Cancer Research High Levels of Hsp90 Cochaperone p23...

Mole

HigProIncr

NataliEllinoLuiz F

Abst

Intro

p23first cone restabil90 (Hsteroidhormo

Author3EnviroNYU SHospitaYork; 6

SciencConnecand 8DPaulo,

Note:Resear

Corresology, NPhone:nyumc.

doi: 10

©2010

Cance8446

Published OnlineFirst September 16, 2010; DOI: 10.1158/0008-5472.CAN-10-1590

Canceresearch

cular and Cellular Pathobiology

h Levels of Hsp90 Cochaperone p23 Promote Tumorgression and Poor Prognosis in Breast Cancer by

R

easing Lymph Node Metastases and Drug Resistance

e E. Simpson1, W. Marcus Lambert2, Renecia Watkins2, Shah Giashuddin5, S. Joseph Huang6,
r Oxelmark2, Rezina Arju2, Tsivia Hochman3, Judith D. Goldberg3, Robert J. Schneider2, ernando Lima Reiz7, Fernando Augusto Soares8, Susan K. Logan1,4, and Michael J. Garabedian2,4
ractp23

lizes udynamcells (MwhichTo gagene epresenidentifPMP22by p23MCF-7contropatien

receptone-bindin

s' Affiliationmental Mchool of Ml Queens aDepartmees, Yaleticut; 7InstepartmentSP, Brazil

Supplemench Online (h

ponding AuYU Schoo212-263-76org.

.1158/0008-

American A

r Res; 70(

Download

is a heat shock protein 90 (Hsp90) cochaperone located in both the cytoplasm and nucleus that stabi-nliganded steroid receptors, controls the catalytic activity of certain kinases, regulates protein-DNAics, and is upregulated in several cancers. We had previously shown that p23-overexpressing MCF-7CF-7+p23) exhibit increased invasion without affecting the estrogen-dependent proliferative response,suggests that p23 differentially regulates genes controlling processes linked to breast tumor metastasis.in a comprehensive view of the effects of p23 on estrogen receptor (ER)-dependent and -independentxpression, we profiled mRNA expression from control versus MCF-7+p23 cells in the absence andce of estrogen. A number of p23-sensitive target genes involved in metastasis and drug resistance wereied. Most striking is that many of these genes are also misregulated in invasive breast cancers, including, ABCC3, AGR2, Sox3, TM4SF1, and p8 (NUPR1). Upregulation of the ATP-dependent transporter ABCC3conferred resistance to the chemotherapeutic agents etoposide and doxorubicin in MCF-7+p23 cells.+p23 cells also displayed higher levels of activated Akt and an expanded phosphoproteome relative tol cells, suggesting that elevated p23 also enhances cytoplasmic signaling pathways. For breast cancerts, tumor stage together with high cytoplasmic p23 expression more accurately predicted disease recur-and mortality than did stage alone. High nuclear p23 was found to be associated with high cytoplasmic
rence
p23, therefore both may promote tumor progression and poor prognosis by increasing metastatic potentialand drug resistance in breast cancer patients. Cancer Res; 70(21); 8446–56. ©2010 AACR.

p23 ismice dmentreceptrecruiincludHsp90

duction

is a ubiquitous and evolutionarily conserved proteinharacterized as a component of unliganded progester-ceptor complexes (1, 2). p23 acts as a cochaperone,izing the ATP-bound form of heat shock proteinsp90), thereby stabilizing Hsp90 interactions with

rs (3) and assisting Hsp90 in increasing theg affinities of many steroid receptors (4–6).

teinsThereprostacontroreducep23

viouslbindindepenshownbreastinto mliferattargetresponTFF1c-myc

ns: Departments of 1Pharmacology, 2Microbiology,edicine, and 4Urology, and NYU Cancer Institute,edicine, and 5Department of Medicine, New Yorknd Weill Cornell Medical College, New York, Newnt of Obstetrics, Gynecology, and ReproductiveUniversi ty School of Medicine, New Haven,ituto de ensino e pesquisa, Hospital Sirio Libanes,of Anatomic Pathology, Hospital AC Camargo, Sao

tary data for this article are available at Cancerttp://cancerres.aacrjournals.org/).

thor: Michael J. Garabedian, Department of Microbi-l of Medicine, 550 First Avenue, New York, NY 10016.62; Fax: 212-263-8276; E-mail: michael.garabedian@

5472.CAN-10-1590

ssociation for Cancer Research.

21) November 1, 2010

Research. on January 29, 20cancerres.aacrjournals.org ed from

required for survival in mammals because p23-nullie perinatally and fail to undergo proper lung develop-due in part to decreased activity of the glucocorticoidor (7). p23 regulates steroid receptors and telomerasetment to DNA (8–11), and stabilizes some kinases,ing PKR, FAK-2, Flt3, and Raf (12–15). In addition to-dependent events, p23 might also regulate client pro-autonomously via its innate chaperone activity (16).is also evidence that mammalian p23 is a cytoplasmicglandin E synthetase (17), although this function isversial because p23 knockout mice do not exhibitd prostaglandin E synthetase activity (18).has many implicated roles in cancer. We have pre-y shown that p23 overexpression increases ligandg to the estrogen receptor α (ER) and enhances ER-dent transcriptional activation (19, 20). We have alsothat p23 overexpression in human ER-positive MCF-7cancer cells promotes a transition from noninvasiveigratory and invasive cells, without affecting cell pro-ion (10). Mechanistically, p23 selectively increases ER-gene transcription and ER recruitment to the estrogense element (ERE) of the metastasis-associated gene
(pS2), rather than to that of proliferative genes like(10).
20. © 2010 American Association for Cancer

http://cancerres.aacrjournals.org/

p23(10) aalso rtelomestagesmatiotion (2(27, 28treatmagentdanamteins (Give

a modstrongand fucells (1gene eof p23come

Mate

Cell liMC

(MCFviouslcloneselectependewith s

MicroTot

MCF-7eitherRNAtranschybridwasheferentand nThe pViewetest wprobewere cand dvaluesL2L owashingene o

RNA iCell

presen(TAM)

contr(InvitrsynthreversamersinstruTaq RMix (Rtion Sreal-ti

ChromCell

the etformato sontion oprecip

Effects of p23 on Tumor Progression in Breast Cancer

www.a


protein expression increases with breast tumor stagend is upregulated in metastatic cancers (21, 22). p23egulates the binding and activity of the oncoproteinrase (11, 23, 24), known to impact both early and lateof tumorigenesis, including epithelial cell transfor-

n (25), human mammary epithelial cell immortaliza-6), and metastasis in ER-negative breast cancer cells). In addition, p23 desensitizes mammalian cells toent with the Hsp90 inhibitor and chemotherapeuticgeldanamycin, thereby protecting Hsp90 from gel-ycin inhibition and stabilizing Hsp90 client oncopro-15).n the implicated role of p23 in cancer and the fact thatest increase in p23 protein expression causes suchestrogen-dependent and -independent transcriptionalnctional effects on migration and invasion in MCF-70), we further characterized the effect of p23 on globalxpression in MCF-7 cells, and examined the expression
protein in primary tumor samples with appended out- modif
was su(HC20K9/14rabbitDNA wused f

RNA iON

(L-003pool sto thehormotransfthe m(RNAiafter tminedanalyzprecipantiboα-tubuMA3-424 hoVerseafter pdoxoring wuntreawas co

ChemCel

suppleride (A(E1383Sigma

data from women with breast cancer.

rials and Methods

nesF-7 control and p23-overexpressing MCF-7 cells-7+p23) were cultured and hormone-starved as pre-y described (10). A single MCF-7+p23 overexpressing(clone 7) was used in the experiments. Validation ofd p23-sensitive genes was performed using an inde-nt, previously described MCF-7+p23 clone (clone 8)imilar results (Supplementary Fig. S1; ref. 10).

array analysisal RNA was isolated using the RNeasy kit (Qiagen) fromcontrol and MCF-7+p23 cells treated 16 hours withethanol vehicle or 1 nmol/L 17-β-estradiol (E2). Totalwas reverse transcribed into cDNA and in vitroribed to cRNA. The cRNA was fragmented, labeled,ized to the human U133A 2.0 Affymetrix GeneChip,d, and scanned. Data are representative of four dif-conditions performed in duplicate (a total of eight)ormalized using Robust Multichip Average Express.rimary data were analyzed using MultiExperimentr to determine hierarchical clustering, and the ANOVAas performed on each of the conditions to find thosesets that were considered significant (P < 0.05). Genesonsidered significant if they passed the ANOVA testisplayed a fold-change of >2.0 when the logarithmicin MCF-7+p23 and control cells were compared. Thenline microarray data analysis tool (http://depts.gton.edu/l2l/) was used to analyze the gene sets forntology.

solation and quantitative real-time PCRs were incubated in the absence (ethanol vehicle) or
ce of 1 nmol/L E2 or 100 nmol/L 4-hydroxy-tamoxifenfor 16 hours before harvesting. Total RNA from MCF-7
in quamine

acrjournals.org

Research. on January 29, 20cancerres.aacrjournals.org Downloaded from

ol and MCF-7+p23 cells was extracted with TRIzologen) as described by the manufacturer. cDNA wasesized from 1 μg of RNA using the Superscript IIIe transcriptase (Invitrogen) and random primer hex-(Amersham Biosciences) following the manufacturer'sctions. cDNAs were amplified with the SYBR Greeneady Mix (USB) or Fast Start Universal SYBR Masteroche) using MyiQ Single-Color Real-Time PCR Detec-ystem from Bio-Rad. The primers used for quantitativeme PCR are listed in Supplementary Table S1.

atin immunoprecipitation experimentss were hormone-starved for 3 days and treated withhanol vehicle or 10 nmol/L E2 for 45 minutes and thenldehyde cross-linked, lysed, washed, and pelleted priorication according to methods described (10). Sonica-f chromatin, preparation of protein A beads, immuno-itation of chromatin, and semiquantitative PCR wereied from methods described (29). Chromatin (125 μg)bjected to immunoprecipitation with 6 μg ER antibody, Santa Cruz Biotechnology), 2 μg acetyl-histone H3-antibody (Millipore, Inc.), or equivalent amount ofIgG (I-8140, Sigma). Following cross-link reversal,as purified using Qiagen PCR purification kit. Primersor PCR are listed in Supplementary Table S2.

nterference-TARGET plus Non-Targeting (D-001810-10-05), ER401-00-0005), and ABCC3 (L-007312-00-0005) SMARTiRNA were purchased from Thermo Scientific. PriorsiER and siABCC3 RNA transfections, cells werene-starved for 1 and 3 days, respectively. Cells wereected using oligofectamine (Invitrogen) according toanufacturer's instructions. For ER RNA interference) experiments, RNA and protein were isolated 36 hourshe transfection. mRNA expression levels were deter-by quantitative real-time PCR, and protein levels wereed following extraction from cells using radioimmuno-itation assay buffer (10) and immunoblotting usingdies for ER (1:1,000; HC20, Santa Cruz Biotechnology),lin (1:2,000; MMS-489P, Covance), and p23 (1:2,000; JJ3;14, Thermo Fisher). For ABCC3 RNAi experiments,urs posttransfection cells were dissociated usingne-EDTA and plated at the same density. Two dayslating, cells were treated for 24 hours with 100 ng/mLubicin. Percent survival was determined by cell count-ells in each condition. RNA was extracted from theted cells and knockdown of ABCC3 mRNA expressionnfirmed.

otherapeutic resistance assaysls were hormone-starved for three days. Media weremented with 0.01 to 1 μg/mL doxorubicin hydrochlo-driamycin; D1515, Sigma), 0.02 to 20 μmol/L etoposide, Sigma), or 0.04 to 40 μmol/L camptothecin (C9911,) for 48 hours and percent survival was determined
druplicate for each condition using the sulphorhoda-B assay (30).
Cancer Res; 70(21) November 1, 2010 8447



ImmumicroTiss

who hLudwiand 1coresmen a(JJ3; Mgists sin a bl(0–3)scoresbut costrongused tdiseasp23 leUnivarels weels wedifferefor mo

Resu

Higheexprein invOur

amoundepenexpresWe prmicrohormo2-foldidentiin thewere(genesmentaIn f

genesgenesMCF-7also mFig. SPMP223.7-focells)14.2-fodent MPMP2in celland isCharc

ABC tposidetherapinhibiin MCshownDownobservcanceTM4SmotilisupprT-cellThe

were vand Arespecells.fold, aupregdegre

Simpson et al.

Cance8448


nohistochemistry and tissuearray analysisue arrays contained tissue specimens from individualsad undergone surgical removal of breast tumors atg Cancer Center in São Paulo, Brazil, between 1983993 with a 12-year follow up. Representative tissue0.6 mm in diameter were extracted from each speci-nd mounted in paraffin blocks and stained for p23A3-414, Thermo Fisher). Two board-certified patholo-cored the specimens from the breast tissue microarrayinded fashion. Each was scored for intensity of stainingin the nucleus and in the cytoplasm. For intensity, no staining above background was scored as 0, weaknvincing staining as 1, moderate staining as 2, andstaining as 3. Spearman correlation coefficients wereo determine associations between variables. Age ande characteristics were compared between cytoplasmicvels using χ2 tests, t-tests, and Fisher's exact test.iate and multivariable Cox proportional hazards mod-re developed to predict recurrence or mortality. Mod-re selected using stepwise procedures in SAS with
nces in the likelihood ratio test statistics as the criteria depen
p23 inor 17-and usion ofor pSBy corathergestsexprementsdisrupplemeexpredata ssion othat cHsp90We

agonisless rementrevealamplechangof p23clusteinvolvin clusadhesrangesponsa groucessesion tr

del selection.

lts

r p23 protein levels cause changes in genession in MCF-7 cells that are also observedasive breast cancersprevious study showed that a small increase in thet of p23 protein expression caused specific estrogen-dent and -independent changes in MCF-7 cell genesion and function favoring tumor progression (10).ofiled genes globally regulated by p23 using Affymetrixarray technology (Fig. 1A) on RNA extracted fromne-starved control or MCF-7 cells stably expressingmore p23 protein (MCF-7+p23; ref. 10). The analysisfied 359 known human genes sensitive to p23 levelsabsence of estrogen treatment. Of these genes, 179

upregulated and 180 were downregulated by >2-foldand corresponding probe sets are listed in Supple-ry Tables S5 and S6).act, 50% of the genes most upregulated and 61% of themost downregulated by p23 overexpression (i.e., thoseexhibiting >5-fold changes in expression between-control and MCF-7+p23 cells in the microarray) wereisregulated in invasive breast cancers (Supplementary2). Many of these genes were validated, including2, AGR2, and ABCC3 (upregulated in the microarrayld, 5.2-fold, and 5.1-fold, respectively, in MCF-7+p23and p8/NUPR1, TM4SF1, and Sox3 (downregulatedld, 9.2-fold, and 7.9-fold, respectively) from indepen-CF-7+p23 lines (Fig. 1B and Supplementary Fig. S1).

2 is an integral membrane protein known to play a role-cell and cell-matrix interactions in multiple cell types,
implicated in hereditary human neuropathies such asot-Marie-Tooth syndrome (31). ABCC3 (MRP3) is an
Althoucharac

r Res; 70(21) November 1, 2010


ransporter with specificity for cellular export of eto-s and antifolates, both of which are used as chemo-eutic agents (32). AGR2 is a secretory protein, p53tor, and late-stage breast cancer marker (33, 34). AsF-7+p23 cells, PMP22, ABCC3, and AGR2 have also beento be upregulated in advanced cancers (32–35).

regulation of p8, TM4SF1, and Sox3, similar to what ised in MCF-7+p23 cells, is also associated with breastr progression (36, 37). p8 is a nuclear phosphoprotein,F1 is a tetraspanin family member implicated in cellty, and studies in Xenopus allude to Sox3 as a tumoressor through its ability to bind to β-catenin and inhibitfactor signaling (38).gene expression changes observed in the microarrayalidated by quantitative real-time PCR. PMP22, ABCC3,GR2 were upregulated 33-fold, 3.5-fold, and 1.5-fold,ctively, in MCF-7+p23 cells compared with controlp8, TM4SF1, and Sox3 were reduced by 8.6-fold, 30.8-nd 6.1-fold, respectively (Fig. 1C). Interestingly, genesulated by p23 overexpression also displayed varyinges of dependency on Hsp90. As expected for a genedent upon Hsp90-p23 complexes, disruption of Hsp90-teractions in MCF-7+p23 cells with either geldanamycinAAG, the geldanamycin analogue with lower toxicitysed in chemotherapeutic regimens, decreased expres-f ABCC3 (Supplementary Fig. S3). The same held true2 and c-myc, both well characterized ER target genes.ntrast, expression of AGR2 was not reduced, butwas enhanced upon inhibition of Hsp90. This sug-that an Hsp90-sensitive pathway may repress AGR2ssion. Surprisingly, geldanamycin or 17-AAG treat-had no effect on the expression of PMP22, despitetion of Hsp90-p23 interaction by geldanamycin (Sup-ntary Fig. S3). Thus, the effect of p23 on PMP22 genession seems Hsp90 independent. Collectively, thesehow that p23 protein levels can modulate the expres-f genes misregulated in advanced breast cancers andertain p23-sensitive changes in gene expression areindependent.also found that 357 genes were sensitive to the ERt E2 in MCF-7 cells, many of which became more orsponsive to E2 by p23 overexpression (Fig. 1D, Supple-ary Fig. S4, Supplementary Table S7). Heat mapsed functional similarities between these genes. For ex-, cluster A (Fig. 1D, yellow bar) represents genes thated from E2 responsive to E2 unresponsive as a resultoverexpression and can be divided into two sub-

rs based on function. In cluster A1, 23% of genes areed in transcriptional regulation, including Sox11, andter A2, 61% of genes are involved in migration and cellion. Cluster B (red bar) represents genes involved in aof biological processes that changed from E2 unre-ive to E2 responsive. Cluster C (green bar) representsp of genes also involved in a variety of biological pro-that are inversely regulated by p23 and E2, includingansport (16%) and regulation of transcription (13%).
gh they did not fall into specific clusters, other genesteristically involved in migration and invasion are also
Cancer Research



sensitmolecsuppreGen

p23 ogenesthe hian enzupregmetasp23 inexpres

HigheH3 acTo

increaactivachromperforand 1The Pternat

FigureoverexpexpressA, depito examWT, wiheat mgenes mbreastchangeMCF-7cells intreatmequantitData arindepennormalipresent(RFC) fcells, sD, clusp23 oveor presHeat mrendereA1, A2,(clusteroverexpbar), geby p23genes mcells ininvasionand MTSS1.


www.a


ive to p23 overexpression, and include cell adhesionule 1 (CADM1), laminin B1 (LAMB1), and metastasisssor 1 (MTSS1; Fig. 1B).e ontology of the genes induced by p23 revealed thatverexpression particularly affects the expression ofinvolved in metabolism (Supplementary Table S3) withghest significance attributed to kynureninase (KYNU),yme involved in tryptophan metabolism. KYNU is alsoulated in ER-negative breast cancer cells that havetasized to the lung (39), further supporting a role of
promoting tumor progression through altering genesion patterns.
scriptwe pr

acrjournals.org


r p23 protein levels result in increased histoneetylation at PMP22 and ABCC3 gene promoterstest whether the mechanism of the p23-mediatedse in PMP22 or ABCC3 gene expression is throughtion of transcription at corresponding promoters,atin immunoprecipitation (ChIP) experiments weremed to detect acetylation of histone H3 lysines 94 (H3-AcK9/14), marks of transcriptional activation.MP22 locus encodes two transcripts that contain al-ive exons 1A and 1B (Fig. 2A). Because the 1B tran-

1. The effects of p23ression on geneion in MCF-7 cells.ction of the procedure usedine p23-responsive genes.

ld type. B, microarrayap and fold-change of sixisregulated in invasive

cancers with >5-foldin expression betweencontrol and MCF-7+p23the absence of E2nt and verified byative real-time PCR (C).e means of at least threedent experiments,zed to GAPDH, anded as relative fold-changerom that of MCF-7 controlet to 1. Error bars, SE.ters of genes affected byrexpression in the absenceence of E2 treatment.ap reveals clusters of genesd unresponsive (clustersyellow bar) or responsiveB, red bar) to E2 by p23ression. Cluster C (greennes inversely regulatedand E2. Also listed areisregulated in MCF-7+p23

volved in migration and, includingCADM1, LAMB1,

is predominantly expressed in nonneuronal cells (40),edicted that increases in histone acetylation at the




PMP2wouldHistonthe P(Fig. 2in hispromohad bdeteccontrop23-mcreasehyperascripti

Downby E2Our

regulaare neplorep23, wE2 ormeasutreatmroughland Blevelsand Erepresgenes

mentin botThus,theseof PMreversRec

depentargetPMP2(42), amoterIn MCPMP2(Fig. 3tary FrelativUPS rER binPMP2catinging. TrecruiPMP2upon pto whHoweactivaInte

Figuresites (Tperformboth seas relat

Simpson et al.

Cance8450


2 locus in MCF-7+p23 cells relative to control cellsbe most likely to occur at the promoter of 1B (P2).e H3-AcK9/14 levels relative to input were higher atMP22 P2 in MCF-7+p23 cells than in control cellsB and C, top). A similar, albeit less robust increasetone H3-AcK9/14 was also observed at the ABCC3ter (Fig. 2B and C, bottom). The ABCC3 promoterasal levels of histone acetylation, which could beted above background (nonspecific IgG) levels inl cells (Fig. 2B). This suggests that a mechanism ofediated transcriptional activation is through in-d recruitment of chromatin-modifying enzymes thatcetylate histone H3-K9/14 and reinitiate gene tran-on from active promoters.

regulation of PMP22 and ABCC3 gene expressionis associated with ER recruitment to distal EREsmicroarray data revealed that many genes positivelyted by p23 overexpression (e.g., PMP22 and ABCC3)gatively regulated by E2 treatment. To further ex-the E2-mediated repression of genes upregulated bye treated MCF-7 control and MCF-7+p23 cells withthe selective estrogen receptor modulator TAM andred mRNA expression of PMP22 and ABCC3. E2ent repressed PMP22 and ABCC3 gene expressiony 2-fold in both control and MCF-7+p23 cells (Fig. 3A, left). Moreover, when comparing total expressionof PMP22 and ABCC3 between untreated control cells2-treated MCF-7+p23 cells, it is evident that E2
sion could not fully reverse the activation of theseelicited by p23 overexpression. Interestingly, treat-
ERE aof exo

miquantitative (B) and quantitative real-time PCR (C). Data are averages from threeive fold-enrichment of histone H3-AcK9/K14 at corresponding promoters of MCF+



with TAM increased expression of PMP22 and ABCC3h control and MCF-7+p23 cells (Fig. 3A and B, right).p23 overexpression furthers induces the expression ofgenes in the presence of TAM, whereas activationP22 and ABCC3 by p23 overexpression is partiallyed by E2.ent data suggest that repression by ER requires E2-dent recruitment of the receptor to the EREs of thegene (41). Therefore, we examined ER recruitment to2 and ABCC3 EREs predicted from ChIP-on-Chip datas well as to regions upstream of PMP22 and ABCC3 pro-s (UPS) that were not predicted to bind ER (Fig. 3C).F-7+p23 cells, E2-dependent ER recruitment to

2 and ABCC3 EREs is enriched relative to control cellsD, compare black with dark grey bars, and Supplemen-ig. S5, compare differences between columns 14 and 6e to respective inputs), but not to the PMP22 or ABCC3egions (Fig. 3D and Supplementary Fig. S5). In fact, noding could be detected above nonspecific IgG levels at2 and ABCC3 UPS regions (Supplementary Fig. S5), indi-both as suitable negative control regions for ER bind-hus, p23 overexpression increases E2-dependent ERtment specifically to the EREs of the E2-repressed genes2 and ABCC3. This trend of increased ER recruitment23 overexpression to PMP22 and ABCC3 EREs is similarat has been observed previously at the pS2 ERE (10).ver, for pS2, ER recruitment is associated with genetion, rather than repression.restingly, ER recruitment was evident at the PMP22
nd more so at the ABCC3 ERE, even in the absencegenously added E2 in MCF-7+p23 cells compared with
2. Increased histone H3 acetylation at genes upregulated by p23 overexpression. A, schematic of PMP22 and ABCC3 loci. Arrows, transcriptional startSS); white boxes, exons; black bars, DNA promoter regions amplified by PCR to determine the level of histone H3-AcK9/K14. ChIP assays wereed using specific antibodies against histone H3-AcK9/14 or negative control IgG. Relative levels of histone H3-AcK9/K14 were determined using

independent experiments, normalized to input, and presentedp23 over MCF-7 control cells, set to 1. Error bars, SD.

Cancer Research



controand Scolumthat EABCClevelsPMP22sion ucells aabsencinductcells umodeThereunder

UpregselectBec

(Fig. 1pressichemorubicicells wmothestain fapeuticells iand C

FigureEREs isof PMPexpressbars) anbars) w3 days,1 nmol100 nmand mRquantitare meexperimFig. 1Cand ABtranscrER binddetermanalysiscontrolblack bamplifiethe leveD, the rbindingabsencby ChIPreal-timthe avereplicatto inpuexpressrecruitmUPS into 1. E,RNAi usiRNAsimmunoF, ABCexpressER werreal-timaverageindepennormalipresent(RFC) bsiContrwith siC


www.a


l cells (Fig. 3D, compare light grey with white bars,upplementary Fig. S5, compare differences betweenns 10 and 2 relative to respective inputs). This suggestsR may also function as a repressor of PMP22 and3 gene expression under basal conditions when E2are very low. To assess whether ER represses “basal”and ABCC3 gene expression, we silenced ER expres-sing siRNA in both MCF-7 control and MCF-7+p23nd measured ABCC3 and PMP22 mRNA levels in thee of exogenous E2. Loss of ER resulted in a significantion of ABCC3 in both MCF-7 control and MCF-7+p23nder such basal conditions (Fig. 3E and F). Only ast effect was observed on PMP22 gene expression.

ontrol cells set to 1.

fore, ER represses PMP22 and ABCC3 gene expressionbasal and high E2 situations.

were na chem

acrjournals.org


ulation of ABCC3 in MCF-7+p23 cells impartsive chemotherapeutic resistanceause MCF-7+p23 cells express more ABCC3 mRNAB–D) compared with control cells, and ABCC3 overex-on confers multidrug resistance by exporting from cellstherapeutic agents including etoposide (43) and doxo-n (Adriamycin; ref. 44), we suspected that MCF-7+p23ould be more resistant than control cells to these che-rapeutic agents. Using a sulphorhodamine B dye toor surviving cells following treatment with chemother-c agents, MCF-7+p23 cells survived better than controln the presence of etoposide and doxorubicin (Fig. 4A). This resistance was selective in that MCF-7+p23 cells

3. ER recruitment to distalassociated with repression22 and ABCC3 geneion. MCF-7 control (whited MCF-7+p23 cells (blackere hormone-starved fortreated for 16 hours with

/L E2 (A and B, left) orol/L TAM (A and B, right),NA was analyzed byative real-time PCR. Dataans from six independentents and presented as in. C, schematic of PMP22CC3 loci containingiption start sites (arrows),ing elements asined by global ChIP-Chip(ERE, striped boxes), andupstream regions (UPS,oxes). Black bars, regionsd by PCR to determinel of ER recruitment.elative amount of ER, in the presence ore of E2, was determined, followed by quantitativee PCR. Data are shown asrages, with SE betweene experiments, normalizedt for each sample, anded relative to ERent at the correspondingMCF-7 control cells, setdepletion of ER protein bysing ER (siER) and control(siControl) verified byblotting. WB, Western blot.C3 and PMP22 geneion levels upon depletion ofe obtained by quantitativee PCR. Shown are thes obtained from threedent experimentszed to GAPDH anded as relative fold-changeetween siER (+) andol (-) transfected cells,

o more resistant than control cells to camptothecin,otherapeutic agent not exported by ABCC3 (Fig. 4B;




ref. 43camptnot shrubicimetasexpresof MCthat MrubiciindicaABCC3otherdoxoru

Highfree sBec

highersuspecwouldlyzedfor in

on a s(SuppUsin

tween(T), noWe fowith0.37),(ρ = 0(ρ = 0Hig

stageate p2Tableexpreonly 4cers atent wmigraWe

Figurewere trof drugexpresstransfecis showand siC

Simpson et al.

Cance8452


). Resistance to etoposide and doxorubicin, but notothecin, was validated using clonogenic assays (dataown). To assess the requirement for ABCC3 in doxo-n resistance, because it is commonly used to treattatic and recurrent breast cancers, we silenced thesion of ABCC3 using siRNA and measured sensitivityF-7+p23 cells to doxorubicin. Our analysis revealedCF-7+p23 cells are nearly 50% more sensitive to doxo-n when ABCC3 expression is reduced by 80% (Fig. 4D),ting that doxorubicin resistance is mediated in part by. Lack of complete sensitization can be attributed tomultidrug resistance proteins capable of exportingbicin (43).

p23 expression correlates with shorter disease-urvival times in breast cancer patientsause breast cancer cells that express more p23 have ainvasive potential and are more drug resistant, weted that patients expressing high levels of p23 proteinhave shorter periods of disease-free survival. We ana-
213 specimens from a human breast tissue microarray individ
n (left). Data from quadruplicate points were averaged from two independent exontrols were measured by quantitative real-time PCR (right).



cale of 0 to 3, with 3 as the highest intensity staininglementary Fig. S6).g univariate analysis we determined associations be-cytoplasmic p23, nuclear p23, stage, and tumor sizedal involvement (N), and metastasis (M; TNM status).und that cytoplasmic p23 is significantly associatednuclear p23 expression (Spearman correlation ρ =as well as with stage (ρ = 0.23) and nodal involvement.26), but does not correlate significantly with tumor size.16) or metastasis (ρ = 0.12; Supplementary Table S4).h cytoplasmic p23 (3+) is also associated with later-tumors (68% are stage III/IV) compared with moder-3 (2+) expression (36% are stage III/IV; χ2, P < 0.01;1). Interestingly 73% of the high cytoplasmic p23–ssing tumors displayed nodal involvement, whereas1% of the moderate cytoplasmic p23–expressing can-re node positive (χ2, P < 0.01; Table 1). This is consis-ith our observation that MCF-7+p23 cells are moretory and invasive than control MCF-7 cells (10).used Cox proportional hazards models for each variable
ually to examine the association of that variable with
tensity of p23 staining in the nucleus and cytoplasm recurrence or death (Table 2). High levels of cytoplasmic p23,

4. Increased ABCC3 gene expression in MCF-7+p23 cells mediates doxorubicin resistance. A, MCF-7 control (blue) and MCF-7+p23 cells (pink) cellseated with the indicated concentrations of etoposide, camptothecin (Cpt; B); or doxorubicin (C). Percent survival upon increasing concentrationss was determined using the sulphorhodamine B dye, which coats cells evenly so optical density (OD) values are proportional to cell number. Data areed as averages of at least three independent experiments with untreated cells set at 100% survival. Error bars, SE. D, MCF-7+p23 cells wereted with siRNA against ABCC3 (siABCC3) or a control siRNA (siControl), and percent survival of doxorubicin treated relative to untreated cells

periments; error bars, SE. ABCC3 mRNA levels from siABCC3

Cancer Research



stage Iassocivariabplasmrecurrthat inp23 (3likelypatienTableand cyor sta(stagehave d(Tablediseasis estiindivip23low

Althoudeathp23 excantlymary,III/IV disease are more likely to exhibit lymph nodemetastasesand ex

Discu

Weoverexvancetivitiewhichto ant

Tab

Sam

AgeM

Nuc0123

StagIIIIIIV

T1234M

N0123M

M01

*t-te†Fis‡χ2


www.a


II/IV disease, and high T, N, and M were all individuallyated with increased recurrence and death. When theseles were considered jointly (multivariable), only cyto-ic p23 expression and stage were associated with diseaseence or death (from Cox models). Our findings indicatedividuals with stage I/II disease and high cytoplasmic+) expression (stage I/II-p23high) are 2.35 times moreto have recurrent disease or death than the same staget with low p23 (1+ or 2+) expression (stage I/II-p23low;2 and Fig. 5). Moreover, patients with stage III/IV diseasetoplasmic p23 1+ or 2+ expression (stage III/IV-p23low),ge III/IV disease and cytoplasmic p23 3+ expressionIII/IV-p23high) are 3.34 and 7.85 times more likely toisease recurrence than those stage I/II-p23low individuals2 and Fig. 5). At 150months postsurgery, the cumulativee-free survival for patients with stage I/II-p23low tumorsmated at 80%; the corresponding estimates are 50% of
duals with stage I/II-p23high, 44% with stage III/IV-, and 18% with stage III/IV-p23high tumors (Fig. 5).
ABCCHsp90

le ion in re

p

, yea 46.0 ± 50.91lea

2 (10 4 (1+++ 0 1 (5e

0 3 (1

I0

0 3 (11 (50 9 (4

is

1 (50 13 (51 (50 6 (2

is

test.

acrjournals.org


gh stage alone is highly associated with recurrence or(likelihood ratio = 45.77), the addition of cytoplasmicpression to the Cox proportional hazard models signifi-improves the model fit (likelihood ratio = 52.15). In sum-individuals with high levels of cytoplasmic p23 and stage

perience disease recurrence and mortality.

ssion

have determined that many genes affected by p23pression in MCF-7 cells are also misregulated in ad-d breast cancers. These genes display varying sensi-s to treatments with Hsp90 inhibitors, TAM, or E2,could play a role in the responsiveness of tumor cellsiestrogen treatment or chemotherapy. For example,
3 is upregulated by p23 overexpression alone, isdependent, is further upregulated by TAM treatment,
oarray
1. p23 express lation to stage and TNM from the breast tumor micr
55.12

2 (

85 (

5 (

21 (

7 (54 (

1

50 (46 (

Cancer Res; 70(21)

20. © 2010 American Association

les (n = 213) Cytoplasm p23 P1+ (n = 2) 2+ (n = 22) 3+ (n = 189) (Tests 1+ & 2+ vs. 3+)

ears
n ± SD 42.24 ± 10.27 ± 13.21 0.10* r p23
<0.0001†
0%) 8%) 1%) 5%) 17%) 0 10 (4 33 (
0 7 (32%) 69 (37%)
%) 45%)
<0.01†
4%) 3%) %) 0%) 29%) 1 (50 11 (5 55 (
1 (50%) 8 (36%) 108 (57%)
0 11%)
0.09†
4%) 4%) %) 3%) 31%) %) 4%) 17%) 1 (50 3 (1 30 (0 6 (29%) 81 (47%)
sing
1 7
<0.01‡
%) 9%) 27%) %) 7%) 24%)
4%)
28%) 0 3 (1 52 (0 0 40 (21%)
sing
1
2 (100%) 22 (100%) 168 (89%) 0.14†

0 0 21 (11%)

st.her's exact test.

November 1, 2010 8453

for Cancer


and isconferMCF-7observic depsensitshownstudiepeuticnegatiHow

been ssteroidindicaE2-repER-actscriptefficienWe

are prand Cthe prcells wp23 massocision. T

expre(HATsratherp300;activeInd

and thyeast,waysK9/14furthetion atwo mwhichcanceand penzymsensitbindinregulaulatetransc iption from active loci to increase gene expression.Furth , amplification of gene regulation in situations of

Table 2. Disease recurre tage using

Age 1.0 0.16Cytop

1+,3+ 2.9 <0.01 2.3

Nuclea0, 12+, 1.3

StageI, IIIII, I 3.4 3.3

T1234 17.3 <0.01

N0123 4.0 <0.000

M01

Abb

Simpson et al.

Cance8454


downregulated by E2 treatment. Functionally, ABCC3s selective resistance to chemotherapeutic agents in+p23 cells when compared with control cells, and weed preliminarily that siRNA-mediated or pharmacolog-letion of ER, using ICI 182,720 treatment, decreases theivity of MCF-7 control cells to doxorubicin (data not). Thus, both gene expression data and functionals show that p23 via ABCC3 may regulate chemothera-resistance acquired in metastatic and recurrent ER-ve breast tumors.might p23 selectively affect gene expression? p23 has

uggested to play a role in the binding and clearance ofreceptors at response elements (8–10). Our results alsote that p23 increases recruitment of ER to EREs ofressed genes ABCC3 and PMP22 (Fig. 3) as well as toivated genes pS2 and cathepsin D (10). Thus, certain tran-ion factors, like steroid receptors, may utilize p23 totly cycle on and off DNA tomodulate gene transcription.find that “active” histone modifications (H3-AcK9/14)esent at the promoters of p23-sensitive genes (Fig. 2B). The fact that these modifications are also present atomoters of some p23-sensitive genes in MCF-7 controlhere p23 is not overexpressed (Fig. 2B) indicates thatay not need to overcome the transcriptional barriers

reviation: 95% CI, 95% confidence interval.

ated with gene silencing to upregulate gene expres-herefore, a mechanism p23 may use to increase gene 9 E. Oxe



ssion could be to recruit histone acetyltransferases) implicated in transcription reinitiation (i.e., GCN5),than those involved in transcription initiation (i.e.,

ref. 45) to raise the transcriptional output from alreadypromoters.eed, we observed a functional interaction between p23e HAT GCN5 by conducting a synthetic lethal screen insuggesting that p23 affects the activity of HAT path-in an overlapping fashion with GCN5.9 Histone H3-is a target for acetylation by GCN5 (39, 46, 47), whichr implicates GCN5 as a HAT-mediating hyperaceyla-t the promoters of p23-upregulated genes. Interestingly,etabolic genes regulated by p23, ACSL1 and KYNU,are upregulated in metastatic and recurrent breastrs, are also known to regulate the activity of HATsoly(ADP-ribose) polymerase chromatin remodelinges (39, 46, 47). Therefore, a plausible model for p23-ive gene regulation is that p23 may influence the DNAg dynamics of specific transcription factors to up-te or downregulate gene expression. p23 may also mod-the recruitment of coactivators necessary to reinitiate
rer
nce as function of p23 expression and s

lmark and M.J. Ga

20. © 2010 A

a Cox proportional hazard model

rabedian, unpublished.

Cancer Resea

merican Association for Cancer

Univariate

d ratio (95% CI)

Multivariable

Hazar P Hazard ratio (95% CI) P

1 (0.99-1.02)
lasm p23 2+ 1.00 <0.01 1.00 0.02
7 (1.46-6.06)
5 (1.15-4.82) r p23 + 1.00 0.17 3+ 3 (0.88-2.01)
1.00 <0.0001 1.00 <0.0001
V 7 (2.34-5.15) 4 (2.22-5.03)
1.00
0 (0.97-51.75) 7.1 0.05
8.04 (1.08-59.81) 0.04
8 (2.41-125.20)
1.00
2 (0.91-2.53) 1.5 0.11
6.81 (4.25-10.90) <0.0001
4 (2.43-6.72) 1
1.00 <0.000111.41 (6.76-19.27)

rch


p23 ovp23-seof chrrecruip23

cells (correlcanceresis bfact, wand anucleuing threctly,(Supp

survivpressiwe prsignalp23-seOur

canceto exhrecurrmodephenoas brethe mimporlymphalso mposidein ThChinep23 btargetin bre

Discl

No p

Ackn

We tthe man

Grant

VilceDOD CeBreastFoundadPhilip

Theof pageaccorda

Refe1. Joh

23-Mo

2. JohtorCh

3. Mcarr20

4. BoantMo

5. Frechaent

FigureKaplansurvivap23 (restage IIand higmultivapatientslevels (


www.a


erexpression may occur through the enhancement ofnsitive metabolic pathways that affect the activityomatin-modifying enzymes. Whether or not p23 isting these factors directly is still under investigation.localizes to both the nucleus and cytoplasm in MCF-7Supplementary Fig. S7), and high cytoplasmic p23ates with shorter disease-free survival times for breastpatients, indicating that p23 can promote tumorigen-

y functioning in multiple cellular compartments. Ine observe increased activated Akt1 in the cytoplasmn expanded phosphoproteome in the cytoplasm ands of MCF-7+p23 compared with control cells, suggest-at p23 overexpression facilitates, either directly or indi-

classified according to stage (I/II or III/IV) and p23 expressionlow, 1+ or 2+; high, 3+).

enhanced kinase activity and protein phosphorylation Rece

eman BC, Felts SJ, Toft DO, Yamamoto KR. The p23 molecularperones act at a late step in intracellular receptor action to differ-ially affect ligand efficacies. Genes Dev 2000;14:422–34.

6. Hotra

7. Gress

8. Freco

9. StagluscrBio

10. OxGageinv

11. To

acrjournals.org


al signal, our findings also suggest that tumors overex-ng p23 would have a survival advantage (48). Therefore,opose that cancers with high levels of p23 activateing pathways in the cytoplasm, which can also amplifynsitive gene transcription in the nucleus.findings strongly suggest that individuals with breast

rs that display high p23 protein levels are more likelyibit lymph node involvement and experience diseaseence and mortality. This is consistent with our in vitrol that p23 overexpression elicits gene expression andtypic changes in MCF-7 cells akin to those that occurast tumors become invasive. Given that nodal status isost significant factor in predicting survival (49), it will betant to find ways to reduce p23 expression to decreasenode metastases. Lowering p23 expression shoulditigate p23-mediated chemotherapeutic resistance (eto-, doxorubicin). Recently, celastrol, the active compoundunder of God Vine root extracts used in traditionalse medicine, has been shown to bind and inactivatey triggering its oligomerization into fibers (50). Thus,ing p23 should reduce invasiveness and drug resistanceast cancer cells and prevent disease recurrence.

osure of Potential Conflicts of Interest

otential conflicts of interest were disclosed.

owledgments

hank Naoko Tanese, Keren Imberg, and ChaoweiWu for critically readinguscript, and Diego Acosta-Alvear for guidance with the ChIP assays.

Support

k Foundation (N.E. Simpson), NIH/NCI P30 CA16087 (J.D. Goldberg),nter of Excellence (W81XWH-04-1-0905; R.J. Schneider/J.D. Goldberg),Cancer Research Foundation (R.J. Schneider/J.D. Goldberg), Avontion (R.J. Schneider), NIH/NCRR M01 RR000096 (R.J. Schneider), an-Morris USA (M.J. Garabedian).costs of publication of this article were defrayed in part by the paymentcharges. This article must therefore be hereby marked advertisement innce with 18 U.S.C. Section 1734 solely to indicate this fact.

ived 05/03/2010; revised 08/12/2010; accepted 08/26/2010; published
lementary Fig. S8). Given tha t activated Akt is a potent OnlineFirst 09/16/2010.
rencesnson JL, Beito TG, Krco CJ, Toft DO. Characterization of a novelkilodalton protein of unactive progesterone receptor complexes.l Cell Biol 1994;14:1956–63.nson JL, Toft DO. A novel chaperone complex for steroid recep-s involving heat shock proteins, immunophilins, and p23. J Biolem 1994;269:24989–93.Laughlin SH, Sobott F, Yao ZP, et al. The co-chaperone p23ests the Hsp90 ATPase cycle to trap client proteins. J Mol Biol06;356:746–58.hen SP. Genetic and biochemical analysis of p23 and ansamycinibiotics in the function of Hsp90-dependent signaling proteins.l Cell Biol 1998;18:3330–9.

lley SJ, Yamamoto KR. A role for Hsp90 in retinoid receptor signalnsduction. Mol Biol Cell 1995;6:1833–42.ad I, McKee TA, Ludwig SM, et al. The Hsp90 cochaperone p23 isential for perinatal survival. Mol Cell Biol 2006;26:8976–83.eman BC, Yamamoto KR. Disassembly of transcriptional regulatorymplexes by molecular chaperones. Science 2002;296:2232–5.vreva DA, Muller WG, Hager GL, Smith CL, McNally JG. Rapidcocorticoid receptor exchange at a promoter is coupled to tran-iption and regulated by chaperones and proteasomes. Mol Celll 2004;24:2682–97.elmark E, Roth JM, Brooks PC, Braunstein SE, Schneider RJ,rabedian MJ. The cochaperone p23 differentially regulates estro-

5. p23 expression and clinical outcome in human breast cancer.-Meier curves plotting time to recurrence (in months) versusl probability are shown for patients with stage I/II disease and lowd solid line), stage I/II disease and high p23 (blue dotted line),I/IV disease and low p23 (green dotted line), or stage III/IV diseaseh p23 (purple dotted line). Hazard ratios were calculated usingriable Cox proportional hazards models and are shown for

n receptor target genes and promotes tumor cell adhesion andasion. Mol Cell Biol 2006;26:5205–13.ogun OA, Zeiger W, Freeman BC. The p23 molecular chaperone




protion

12. DoboPK

13. Scanele20

14. Yaexpcultra

15. Fop2intr

16. Boass

17. Tanideallybio

18. Lococgla

19. Knon19:

20. OxGesur365

21. Kreproph

22. Mocolink

23. HoHs

24. Foastel

25. HaWeele

26. StYaindinU

27. Hoangro12

28. GoThbreeff

29. ActypMo

30. HoLe

toed

31. Hode20

32. KocMtidce

33. LiuHuind37

34. Smpro

35. NieExhu

36. JiaFoits11

37. Abhuof

38. ZoVaan

39. Mica

40. Mapein80

41. Ellthrpro

42. Cage

43. Krumarec20

44. Liua dco

45. Meordna

46. PitMUresAc

47. Taacglo

48. LiuFro

49. Shprotor

Simpson et al.

Cance8456


motes functional telomerase complexes through DNA dissocia-. Proc Natl Acad Sci U S A 2007;104:5765–70.nze O, Abbas-Terki T, Picard D. The Hsp90 chaperone complex isth a facilitator and a repressor of the dsRNA-dependent kinaseR. EMBO J 2001;20:3771–80.hmidt MH, Chen B, Randazzo LM, Bogler O. SETA/CIN85/Rukd its binding partner AIP1 associate with diverse cytoskeletalments, including FAKs, and modulate cell adhesion. J Cell Sci03;116:2845–55.o Q, Nishiuchi R, Li Q, Kumar AR, Hudson WA, Kersey JH. FLT3ressing leukemias are selectively sensitive to inhibitors of the mole-ar chaperone heat shock protein 90 through destabilization of signalnsduction-associated kinases. Clin Cancer Res 2003;9:4483–93.rafonov F, Toogun OA, Grad I, Suslova E, Freeman BC, Picard D.3/Sba1p protects against Hsp90 inhibitors independently of itsinsic chaperone activity. Mol Cell Biol 2008;28:3446–56.se S, Weikl T, Bugl H, Buchner J. Chaperone function of Hsp90-ociated proteins. Science 1996;274:1715–7.ioka T, Nakatani Y, Semmyo N, Murakami M, Kudo I. Molecularntification of cytosolic prostaglandin E2 synthase that is function-coupled with cyclooxygenase-1 in immediate prostaglandin E2synthesis. J Biol Chem 2000;275:32775–82.vgren AK, Kovarova M, Koller BH. cPGES/p23 is required for glu-orticoid receptor function and embryonic growth but not prosta-ndin E2 synthesis. Mol Cell Biol 2007;27:4416–30.oblauch R, Garabedian MJ. Role for Hsp90-associated cochaper-e p23 in estrogen receptor signal transduction. Mol Cell Biol 1999;3748–59.elmark E, KnoblauchR, Arnal S, Su LF, SchapiraM,GarabedianMJ.netic dissection of p23, an Hsp90 cochaperone, reveals a distinctface involved in estrogen receptor signaling. J Biol Chem2003;278:47–55.bs J, Saremaslani P, Caduff R. ALG-2: a Ca2+ -binding modulatortein involved in cell proliferation and in cell death. Biochim Bio-ys Acta 2002;1600:68–73.llerup J, Krogh TN, Nielsen PF, Berchtold MW. Properties of the-chaperone protein p23 erroneously attributed to ALG-2 (apoptosis-ed gene 2). FEBS Lett 2003;555:478–82.lt SE, Aisner DL, Baur J, et al. Functional requirement of p23 andp90 in telomerase complexes. Genes Dev 1999;13:817–26.rsythe HL, Jarvis JL, Turner JW, Elmore LW, Holt SE. Stablesociation of hsp90 and p23, but not hsp70, with active humanomerase. J Biol Chem 2001;276:15571–4.hn WC, Counter CM, Lundberg AS, Beijersbergen RL, Brooks MW,inberg RA. Creation of human tumour cells with defined geneticments. Nature 1999;400:464–8.ampfer MR, Garbe J, Levine G, Lichtsteiner S, Vasserot AP,swen P. Expression of the telomerase catalytic subunit, hTERT,uces resistance to transforming growth factor β growth inhibitionp16INK4A(-) human mammary epithelial cells. Proc Natl Acad SciS A 2001;98:4498–503.chreiter AE, Xiao H, Goldblatt EM, et al. Telomerase templatetagonist GRN163L disrupts telomere maintenance, tumorwth, and metastasis of breast cancer. Clin Cancer Res 2006;:3184–92.ldblatt EM, Gentry ER, Fox MJ, Gryaznov SM, Shen C, Herbert BS.e telomerase template antagonist GRN163L alters MDA-MB-231ast cancer cell morphology, inhibits growth, and augments theects of paclitaxel. Mol Cancer Ther 2009;8:2027–35.osta-Alvear D, Zhou Y, Blais A, et al. XBP1 controls diverse celle- and condition-specific transcriptional regulatory networks.
l Cell 2007;27:53–66.ughton P, Fang R, Techatanawat I, Steventon G, Hylands PJ,e CC. The sulphorhodamine (SRB) assay and other approaches
50. Chingp2



testing plant extracts and derived compounds for activities relat-to reputed anticancer activity. Methods 2007;42:377–87.ulden H, Reilly MM. Molecular genetics of autosomal-dominantmyelinating Charcot-Marie-Tooth disease. Neuromolecular Med06;8:43–62.ol M, de Haas M, Scheffer GL, et al. Analysis of expression ofOAT (MRP2), MRP3, MRP4, and MRP5, homologues of the mul-rug resistance-associated protein gene (MRP1), in human cancerll lines. Cancer Res 1997;57:3537–47.D, Rudland PS, Sibson DR, Platt-Higgins A, Barraclough R.man homologue of cement gland protein, a novel metastasisucer associated with breast carcinomas. Cancer Res 2005;65:96–805.irnov DA, Zweitzig DR, Foulk BW, et al. Global gene expressionfiling of circulating tumor cells. Cancer Res 2005;65:4993–7.s AT, Konig J, Pfannschmidt M, Klar E, Hofmann WJ, Keppler D.pression of the multidrug resistance proteins MRP2 and MRP3 inman hepatocellular carcinoma. Int J Cancer 2001;94:492–9.ng WG, Watkins G, Douglas-Jones A, Mokbel K, Mansel RE,dstad O. Expression of Com-1/P8 in human breast cancer andrelevance to clinical outcome and ER status. Int J Cancer 2005;7:730–7.ba MC, Drake JA, Hawkins KA, et al. Transcriptomic changes inman breast cancer progression as determined by serial analysisgene expression. Breast Cancer Res 2004;6:R499–513.rn AM, Barish GD, Williams BO, Lavender P, Klymkowsky MW,rmus HE. Regulation of Wnt signaling by Sox proteins: XSox17 α/βd XSox3 physically interact with β-catenin. Mol Cell 1999;4:487–98.nn AJ, Gupta GP, Siegel PM, et al. Genes that mediate breastncer metastasis to lung. Nature 2005;436:518–24.ier M, Castagner F, Berger P, Suter U. Distinct elements of theripheral myelin protein 22 (PMP22) promoter regulate expressionSchwann cells and sensory neurons. Mol Cell Neurosci 2003;24:3–17.ison-Zelski SJ, Solodin NM, Alarid ET. Repression of ESR1ough actions of estrogen receptor α and Sin3A at the proximalmoter. Mol Cell Biol 2009;29:4949–58.rroll JS, Meyer CA, Song J, et al. Genome-wide analysis of estro-n receptor binding sites. Nat Genet 2006;38:1289–97.h GD, Belinsky MG, Gallo JM, Lee K. Physiological and phar-cological functions of Mrp2, Mrp3 and Mrp4 as determined froment studies on gene-disrupted mice. Cancer Metastasis Rev07;26:5–14.Y, Peng H, Zhang JT. Expression profiling of ABC transporters inrug-resistant breast cancer cell line using AmpArray. Mol Pharma-l 2005;68:430–8.tivier R, Penot G, Hubner MR, et al. Estrogen receptor-α directsered, cyclical, and combinatorial recruitment of cofactors on atural target promoter. Cell 2003;115:751–63.roda SP, Khodarev NN, Beckett MA, Kufe DW, Weichselbaum RR.C1-induced alterations in a lipid metabolic gene network predictponse of human breast cancers to tamoxifen treatment. Proc Natlad Sci U S A 2009;106:5837–41.kahashi H, McCaffery JM, Irizarry RA, Boeke JD. Nucleocytosolicetyl-coenzyme a synthetase is required for histone acetylation andbal transcription. Mol Cell 2006;23:207–17.W, Bagaitkar J, Watabe K. Roles of AKT signal in breast cancer.nt Biosci 2007;12:4011–9.ek LL, Godolphin W. Model for breast cancer survival: relativegnostic roles of axillary nodal status, TNM stage, estrogen recep-concentration, and tumor necrosis. Cancer Res 1988;48:5565–9.adli A, Felts SJ, Wang Q, et al. Celastrol inhibits Hsp90 chaperon-
of steroid receptors by inducing fibrillization of the Co-chaperone3. J Biol Chem 2010;285:4224–31.
Cancer Research



2010;70:8446-8456. Published OnlineFirst September 16, 2010.Cancer Res Natalie E. Simpson, W. Marcus Lambert, Renecia Watkins, et al. Increasing Lymph Node Metastases and Drug ResistanceProgression and Poor Prognosis in Breast Cancer by High Levels of Hsp90 Cochaperone p23 Promote Tumor

Updated version

10.1158/0008-5472.CAN-10-1590doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2010/09/16/0008-5472.CAN-10-1590.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/70/21/8446.full#ref-list-1

This article cites 50 articles, 32 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/70/21/8446.full#related-urls

This article has been cited by 4 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/70/21/8446To request permission to re-use all or part of this article, use this link

Research. on January 29, 2020. © 2010 American Association for Cancercancerres.aacrjournals.org Downloaded from


http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-10-1590

http://cancerres.aacrjournals.org/content/suppl/2010/09/16/0008-5472.CAN-10-1590.DC1

http://cancerres.aacrjournals.org/content/70/21/8446.full#ref-list-1

http://cancerres.aacrjournals.org/content/70/21/8446.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/70/21/8446


Cancer Research High Levels of Hsp90 Cochaperone p23...

Documents

Transcript of Cancer Research High Levels of Hsp90 Cochaperone p23...