Nc

AGT

*SR

BttptiMbigacfaptcklggcCpsiarftbtCnEgc

K

GASTROENTEROLOGY 2010;138:1189–1199

FAT-Induced Histone Acetylation Relay Switch Promotes-Myc-Dependent Growth in Pancreatic Cancer Cells

LEXANDER KÖENIG,* THOMAS LINHART,* KATRIN SCHLENGEMANN,* KRISTINA REUTLINGER,* JESSICA WEGELE,‡

UIDO ADLER,‡ GARIMA SINGH,* LEONIE HOFMANN,* STEFFEN KUNSCH,* THOMAS BÜCH,§ EVA SCHÄFER,§

HOMAS M. GRESS,* MARTIN E. FERNANDEZ–ZAPICO,� and VOLKER ELLENRIEDER*

Department of Gastroenterology and Endocrinology, Philipps-University of Marburg, Marburg; ‡Department of Gastroenterology, University of Ulm, Ulm; §Walther-traub-Institute of Pharmacology and Toxicology, Ludwig-Maximilian University, Munich, Germany; and �Schulze Center for Novel Therapeutics, Mayo Clinic,

ochester, Minnesota

Ad(stuntnoImatttadtorpssIgsmhatp

ttf

BA

SIC–L

IVER

,PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

ACKGROUND & AIMS: Induction of immediate earlyranscription factors (ITF) represents the first transcrip-ional program controlling mitogen-stimulated cell cyclerogression in cancer. Here, we examined the transcrip-ional mechanisms regulating the ITF protein c-Myc andts role in pancreatic cancer growth in vitro and in vivo.

ETHODS: Expression of ITF proteins was examinedy reverse-transcription polymerase chain reaction and

mmunoblotting, and its implications in cell cycle pro-ression and growth was determined by flow cytometrynd [3H]-thymidine incorporation. Intracellular Ca2�

oncentrations, calcineurin activity, and cellular nuclearactor of activated T cells (NFAT) distribution werenalyzed. Transcription factor complex formations andromoter regulation were examined by immunoprecipi-ations, reporter gene assays, and chromatin immunopre-ipitation. Using a combination of RNA interferencenockdown technology and xenograft models, we ana-

yzed the significance for pancreatic cancer tumorrowth. RESULTS: Serum promotes pancreatic cancerrowth through induction of the proproliferative NFAT/-Myc axis. Mechanistically, serum increases intracellulara2� concentrations and activates the calcineurin/NFATathway to induce c-Myc transcription. NFAT binds to aerum responsive element within the proximal promoter,nitiates p300-dependent histone acetylation, and creates

local chromatin structure permissive for the inducibleecruitment of Ets-like gene (ELK)-1, a protein requiredor maximal activation of the c-Myc promoter. The func-ional significance of this novel pathway was emphasizedy impaired c-Myc expression, G1 arrest, and reducedumor growth upon NFAT depletion in vitro and in vivo.ONCLUSIONS: Our study uncovers a novel mecha-ism regulating cell growth and identifies the NFAT/LK complex as modulators of early stages of mito-en-stimulated proliferation in pancreatic cancerells.

ey words: Pancreatic Cancer; c-Myc; NFAT; Transcription.

hallmark in cell cycle reentry and G1 progression isthe rapid, sustained, and protein-synthesis indepen-

ent induction of immediate early transcription factorsITF).1,2 They function as mediators coupling externalhort-term signals to G1 transition through their abilityo induce specific transcriptional programs resulting inp-regulation of D-type cyclins and their partnering ki-ases CDK4 and CDK6.3 Immediate inducible transcrip-ion factors themselves are tightly controlled by mitoge-ic- and antiproliferative-signaling pathways on the levelsf promoter regulation and proteasomal degradation.4 – 6

nduction of ITF gene products is caused by direct pro-oter interaction, occurs within minutes upon stimul-

tion, and requires activation of latent signaling regulatedranscription factors already present in the cell.1,4,6 Bothhe composition and activation level of these upstreamranscriptional regulators (eg, SRF and Smads) as wells the expression pattern of downstream ITF proteins areependent on the cell type and the activation status ofhe cell at a given time.5,7–10 In cancer cells, activation ofncogenic signaling and transcription pathways can over-ide growth suppressor activities and stimulate cell cyclerogression, either through rapid induction of ITF sub-ets or via stimulation of single master ITF proteins. Noturprisingly, induction and transcriptional activation ofTF factors have frequently been associated with fastrowing malignancies.11–13 Therefore, understanding theignaling and transcriptional mechanisms underlying

itogen-induced ITF expression and activation not onlyelps to better understand cancer growth promotion butlso can provide a platform for the development of novelherapeutic options in cancer therapy. This might bearticularly true for pancreatic cancer, which is charac-

Abbreviations used in this paper: ChIP, chromatin immunoprecipi-ation; CsA, cyclosporine A; FCS, fetal calf serum; HAT, histone acetyl-ransferase; ITF, immediate early transcription factors; NFAT, nuclearactor of activated T cells; PCR, polymerase chain reaction.

© 2010 by the AGA Institute0016-5085/10/$36.00

doi:10.1053/j.gastro.2009.10.045

tic

pptIcpiacsstnN�asgcbmadtvp

afp

w(aCsctCeOoas

Una(3hss

tppMwctESL

dolhpoNanfw(Pftp

cAM

dcssM

BA

SIC–LIV

ER,

PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

1190 KÖENIG ET AL GASTROENTEROLOGY Vol. 138, No. 3

erized by a rapid disease progression, an aggressivelynvasive tumor phenotype, and a general resistance toonventional chemotherapy.14,15

With the final goal to develop novel innovative thera-eutic strategies in pancreatic cancer, we conducted theresented study and examined the expression, transcrip-ional regulation, and function of the most prominentTF members in pancreatic cancer. We identified the-Myc proto-oncogene as a master ITF in G1-S cell cyclerogression and growth of pancreatic cancer. Serum-

nduced c-Myc expression and cancer growth are medi-ted by fast and sustained stimulation of the Ca2�/cal-ineurin/nuclear factor of activated T cells (NFAT)ignalling and transcription pathway. Mechanistically,erum rapidly enhances intracellular Ca2� concentra-ions, leading to calcineurin activation and subsequentuclear translocation and c-Myc promoter binding ofFAT factors. NFAT binding to the serum responsive84/�63 promoter region enhances the net histone

cetylase activity associated with the c-Myc promoter andubsequently allows recruitment of the cofactor Ets-likeene (ELK)-1. Depletion of NFAT in pancreatic cancerells prevents mitogen-induced c-Myc expression andlocks tumor cell growth in vitro and in a xenograftouse model, confirming a key role of the NFAT/c-Myc

xis in pancreatic cancer growth. Together, these findingsefine a novel insight into the network of events con-rolling early stages of cell cycle progression and pro-ide a platform to develop novel therapeutic ap-roaches for pancreatic cancer.

Materials and MethodsCell Lines, Plasmids, Reverse-TranscriptionPolymerase Chain Reaction Analysis, andSmall Interfering RNA OligonucleotidesDetailed descriptions on cell lines, plasmid gener-

tion, small interfering RNA oligonucleotides, and trans-ection protocols used in this study are available in Sup-lementary Materials and Methods.

Immunohistochemistry and FluorescenceMicroscopyPaTu8988 cells grown on chambered coverslips

ere left untreated or treated with 10% fetal calf serumFCS) for 60 minutes. Cells were washed, fixed, blocked,nd probed with anti-NFATc2 antibody (1:150; Santaruz Biotechnology, Santa Cruz, CA) as previously de-

cribed.16 NFATc2 was detected with a fluorochrome-onjugated secondary antibody, and nuclei were coun-erstained with 4=6-diamidino-2-phenylindole (DAPI).overslips were mounted on glass slides, and cells were

valuated with a fluorescence microscope (Carl Zeiss, Inc,berkochen, Germany). Immunohistochemical analysis

f tumors explanted from the nude mice was performeds previously described.16 Briefly, paraffin sections were

tained with anti-NFATc1 (1:150; Abcam, Cambridge, e

K) or anti-c-Myc antibody (1:250; Santa Cruz Biotech-ology). Antibody binding was visualized using a biotinyl-ted secondary antibody, avidine-conjugated peroxidaseABC method; Vector Laboratories, Burlingame, CA),,3=diaminobenzidine tetrachloride as a substrate, andematoxylin as counterstain. The sections were counter-tained with hematoxylin, and slides were evaluated bytandard light microscopy.

Subcellular Fractionation,Coimmunoprecipitation, and ImmunoblottingSubcellular fractionation, coimmunoprecipita-

ion, and Western blotting were performed as describedreviously.16,17 Detailed descriptions of these technicalrocedures are available in Supplementary Materials andethods. For immunoblotting, membranes were probedith antibodies against NFATc1 (Abcam), NFATc2,

-Myc, c-Jun, c-Fos, p300, and ELK-1 (Santa Cruz Bio-echnology). CdK4, CdK6, cyclin D1, cyclin D3, cyclin E,GR-1, anti-HA, and lamin a/c were purchased from Cellignaling (Danvers, MA) or �-actin (Sigma-Aldrich, Saintouis, MO).

Chromatin Immunoprecipitation AssaysThe chromatin immunoprecipitation (ChIP) was

one in PaTu8988t and Panc-1 cells treated with serumr untreated for indicated time periods. Cells were cross-

inked with 1% formaldehyde for 10 minutes at 37°C andarvested in sodium dodecyl sulfate lysis buffer (Milli-ore, Billerica, MA), and DNA was shredded to fragmentsf 500 base pair by sonication. Antibodies againstFATc1, NFATc2, AcH3 (Millipore), or ELK-1 were

dded, and precleared chromatin was incubated over-ight. Protein G agarose beads were added and incubated

or 1.5 hours at 4°C. After reversing the cross-links, DNAas isolated and used for polymerase chain reaction

PCR). Specific primer pairs were designed with therimerExpresss 3.0 (Applied Biosystems, Austin, TX) asollowed: 5=-gagggatcgcgctgagtat-3= and 5=-tctaactcgctg-agtaattccagc-3= for qualitative or quantitative PCR am-lifying the �84/�63 element of the c-Myc promoter.

[3H]-Thymidine Assay and Flow CytometryCell growth was measured by [3H]-thymidine in-

orporation and flow cytometry as described previously.16

detailed protocol is available in the Supplementaryaterials and Methods.

Reporter Gene AssaysLuciferase reporter gene assays were performed as

escribed previously.18 Cells were seeded in 24-well tissueulture dishes and transfected with the indicated con-tructs before treatment. Luciferase activity was mea-ured using the Lumat LB 9501 (Berthold Technologies,

annheim, Germany) luminometer and the Dual-Lucif-

rase-Reporter Assay System (Promega, Madison, WI) ac-

ceau

piaCoCadsaHmm5udb

sK(w

t�

HsNstvcnm(wtC

glPt1iaa

Fweaf1W

BA

SIC–L

IVER

,PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

March 2010 NFAT INDUCED HAT RELAY SWITCH 1191

ording to the manufacturer’s instructions. Firefly lucif-rase values were normalized to Renilla luciferase activitynd were expressed as mean “fold induction.” Mean val-es are displayed � standard deviations.

Generation of Aequorin Expressing Cells andMeasurement of Intracellular Ca2�

ConcentrationsFor calcium measurements, a green fluorescent

rotein-aequorin (G5A) fusion protein was expressed us-ng a lentiviral expression system.19 Detailed descriptionsre available in Supplementary Materials and Methods.a2� concentrations were measured as described previ-usly.20 In brief, PaTu8988t cells stably expressing thea2� sensitive photoprotein (green fluorescent protein-)equorin were incubated for 30 minutes in culture me-ium containing 5 �mol/L coelenterazine (Biaffin, Kas-el, Germany), which constitutes the chromophore ofpoaequorin. The cells were then washed twice withank’s balanced salt solution (HBSS) buffer (140mol/L NaCl, 5 mmol/L KCl, 1 mmol/L MgSO4, 2mol/L CaCl2, 20 mmol/L KH2PO4, 20 mmol/L HEPES,

.5 mmol/L glucose, pH 7.4). Afterward, cells were stim-lated as indicated, and the luminescence signal wasetected in a PolarSTAR plate reader from BMG (Offen-urg, Germany).

Calcineurin Activity AssayCellular calcineurin phosphatase activity was mea-

ured with the Biomol Green Cellular Calcineurin Assayit Plus according to the manufacturer’s instructions

Enzo Life Sciences GmbH, Lörrach, Germany). Cellsere left untreated or treated with 10% FCS for indicated

igure 1. Serum promotes pancreatic cancer growth through accelerith 10% FCS for 24 hours. Thymidine incorporation was measured andxperiment. Data are representative of triplicate experiments and displafter treatment with 10% FCS or without FCS for 24 hours were performerom 1 representative experiment (n � 4). (C) Time course analysis of seru0% FCS for indicated time periods and analyzed by flow cytometry.

estern blot analysis to determine expression of cyclins following serum stim

ime periods before harvesting. Mean values are displayedstandard deviations.

Tumor XenograftsPanc-1 cells were transfected with the pSilencer3.1-H1

ygro vector (Applied Biosystems) bearing a hairpin-tructured short hairpin RNA (shRNA) cassette forFATc1 silencing (5=-ggactccaaggtcattttctt-3=) or control

hRNA. Stable clones were selected in Hygromycin con-aining media, and the efficacy of NFATc1 silencing wasalidated by Western blot. We injected 2 � 106 Panc-1ells subcutaneously into the flank of a male athymicude mouse and measured the tumor weekly.21 We esti-ated tumor volume (V) from the length (l) and width

w) of the tumor using the formula: V � (�/6) � [(l �)/2]3. Animal experiments were carried out using pro-

ocols approved by the Institutional Animal Care and Useommittee at the University of Ulm.

ResultsTime-Dependent Cell Cycle Propagation andGrowth in Pancreatic Cancer CellsStimulated With SerumThe effect of serum on pancreatic cancer cell

rowth was determined in PaTu8988t and Panc-1 cellines by thymidine assays and flow cytometry analysis.ancreatic cancer cells were arrested by serum depriva-ion, and growth was stimulated through application of0% FCS. Serum stimulation caused a significant increase

n cell proliferation (Figure 1A) resulting from acceler-ted cell cycle propagation (Figure 1B). Flow cytometrynalysis revealed a significant shift from G1 to S phase

cell cycle transition. (A) Cell proliferation was examined after treatmentlated relative to untreated control, which was arbitrarily set to 1 for eachSD. (B) Cell cycle analyses of PaTu8988t (left) and Panc-1 (right) cells

propidium iodide staining and following flow cytometry. Data are shownduced cell cycle progression in PaTu8988t cells. Cells were treated withare representative of triplicate experiments and displayed � SD. (D)

atedcalcuyed �d bym inData

ulation over time.

asDeai

sS1fs2atcgtf

dfkmtcic

camnBicwcnawlpnwnCtrigpioctebsttpbcpchsac3

FtaEtewpS

BA

SIC–LIV

ER,

PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

1192 KÖENIG ET AL GASTROENTEROLOGY Vol. 138, No. 3

fter 8 to 16 hours upon treatment, and this effect was as-ociated with induction of D-type cyclins (Figure 1C and). Thus, serum induction of cell growth involves the

xpression of G1-promoting cyclins, cell cycle re-entry ofrrested cells, and, subsequently, G1-S phase propagationn pancreatic cancer cells.

Expression and Induction of ITF inPancreatic Cancer CellsWe next investigated the expression status and

erum responsiveness of ITFs in pancreatic cancer cells.erum-starved pancreatic cancer cells were treated with0% FCS over time, and the expression was determinedor the most prominent members of this group of tran-cription factors: EGR-1, c-Myc, c-Jun, and c-Fos. FigureA displays moderate to high expression levels of c-Junnd c-Fos in serum-starved pancreatic cancer cells, andhese levels remained unaffected by serum stimulation. Inontrast, EGR-1 and c-Myc were modestly expressed inrowth-arrested cells, but significantly induced upon mi-ogen stimulation (Figure 2A and B). To investigate theunctional implication of c-Myc and EGR-1 in serum-

igure 2. Serum-induced cell proliferation depends on c-Myc induc-ion. (A) PaTu8988t (left) and Panc-1 (right) cells were treated with FCSs indicated, and the expression levels of c-Jun, c-Fos, c-Myc, andGR-1 were examined by immunoblotting. (B) PaTu8988t cells were

reated with FCS to determine EGR-1 (left) and c-Myc (right) mRNAxpression by quantitative RT-PCR. (C) Specific siRNA oligonucleotidesere transfected to either silence EGR-1 (left panel) or c-Myc (rightanel), and [3H]-thymidine incorporation was measured in Panc-1 cells.

Nuccessful gene silencing was confirmed by immunoblotting.

riven cell growth, we performed [3H]-thymidine assaysollowing genetic silencing of either factor. Successfulnockdown of EGR-1 by RNA interference caused only aoderate inhibition of cell proliferation, whereas deple-

ion of c-Myc resulted in a strong reduction of pancreaticancer growth (Figure 2C). Together, these experimentsdentify c-Myc as the early mediator of serum-inducedell cycle progression in pancreatic cancer cells.

Serum Stimulates c-Myc Promoter ActivationThrough Activation of the Ca2�/Calcineurin/NFAT Signaling and Transcription PathwayDepending on the cell type, mitogen stimulation

an utilize different intracellular signaling cascades toctivate the ITF response. Here, we focused on 2 majoritogenic pathways: the mitogen-activated protein ki-

ase (MAPK) network and the Ca2�-dependent signaling.oth pathways exert oncogenic potential with known

mplications in cell proliferation and progression of pan-reatic cancer.16,22 Serum-starved cells were pretreatedith either cyclosporine A (CsA) to block Ca2�/cal-

ineurin signaling, the extracellular signal-regulated ki-ase/MAPK inhibitor U0126, or a combination of bothgents. Cells were then stimulated, and c-Myc expressionas determined. Again, serum treatment enhanced the

evel of c-Myc expression in 8988t cells (Figure 3A). Sur-risingly, pharmacologic inactivation of extracellular sig-al-regulated kinase/MAPK only marginally interferedith serum inducibility on both messenger RNA (dataot shown) and protein level (Figure 3A). In contrast,sA prevented c-Myc up-regulation, indicating a role of

he Ca2�/calcineurin cascade in c-Myc induction by se-um. Consequently, we tested whether serum affectsntracellular Ca2� concentrations in cancer cells andenerated 8988t cells stably expressing aequorin, ahotoprotein from jellyfish that allows monitoring of

ntracellular Ca2� fluctuations by luminescence technol-gy. Serum treatment of growth-arrested cancer cellsaused a rapid increase in intracellular Ca2� concentra-ions (Figure 3B), and this effect was independent fromxtracellular signal-regulated kinase/MAPK activationut resulted in a tremendous induction of the down-tream phosphatase calcineurin (Figure 3C). Transcrip-ion downstream of calcineurin is in large part funnelledhrough the transcription factor NFAT, a heavily phos-horylated protein that is cytoplasmic in resting cellsut enters the nucleus when dephosphorylated by cal-ineurin. We have most recently demonstrated high ex-ression levels of NFATc1 and NFATc2 in pancreaticancer cells.16 Interestingly, all tested cell lines displayedigh expression of either NFATc1 or NFATc2, but nonehowed significant expression of both factors. In growth-rrested 8988t cells, NFATc2 is primarily localized in theytoplasm, suggesting inactivation of the protein (FigureD). Serum stimulation promoted rapid and sustained

FATc2 translocation into the nucleus for several hours,

aceoMtrvwifnmpeu

msst

i(ttevcSpsFwtsfCdcTi

Fcww3muSssti are ex

BA

SIC–L

IVER

,PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

March 2010 NFAT INDUCED HAT RELAY SWITCH 1193

s revealed by immunofluorescence microscopy and sub-ellular fractionation assays (Figure 3D and E). Similarffects were observed in Panc-1 cells with accumulationf nuclear NFATc1 upon serum application (Figure 3E).oreover, nuclear translocation of both factors was an-

agonized following CsA treatment, indicating that se-um stimulates nuclear shuttling of NFAT through acti-ation of the Ca2�/calcineurin cascade. We then analyzedhether calcineurin/NFAT signaling mediates serum

nduction on the level of promoter regulation and per-ormed reporter gene assays using a 2.8 kilobase span-ing c-Myc promoter reporter construct. Serum treat-ent caused a 4-fold induction of the wild-type c-Myc

romoter, and, consistent with our expression data, thisffect was diminished upon application of CsA (Fig-re 3F).To identify the serum responsive element of the hu-an c-Myc promoter, we performed reporter assays using

equential deletion fragments fused to luciferase. Analy-is of the promoter fragments revealed that sequences up

igure 3. Serum utilizes the calcium/calcineurin/NFAT pathway to indells depending on treatment with either 1 �mol/L CsA, 10 �mol/L U01as used for loading control. (B) 8988t Cells stably expressing the calciuas continuously recorded following application of FCS � U0126 that windependent experiments. The light and dark gray areas indicate SEeasurement of phosphate released from the RII phosphopeptide follo

es � SD. (D) Immunocytologic detection of NFATc2 localization in 8988erum stimulation led to a rapid translocation of NFATc2 (red signals) frignals). (E) 8988t and Panc-1 cells were either pretreated with CsA to bubcellular fractionation, nuclear and cytoplasmic fractions were analyzing. (F) The luciferase activity of the human wild-type c-Myc promoter cn the presence or absence of 1 �mol/L CsA. Reporter gene activities

o �109 base pair did not significantly decrease serum t

nducibility, whereas further deletion up to position �26relative to the P2 transcription initiation site) renderedhe promoter insensitive to serum (Figure 4A). The �109o �26 base pair region of the human c-Myc promoterncompasses a short region that is known as a key con-ergence node for multiple signalling and transcriptionascades (located between �84 and �63 base pair).23

imilar to the full-length c-Myc promoter, the �84/�63romoter region was responsive to serum treatment andhowed an approximate 4-fold induction (Figure 4B).urthermore, inactivation of the calcineurin/NFAT path-ay through application of CsA or mutational disrup-

ion of the NFAT-binding sequence abolished serum re-ponsiveness, indicating that NFAT binding is requiredor c-Myc promoter induction by serum (Figure 4B and). Support for this assumption came from ChIP assaysemonstrating inducible NFAT binding to the proximal-Myc promoter upon serum stimulation (Figure 4D).hus, NFAT binds to the serum responsive element to

nduce transcription from the c-Myc promoter. Consis-

-Myc. (A) Western blot analysis to illustrate c-Myc expression in 8988tr a combination of both agents before application of 10% FCS. �-actinnsor protein aequorin were loaded with coelenterazine. Luminescenceiven 30 minutes before stimulation (gray line). Data represent means of) Calcineurin phosphatase activity was determined in 8988t cells by

treatment with FCS for the indicated time periods. Data are mean val-s. Cells were serum deprived and then treated with FCS for 60 minutes.e cytosol to the nucleus. Nuclei were visualized by DAPI staining (blue

calcineurin or directly stimulated with 10% FCS as indicated. Followingparately for the presence of NFATc2/NFATc1 proteins by immunoblot-uct was measured upon treatment of 8988t cells with FCS for 24 hourspressed as “fold activation” relative to untreated controls.

uce c26, om seas gM. (Cwingt cellom thlock

ed seonstr

ently, depletion of NFAT impaired serum-induced ex-

pTcc

qNeamw

taiNbftswlaf

F(e(otwCPNW r stim

BA

SIC–LIV

ER,

PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

1194 KÖENIG ET AL GASTROENTEROLOGY Vol. 138, No. 3

ression of c-Myc in pancreatic cancer cells (Figure 4E).hese data strongly indicate that serum utilizes the Ca2�/alcineurin/NFAT pathway to induce c-Myc expression inancer cells.

NFAT Recruits ELK-1 to Induce MaximalActivation of the c-Myc PromoterSufficient NFAT promoter induction often re-

uires recruitment of other DNA binding proteins.24 TheFAT binding region (�75/�71) of the c-Myc promoter

ncompasses the core GGAA/T consensus sequence forvian erythroblastosis virus E26 (v-ets) oncogene ho-olog-domain transcription factors. We have tested

igure 4. Serum promotes NFAT promoter binding to induce c-Myc. (Ac-Myc del-I) or various deletion constructs were treated with FCS, andxpressed as “fold activation” relative to the activity of the untreated fulB) The (�84/�63) c-Myc promoter fragment was transfected into 8988f 1 �mol/L CsA. Reporter gene activities are expressed as “fold inductiohe wild-type (�84/�63 wt) promoter or a mutant that lacks the NFATere normalized to Renilla luciferase activity and expressed as fold indhromatin immunoprecipitations were performed with specific NFATc1CR using primers specific for the c-myc promoter region harboring tFATc1 to the c-Myc �84/�63 element upon addition of serum. (E)estern blot analysis were performed to detect c-Myc expression afte

hether NFAT and ELK-1 cooperate in c-Myc transcrip- C

ion and precipitated endogenous NFAT proteins to an-lyze inducible complex formation with ELK-1. Figure 5Allustrates our findings and shows the activation of

FATc1 by serum application (lanes 2 and 3), as indicatedy the shift of the protein from a slow to a fast migratingorm, representing dephosphorylation and activation ofhe factor. NFAT and ELK modestly precipitated in un-timulated cells (Figure 5A, lane 2), but this interactionas strongly enhanced by serum treatment (Figure 5A,

ane 3). Moreover, NFAT/ELK-1 complex formation wasssociated with increased binding of both transcriptionactors to the �84/�63 promoter region, as revealed by

u8988t cells transfected with the full-length c-Myc promoter constructase activities were assayed after 24 hours. Reporter gene activities areth promoter. Promoter activities are expressed as mean values � SD., and serum responsiveness was measured in the presence or absencelative to the untreated control. (C) PaTu8988 cells were transfected withing site before treatment with FCS. Firefly luciferase reporter activitiesrelative to the activity of untreated �84/�63 wt c-Myc promoter. (D)odies, and in vivo binding to the c-Myc promoter was determined by

84/�63 region. ChIP assay demonstrates time-dependent binding ofwere transfected with siRNA against NFATc1 or control siRNA, andulation with 10% FCS over time.

) PaTluciferl-lengt cellsn” re-binduction

antibhe �Cells

hIP experiments (Figure 5B). Further analysis demon-

smNptcpbdNt

cgctir6w

apbtfm(ctpiTtlm

cTao

FftEccFububcsbmiprCacdE(pNPwap

BA

SIC–L

IVER

,PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

March 2010 NFAT INDUCED HAT RELAY SWITCH 1195

trated that ELK-1 functions as a cofactor to ensureaximal transcription from the c-Myc promoter byFAT. In fact, whereas ELK-1 alone barely induced c-Mycromoter activation, it strongly enhanced NFAT-drivenranscription. Moreover, recruitment of ELK-1 to the-Myc promoter is strictly dependent on the presence andromoter binding of NFAT (Figure 5C and E), as evidencedy the loss of ELK-1 promoter occupancy following NFATepletion. Together, these studies indicate that activatedFAT factors recruit ELK-1 to induce maximal activation of

he c-Myc promoter in pancreatic cancer cells.

NFAT/ELK-1 Transcriptional ComplexFormation Requires the HistoneAcetyltransferase p300 to Activate thec-Myc Promoterp300 is commonly recruited to transcriptional

omplexes to bridge DNA-binding partners and enhanceene transcription through creation of a hyperacetylatedhromatin state.25 ChIP assays indeed displayed a sequen-ial occupancy of the proximal c-Myc promoter that isnitiated by NFAT factors and followed by consecutiveecruitment of acetylated histones and ELK-1 (FigureA). In detail, serum rapidly induced NFAT binding

igure 5. NFAT recruits ELK-1orcooperativec-Mycpromoterac-ivation. (A) Endogenous NFATc1/LK-1 interaction was examined byoimmunoprecipitation in Panc-1ells. Cells were treated with 10%CS before NFATc1 was isolatedsing specific antibodies. NFATc1ound ELK-1 was identified by these of a specific anti-ELK-1 anti-ody. (B) Chromatin immunopre-ipitations were performed withpecific NFATc1 and ELK-1 anti-odies to determine inducible pro-oter binding. PCR was done us-

ng primers specific for the c-mycromoter region harboring the se-um responsive region. (C and D)hIPanalysis todemonstrateNFATndELK-1binding to the �84/�63-Myc promoter region followingepletion of either NFATc1 (C) orLK-1 (D) in Panc-1 cells. (E) The

�84/�63) c-Myc promoter re-orterwascotransfectedwitheitherFATc2,ELK-1,orboth factors intoanc-1 cells, and luciferase activityasexpressedas fold induction rel-tive to basal (�84/�63) c-Mycromoter activity.

ithin 1 hour, followed by enhanced local promoter N

cetylation and ELK-1 binding. NFAT-promoter bindingrecedes histone acetylation and thus remains unaffectedy increased global acetylation through tryptic soy agarreatment (Figure 6B). In contrast, NFAT/ELK-1 complexormation and ELK-1’s recruitment to the c-Myc pro-

oter require the p300 histone acetyltransferase (HAT)Figure 6C and D). Hence, depletion of p300 severelyompromised both transcription factor complex forma-ion and promoter binding, whereas introduction of300 stimulates binding to NFAT and enhances its abil-

ty to activate the c-Myc promoter (Figure 6C and E).ogether, NFAT binding to the serum responsive element is

he initial event in c-Myc promoter activation that stimu-ates HAT-induced histone acetylation at the c-myc pro-

oter and consequently allows accessibility of ELK-1.

NFAT Mediates Mitogen-Induced PancreaticCancer Growth In Vitro and In VivoFinally, we analyzed the relevance of the NFAT/

-Myc axis in serum-induced pancreatic cancer growth.o this end, we arrested cells by serum deprivation andllowed cells to re-enter the cell cycle through applicationf serum. These experiments, performed in control and

FAT-depleted Panc-1 and 8988t cells, revealed a critical

fc1craecrpmsrswjPNaswiN

acegtg

cncrWrdtbtcC

FNuccwp(s d as

BA

SIC–LIV

ER,

PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

1196 KÖENIG ET AL GASTROENTEROLOGY Vol. 138, No. 3

unction of NFAT proteins in serum-driven pancreaticancer growth. Consistent with the data shown in Figure, serum significantly stimulates proliferation in cancerells (Figure 7A). However, depletion of NFAT proteinsendered cells refractory to serum and prevented pancre-tic cancer growth through induction of a G1 arrest, asvidenced by a significant down-regulation of D-typeyclins and their partner kinases (Figure 7B). The biologicelevance of NFAT proteins in c-Myc expression andancreatic cancer growth was examined in a xenograftouse model using Panc-1 cells stably transfected with

hRNA against NFATc1 or control shRNA. Similar to ouresults using small interfering RNA against NFAT, stablehRNA-mediated knockdown of NFATc1 was associatedith reduced levels of c-Myc (Figure 7C). We then in-

ected untransfected Panc-1 control cells, control shRNAanc-1 cells, NFATc1 shRNA clone number 2 andFATc1 shRNA clone number 4 cells in the flanks of

thymic nude mice and determined the effect of NFATilencing on tumor growth. After 5 weeks, the animalsere killed and the tumors extracted. The results, shown

n Figure 7D–F, demonstrate that genetic silencing of

igure 6. NFAT binding to the c-Myc promoter induces sequential hFATc1, ELK-1, and acetylated histone 3 (acH3) to the �84/�63 c-Msing primers specific for the c-myc promoter region harboring the �8-Myc promoter in response to TSA treatment. (C) ChIP analysis to studyells following siRNA-mediated depletion of the histone acetyltransferasas examined by coimmunoprecipitation following overexpression (leftromoter was cotransfected with NFAT and ELK-1 alone or together wi

�84/�63) c-Myc promoter activity. (F) Cells were transfected with theiRNA against p300 or control siRNA. Luciferase activity was expresse

FATc1 caused a significant reduction in tumor mass c

nd that this effect was closely associated with reduced-Myc expression (Figure 7E and F). A representativexample of the tumor mass is shown in Figure 7D. To-ether, these experiments suggest an important role ofhe NFAT/c-Myc axis in mitogen-driven cell cycle pro-ression and tumor growth in the pancreas.

DiscussionIn this study, we examined the expression, trans-

riptional regulation, and function of the most promi-ent ITF members in pancreatic cancer and identified the-Myc proto-oncogene as the predominant factor in se-um-stimulated G1 cell cycle transition and growth.

hereas other ITF proteins have proven to be eitheresistant to mitogen stimulation (c-jun and c-fos) orispensable (egr-1) for pancreatic cancer cell prolifera-ion, only c-Myc fulfilled both criteria. c-Myc inductiony serum, on the other hand, requires prior activation ofhe Ca2�/calcineurin/NFAT signaling and transcriptionascade. In fact, serum rapidly enhances intracellulara2� concentrations in pancreatic cancer cells and, as a

e acetylation and ELK-1 recruitment. (A) Time-dependent binding ofmoter was tested by ChIP analysis. Quantitative PCR was performed3. (B) ChIP analysis to test NFAT/ELK-1 and acH3 occupancy of theT and ELK-1 binding to the �84/�63 c-Myc promoter region in Panc-1T) p300. (D) p300-Dependent interaction between NFATc1 and ELK-1l) or genetic depletion (right panel) of p300. (E) The (�84/�63) c-Myc

00. Luciferase activity was expressed as fold induction relative to basal/�63) c-Myc promoter construct along with NFATc2/ELK-1 and eitherfold induction relative to basal (�84/�63) c-Myc promoter activity.

istonyc pro4/�6NFA

e (HApane

th p3(�84

onsequence of this, causes activation of the calcium

rtiqtrfdniN

iptfilcpWto

Fdcvcaa1ltlSdbdDaNtkrmla(winmstuwamP

BA

SIC–L

IVER

,PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

March 2010 NFAT INDUCED HAT RELAY SWITCH 1197

esponsive phosphatase calcineurin. Calcineurin activa-ion in turn promotes nuclear translocation of preexist-ng cytoplasmic NFAT proteins and allows their subse-uent transcriptional activation. The NFAT family ofranscription factors comprises a group of 5 proteinselated to the Rel family that play a critical role in theunction of many vertebrate tissues.26 Beside their well-ocumented role in T-cell activation, NFAT proteins areow being increasingly recognized for their implications

n tumorigenesis.27–29 This is particularly true for

igure 7. NFAT depletion re-uces c-Myc expression and pan-reatic cancer growth in vitro and inivo. (A) Proliferation of PaTu8988tells after treatment with siRNAgainst NFATc2 or control siRNAnd in the presence or absence of0% FCS for 24 hours was ana-

yzed by [3H]-thymidine incorpora-ion. Data are representative of trip-icate experiments and displayed �D. Successful NFATc2 knock-own is demonstrated by Westernlot analysis. (B) Immunoblotting toetermine the expression of cyclin1 and its related kinases CdK4nd CdK6 upon depletion ofFATc1. (C) Western blot analyses

o verify successful stable NFATc1nockdown in Panc-1 cells. (D) Aepresentative example of tumorass is shown for the Panc-1 cell

ine transfected with either shRNAgainst NFATc1 or control shRNA.

E) Panc-1 cells were transfectedith the indicated constructs and

njected in the flank of male athymicude mice (n � 6 for each experi-ental condition). Weekly mea-

urements were taken from theumors, and the mean tumor vol-me was determined after 5eeks. (F) Expression of NFATc1nd c-Myc was determined by im-unohistochemical analysis ofanc-1 tumors.

FATc1 and NFATc2, which are frequently deregulated c

n cancer. Depending on the cell type and organ, NFATroteins can have similar or opposing functions duringumor development and progression.16,29,30 In NIH-3T3broblasts, for instance, NFATc1 acts as a positive regu-

ator of cell proliferation, inhibits cell death, and inducesell transformation, whereas NFATc2 blocks cell cyclerogression and subverts oncogenic transformation.29

e could most recently confirm a proproliferative func-ion of NFATc1 in pancreatic cancer. In fact, high levelsf NFATc1 were found in pancreatic intraepithelial pre-

ursor lesions (PanIN lesions) and with a high frequency

italtcpmuvwivctalgutobc

tsbUwtw

gar(mactmtlpoasttptHigt

aN

tcwpra

aG1

1

1

1

1

1

1

1

1

1

BA

SIC–LIV

ER,

PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

1198 KÖENIG ET AL GASTROENTEROLOGY Vol. 138, No. 3

n advanced pancreatic cancers.16 We have also shownhat NFATc1 is highly active in pancreatic cancer cellsnd induces anchorage-independent growth and cell pro-iferation through activation of the c-Myc oncogene. Inhis study, we extended our knowledge on the role of thealcineurin/NFATc1-c2/c-Myc axis in the regulation ofancreatic carcinogenesis and identified this signaling-odulated transcriptional mechanism as a central mod-

le in the mitogen-driven pancreatic cancer growth initro and in vivo. Accordingly, inactivation of this path-ay, either through application of CsA or through RNA

nterference-mediated silencing of NFAT factors, pre-ented c-Myc up-regulation by serum and arrested pan-reatic cancer cells at G1 cell cycle phase. The relevance ofhe NFAT/c-Myc axis in pancreatic cancer growth waslso emphasized in xenograft mouse models confirmingoss of c-Myc expression and significantly reduced tumorrowth of subcutaneously injected pancreatic cancer cellspon stable silencing of NFAT. Therefore, disruption ofhe NFAT/c-Myc interplay through specific inactivationf the calcineurin/NFAT pathway in tumor cells mighte an attractive target for future approaches in pancreaticancer therapy.

To understand better how NFAT confers serum induc-ion of c-Myc, we performed extensive c-Myc promotertudies and identified a proximal region that is occupiedy NFAT and responsible for serum responsiveness.pon mitogenic activation of the Ca2�/calcineurin path-ay, NFAT factors rapidly bind to and induce transcrip-

ion from a consensus promoter sequence localizedithin the serum responsive element.Mechanistically, NFAT binding to this promoter re-

ion enhances the net histone acetylase activity associ-ted with the c-Myc promoter and subsequently allowsecruitment of ELK-1, an avian erythroblastosis virus E26v-ets) oncogene homolog-like factor required for maxi-

al activation of the promoter. Detailed sequential ChIPnalysis revealed that NFAT binding is the initial event in-Myc promoter activation that stimulates the HAT p300o create a hyperacetylated chromatin state at the proxi-

al promoter that allows accessibility of ELK-1. Consis-ently, depletion of the HAT p300 severely compromisedocal histone acetylation, prevented NFAT/ELK-1 com-lex formation, and abolished inducible ELK-1 promoterccupancy, resulting in inappropriate c-Myc promoterctivation. Together, these studies extend our under-tanding of how NFAT proteins regulate gene expressionhrough interaction with partnering transcription fac-ors. Based on their weak DNA-binding affinities, NFATroteins frequently interact with site-specific transcrip-ional regulators for sufficient promoter transactivation.ere, we show for the first time that this protein/protein

nteraction (NFAT/ELK-1) involves local chromatin reor-anization and requires p300-mediated histone acetyla-

ion. Future approaches will analyze whether this mech-

nism is c-Myc promoter specific or holds true for otherFAT-regulated target promoters.Taken together, this study uncovered a master role of

he NFAT/p300/ELK axis in mitogen-induced pancreaticancer growth in vitro and in vivo. Based on our findings,e propose a model in which NFAT binding to targetromoters can trigger a HAT relay switch and thus allowecruitment of additional transcription factors throughlteration of the local chromatin structure.

Supplementary Material

Note: To access the supplementary materialccompanying this article, visit the online version ofastroenterology at www.gastrojournal.org, and at doi:0.1053/j.gastro.2009.10.045.

References

1. Almendral JM, Sommer D, Macdonald-Bravo H, et al. Complexityof the early genetic response to growth factors in mouse fibro-blasts. Mol Cell Biol 1988;8:2140–2148.

2. Bravo R. Growth factor-responsive genes in fibroblasts. CellGrowth Differ 1990;1:305–309.

3. Lee MH, Yang HY. Regulators of G1 cyclin-dependent kinases andcancers. Cancer Metastasis Rev 2003;22:435–449.

4. Murphy LO, MacKeigan JP, Blenis J. A network of immediate earlygene products propagates subtle differences in mitogen-acti-vated protein kinase signal amplitude and duration. Mol Cell Biol2004;24:144–153.

5. Chen CR, Kang Y, Siegel PM, et al. E2F4/5 and p107 as Smadcofactors linking the TGF-� receptor to c-myc repression. Cell2002;110:19–32.

6. Sng JC, Taniura H, Yoneda Y. A tale of early response genes. BiolPharm Bull 2004;27:606–612.

7. Yordy JS, Muise-Helmericks RC. Signal transduction and the Etsfamily of transcription factors. Oncogene 2000;19:6503–6513.

8. Herold S, Herkert B, Eilers M. Facilitating replication understress: an oncogenic function of MYC? Nat Rev Cancer 2009;9:441–444.

9. Murphy LO, Blenis J. MAPK signal specificity: the right place at theright time. Trends Biochem Sci 2006;31:268–275.

0. Hsu T, Trojanowska M, Watson DK. Ets proteins in biologicalcontrol and cancer. J Cell Biochem 2004;91:896–903.

1. Thiel G, Cibelli G. Regulation of life and death by the zinc fingertranscription factor Egr-1. J Cell Physiol 2002;193:287–292.

2. Jochum W, Passegue E, Wagner EF. AP-1 in mouse developmentand tumorigenesis. Oncogene 2002;20:2401–2412.

3. Eilers M, Eisenman RN. Myc’s broad reach. Genes Dev 2008;22:2755–2766.

4. Ghaneh P, Costello E, Neoptolemos JP. Biology and managementof pancreatic cancer. Postgrad Med J 2008;84:478–497.

5. Schneider G, Siveke JT, Eckel F, et al. Pancreatic cancer: basicand clinical aspects. Gastroenterology 2005;128:1606–1625.

6. Buchholz M, Schatz A, Wagner M, et al. Overexpression of c-mycin pancreatic cancer caused by ectopic activation of NFATc1 andthe Ca2�/calcineurin signaling pathway. EMBO J 2006;25:3714–3724.

7. Buck A, Buchholz M, Wagner M, et al. The tumor suppressorKLF11 mediates a novel mechanism in transforming growth fac-tor �-induced growth inhibition that is inactivated in pancreaticcancer. Mol Cancer Res 2006;4:861–872.

8. Ellenrieder V, Zhang JS, Kaczynski J, et al. Signaling disruptsmSin3A binding to the Mad1-like Sin3-interacting domain of

TIEG2, an Sp1-like repressor. EMBO J 2002;21:2451–2460.

1

2

2

2

2

2

2

2

2

2

2

3

R

DUe

A

C

F

TG3DDC

March 2010 NFAT INDUCED HAT RELAY SWITCH 1199

9. Baubet V, Le Mouellic H, Campbell AK, et al. Chimeric greenfluorescent protein-aequorin as bioluminescent Ca2� reportersat the single-cell level. Proc Natl Acad Sci U S A 2000;97:7260–7265.

0. Jones K, Hibbert F, Keenan M. Glowing jellyfish, luminescenceand a molecule called coelenterazine. Trends Biotechnol 1999;17:477–481.

1. Fernandez-Zapico ME, Gonzalez-Paz NC, Weiss E, et al. Ectopicexpression of VAV1 reveals an unexpected role in pancreaticcancer tumorigenesis. Cancer Cell 2005;7:39–49.

2. Dunn KL, Espino PS, Drobic B, et al. Ras-MAPK signal transduc-tion pathway, cancer and chromatin remodeling. Biochem CellBiol 2005;83:1–14.

3. Yagi K, Furuhashi M, Aoki H, et al. c-myc Is a downstream targetof the Smad pathway. J Biol Chem 2002;277:854–861.

4. Rao A, Luo C, Hogan PG. Transcription factors of the NFAT family:regulation and function. Annu Rev Immunol 1997;15:707–747.

5. Wang L, Tang Y, Cole PA, et al. Structure and chemistry of thep300/CBP and Rtt109 histone acetyltransferases: implicationsfor histone acetyltransferase evolution and function. Curr OpinStruct Biol 2008;18:741–747.

6. Crabtree GR, Olson EN. NFAT signaling: choreographing the so-cial lives of cells. Cell 2002;109:67–79.

7. Neal JW, Clipstone NA. A constitutively active NFATc1 mutantinduces a transformed phenotype in 3T3-L1 fibroblasts. J BiolChem 2003;278:17246–17254.

8. Buchholz M, Ellenrieder V. An emerging role for Ca2�/cal-cineurin/NFAT signaling in cancerogenesis. Cell Cycle 2007;6:

16–19. M

9. Robbs BK, Cruz AL, Werneck MB, et al. Dual roles for NFATtranscription factor genes as oncogenes and tumor suppressors.Mol Cell Biol 2008;28:7168–7181.

0. Jauliac S, López-Rodriguez C, Shaw LM, et al. The role of NFATtranscription factors in integrin-mediated carcinoma invasion.Nat Cell Biol 2002;4:540–544.

Received June 28, 2009. Accepted October 29, 2009.

eprint requestsAddress requests for reprints to: Volker Ellenrieder, MD,

epartment of Gastroenterology and Endocrinology, Philipps-niversity of Marburg, Baldingerstrasse, 35043 Marburg, Germany.-mail: ellenrie@med.uni-marburg.de.

cknowledgmentsA.K and T.L. contributed equally to the paper.

onflicts of interestThe authors disclose no conflicts.

undingSupported by the Deutsche Forschungsgemeinschaft (DFG, SFB-

R17; to V.E., A.G., and J.M.) and the Max Eder program of theerman Cancer Research Foundation (Deutsche Krebshilfe, 70-022-El I; to V.E.) and by the Miles and Shirley Fiterman Center forigestive Diseases, Division of Gastroenterology and Hepatology,ivision of Oncology Research, Mayo Clinic Cancer Center, Mayolinic Pancreatic SPORE P50 CA102701 and CA136526 (to

.E.F.-Z.).

BA

SIC–L

IVER

,PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

P(btPmsgbtcMcN(twtsgciam�CJaacptTttumtTa

u2cug5csMf

aBc

afsuwc3gfw

ppplmwpseawpfiirTHt

sicpmmsn1p

1199.e1 KÖENIG ET AL GASTROENTEROLOGY Vol. 138, No. 3

Supplementary Materials and Methods

Cell Lines, Plasmids, and Transfection

The human pancreatic adenocarcinoma cell linesaTu8988t and Panc-1 were obtained from Dr Elsasser

Department of Cell Biology, Philipps University, Mar-urg, Germany) and the American Type Culture Collec-ion (ATTC, Manassas, VA), respectively. PaTu8988t andanc-1 cells were maintained in Dulbecco’s modifiedinimal essential medium (Invitrogen, Carlsbad, CA)

upplemented with 10% fetal calf serum (FCS) (Invitro-en) and 50 �g/mL Gentamycin (Amaxa Inc, Gaithers-urg, MD). The full-length human nuclear factor of ac-ivated T cells (NFAT) c1 and NFATc2 expressiononstructs were kindly provided by Dr A. Rao (Harvard

edical School, Boston, MA). The HA-tagged NFATc1onstruct was generated by insertion of wild-typeFATc1 into the pCMV-HA vector supplied by Clontech

Mountain View, CA) using the XhoI and BamHI restric-ion sites. The Ets-like gene (ELK)-1 expression constructas provided by A.D. Sharrocks (University of Manches-

er, Manchester, UK). The wild-type c-Myc reporter con-truct (�2446 to �334) and its deletions were kindlyifted by Joan Massague (Memorial Sloan-Kettering Can-er Center, New York, NY). All constructs were insertednto a pGL3 vector (Promega, Madison, WI) using KpnInd SpeI restriction sites. To generate the single pointutations of the NFAT consensus sequence within the84/�63 c-Myc promoter construct, we used the Quick-hange site-directed mutagenesis kit (Stratagene, La

olla, CA) according to the manufacturer’s instructionsnd with the following primers: NFAT, 5=-ctc-gaggcttggcgggcccaagaacggagggag-3= and its congruentomplementary strands. For transient transfection of ex-ression constructs, Panc-1 and PaTu8988t cells wereransfected 12 hours after seeding at 70% density usingransfast (Promega, Mannheim, Germany) as transfec-

ion reagent according to the manufacturer’s instruc-ions. Small interfering RNA (siRNA) was transfectedsing the Transmessenger reagent (Qiagen, Hilden, Ger-any) according to the manufacturer’s instructions. For

ransfection of siRNA or short hairpin RNA (shRNA),ransmessenger (Qiagen) was used as transfection re-gent.

The following RNAi oligonucleotide sequences weresed: NFATc1 No. 1, 5=-ggucauuuucguggagaaatt-3=; No., 5=-ggacuccaaggucauuuuctt-3=; NFATc2 No. 1, 5=-cauuaaacaggagcagaatt-3=, No. 2, 5=-gcugaugagcggauccu-att-3=; p300, 5=-cagagcaguccuggauuagtt-3=, 5=-cuaaucca-gacugcucugtt-3=; c-Myc, 5=-ggaacgagcuaaaacggagtt-3=,=-cuccguuuuagcucguucctc-3=; EGR1, 5=-cgucgguggccac-acgua, 5=-uacgugguggccaccgacg-3=; ELK-1 smartpooliRNA (L-003885-00-0020; Thermo Scientific, Waltham,

A). As a negative control, the silencer negative-control

rom Ambion (Applied Biosystems) was used. N

Antibodies and ReagentsCyclosporin A was purchased from Sigma-Aldrich

nd U0126 was from Calbiochem-Novabiochem, EMDiosciences (San Diego, CA). Tryptic soy agar was pur-hased from Sigma-Aldrich.

Quantitative Real-time Polymerase ChainReaction AnalysisRNA was extracted using the RNeasy Mini Kit,

nd first-strand complementary DNA was synthesizedrom 2 �g total RNA using random primers and Omni-cript reverse transcriptase (Qiagen). Cyclophilin wassed as housekeeping gene. The following primer pairsere used: Cyclophilin: 5=-caccgtgttcttcgacatca-3=, 5=-ag-

atttgccatggacaagat-3=; c-Myc: 5=-gctcctggcaaaaggtcaga-=, 5=-cagtgggctgtgaggaggtt-3=; EGR-1: 5=-tttcacgtcttggt-cctttt-3=, 5=-cactccctgcccccctta-3=. All primers were obtainedrom Biomers (Ulm, Germany). Real-time PCR experimentsere performed in triplicate and are displayed � SD.

Generation of Aequorin-Expressing Cells byLentiviral TransfectionFor calcium measurements, a green fluorescent

rotein (GFP)-aequorin (G5A) fusion protein was ex-ressed using a lentiviral expression system. G5A wasackaged into lentivirus using HEK293 cells and the

entiviral packaging plasmid psPAX2 and envelope plas-id. For this purpose, 20 �g/10-cm dish of the cis-vectoras cotransfected into HEK293 cells together with 15 �gsPAX2 (“trans plasmid”) and with 6 �g pMD2G by thetandard calcium phosphate precipitation method. Six toight hours after transfection, medium was changed; and,fter 48 hours, supernatant containing lentiviral particlesas collected and centrifuged at 3000 rpm at room tem-erature. Subsequent to filtration through a 0.45-mmlter, virus particles were concentrated on a sucrose cush-

on (20% sucrose in Tris-buffered saline buffer; 26,000pm 2 hours) and were resuspended in RPMI medium.he viral titer was determined by serial dilution inEK293 cells. PaTu8988t cells were transduced with 20

ransducing units of recombinant virus.

Subcellular Fractionation,Coimmunoprecipitation, and ImmunoblottingSubcellular fractionation was performed as de-

cribed previously.1 Briefly, cells were washed twice withce-cold phosphate-buffered saline (PBS) and collected byentrifugation at 2500g at 4°C. Lysates were then resus-ended in buffer A (10 mmol/L HEPES, ph 7.9, 10mol/L KCl, 0.1 mmol/L EDTA, 0.1 mmol/L EGTA, 1mol/L DTT, proteinase inhibitors) for 15 minutes and

ubsequently centrifuged for 2 minutes at 5000g. Super-atants were transferred to new cups and centrifuged at2,000g for an additional 20 minutes. Pellets were resus-ended in buffer C (20 mmol/L HEPES, ph 7.9, 0.4 mol/L

aCl, 1 mmol/L EDTA, 1 mmol/L EGTA, 1 mmol/L

Dcf

wibBFw(iale�(jt

cPtf

cata1ead

pTmrP�s5m

1

2

March 2010 NFAT INDUCED HAT RELAY SWITCH 1199.e2

TT, proteinase inhibitors) and incubated on ice. A finalentrifugation step at 12,000g for 20 minutes was per-ormed to separate nuclear proteins from cellular debris.

Upon protein extraction and gel transfer, membranesere washed in Tris-buffered saline washing buffer and

ncubated with peroxidase-conjugated secondary anti-odies. Western Blotting Detection Reagent (Amershamiosciences, Aylesbury, UK) was used for visualization.or coimmunoprecipitation analyses, cells were lysedith immunoprecipitation buffer (10 mmol/L Tris/HCl

pH 7.4), 150 mmol/L NaCl, 1% Triton X-100, 1% Non-det P-40) containing protease inhibitors (400 Amol/Lprotinin, 50 �mol/L leupeptin, and 0.5 mmol/L Pefab-oc (Roche Diagnostics, Mannheim, Germany). Homog-nized lysates (500 �g) were immunoprecipitated with 4L of the indicated antibody and protein G-agarose

Roche Diagnostics). The immunoprecipitates were sub-ected to sodium dodecyl sulfate-polyacrylamide gel elec-rophoresis analysis.

[3H]-Thymidine Assay and Flow CytometryCell growth was measured by [3H]-thymidine in-

orporation as described previously.2 Panc-1 andaTu8988t cells were seeded in 24-well plates and cul-ured in medium containing 10% FCS. Cells were trans-

ected with indicated siRNA and incubated in medium

ontaining 10% FCS. [3H]-thymidine (0.5 �Ci/well) wasdded during the last 5 hours of incubation. Cells werehen washed with cold 5% trichloroacetic acid, and thecid-insoluble fraction was dissolved by incubation withmol/L NaOH for 30 minutes at 37°C. Radioactivity was

valuated with a scintillation counter. All proliferationssays were performed in triplicate in at least 2 indepen-ent experiments.For flow cytometric analysis, cells were seeded in 6-well

lates and cultured in medium containing 10% FCS.wenty-four hours after seeding, cells were treated withedium containing 10% FCS or without FCS. For fluo-

escein diacetate staining, cells were resuspended in 1 mLBS at 106 cells/mL, and 10 �L of fluorescein diacetate (1g/mL) was added to the cells for 10 minutes. For FITC

taining, cells were resuspended in 1 mL PBS containing�L FITC (1 �g/mL), 20 �L propidium iodide (2 mg/L), and 50 �L RNase (100 �g/mL) for 30 minutes.

References

. Buck A, Buchholz M, Wagner M, et al. The tumor suppressor KLF11mediates a novel mechanism in transforming growth factor �-in-duced growth inhibition that is inactivated in pancreatic cancer.Mol Cancer Res 2006;4:861–872.

. Ellenrieder V, Zhang JS, Kaczynski J, et al. Signaling disruptsmSin3A binding to the Mad1-like Sin3-interacting domain of TIEG2,

an Sp1-like repressor. EMBO J 2002;21:2451–2460.