analysis of eicosanoids profiles is farmore significant than individual biomarker(s) forevaluating arachidonic acid metabolic pathways.Simply, the ratios of fatty acid epoxides (EpOMEs)to their corresponding diols (DiHOMEs) and EETs toDHETs are the most crucial biomarkers for deter-mining sEH activity. With LC/MS-MS approach, weanalyzed plasma levels of eicosanoid profile in theseanimals. Compared to wild-type mice, sEH(�/�)exhibited a significant increase of EETs, including11(12)-, 14(15)-EET, and the fatty acid epoxides(EpOME), including 9(10)- and 12(13)-EpOME, aswell as a significantly decrease of corresponding fat-ty acid diols, including 9(11)-DHOME and 12(13)-DHOME, and the DHETs, including 11(12)-DHET,and 14(15)-DHET, as seen in Figure 6. As expected,ratios of EETs to DHETs and EpOMEs to DiHOMEswere significantly increased in sEH(�/�) mice.

In IL-10(�/�) mice (IBD mice), a significant in-crease of fatty acid diols, including 9(11)-DHOME

Figure 4. Quantitative real-time PCR analysis of TNF-a, IFN-g, MCP-1, IL-12, IL-17, and IL-23 mRNA expres-sions in the colon in wild-type, sEH(�/�). IL-10(�/�), sEH(�/�)/IL-10(�/�) mice. The statistically significantdifference of these cytokines and chemokines was observed between IL-10(�/�) and sEH(�/�)/IL-10(�/�) mice(n ¼ 6/group, mucosa specimens from 3 males and 3 females chosen randomly from each group). Results werereported as mean � SE.



and 12(13)-DHOME, and the DHETs, including11(12)-DHET, and 14(15)-DHET was observed ascompared to wild-type mice (P < 0.05, Figure 6) Butno change was seen in the levels of EETs (including11(12)-, 14(15)-EET) and the fatty acid epoxides(EpOME) (including 9(10)- and 12(13)-EpOME), northe ratios of EETs to DHETs and EpOMEs toDiHOMEs (Figure 6). Compared to IL-10(�/�) mice,sEH(�/�)/IL-10(�/�) mice displayed a significant el-evation of the levels of EETs (including 11(12)-,14(15)-EET) and fatty acid epoxides (EpOME, includ-ing 9(10)- and 12(13)-EpOME), and a significant de-crease of the levels of corresponding fatty acid diols(including 9(11)-DHOME and 12(13)-DHOME) andDHETs (including 11(12)-DHET and 14(15)-DHET),as seen in Figure 6. sEH(�/�)/IL-10(�/�) miceexhibited a much more significant increase in theratios of EETs to DHETs and EpOMEs to DiHOMEsas compared to IL-10(�/�) mice (Figure 6).

To investigate if there was an effect of sEH genedeficiency on modulating other pathways of arachi-donic acid metabolism, the key cyclooxygenase(COX2) and lipoxygenase (LOX)-mediated metabo-lites were analyzed, including PGE2 (prostaglandinE2), PGD2 (prostaglandin D2), and TXB2 (thrombox-ane B2) as products of COX2 pathway, andLTB4 (leukotriene B4) and 5-HETE (5-hydroxyeicosa-tetraenoic acid) as metabolites of LOX pathway. Asshown in Figure 7, the level of PGE2 was significant-ly increased in IBD mice (both IL-10(�/�) andsEH(�/�)/IL-10(�/�) mice) as compared to wild-type and sEH(�/�) mice; but a slight decrease ofPGE2 was seen sEH(�/�)/IL-10(�/�) mice as com-pared to IL-10(�/�) mice (no statistically signifi-cant). The level of LTB4 was significantly increasedin sEH knockout mice (both sEH(�/�) mice andsEH(�/�)/IL-10(�/�) mice) as compared to wild-type and IL-10(�/�) mice (Figure 7). PGD2 was

markedly increased in sEH(�/�)/IL-10(�/�) mice ascompared to IL-10(�/�) mice (Figure 7, no statisti-cally significant). TXB2 and 5-HETE was mildly in-creased in sEH knockout mice (both sEH(�/�) miceand sEH(�/�)/IL-10(�/�) mice) as compared towild-type and IL-10(�/�) mice (Figure 7, no statisti-cally significant).

DISCUSSION

The studies have demonstrated the beneficialeffects of sEH gene deficiency on protection againstmyocardial ischemia-reperfusion injury, the inflam-matory response to cerebral ischemia, and on im-provement of glucose homeostasis using a powerfulsEH gene knockout model in mice [16–18]. In thepresent study, we have first reported that sEH genedeficiency in IL-10 knockout mice led to the reduc-tion of IBD inflammatory activity and IBD-inducedtumor development in the bowel.Long-standing chronic active IBD is at high risk

for cancer development. Several molecular eventsinvolved in chronic active inflammatory processescontribute to cancer development, includingthe overproduction of reactive oxygen and nitrogenspecies, aberrant metabolites of key arachidonicacid metabolism, overproduction of cytokines/chemokines, and dysfunction of immunity system[2]. Inflammatory cells (including neutrophils, mac-rophages, lymphocytes, and plasma cells, as well aseosinophils and mast cells) play a central role inthese molecular events. Reduction of inflammatorycell infiltration is crucial for suppressing inflamma-tory activity and inhibiting inflammation-inducedcarcinogenesis. sEH gene deficiency in IL-10(�/�)mice resulted in a significant decrease of inflamma-tory cell infiltration (mainly including neutrophilsand macrophages) in the inflamed areas ofbowel which was correlated with the reduction of

Figure 5. Western blot and densitometry analysis for NF-kB p65, phosphorylated NF-kB p65 (S276), IKK-a,and VCAM-1 for freshly collected colon mucosa from IL-10(�/�) and sEH(�/�)/IL-10(�/�) mice. (A) Westernblot assay, and (B) histogram of densitometry analysis. Specimens from male and female animals were pooledtogether within each group (n ¼ 6/group, 3 males and 3 females chosen randomly from each group). Thesignificant decreases of NF-kB p65, phosphorylated NF-kB p65 (S276), IKK-a, and VCAM-1 were seen in sEH(�/�)/IL-10(�/�) mice as compared with IL-10(�/�) mice. b-actin was used as internal control. Results werereported as mean � SD.

734 ZHANG ET AL.


IBD-induced tumor development. A simple mecha-nism involved in inhibiting inflammatory cell infil-tration in sEH(�/�)/IL-10(�/�) mice directly linksto abolish the hydrolysis of EETs by sEH knockoutthat results in an increase of EETs and further leadsto inhibit NF-kB and IKK signals, VCAM-1 expres-sion, and inflammatory cell infiltration [10].Our qPCR analysis further revealed a significant

decrease of mRNA expression of several key cyto-kines and chemokines including TNF-a, MCP-1, IL-12, IL-17, and IL-23. Although the mechanism forinhibition of these cytokines and chemokines bysEH knockout is not known, these inflammatorymediators play an important role in IBD. Studies inclinical and experimental IBD have shown the upre-gulation of various chemokines including TNF-aand MCP-1 in the colonic mucosal tissue. There area significant attenuation of colitis and reduction ofmortality in TNF-a and MCP-1-deficient mice. IL-12

is a well-studied proinflammatory cytokine thatdrives the pathogenic CD4 T cells respondingagainst the bacterial microflora [30]. IL-23 is an ad-ditional member of IL-12 family [31]. Studies in ani-mal models of IBD have revealed that IL-23 plays akey role in chronic intestinal inflammation. Selec-tive depletion of IL-23 using monoclonal antibodiesspecific for the p19 subunit or by genetic ablationof the p19 gene greatly attenuates T cell-dependentcolitis in T cell transfer model in mice [32,33]and inhibits spontaneous IBD development in IL-10(�/�) mice [34]. IL-23 drives an increased produc-tion of IL-17 and interferon (IFN)-g by both Tcells and non-T cells. In IBD, the major action of IL-17 is to promote additional inflammatory cascadesby stimulating the production of chemokinesleading to activate granulocytes[35,36], and to pro-mote development of IBD-associated colorectalcancer [37].

Figure 6. Metabolites of eicosanoid profile in plasma specimensanalyzed using an LC/MS-MS method. The levels of epoxygenase-dependent metabolites in plasma (A) EETs (including 11(12)-,14(15)-EET), (B) DHETs (including 11(12)-DHET, and 14(15)-DHET),and (C) the ratio of total EETs to DHETs. (D) EpOMEs (including9(10)-, 11(12)-EpOME), (E) DiHOMEs (including 9(10)-DiHOME, and11(12)-DiHOME), and (F) the ratio of total EpOME to DiHOME. Thestatistically significant difference was labeled in the figures in

comparison between wild-type mice and sEH (�/�) mice and be-tween sEH(�/�)/IL-10(�/�) mice and IL-10(�/�) mice. Especially,both EETs and EpOMEs were significantly increased, while DHETsand DiHOME were significantly decreased in sEH(�/�)/IL-10(�/�)mice compared to IL-10(�/�) mice (n ¼ 20 per group), therefore,the ratios of EETs/DHETs and EpOMEs/DiHOME were much more sig-nificantly increased. Results were reported as mean � SE.



NF-kB has been proposed to be a main molecularlink between inflammation and carcinogenesis. NF-kB is not only a key signal transmitting cytokine sig-nal cascade to the targeted cells such as endothelialcells for recruiting inflammatory cells into inflamedtissues, but also involved in driving tumor invasionand metastasis, particularly via its downstreamproducts of matrix metalloproteinases and serineproteases, as well as anti-apoptotic factors such asBcl-2 and Bcl-XL, etc [38]. Thus, downregulation ofNF-kB and IKK signals in IBD in sEH deficient micemay play a critical role in IBD and its induced tumordevelopment [39].

In addition to stabilization of EETs by abolishingsEH function through its gene knockout, whether ornot there is any effect on modulating other path-ways in arachidonic acid metabolism is an impor-tant concern. Our result revealed that LOX-mediated pathway was mildly regulated by sEH geneknockout and showed a mild increase of LTB4 and5-HETE in sEH(�/�) mice and sEH(�/�)/IL-10(�/�)mice. This finding may represent a feedback regula-tion of inflammatory metabolites of arachidonicacid metabolism. However, COX mediated keyproduct PGE2 appears no change between wild-type mice and sEH(�/�) mice. PGE2 production viaCOX-1 and COX-2 is a key event to promote inflam-mation-induced carcinogenesis [40,41] and PGE2production via COX-1 promotes intestinal stemcell survival and proliferation [42]. Our findingshowed that the high elevation of PGE2 in IBD inIL-10(�/�) mice was mildly decreased in IL-10(�/�)mice with sEH gene deficiency.

A recent report indicates that endothelium-derived EETs and its induced VEGF appear to pro-mote angiogenesis and tumor metastasis [43], whichis contradictory to our findings. However, severalstudies in the literatures and our present results in-dicate that targeting sEH may be more beneficial forpreventing inflammation-induced carcinogenesisrather than promoting it. Evidence includes (1) atleast one-third of human cancer are associated withchronic inflammatory process in which active in-flammatory cells (neutrophils and macrophages)play a central role in generating nitro-oxidativestress and inflammatory cytokines (such as TNF-a).Increase of EETs leads to decrease a production ofproinflammatory cytokines and reduce an infiltra-tion of inflammatory cells, which appears an impor-tant step for preventing inflammation-inducedcancer. (2) Endothelium-derived EETs and itsassociated angiogenesis are also important for theprocess of regeneration to repair inflammation-induced injury, particularly for IBD-induced erosionand ulceration. (3) Targeting TNF-a-induced VCAM-1 expression in endothelial cells by EETs is throughinhibition of transcriptional factor NF-kB and Ikk ki-nase [10]. Activation of NF-kB and Ikk kinase iscommonly observed in tumorigenesis [38]. Inhibi-tion of NF-kB and Ikk kinase by EETs or sEH knock-out is a crucial mechanism for preventing tumordevelopment. (4) Sorafenib is a multikinase inhibi-tor, including sEH [44]. It has been demonstratedthe beneficial effects of sorafenib on treatinghighly angiogenic malignances such as renal cellcarcinoma, as well the effect on blocking tumor

Figure 7. Representative metabolites of eicosanoid profile in COX and LOX pathways analyzed using an LC/MS-MS method. The levels of COX-dependent metabolites in plasma (A) PGE2, and PGD2, and (B) TXB2. Thelevels of LOX-dependent metabolites in plasma (C) LTB4 and (D) 5-HETE. Results were reported as mean � SE.The statistically significant difference was labeled in the figure.

736 ZHANG ET AL.


growth, angiogenesis and metastatic potential inpreclinical models [45,46]. (5) In the present study,using IBD-induced carcinogenesis in mice, we haveobserved a significant reduction of IBD-induced tu-mor development and no metastasis observedthrough extensively histopathological analysis forall key organs in IL-10(�/�) mice with sEHdeficiency.In conclusion, our results imply that sEH plays an

important role in IBD and its-induced carcinogene-sis. Simply, an increase of EETs to DHETs ratio in-cluding the stabilization of EETs and reduction ofDHETs, is a key function for sEH gene deficiency.These functional changes in sEH knockout micelead to attenuation of IBD inflammatory activitiesand its induced carcinogenesis, mechanistically viadownregulation of cytokines and chemokines, NF-kB signals and VCAM1. Our results may havebrought up a new direction of investigation on pre-ventive and therapeutic strategy for IBD via target-ing sEH.

ACKNOWLEDGMENTS

This study was supported by NIH R01 CA137467to GYY. Partial supports were provided by NIEHSR01 ES002710 and NIEHS Superfund Program P42ES004699 to BDH. BDH is a George and Judy MarcusSenior Fellow of The American Asthma Foundation.

REFERENCES

1. Norwood S, Liao J, Hammock BD, Yang GY. Epoxyeicosatrie-noic acids and soluble epoxide hydrolase: Potential therapeu-tic targets for inflammation and its induced carcinogenesis.Am J Transl Res 2010;2:447–457.

2. Yang GY, Taboada S, Liao J. Inflammatory bowel disease: Amodel of chronic inflammation-induced cancer. MethodsMol Biol 2009;511:193–233.

3. Chiamvimonvat N, Ho CM, Tsai HJ, Hammock BD. The solu-ble epoxide hydrolase as a pharmaceutical target for hyper-tension. J Cardiovasc Pharmacol 2007;50:225–237.

4. Huang H, Morisseau C, Wang J, et al. Increasing or stabiliz-ing renal epoxyeicosatrienoic acid production attenuates ab-normal renal function and hypertension in obese rats. Am JPhysiol Renal Physiol 2007;293:F342–F349.

5. Zhang W, Koerner IP, Noppens R, et al. Soluble epoxidehydrolase: A novel therapeutic target in stroke. J CerebBlood Flow Metab 2007;27:1931–1940.

6. Inceoglu B, Schmelzer KR, Morisseau C, Jinks SL, HammockBD. Soluble epoxide hydrolase inhibition reveals novel bio-logical functions of epoxyeicosatrienoic acids (EETs). Prosta-glandins Other Lipid Mediat 2007;82:42–49.

7. Schmelzer KR, Kubala L, Newman JW, Kim IH, Eiserich JP,Hammock BD. Soluble epoxide hydrolase is a therapeutictarget for acute inflammation. Proc Natl Acad Sci USA2005;102:9772–9777.

8. Smith KR, Pinkerton KE, Watanabe T, Pedersen TL, Ma SJ,Hammock BD. Attenuation of tobacco smoke-induced lunginflammation by treatment with a soluble epoxide hydrolaseinhibitor. Proc Natl Acad Sci USA 2005;102:2186–2191.

9. Zhao X, Yamamoto T, Newman JW, et al. Soluble epoxidehydrolase inhibition protects the kidney from hypertension-induced damage. J Am Soc Nephrol 2004;15:1244–1253.

10. Node K, Huo Y, Ruan X, et al. Anti-inflammatory propertiesof cytochrome P450 epoxygenase-derived eicosanoids. Sci-ence 1999;285:1276–1279.

11. Newman JW, Morisseau C, Hammock BD. Epoxide hydro-lases: Their roles and interactions with lipid metabolism.Prog Lipid Res 2005;44:1–51.

12. Morisseau C, Hammock BD. Epoxide hydrolases: Mecha-nisms, inhibitor designs, and biological roles. Annu RevPharmacol Toxicol 2005;45:311–333.

13. Fleming I. DiscrEET regulators of homeostasis: Epoxyeicosa-trienoic acids, cytochrome P450 epoxygenases and vascularinflammation. Trends Pharmacol Sci 2007;28:448–452.

14. Inceoglu B, Jinks SL, Schmelzer KR, Waite T, Kim IH, Ham-mock BD. Inhibition of soluble epoxide hydrolase reducesLPS-induced thermal hyperalgesia and mechanical allodyniain a rat model of inflammatory pain. Life Sci 2006;79:2311–2319.

15. Sinal CJ, Miyata M, Tohkin M, Nagata K, Bend JR, GonzalezFJ. Targeted disruption of soluble epoxide hydrolase revealsa role in blood pressure regulation. J Biol Chem 2000;275:40504–40510.

16. Manhiani M, Quigley JE, Knight SF, et al. Soluble epoxidehydrolase gene deletion attenuates renal injury and inflam-mation with DOCA-salt hypertension. Am J Physiol RenalPhysiol 2009;297:F740–F748.

17. Motoki A, Merkel MJ, Packwood WH, et al. Soluble epoxidehydrolase inhibition and gene deletion are protectiveagainst myocardial ischemia-reperfusion injury in vivo. Am JPhysiol Heart Circ Physiol 2008;295:H2128–2134.

18. Luria A, Bettaieb A, Xi Y, et al. Soluble epoxide hydrolasedeficiency alters pancreatic islet size and improves glucosehomeostasis in a model of insulin resistance. Proc Natl AcadSci USA 108:9038–9043.

19. Kuhn R, Lohler J, Rennick D, Rajewsky K, Muller W. Inter-leukin-10-deficient mice develop chronic enterocolitis. Cell1993;75:263–274.

20. Berg DJ, Zhang J, Weinstock JV, et al. Rapid developmentof colitis in NSAID-treated IL-10-deficient mice. Gastroenter-ology 2002;123:1527–1542.

21. Berg DJ, Davidson N, Kuhn R, et al. Enterocolitis and coloncancer in interleukin-10-deficient mice are associated withaberrant cytokine production and CD4(þ) TH1-likeresponses. J Clin Invest 1996;98:1010–1020.

22. Seril DN, Liao J, Lu G, Yang CS, Yang G-Y. Increased sus-ceptibility of ulcerative colitis-associated carcinoma develop-ment in Ogg1 knockout mice. The Proceeding of 96thAACR Annual Meeting, 2005.

23. Seril DN, Liao J, Yang CS, Yang GY. Systemic iron supple-mentation replenishes iron stores without enhancing coloncarcinogenesis in murine models of ulcerative colitis: Com-parison with iron-enriched diet. Dig Dis Sci 2005;50:696–707.

24. Mahler M, Bristol IJ, Leiter EH, et al. Differential susceptibili-ty of inbred mouse strains to dextran sulfate sodium-in-duced colitis. Am J Physiol 1998;274:G544–G551.

25. Vogel CF, Nishimura N, Sciullo E, Wong P, Li W, MatsumuraF. Modulation of the chemokines KC and MCP-1 by2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in mice. ArchBiochem Biophys 2007;461:169–175.

26. Deng X, Li H, Tang YW. Cytokine expression in respiratorysyncytial virus-infected mice as measured by quantitative re-verse-transcriptase PCR. J Virol Methods 2003;107:141–146.

27. Yang J, Schmelzer K, Georgi K, Hammock BD. Quantitativeprofiling method for oxylipin metabolome by liquid chroma-tography electrospray ionization tandem mass spectrometry.Anal Chem 2009;81:8085–8093.

28. Narushima S, Spitz DR, Oberley LW, et al. Evidence for oxi-dative stress in NSAID-induced colitis in IL10�/� mice. FreeRadic Biol Med 2003;34:1153–1166.



29. Luria A, Morisseau C, Tsai HJ, et al. Alteration in plasmatestosterone levels in male mice lacking soluble epoxide hy-drolase. Am J Physiol Endocrinol Metab 2009;297:E375–E383.

30. Bouma G, Strober W. The immunological and genetic basisof inflammatory bowel disease. Nat Rev Immunol 2003;3:521–533.

31. Hunter CA. New IL-12-family members: IL-23 and IL-27,cytokines with divergent functions. Nat Rev Immunol 2005;5:521–531.

32. Hue S, Ahern P, Buonocore S, et al. Interleukin-23 drivesinnate and T cell-mediated intestinal inflammation. J ExpMed 2006;203:2473–2483.

33. Kullberg MC, Jankovic D, Feng CG, et al. IL-23 plays a keyrole in Helicobacter hepaticus-induced T cell-dependent co-litis. J Exp Med 2006;203:2485–2494.

34. Yen D, Cheung J, Scheerens H, et al. IL-23 is essential for Tcell-mediated colitis and promotes inflammation via IL-17and IL-6. J Clin Invest 2006;116:1310–1316.

35. Alsabti EA. Serum alphafetoprotein in bladder carcinoma.Oncology 1977;34:78–79.

36. Ljujic B, Radosavljevic G, Jovanovic I, et al. Elevatedserum level of IL-23 correlates with expression of VEGF inhuman colorectal carcinoma. Arch Med Res 2010;41:182–189.

37. Fantini MC, Pallone F. Cytokines: From gut inflammation tocolorectal cancer. Curr Drug Targets 2008;9:375–380.

38. Karin M, Lin A. NF-kappaB at the crossroads of life anddeath. Nat Immunol 2002;3:221–227.

39. Karin M, Greten FR. NF-kappaB: Linking inflammation andimmunity to cancer development and progression. Nat RevImmunol 2005;5:749–759.

40. Boolbol SK, Dannenberg AJ, Chadburn A, et al. Cyclooxy-genase-2 overexpression and tumor formation are blockedby sulindac in a murine model of familial adenomatous pol-yposis. Cancer Res 1996;56:2556–2560.

41. Williams CS, Luongo C, Radhika A, et al. Elevated cyclooxy-genase-2 levels in Min mouse adenomas. Gastroenterology1996;111:1134–1140.

42. Cohn SM, Schloemann S, Tessner T, Seibert K, Stenson WF.Crypt stem cell survival in the mouse intestinal epithelium isregulated by prostaglandins synthesized through cyclooxy-genase-1. J Clin Invest 1997;99:1367–1379.

43. Panigrahy D, Edin ML, Lee CR, et al. Epoxyeicosanoids stim-ulate multiorgan metastasis and tumor dormancy escape inmice. J Clin Invest 2012;122:178–191.

44. Liu JY, Park SH, Morisseau C, Hwang SH, Hammock BD,Weiss RH. Sorafenib has soluble epoxide hydrolase inhibito-ry activity, which contributes to its effect profile in vivo. MolCancer Ther 2009;8:2193–2203.

45. Hutson TE, Figlin RA, Kuhn JG, Motzer RJ. Targeted thera-pies for metastatic renal cell carcinoma: An overview of tox-icity and dosing strategies. Oncologist 2008;13:1084–1096.

46. Pignochino Y, Grignani G, Cavalloni G, et al. Sorafenibblocks tumour growth, angiogenesis and metastatic poten-tial in preclinical models of osteosarcoma through a mecha-nism potentially involving the inhibition of ERK1/2, MCL-1and ezrin pathways. Mol Cancer 2009;8:118.

738 ZHANG ET AL.


reduction of inflammatory bowel diseaseinduced tumor ...molecular carcinogenesis 52:726–738 (2013)...

Documents