Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820

Contents lists available at ScienceDirect

Journal of Steroid Biochemistry and Molecular Biology

journa l homepage: www.e lsev ier .com

Gene network signaling in hormone responsivenautophagy in breast cancer cells

Robert C a, YJianhua X etta Department o l of Meb Department o , USAc Department o ity, Arl

a r t i c l

Article history:Received 25 November 2008Accepted 31 December 2008

Keywords:AntiestrogenAutophagyApoptosisBreast cancerCell signalingEndoplasmic rEstrogensGene networkUnfolded prot

Resistance to endocrine therapies, whether de novo or acquired, remains a major limitation in the abil-ity to cure many tumors that express detectable levels of the estrogen receptor alpha protein (ER).While several resistance phenotypes have been described, endocrine unresponsiveness in the context oftherapy-induced tumor growth appears to be the most prevalent. The signaling that regulates endocrine

1. Introdu

Over 40,incidence issidered as awomen wiU.S., almos(ER+; HUGsporadic br

Lecture prBiochemistry a

CorrespoOncology, Geoington, DC 200

E-mail add

0960-0760/$ doi:10.1016/j.jeticulum

sein response

resistant phenotypes is poorly understood but it involves a complex signaling network with a topologythat includes redundant anddegenerative features. To be relevant to clinical outcomes, themost pertinentfeatures of this network are those that ultimately affect the endocrine-regulated components of the cellfate and cell proliferationmachineries.We showthat autophagy, as supportedby the endocrine regulationofmonodansylcadaverine staining, increased LC3 cleavage, and reduced expression of p62/SQSTM1, playsan important role in breast cancer cells responding to endocrine therapy. We further show that the cellfate machinery includes both apoptotic and autophagic functions that are potentially regulated throughintegrated signaling that ows through key members of the BCL2 gene family and beclin-1 (BECN1). Thissignaling links cellular functions in mitochondria and endoplasmic reticulum, the latter as a consequenceof induction of the unfolded protein response. We have taken a seed-gene approach to begin extractingcritical nodes and edges that represent central signaling events in the endocrine regulation of apoptosisandautophagy. Three seednodeswere identied fromglobal geneorprotein expression analyses and sup-portedby subsequent functional studies that established their abilities to affect cell fate. The seednodes ofnuclear factor kappa B (NFB), interferon regulatory factor-1 (IRF1), and X-box binding protein-1 (XBP1)are linked by directional edges that support signal ow through a preliminary network that is grownto include key regulators of their individual function: NEMO/IKK, nucleophosmin and ER respectively.Signaling proceeds through BCL2 gene family members and BECN1 ultimately to regulate cell fate.

2009 Elsevier Ltd. All rights reserved.

ction

000Americanwomendie of breast cancer each year [1];broadly similar across the European Union when con-percentage of the population. In 2008, over 178,000

ll be diagnosed with invasive breast cancer in thet 70% of which will be estrogen receptor- positiveO Gene Symbol = ESR1) [2,3]. The percentage of ER+east cancers increases linearly with age but even in pre-

esented at the 18th International Symposiumof the Journal of SteroidndMolecular Biology, 1821 September 2008, Seefeld, Tyrol, Austria.

nding author at: Room W405A Research Building, Department ofrgetownUniversity School ofMedicine, 3970Reservoir Rd,NW,Wash-57, USA. Tel.: +1 202 687 3755; fax: +1 202 687 7505.ress: clarker@georgetown.edu (R. Clarke).

menopausal cases the proportion is high; 62% at age 35 and 72%by age 49 [24]. Data from randomized trials and meta-analysesclearly show that all breast cancer patients derive a statisticallysignicant survival benet from adjuvant chemotherapy, and thatall hormone receptor positive breast cancer patients benet fromadjuvant endocrine therapy [59]. For postmenopausalwomen, thebenet from adjuvant Tamoxifen (TAM) is comparable to that seenfor cytotoxic chemotherapy. While 5 years of adjuvant TAM pro-duces a 26% proportional reduction in mortality [8], many ER+tumors eventually recur [10]. Since advanced ER+ breast cancerlargely remains an incurable disease, resistance to endocrine ther-apies is a signicant clinical problem.

Endocrine therapy is administered as an antiestrogen (AE) likeTamoxifen (TAM) or Fulvestrant (FAS; Faslodex; ICI 182,780), or asan aromatase inhibitor (AI) such as Letrozole or Exemestane. It isless toxic andpotentiallymoreeffective therapy in themanagementof hormone-dependent breast cancers. Antiestrogens, and TAM in

see front matter 2009 Elsevier Ltd. All rights reserved.sbmb.2008.12.023larkea,b,, Ayesha N. Shajahana, Rebecca B. Rigginsuanc, Yue Wangc, Alan Zwarta, Ruchi Nehraa, Min

f Oncology and Lombardi Comprehensive Cancer Center, Georgetown University Schoof Physiology & Biophysics, Georgetown University School of Medicine, Washington, DCf Electrical and Computer Engineering, Virginia Polytechnic Institute and State Univers

e i n f o a b s t r a c t/ locate / j sbmb

ess modies apoptosis and

ounsook Choa, Anatasha Crawforda,a C. Liua

dicine, Washington, DC, USA

ington, VA, USA

R. Clarke et al. / Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820 9

particular, have been the gold standard rst line endocrine ther-apy for over 30 years [11], clinical experience with this drug likelyexceeding over 15 million patient years [10]. TAM increases bothdisease free and overall survival from early stage breast cancer, andit also reduccer in high-effective inosteoporosiantiestrogeand some Aas viable altmenopausaare generallsingle agenandpostmeand all of th

Of curreschedulingin disease fan AE usuaHowever, thoverall surv[15]. In termrates withTAM [21,22options, theing of agentway these cAIs andAEsbreast cancto cure man

1.1. Endocri

Severalmental modphenotypeshave been freceptor fodegree of tring to receER+/PGR+,metastatic bin truly ERrelevance, amay be verEarly Breasshow that Tthe 10-yearof death froreported [8to this clinicbecome routherapy basER tumorsand/or PR+ever, it remawill receivetion of anyendocrine r

1.2. Endocr

Severalexperimentin vitro or o

phenotypes include (i) estrogen-independent (which appearsequivalent to AI resistance but is not so for antiestrogen resistance[23]some breast cancers can become resistant to an AE but stillrespond to an AI and vice versa); (ii) estrogen-inhibited (recently

ed ist inve [2ed inps thlly invariape o

inicace p

aininarmauslyawalt ingicanumre, Tn TApreduse

-estra], sudeedse stR+ brf ERto esurprl respses).thattingl pheeno

cance numnicalstrostanossrcannura

sivenf n=thorsimatmostolnnicalimpomose eall e

eceptone)plainimateredgicaes the incidence of invasive andnoninvasive breast can-risk women [8,9]. Raloxifene, another antiestrogen, isreducing the rate of postmenopausal bone loss froms aswell as the rate of invasive breast cancer [12]. Newerns such as FAS show signicant activity relative to TAMIs [13,14]. Third generation AIs are now widely acceptedernatives to AEs for rst line endocrine therapy in post-lwomenwithmetastatic disease; overall response ratesy greater for AIs [15]. Importantly, Tamoxifen is the onlyt with demonstrated efcacy in both premenopausalnopausalwomenwith invasive breast cancer. Other AEseAIs require the complete cessation of ovarian function.nt interest is identication of the optimum choice andof AEs and AIs. Evidence clearly shows improvementsree survival for combined adjuvant therapy (an AI andlly given sequentially) over single agent TAM [1620].e ability of AIs to induce a signicant improvement inival compared with 5 years of TAM alone is uncertains ofmetastatic disease, recent data imply that responsean AI are either equivalent with or higher than with]. Given the increasing number of endocrine treatmentre is a clear need to optimize the selection and schedul-s for both early stage and advanced disease. Whicheverontroversies are eventually resolved, it is clear that bothwill remain as keymodalities in themanagement of ER+ers. Unfortunately, the inability of endocrine therapiesy women with ER+ disease will also remain.

ne resistance: receptor phenotypes

resistance phenotypes are evident from both experi-els and clinical observations. The two primary receptorare ER+ and ER. These receptor-based phenotypes

urther stratied by addition of the estrogen-regulatedr progesterone (PGR; HUGO Gene Symbol = PGR). Theeatment benet from endocrine therapy varies accord-ptor phenotype. For example, approximately 75% of33% of ER+/PGR, and 45% of ER/PGR+ cases ofreast cancer respond to TAM [10]. Endocrine responsestumors are probably relatively rare and of uncertain

s they most likely reect incorrect assessments of whaty low ER and/or PGR expression values. Data from thet Cancer Trialists Collaborative Group meta-analysesAM therapy generates a non-signicant 6% reduction inrisk of recurrence. A non-signicant increase in the riskm any cause in patientswith ER breast cancer alsowas,9]. The real value of PGR, which is the only modicational prediction scheme for directing endocrine therapy totine in over 30 years (the value of directing endocrineed on HER2 is still controversial), is largely limited to. It is general practice in theUnited States to treat all ER+invasive breast tumors with endocrine therapy. How-ins impossible to predictwhether an individual patientbenet from treatment and the magnitude or dura-benet. Better predictors of each individual patientsesponsiveness are clearly needed.

ine resistance: pharmacological phenotypes

pharmacological phenotypes have been identied inal models of either human breast cancer cells growingf xenografts in immune-decient rodents [10]. These

identied rsensitiobserv(perhaclinicaOtherthe sco

1.3. Clresistan

Obtthe phpreviowithdrat leasmacoloas thethermobetweeshouldt becaof 17tus [10[24]. Inresponof all Emore oaccess

Inoclinicaresponmatedsuggesclinicatant phbreastnotablwas clicases cal resifully crto AIs,

Nomrespontures othe auapproxand alestraditwo cli

It isgens frHigh dAswithcers (rwere dably exapproxconsidmacolon MCF-7 models [24]); (iii) TAM-stimulated (identi-MCF-7 xenografts [25,26]); TAM-unresponsive but FAS7] (identied rst in MCF-7 models and subsequentlyclinical trials [13]); TAM and FAS crossresistant [28]

is is truly antiestrogen crossresistant and it is seen bothpatients and experimentally in MCF-7 models [13,29]).tions on these phenotypes likely occur but are beyondf our discussion.

l evidence for the prevalence of pharmacologicalhenotypes

g direct clinical evidence for the prevalence of each ofcological resistance phenotypes is challenging.Wehavenoted the utility of applying clinical responses to TAMin metastatic breast cancer as one means to dene,

broad terms, the likely relevance of a series of phar-l phenotypes [29]. This approach is somewhat limited,ber of cases across all studies is modest (n=241). Fur-AM withdrawal responses cannot readily distinguishM-stimulation and estrogen-inhibition because eachict for a clinical benet. The latterwould induce a bene-many breast cancers contain signicant concentrationsdiol, independent of both menopausal and ER/PGR sta-fcient to produce the estrogen-inhibited phenotype, the superiority of AIs over TAM in inducing clinical

rongly implies that over 75% of ER+/PGR+, at least 50%east cancers irrespective of PGR expression, and 45% or/PGR+ breast tumors are probably driven by adequate

trogen.ior assessment, almost 9%ofpatients receivedanoverallonse to TAM withdrawal (partial responses + completeWhen disease stabilizations were included we esti-less than 20% of patients received clinical benet [29],that the sumof TAM-stimulated plus estrogen-inhibitednotypes may not account for the majority of resis-types in women. Of course, given the number of ER+ers arising every year, these phenotypes are relevant to aber of women. The major response to TAM withdrawal

ly detectable disease progression greater than 80% ofngly implicating unresponsiveness as the primary clini-cemechanism to TAM.Whether these breast cancers areesistant to all endocrine therapies, or retain sensitivityot be determined from this simple analysis.et al. [30] took a different approach and assessed theess to estrogen and TAM in short-termprimary cell cul-153 ER+ breast cancer biopsies. This approach allowedto separate the various pharmacological phenotypes;ely 7% of ER+ primary cultures were stimulated by TAM3% were inhibited by physiological concentrations ofotably close to our estimate of 9% for the sum of thesephenotypes.

ortant here to separate responses to physiological estro-those produced by pharmacological estrogen therapy.strogen therapy was used prior to the advent of TAM.ndocrine therapies, approximately30%of all breast can-or status was not available when most of these studiesresponded [31,32]. Side effects were unfavorable, prob-ing the switch to TAM that also induces responses inely 30% of all breast cancers (when receptor status is not). It is also likely that the mechanisms of action of phar-l and physiological dose estrogens differ. Over 15 years

10 R. Clarke et al. / Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820

ago, we were the rst to show that pharmacological concentrationsof both estradiol and TAM induce changes in the membrane uid-ity of breast cancer cells and that this correlates with changes incell growth [33]. It is unlikely that membrane uidity changes aremajor contrphysiologicthe prodeat

2. Cell fate

Therapebalance bet(but ideallyconsistentlyThe relativeunclear. Todeath and tarrest to co

Cell deamitotic catadeath are recleavage) anever, knowevents, andknown to bulated in paof cell death

2.1. Apopto

Apoptoscriteria relamentationformation ocell membrintrinsic (mand antiapchanges inof cytochromof the plasmbrane integis dependefactor- (TNThe intrinscutionerscytoskeletotimeearlyoften descrChanges inprecede chof phosphaearly-to-mimitochondrminal transthe DNA reor necrosis

2.2. Autoph

Autophaare degradesomes, normorganelles,lation of auare now kn

interacting proteins such as beclin-1/ATG6 (BECN1) [42]. BCL2 anti-apoptotic proteins can block autophagy by inhibiting BECN1 [36].Since monoallelic loss of the BECN1 locus is seen in >40% of breastcancers [43] (and in MCF-7 cells), modulating BCL2 may be an

ve magy

atin,n elee oSTMble;ell deupre

ptosiadoxextrarnates l/nutagoson on nocelernden

itotic

lty Dointtiond cellideat caacter

ecros

rosision oasso

ntente orsatio

].

nesc

escees and ansidauccelularcanres a[64

docr

cisely(ort bes DNls reutat], soibutors to the action, either prosurvival or prodeath, ofal estrogen exposures but they likely do contribute toh effects of pharmacological exposures.

in the context of endocrine responsiveness

utic strategies forbreast cancergenerally aimtoalter theween cell death and cell survival such that cancer cellsnot normal cells) die. However, endocrine therapiesalso induce anotable growtharrest in sensitive tumors.importance of growth arrest and cell death remains

explore this issue, we will rst discuss the forms of cellhen compare the potential for cell death and cell growthntribute to endocrine responsiveness.th pathways include signaling to apoptosis, autophagy,strophe, necrosis, and senescence. Late events in cellasonably well dened at the molecular (such as PARPd cellular levels (including DNA disintegration). How-

ledge of the regulatory signaling upstream of thesehow this signaling is integrated and processed, is now

e incomplete. Mitochondrial function and integrity, reg-rt by BCL2 family members, are central to several forms[3436].

sis

is is a programmed cell death dened by morphologicalted to organized chromatin condensation and frag-of the cell nucleus, accompanied by cleavage of DNA,f apoptotic bodies, cell shrinkage, and rufing of theane [35,37,38]. Two major pathways are involved. Theitochondrial) pathway is regulated by the proapoptoticoptotic BCL2 family members; this pathway involvesmitochondrial membrane permeability (MMP), release

e c, exposureofphosphatidylserineon theouter leaeta membrane, and the eventual loss of plasma mem-

rity [39]. The extrinsic (cell surface receptor) pathwaynt upon extracellular signals including tissue necrosisF), Fas ligand, and TNF-related ligand TRAIL [37,38].

ic and extrinsic pathways activate caspases, the exe-of apoptosis, which cleave DNA and catabolize then. Apoptosis is not a discrete process and occurs over(418h), middle (1836h), and late stages (36h) areibed based largely on data from cell culture models.specic BCL2 family members (early events that cananges in MMP), changes in MMP, and the exposuretidylserine are generally interpreted as representingddle apoptosis. Cytoplasmic cytochrome c release fromia, changes in propidium iodide staining, increased ter-ferase dUTP nick end labeling (TUNEL) and cleavage ofpair enzyme PARP-1 are associated with late apoptosis[35].

agy

gy is a lysosomal pathway where cytoplasmic contentsd by double/multi-membrane vacuoles or autophago-ally resulting in the removal of defective or damaged

e.g., mitochondria. A better understanding of the regu-tophagy has recently begun to emerge; key regulatorsown to include BCL2 family members [40,41] and their

effectiAutophchrommissiocleavagp62/SQreversiother c[49] orto apo

Parwhenan alteoutcomenergyautophdigesticell cacan acindepe

2.3. M

FaucheckpDisrupin rapican div[61] this char

2.4. N

Necuolizatand ancell copan bluconden[44,62

2.5. Se

Senvacuolarrestegalactotime, sto celof p53telomecrisis

2.6. En

Predrawalmay noinvolvecer celhave m[6568echanism for regulating BECN1-activated autophagy.can be identied by the absence of marginated nuclearthe presence of cytoplasmic vacuoles using trans-ctron microscopy or monodansylcadaverine [44,45],f the LC3B protein [46,47], and regulation of the1 protein [48]. Early events in autophagy may belater events may (or appear to) share mechanisms withath pathways. For example, cleavage of ATG5by caplaingulation of BID [41] can cause a switch from autophagys.ically, autophagy can act as a cell survival mechanismcellular nutrients or growth factors are limited, or asive cell death pathway to apoptosis [50]. Prosurvivalikely reect an adequate adjustment to stress, withrients recovered from the organelles digested in theomes. Prodeath outcomes may arise when the self-f autophagy leads to such a loss of organelles that thelonger survive. In cancer cells, autophagy induction

ate cell death [5155] or promote cell survival [5658],tly or in response to treatment with cytotoxic agents.

catastrophe

NA structure checkpoints, or the spindle assembly, are key components of this form of cell death [59,60].of the normal segregation ofmany chromosomes resultsdeath [59].When this cell death does not occur, the cellasymmetrically and produce aneuploid daughter cellsn become neoplastic [59,61]. Thus, mitotic catastropheized by multinucleation.

is

is a chaotic process marked by cellular edema, vac-f the cytoplasm, breakdown of the plasma membrane,ciated inammatory response caused by the release ofs into the surroundings. Increased permeability to try-other vital dyes, in the absence of organized chromatinn and DNA fragmentation, is characteristic of necrosis

ence

nt cells are characteristically enlarged, attened withd a large nucleus, be come permanently cell cycled unresponsive to mitogenic stimuli and express -se [45,63]. Normally, as telomerase activity falls overssive telomere shortening limits proliferation and leadssenescence or mortality stage 1 (M1). Inactivationby bypass M1 growth arrest, producing critically shortnd massive cell death called mortality stage 2 (M2) or].

ine-induced cell death in breast cancer

how breast cancer cells die following estrogen with-AI treatment) or AE treatment is unclear. Senescencethe dominant mechanism, since this process frequentlyA damage and p53 activation [38,45] but breast can-

spond to AEs and to estrogen withdrawal even if theyed p53 [35,65]. While apoptosis is clearly implicatedme of the apoptosis endpoints in prior studies may not

R. Clarke et al. / Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820 11

Fig. 1. Autophagy is enhanced upon FAS treatment in ER+ breast cancer cell lines. MCF-7 cells were tautophagy inducer tunicamycin (TUN), or ethanol control (vehicle) prior to staining with monodansylhas been induced.

distinguish among earlier events more closely implicated with sig-naling initiated through autophagy. Autophagyhas been implicatedin response to endocrine therapy [6971] andwealso see the induc-tion of sign

Fig. 1 shober of autopof cytoplasming [44,45]and reducedpreviously swithdrawaltosis and grAE sensitivity is reec(or estrogeproliferatiowith otherated with tbreast cancing to endocell growthautophagy.

2.7. Prolifer

One of tmodels in vity of endosensitive brincreased centirely clegrowth arre

Fig. 2. Autophlines. MCF7/Land immunobdecreased p62

contribute tanchorage-formation,

e conrowdingchan[66],ptosistuded tosL:Fale BtreerapBCL2meaath mll decleain thor e

timeres oes ofandawaltopherprwitagyer thsis mbe aof c

onsidicant autophagy associated with endocrine therapies.ws our ability to detect signicant changes in the num-hagosomes as measured by an increase in the presenceic vacuoles identied bymonodansylcadaverine stain-

(Fig. 1), increased cleavage of the LC3 protein [46,47],expressionofp62/SQSTM1[48,7274] (Fig. 2).Wehavehown, as have others, that AE treatment and estrogenare also accompanied by increases in the level of apop-owth arrest in sensitive cells. Indeed, when restoringity in resistant cells we frequently see that sensitiv-ted in the restoration of an ability of the antiestrogenn withdrawal) to both increase apoptosis and reducen [75,76]. As shown in Figs. 1 and 2, and consistentreports [6971], prodeath autophagy also is associ-he growth inhibitory effects of endocrine therapies iner cells. Thus in experimental models, cells respond-crine therapies concurrently experience an increase inarrest accompanied by both apoptosis and a prodeath

ation, cell death, and endocrine responsiveness

he most consistent observations in both experimentalitro and in vivo and in clinical specimens is the abil-crine therapies to induce a profound growth arrest ineast cancer cells. However, the relative importance ofell death compared with reduced proliferation is notar. In most endocrine sensitive experimental models,st and cell death concurrently occur and both clearly

relativapies. Gresponrobustdo notof apotrast tois relatthe Famolecuof TAMTAM thabove,and itscell de

If cevationexistportingat thismeasuing ratFigs. 1withdrand auWe intsistentautophwhethapoptoit mayportionoften cagy is enhanced upon FAS treatment in ER+ breast cancer cellCC1 cells were treated with FAS, TUN, or vehicle prior to lysislotting using standard procedures. Increased LC3BII (asterisk) and/SQSTM1 expression both indicate that autophagy has been induced.

versible comwith endpoapoptosis. Awith ultimaof the ratesis is as fastbefore 243events. Thesis and be ahave returnresponse mapoptosis. Ithe numbereated with FAS (ICI 182,780), the endoplasmic reticulum stress andcadaverine (MDC). Increased MDC staining indicates that autophagy

o the ability of endocrine therapies to affect changes independent cell number, anchorage-independent colonyor tumorigenesis over time [27,77,78]. Less clear is theirtribution indriving clinical responses to endocrine ther-th arrest appears to be readily detected in breast tumorsto endocrine therapy. Less clear is the ability to detectges in apoptosis. Some investigators do [79], and somesee an association of apoptosis or a molecular maker(s)s with clinical response. The latter is in marked con-ies in experimental models. For some studies, responsemolecular markers of apoptosis such as BCL2 [79] or

s ratio [80]. Notably, expression of the anti-apoptoticCL2 is reduced in responsive breast tumors by 3 monthsatment [79], while in breast tumors that remain aftery BCL2 expression is elevated [81]. However, as notedcan affect both an apoptotic and autophagic cell death

surement alone is likely a poor predictor of any specicechanism.ath does not occur in clinical breast cancer this obser-rly requires explanation. Several possible explanationse absence of compelling experimental/clinical data sup-liminating these explanations we make no assessmenton their relative merits. Firstly, it should be noted that

f apoptosis are usually the primary endpoints for assess-cell death. Our previously published results, the data in2, and the work of others [6971] show that estrogenor antiestrogens increase both the rates of apoptosisagy in breast cancer models responding to treatment.et this as a prodeath autophagy in sensitive cells, con-h other reports [6971]. It remains unclear whetheror apoptosis dominates as the cell death mechanism oris varies among different breast cancer cells. Measuringay be the wrong measure of cell death in tumors, orn inadequate measure if it represents only some pro-ells that die through this process. Secondly, apoptosis isered to comprise early,mid and late stages, and an irre-

mitment to cell death may not be robustly associated

ints other than those denitively reecting late stagemeasure of apoptosis that is not robustly associated

te cell death could provide an incomplete assessmentor extent of cell death. Thirdly, if the timing of apopto-in patient tumors as it is in vitro, measurements taken6h and/or after 3648h could miss many of the keymost sensitive cells would have been through apopto-lready dead and gone, and the rate of apoptosis coulded to the basal level. Fourthly, duration of the apoptoticay differ between basal apoptosis and drug-inducedf drug-induced apoptosis leads to a more rapid death,r of cells processing though apoptosis could increase

12 R. Clarke et al. / Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820

without any detectable change across time in the apparent rate ofapoptosis.

Finally, a reduction in cell proliferation alone could be suf-cient to account for some shrinkage of tumor size, as the rate ofcell replaceloss from eand metastalso undergarrest alonein tumor sizof presumatreatment).good responbasal rate othis applieswhether bethe growthwhere mathcould compliferation antime.

While thmary cell dbreast tumotherapy-indtherearepoogy of thein tumor shmany cellsthe ability tovercome tresistant. Twomen whreduction in

Whethetually becomsatisfying hthe abilitylonged perieventuallyThis procesof growth atern consistthat initialldecade or m

3. Molecul

The preremain uncnaling affechormonal rthe mechantion elsewhmolecular sintegratedaprimarily oation in theof a separat

The primenable gloexpressionDNA sequereached athe genera

whether the read-outbegenomicor transcriptomicdata; this also istrue of the emerging high-throughput proteomic technologies. Theremarkable volume of data, and the diversity of biological infor-mation that informs the interpretation of these data, has begun to

rm tics. Hndersalyse of

l addwork

netw

ogennd b85].morspreviokesalingc signctiosuchalin

survie maThision otion,we etworing tine rre chinan

s cellrtainn attsful ae. This likof aork

rk of

sic c

lularthanrk isns (e-oned oraineds onnginuilt sh dind salsothe te actin thin thment might no longer be sufcient to account for cellither a basal rate of cell death and/or loss to migrationasis. However, unless almost all growth arrested cellso some form of cell death, it is unclear why growthshould lead to large and relatively rapid reductions

e (over several weeks compared with often many yearsbly much longer growth prior to clinical detection andGrowth arrest alone may be sufcient to account forses in some tumors, particularly where there is a highf cell death. However, it is not immediately clear howto tumors with an inherently low rate of proliferation,cause the growth fraction is large but cycling slowly orfraction is small but proliferating rapidly. This is an areaematical modeling could be particularly useful, since itare the effect sizes needed for relative changes in pro-d cell death to affect predicted overall tumor size over

ere is currently no denitive understanding of the pri-eath mechanisms in either experimental models or inrs in women, or of the relative importance of endocrineuced changes inproliferation comparedwith cell death,tentially important implications for theunderlyingbiol-cancer cells. If the primary driver of response as seenrinkage is a reduction in proliferation, this will leave

alive and still metabolically active. Surviving cells haveo adapt to the endocrine-induced stress and eventuallyhe proliferative blockade and growthey will becomehis process seems unlikely to occur in many of thoseo receive the clear long term benet of a signicantthe risk of death [8,9].

r it is the growth arrested but surviving cells that even-e resistant is unknown but it is certainly an intuitively

ypothesis. Moreover, this hypothesis is supported byto take sensitive cells in culture, expose them for pro-ods to either estrogen withdrawal or AE treatment, andinduce an acquired resistant phenotype [27,28,77,82].s is accompanied by a profound and prolonged periodrrest prior to the emergence of resistant cells, a pat-ent with the clinical progress of the disease in tumorsy respond to therapy but that eventually recuroften aore after the initiation of TAM treatment.

ar signaling and resistance

cise mechanisms of resistance to an AE and/or an AIlear, reecting an incomplete understanding of the sig-ting cell proliferation, survival, and death and their

egulation in breast cancer.Wehave previously reviewedisms of resistance to AEs and to estrogen depriva-ere in some detail [10,23,29], so we focus here on theignaling aspects of resistance and how these may bendexploredusingemerging technologies.Wewill focusn signaling to cell deathsignaling to regulate prolifer-context of endocrine responsiveness will be the subjecte review.ary technologies that have matured sufciently to

bal approaches to network modeling include genemicroarrays, ChIP-on-chip, SNP chips, high-throughputncing, and array CGH. Each of these technologies hashigh level of maturity, and each is characterized bytion of very high dimensional data on each sample

transfoformatfully udata aning somwe wilfor net

3.1. A

Estrtypes a[10,33,ER+ tumanysis invin signspeciER fuagents

Signdeath/versiblpoint.activatintegraHencethis nereectendocrwork aor dom

Thiand ceing in asuccesand dinalingan areaof netwnetwoness.

3.2. Ba

Celworksnetwonectioone-toforwarconstrbehaveto chaoften bare higconfoueventsaffectmay bmentsis seenhe elds of biostatistics, computer science, and bioin-owever, the properties of these datasets are often nottood nor are the challenges these properties provide foris and network modeling. Readers interested in explor-these challenges can read recent reviews [83,84]. Hereress briey several approaches to the use of these datamodeling.

ork signaling hypothesis of endocrine responsiveness

-independence and AE resistance are complex pheno-oth genomic and non-genomic activities are implicatedWe consider it unlikely that endocrine resistance inis driven by a single gene/signaling pathway. Unlike

ous single gene/pathway studies, our central hypothe-a gene network that confers diversity and redundancy[10,86]. The cell death/survival network incorporates

naling as affected by estrogen and AE modication ofn. Thus, AEs regulate this networkdifferently thanotheras cytotoxic drugs.

g leads rst to the reversible initiation of several cellval signaling pathways within the network. The irre-chinery of cell destruction is activated at some latermachinery may induce common outcomes such asf effector caspases and DNA/plasma membrane dis- independent of the early specic initiating signals.nvisionmultiple concurrent signals processing throughk, some prosurvival and some prodeath, with cell fatehe dominant signaling. In endocrine resistant cells,egulation and/or function of components of this net-anged andprodeath signals are either no longer inducedt.fate signaling network hypothesis is intuitively logical

ly testable. Evidence that cells induceprosurvival signal-empt to circumvent stressors implies that some cells arend ultimately survive whereas others are unsuccessfulus, the balance between prosurvival and prodeath sig-ely the nal arbiter of cells fate [83]. While this remainsctive investigation, we rst discuss the basic principlesmodeling and then provide an example of a seed-geneendocrine-regulated signaling in endocrine responsive-

oncepts of gene networks

signaling occurs more in the context of interactive net-through linear pathways [83]. The basic topology of adened by nodes (genes/proteins) and their intercon-dges). Interconnections are multi-faceted and include, one-to-many, or many-to-one relationships, and feed-feed-back loops. The dynamic activity of a network isby the various forms of interactions, and the network

ly in certain ways and controlled manners in responseg cellular conditions or external stimuli [87]. Whileolely from gene expression microarray data, these datamensional and contain spurious correlations that canimple solutions for network building [83,84]. Relevantoccur in the genome and proteome, some of which canranscriptome. For example, a transcription factor (TF)ivated by phosphorylation and bind to responsive ele-e genome but the regulation of its downstream targetse transcriptome [83]. An example of this relationship is

R. Clarke et al. / Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820 13

Fig. 3. Illustraidentied as band gene exprto identify rela

the ligand-ilation on SE

Simplisting of high dis probablycommerciatask to perbioinformationsof gencellular funassigned; th

The soluare often chnodes linkepretation chconfoundincurse of dimporate the cof self-fulchallenges aand the annthese algoriwhen carefwhen expercombine exsuch approaesis generatmeaningful

3.3. The se

The bottgene approthe extractimarydata; tways. We wbut providebe applied tnovo. Local

the initial seed genes. A simple approach is the incorporation of acanonical pathway (which may be a subnetwork in what would bea nal and much broader network) when it is known to be relevantin the cellular context under study and where incorporating the

andatistwledher napplwn (corpotreamhmsnal dnteratheimos

and ed knavanal dtandpresshe eme, outioinveenerainvese semetnetween pene], Bas [31netwexprgratiptioativeincor1] hation of the complex and challengingnature of pathway analysis. Geneseing differentially expressed in resistant MCF7/LCC9 cells by SAGEession microarray were analyzed by Pathway Architect (Stratagene)tionships in silico.

ndependent activation of ER following its phosphory-R118 by MAPK [88].ically, thereare twobasic approaches tonetworkmodel-imensional data: top-downandbottom-up. The formerthe most widely used approach as several accessible

l software packages are available that make this an easyform without the need for training in biostatistics ortics. These packages often apply various implementa-eontologic and semantic searchalgorithms that identifyctions and pathways to which individual nodes areese data are then graphically represented.tions producedby several popular top-downalgorithmsaracterized by representations of tens-to-hundreds ofd by hundreds-to-thousands of edges, making inter-allenging (Fig. 3). Whether the algorithms address theg properties of high dimensional spaces, such as the

nodeswith st

Knoof anotcan bebe grothat indownsalgoritfunctioThus, iality of

Thenodesderivealreadyfunctiousing soverexwith tribozycontribenabletally gunder

Onclogicalneuralhave bfrom g[2426modelneuralmRNA

IntetranscrAn iteror thedata [4ensionality or the confound of multimodality, or incor-ritical aspects of cellular context and alleviate the traplling prophesy, is not clear [83]. Among the additionalre the incompleteness of relevant biological knowledgeotation error rate in the source databases searched bythms [83]. Nonetheless, these approaches can be usefulully applied and their limitations fully understood, andts from both the biological and mathematics domainspertise to assess the validity of the solutions. Currently,ches probably have most to offer in the area of hypoth-ion, rather than in the construction of truly biologicallysignal transduction networks.

ed-gene approach to network modeling

om-up approach is generally referred to as the seed-ach to network modeling [89]. This approach requireson of a small number of seed genes from within the pri-hese genes are thenused to grow thenetwork in severalill not address all the various approaches in this reviewa few brief examples. Various modeling methods cano nd and link adjacent nodes, growing the network desubnetworks can be identied and overlaid or linked to

networks. Sods have als(NCA) is a nmethods catems more

Other lilogical knowdirectionalitional datapartially) asources areof differentcal knowledcontains falinference.

4. Seed-gecell fate

While weling, we hexpanding dedges of the canonical pathway members is consistentical properties of the growing model topology.ge of howa gene (node) affects the expression/functionode provides directional connectivity information that

ied to the interacting nodes. Transcription networks canor transcriptional edges between nodes in a networkrates otherbiological knowledge) by linkingTFs to theirtargets. These targets can be predicted using specic

[9093]; where possible it is preferable to incorporateata such as that obtained fromChIP-on-chip arrays [91].cting nodes can be identied along with the direction-r edges as the seed-gene network is grown.t labor intensive approach is to derive experimentallydges, growing the network using denitive laboratory-owledge. Where additional high-throughput data areilable, such as ChIP-on-chip, this is preferable. Currently,ata is probably more often obtained one gene at a time,ard molecular methods such as gene knock-down andion. This laborious approach is becoming supplanted

merging functional genomic methods such as siRNA,r antisense libraries that can test experimentally then of hundreds to thousands of genes. These methodsstigators to extract concurrently nodes that experimen-te biologically appropriate changes in the phenotypetigation.eds and their edges are identied, and functional bio-adata obtained, interactive models can be grown usingork and other machine learning tools. Several modelsroposed to reveal the behaviors of regulatory networksexpression data [22,23] including Boolean networksyesian networks [2730], linear additive regulation

,32], state-space models (SSMs) [33,34], and recurrentorks (RNN) [35,36]. However, these methods use onlyession data to infer networks.ed approaches have been recently proposed to learnnal regulation from various data sources [27,30,3743].searchonmRNAexpression andChIP-on-chipdata [37],poration of expression proles, ChIP-on-chip, and motifve each been used in yeast to discover transcriptionaleveral linear models or matrix decomposition meth-o been proposed [4346]. Network component analysisotably powerful approach [45] but NCA and these othernnot easily infer regulatory networks in biological sys-complex than yeast.mitations exist in network modeling. Complete bio-ledge for topology estimation (nodenode edges and

ty), such as high-throughput ChIP-on-chip data or func-from laboratory experiments, are often not (or onlyvailable for human cells. When heterogeneous dataintegrated for computational inference, the consistencydata sources is often inadequate or unknown. Topologi-ge also comes frombiological experiments,whichoftensepositives/negatives that can lead to incorrect network

ne model for cell signaling and the regulation of

e continue to develop new methods for network mod-ave yet to report our modeling approaches to our ownata sets. Hence,wewill here describe our initial studies

14 R. Clarke et al. / Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820

on the use of seed genes and experimental data to construct a sim-ple wiring-diagram of our initial seed-gene network. The inabilityto induce signaling to irreversible cell death is a central componentof drug resistance [94]. Thus, we propose that cells possess a com-mon cell deand/or inte

Prior totial nodes (Since ER isinuence ennumber ofThe full listis unknownhypothesis[75,95] arework. Moreresistant brand directicondence.

4.1. X-box bresponse (U

UPR is a[98], an adaaccumulatilumen [99]recover normay inducesors suchasIRE1, ATF6the cytosol)ATF6-inducby IRE1)survival oractivation a[103].

XBP1 isregion/leucXBP1(U), hanegative ofa moleculaand createstion by XBspecic cAMcore sequenUPR elemening plasmaneurologication, and livin responsework of othactivate ER

We have(effectivelyand FAS crocancer cellsXBP1(S), asMCF-7 or T4This observadequate anonly knowncan splicemtion primarrespond toof UPR.

hysical association of XBP1 and ER is accompanied by robust ERE-driventional activity in MCF7/XBP1 cells. (A) MCF-7 cells stably expressing XBP1the empty vector control (c) were treated with FAS or ethanol control (ctrl.)rior to lysis and immunoblotting (lanes 1 and2) or co-immunoprecipitationand ER (lanes 3 and 4) using standard procedures. (B) MCF7/c and

BP1 cells were transiently co-transfected with plasmids encoding 3xERE-e and phRLSV40-Renilla for 24h prior to lysis and promoterreportere assay by standard methods. Data are presented as mean relative ERE-e activity SE for a representative experiment performed in triplicate,1.

terferon regulatory factor-1 (IRF1)

P linkage analysis assigned the IRF1 gene to 5q23-31; moreive studies identied the locus as 5q31.1 [114]. IRF1 was ini-dentied because of its transcriptional activation of typeferon (IFN) genes. We rst showed the ability of interfer-sensitize breast cancer cells to TAM over 20 years ago

More recently, IRF1 was implicated in T-cell developmentand it is now known also to coordinate expression of theoproteasome [117], to regulate human telomerase activ-8,119], and to regulate key aspects of DNA damage repair1]. Loss of IRF1 increases tumorigenicity in mouse mod-ven by ras or loss of p53 [122]. These activities may reectability to signal to apoptosis [123], which can occur in apendent or -independentmanner [120,124],withorwithoution of p21cip1 [124] or p27kip1 [125], and through caspase-1aspase-3 [96], caspase-7 [96,126], caspase-8 [96,127], and/or28].owing our initial observations of IRF1s likely role in breast[129131] and antiestrogen resistance [129], we conrmedctional involvement using a dominant negative approach1) [65]. IRF1 and dnIRF1 induce opposing effects on pro-on in vitro and tumorigenesis in vivo through regulationases-3/7 and caspase-8 activities [96]. These observations

nsistent with the effects of inoculating an adenoviral vec-ath/survival regulatory decision network of integratedracting pathways (see above).building network models, it is necessary to extract ini-seed genes) from which a network can be built [89].a TF and regulates other functionally relevant TFs thatdocrine responsiveness and cell fate, selecting a small

TFs as seed genes is reasonable for network modeling.of relevant ER-regulated TFs that may affect cell fate. Nonetheless, our published data support the centralthat that IRF1 [65,9597], XBP1 [76,95] andNFB (RELA)key regulatory nodes or control key modules in this net-over, our experimental data in endocrine sensitive andeast human cancer cells nowallowus tomap their edgesonality, in an appropriate cellular context, with some

inding protein-1 (XBP1) and the unfolded proteinPR)

central component of the endoplasmic stress responseptive signaling pathway that allows cells to survive theon of unfolded proteins in the endoplasmic reticulum. Initially a compensatory mechanism allowing cells tomal endoplasmic reticulum function, a prolonged UPRcell death. UPR, which can be induced by cellular stres-hypoxia, is activatedbyeachof threemolecular sensors:, PERK [100]. XBP1s unconventional splicing (occurs inby IRE1 is an obligate component in both IRE1- anded UPR [100,101]. The UPR (initiated by XBP1 splicingcan activate autophagy [102]. Whether this is a pro-prodeath form of autophagy is unknown, since UPRlso can induce both prodeath and prosurvival outcomes

a transcription factor that belongs to the basicine zipper (bZIP) family [104,105]. The unspliced form,s amolecularweight of 33kDa and acts as a dominantspliced XBP1 [106,107]. The spliced form, XBP1(S), hasr weight of 54kDa; splicing removes a 26bp introna translational frame-shift. Regulation of transcrip-

P1(S) is a consequence of its homodimers activatingP response elements (CREs) with a conserved ACGT

ceGATGACGTG(T/G)NNN(A/T)Tsometimes called thet [103,104,108]. XBP1(S), which is implicated in affect-cell differentiation [109], is essential for fetal survival,l development, bone growth, immune system activa-er development [110,111]. XBP1 is also rapidly inducedto estrogen-stimulation [112,113]. Consistent with theers [108], we have shown that XBP1(S) can bind to andin a ligand-independent manner (Fig. 4).

recently shown that XBP1(S) confers E2-independencean AI resistant phenotype) and AE crossresistance (TAMssresistance) in both MCF-7 and T47D human breast[76]. This activity appears to be driven primarily byintroduction of the full-length XBP1 cDNA in either7D cells generates predominately the XBP1(S) protein.

ation suggests that the basal activity of IRE1 is alreadyd that XBP1(S) is the rate limiting protein. XBP1 is thesubstrate for the IRE1 endonuclease and only IRE1ammalian XBP1. Since XBP1 splicing is thought to func-ily within the UPR, breast cancer cells may be primed tomultiple stressors by activating a prosurvival induction

Fig. 4. PtranscripcDNA orvehicle pof XBP1MCF7/Xluciferasluciferasluciferas*p

R. Clarke et al. / Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820 15

tor containing IRF1 directly into mouse mammary tumors [132].While p53-dependent apoptosis occurs in the breast [133], T47Dcells express mutant p53 and our data show that intact p53 is notrequired for the proapoptotic actions of IRF1 [65,96]. In AE sen-sitive breasdnIRF1 is acobservationand extendbreast cell cbothp53-devides a neintegratingAEs.

4.3. Nuclea

The NFogous protechromosomtained in thof the IBnuclear tranthe IKK kinubiquitinatcated in seregulationare involvegland deveNFB activrat [144] aNFB is assinant NFBp52 family[146].

We haveand AE crosin vitro anding activityfunctional i(NFB reprare derivedtive to AEsto estrogenindependen

LCC9 ceexhibit a fuative to LCof NEMO (of activitybut not AEof its antiethese cellsThus, we coto vehicleinhibitor ofinhibition omately 600In markedcells at eitparthenolidtosis. FAS a[75]. Sinceat least athe LCC9 cminant [75reliance up

of NFB restores their sensitivity to apoptosis induced by FAS[95].

pression of ER, PGR, XBP1, NFB and IRF1 in breast tumors

nggestruccell l+; 2tumoes anted gs [151eastf there apine-rary televaeginstudibreaoexpell li480liedenessibileachientcorrourthignaconng ths. XBbrea

diesble i

canocrincancthe

s [146ressis. IRFBP1.ation, thesionle, wthatxpre

152])ith thWhecorrns reing sR+ bromnctiodelsh twlasmt cancer cells, inhibition of AE-induced IRF1 activity bycompaniedby reducedproapoptotic activity [65]. Theses on IRF1 and AE responsiveness have been conrmeded by others in both normal [134] and other neoplasticulture models [135,136]. IRF1, which can signal throughpendent and -independentmechanisms [120,124], pro-

w and potentially important signaling molecule forand regulating breast cancer cell survival in response to

r factor kappa B (NFB)

B p50/p65 heterodimer complex comprises two homol-ins; the p50 product of its p105 precursor (NFB1;e 4q24) and the p65 (RELA; 11q13). NFB is main-e cytosol in an inactive state, bound with members

family that inhibit nuclear transport or block NFBsslocation signal [137]. Activation usually proceeds byase complex phosphorylating IB, resulting in IB

ion and degradation [138]. NFB (RELA/NFB1) is impli-veral critical cellular functions [139]. Reecting itsby both estrogen and growth factors [140,141] thatd in endocrine resistance [10,142], normal mammarylopment is dependent upon NFB [143]. Increasedity arises during neoplastic transformation in thend mouse mammary gland [145]. Upregulation ofociated with E2-independence [140,143]. The predom-form in breast cancer cell lines is RELA/NFB1; themember also is expressed in some breast cancers

shown that NFB can confer estrogen-independencesresistance [75,95,147]. Estrogen-independent growthin vivo is supportedby increases inbothNFBDNAbind-and expression of BCL3 [147]. This study highlights themplications of NFB in AI resistance. Expression of IBessor) in estrogen-independent LCC1 cells (LCC1 cellsfrom MCF-7 and are estrogen-independent but sensi-[148]), which have increased NFB activation relative-dependent MCF-7 cells, eliminates their estrogen-ce in vivo.lls (TAM and FAS crossresistant variant of LCC1 [28])rther increase in NFB expression and activation rel-C1 cells, apparently driven by increased expressionIKK) [75]. These observations imply that the levelin LCC1 cells is adequate for estrogen-independenceresistance. Increased activation of NFB [95] and lossstrogenic regulation in LCC9 cells [75] suggest thatmight be dependent upon NFB for survival/growth.mpared the growth response of LCC1 and LCC9 cells

or parthenolide (300 and 600nM), a small moleculeNFB [149]. Parthenolide produces a dose-dependentf MCF7/LCC9 cells with an apparent IC50 of approxi-nM (p

16 R. Clarke et al. / Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820

Fig. 5. Endocr XBP1,cellular contex

4.5. Simpleand IRF1 bacellular cont

The expsentation opreceding sthrough thiand NFB asion is reprein sensitivestrogen ressignaling is

In additition of NPMNPM and IRexpressionSince NPM[156], the iand inhibitalso cannotpendent ofto transformIncreased leon TAM 6-m

IRF1 anddirectly gennitric oxideprimarily ththeir subcetering one ooccur), thisthe inverseand in somof any remacells and/ora dominancof resistantrelevance o

We haveantiestrogeNEMO/IKKwell documvival remai

ers on-inNFBet toin plateder boore

o pre

centivatrogelthos unss. Au. BECck Breguhersagyir sigloca

enprphagignaline resistance seed-gene network. Simple representation of a seed-gene network oft (resistant MCF7/LCC9 cells).

representation of a seed-gene network of XBP1, NFBsed on functional data obtained from an appropriateext

erimental data supporting the wiring-diagram repre-f the network model shown in Fig. 5 are discussed theections. Here we discuss how the signals may ows network. The three primary seed genes of IRF1, XBP1,re evident as previously proposed [95]. IRF1 expres-ssed in resistant cells [95] but induced by antiestrogenscells [65]. A dominant negative IRF1 confers an antie-istant phenotype, implying that IRF1-driven prodeathkey to the regulation of cell fate [65].on to changes in the expression of IRF1, the upregula-expression [95,155] could also affect IRF1 action. BothF1 are estrogen-regulated genes in MCF-7 cells, IRF1being suppressed, whereas NPM is induced [129,155].inhibits the transcription regulatory activities of IRF1ncrease in NPM expression could bind remaining IRF1its ability to initiate an apoptotic caspase cascade. Weexclude the possibility that NPM has activities inde-blocking IRF1, since NPM overexpression is sufcientNIH 3T3 cells in a standard oncogenesis assay [156].

vels of serumautoantibodies toNPMpredict recurrenceonths prior to clinical detection [157].NFB are known to form heterodimers and to regulate

membestrogeWhilehave yies areupreguto conf[76]. Mshow t[154].

TheUPR acantiest[102], aremainsiveneBECN1can blo

Theably otautophof theis onebetweis autolinks se expression [158,159] including that of the induciblesynthase promoter [158]. Since we do not know if it ise gene regulatory effects of these heterodimers, or if

llular location is key (they act by preferentially seques-r the other so that transcriptional regulation does notis shown as a dotted line. We would predict, based onexpression between NFB and IRF1 in LCC9 cells [95]e breast cancers [152], that either the prodeath effectsining IRF1 are being sequestered by NFB in resistantthat the overexpression and activation of NFB leads toe of its prosurvival activities. The increased sensitivitycells to parthenolide is consistent with the functionalf at least the latter signaling outcome [75].previously shown that the upregulation of NFB in

n resistant cells [95] is likely driven in part by increasedactivity [75]. The prosurvival activities of NFB areented [160]. Precisely how NFB regulates cell sur-

ns to be fully established but activation of prosurvival

mitochondrthe contextresponsivenrepresentsertheless, tfunctional dseed genesbegin a mostanding thcandidateshow individ

References

[1] A. Jemastatistic

[2] W.D. FoTung, OEstrogeNFB and IRF1 based on functional data obtained from an appropriate

f the BCL2 gene family are involved in both acquireddependence [147] and antiestrogen resistance [75,76].is predicted to induce transcription of XBP1 [109], we

report this direct regulation in breast cancer cells (stud-rogress). Whether or not this occurs, XBP1 is clearlyin resistant cells [95] and this activity is sufcient

th estrogen-independence and antiestrogen resistancerecently, increased XBP1 mRNA expression has beendict for a poor response to TAM in breast cancer patients

tral role of XBP1 within the UPR clearly implicatesion in responsiveness to both estrogen-withdrawal andn treatment [76].UPRalso is knownto induceautophagyughwhether this is a prosurvival or prodeath autophagyclear in the context of determining endocrine respon-tophagy is regulated, at least in part, by the action of

N1 activity is regulated byBCL2,which binds BECN1andECN1-mediated autophagy [36].lation of BCL2 family members (BCL2, BCL3, and prob-) whether by IRF1, NFB, and/or XBP1, can affect bothand the intrinsic apoptosis pathway. The intersectionnaling at BCL2 family members, as shown in Fig. 5,tion within the broader network where the balanceodeath andprosurvival signaling, andwhetherprodeathic or apoptotic, is determined. This intersection alsoing through the UPR and endoplasmic reticulum to the

ia with the cell fate decision mechanismsat least inof determining cell fate in the context of endocrineess in breast cancer. The signaling depicted in Fig. 5only a small component of this broader network. Nev-his initial wiring-diagram is consistent with a body ofata in experimental models and it provides sufcient

, their edges, and the directionality of these edges, tore detailed exploration of this central network. Under-is networks topology and function will lead to betterfor drug discovery and to better algorithms to predictual tumorswill respond to specic endocrine therapies.

l, R. Siegel, E. Ward, T. Murray, J. Xu, C. Smigal, M.J. Thun, Cancers, 2006, CA Cancer J. Clin. 56 (2006) 106130.ulkes, K. Metcalfe, P. Sun, W.M. Hanna, H.T. Lynch, P. Ghadirian, N..I. Olopade, B.L.Weber, J.McLennan, I.A. Olivotto, L.R. Begin, S.A. Narod,n receptor status inBRCA1-andBRCA2-relatedbreast cancer: the inu-

R. Clarke et al. / Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820 17

ence of age, grade, and histological type, Clin. Cancer Res. 10 (2004) 20292034.

[3] S.M. Thorpe, Estrogen and progesterone receptor determinations in breastcancer. Technology, biology and clinical signicance, Acta Oncol. 27 (1988)119.

[4] W.D. Fotive of eInst. 96

[5] Early Brbreast c

[6] B. Fishemond, NAbramsate andwith estgical Adndingsvention(1996) 1

[7] E.G. MaG. Falksnegative16 (1998

[8] Early Brcancer:

[9] Early Brbreast c115.

[10] R. Clarkmacolog2571.

[11] M.P. Colcancer. A

[12] S.R. CumNorton,GlusmainpostmOutcom

[13] A. Howea speciLancet 3

[14] A. HoweKleeberICI 182,advanceOncol. 2

[15] G. FerreRuggeriTerzoli,line enda pooled

[16] B. ThurlR. ParidWardleypostmen274727

[17] R.C. CooI. Alvareman, L.JStewartrandomin postm(2004) 1

[18] F. Boccaris, G. CraminoglmenopaClin. On

[19] R. JakesKwasnyWolfganearly breresults o

[20] M. BaumAnastrofor adjucancer:trial ef

[21] J. Bonnevon Eultamoxifcarcinom

[22] H. MourJ. ApffelBecquarH.A. Cha

asrst-line therapy for postmenopausalwomenwith advancedbreast cancer:results of a phase iii study of the international letrozole breast cancer group,J. Clin. Oncol. 19 (2001) 25962606.

[23] R. Clarke, N. Brnner, Cross resistance and molecular mechanisms in antie-strogen resistance, Endocr. Relat. Cancer 2 (1995) 5972.

Osipoestradncer I.M. GoCF-7 tncer RK. Osbde min. On. Brng, M.man

rogen. Brnn, M.sistanrogentiestrClark

Briend theNomurogenatl. Ca. Haddvance

.L. Wa949) 7Clarke mem990) 1Pattinncer RRiggind apoPattinhneidtophaZhivoll Bio. Bro

eatme. GoldleasevarianShimompsamme2112. Lamr MCFtopha

. Fururved dppresH. Liavine,ature 4.L. EdtophaE. Broview,.Mizu(2007Kimution ptopha

.J. Klio. Babargam. Brumen, L.Clave

.M. Cuevenisjavahe.T. Eisentes

. Gohlaeb, D..P. Huulkes, Re: Estrogen receptor status of primary breast cancer is predic-strogen receptor status of contralateral breast cancer, J. Natl. Cancer(2004) 10401041.east Cancer Trialists Collaborative Group, Polychemotherapy for earlyancer: an overview of randomised trials, Lancet 352 (1998) 930942.r, J. Dignam, E.P. Mamounas, J.P. Costantino, D.L. Wickerham, C. Red-. Wolmark, N.V. Dimitrov, D.M. Bowman, A.G. Glass, J.L. Atkins, N.

on, C.M. Sutherland, B.S. Aron, R.G. Margolese, Sequential methotrex-uorouracil for the treatment of node-negative breast cancer patientsrogen-receptor-negative tumors: eight year results fromNational Sur-juvant Breast and Bowel Project (NSABP) B-13 and rst report offromNSABPB-19 comparingmethotrexate anduorouracilwith con-al cyclophosphamide,methotrexate, and uorouracil, J. Clin. Oncol. 149821992.nsour, R. Gray, A.H. Shatila, D.C. Tormey, M.R. Cooper, C.K. Osborne,on, Survival advantage of adjuvant chemotherapy in high-risk node-breast cancer: ten-year analysisan intergroup study, J. Clin. Oncol.) 34863492.east Cancer Trialists Collaborative Group, Tamoxifen for early breastan overview of the randomized trials, Lancet 351 (1998) 14511467.east Cancer Trialists Collaborative Group, Systemic treatment of earlyancer by hormonal, cytotoxic, or immune therapy, Lancet 399 (1992)

e, F. Leonessa, J.N. Welch, T.C. Skaar, Cellular and molecular phar-y of antiestrogen action and resistance, Pharmacol. Rev. 53 (2001)

e, C.T.A. Jones, I.D.H. Todd, A new antioestrogenic agent in late breastn early clinical appraisal of ICI 46474, Br. J. Cancer 25 (1971) 270275.mings, S. Eckert, K.A. Krueger, D. Grady, T.J. Powles, J.A. Cauley, L.T. Nickelsen, N.H. Bjarnson, M. Morrow, M.E. Lippman, D. Black, J.E.n, A. Costa, V.C. Jordan, The effect of raloxifene on risk of breast cancerenopausalwomen; results from theMORE randomized trial.Multiplees of Raloxifene Evaluation, J. Am. Med. Assoc. 281 (1999) 21892197.ll, D. DeFriend, J.F.R. Robertson, R.W. Blamey, P. Walton, Response toc antioestrogen (ICI 182,780) in tamoxifen-resistant breast cancer,45 (1995) 2930.ll, J.F. Robertson, A.J. Quaresma, A. Aschermannova, L. Mauriac, U.R.g, I. Vergote, B. Erikstein, A. Webster, C. Morris, Fulvestrant, formerly780, is as effective as anastrozole in postmenopausal women withd breast cancer progressing after prior endocrine treatment, J. Clin.0 (2002) 33963403.tti, E. Bria, D. Giannarelli, A. Felici, P. Papaldo, A. Fabi, C.S. Di, E.M.,M.Milella,M.Ciccarese, F.L. Cecere,A.Gelibter, C.Nuzzo, F. Cognetti, E.P. Carlini, Second- and third-generation aromatase inhibitors as rst-ocrine therapy in postmenopausal metastatic breast cancer patients:analysis of the randomised trials, Br. J. Cancer 94 (2006) 17891796.

imann, A. Keshaviah, A.S. Coates, H. Mouridsen, L. Mauriac, J.F. Forbes,aens, M. Castiglione-Gertsch, R.D. Gelber, M. Rabaglio, I. Smith, A., K.N. Price, A. Goldhirsch, A comparison of letrozole and tamoxifen inopausal women with early breast cancer, N. Engl. J. Med. 353 (2005)57.mbes, E. Hall, L.J. Gibson, R. Paridaens, J. Jassem, T. Delozier, S.E. Jones,z, G. Bertelli, O. Ortmann, A.S. Coates, E. Bajetta, D. Dodwell, R.E. Cole-. Falloweld, E. Mickiewicz, J. Andersen, P.E. Lonning, G. Cocconi, A., N. Stuart, C.F. Snowdon, M. Carpentieri, G. Massimini, J.M. Bliss, Aized trial of exemestane after two to three years of tamoxifen therapyenopausal women with primary breast cancer, N. Engl. J. Med. 3500811092.rdo, A. Rubagotti, D. Amoroso, M. Mesiti, D. Romeo, C. Caroti, A. Far-uciani, E. Villa, G. Schieppati, G. Mustacchi, Sequential tamoxifen andutethimide versus tamoxifen alone in the adjuvant treatment of post-usal breast cancer patients: results of an Italian cooperative study, J.col. 19 (2001) 42094215.z, W. Jonat, M. Gnant, M. Mittlboeck, R. Greil, C. Tausch, J. Hilfrich, W., C. Menzel, H. Samonigg, M. Seifert, G. Gademann, M. Kaufmann, J.g, Switching of postmenopausal women with endocrine-responsiveast cancer to anastrozole after 2 years adjuvant tamoxifen: combinedf ABCSG trial 8 and ARNO 95 trial, Lancet 366 (2005) 455462., A. Buzdar, J. Cuzick, J. Forbes, J. Houghton, A. Howell, T. Sahmoud,

zole alone or in combination with tamoxifen versus tamoxifen alonevant treatment of postmenopausal women with early-stage breastresults of the ATAC (Arimidex, Tamoxifen Alone or in Combination)cacy and safety update analyses, Cancer 98 (2003) 18021810.terre, A. Buzdar, J.M. Nabholtz, J.F. Robertson, B. Thurlimann, M.

er, T. Sahmoud, A. Webster, M. Steinberg, Anastrozole is superior toen as rst-line therapy in hormone receptor positive advanced breasta, Cancer 92 (2001) 22472258.

idsen, M. Gershanovich, Y. Sun, R. Perez-Carrion, C. Boni, A. Monnier,staedt, R. Smith, H.P. Sleeboom, F. Janicke, A. Pluzanska, M. Dank, D.t, P.P. Bapsy, E. Salminen, R. Snyder, M. Lassus, J.A. Verbeek, B. Stafer,udri-Ross, M. Dugan, Superior efcacy of letrozole versus tamoxifen

[24] C.inCa

[25] MMCa

[26] C.nuCl

[27] Nlinhust

[28] Nserestan

[29] R.Oan

[30] Y.stN

[31] Aad

[32] A(1

[33] R.th(1

[34] S.Ca

[35] R.an

[36] S.Scau

[37] B.Ce

[38] J.Mtr

[39] J.Crein

[40] S.Thgr12

[41] MceAu

[42] Nsesu

[43] X.LeN

[44] Aau

[45] L.re

[46] N3

[47] S.raAu

[48] DMBeJ.HChC.ADDNFuAtliW, C. Gajdos, H. Liu, B. Chen, V.C. Jordan, Paradoxical action of fulvestrantiol-induced regression of tamoxifen-stimulated breast cancer, J. Natl.nst. 95 (2003) 15971608.ttardis, V.C. Jordan, Development of tamoxifen-stimulated growth ofumors in athymic mice after long-term antiestrogen administration,es. 48 (1988) 51835187.orne, E.B. Coronado, J.P. Robinson, Human breast cancer in athymicice: cytostatic effects of long-term antiestrogen therapy, Eur. J. Cancercol. 23 (1987) 11891196.ner, T.L. Frandsen, C. Holst-Hansen, M. Bei, E.W. Thompson, A.E. Wake-E. Lippman, R. Clarke, MCF7/LCC2: A 4-hydroxytamoxifen resistantbreast cancer variant which retains sensitivity to the steroidal antie-ICI 182, 780, Cancer Res. 53 (1993) 32293232.ner, B. Boysen, S. Jirus, T.C. Skaar, C. Holst-Hansen, J. Lippman, T. Frand-Spang-Thomsen, S.A.W. Fuqua, R. Clarke, MCF7/LCC9: an antiestrogent MCF-7 variant in which acquired resistance to the steroidal antie-ICI 182, 780 confers an early crossresistance to the non-steroidalogen tamoxifen, Cancer Res. 57 (1997) 34863493.e, M.C. Liu, K.B. Bouker, Z. Gu, R.Y. Lee, Y. Zhu, T.C. Skaar, B. Gomez, K., Y.Wang, L.A. Hilakivi-Clarke, Antiestrogen resistance in breast cancerrole of estrogen receptor signaling, Oncogene 22 (2003) 73167339.ra, H. Tashiro, K. Hisamatsu, Differential effects of estrogen and antie-on in vitro clonogenic growth of human breast cancers in soft agar, J.ncer Inst. 82 (1990) 11461149.ow, J.M. Watkins, E. Paterson, Inuence of synthetic oestrogens upond malignant disease, Br. Med. J. 2 (1944) 393398.lpole, E. Paterson, Synthetic oestrogens in mammary cancer, Lancet 283789.

e, H.W. van denBerg, R.F.Murphy, Tamoxifen and 17-estradiol reducebrane uidity of human breast cancer cells, J. Natl. Cancer Inst. 82

7021705.gre, B. Levine, Bcl-2 inhibition of autophagy: a new route to cancer?es. 66 (2006) 28852888.s,A.H.Bouton,M.C. Liu,R. Clarke,Antiestrogens, aromatase inhibitors,ptosis in breast cancer, Vitam. Horm. 71 (2005) 201237.gre, A. Tassa, X. Qu, R. Garuti, X.H. Liang, N.Mizushima,M. Packer,M.D.er, B. Levine, Bcl-2 antiapoptotic proteins inhibit Beclin 1-dependentgy, Cell 122 (2005) 927939.tovsky, G. Kroemer, Apoptosis and genomic instability, Nat. Rev. Mol.l. 5 (2004) 752762.wn, L.D. Attardi, The role of apoptosis in cancer development andnt response, Nat. Rev. Cancer 5 (2005) 231237.stein, N.J. Waterhouse, P. Juin, G.I. Evan, D.R. Green, The coordinateof cytochrome c during apoptosis is rapid, complete and kineticallyt, Nat. Cell Biol. 2 (2000) 156162.izu, T. Kanaseki, N. Mizushima, T. Mizuta, S. Rakawa-Kobayashi, C.B.on, Y. Tsujimoto, Role of Bcl-2 family proteins in a non-apoptotic pro-d cell death dependent on autophagy genes, Nat. Cell Biol. 6 (2004)28.parska-Przybysz, B. Gajkowska, T. Motyl, BID-decient breast can--7 cells as a model for the study of autophagy in cancer therapy,gy 2 (2006) 4748.

ya, J. Yu, M. Byeld, S. Pattingre, B. Levine, The evolutionarily con-omain of Beclin 1 is required for Vps34 binding, autophagy and tumorsor function, Autophagy 1 (2005) 4652.ng, S. Jackson, M. Seaman, K. Brown, B. Kempkes, H. Hibshoosh, B.Induction of autophagy and inhibition of tumorigenesis by beclin 1,02 (1999) 672676.

inger, C.B. Thompson, Death by design: apoptosis, necrosis andgy, Curr. Opin. Cell Biol. 16 (2004) 663669.ker, F.A. Kruyt, G. Giaccone, Cell death independent of caspases: aClin. Cancer Res. 11 (2005) 31553162.shima, T. Yoshimori,Howto interpret LC3 immunoblotting, Autophagy) 542545.ra, T. Noda, T. Yoshimori, Dissection of the autophagosome matu-rocess by a novel reporter protein, tandem uorescent-tagged LC3,gy 3 (2007) 452460.nsky, H. Abeliovich, P. Agostinis, D.K. Agrawal, G. Aliev, D.S. Askew,, E.H. Baehrecke, B.A. Bahr, A. Ballabio, B.A. Bamber, D.C. Bassham, E.ini, X. Bi, M. Biard-Piechaczyk, J.S. Blum, D.E. Bredesen, J.L. Brodsky,ell, U.T. Brunk, W. Bursch, N. Camougrand, E. Cebollero, F. Cecconi, Y.

S. Chin, A. Choi, C.T. Chu, J. Chung, P.G. Clarke, R.S. Clark, S.G. Clarke,, J.L. Cleveland, P. Codogno, M.I. Colombo, A. Coto-Montes, J.M. Cregg,ervo, J. Debnath, F. Demarchi, P.B. Dennis, P.A. Dennis, V. Deretic, R.J.h, S.F. Di, J.F. Dice, M. Diglia, S. Nesh-Kumar, C.W. Distelhorst, M.ri-Mergny, F.C. Dorsey,W. Droge,M. Dron,W.A. Dunn Jr., M. Duszenko,sa, Z. Elazar, A. Esclatine, E.L. Eskelinen, L. Fesus, K.D. Finley, J.M., J. Fueyo, K. Fujisaki, B. Galliot, F.B. Gao, D.A. Gewirtz, S.B. Gibson,, A.L. Goldberg, R. Gonzalez, C. Gonzalez-Estevez, S. Gorski, R.A. Got-Haussinger, Y.W. He, K. Heidenreich, J.A. Hill, M. Hoyer-Hansen, X. Hu,ang, A. Iwasaki,M. Jaattela,W.T. Jackson, X. Jiang, S. Jin, T. Johansen, J.U.

18 R. Clarke et al. / Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820

Jung, M. Kadowaki, C. Kang, A. Kelekar, D.H. Kessel, J.A. Kiel, H.P. Kim, A. Kim-chi, T.J. Kinsella, K. Kiselyov, K. Kitamoto, E. Knecht, M. Komatsu, E. Kominami,S. Kondo, A.L. Kovacs, G. Kroemer, C.Y. Kuan, R. Kumar, M. Kundu, J. Landry,M. Laporte, W. Le, H.Y. Lei, M.J. Lenardo, B. Levine, A. Lieberman, K.L. Lim, F.C.Lin, W. Liou, L.F. Liu, G. Lopez-Berestein, C. Lopez-Otin, B. Lu, K.F. Macleod,W. MaloMichelsM.N. MNaqvi, TOleinickD.H. PerProikas-K.M. RyaC. SchneA. SibirnK. SpaneSubaustN.J. TalbTermanR. TruanVaqueroJ. YahaloZabirnyof assay151175

[49] S. Yousener, H.Uapoptos

[50] S. Jin, Au2 (2006

[51] M.A. ParM. Graf,MDA-7/4 (2008

[52] A. YacouSarkar, IM.Graf,dependglioma c

[53] Y. Cheninducescancer c

[54] C.H. Yanin cytotActa Pha

[55] A. Guilloeven in

[56] A. Apel,sitizes r148514

[57] M. Li, Xunder a

[58] P. Boya,D. MetiKroeme(2005) 1

[59] M. CasteHorne, Jcatastroaneuplo

[60] I.B. Roniinduced4 (2001

[61] G.J. Kopthe mito

[62] D. Kandtacroce,Bucci, E165170

[63] G.P. DimM. Linskhuman(1995) 9

[64] W.E. WrimplicatCell Bio

[65] K.B. Bouulatorythe ster(2004) 4

[66] N. Kypriduring rtion, Ca

[67] M.F. El Estone ancells, Br

[68] V.T. Gaddy, J.T. Barrett, J.N. Delk, A.M. Kallab, A.G. Porter, P.V. Schoenlein,Mifepristone induces growth arrest, caspase activation, and apoptosis ofestrogen receptor-expressing, antiestrogen-resistant breast cancer cells, Clin.Cancer Res. 10 (2004) 52155225.

[69] W. Bursch, A. Ellinger, H. Kienzl, L. Torok, S. Pandey, M. Sikorska, R. Walker,S. HerI 164autopInbalediateogram.A. Qaacroad enh940

. Bjorkenmatophaol. 171Ichimminag me847Ichim2/A17Riggirthens in an41.P. Gomu, A.th estes, FAClarkppi, Snt tooc. NaClark

.E. Liptiestrman

.A. Camiller, EativeReimed:Fasmoxifs. 4 (2

A. Ellison, Msiduas. Tre. Brunhormpresstions,Clarkopertne anWangta sp28.ClarkolecuochemClark

sistandocriHuantwork2.

E. Gosglioneorton,ed triceptoatl. CaF. Hasrowin004) 1Wangmpon008),Zhan

otif-btegratrni, W. Martinet, K. Matsuoka, J. Mautner, A.J. Meijer, A. Melendez, P., G. Miotto, W.P. Mistiaen, N. Mizushima, B. Mograbi, I. Monastyrska,oore, P.I. Moreira, Y. Moriyasu, T. Motyl, C. Munz, L.O. Murphy, N.I..P. Neufeld, I. Nishino, R.A. Nixon, T. Noda, B. Nurnberg, M. Ogawa, N.L., L.J. Olsen, B. Ozpolat, S. Paglin, G.E. Palmer, I. Papassideri, M. Parkes,lmutter, G. Perry, M. Piacentini, R. Pinkas-Kramarski, M. Prescott, T.Cezanne, N. Raben, A. Rami, F. Reggiori, B. Rohrer, D.C. Rubinsztein,n, J. Sadoshima, H. Sakagami, Y. Sakai, M. Sandri, C. Sasakawa,M. Sass,ider, P.O. Seglen, O. Seleverstov, J. Settleman, J.J. Shacka, I.M. Shapiro,y, E.C. Silva-Zacarin, H.U. Simon, C. Simone, A. Simonsen, M.A. Smith,l-Borowski, V. Srinivas, M. Steeves, H. Stenmark, P.E. Stromhaug, C.S.e, S. Sugimoto, D. Sulzer, T. Suzuki, M.S. Swanson, I. Tabas, F. Takeshita,ot, Z. Talloczy, K. Tanaka, K. Tanaka, I. Tanida, G.S. Taylor, J.P. Taylor, A., G. Tettamanti, C.B. Thompson, M. Thumm, A.M. Tolkovsky, S.A. Tooze,t, L.V. Tumanovska, Y. Uchiyama, T. Ueno, N.L. Uzcategui, d.K. van I, E.C., T. Vellai, M.W. Vogel, H.G. Wang, P. Webster, J.W. Wiley, Z. Xi, G. Xiao,m, J.M. Yang, G. Yap, X.M. Yin, T. Yoshimori, L. Yu, Z. Yue, M. Yuzaki, O.k, X. Zheng, X. Zhu, R.L. Deter, Guidelines for the use and interpretations for monitoring autophagy in higher eukaryotes, Autophagy 4 (2008)., R. Perozzo, I. Schmid, A. Ziemiecki, T. Schaffner, L. Scapozza, T. Brun-. Simon, Calpain-mediated cleavage of Atg5 switches autophagy tois, Nat. Cell Biol. 8 (2006) 11241132.tophagy, mitochondrial quality control, and oncogenesis, Autophagy) 8084.k, A. Yacoub, D. Sarkar, L. Emdad, M. Rahmani, S. Spiegel, C. Koumenis,D.T. Curiel, S. Grant, P.B. Fisher, P. Dent, PERK-dependent regulation ofIL-24-induced autophagy in primary human glioma cells, Autophagy).b, M.A. Park, P. Gupta, M. Rahmani, G. Zhang, H. Hamed, D. Hanna, D..V. Lebedeva, L. Emdad,M. Sauane, N. Vozhilla, S. Spiegel, C. Koumenis,D.T. Curiel, S.Grant, P.B. Fisher, P.Dent, Caspase-, cathepsin-, andPERK-ent regulation of MDA-7/IL-24-induced cell killing in primary humanells, Mol. Cancer Ther. 7 (2008) 297313., E. Millan-Ward, J. Kong, S.J. Israels, S.B. Gibson, Oxidative stressautophagic cell death independent of apoptosis in transformed andells, Cell Death Differ. 15 (2008) 171182., Y.P. Yang, Z.H. Qin, Z.L. Gu, P. Reid, Z.Q. Liang, Autophagy is involvedoxic effects of crotoxin in human breast cancer cell line MCF-7 cells,rmacol. Sin. 28 (2007) 540548.n-Munos, M.X. van Bemmelen, P.G. Clarke, Autophagy can be a killerapoptosis-competent cells, Autophagy 2 (2006) 140142.I. Herr, H. Schwarz, H.P. Rodemann, A. Mayer, Blocked autophagy sen-esistant carcinoma cells to radiation therapy, Cancer Res. 68 (2008)94.. Jiang, D. Liu, Y. Na, G.F. Gao, Z. Xi, Autophagy protects LNCaP cellsndrogen deprivation conditions, Autophagy 4 (2008) 5460.R.A. Gonzalez-Polo, N. Casares, J.L. Perfettini, P. Dessen, N. Larochette,vier, D. Meley, S. Souquere, T. Yoshimori, G. Pierron, P. Codogno, G.r, Inhibition of macroautophagy triggers apoptosis, Mol. Cell Biol. 250251040.do, J.L. Perfettini, T. Roumier, A. Valent, H. Raslova, K. Yakushijin, D.. Feunteun, G. Lenoir, R.Medema,W. Vainchenker, G. Kroemer,Mitoticphe constitutes a special case of apoptosis whose suppression entailsidy, Oncogene 23 (2004) 43624370.nson, E.V. Broude, B.D. Chang, If not apoptosis, thenwhat? Treatment-senescenceandmitotic catastrophe in tumorcells,DrugResist.Updat.) 303313.s, B.A. Weaver, D.W. Cleveland, On the road to cancer: aneuploidy andtic checkpoint, Nat. Rev. Cancer 5 (2005) 773785.uc, A. Mittelman, R. Serpico, E. Sinigaglia, A.A. Sinha, C. Natale, R. San-M.G. Di Corcia, A. Lucchese, L. Dini, P. Pani, S. Santacroce, S. Simone, R.. Farber, Cell death: apoptosis versus necrosis, Int. J. Oncol. 21 (2002).ri, X. Lee, G. Basile, M. Acosta, G. Scott, C. Roskelley, E.E. Medrano,ens, I. Rubelj, O. Pereira-Smith, A biomarker that identies senescentcells in culture and in aging skin in vivo, Proc. Natl. Acad. Sci. U.S.A. 923639367.ight, O.M. Pereira-Smith, J.W. Shay, Reversible cellular senescence:ions for immortalization of normal human diploid broblasts, Mol.l. 9 (1989) 30883092.ker, T.C. Skaar, D.R. Fernandez, K.A. OBrien, R. Clarke, Interferon reg-factor-1 mediates the proapoptotic but not cell cycle arrest effects ofoidal antiestrogen ICI 182, 780 (Faslodex, Fulvestrant), Cancer Res. 640304039.anou, H.F. English, N.E. Davidson, J.T. Isaacs, Programmed cell deathegression of the MCF-7 human breast cancer following estrogen abla-ncer Res. 51 (1991) 162166.treby, Y. Liang, R.W.Wrenn, P.V. Schoenlein, Additive effect ofmifepri-d tamoxifen on apoptotic pathways in MCF-7 human breast cancer

east Cancer Res. Treat. 51 (1998) 149168.

R.ICof

[70] B.mpr

[71] MMan38

[72] GStauBi

[73] Y.Koin22

[74] Y.p6

[75] R.pasi33

[76] B.Zhbolin

[77] R.KodePr

[78] R.Manhu

[79] DMer

[80] T.antaRe

[81] P.streRe

[82] Nofexca

[83] R.prge

[84] Y.da10

[85] R.MBi

[86] R.reEn

[87] S.ne22

[88] P.tiNizreN

[89] R.G(2

[90] C.co(2

[91] Y.minmann, Active cell death induced by the anti-estrogens tamoxifen and384 in human mammary carcinoma cells (MCF-7) in culture: the rolehagy, Carcinogenesis 17 (1996) 15951607., S. Bialik, I. Sabanay, G. Shani, A. Kimchi, DAP kinase and DRP-1membrane blebbing and the formation of autophagic vesicles duringmed cell death, J. Cell Biol. 157 (2002) 455468.dir, B. Kwok, W.H. Dragowska, K.H. To, D. Le, M.B. Bally, S.M. Gorski,utophagy inhibition sensitizes tamoxifen-resistant breast cancer cellsancesmitochondrial depolarization, Breast Cancer Res. Treat. 3 (2008)3.oy, T. Lamark, A. Brech, H. Outzen, M. Perander, A. Overvatn, H.

rk, T. Johansen, p62/SQSTM1 forms protein aggregates degraded bygy and has a protective effect on huntingtin-induced cell death, J. Cell(2005) 603614.ura, T. Kumanomidou, Y.S. Sou, T. Mizushima, J. Ezaki, T. Ueno, E.mi, T. Yamane, K. Tanaka, M. Komatsu, Structural basis for sort-chanism of p62 in selective autophagy, J. Biol. Chem. 283 (2008)22857.ura, E. Kominami, K. Tanaka, M. Komatsu, Selective turnover of0/SQSTM1 by autophagy, Autophagy 4 (2008).ns, A. Zwart, N. Nehra, P. Agarwal, R. Clarke, The NFB inhibitorolide restores ICI 182, 780 (Faslodex; Fulvestrant)-induced apopto-tiestrogen resistant breast cancer cells, Mol. Cancer Ther. 4 (2005)

ez, R. Riggins, A.N. Shajahan, U. Klimach, A. Wang, A.C. Crawford, Y.Zwart, M. Wang, R. Clarke, Human X-Box binding protein-1 confersrogen independence and antiestrogen resistance in breast cancer cellSEB J. 21 (2007) 40134027.e, N. Brnner, B.S. Katzenellenbogen, E.W. Thompson, M.J. Norman, C.. Paik, M.E. Lippman, R.B. Dickson, Progression from hormone depen-hormone independent growth in MCF-7 human breast cancer cells,tl. Acad. Sci. U.S.A. 86 (1989) 36493653.e, N. Brnner, E.W. Thompson, P. Glanz, D. Katz, R.B. Dickson,pman, The inter-relationships between ovarian-independent growth,ogen resistance and invasiveness in the malignant progression ofbreast cancer, J. Endocrinol. 122 (1989) 331340.eron, J.C. Keen, J.M. Dixon, C. Bellamy, A. Hanby, T.J. Anderson, W.R.

ffective tamoxifen therapy of breast cancer involves both antiprolif-and pro-apoptotic changes, Eur. J. Cancer 36 (2000) 845851.r,D. Koczan,H.Muller, K. Friese,H.J. Thiesen, B.Gerber, Tumour Fas lig-ratio greater than 1 is an independent marker of relative resistance toen therapy in hormone receptor positive breast cancer, Breast Cancer002) R9., I.E. Smith, S. Detre, S.A. Burton, J. Salter, R. AHern, G.Walsh, S.R. John-. Dowsett, Reduced apoptosis and proliferation and increased Bcl-2 inl breast cancer following preoperative chemotherapy, Breast Cancerat. 48 (1998) 107116.ner, V. Boulay, A. Fojo, C.E. Freter, M.E. Lippman, R. Clarke, Acquisitionone-independent growth in MCF-7 cells is accompanied by increasedion of estrogen-regulated genes but without detectable DNA ampli-Cancer Res. 53 (1993) 283290.e, H.W. Ressom, A. Wang, J. Xuan, M.C. Liu, E.A. Gehan, Y. Wang, Theies of very high dimensional data spaces: implications for exploringd protein expression data, Nat. Rev. Cancer 8 (2008) 3749., D.J. Miller, R. Clarke, Approaches to working in high-dimensionalaces: gene expression microarrays, Br. J. Cancer 98 (2008) 1023

e, T.C. Skaar, K.B. Bouker, N. Davis, Y.R. Lee, J.N. Welch, F. Leonessa,lar and pharmacological aspects of antiestrogen resistance, J. Steroid. Mol. Biol. 76 (2001) 7184.e, N. Brnner, Acquired estrogen independence and antiestrogence in breast cancer: estrogen receptor-driven phenotypes? Trendsnol. Metab. 7 (1996) 2535.g, Genomics, complexity and drug discovery: insights from Booleanmodels of cellular regulation, Pharmacogenomics 2 (2001) 203

s, J.N. Ingle, S. Martino, N.J. Robert, H.B. Muss, M.J. Piccart, M. Cas-, D. Tu, L.E. Shepherd, K.I. Pritchard, R.B. Livingston, N.E. Davidson, L.E.A. Perez, J.S. Abrams, D.A. Cameron, M.J. Palmer, J.L. Pater, Random-al of letrozole following tamoxifen as extended adjuvant therapy inr-positive breast cancer: updated ndings from NCIC CTG MA. 17, J.ncer Inst. 97 (2005) 12621271.himoto, S. Kim, I. Shmulevich, W. Zhang, M.L. Bittner, E.R. Dougherty,g genetic regulatory networks from seed genes, Bioinformatics 202411247., J. Xuan, L. Chen, P. Zhao, Y.Wang, R. Clarke, E.Hoffman,Motif-directedent analysis for regulatory network inference, BMC Bioinform. 9doi:10.1186/1471-2105-9-S1-S21.g, J. Xuan, B.G. de Los Reyes, R. Clarke, H.W. Ressom, Networkased identication of transcription factor-target gene relationships bying multi-source biological data, BMC Bioinform. 9 (2008) 203.

R. Clarke et al. / Journal of Steroid Biochemistry & Molecular Biology 114 (2009) 820 19

[92] I.B. Jeffery, S.F.Madden, P.A.McGettigan,G. Perriere,A.C. Culhane,D.G.Higgins,Integrating transcription factor binding site informationwith gene expressiondatasets, Bioinformatics 23 (2007) 298305.

[93] L. Chen, J. Xuan, C. Wang, L.-M. Shih, Y. Wang, Z. Zhang, E.P. Hoffman, R.Clarke, Knowledge guided multi-scale independent component analysis forbiomark416.

[94] S. McKein prost(2006) 1

[95] Z. Gu, R.Leonessfactor-1ment bi62 (200

[96] K.B. BouWang, Rsor activof apopt

[97] K.B. BouR. Clarkfrequen6164.

[98] D. Ron, TClin. Inv

[99] K. Zhangway crit

[100] J.B. DuRunfoldestress, M

[101] H. Yoshiby ATF6transcri

[102] M. OgatM. TanigImaizumstress, M

[103] D.E. Feldcompon597605

[104] I.M. Clauproteincontaini

[105] H.C. LiouL.H. Glimto the H

[106] R. Sriburproteinulum, J.

[107] A.H. LeedisruptU.S.A. 10

[108] L. Ding,indepenRes. 31 (

[109] A.M.ReiE.M. Graentiatio

[110] I.M. Claucher, Inzipper pmouse e

[111] A.M. Reton, A. Srole in li152157

[112] D.Y. Waresponsand chaEndocri

[113] S. TozluLidereaugen recreal-tim110911

[114] C.L. Wilvak, H. YIRF-1, mmyelody

[115] H.W. vahumanbreast ceffects o

[116] T. MatsuT.M. KunNarendr

disruption of IRF-1 or IRF-2 results in abnormal type I IFN gene induction andaberrant lymphocyte development, Cell 75 (1993) 8397.

[117] S. Namiki, T. Nakamura, S. Oshima, M. Yamazaki, Y. Sekine, K. Tsuchiya, R.Okamoto, T. Kanai, M. Watanabe, IRF-1 mediates upregulation of LMP7 byIFN-gamma and concerted expression of immunosubunits of the proteasome,

BS LeH. LeeN-gammanBS LeH. Leee, InteducedanscriTamuyamamage995) 5Prost,sregu. Nozaakatsunaka,rrying4012. Tana.S. Lamcogenn fac

. Tanakzawa,rs IRF. MoroomotSanceosph

s via i7233Suk,mmarvicale NF-1531.A. Chgulati1235C. Skaancer,B. BourogenGu, J.roughve vs.8159K.M. Ktto, Tmor g(200Tu, Don comand, M.L. BoL. Seediate438. Yim

ron reowths. 23PizzoM. Albinterfsults i.F. Tamol. Ch. Yaron. Manthe IkBours.P. Meacol. 6. Naksitutivermon

.K. Bisduceder identication, BMC Bioinform. 9 (2008), doi:10.1186/1471-2105-9-

nzie, N. Kyprianou, Apoptosis evasion: the role of survival pathwaysate cancer progression and therapeutic resistance, J. Cell Biochem. 97832.Y. Lee, T.C. Skaar, K.B. Bouker, J.N. Welch, J. Lu, A. Liu, Y. Zhu, N. Davis, F.a, N. Brunner, Y. Wang, R. Clarke, Association of interferon regulatory, nucleophosmin, nuclear factor-kappaB, and cyclic AMP response ele-nding with acquired resistance to faslodex (ICI 182, 780), Cancer Res.2) 34283437.ker, T.C. Skaar, R. Riggins, D.S. Harburger, D.R. Fernandez, A. Zwart, A.. Clarke, Interferon regulatory factor-1 (IRF-1) exhibits tumor suppres-ities inbreast cancer associatedwith caspase activation and inductionosis, Carcinogenesis 26 (2005) 15271535.ker, T.C. Skaar, D.S. Harburger, R. Riggins, D.R. Fernandez, A. Zwart,e, The A4396G polymorphism in interferon regulatory factor-1 istly expressed in breast cancer, Cancer Genet. Cytogenet. 175 (2007)

ranslational control in the endoplasmic reticulum stress response, J.est. 110 (2002) 13831388., R.J. Kaufman, Theunfoldedprotein response: a stress signaling path-ical for health and disease, Neurology 66 (2006) S102S109.ose, A.B. Tam, M. Niwa, Intrinsic capacities of molecular sensors of thed protein response to sense alternate forms of endoplasmic reticulumol. Biol. Cell 17 (2006) 30953107.da, T. Matsui, A. Yamamoto, T. Okada, K. Mori, XBP1 mRNA is inducedand spliced by IRE1 in response to ER stress to produce a highly activeption factor, Cell 107 (2001) 881891.a, S. Hino, A. Saito, K. Morikawa, S. Kondo, S. Kanemoto, T. Murakami,uchi, I. Tanii, K. Yoshinaga, S. Shiosaka, J.A. Hammarback, F. Urano, K.i, Autophagy is activated for cell survival after endoplasmic reticulumol. Cell Biol. 26 (2006) 92209231.man, V. Chauhan, A.C. Koong, The unfolded protein response: a novelent of thehypoxic stress response in tumors,Mol. CancerRes. 3 (2005).ss, M. Chu, J.L. Zhao, L.H. Glimcher, The basic domain/leucine zipper

hXBP-1 preferentially binds to and transactivates CRE-like sequencesng an ACGT core, Nucleic Acids Res. 24 (1996) 18551864., M.R. Boothby, P.W. Finn, R. Davidon, N. Nabavi, L. Zeleznik, J.P. Ting,cher, A newmember of the leucine zipper class of proteins that binds

LA DR alpha promoter, Science 247 (1990) 15811584.i, S. Jackowski, K.Mori, J.W. Brewer, XBP1: a linkbetween theunfoldedresponse, lipid biosynthesis, and biogenesis of the endoplasmic retic-Cell Biol. 167 (2004) 3541., N.N. Iwakoshi, K.C. Anderson, L.H. Glimcher, Proteasome inhibitorsthe unfolded protein response in myeloma cells, Proc. Natl. Acad. Sci.0 (2003) 99469951.J. Yan, J. Zhu, H. Zhong, Q. Lu, Z. Wang, C. Huang, Q. Ye, Ligand-dent activation of estrogen receptor alpha by XBP-1, Nucleic Acids2003) 52665274.mold,N.N. Iwakoshi, J.Manis, P. Vallabhajosyula, E. Szomolanyi-Tsuda,vallese, D. Friend, M.J. Grusby, F. Alt, L.H. Glimcher, Plasma cell differ-n requires the transcription factor XBP-1, Nature 412 (2001) 300307.ss, E.M. Gravallese, J.M. Darling, F. Shapiro, M.J. Glimcher, L.H. Glim-situ hybridization studies suggest a role for the basic regionleucinerotein hXBP-1 in exocrine gland and skeletal development duringmbryogenesis, Dev. Dyn. 197 (1993) 146156.imold, A. Etkin, I. Clauss, A. Perkins, D.S. Friend, J. Zhang, H.F. Hor-cott, S.H. Orkin, M.C. Byrne, M.J. Grusby, L.H. Glimcher, An essentialver development for transcription factor XBP-1, Genes Dev. 14 (2000).ng, R. Fulthorpe, S.N. Liss, E.A. Edwards, Identication of estrogen-ive genes by complementary deoxyribonucleic acid microarrayracterization of a novel early estrogen-induced gene: EEIG1, Mol.nol. 18 (2004) 402411., I. Girault, S. Vacher, J. Vendrell, C. Andrieu, F. Spyratos, P. Cohen, R., I. Bieche, Identication of novel genes that co-cluster with estro-

eptor alpha in breast tumor biopsy specimens, using a large-scalee reverse transcription-PCR approach, Endocr. Relat. Cancer 13 (2006)20.lman, C.E. Sever, M.G. Pallavicini, H. Harada, N. Tanaka, M.L. Slo-amamoto, K. Harada, T.C. Meeker, A.F. List, T. Taniguchi, Deletion ofapping to chromosome 5q31.1, in human leukemia and preleukemicsplasia, Science 259 (1993) 965971.n den Berg, W.J. Leahey, M. Lynch, R. Clarke, J. Nelson, Recombinantinterferon alpha increases oestrogen receptor expression in humanancer cells (ZR-75-1) and sensitises them to the anti-proliferativef tamoxifen, Br. J. Cancer 55 (1987) 255257.yama, T. Kimura, M. Kitagawa, K. Pfeffer, T. Kawakami, N. Watanabe,dig, R. Amakawa, K. Kishihara, A.Wakeham, J. Potter, C.L. Furlonger, A.an, H. Suzuki, P.S. Ohashi, C.J. Paige, T. Taniguchi, T.W. Mak, Targeted

FE[118] S.

IFhuFE

[119] S.Leintr

[120] T.suda(1

[121] S.dy

[122] HNTaca12

[123] NMontio

[124] NAiso

[125] Apr

[126] J.phsi33

[127] K.gacetiv13

[128] Wre35

[129] T.ca

[130] K.st

[131] Z.thsi15

[132] P.botu23

[133] Y.mgl

[134] MV.m87

[135] J.HfegrRe

[136] E.S.ofre

[137] WBi

[138] AMof

[139] V.Mm

[140] Hstho

[141] Dintt. 579 (2005) 27812787., J.W. Kim, S.H. Oh, Y.J. Kim, S.B. Rho, K. Park, K.L. Park, J.H. Lee,ma/IRF-1-induced p27kip1 down-regulates telomerase activity and

telomerase reverse transcriptase expression in human cervical cancer,tt. 579 (2005) 10271033., J.W. Kim, H.W. Lee, Y.S. Cho, S.H. Oh, Y.J. Kim, C.H. Jung, W. Zhang, J.H.rferon regulatory factor-1 (IRF-1) is a mediator for interferon-gammaattenuation of telomerase activity and human telomerase reverse

ptase (hTERT) expression, Oncogene 22 (2003) 381391.ra, M. Ishihara, M.S. Lamphier, N. Tanaka, I. Oishi, S. Alzawa, T. Mat-, T.W. Mak, S. Taki, T. Taniguchi, An IRF-1-dependent pathway of DNA-induced apoptosis in mitogen-activated T lymphocytes, Nature 37696599.C.O.C. Bellamy,D.S. Cunningham,D.J.Harrison,AlteredDNArepair andlation of p53 in IRF-1 null hepatocytes, FASEB J. 12 (1998) 181188.wa, E. Oda, K. Nakao, M. Ishihara, S. Ueda, T. Yokochi, K. Ogasawara, Y.ru, K. Hioki, S. Aizawa, T. Ishikawa,M. Katsuki, T. Muto, T. Taniguchi, N.Loss of transcription factor IRF-1 affects tumor susceptibility in micethe Ha-ras transgene or nullizygosity for p53, Genes Dev. 13 (1999)45.ka, M. Ishihara, M. Kitagawa, H. Harada, T. Kimura, T. Matsuyama,phier, S. Aizawa, T.W. Mak, T. Taniguchi, Cellular commitment to

e-induced transformationor apoptosis is dependent on the trans