Transgenic mice rich in endogenous omega-3fatty acids are protected from colitisChristian A. Hudert*†‡, Karsten H. Weylandt†‡, Yan Lu§, Jingdong Wang*, Song Hong§, Axel Dignass†,Charles N. Serhan§, and Jing X. Kang*¶

*Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, MA 02114; †Department of Gastroenterology,Charite University Medicine, Virchow Campus, 13353 Berlin, Germany; and §Center for Experimental Therapeutics and Reperfusion Injury,Brigham and Women’s Hospital, Harvard Medical School, Boston, MA 02115

Edited by Charles A. Dinarello, University of Colorado Health Sciences Center, Denver, CO, and approved June 9, 2006 (received for reviewFebruary 15, 2006)

Omega-6 (n-6) and omega-3 (n-3) polyunsaturated fatty acids(PUFA) are the precursors of potent lipid mediators and play animportant role in regulation of inflammation. Generally, n-6 PUFApromote inflammation whereas n-3 PUFA have antiinflammatoryproperties, traditionally attributed to their ability to inhibit theformation of n-6 PUFA-derived proinflammatory eicosanoids.Newly discovered resolvins and protectins are potent antiinflam-matory lipid mediators derived directly from n-3 PUFA with distinctpathways of action. However, the role of the n-3 PUFA tissue statusin the formation of these antiinflammatory mediators has not beenaddressed. Here we show that an increased n-3 PUFA tissue statusin transgenic mice that endogenously biosynthesize n-3 PUFA fromn-6 PUFA leads to significant formation of antiinflammatory re-solvins and effective reduction in inflammation and tissue injury incolitis. The endogenous increase in n-3 PUFA and related productsdid not decrease n-6 PUFA-derived lipid mediators such as leuko-triene B4 and prostaglandin E2. The observed inflammation pro-tection might result from decreased NF-�B activity and expressionof TNF�, inducible NO synthase, and IL-1�, with enhanced muco-protection probably because of the higher expression of trefoilfactor 3, Toll-interacting protein, and zonula occludens-1. Theseresults thus establish the fat-1 transgenic mouse as a new exper-imental model for the study of n-3 PUFA-derived lipid mediators.They add insight into the molecular mechanisms of inflammationprotection afforded by n-3 PUFA through formation of resolvinsand protectins other than inhibition of n-6 PUFA-derived eico-sanoid formation.

inflammation � inflammatory bowel disease � lipid mediators �resolvins � protectins

The epithelial surface of the gut is the largest mucosal surface inmammals and is particularly exposed to microbial attacks and

trauma. The inflammatory bowel diseases (IBD) Crohn’s diseaseand ulcerative colitis are characterized by idiopathic relapses andremitting chronic inflammation. Modulating the formation ofproinflammatory mediators and�or antiinflammatory molecules isuseful in the treatment of IBD (1).

The omega-3 (n-3) fatty acids, particularly eicosapentaenoic acid(EPA) and docosahexaenoic acid (DHA), are implicated in treatingIBD (2–4), whereas the eicosanoids derived from the omega-6 (n-6)fatty acid arachidonic acid, such as prostaglandin E2 (PGE2) andleukotriene B4 (LTB4), are potent proinflammatory mediators(5–7). The increased incidence of IBD in humans correlates with anincreased dietary content of n-6 fatty acids (8, 9). In most West-erners the amounts of tissue n-3 fatty acids appear to be low,whereas the levels of n-6 fatty acids are high, with an n-6�n-3 ratioranging from 10:1 to 20:1 (10).

Newly identified lipid mediators produced from n-6 and n-3polyunsaturated fatty acids (PUFA), the aspirin-triggered lipoxins(11, 12) and their stable analogues (13), as well as the resolvin E1(RvE1) generated from EPA, have inflammation-dampening ef-fects in models of inflammation (14). In light of these findings the

direct contribution of n-3 fatty acid status itself to chronic diseaseprogression such as in IBD and to the generation of local inflam-matory mediators remains of interest.

Transgenic fat-1 mice, engineered to express the Caenorhabditiselegans fat-1 gene encoding an n-3 fatty acid desaturase, are capableof producing n-3 PUFA from n-6 PUFA and thereby have a low orbalanced ratio of n-6�n-3 fatty acids in their tissues and organswithout the need of dietary interventions (15). This model allowscarefully controlled studies to be performed in the absence ofrestricted diets, which can create confounding factors that limitstudies of this nature. Therefore, the transgenic mice offer theopportunity to address the molecular events underlying the bene-ficial impact of n-3 fatty acids.

Dextran sodium sulfate (DSS)-induced colitis is a well estab-lished experimental model of IBD used to study cytokine-triggeredinflammation and injury in the colon (16, 17) as well as othermechanisms of colitis such as thrombin-triggered pathways ofinflammation (18). DSS colitis is characterized histologically byinfiltration of inflammatory cells into the lamina propria, withlymphoid hyperplasia, focal crypt damage, and epithelial ulceration(16–19). These pathological changes are thought to develop as aresult of a barrier-destructive effect of DSS on the epithelium,subsequent phagocytosis of lamina propria cells, and production ofcytokines (16–19). Although the relationship of murine DSS-induced colitis to the human disease remains to be established, thiswidely used IBD model has a number of advantages, includingsimplicity, high degree of uniformity of the lesions, and leukocyteinfiltration (19). In the present report we used this model toexamine the impact of enhanced n-3 PUFA tissue status on thedevelopment of colitis in the fat-1 transgenic mice and relationshipto resolvins and protectins.

ResultsFatty Acid Profiles of Colon Tissues. Both WT and fat-1 transgeniclittermates born to the same mother were maintained on a diet(10% safflower oil) high in n-6 and low in n-3 PUFA (n-6�n-3 �20). During this dietary regime, fat-1 transgenic mice had signifi-cantly higher amounts of n-3 PUFA, such as EPA, DPA, and DHA,in all organs and tissues including the colon compared with WTmice (Fig. 1 and Table 1). The ratio of the long-chain n-6 PUFA(20:4 n-6, 22:4 n-6, and 22:5 n-6) to the long-chain n-3 PUFA (20:5

Conflict of interest statement: No conflicts declared.

This paper was submitted directly (Track II) to the PNAS office.

Abbreviations: n-3, omega-3; n-6, omega-6; PUFA, polyunsaturated fatty acids; EPA, eico-sapentaenoic acid; DHA, docosahexaenoic acid; IBD, inflammatory bowel disease; RvE1,resolvin E1; RvD3, resolvin D3; NPD1, neuroprotectin D1; PD1, protectin D1; LTB4, leuko-triene B4; LTB5, leukotriene B5; PGE3, prostaglandin E3; PGE2, prostaglandin E2; TFF3,trefoil factor 3; Tollip, Toll-interacting protein; ZO-1, zonula occludens-1; DSS, dextransodium sulfate.

‡C.A.H. and K.H.W. contributed equally to this work.

¶To whom correspondence should be addressed at: Massachusetts General Hospital, 14913th Street, Room 4433, Charlestown, MA 02129. E-mail: jxkang@partners.org.

© 2006 by The National Academy of Sciences of the USA

11276–11281 � PNAS � July 25, 2006 � vol. 103 � no. 30 www.pnas.org�cgi�doi�10.1073�pnas.0601280103

Dow

nloa

ded

by g

uest

on

Nov

embe

r 25

, 202

0

n-3, 22:5 n-3, and 22:6 n-3) was 1.7 in fat-1 transgenics and 30.1 inWT mice. Apparently, although both WT and fat-1 eat the samediet, their body fatty acid profiles are distinct.

fat-1 Transgenic Mice Are Protected Against DSS-Induced Colitis.Induction of colitis resulted in significant changes in body weight,stool consistence, appearance of fecal blood, and general status,typically associated with human and experimental DSS colitis.fat-1 mice showed significantly less body weight loss (Fig. 2D)and a delayed progression of diarrhea but no apparent change infecal bleeding. Interestingly, fat-1 transgenic mice showed arecovery beginning from the second day after stop of DSSexposure whereas WT mice exhibited a continued loss of bodyweight during the 3 days after cessation of DSS (Fig. 2D). Theseclinical manifestations were reflected in the macroscopic patho-logical changes. Multiple adhesions, strictures, and a massivethickening of the colon were observed in WT mice but not infat-1 animals (Fig. 2 A). Furthermore, colon shorteningamounted to 35% in WT mice but only 15% in fat-1 transgenicmice when compared with that of untreated control mice (Fig.2B). Microscopic assessment of the distal part of the colon

revealed that severity and thickness of the inflammatory infil-trate as well as the extent of epithelial damage were significantlyalleviated in fat-1 mice (Fig. 2 A). All hallmarks of colitis werereduced in fat-1 mice except for minor punctate erosions and fewulcerations. In contrast, WT mice showed a massive fibrinousexudate on the luminal surface and marked epithelial infiltrateof leukocytes, as well as severe submucosal edema and diffuseulcerations of the mucosa (Fig. 2C). These findings indicate thatfat-1 transgenic mice, rich in n-3 fatty acids, are protected frominflammation.

Formation of n-3-Derived Antiinflammatory Mediators. Newly iden-tified potent n-3 fatty acid-derived mediators such as the resolvins,including RvE1 and resolvin D3 (RvD3), and protectins, i.e.,neuroprotectin D1 (NPD1)�protectin D1 (PD1), are antiinflam-matory (20, 21). We assessed both n-6 and n-3 PUFA-derivedmediators from colons using liquid chromatography–UV–tandemMS mediator informatics to determine whether the difference inDSS-induced colitis observed between WT and fat-1 mice wasassociated with these pathways. The n-3-derived mediators, includ-ing RvE1, RvD3, and NPD1�PD1, were identified in physiologicallyactive levels within colons of fat-1 transgenics (Fig. 3). Thesemediators were not found in the WT colons. Also, both 17-hydroxy-DHA and 14-hydroxy-DHA as pathway markers of DHA utiliza-tion (22) were identified in fat-1 mice. In addition to the resolvinlipid mediators, significant amounts of n-3 PUFA-derived prosta-glandin E3 (PGE3) and leukotriene B5 (LTB5) were formed in fat-1mice. There were no significant differences in the levels of LTB4,a potent chemoattractant, and the proinflammatory PGE2. Simi-larly, there was no significant difference in the formation of15-hydroxyeicosatetraenoic acid, the precursor for the n-6 PUFA-derived antiinflammatory lipoxin A4, between WT and fat-1 mice.

Expression of Genes Involved in Inflammation and Colitis Pathogen-esis. We next examined whether the protection from colitis ob-served in fat-1 transgenics had an impact on inflammation-relatedgene expression. TNF� plays a critical role in IBD, and its over-expression is associated with an IBD-like phenotype in mice (23).Concordant with the protective action of the increased n-3 PUFAstatus was a decrease in NF-�B protein activity, as determined byactivated p65 protein (Fig. 4A) as well as in TNF� mRNA levels(Fig. 4B). In addition, transcription of other prominent inflamma-tory markers, such as inducible NO synthase and IL-1� (24), wasdampened in the transgenic fat-1 mice (Fig. 4 C and D).

In addition, we observed that the mRNA levels of intestinaltrefoil factor 3 (TFF3), a factor important in maintenance andrepair of the intestinal mucosa (25), was increased in the colonsof fat-1 mice (Fig. 4F). The intercellular tight junction protein

Table 1. PUFA profiles of colons from WT and fat-1 mice

PUFAWT

(n � 3)fat-1

(n � 4)

n-6 (%)AA (20:4 n-6) 12.66 � 2.94 12.47 � 1.94DTA (22:4 n-6) 3.02 � 0.33 2.17 � 0.38DPA (22:5 n-6) 2.95 � 1.13 0.57 � 0.22Total 18.62 � 4.85 15.21 � 2.22

n-3 (%)EPA (20:5 n-3) 0.06 � 0.08 2.03 � 0.47DPA (22:5 n-3) 0.12 � 0.15 2.45 � 0.55DHA (22:6 n-3) 0.43 � 0.08 4.68 � 0.36Total 0.62 � 0.15 9.86 � 1.17

n-6�n-3 (of total fractions) 30.13 1.66

DTA, docosatetraenoic acid; DPA, docosapentaenoic acid; AA, arachidonicacid.

Fig. 1. Differential fatty acid profiles in WT and fat-1 transgenic mice.Whereas high levels of n-6 fatty acids characterize WT samples (A), n-3 fattyacids are nearly absent. In contrast, an abundance of EPA (20:5 n-3), docosa-pentaenoic acid (22:5 n-3), and DHA (22:6 n-3) can be found in fat-1 transgenicmice (B). The n-3 PUFA are marked with asterisks.

Hudert et al. PNAS � July 25, 2006 � vol. 103 � no. 30 � 11277

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Nov

embe

r 25

, 202

0

zonula occludens 1 (ZO-1), which is important in epithelialintegrity (26), was also sustained in fat-1 transgenic animals (Fig.4G). Furthermore, mRNA levels of Toll-interacting protein(Tollip), a downstream inhibitor of the Toll-like receptor path-way that mediates inflammatory response (27), were higher infat-1 transgenic mice (Fig. 4E). These results suggest an en-hanced defense status in the fat-1 mice.

DiscussionThe results presented here clearly show that the inflammation ofcolon induced by DSS, in terms of both clinical manifestation and

pathology, is significantly less severe in fat-1 transgenic mice thanthat in WT littermates. The protection from colitis in fat-1 mice iscorrelated with the formation of antiinflammatory derivates of n-3fatty acids, down-regulation of proinflammatory cytokines, andup-regulation of mucoprotective factors in the colons of theseanimals. These findings suggest a role played by an increased tissuestatus of n-3 fatty acids in protection against colitis throughalterations of gene expression mediated, probably, by antiinflam-matory lipid mediators of the n-3 fatty acids.

Although a number of previous studies have examined theeffectiveness of n-3 fatty acids in prevention and treatment of colitis(2–4), the outcomes were inconsistent or conflicting. The discrep-ancy may be caused by the confounding factors of diet or n-3 fattyacid supplements. In fact, many variables can arise from the dietsand the feeding procedure, including the impurity or unwantedcomponents of the oils used (e.g., fish oil versus corn oil), flavor,sensitivity to oxidation, diet storage, duration of diet change, etc.,which can impose confounding effects on the fatty acid ratio. Incontrast, the genetic approach using the fat-1 gene, as presentedhere, only modifies the n-6�n-3 fatty acid ratio (converts n-6 to n-3)endogenously and thereby allows experimental animals (i.e., WTand transgenic littermates) to be fed with a single diet. Therefore,the results obtained by using this model are more reliable.

The n-3 fatty acids may exert an antiinflammatory effect viacompetitive inhibition of the n-6 (arachidonic acid)-derived proin-flammatory eicosanoids, most notably LTB4 and PGE2, and this hasbeen the mechanism mostly proposed in the past to explain thebiological effectiveness of n-3 fatty acids (28). In this study we foundno significant differences in the content of arachidonic acid, LTB4,and PGE2 between fat-1 transgenic and WT mice. There was aremarkable difference in the amounts of n-3 fatty acids (EPA,DHA, and precursors) and their potent bioactive products, theresolvins and protectins (RvE1, RvD3, and PD1�NPD1), as well asthe n-3 PUFA-derived LTB5 and PGE3 (Fig. 3). It is possible thatLTB5 and PGE3 could exert an antiinflammatory effect throughcompetition with LTB4 and PGE2. However, given the much higherabsolute concentrations of LTB4 and PGE2 present in the fat-1 micethis may not be the major underlying mechanism.

The newly identified n-3-derived resolvins and protectins arepotent antiinflammatory mediators in various settings, includingtrinitrobenzene sulfonate colitis and periodontitis (20, 21, 29, 30).These lipid mediators have been shown to decrease formation ofinflammatory cytokines such as TNF�, IL-6, and others (31, 32).Our data confirm decreased levels of cytokines, which may be dueto the documented lower NF-�B activity in fat-1 mice shown here.Indeed, RvE1 has been shown to inhibit NF-�B activation throughits specific G protein-coupled receptor, ChemR23 (14). Our dataare consistent with this mechanism of action. In view of these resultsthe n-3 fatty acid-derived mediators documented in the fat-1transgenic mice might link an enhanced n-3 PUFA status toinflammation dampening.

Although the pathogenesis of DSS colitis appears to be Tcell-independent (33), cytokine production plays a key role in thedevelopment of colitis in this model (16, 17). In fact, cytokinesare important mediators of inflammation (34). Increased pro-duction of proinflammatory cytokines such as TNF�, IL-1�,IL-12, and IFN-� are found in inflamed colons from patientswith IBD (35) as well as DSS colitis animals (16, 17). Furtherevidence for the involvement of these cytokines came from theobservations that antibodies against TNF� and IL-12 reducedthe severity of the disease in the animal model of DSS-inducedcolitis (36, 37) as well as patients with Crohn’s disease (38). Thus,reduction in the production of proinflammatory cytokines ap-pears to be an effective approach to the prevention and treat-ment of IBD. In the present study, the colons of fat-1 mice, richin n-3 fatty acids, exhibited significantly lower amounts of theproinflammatory cytokines TNF� and IL-1� than those fromWT animals and were protected from DSS-induced colitis.

Fig. 2. Colon inflammation activity in WT and fat-1 transgenic mice. (A)Macroscopic view (Upper) and microscopic hematoxylin and eosin staining(Lower) of the distal colon in WT control mice (Left), DSS-treated WT non-transgenic littermates (Center), and fat-1 mice (Right). (B) Colon shortening asa hallmark of DSS-induced colonic damage is reduced in fat-1 mice. *, P � 0.01versus WT DSS-treated animals. (C) Histopathological scores for colonic in-flammatory infiltration and epithelial damage in WT and fat-1 mice. *, P �0.01 versus WT DSS. (D) Body weight change from 100% baseline over 8 daysin fat-1 mice and WT littermates (n � 6 for each group), 5 days of DSStreatment and 3 days of normal drinking water. *, P � 0.05 versus WT DSS; **,P � 0.01 versus WT DSS. Mice were killed on day 8 (arrow), and samples weretaken for further analysis.

11278 � www.pnas.org�cgi�doi�10.1073�pnas.0601280103 Hudert et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 25

, 202

0

Along these lines, earlier animal and human results showed thatdietary supplementation with n-3 fatty acids reduced the pro-duction of cytokines, including TNF� and IL-1 (39). This findingsuggests that the inflammation protection observed in fat-1 mice

may be, in part, the result of the reduction in cytokine productionand action by n-3-derived resolvins.

The results of the present study also support a role for n-3 fattyacids in the maintenance of intestinal integrity, as demonstrated by

Fig. 3. LC–UV–tandem MS profiles of n-3 PUFA-derived lipid mediators. (A) DHA-derived resolvins and protectins (main pathway products identified were RvD3and PD1�NPD1). (B) EPA-derived bioactive lipid mediators (identified mediators include RvE1, PGE3, and LTB5 ). (C) Arachidonic acid-derived bioactive mediators[PGE2, LTB4, and 15-hydroxyeicosatetraenoic acid (15-HETE) as precursor for the n-6 PUFA-derived lipoxin A4 (LXA4)]. (D) Presence of different lipid mediatorsin colon samples of fat-1 transgenic mice (n � 6) and WT animals (n � 6). **, P � 0.01; *, P � 0.05. Note the different scale for 15-HETE and PGE2 (on the right).

Hudert et al. PNAS � July 25, 2006 � vol. 103 � no. 30 � 11279

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Nov

embe

r 25

, 202

0

higher levels of the mucosal protective factors Tollip, TFF3, andZO-1 in the colon tissues of fat-1 transgenic mice. The innateimmune system maintains a steady state of physiologic inflamma-tion in coexistence with the luminal commensal bacteria. Toll-likereceptors sense components of these microorganisms (e.g., LPS)and lead to a delicately regulated downstream signaling cascade thatbalances an appropriate mucosal response by production of pro-tective factors (40) or inflammatory mediators (41). In the unin-flamed colon a state of reduced sensitivity to bacterial products likeLPS inhibits an exaggerated activation of the transcriptional factorNF-�B and the consecutive proinflammatory stimuli (27). A dys-regulation of this balancing system may contribute to the severityand chronification of intestinal inflammation. It is possible that n-3fatty acids may preserve this system, because our results showedthat the inhibitory Tollip, a downstream regulator of the Toll-likereceptor pathway, was markedly reduced in WT mice but sustainedin fat-1 transgenic mice. Intestinal TFF3 is secreted by goblet cellsthroughout the entire colon onto the luminal surface in physiologicconditions. Under disease conditions TFF3 promotes epithelial cellmigration into damaged areas to subserve the reestablishment ofmucosal integrity. In this context, mice lacking intestinal trefoilfactor may suffer from an impaired epithelial defense and are morevulnerable to inflammatory injury. Our results show a favorableeffect of n-3 fatty acids on this system, as evidenced by theup-regulation of TFF3 in fat-1 transgenic mice. Thus, it seems thatup-regulation or maintenance of mucoprotective factors (TFF3 andTollip) in the n-3 PUFA-enriched tissues may be one of theunderlying mechanisms for the observed protection against colitisin fat-1 mice. However, the molecular mechanisms remain forfurther study.

In short, the present results demonstrate that colon tissue with anincreased n-3 PUFA status generates higher levels of bioactive n-3

PUFA-derived lipid mediators (resolvins and protectins), whichmay, on one hand, suppress the inflammatory response and, on theother hand, enhance mucoprotection (defense of intestinal mu-cosa) and is thereby protected against inflammation and injury.

MethodsMice. Transgenic fat-1 mice were created as in ref. 15 and subse-quently backcrossed onto a C57BL background. Generations ofheterozygous fat-1 mice were then mated to obtain WT andheterozygous�homozygous transgenic mice. In this study, all trans-genic fat-1 mice used were heterozygous. Animals were kept underspecific pathogen-free conditions in standard cages and were fed aspecial diet (10% safflower oil) high in n-6 and low in n-3 fatty acidsuntil the desired age (9–10 weeks) and weight (19–21 g). Each cagehoused two weight-matched female mice, combining one WT andone fat-1 transgenic mouse. All studies were approved by theMassachusetts General Hospital Subcommittee on Research An-imal Care.

Induction of Colitis. Colitis was induced in both WT and transgenicmice by addition of 3% (wt�vol) DSS (molecular weight 35,000–40,000; ICN Biomedicals) to sterile drinking water. On day 5, DSSsupplementation was discontinued, and mice were killed on day 8(3 days after cessation of DSS administration). Clinical assessmentof all DSS-treated animals for body weight, stool consistency, rectalbleeding, and general appearance was performed daily. Mice wereweighed twice at designated time points each day. Mice were killedon day 8. Colons were excised, and their length and thickening weredocumented. Histological examination was performed in a blindedmanner, and the degree of inflammation and epithelial damage onmicroscopic cross-sections of the colon was graded by an experi-enced pathologist. The inflammation score is a combined score of

Fig. 4. Markers of inflammation and mucoprotection. (A) NF-�B activation reflected in p65 ELISA activity shows significant differences in control baselines andin disease between WT and fat-1 mice. *, P � 0.05 versus WT DSS; **, P � 0.05 versus WT control. (B–F) Semiquantitative real-time PCR analysis of mRNA expressionlevels of inflammatory mediators TNF�, inducible NO synthase (iNOS), and IL-1� (B–D) and mucoprotective factors Tollip and TFF3 (E and F) in colons from WTand fat-1 mice after DSS exposure and fat-1 control mice, normalized as fold increase to the baseline of WT controls (dashed line). *, P � 0.05 versus WT DSS;

**, P � 0.01 versus WT DSS. (G) ZO-1 expression profile. Compared with WT mice without treatment (Left), ZO-1 expression is down-regulated on the luminalepithelial surface in WT mice on day 4 (Center), whereas luminal continuity of expression is sustained in fat-1 mice (Right).

11280 � www.pnas.org�cgi�doi�10.1073�pnas.0601280103 Hudert et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 25

, 202

0

(i) severity of inflammation (0 � no inflammation, 1 � mild, 2 �moderate, and 3 � severe) and (ii) thickness of inflammatoryinvolvement (0 � no inflammation, 1 � mucosa, 2 � mucosa plussubmucosa, and 3 � transmural); epithelial damage score consistsof character (0 � intact epithelium, 1 � disruption of architecturalstructure, 2 � erosion, and 3 � ulceration) and extent of lesions(0 � no lesions, 1 � punctuate, 2 � multifocal, and 3 � diffuse).

Immunofluorescence. Colon tissue was fresh-frozen in OCT me-dium, and sections were cut at 4-�m thickness. After air-drying,they were incubated with ZO-1 primary antibody (1:100 dilution;Zymed) for 30 min at room temperature in a moist chamber,rinsed with PBS, and incubated with an Alexa Fluor 488FITC-conjugated secondary antibody (1:50 dilution; MolecularProbes) in the same manner. Sections were mounted withGlycergel mounting medium (Dako) and evaluated with a ZeissLSM 5 Pascal confocal microscope.

Semiquantitative Real-Time PCR. Total RNA was isolated fromwhole colon tissue using TRIzol reagent (Invitrogen Life Technol-ogies) following the manufacturer’s instructions. RNA concentra-tions and purity were determined spectrometrically by their absor-bance at 260 nm in relation to the absorbance at 280 nm. Reversetranscription of mRNA was performed by using random hexamerprimers. Real-time PCR was carried out by using SYBR Green ina PRISM 9000 Light Cycler (Applied Biosystems) following themanufacturer’s protocol. All samples were processed in triplicate,and means were standardized to GAPDH values. Results areexpressed as a fold induction of the WT controls.

NF-�B Activation. To quantify the activated p65�RelA protein,TransAM NF-�B p65 Activation Assay (Active Motif) was per-formed according to the manufacturer’s instructions. Nuclear ex-tracts from whole colon tissues were collected by using NE-PER(Pierce), and protein concentrations were determined by using aCoomassie Plus Assay. Lysates (13 �g of total protein) wereincubated at room temperature for 1 h in 96-well dishes containingimmobilized oligonucleotides that comprise the NF-�B consensusDNA binding site (5�-GGGACTTTCC-3�). Consecutively, theprimary antibody against p65 and the horseradish peroxidase-

conjugated secondary antibody were incubated in the same man-ner, separated by washing steps. The reaction was developed for 5min at room temperature, and its intensity was measured immedi-ately at 450 nm by using a microplate reader.

Analysis of PUFA and Lipid Mediators. Fatty acid profiles wereanalyzed by using gas chromatography as described previously (42).Briefly, fresh colon tissues were grounded to powder under liquidnitrogen and subjected to extraction of total lipids and fatty acidmethylation by heating at 100°C for 1 h under 14% boron trifluo-ride–methanol reagent. Fatty acid methyl esters were analyzed bygas chromatography using a fully automated HP5890 systemequipped with a flame-ionization detector. Peaks of resolved fattyacids were identified by comparison with fatty acid standards (NuChek Prep, Elysian, MN), and area percentage for all resolvedpeaks was analyzed by using a PerkinElmer M1 integrator.

Lipid mediators from n-3 and n-6 fatty acids were measured byusing LC–tandem MS methods as in ref. 14. Samples were extractedwith 2 ml of cold methanol and analyzed by LC–UV–tandem MSusing a Finnigan LCQ liquid chromatography ion trap tandem massspectrometer equipped with a LUNA C18–2 (100 � 2 mm � 5 �m)column and UV diode array detector using mobile phase (meth-anol:water:acetate at 65:35:0.01) with a 0.2 ml�min flow rate.

Statistical Analysis. Data analysis was performed with Prism 3.02software (GraphPad). Comparison was made by using the Stu-dent t test. All values are presented as the mean � SEM or asindicated. Statistical significance was set at P � 0.05.

We thank Anett Rexin, Trisha della Pelle, Jan Niess, Dr. Ian Sanderson,and Dr. Atul Bhan for the kind support. This work was supported bygrants from the German Academic Exchange Service (BiomedicalExchange Program, International Academy of Lifesciences) and a Stu-dent Research Fellowship Award of the Crohn’s and Colitis Foundationof America (to C.A.H.), grants from the American Cancer Society(RSG-03-140-01-CNE) and the Center for the Study of InflammatoryBowel Disease at Massachusetts General Hospital (DK43351) (toJ.X.K.), and a grant from the National Institutes of Health (P50-DE016191) (to C.N.S.).

1. Podolsky, D. K. (2002) N. Engl. J. Med. 347, 417–429.2. Belluzzi, A., Brignola, C., Campieri, M., Pera, A., Boschi, S. & Miglioli, M. (1996) N. Engl.

J. Med. 334, 1557–1560.3. Endres, S., Lorenz, R. & Loeschke, K. (1999) Curr. Opin. Clin. Nutr. Metab. Care 2, 117–120.4. Stenson, W. F., Cort, D., Rodgers, J., Burakoff, R., DeSchryver-Kecskemeti, K., Gramlich,

T. L. & Beeken, W. (1992) Ann. Intern. Med. 116, 609–614.5. Haeggstrom, J. Z. (2004) J. Biol. Chem. 279, 50639–50642.6. Samuelsson, B., Granstrom, E., Green, K., Hamberg, M. & Hammarstrom, S. (1975) Annu.

Rev. Biochem. 44, 669–695.7. Kunikata, T., Yamane, H., Segi, E., Matsuoka, T., Sugimoto, Y., Tanaka, S., Tanaka, H.,

Nagai, H., Ichikawa, A. & Narumiya, S. (2005) Nat. Immunol. 6, 524–531.8. Shoda, R., Matsueda, K., Yamato, S. & Umeda, N. (1996) Am. J. Clin. Nutr. 63, 741–745.9. Sakamoto, N., Kono, S., Wakai, K., Fukuda, Y., Satomi, M., Shimoyama, T., Inaba, Y.,

Miyake, Y., Sasaki, S., Okamoto, K., et al. (2005) Inflamm. Bowel Dis. 11, 154–163.10. Simopoulos, A. P. (2002) Biomed. Pharmacother. 56, 365–379.11. Gewirtz, A. T., Collier-Hyams, L. S., Young, A. N., Kucharzik, T., Guilford, W. J., Parkinson,

J. F., Williams, I. R., Neish, A. S. & Madara, J. L. (2002) J. Immunol. 168, 5260–5267.12. Paul-Clark, M. J., Van Cao, T., Moradi-Bidhendi, N., Cooper, D. & Gilroy, D. W. (2004)

J. Exp. Med. 200, 69–78.13. Fiorucci, S., Wallace, J. L., Mencarelli, A., Distrutti, E., Rizzo, G., Farneti, S., Morelli, A.,

Tseng, J. L., Suramanyam, B., Guilford, W. J. & Parkinson, J. F. (2004) Proc. Natl. Acad.Sci. USA 101, 15736–15741.

14. Arita, M., Bianchini, F., Aliberti, J., Sher, A., Chiang, N., Hong, S., Yang, R., Petasis, N. A.& Serhan, C. N. (2005) J. Exp. Med. 201, 713–722.

15. Kang, J. X., Wang, J., Wu, L. & Kang, Z. B. (2004) Nature 427, 504.16. Siegmund, B., Fantuzzi, G., Rieder, F., Gamboni-Robertson, F., Lehr, H. A., Hartmann, G.,

Dinarello, C. A., Endres, S. & Eigler, A. (2001) Am. J. Physiol. 281, R1264–R1273.17. Dieleman, L. A., Palmen, M. J., Akol, H., Bloemena, E., Pena, A. S., Meuwissen, S. G. &

Van Rees, E. P. (1998) Clin. Exp. Immunol. 114, 385–391.18. Vergnolle, N., Cellars, L., Mencarelli, A., Rizzo, G., Swaminathan, S., Beck, P., Steinhoff,

M., Andrade-Gordon, P., Bunnett, N. W., Hollenberg, M. D., et al. (2004) J. Clin. Invest. 114,1444–1456.

19. Elson, C. O., Sartor, R. B., Tennyson, G. S. & Riddell, R. H. (1995) Gastroenterology 109,1344–1367.

20. Serhan, C. N., Gotlinger, K., Hong, S. & Arita, M. (2004) Prostaglandins Other Lipid Mediat.73, 155–172.

21. Serhan, C. N., Gotlinger, K., Hong, S., Lu, Y., Siegelman, J., Baer, T., Yang, R., Colgan, S. P.& Petasis, N. A. (2006) J. Immunol. 176, 1848–1859.

22. Hong, S., Gronert, K., Devchand, P. R., Moussignac, R. L. & Serhan, C. N. (2003) J. Biol.Chem. 278, 14677–14687.

23. Neurath, M. F., Fuss, I., Pasparakis, M., Alexopoulou, L., Haralambous, S., Meyer zumBuschenfelde, K. H., Strober, W. & Kollias, G. (1997) Eur. J. Immunol. 27, 1743–1750.

24. Beck, P. L., Xavier, R., Wong, J., Ezedi, I., Mashimo, H., Mizoguchi, A., Mizoguchi, E.,Bhan, A. K. & Podolsky, D. K. (2004) Am. J. Physiol. 286, G137–G147.

25. Mashimo, H., Wu, D. C., Podolsky, D. K. & Fishman, M. C. (1996) Science 274, 262–265.26. Kucharzik, T., Walsh, S. V., Chen, J., Parkos, C. A. & Nusrat, A. (2001) Am. J. Pathol. 159,

2001–2009.27. Otte, J. M., Cario, E. & Podolsky, D. K. (2004) Gastroenterology 126, 1054–1070.28. James, M. J., Gibson, R. A. & Cleland, L. G. (2000) Am. J. Clin. Nutr. 71, 343S–348S.29. Arita, M., Yoshida, M., Hong, S., Tjonahen, E., Glickman, J. N., Petasis, N. A., Blumberg,

R. S. & Serhan, C. N. (2005) Proc. Natl. Acad. Sci. USA 102, 7671–7676.30. Hasturk, H., Kantarci, A., Ohira, T., Arita, M., Ebrahimi, N., Chiang, N., Petasis, N. A.,

Levy, B. D., Serhan, C. N. & Van Dyke, T. E. (2006) FASEB J. 20, 401–403.31. Bannenberg, G. L., Chiang, N., Ariel, A., Arita, M., Tjonahen, E., Gotlinger, K. H., Hong,

S. & Serhan, C. N. (2005) J. Immunol. 174, 4345–4355.32. Ariel, A., Li, P. L., Wang, W., Tang, W. X., Fredman, G., Hong, S., Gotlinger, K. H. &

Serhan, C. N. (2005) J. Biol. Chem. 280, 43079–43086.33. Dieleman, L. A., Ridwan, B. U., Tennyson, G. S., Beagley, K. W., Bucy, R. P. & Elson, C. O.

(1994) Gastroenterology 107, 1643–1652.34. Dinarello, C. A. (2000) Chest 118, 503–508.35. Pullman, W. E., Elsbury, S., Kobayashi, M., Hapel, A. J. & Doe, W. F. (1992) Gastroen-

terology 102, 529–537.36. Murthy, S., Cooper, H. S., Yoshitake, H., Meyer, C., Meyer, C. J. & Murthy, N. S. (1999)

Aliment. Pharmacol. Ther. 13, 251–260.37. Neurath, M. F., Fuss, I., Kelsall, B. L., Stuber, E. & Strober, W. (1995) J. Exp. Med. 182,

1281–1290.38. Sandborn, W. J. & Hanauer, S. B. (1999) Inflamm. Bowel Dis. 5, 119–133.39. Meydani, S. N. & Dinarello, C. A. (1993) Nutr. Clin. Pract. 8, 65–72.40. Rakoff-Nahoum, S., Paglino, J., Eslami-Varzaneh, F., Edberg, S. & Medzhitov, R. (2004)

Cell 118, 229–241.41. Cario, E. (2003) Expert Rev. Mol. Med. 2003, 1–15.42. Kang, J. X. & Wang, J. (2005) BMC Biochem. 6, 5.

Hudert et al. PNAS � July 25, 2006 � vol. 103 � no. 30 � 11281

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Nov

embe

r 25

, 202

0