C

teA

mt

D

sd

GASTROENTEROLOGY 2013;144:989–1000

The Unfolded Protein Response and Chemical Chaperones ReduceProtein Misfolding and Colitis in Mice

STEWART SIYAN CAO,1,2 ELLEN M. ZIMMERMANN,3 BRANDY–MENGCHIEH CHUANG,1 BENBO SONG,4

ANOSIKE NWOKOYE,5 J. ERBY WILKINSON,6 KATHRYN A. EATON,6,7 and RANDAL J. KAUFMAN1,2

1Del E. Webb Neuroscience, Aging and Stem Cell Research Center, Sanford Burnham Medical Research Institute, La Jolla, California; Departments of 2Biologicalhemistry and 7Microbiology and Immunology and 3Division of Gastroenterology, Department of Internal Medicine, University of Michigan Medical Center, Ann Arbor,

4NGM Biopharmaceuticals, Inc, South San Francisco, California; 5University of Michigan Comprehensive Cancer Center Tissue Core, Ann Arbor, Michigan;

Michigan;and 6Unit for Laboratory Animal Medicine, University of Michigan, Ann Arbor, Michigan

aeItfaepga

spspp

Ist

fIi1po

BA

SIC

AN

DTR

AN

SLA

TIO

NA

LA

T

BACKGROUND & AIMS: Endoplasmic reticulum(ER) stress has been associated with development ofinflammatory bowel disease. We examined the effects of ERstress–induced chaperone response and the orally activechemical chaperones tauroursodeoxycholate (TUDCA) and4-phenylbutyrate (PBA), which facilitate protein folding andreduce ER stress, in mice with colitis. METHODS: Weused dextran sulfate sodium (DSS) to induce colitis inmice that do not express the transcription factorATF6� or the protein chaperone P58IPK. We examinedhe effects of TUDCA and PBA in cultured intestinalpithelial cells (IECs); in wild-type, P58IPK�/�, andtf6��/� mice with colitis; and in Il10�/� mice.

RESULTS: P58IPK�/� and Atf6��/� mice developedore severe colitis following administration of DSS

han wild-type mice. IECs from P58IPK�/� mice hadexcessive ER stress, and apoptotic signaling was acti-vated in IECs from Atf6��/� mice. Inflammatory stimuliinduced ER stress signals in cultured IECs, which werereduced by incubation with TUDCA or PBA. Oral ad-ministration of either PBA or TUDCA reduced featuresof DSS-induced acute and chronic colitis in wild-typemice, the colitis that develops in Il10�/� mice, and

SS-induced colitis in P58IPK�/� and Atf6��/� mice.Reduced signs of colonic inflammation in these micewere associated with significantly decreased ER stress incolonic epithelial cells. CONCLUSIONS: The unfoldedprotein response induces expression of genes that en-code chaperones involved in ER protein folding; thesefactors prevent induction of colitis in mice. Chemicalchaperones such as TUDCA and PBA alleviate differentforms of colitis in mice and might be developed fortreatment of inflammatory bowel diseases.

Keywords: IBD; Mouse Model; Ulcerative Colitis; Thera-peutic Agent.

The processes of protein folding, modification, andmaturation in the endoplasmic reticulum (ER) are

ensitive to environmental changes and multiple cellularisturbances, including ER Ca2� depletion, defective gly-

cosylation, metabolic stimuli, altered redox status, energydeprivation, inflammatory stimuli, and increased protein

secretion. When ER protein folding is perturbed or when

cells are stimulated to secrete large amounts of protein,unfolded/misfolded proteins accumulate in the ER lu-men, a condition called ER stress.1 To restore ER function

nd improve protein-folding homeostasis (proteostasis),ukaryotes evolved the unfolded protein response (UPR).n mammalian cells, the UPR is initiated by 3 ER-localizedransmembrane protein sensors: activating transcriptionactor 6� (ATF6�), inositol-requiring kinase 1� (IRE1�),nd PKR-like ER kinase.2– 4 UPR signaling can lead toither adaptation or apoptosis. In the adaptive UPR, ERrotein folding is remodeled through transactivation ofenes encoding ER chaperones, ER trafficking machinerynd ER-associated protein degradation, and eIF2� phos-

phorylation-mediated global translation attenuation.5– 8

Alternatively, prolonged and/or severe ER stress leads tothe activation of the proapoptotic UPR, including thetranscription factor CHOP and the IRE1�-activated c-Jun-N-terminal kinase (JNK) pathway.9 Moreover, chronic ERstress impairs cellular homeostasis through energy deple-tion, leakage of ER Ca2�, mitochondrial damage, oxidativetress, and activation of caspases.10 Therefore, persistentrotein misfolding in response to chronic environmentaltress and/or ineffective adaptive UPR signaling can com-romise cell function and homeostasis and induce apo-tosis.11

Recent studies link ER stress to the pathogenesis ofinflammatory bowel disease (IBD). For example, pa-tients with active Crohn’s disease and ulcerative colitisexhibit signs of ER stress in their ileal and/or colonicepithelium.12–15 In addition, human genetic studies ofBD have identified primary genetic abnormalities ineveral genes, including XBP1, AGR2, and ORMDL3,hat encode proteins associated with ER stress.15–19

Abbreviations used in this paper: ATF6�, activating transcriptionactor 6�; DSS, dextran sulfate sodium; ER, endoplasmic reticulum;BD, inflammatory bowel disease; IEC, intestinal epithelial cell; IHC,mmunohistochemistry; IL, interleukin; IRE1�, inositol-requiring kinase�; JNK, c-Jun-N-terminal kinase; PBA, 4-phenylbutyrate; PBS, phos-hate-buffered saline; TNF, tumor necrosis factor; TUDCA, tauroursode-xycholate; UPR, unfolded protein response.

© 2013 by the AGA Institute0016-5085/$36.00

http://dx.doi.org/10.1053/j.gastro.2013.01.023

smrts

E

A

sisci

1

n(

Estmt

dftfw

tc

c

BA

SICA

ND

TRA

NSLA

TION

AL

AT

990 CAO ET AL GASTROENTEROLOGY Vol. 144, No. 5

Previous studies have indicated that cells with a highload of protein folding and secretion are sensitive toaltered ER homeostasis and this can induce inflamma-tory response gene expression.5,20,21 Intestinal microbi-ota and their molecules stimulate intestinal epithelialcells (IECs) to increase secretion of mucins and antimi-crobial peptides that can overwhelm their protein se-cretory capacity. On the other hand, exposure to in-flammatory stimuli can cause ER stress, although theprecise mechanism is not well understood.21 On expo-ure to high levels of exogenous antigens and inflam-

atory cytokines in the intestinal lumen, IECs mayequire efficient UPR-mediated ER chaperone induc-ion to survive the heavy burden of protein folding andecretion.

In this study, we show that protein misfolding in theR caused by deletion of the ER cochaperone gene P58IPK/

Dnajc3 exacerbates experimental colitis in mice. ATF6� isa potent transcriptional activator for a number of ERchaperone genes, including BiP, Grp94, and P58IPK in manycell types.2– 4 Although whole body deletion of Atf6� doesnot generate an obvious phenotype under normal condi-tions, it is required for cells to survive chemical-inducedER stress.8 We found that in the absence of P58IPK or

TF6�, mice are sensitive to colitis and exhibit reducedinduction of ER chaperone genes and hyperactivation ofproapoptotic UPR signaling in colonic IECs. The chemicalchaperones tauroursodeoxycholate (TUDCA) and 4-phe-nylbutyrate (PBA) are Food and Drug Administration–approved bioactive small molecules that function to facil-itate protein folding and reduce ER stress both in vitroand in vivo by stabilizing protein-folding intermediatesand preventing protein aggregation.22–28 In this study, wehow that oral delivery of either TUDCA or PBA dramat-cally decreases the clinical, histologic, and biochemicaligns of inflammation in both innate immunity– and Tell– dependent colitis through reducing ER stress signal-ng in colonic IECs.

Materials and MethodsMiceAtf6��/� and P58IPK�/� mice (C57BL/6J background)

were described previously.8,29 Wild-type C57BL/6J mice werepurchased from the Jackson Laboratory (Bar Harbor, ME). Allanimal care and procedures were conducted according to theprotocols and guidelines approved by the University of MichiganUniversity Committee on Use and Care of Animals and theSanford-Burnham Medical Research Institute Institutional An-imal Care and Use Committee.

Cell CultureIEC-6 cells (passages 8 –10) were kindly provided by Dr

Linda Samuelson (Department of Physiology, University ofMichigan Medical Center, Ann Arbor, MI). The cells were main-tained in Dulbecco’s modified Eagle medium containing 4.5mg/mL glucose and supplemented with 2 mmol/L L-glutamine,

0 mmol/L HEPES, 100 �g/mL streptomycin, and 100 U/mL

penicillin. Cells were seeded 5 � 105 cells/well in 12-well plates, i

and all experiments were performed 1 day after cultures reachedconfluence. Cells were treated with 100 ng/mL rat tumor necro-sis factor (TNF)-� (R&D Systems, Minneapolis, MN), 100

g/mL rat MCP-1 (R&D Systems), 25 ng/mL rat interleukinIL)-1� (R&D Systems), and 5 mmol/L PBA (Scandinavian For-

mulas, Inc, Sellersville, PA) or TUDCA (EMD Chemicals, Bil-lerica, MA).

ResultsDextran Sodium Sulfate–Induced ColitisInduces ER Stress in Colonic EpitheliumTo characterize the role of ER stress and the

UPR in IECs during intestinal inflammation, we usedthe dextran sodium sulfate (DSS)-induced colitis mu-rine model. C57BL/6J mice at 8 weeks of age were fed3% DSS in their drinking water for 3 or 5 days todevelop colitis in the large intestine. The purity ofisolated colonic epithelial cells was greater than 95%,measured by flow cytometry using an antibody againstmurine epithelial cell adhesion molecule (Supplemen-tary Figure 1A). Analysis of UPR markers, including theER chaperone BiP, phosphorylation of the ER stresssensor PKR-like ER kinase, and the UPR transcriptionfactors ATF4, CHOP, and spliced XBP1 showed a time-dependent induction in colonic epithelium that coin-cided with progression of DSS-induced colitis. A simi-lar messenger RNA induction pattern of UPR genes wasobserved by quantitative reverse-transcription polymer-ase chain reaction analysis and immunohistochemistry(IHC) staining (Supplementary Figure 1C and D). The

R stress induction in the DSS-induced colitis model isimilar to that observed in intestinal tissues from pa-ients with active IBD,12–15 suggesting that this experi-

ental colitis model is valid for studies of ER stress andhe UPR in IECs on intestinal inflammation.

The ER Cochaperone P58IPK Protects FromDSS-Induced ColitisTo determine the requirement for proper protein

folding in the ER for IEC function, we analyzed micewith deletion in P58IPK. P58IPK is a heat-shock 40-kilo-

alton protein that belongs to the DNAJ chaperoneamily and resides in the ER lumen in association withhe ER chaperone BiP and promotes proper proteinolding.30 –32 The UPR induces transcription of P58IPK,hile cells and mice deleted in P58IPK display slight

protein misfolding and are sensitive to ER stress.29,30

The role of P58IPK in colonic epithelia was studied usingbone marrow chimeras of P58IPK�/� and P58IPK�/� miceo exclude the effect of P58IPK deletion in hematopoieticells. On challenge with DSS, P58IPK�/� mice displayed

severe body weight loss, rectal bleeding, and shortening ofthe large intestine (Figure 1A and B and SupplementaryFigure 2A). Consistently, the P58IPK�/� mice showed signifi-antly more severe mucosal damage, loss of goblet cells, and

nflammatory cell infiltration in the colon compared with

PPCftP(fr

Dra

i

wbsC

.00

BA

SIC

AN

DTR

AN

SLA

TIO

NA

LA

T

May 2013 MOLECULAR/CHEMICAL CHAPERONES ALLEVIATE COLITIS 991

their heterozygous littermates (Figure 1C and Supple-mentary Figure 2B). After DSS challenge, immunoblotsshowed up-regulation of ER stress markers BiP andphosphor-IRE1� in isolated colonic IECs from

58IPK�/� mice compared with both P58IPK�/� and58IPK�/� mice. The proapoptotic transcription factorHOP was also induced in P58IPK�/� colonic IECs be-

ore and after treatment with DSS (Figure 1E). Consis-ently, IHC indicated increased expression of CHOP in58IPK�/� colonic epithelium with DSS-induced colitis

Figure 1F). Bim, a BH3-only member of the Bcl-2amily that is transactivated by CHOP, plays a criticalole in ER stress–induced apoptosis.33 On DSS chal-

lenge, Bim and cleaved caspase-3 were highly induced in

Figure 1. Loss of P58IPK exacerbates DSS-induced colitis in mice dueith wild-type bone marrow cells were fed 2.5% DSS in drinking water fleeding were measured over 7 days. (C) After administration of DSS, thehown (original magnification 100�). (D) Histologic scores were measureHOP was induced whereas antiapoptotic Bcl2 was reduced in P58IPK�/

DSS-induced colitis. n � 7 for each group. *P � .05, **P � .01, ***P �

the colonic epithelium of P58IPK�/� mice compared with p

both P58IPK�/� and P58IPK�/� mice (Figure 1E). These datasuggest that the elevated susceptibility of P58IPK�/� mice to

SS-induced colitis is due to a hyperactivated ER stressesponse and proapoptotic UPR signaling, including CHOPnd Bim, in colonic IECs during inflammation.

ATF6�-Mediated ER Chaperone InductionProtects Against DSS-Induced ColitisTo further elucidate the role of ER chaperones

induced during intestinal inflammation, we studied micewith a deletion in Atf6�, the master regulator of ERchaperone gene expression. The colon of Atf6��/� mice isndistinguishable from wild-type, as indicated by H&E and

hyperactivated proapoptotic UPR. P58IPK�/� and P58IPK�/� littermatesdays, followed by 2 days of fresh water. (A) Body weight and (B) rectal

lons were isolated and fixed for H&E staining. Representative images aremice with DSS-induced colitis. (E) The proapoptotic transcription factorCs. (F) IHC shows CHOP is induced in P58IPK�/� colonic epithelium with1.

to aor 5cod in� IE

eriodic acid–Schiff staining (Supplementary Figure 3A).

Fi

UbiiIcc

BA

SICA

ND

TRA

NSLA

TION

AL

AT

992 CAO ET AL GASTROENTEROLOGY Vol. 144, No. 5

Age- and sex-matched Atf6��/�, Atf6��/�, and Atf6��/�

littermate mice were reconstituted with wild-type bonemarrow cells. The genetic ablation of Atf6� exacerbatedsymptoms of DSS-induced colitis, including severebody weight loss, rectal bleeding (Figure 2A and B), andsignificantly greater mucosal damage, goblet cell loss,

Figure 2. Loss of ATF6� exacerbates DSS-induced colitis in mice duePR. Atf6��/�, Atf6��/�, and Atf6��/� littermates with wild-type bony 2 days of fresh water. (A) Body weight and (B) rectal bleeding we

solated and fixed for H&E staining. Representative images are shown mice with DSS-induced colitis. (E) The expression of ER chapeRE1�-JNK pathway and caspase-3 are activated in Atf6��/� IEChaperone BiP is impaired whereas the proapoptotic transcription facolitis. n � 14 or 15. *P � .05, **P � .01, ***P � .001.

and macrophage infiltration in the colon (Figure 2C

and D and Supplementary Figure 3B). Atf6��/� micedisplayed reduced expression of ER chaperone genes,including BiP, Grp94, and P58IPK, in both protein andmessenger RNA levels (Figure 2E and Supplementary

igure 3C), indicating that the adaptive UPR signalingn colonic IECs of Atf6��/� mice is compromised in

defective ER chaperone induction and a hyperactivated proapoptoticarrow cells were fed 3% DSS in drinking water for 5 days, followedeasured over 7 days. After administration of DSS, the colons were

; original magnification 100�). (D) Histologic scores were measuredes BiP, GRP94, and P58IPK is reduced whereas the proapoptoticith DSS-induced colitis. (F) IHC shows that expression of the ERCHOP is increased in Atf6��/� colonic epithelium with DSS-induced

toe m

re mn (Crons wtor

response to intestinal inflammation. In contrast, the proapo-

ecffbe

upsA

rI

iT

gBA

SIC

AN

DTR

AN

SLA

TIO

NA

LA

T

May 2013 MOLECULAR/CHEMICAL CHAPERONES ALLEVIATE COLITIS 993

ptotic IRE1�-JNK pathway was induced in inflamed colonicpithelium of Atf6��/� mice (Figure 2E). After induction ofolitis, the apoptotic markers cleaved caspase-3 and DNAragmentation (by terminal deoxynucleotidyl trans-erase–mediated deoxyuridine triphosphate nick-end la-eling [TUNEL] staining) were also increased in colonicpithelium of Atf6��/� mice compared with Atf6��/�

and Atf6��/� mice (Figure 2E and Supplementary Fig-re 3D and E). Consistently, IHC showed reduced ex-ression of the ER chaperone BiP and increased expres-ion of the proapoptotic transcription factor CHOP in

Figure 3. TUDCA and PBA alleviate inflammatory stimuli-inducededucing ER stress in colonic epithelium in vivo. (A) IEC-6 cells were trL-1�) for 8 hours or pretreated with 5 mmol/L TUDCA or PBA for 4mmol/L TUDCA or PBA for 8 hours. The cells were then collected forreaction. The messenger RNA levels were normalized to the expressiowater and received 500 mg/kg body wt TUDCA or the same amount o8 or 10 per group). (B) Body weight and (C) rectal bleeding were msolated and fixed for H&E staining. Representative images are shoUDCA-treated and control mice with DSS-induced colitis. (F) Exp

apoptosis in colonic mucosa as well as ER stress markers in colonicroup; *P � .05, **P � .01, ***P � .001.

tf6��/� colonic epithelium (Figure 2F). These data are

consistent with the notion that ATF6� is an importanttransactivator of ER chaperone genes in colonic IECsduring colitis. The impaired ER chaperone induction inAtf6��/� mice leads to unresolved ER stress, whichinduces proapoptotic UPR signaling in colonic IECsand exacerbates DSS-induced colitis.

TUDCA and PBA Alleviate Inflammation-Induced ER Stress in an IEC LineGiven the protective role of the ER chaperone

stress in IECs in vitro; TUDCA ameliorates DSS-induced colitis byd with a combination of inflammatory cytokines (TNF-�, MCP-1, andrs, followed by treatment with the same inflammatory signals with 5A extraction and quantitative reverse-transcription polymerase chainf 18S ribosomal RNA. Wild-type mice were fed 2.5% DSS in drinkinghosphate-buffered saline (PBS) without TUDCA daily by gavage (n �ured over 8 days. (D) After administration of DSS, the colons were(original magnification 40�). (E) Histologic scores are shown fromsion of genes associated with inflammation, oxidative stress, ands is shown (normalized to the expression of Gapdh). n � 8 for each

EReatehouRNn of p

easwnresIEC

response in IECs against intestinal inflammation, we

ismTtism

a

BA

SICA

ND

TRA

NSLA

TION

AL

AT

994 CAO ET AL GASTROENTEROLOGY Vol. 144, No. 5

tested whether the chemical chaperones TUDCA andPBA can reduce ER stress in IECs and alleviate colitis inmice. We first analyzed the effect of PBA and TUDCA inthe nontransformed rat IEC line IEC-6 that was treatedwith physiologically relevant stimuli to cause ER stress.The inflammatory cytokines TNF-�, MCP-1, and IL-1�are highly up-regulated in animal models of enteroco-litis and patients with IBD. We found that a combinedcocktail of TNF-�, MCP-1, and IL-1� induces ER stressn IEC-6 cells, as monitored by the up-regulation of ERtress markers BiP and CHOP (Figure 3A). Prior treat-

ent and cotreatment of IEC-6 cells with either PBA orUDCA reduced CHOP and BiP expression in response

o the inflammatory stimuli (Figure 3A). These datandicate that proinflammatory cytokines induce ERtress in IECs in vitro, and this cellular stress can be

Figure 4. PBA alleviates DSS-induced colitis by reducing ER stress ind received 500 mg/kg body wt PBA or the same amount of PBS wi

measured over 10 days. (C) Colon lengths were measured after indisolated and fixed for H&E staining. Representative images are shownand PBA-treated mice with DSS-induced colitis. (F) Expression of gcolonic mucosa as well as ER stress markers in colonic IECs is show.05, **P � .01, ***P � .001.

itigated by treatment with either PBA or TUDCA.

TUDCA and PBA Alleviate Signs of DSS-Induced Colitis by Reducing ER StressSignaling in Colonic Epithelial Cells

The therapeutic potential of chemical chaper-ones was tested in a prevention paradigm in the DSS-induced colitis murine model. C57BL/6J mice were fed2.5% DSS in their drinking water for 8 days. TUDCAwas administrated orally by gavage at 500 mg/kg bodywt per day (single dose) throughout the whole period.Compared with mice subjected to DSS challenge only,feeding of TUDCA significantly ameliorated the symp-toms of DSS-induced colitis, as indicated by the lowerclinical scores (Figure 3B and C and SupplementaryFigure 4). TUDCA dramatically reduced the histologicmanifestations of DSS-induced colitis (Figure 3D and

olonic epithelium. Wild-type mice were fed 2% DSS in drinking waterut PBA daily by gavage. (A) Body weight and (B) rectal bleeding wereion of DSS colitis. (D) After administration of DSS, the colons wereinal magnification 40�). (E) Histologic scores are shown from control

es associated with inflammation, oxidative stress, and ER stress innormalized to the expression of Gapdh). n � 8 for each group; *P �

n cthouct(origenn (

E). In parallel, the expression of proinflammatory cyto-

C

BA

SIC

AN

DTR

AN

SLA

TIO

NA

LA

T

May 2013 MOLECULAR/CHEMICAL CHAPERONES ALLEVIATE COLITIS 995

kines IL-1� and TNF-� in the colon was significantlyreduced by feeding of TUDCA (Figure 3F). Additionally,the induction of oxidative/nitrosative stress and apo-ptotic signaling was inhibited by treatment withTUDCA, as indicated by diminished expression of genesencoding iNOS, NOX2, and Bim (Figure 3F). As ex-pected, the induction of ER stress markers BiP, P58IPK,

HOP, GADD34, and ERO1� in colonic IECs was con-siderably reduced in mice fed with TUDCA during theinduction of DSS colitis (Figure 3F), which is consis-tent with the observations in IEC-6 cells treated withinflammatory signals and TUDCA. Similarly, feeding ofPBA dramatically ameliorated the symptoms of DSS-induced colitis (Figure 4A–E). Induction of ER stress incolonic epithelium was significantly reduced on admin-istration of PBA during the induction of DSS colitis

Figure 5. Either TUDCA or PBA mitigates inflammation in mice with ccolitis received 300 mg/kg body wt TUDCA or the same amount of Pa mouse with chronic colitis fed with PBS alone (control), showing s(mucodepletion; bracket), and transmural inflammatory infiltrate (arro(bracket). (b) Mice with chronic colitis treated with TUDCA display re100� for inset). (B) Histologic scores, including damaged area involvefrom TUDCA-treated and control mice with chronic DSS-induced colitDSS-induced colitis received 500 mg/kg body wt PBA or the same amfrom a mouse with chronic colitis fed with PBS alone (control) showiloss of goblet cell morphology (bracket). (b) Cecum from a mouse withinflammation (original magnification 40�). (D) Histologic scores arecolitis. n � 9 or 13 for each group. *P � .05, **P � .01.

(Figure 4F).

TUDCA and PBA Reverse DSS-InducedChronic Colitis

We then examined whether the chemical chaperonescan reverse the symptoms of chronic colitis. The mice werefed with 3 cycles of 2% DSS in drinking water and thenreceived 300 mg/kg body wt TUDCA daily (double dose, 150mg/kg body wt per dose) by oral administration for 10 days.Dramatically, the histologic scores of the large intestine,including damaged area involved, ulceration, mucodepletionof glands, and inflammatory cell infiltration, were signifi-cantly reduced after the administration of TUDCA com-pared with the controls (Figure 5A and B). For delivery ofPBA, C57BL/6J mice with chronic colitis were fed 500 mg/kgbody wt PBA daily (double dose, 250 mg/kg body wt perdose) for 10 days. PBA-treated mice also showed a similar

nic colitis. (A) Wild-type mice with established chronic DSS-inducedwithout TUDCA (control) daily by gavage for 10 days. (a) Colon fromre epithelial ulceration (arrowheads), loss of goblet cell morphologyInset shows transmural inflammatory infiltrate at the edge of an ulcerced mucosal damage and inflammation (original magnification 40�;

ucodepletion of glands, and inflammatory cell infiltration, are shown� 6 or 8 for each group. (C) Wild-type mice with established chronic

t of PBS without PBA (control) daily by gavage for 10 days. (a) Cecumevere ulceration (arrowheads), inflammatory infiltrates (arrows), and

ronic colitis treated with PBA showing reduced mucosal damage andwn from PBA-treated and control mice with chronic DSS-induced

hroBSevew).dud, mis. nounng schsho

recovery from chronic colitis (Figure 5C and D).

PD e

dmst[ss

PDPt.

BA

SICA

ND

TRA

NSLA

TION

AL

AT

996 CAO ET AL GASTROENTEROLOGY Vol. 144, No. 5

TUDCA and PBA Complement theRequirement for P58IPK and ATF6� inPreventing DSS-Induced ColitisTo further explore the molecular mechanisms of

how chemical chaperones function in alleviating intes-tinal inflammation, we tested the 2 compounds onDSS-induced colitis in the P58IPK or Atf6� mice. Where

58IPK�/� and Atf6��/� mice were more susceptible toSS-induced colitis, as shown in Figures 1 and 2, feed-

ing of TUDCA or PBA reduced the clinical and histo-logic scores of P58IPK�/� and Atf6��/� mice to levelssimilar to those of their littermate controls (Figure6A–F). These data indicate that the chemical chaper-ones can complement the requirement for molecularchaperones, correct the protein folding defects in ani-mals with an impaired ER chaperone response, and

Figure 6. Feeding of TUDCA or PBA corrects the defects of P58IPK

58IPK�/�, and P58IPK�/� mice and Atf6��/�, Atf6��/�, and Atf6��/�

SS in drinking water for 5 days, followed by 2 days of fresh water. DuBA, or the same amount of PBS without TUDCA/PBA daily by gavage

reatment, the colons were isolated and fixed for H&E staining and h001.

protect mice against intestinal inflammation. (

TUDCA and PBA Dramatically MitigateColitis in Il10�/� Mice

Because TUDCA and PBA can alleviate both acuteand chronic colitis induced by DSS, we then examinedwhether the 2 chemical chaperones can ameliorate non-steroidal anti-inflammatory drug–induced colitis in Il10�/�

mice, a T cell–dependent IBD model. Il10�/� mice withstablished colitis received either TUDCA or PBA in therinking water at a concentration of 2 mg/mL (5.2 mg/ouse per day) for 3 weeks. Dramatically, the histologic

cores of the large intestine were dramatically reduced afterhe administration of either TUDCA or PBA (Figure 7Aupper and middle panels], B, and C). Furthermore, trichrometaining indicated that fibrosis in the large intestine wasignificantly reduced by the feeding of TUDCA or PBA

and Atf6��/� mice in response to DSS-induced colitis. P58IPK�/�,ce were reconstituted with wild-type bone marrow cells and fed 3%g the same period, the animals received 500 mg/kg body wt TUDCA,and C) Body weight was measured over the 7-day period. (B and D) After

logic scoring. n � 7–14 for each group; *P � .05, **P � .01, ***P �

�/�

mirin. (A

isto

Figure 7A [lower panel] and D). In the colonic epithelia

mtmauhi

pap

dc

chm

o

cec

BA

SIC

AN

DTR

AN

SLA

TIO

NA

LA

T

May 2013 MOLECULAR/CHEMICAL CHAPERONES ALLEVIATE COLITIS 997

of Il10�/� mice with colitis, the induction of ER stressarkers including BiP, phospho-eIF2�, and CHOP, mi-

ochondrial UPR marker HSP60, as well as apoptoticarkers including cleaved caspase-3/12 were consider-

bly reduced after treatment with TUDCA or PBA (Fig-re 7E). These data show that the chemical chaperonesave potent anti-inflammatory and antifibrotic effects

n the ll10�/� colitis model through the suppression ofER stress and cell death in colonic epithelium.

DiscussionRecent studies indicate that inflammatory con-

ditions in the gastrointestinal tract can activate theUPR in IECs.12–15 However, it is unknown whether these

athway(s) function to disrupt cellular homeostasisnd induce apoptosis or to restore ER function andrevent cell death. Previous studies showed that P58IPK

binds to newly synthesized secretory proteins in the ER

Figure 7. Either TUDCA or PBA alleviates chronic colitis in Il10�/� mr TUDCA in the drinking water for 3 weeks. (A) Feeding of TUDCA or

(PAS) staining shows mucin in goblet cells, and trichrome staining indusing the standard of Otuska et al.49 (C) Histologic scores are shownolonic fibrosis. n � 10 –11 for each group. (E) The mice were killextraction and Western blotting. Representative immunoblots are sholitis ¡ PBA. *P � .01, **P � .001, ***P � .0001.

and promotes protein folding/maturation in cells.30 –32

Consistent with a role for P58IPK in reducing proteinmisfolding in the cell, we showed that P58IPK prevents

ysfunction of IECs and progression of DSS-inducedolitis. During the development of colitis, P58IPK�/�

mice display increased expression of the proapoptoticfactor CHOP and reduced expression of the prosurvivalprotein Bcl2 in IECs due to an unresolved/prolongedER stress. CHOP is a major cell death–inducing factorduring the UPR9,15 and has been shown to exacerbateolitis in mice.34 Previous studies have shown that aypomorphic mutation in the gene encoding S1P inice enhances the sensitivity to DSS-induced colitis.35

However, given that S1P targets several ER stress–in-duced bZIP transcription factors, including Luman,OASIS, and CREBH, as well as the SREBPs,36 it was notclear whether the increased susceptibility is attributedto reduced activation of ATF6 or other transcriptionfactors. In this study, we showed that colonic IECs from

. Il10�/� mice with piroxicam-induced colitis received 2 mg/mL PBAA reduces signs of chronic colitis in Il10�/� mice. Periodic acid–Schifftes collagen deposition in the colon. (B) Histologic scores are showning the standard of Berg et al.50 (D) Histologic scores are shown forfter the experiment, and the colonic IECs were isolated for proteinn. 1, no piroxicam-induced colitis; 2, colitis; 3, colitis ¡ TUDCA; 4,

icePBicaus

d aow

Atf6��/� mice have reduced adaptive ER chaperone

IItouaptetceltdw

Icmswd

d

EsMTcaoiv

1

1

1

1

BA

SICA

ND

TRA

NSLA

TION

AL

AT

998 CAO ET AL GASTROENTEROLOGY Vol. 144, No. 5

expression and increased proapoptotic UPR signaling,including the IRE1�-JNK pathway.9,15 In mice with anEC-specific deletion of Xbp1, hyperphosphorylatedRE1� activates JNK and induces spontaneous inflamma-ion in the ileum.15,37 Therefore, compromised ER chaper-ne expression, by loss of either an ER chaperone itself or thepstream transactivator, leads to unresolved ER stress andctivation of proapoptotic signaling and therefore im-airs cell function and exacerbates inflammation. Inhese experiments, we used bone marrow chimeras toxclude the effect of P58IPK or Atf6� deletion in hema-opoietic cells, including macrophages, neutrophils, Tells, and B cells, during intestinal inflammation. How-ver, bone marrow reconstitution is not able to replaceamina propria fibroblasts and smooth muscle cells inhe gut, which may still contribute to colitis. Villin-Cre–irected conditional deletion models, if available,ould be ideal for this study.Consistent with the protective role of ER chaperones in

ECs for intestinal homeostasis, we showed that the chemi-al chaperones TUDCA and PBA, which promote ER ho-eostasis and increase ER folding capacity, reduce ER stress

ignaling in IECs and alleviate colitis in mice. Furthermore,e showed that feeding of TUDCA or PBA corrected theefects in P58IPK�/� and Atf6��/� mice with impaired ER

chaperone induction during DSS-induced colitis, suggestingthat TUDCA and PBA resolve intestinal inflammation dueto their function in promoting protein folding and alleviat-ing ER stress.

Recent studies showed that TUDCA inhibits the expres-sion of UPR genes in an intestinal epithelial cell line inducedby the widely used ER stressor tunicamycin.38 In our study,we showed that either TUDCA or PBA reduces ER stressgene expression in IECs induced by inflammatory stimulithat are physiologically relevant to IBD. More strikingly, thefeeding of either TUDCA or PBA dramatically protected theintestinal mucosa in both a DSS-induced model and a Tcell–dependent genetic model of colitis. These data suggestthat TUDCA and PBA exert an epithelial-protective effectduring the intestinal inflammation that is predominated byeither the innate or adaptive immune response in the mu-cosa.

The chemical chaperones TUDCA and PBA are Food andDrug Administration–approved drugs that have outstand-ing safety profiles in humans. TUDCA is safely used as ahepatoprotective drug for the treatment of primary biliarycirrhosis.39 PBA is approved for clinical use in urea cycle

isorders.40 Both compounds are in clinical trials for thetreatment of a number of diseases that are associated withprotein misfolding in the ER, including cystic fibrosis, amyo-trophic lateral sclerosis, spinal muscular atrophy, Huntington’sdisease, and type 2 diabetes (http://clinicaltrialsfeeds.org/clinical-trials/results/term�TUDCA; http://clinicaltrialsfeeds.org/clinical-trials/results/intr�4-phenylbutyric�acid).2,41– 44

Current medications used to treat IBD, including corticoste-roids, immunosuppressants, and biologics, have significant

risks and adverse effects.45–47 If efficacy can be shown in

patients with IBD, given the safety profile and potential fororal delivery of TUDCA and PBA, this therapy could fill animportant gap in our current therapeutic armamentarium.Ursodeoxycholate, the unconjugated bile salt of TUDCA, isa promising drug for chemoprevention of colorectal can-cer.48 Although TUDCA can inhibit inflammation-induced

R stress in nontransformed IEC-6 cells, it exacerbates ERtress in some colon cancer cell lines (unpublished results

ay, 2012.) It would be worthwhile to determine whetherUDCA is able to suppress the growth of carcinogenicolonocytes while protecting normal colonic epithelial cellsgainst ER stress in patients with ulcerative colitis. Based onur findings in multiple murine models of colitis, the chem-

cal chaperones TUDCA and PBA may warrant clinical in-estigation as a novel treatment for IBD.

Supplementary Material

Note: To access the supplementary materialaccompanying this article, visit the online version ofGastroenterology at www.gastrojournal.org, and at http://dx.doi.org/10.1053/j.gastro.2013.01.023.

References

1. Dorner AJ, Wasley LC, Kaufman RJ. Increased synthesis of se-creted proteins induces expression of glucose-regulated proteinsin butyrate-treated Chinese hamster ovary cells. J Biol Chem1989;264:20602–20607.

2. Hotamisligil GS. Endoplasmic reticulum stress and the inflamma-tory basis of metabolic disease. Cell 2010;140:900–917.

3. Kroemer G, Mariño G, Levine B. Autophagy and the integratedstress response. Mol Cell 2010;40:280–293.

4. Zhang K, Kaufman RJ. From endoplasmic-reticulum stress to theinflammatory response. Nature 2008;454:455–462.

5. Rutkowski DT, Arnold SM, Miller CN, et al. Adaptation to ER stressis mediated by differential stabilities of pro-survival and pro-apop-totic mRNAs and proteins. PLoS Biol 2006;4:e374.

6. Wang S, Kaufman RJ. The impact of the unfolded protein responseon human disease. J Cell Biol 2012;197:857–867.

7. Cao SS, Kaufman RJ. Unfolded protein response. Curr Biol 2012;22:R622–R626.

8. Wu J, Rutkowski DT, Dubois M, et al. ATF6alpha optimizes long-term endoplasmic reticulum function to protect cells from chronicstress. Dev Cell 2007;13:351–364.

9. Tabas I, Ron D. Integrating the mechanisms of apoptosis in-duced by endoplasmic reticulum stress. Nat Cell Biol 2011;13:184–190.

0. Malhotra JD, Miao H, Zhang K, et al. Antioxidants reduce endo-plasmic reticulum stress and improve protein secretion. Proc NatlAcad Sci U S A 2008;105:18525–18530.

1. Malhotra JD, Kaufman RJ. Endoplasmic reticulum stress and ox-idative stress: a vicious cycle or a double-edged sword? AntioxidRedox Signal 2007;9:2277–2293.

2. Shkoda A, Ruiz PA, Daniel H, et al. Interleukin-10 blockedendoplasmic reticulum stress in intestinal epithelial cells: im-pact on chronic inflammation. Gastroenterology 2007;132:190–207.

3. Hu S, Ciancio MJ, Fujiya M. et al. Translational inhibition ofcolonic epithelial heat shock proteins by IFN-gamma and TNF-alpha in intestinal inflammation. Gastroenterology 2007;133:1893–1904.

14. Heazlewood CK, Cook MC, Eri R, et al. Aberrant mucin assem-bly in mice causes endoplasmic reticulum stress and sponta-neous inflammation resembling ulcerative colitis. PLoS Med

2008;5:e54.

4

4

4

5

BA

SIC

AN

DTR

AN

SLA

TIO

NA

LA

T

May 2013 MOLECULAR/CHEMICAL CHAPERONES ALLEVIATE COLITIS 999

15. Kaser A, Lee AH, Franke A, et al. XBP1 links ER stress to intes-tinal inflammation and confers genetic risk for human inflamma-tory bowel disease. Cell 2008;134:743–756.

16. McGovern DP, Gardet A, Torkvist L, et al. Genome-wide associa-tion identifies multiple ulcerative colitis susceptibility loci. NatGenet 2010;42:332–337.

17. Zheng W, Rosenstiel P, Huse K, et al. Evaluation of AGR2 andAGR3 as candidate genes for inflammatory bowel disease. GenesImmun 2006;7:11–18.

18. Park SW, Zhen G, Verhaeghe C, et al. The protein disulfideisomerase AGR2 is essential for production of intestinal mucus.Proc Natl Acad Sci U S A 2009;106:6950–6955.

19. Zhao F, Edwards R, Dizon D, et al. Disruption of Paneth and gobletcell homeostasis and increased endoplasmic reticulum stress inAgr2�/� mice. Dev Biol 2010;338:270–279.

20. Bertolotti A, Wang X, Novoa I, et al. Increased sensitivity todextran sodium sulfate colitis in IRE1beta-deficient mice. J ClinInvest 2001;107:585–593.

21. Zhang K, Shen X, Wu J, et al. Endoplasmic reticulum stressactivates cleavage of CREBH to induce a systemic inflammatoryresponse. Cell 2006;124:587–599.

22. Burrows JA, Willis LK, Perlmutter DH. Chemical chaperones me-diate increased secretion of mutant alpha 1-antitrypsin (alpha1-AT) Z: A potential pharmacological strategy for prevention of liverinjury and emphysema in alpha 1-AT deficiency. Proc Natl Acad SciU S A 2000;97:1796–1801.

23. Singh OV, Pollard HB, Zeitlin PL. Chemical rescue of deltaF508-CFTR mimics genetic repair in cystic fibrosis bronchial epithelialcells. Mol Cell Proteomics 2008;7:1099–1110.

24. Powers ET, Morimoto RI, Dillin A, et al. Biological and chemicalapproaches to diseases of proteostasis deficiency. Annu RevBiochem 2009;78:959–991.

25. Ozcan U, Yilmaz E, Ozcan L, et al. Chemical chaperones reduce ERstress and restore glucose homeostasis in a mouse model of type2 diabetes. Science 2006;313:1137–1140.

26. Ozcan L, Ergin AS, Lu A, et al. Endoplasmic reticulum stress playsa central role in development of leptin resistance. Cell Metab2009;9:35–51.

27. Xiao C, Giacca A, Lewis GF. Sodium phenylbutyrate, a drug withknown capacity to reduce endoplasmic reticulum stress, partiallyalleviates lipid-induced insulin resistance and beta-cell dysfunc-tion in humans. Diabetes 2011;60:918–924.

28. Kars M, Yang L, Gregor MF, et al. Tauroursodeoxycholic Acidmay improve liver and muscle but not adipose tissue insulinsensitivity in obese men and women. Diabetes 2010;59:1899–1905.

29. Ladiges WC, Knoblaugh SE, Morton JF, et al. Pancreatic beta-cellfailure and diabetes in mice with a deletion mutation of theendoplasmic reticulum molecular chaperone gene P58IPK. Diabe-tes 2005;54:1074–1081.

30. Rutkowski DT, Kang SW, Goodman AG, et al. The role of p58IPKin protecting the stressed endoplasmic reticulum. Mol Biol Cell2007;18:3681–3691.

31. Petrova K, Oyadomari S, Hendershot LM, et al. Regulated asso-ciation of misfolded endoplasmic reticulum lumenal proteins withP58/DNAJc3. EMBO J 2008;27:2862–2872.

32. Tao J, Petrova K, Ron D, et al. Crystal structure of P58(IPK) TPRfragment reveals the mechanism for its molecular chaperoneactivity in UPR. J Mol Biol 2010;397:1307–1315.

33. Puthalakath H, O’Reilly LA, Gunn P, et al. ER stress triggersapoptosis by activating BH3-only protein Bim. Cell 2007;129:1337–1349.

34. Namba T, Tanaka K, Ito Y, et al. Positive role of CCAAT/enhancer-binding protein homologous protein, a transcriptionfactor involved in the endoplasmic reticulum stress response inthe development of colitis. Am J Pathol 2009;174:1786–1798.

35. Brandl K, Rutschmann S, Li X, et al. Enhanced sensitivity to DSS

colitis caused by a hypomorphic Mbtps1 mutation disrupting the

ATF6-driven unfolded protein response. Proc Natl Acad Sci U S A2009;106:3300–3305.

36. Ron D, Walter P. Signal integration in the endoplasmic reticu-lum unfolded protein response. Nat Rev Mol Cell Biol 2007;8:519–529.

37. Garrett WS, Gordon JI, Glimcher LH. Homeostasis and Inflamma-tion in the intestine. Cell 2010;140:859–870.

38. Berger E, Haller D. Structure-function analysis of the tertiary bileacid TUDCA for the resolution of endoplasmic reticulum stress inintestinal epithelial cells. Biochem Biophys Res Commun 2011;409:610–615.

39. Kaplan MM, Gershwin ME. Primary biliary cirrhosis. N Engl J Med2005;353:1261–1273.

40. Rottner M, Freyssinet JM, Martínez MC. Mechanisms of thenoxious inflammatory cycle in cystic fibrosis. Respir Res 2009;10:23.

41. Ribeiro CM, Boucher RC. Role of endoplasmic reticulum stress incystic fibrosis-related airway inflammatory responses. Proc AmThorac Soc 2010;7:387–394.

42. Nassif M, Matus S, Castillo K, et al. Amyotrophic lateral sclerosispathogenesis: a journey through the secretory pathway. AntioxidRedox Signal 2010;13:1955–1989.

43. Gkogkas C, Middleton S, Kremer AM, et al. VAPB interacts withand modulates the activity of ATF6. Hum Mol Genet 2008;17:1517–1526.

44. Vidal R, Caballero B, Couve A, et al. Converging pathways in theoccurrence of endoplasmic reticulum (ER) stress in Huntington’sdisease. Curr Mol Med 2011;11:1–12.

45. Rogler G. Gastrointestinal and liver adverse effects of drugsused for treating IBD. Best Pract Res Clin Gastroenterol 2010;24:157–165.

46. Coffin CS, Fraser HF, Panaccione R, et al. Liver diseases associ-ated with anti-tumor necrosis factor-alpha (TNF-�) use for inflam-matory bowel disease. Inflamm Bowel Dis 2011;17:479–484.

7. Dahan A, Amidon GL, Zimmermann EM. Drug targeting strategiesfor the treatment of inflammatory bowel disease: a mechanisticupdate. Expert Rev Clin Immunol 2010;6:543–550.

8. Solimando R, Bazzoli F, Ricciardiello L. Chemoprevention of colo-rectal cancer: a role for ursodeoxycholic acid, folate and hormonereplacement treatment? Best Pract Res Clin Gastroenterol 2011;25:555–568.

9. Otsuka M, Kang YJ, Ren J, et al. Distinct effects of p38alphadeletion in myeloid lineage and gut epithelia in mouse models ofinflammatory bowel disease. Gastroenterology 2010;138:1255–1265, 1265.e1–9.

0. Berg DJ, Zhang J, Weinstock JV, et al. Rapid development of colitisin NSAID-treated IL-10-deficient mice. Gastroenterology 2002;123:1527–1542.

Author names in bold designate shared co-first authorship.Received March 14, 2012. Accepted January 13, 2013.

Reprint requestsAddress requests for reprints to: Randal J. Kaufman, PhD, Del E.

Webb Neuroscience, Aging and Stem Cell Research Center, SanfordBurnham Medical Research Institute, 10901 North Torrey PinesRoad, La Jolla, California 92037-1062. e-mail: rkaufman@sanfordburnham.org; fax: (858) 795-5273.

AcknowledgmentsThe authors thank Dr Michael Katze (Department of Comparative

Medicine, University of Washington, Seattle, WA) for providing P58IPK�/�

mice, Dr Linda Samuelson (University of Michigan Medical Center, AnnArbor, MI) for providing IEC-6 cells, University of MichiganComprehensive Cancer Center Tissue Core and Histopathology Core atSanford Burnham Medical Research Institute (SBMRI; La Jolla, CA) forhistologic analysis, Jian Xing of SBMRI Gene Analysis Facility fortechnical support of quantitative reverse-transcription polymerase chain

reaction, Yoav Altman of SBMRI Flow Cytometry Core for flow cytometry

BA

SICA

ND

TRA

NSLA

TION

AL

AT

1000 CAO ET AL GASTROENTEROLOGY Vol. 144, No. 5

analysis, Jenna Rousseau and Joseph Burzynski at University ofMichigan, and Francis Lee and Kinh-Vy Nguyen at SBMRI for generaltechnical support.

Conflicts of interest

The authors disclose no conflicts.

Funding

Supported by National Institutes of Health grants P01HL057346, R37 DK042394, R01 DK088227, and R01 HL052173and a Crohn’s and Colitis Foundation of America Senior Research

Award 3800 (to R.J.K.).

fitbac

wmcPp

GSPawssd

MmTtmHfcmm(Tv

Ip

PisTaoCoLH

May 2013 MOLECULAR/CHEMICAL CHAPERONES ALLEVIATE COLITIS 1000.e1

Supplementary Methods

Generation of Bone Marrow ChimerasLittermate mice of Atf6��/�, Atf6��/�, Atf6��/�

and P58IPK�/�, P58IPK�/�, and P58IPK�/� were lethally irra-diated with 950 rad ionizing irradiation. After two hours,5 � 106 bone marrow cells isolated from the tibias andemurs of wild-type mice were injected into the lethallyrradiated recipient mice through the tail vein. Afterransplantation, mice were treated with 4 weeks of anti-iotics in their drinking water and allowed to recover fornother 4 weeks prior to treatment with DSS to induceolitis.

DSS-induced Colitis, Piroxicam-inducedColitis in II10�/� Mice, and DrugAdministrationFor acute colitis, mice received 2–3% (w/v) DSS

(MW 36,000 –50,000; MP Biomedicals, Solon, OH) indrinking water for indicated days. To develop long-last-ing chronic DSS colitis, 8-week old C57BL/6J mice re-ceived 3 cycles of 2% (w/v) DSS (7 days DSS, 14 dayswater) as previously described.1

Six-month old II10�/� mice (C57BL/6J background)ere fed 200 ppm piroxicam, a non-steroidal anti-inflam-atory drug (NSAID), in the diet for two weeks to induce

olitis as previously described.2 The mice were subject toBA/TUDCA treatment for three weeks after feeding ofrixicam-diet.For administration of TUDCA (EMD Chemicals USA,

ibbstown, NJ) and PBA (Scandinavian Formulas, Inc.,ellersville, PA), the sodium salt was dissolved in sterileBS (Invitrogen, Carlsbad, CA) in a volume of 200 �Lnd given to mice by gavage, or dissolved in drinkingater at a concentration of 2 mg/mL. Water intake mea-

urements showed that II10�/� mice with colitis con-umed an average of 2.6 mL of water with or withoutrug every day. Therefore, each Il10�/� mouse received an

average of 5.2 mg of PBA or TUDCA per day.

Isolation of Colonic Epithelial CellsColons were cut open longitudinally and feces

were removed by washing with ice-cold PBS. Colons werethen cut into 2–3 mm pieces and incubated in Ca2�,

g2� - free PBS buffer containing 10 mM EDTA in a 50L conical tube at 4°C for 1hr with gentle rotation.hen the tubes were rigorously shaken to elute the epi-

helium from colon sections. The supernatant was re-oved and sieved through a cell strainer (500 um; FisherealthCare, Houston, TX). The flow-through was centri-

uged; the resulting cell pellet was washed twice in ice-old PBS.3 The purity of isolated colonocytes was deter-ined by flow cytometry using an antibody againsturine epithelial cell adhesion molecule from BioLegend

San Diego, CA) following the manufacturer’s protocol.rypan blue staining confirmed the presence of � 85%

iable epithelial cells after the 2-hr isolation procedure.

solated cells were snap-frozen in liquid nitrogen forrotein and RNA extraction.

Quantitative Real-Time PCRRNAs from isolated murine colonic epithelial cells

and IEC-6 cells were extracted by using RNeasy kit (Qia-gen, Valencia, CA); RNA from 5 mm distal colons wasisolated by using TRIzol reagent (Invitrogen, Carlsbad,CA). Q-RT-PCR was performed as previously described.4,5

Q-RT-PCR results were normalized to the level of 18SrRNA or mRNA encoding glyceraldehyde 3-phosphatedehydrogenase (GAPDH). Primer sequences are listed inSupplemental Table 1.

Western BlottingThe isolated colonic epithelial cells were lysed in

RIPA buffer for protein extraction. Protein content wasmeasured by Protein Assay (Bio-Rad, Hercules, CA), andprotein samples were analyzed by 10 –20% reducing SDS/PAGE (Tris-HCl precast gels, Bio-Rad) and were detectedwith the following antibodies: anti-GRP94/BiP (Enzo LifeSciences International, INC, Plymouth Meeting, PA), anti-phosphor-PERK (Cell Signaling Technology, Danvers,MA), anti-ATF4 (Santa Cruz Biotechnology, Santa CruzCA), anti-CHOP (Santa Cruz Biotechnology), anti-XBP1(Santa Cruz Biotechnology), anti-Bim (Cell SignalingTechnology), anti-glyceraldehyde 3-phosphate dehydro-genase (GAPDH, Millipore Corporation, Billerica, MA),and anti-tubulin (Cell Signaling Technology). After over-night incubation of primary antibodies at 4°C, the mem-branes were washed and then incubated with secondaryHRP-coupled antibodies (GE Healthcare, Piscataway, NJ)and developed with chemiluminescent detection system(GE Healthcare).

Immunohistochemical StainingColonic tissues were fixed overnight in 10% for-

malin, embedded in paraffin, and cut into 5 �m sections.araffin-embedded tissue sections were rehydrated, heat-

nduced antigen retrieval was performed either in 10 mModium citrate, 0.05% Tween-20, pH 6.0, or in 10 mMris, 1 mM EDTA, pH 8.0, for 10 minutes. Primaryntibodies: anti-BiP antibody (Santa Cruz Biotechnol-gy) at a dilution of 1:200 at room temperature; anti-HOP antibody (Santa Cruz Biotechnology) at a dilutionf 1:50 at room temperature; anti-XBP1 antibody (Bio-egend) at a dilution of 1:50 overnight at 4° temperature.ematoxylin was performed for counterstaining.

Histological ScoringHematoxylin and eosin (H&E) stained colon sec-

tions were scored in a blinded manner by board certifiedveterinary pathologists as previously described by Berg et

al,2 Otuska et al,6 and Lee et al.7

5

6

7

1000.e2 CAO ET AL GASTROENTEROLOGY Vol. 144, No. 5

References

1. Wirtz S, Neufert C, Weigmann B, et al. Chemically induced mousemodels of intestinal inflammation. Nat Protoc 2007;2:541–546.

2. Berg DJ, Zhang J, Weinstock JV, et al. Rapid development of colitisin NSAID-treated IL-10-deficient mice. Gastroenterology 2012;123:1527–1542.

3. Cao SS, Song B, Kaufman RJ. PKR protects colonic epitheliumagainst colitis through the unfolded protein response and prosur-vival signaling. Inflamm Bowel Dis 2012;18:1735–1742.

4. Back SH, Scheuner D, Han J, et al. Translation attenuationthrough eIF2alpha phosphorylation prevents oxidative stress andmaintains the differentiated state in beta cells. Cell Metab 2009;

10:13–26.

. Nassif M, Matus S, Castillo K, et al. Amyotrophic lateral sclerosispathogenesis: a journey through the secretory pathway. AntioxidRedox Signal 2010;13:1955–1989.

. Otsuka M, Kang YJ, Ren J, et al. Distinct effects of p38alphadeletion in myeloid lineage and gut epithelia in mouse models ofinflammatory bowel disease. Gastroenterology 2010;138:1255–1265, 1265.e1–9.

. Lee J, Mo JU, Katakura K, et al. Maintenance of colonic homeo-stasis by distinctive apical TLR9 signalling in intestinal epithelialcells. Nat Cell Biol 2006;8:1327–1336.

Author names in bold designate shared co-first authorship.

imum of 6 individual mice.

May 2013 MOLECULAR/CHEMICAL CHAPERONES ALLEVIATE COLITIS 1000.e3

Supplementary Figure 1. DSS colitis induces ER stress in colonic epby flow cytometry using an antibody against murine EpCAM. (b) Wild-typadministration for the indicated days, the mice were euthanized andRepresentative immunoblots of colon tissues from mice are shown. (c) WAfter DSS administration for the indicated days, the mice were euthanizemRNA levels were normalized to the expression of Gapdh. (d) Wild-typeeuthanized and the colons were removed, fixed and paraffin embeddedimmunohistochemical images are shown; each group contained a min

ithelial cells. (a) The purity of isolated colonic epithelial cells was determinede mice were fed 3% DSS in the drinking water for 0, 3 and 5 days. After DSS

the colonic IECs were isolated for protein extraction and western blotting.ild-type mice were fed 3% DSS in the drinking water for 0, 2, 4 and 6 days.

d and the colonic IECs were isolated for RNA extraction and Q-RT-PCR. Themice were fed 3% DSS in the drinking water for 6 days, then the mice werefor immunohistochemical staining of BiP, CHOP and XBP1. Representative

1000.e4 CAO ET AL GASTROENTEROLOGY Vol. 144, No. 5

Supplementary Figure 2. P58IPK�/� mice exhibit shorter colonlengths (a) and higher levels of macrophage marker F4/80 in colonic

mucosa (b) upon DSS colitis.

Am

r

May 2013 MOLECULAR/CHEMICAL CHAPERONES ALLEVIATE COLITIS 1000.e5

Supplementary Figure 3. (a) Atf6��/� mice display normal colon morphology under normal conditions. Colons were isolated from two-month oldtf6��/� and Atf6��/� littermates mice for H&E and PAS staining. (b) Atf6��/� mice showed higher levels of macrophage marker F4/80 in colonicucosa upon DSS colitis. (c) Atf6��/� mice exhibit an impaired induction of ER chaperone genes including Bip, Grp94, P58IPK and Pdi in colonic IECs

upon DSS colitis. (d) Terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL)-staining of apoptotic epithelial cells in the colon sectionsfrom Atf6��/� and Atf6��/� mice with DSS colitis. (e) The apoptotic indices were calculated as the number of TUNEL-positive epithelial cells per 100

andomly selected crypts in the colon sections. 4 mice per group; * p � 0.05.

Supplementary Figure 4. Shortening of colon lengths upon DSS

colitis was significantly mitigated by the feeding of TUDCA.

1000.e6 CAO ET AL GASTROENTEROLOGY Vol. 144, No. 5

Supplementary Table 1. Sequence of primers used in thisstudy

Primer name Oligo sequence (5= to 3=)

rat CHOP (f) AGAGTGGTCAGTGCGCAGCrat CHOP (r) CTCATTCTCCTGCTCCTTCTCCrat BiP (f) TGGGTACATTTGATCTGACTGGArat BiP (r) CTCAAAGGTGACTTCAATCTGGGrat 18S (f) AGTCCCTGCCCTTTGTACACArat 18S (r) GATCCGAGGGCCTCACTAAACmouse IL-1beta (f) CAACCAACAAGTGATATTCTCCATGmouse IL-1 beta (r) GATCCACACTCTCCAGCTGCAmouse TNF-alpha (f) CCCTCACACTCAGATCATCTTCTmouse TNF-alpha (r) GCTACGACGTGGGCTACAGmouse iNOS (f) CAGCTGGGCTGTACAAACCTTmouse iNOS (r) CATTGGAAGTGAAGCGTTTCGmouse BiP (f) TCATCGGACGCACTTGGAmouse BiP (r) CAACCACCTTGAATGGCAAGAmouse GRP94 (f) AATAGAAAGAATGCTTCGCCmouse GRP94 (r) TCTTCAGGCTCTTCTTCTGGmouse Chop (f) GTCCCTAGCTTGGCTGACAGAmouse Chop (r) TGGAGAGCGAGGGCTTTGmouse ERdj4 (f) CCCCAGTGTCAAACTGTACCAGmouse ERdj4 (r) AGCGTTTCCAATTTTCCATAAATTmouse P58IPK (f) TCCTGGTGGACCTGCAGTACGmouse P58IPK (r) CTGCGAGTAATTTCTTCCCCmouse Ero1a (f) GCATTGAAGAAGGTGAGCAAmouse Ero1a (r) ATCATGCTTGGTCCACTGAAmouse Gadd34 (f) CCCGAGATTCCTCTAAAAGCmouse Gadd34 (r) CCAGACAGCAAGGAAATGGmouse Nox2 (f) CCCTTTGGTACAGCCAGTGAAGmouse Nox2 (r) CAATCCCAGCTCCCACTAACATmouse Bim (f) GGAGATACGGATTGCACAGGAGmouse Bim (r) CCTTCTCCATACCAGACGGAAGmouse GAPDH (f) TTCAACGGCACAGTCAAGG

mouse GAPDH (r) CATGGACTGTGGTCATGAG