Dc

TGAa

b

c

d

e

f

g

a

ARRA

KACBIq

I

fipmTd

aS

h0

Acta Histochemica 116 (2014) 1178–1184

Contents lists available at ScienceDirect

Acta Histochemica

jo ur nal homepage: www.elsev ier .de /ac th is

ynamic dialog between cytokeratin 18 and annexin A1 in breastancer: A transcriptional disequilibrium

haise G. Araujoa,∗, Karina Marangonia, Rafael M. Rochab, Yara C.P. Maiac,alber R. Araujoa, Tânia M. Alcântard, Patrícia T. Alvesa, Luanda Calábriae,driana F. Neves f, Fernando A. Soaresb, Luiz R. Goularta,g

Laboratory of Nanobiotechnology, Institute of Genetics and Biochemistry, Federal University of Uberlandia, Uberlandia, MG, BrazilAC Camargo Cancer Hospital, São Paulo, SP, BrazilSchool of Medicine, Federal University of Uberlandia, Uberlandia, MG, BrazilDepartment of Pathology, Clinical Hospital of Uberlandia, Federal University of Uberlandia, Uberlandia, MG, BrazilObstetrics Division, Internal Medicine, University Hospital, Federal University of Uberlandia, Uberlandia, MG, BrazilLaboratory of Genetics and Biotechnology, Federal University of Goias, Catalao, GO, BrazilDepartment of Medical Microbiology and Immunology, University of California Davis, Davis, CA, USA

r t i c l e i n f o

rticle history:eceived 11 April 2014eceived in revised form 19 June 2014ccepted 23 June 2014

eywords:nnexin A1ytokeratin 18reast cancer

mmunohistochemistryPCR

a b s t r a c t

Cytokeratins (CKs) constitute the cytoskeletal network and are regulated by post-translational modi-fications, acting not only as a mechanical support, but also in cell signaling and regulatory processes.Signaling is mediated by CK-associated proteins, such as Annexin A1 (ANXA1), a ligand of the CK18/CK8complex. ANXA1 has a pivotal role in cellular and immunological responses, and together with CK18 havebeen implicated in several processes related to malignant transformation in breast cancer (BC). Our aimwas to demonstrate how their interaction might be linked to BC development. We investigated tran-script levels, protein expression and distribution for both targets in breast tissues of 92 patients (42 BCsand 50 benign diseases) using qPCR and immunohistochemistry, respectively. ANXA1 and CK18 mRNAswere inversely correlated, and their ratio in each TNM stage significantly differentiated BC from benigndiseases (OR = 5.62). These differences did not mirror tissue protein levels, but a significant dichotomousprotein distribution in tumor tissues was observed, differing from the expected co-localization observed

during cell homeostasis. The disequilibrium of transcriptional levels between ANXA1/CK18 and alter-ations in their tissue distribution are present either in initial events or tumor progression, which suggesta critical event in BC. The broken dialog between ANXA1 and CK18 in normal breast tissues may playa critical role in BC development, and together may be used as combined targets for BC diagnostics.

ntroduction

The intermediate filaments (IF) are cytoskeletal structures thatunction not only as mechanical support, but also are dynamicallynvolved in cell signaling pathways, and are actively regulated byhosphorylation and other posttranslational modifications, deter-

ining either cell survival or apoptosis (Pallari and Eriksson, 2006).

he IF network in simple glandular epithelial cells consists pre-ominantly of heterotypic complexes of cytokeratin 8 (CK8) and

∗ Corresponding author at: Laboratory of Nanobiotechnology, Institute of Geneticsnd Biochemistry, Federal University of Uberlandia, Campus Umuarama, Bloco 2E,ala 248, 38400-902, Uberlandia, MG, Brazil.

E-mail address: tgaraujo@ingeb.ufu.br (T.G. Araujo).

ttp://dx.doi.org/10.1016/j.acthis.2014.06.008065-1281/© 2014 Elsevier GmbH. All rights reserved.

© 2014 Elsevier GmbH. All rights reserved.

cytokeratin 18 (CK18). In cancer, their persistent expression isaccompanied by changes in cell morphology and alterations in theIF network (Ditzel et al., 2002).

Among several important signaling molecules, we have focusedon Annexin A1 (ANXA1) that has been shown to be an IF-associatedprotein, specifically linked to CK8 and CK18 (Croxtall et al., 1998).ANXA1 is a calcium- and phospholipid-binding protein involvedin many membrane-related events, such as membrane organiza-tion domains and membrane-cytoskeleton signaling (Gerke et al.,2005). ANXA1 also seems to be a key molecule involved in thedevelopment of many types of cancer, and a modulator of the

epithelial–mesenchymal transition (EMT) associated with highlyinvasive breast carcinomas (BC) (de Graauw et al., 2010).

ANXA1, however, is a part of a complex and unknown network,and besides its function as membrane domains organizer, it also

ochem

pI(Mfmt

amwAdbiCaMmedtte

fstaethaltrttri

M

S

v

Fmd

T.G. Araujo et al. / Acta Hist

rovides recruitment platforms for proteins with which it interacts.nterestingly, cytoskeletal disruptive drugs have anti-inflammatoryCronstein et al., 1995) and anti-proliferative roles (Dustin, 1963;

iquel et al., 1996; Tiozzo et al., 1996). Similarly ANXA1 alsounctions as an anti-inflammatory, anti-proliferative and apoptotic

olecule (Flower and Rothwell, 1994; Perretti and Gavins, 2003),hrough post-translational modifications (Solito et al., 2006).

It has been shown that expression of CKs is tightly regulatednd correlates with the origin of the cells in the ducts of mam-ary glands (Ciocca et al., 2006), and CK8/CK18 are co-localizedith ANXA1 in the human alveolar squamous epithelial cell line549 (Croxtall et al., 1998); however, it is not known their jointistribution in benign and breast cancer human tissues. CKs haveeen recognized for more than 20 years as epithelial markers

n histopathological diagnostics, in which basal like cells expressK5, CK6, and CK14 and/or CK17 (basal/myoepithelial phenotype)nd luminal-like express CK8, CK18, and CK19 (Trask et al., 1990;alzahn et al., 1998; Linder et al., 2004). Two key observations ofany studies is that CK expression changes rapidly during differ-

ntiation, tissue injury, and metastasis (Magin et al., 2007), andespite the identification of ANXA1 as one of several cellular pro-eins that are differentially expressed during the progression ofumors, its role in carcinogenesis has not yet been elucidated (Shent al., 2006).

Therefore, we hypothesize that both CK18 and ANXA1 may play aar more dynamic and joint role in tumor development and progres-ion than thought previously, and we believe that the link betweenhem may provide novel clues and approaches to the current char-cterization of BC (Parton et al., 2001; Gerke et al., 2005; Kakehashit al., 2009; Ratkaj et al., 2010). Considering that most BCs areightly linked to the CK18 profile in tissues, and that ANXA1 mayave considerable relevance in the cell signaling process through itsssociation with CK18, we characterized the transcript and proteinevels of both molecules, as well as their distribution in benign andumor tissues, and showed striking differences in their transcriptatios and protein distribution in tumor tissues. We present impor-ant evidence of concerted action of both molecules in benign andumor tissues, and the loss of homeostasis between them may rep-esent a link of important cellular and molecular events for tumornitiation and progression.

aterials and methods

tudy design and sample collection

Total of 92 patients were selected from the Obstetric Ser-ice at University Hospital at Federal University of Uberlandia,

ig. 1. ANXA1 and CK18 mRNA expression levels, their ratios and odds ratios obtained iessenger RNA (mRNA) in breast tissues and benign samples and ANXA1/CK18 transcrip

eviation. In (B) the odds ratios estimated according to the detection limit for each marke

ica 116 (2014) 1178–1184 1179

and grouped into two groups of patients: 42 BCs and 50with benign breast diseases (BBD). The surgical procedures ofuntreated patients consisted of axillary dissection, either by rad-ical mastectomy or quadrantectomy, depending on the size of thetumor.

The average age of the patients investigated was 47.7 years(range 31–89 years) for BC group and 46.8 years (range 18–80 years)for BBD group. There were 12 (29%), 18 (43%), 5 (12%), and 7 (16%)breast tumors classified as TNM stages T1, T2, T3 and T4, respec-tively. The histological grading according to the Nottingham systemconsisted of grade I (GI) in 4 (10%), GII in 24 (57%), and GIII in 14(33%).

Hormone receptors, estrogen receptor (ER) was positive in 28(67%), negative in 9 (21%) and not evaluated in 5 (12%); proges-terone receptor (PgR) was positive in 27 (64%), negative in 10 (24%)and not analyzed in 5 (12%). HER-2 status was considered as posi-tive (score 3+) and negative (score 0–1+); scores 2+ were excludedfrom the analyses. HER-2 status was positive in 12 cases (29%),negative in 25 (60%) and score 2+ or not analyzed in another 5(11%).

The Ethics Committee of the Institutional Research Boardapproved all procedures (# 176/2008) and informed consent wasobtained from all participants.

Quantitative RT-PCR and immunohistochemistry

For transcriptional analysis, RNA was extracted using the Trizolreagent (Invitrogen, Life Technologies, Carlsbad, CA, USA) accord-ing to the manufacturer’s recommendations and the CK18, ANXA1,and B2M transcripts were analyzed by qPCR with SybrGreen detec-tion in an ABI PRISM 7300 (Applied Biosystems, Foster City, CA,USA). The primers’ sequences were: GATTCAGATGCCAGGGCCT andCACTCTGCGAAGTTGTGGAT for ANXA1; GCTCTGGGTTGACCGTGGand GTGGTGCTCTCCTCAATCTGC for CK18; CCTGCCGTGTGAAC-CATGT and GCGGCATCTTCAAACCTCC for B2M.

For protein detection of ANXA1 and CK18 and their tissuedistribution, immunohistochemistry were carried out in TissueMicroArrays (TMA) by using monoclonal antibodies for ANXA1(Becton Dickinson, Franklin Lakes, NJ, USA; 1:250) and for CK18(Cell Marque, Rocklin, CA, USA; 1:50), and the procedures wereperformed according to standard protocols with EnVision PlusSys-tem HRP (Dako, Glostrup, Denmark). The final scores were obtainedaccording to immunostaining intensity in epithelial cells and were

designated as negative (score 0) or positive (score 1–3). The analy-sis was carried out by three observers (FAS, RMR, and TMA) and thesamples were scored blinded with respect to clinical patient data.In case of discrepant recording, a consensus score was used.

n a comparison between BC and BBD. Relative quantification of ANXA1 and CK18ts ratio (A) in which the values above the bars represent ANXA1/CK18 ± standardr. The values were calculated for breast cancer vs. benign breast disease. *P < 0.05.

1 ochemica 116 (2014) 1178–1184

S

rcStawI

R

Ad

bwCdimos(

Abois2

Dd

APamtsiwci

AoeirfiCt(

trCtb

Fig. 2. Transcriptional quantification of ANXA1 and CK18 and protein expressionin benign samples and according to TNM system. In (A) relative mRNA levels ofANXA1 and CK18 are shown. The values above the bars represent ANXA1/CK18ratio for each histopathological classification ± standard deviation. The transcriptswere significantly different in pT3 classification compared to all others stages.Immunohistochemical staining for annexin A1 (ANXA1) and cytokeratin 18 (CK18)

180 T.G. Araujo et al. / Acta Hist

tatistical analysis

For statistical analysis we used Mann–Whitney test, Kendall’sank correlation (�), McNemar’s test, and Pearson’s correlation forlinical parameters; patient age at diagnosis, TNM system staging,carff-Bloom–Richardson (SBR) grading, and hormonal status. Sta-istical significance was considered when P < 0.05. The statisticalnalyses were carried out using GraphPad Prism 5 (GraphPad Soft-are, La Jolla, CA, USA) and SPSS version 17.0 (SPSS, IBM, Chicago,

L, USA).

esults

NXA1 and CK18 gene expression levels are inisequilibrium in breast cancer

No difference was observed for ANXA1 transcripts betweenreast cancer and benign tissues, but lower CK18 mRNA levelsere detected in BC (P = 0.02) (Fig. 1A) Relative quantification of

K18 mRNA levels was 2.08-fold higher in BBD than in BC. Aetection limit was estimated for CK18 and ANXA1 gene accord-

ng to maximal differences between BC and BBD groups in theiredians and percentiles. These limits were used to calculate the

dds ratios values. The cut-off value for the relative levels of tran-cript revealed a down regulation of CK expression in BC samplesFig. 1B).

However, the most striking differences were observed forNXA1:CK18 ratios in breast cancer stages when compared toenign tissues (P < 0.05) (Fig. 1A). Considering the significant ratiosbserved between each tumor stage and BBD (>1.5 or <0.5), whichndicated disequilibrium between markers, we have obtained aignificant odds ratio of 5.62-fold toward BC occurrence (CI95%.1–15.1, P = 0.0009).

ichotomous ANXA1 and CK18 expression and distributionuring cancer development and progression

A significant moderate negative correlation between CK18 andNXA1 protein levels was found in breast samples (r = −0.31,

= 0.01). Interestingly, although CK18 transcript levels are gener-lly reduced across tumor stages, its protein expression becomesore intense as tumor progresses, suggesting a loss of cell architec-

ure, as observed in Fig. 2. Additionally, CK18 transcripts presentedignificant higher levels in pT3 tissues (Fig. 2A). It was not verifiedf there are modifications in ANXA1/CK18 ratio in benign tissues,

ith a concert expression of both markers. However, during can-er development and progression this proportion was altered, evenn lower tumor staging.

Analysis of the concomitant expression and distribution ofNXA1 and CK18 in benign tissues demonstrated co-localizationf both in ductal and glandular epithelial cells (Fig. 2B). How-ver, during malignant transformation, the ANXA1 expressionn BC tissues was reduced in the malignant epithelium, butetained in the myoepithelium. Other stromal cells such asbroblasts were also positive. Furthermore, the antibody againstK18 reacted with epithelial cells in both benign and tumorissues, but with increasing intensities as tumor progressesFig. 2B–F).

Considering BC molecular subtypes (Fig. 3), ANXA1 and CK18ranscript levels presented similar behavior, although with greater

atio discrepancies in triple-negative BCs. However, the ANXA1 andK18 expression in tissues were differentially expressed and dis-ributed across molecular subtypes, with greater differences in theasal-like phenotype, in which ANXA1 was highly expressed, while

in epithelial cells from BBD (B) and in breast tumor cells classified as pT1(C); pT2(D); pT3 (E) and pT4 (F). BBD, benign breast disease; BC, breast cancer. *P < 0.05.

T.G. Araujo et al. / Acta Histochemica 116 (2014) 1178–1184 1181

Fig. 3. Transcripts and protein levels of ANXA1 and CK18 according to molecular subtypes. In (A) relative mRNA levels are shown. Immunostaining for ANXA1 and CK18breast cancer tissues classified according to molecular subtypes characterization are presented: luminal A (B); luminal B (C); triple-negative (D) and Her-2 overexpression(E). CK18 transcripts were significantly lower in Her-2 tumors compared to other molecular subtypes. *P < 0.05; TNBC: triple-negative breast cancer.

1 ochemica 116 (2014) 1178–1184

Ct

pecttA(

D

dawAictomq

rrpttsbw(saiere

f(aabtba2tsws

Bsdnee

pwp

Fig. 4. Relative quantification of messenger RNA (mRNA) and immunohistochem-istry of ANXA1 and CK18 according to tumor grade. In (A) transcripts levels of ANXA1and CK18 in benign, grade 1(G1), grade 2 (G2), and grade 3 (G3) tissues are shown.

182 T.G. Araujo et al. / Acta Hist

K18 was down-regulated, showing an epithelial–mesenchymalransition (Fig. 3D).

Alteration in ANXA1/CK18 ratio also characterized the tumorrogression and cell differentiation (Fig. 4). Transcriptional lev-ls of CK18 were higher in benign tissues compared to breastancer samples and it was observed a down-regulation of ANXA1ranscripts according to histological grade (Fig. 4A). CK18 pro-ein staining was lower in epithelial cells as tumor progresses andNXA1 immunostaining was higher in myoepithelial cells in G3

Fig. 4C–E).

iscussion

The cell core machinery for malignant cell transformation andevelopment of breast cancer is intrinsically associated with imbal-nce between proliferation and apoptosis (Parton et al., 2001),ith an intimate link between cell architecture and signaling.NXA1 and CK18 are two important molecules deeply involved

n these two processes, and our study demonstrated its con-omitant association in breast cancer. We presented, for the firstime, the importance of ANXA1 and CK18 balance in BC devel-pment and progression, not as single markers but as molecularultifaceted events with transcriptional and translational conse-

uences.Differential expression of ANXA1 in human cancers has been

eviewed elsewhere (Debret et al., 2003), generally showingeduced protein levels in head and neck cancer, esophageal androstate cancer, and in B-cell non-Hodgkin’s lymphoma. However,emporal and spatial changes in expression of ANXA1 in breastumors are not well defined and still controversial. ANXA1 expres-ion in ductal cells from both primary breast cancer tissue andreast cancer with lung metastases are shown to be increasedhen compared to ductal luminal cells from normal breast tissue

Wallner et al., 1986). On the other hand, reduced ANXA1 expres-ion has also been observed in both ductal carcinoma in situ (DCIS)nd invasive breast cancer tissue (Bruggers et al., 1999). In ournvestigation, we did not detect ANXA1 expression differences inither cancer types or stages, except in pT3, which was significantlyeduced, probably indicating a set point for metastasis (Maschlert al., 2010).

Our interest in this joint analysis with CK18 relies on theact that ANXA1 is a specific ligand of both CK8 and CK18Croxtall et al., 1998). These intermediate filaments (cytoker-tins) have been used as molecular markers in diagnostic andn aberrant expression of individual CKs results in abnormal cellehavior (Willipinski-Stapelfeldt et al., 2005). We have shownhat CK18 transcript levels were able to distinguish betweenenign and tumors tissues, and its down-regulation was gener-lly correlated with breast cancer development (Woelfle et al.,004). This is corroborated by our immunohistochemical data andranscriptional profiles of CK18 across disease stages, which is con-tantly reduced during BC progression, except at the pT3 stage,ith a profound difference in ANXA1:CK18 ratios in cancer tis-

ues.Considering ANXA1 transcript levels across tumor stages and

BD, we have shown a consistent expression pattern, and besides amall reduction in pT3, its expression alone did not explain tumorevelopment. However, its tissue distribution and expression areot correlated with their mRNA levels, and showed not only differ-ntial protein expression, but also presented a redistribution frompithelial to stromal compartments.

Our data for basal-like subtypes and pT3 and pT4 expressionatterns suggest a switch of expression pattern of CK18 and ANXA1,ith a bimodal behavior that seems to be essential for an invasivehenotype. During cancer development, specifically at pT1 and pT2

ANXA/CK18 ratio is annotated above the bars, with standard deviation (±) and in(B) is demonstrated the odds ratio for ANXA/CK18 ratio calculated for histologicalgrade vs benign breast disease. Breast cancer tissues classified as G1, G2, and G2and staining against ANXA1 and CK18 antibodies are represented in (C), (D) and (E),respectively. *P < 0.05 and **P < 0.005.

ochem

stgpasaera

dtdCtldaiA

nsgoiwsdtmTqci

A

lttcETmR

R

B

C

C

C

T.G. Araujo et al. / Acta Hist

tages, CK18 expression was down-regulated, suggesting a dysfunc-ional cytoskeleton assembly without a concerted signaling at thelandular epithelial cell level, which is one of the first BC events, arofile that was maintained at pT4, after the CK18 expression switcht the pT3. However, these instant changes in the CK18 expres-ion during malignant progression may be critical to EMT switchnd to tumor cells to become mobile and invasive, with alteredxpression ratio of ANXA1/CK18. This is also corroborated by theeduced expression of ANXA1 in pT3 that is associated with EMTnd metastasis (Maschler et al., 2010).

The intimate connection between ANXA1 and CK18 has beenemonstrated by their co-localization in normal epithelial breastissue, with a striking dichotomous redistribution during cancerevelopment. ANXA1 was found in the stromal of tumor tissues, theK18 was mainly found in the epithelium. Therefore, different fromissue homeostasis and mechanical equilibrium in normal physio-ogical conditions, the disrupted tissue structure and organizationuring tumor evolution has led to unbalanced physical forces andltered properties of tumor components (Yu et al., 2011), whichs evidenced in this study by a disequilibrium between CK18 andNXA1.

In brief, we showed that benign tissues present a homoge-eous distribution of both CK18 and ANXA1 in the ducts withimilar transcript levels, but during cancer development and pro-ression this stable transcription is lost. This is the first descriptionf the concomitant protein distribution of both ANXA1 and CK18n cancer showing a surprising discrepancy in their distribution

ith different localizations, which is further corroborated by theignificant alterations in their transcriptional expression ratiosuring cancer progression (ANXA1:CK18), supporting the notionhat multiple markers balance are necessary to maintain the nor-

al breast, and modifications may lead to tumor development.he combined use of both molecules in a reverse transcription-PCR assay has proved to be an important diagnostic strategy, andould be used as an auxiliary tool for disease diagnostics and stag-ng.

cknowledgements

The authors would like to thank the patients and their fami-ies for the direct collaboration in this work, the medical staff fromhe Gynecology Division of the University Hospital for providinghe biological samples and the clinical parameters, and the finan-ial support by the Coordination for the Improvement of Higherducation Personnel (CAPES) National Council for Scientific andechnological Development (CNPq), Coordination for the Improve-ent of Higher Education Personnel (CAPES), and Foundation for

esearch Support of Minas Gerais (FAPEMIG).

eferences

ruggers CS, Fults D, Perkins SL, Coffin CM, Carroll WL. Coexpres-sion of genes involved in apoptosis in central nervous systemneoplasms. J Pediatr Hematol Oncol 1999;21:19–25.

iocca V, Bombonati A, Gatalica Z, Di Pasquale M, Milos A, Ruiz-Orrico A, et al. Cytokeratin profiles of male breast cancers.Histopathology 2006;49:365–70.

ronstein BN, Molad Y, Reibman J, Balakhane E, Levin RI, Weiss-mann G. Colchicine alters the quantitative and qualitativedisplay of selectins on endothelial cells and neutrophils. J ClinInvest 1995;96:994–1002.

roxtall JD, Wu HL, Yang HY, Smith B, Sutton C, Chang BI,et al. Lipocortin 1 co-associates with cytokeratins 8 and 18in A549 cells via the N-terminal domain. Biochim Biophys1998;1401:39–51.

ica 116 (2014) 1178–1184 1183

de Graauw M, van Miltenburg MH, Schmidt MK, Pont C, Lalai R,Kartopawiro J, et al. Annexin A1 regulates TGF-beta signalingand promotes metastasis formation of basal-like breast cancercells. Proc Natl Acad Sci USA 2010;107:6340–5.

Debret R, El Btaouri H, Duca L, Rahman I, Radke S, Haye B, et al.Annexin A1 processing is associated with caspase-dependentapoptosis in BZR cells. FEBS Lett 2003;546:195–202.

Ditzel HJ, Strik MC, Larsen MK, Willis AC, Waseem A, Kejling K,et al. Cancer-associated cleavage of cytokeratin 8/18 heterotypiccomplexes exposes a neoepitope in human adenocarcinomas. JBiol Chem 2002;277:21712–22.

Dustin P Jr. New aspects of the pharmacology of antimitotic agents.Pharmacol Rev 1963;15:449–80.

Flower RJ, Rothwell NJ. Lipocortin-1: cellular mechanisms and clin-ical relevance. Trends Pharmacol Sci 1994;15:71–6.

Gerke V, Creutz CE, Moss SE. Annexins: linking Ca2+ signalling tomembrane dynamics. Nat Rev Mol Cell Biol 2005;6:449–61.

Kakehashi A, Inoue M, Wei M, Fukushima S, Wanibuchi H. Cytoker-atin 8/18 overexpression and complex formation as an indicatorof GST-P positive foci transformation into hepatocellular carci-nomas. Toxicol Appl Pharm 2009;238:71–9.

Linder S, Havelka AM, Ueno T, Shoshan MC. Determining tumorapoptosis and necrosis in patient serum using cytokeratin 18 asa biomarker. Cancer Lett 2004;214:1–9.

Magin TM, Vijayaraj P, Leube RE. Structural and regulatory func-tions of keratins. Exp Cell Res 2007;313:2021–32.

Malzahn K, Mitze M, Thoenes M, Moll R. Biological and prog-nostic significance of stratified epithelial cytokeratins ininfiltrating ductal breast carcinomas. Virchows Arch 1998;433:119–29.

Maschler S, Gebeshuber CA, Wiedemann EM, Alacakaptan M,Schreiber M, Custic I, et al. Annexin A1 attenuates EMTand metastatic potential in breast cancer. EMBO Mol Med2010;2:401–14.

Miquel K, Pradines A, Favre G. Farnesol and geranylgeraniolinduce actin cytoskeleton disorganization and apoptosis in A549lung adenocarcinoma cells. Biochem Biophys Res Commun1996;225:869–76.

Pallari HM, Eriksson JE. Intermediate filaments as signal-ing platforms. Sci STKE: Signal Transduct Knowl Environ2006;2006:pe53.

Parton M, Dowsett M, Smith I. Studies of apoptosis in breast cancer.BMJ 2001;322:1528–32.

Perretti M, Gavins FN. Annexin 1: an endogenous anti-inflammatory protein. News Phys Sci 2003;18:60–4.

Ratkaj I, Stajduhar E, Vucinic S, Spaventi S, Bosnjak H, Pavelic K,et al. Integrated gene networks in breast cancer development.Funct Integr Genomics 2010;10:11–9.

Shen DJ, Nooraie F, Elshimali Y, Lonsberry V, He JB, Bose S, et al.Decreased expression of annexin A1 is correlated with breastcancer development and progression as determined by a tissuemicroarray analysis. Hum Pathol 2006;37:1583–91.

Solito E, Christian HC, Festa M, Mulla A, Tierney T, Flower RJ, et al.Post-translational modification plays an essential role in thetranslocation of annexin A1 from the cytoplasm to the cell sur-face. FASEB J 2006;20:1498–500.

Tiozzo R, Monti D, Straface E, Capri M, Croce MA, Rainaldi G, et al.Antiproliferative activity of 3-aminobenzamide in A431 carci-noma cells is associated with a target effect on cytoskeleton.Biochem Biophys Res Commun 1996;225:826–32.

Trask DK, Band V, Zajchowski DA, Yaswen P, Suh T, Sager R. Ker-atins as markers that distinguish normal and tumor-derived

mammary epithelial-cells. Proc Natl Acad Sci USA 1990;87:2319–23.

Wallner BP, Mattaliano RJ, Hession C, Cate RL, Tizard R, Sin-clair LK, et al. Cloning and expression of human lipocortin, a

1 ochem

W of human breast cancer. Clin Cancer Res 2004;10:2670–4.

184 T.G. Araujo et al. / Acta Hist

phospholipase A2 inhibitor with potential anti-inflammatoryactivity. Nature 1986;320:77–81.

illipinski-Stapelfeldt B, Riethdorf S, Assmann V, Woelfle U, Rau

T, Sauter G, et al. Changes in cytoskeletal protein compositionindicative of an epithelial–mesenchymal transition in humanmicrometastatic and primary breast carcinoma cells. Clin CancerRes 2005;11:8006–14.

ica 116 (2014) 1178–1184

Woelfle U, Sauter G, Santjer S, Brakenhoff R, Pantel K. Down-regulated expression of cytokeratin 18 promotes progression

Yu H, Mouw JK, Weaver VM. Forcing form and function: biome-chanical regulation of tumor evolution. Trends Cell Biol2011;21:47–56.