RESEARCH ARTICLE

Overexpression of HER-2/neu protein attenuates the oxidativesystemic profile in women diagnosed with breast cancer

Vanessa J. Victorino & Fernanda C. Campos & Ana C. S. A. Herrera &

Andréa N. Colado Simão & Alessandra L. Cecchini &Carolina Panis & Rubens Cecchini

Received: 24 September 2013 /Accepted: 5 November 2013# International Society of Oncology and BioMarkers (ISOBM) 2013

Abstract About 20 % of breast cancer patients over-expressthe human epidermal growth factor receptor-2 (HER2), whichis associated with enhanced tumor malignancy. The influenceof HER2 overexpression on oxidant/antioxidant parameters inhumans remains unknown; therefore, we investigated theoxidative profile in women according to their HER2 status.Fifty-two controls and 52 breast cancer (BC) patients wereenrolled. The BC patients were subdivided into HER−, nega-tive for HER2 overexpression, and HER+, positive for HER2overexpression. Oxidative stress profilling was measured bymalondialdehyde (MDA), free 8-isoprostane F2, protein car-bonyl content, nitric oxide (NO), total radical antioxidantparameter (TRAP), superoxide dismutase (SOD), catalaseactivity, and glutathione (GSH) levels. Total thiol contentand lipoperoxidation were evaluated in HCC1954 and MCF-7. Cells overexpressing HER2 presented enhanced oxidativestress. Increased erythrocyte lipoperoxidation was found inBC patients, while plasma lipoperoxidation was detected inboth the BC and HER− groups. Decreased MDA levels were

found in the HER+ group, suggesting that HER2 overexpres-sion may protects against plasma lipoperoxidation. No alter-ation was found for 8-isoprostane F2, NO, and carbonylcontent. TRAP was decreased in BC patients, while HER2overexpression increased SOD and prevented decreased GSHlevels. These data help to understand the HER2 overexpres-sion in oxidative signaling andmay enable the development ofnew strategies for anti-HER2 therapy.

Keywords HER2 . Oxidative stress . Pro-oxidative profile .

Antioxidant profile

Introduction

Breast cancer is the most lethal malignancy in women world-wide, and its incidence is high in both developed and devel-oping countries [1]. About 20 % of breast cancer patientsoverexpress the human epidermal growth factor receptor-2(HER2/neu, also known as ErbB2) protein [2]. This overex-pressionmay be due to genetic amplification or transcriptionalderegulation. HER2 is a subclass I tyrosine-kinase receptorthat belongs to the HER family, which contains 4 receptors:HER1, HER2, HER3, and HER4 [3].

HER2 is present as dimers on the cell surface, as either ahomodimer with HER2 or a heterodimer with HER1, HER3,or HER4. Overexpression of these receptors results inprolonged signal transduction through distinct pathways. The-se pathways activate transcription factors that decrease apo-ptosis [4] and augment cell survival [5, 6], increase cellmigration, favor tumor progression and growth [7, 8], induceprotein synthesis [9], and stimulate proliferation and cell cycleactivation [8, 10]. Therefore, it is apparent that overexpressionof HER2 in breast cancer is associated with tumor malignancyand bad prognostic.

Carolina Panis and Rubens Cecchini equally contributed to this study.

V. J. Victorino (*) : F. C. Campos :A. C. S. A. Herrera : C. Panis :R. CecchiniLaboratory of Pathophysiology and Free Radicals, Department ofGeneral Pathology, State University of Londrina,86051-990 Londrina, Paraná, Brazile-mail: vanjacob@gmail.com

A. L. CecchiniLaboratory of Molecular Pathology, Department of GeneralPathology, State University of Londrina, Londrina, Paraná, Brazil

C. PanisStem Cells Laboratory, National Institute of Cancer, Rio de Janeiro,Brazil

A. N. Colado SimãoClinical Analyses and Toxicology, Center of Applied Pathology,State University of Londrina, Londrina, Paraná, Brazil

Tumor Biol.DOI 10.1007/s13277-013-1391-x

Studies have highlighted the interaction between HER2overexpression and oxidative stress. However, most studieswere conducted in vitro and were based on non-cancerouscells. Cardiomyocytes have been extensively used, as HER2is constitutively expressed and have a physiological functionin cell contractility. Azambuja et al. suggest that HER2/HER4heterodimer stimulation minimizes oxidative stress incardiomyocytes [11]. Literature also shows that cellular treat-ment inhibiting HER1/HER2 heterodimer increases reactiveoxygen species (ROS) leading to apoptosis. The ROS-inducedapoptosis can be reversed by the presence of SOD mimicadjuvant to treatment of breast cancer [12]. Besides, cellularblockage of HER2 with Herceptin decreases superoxide dis-mutase (SOD) and catalase activity, as well as glutathionelevels due to ROS increase [13]. Timolati and collaboratorssuggested that stimulation of HER2 through Akt pathway byneuregulin-1 beta can exhibit an antioxidant role in cardiaccells [14]. It has also been shown that deleterious effect oftrastuzumab treatment occurs through a mitochondrial path-way and is mediated by ROS in cardiomyocytes [15]. HER2inhibition downregulates the anti-apoptotic protein BCL-XLand upregulates the pro-apoptotic protein BCL-XS conse-quently leading to mitochondrial dysfunction andcardiomyocytes death. Furthermore, chemotherapy inhibitionof HER2 increases ROS upregulating Angiotensin II, whichcan activate NADPH oxidase producing superoxide radicaland inhibits neuregulin signaling [16].

Pathways involved in HER2 signaling can be modulated bythe products of lipid peroxidation, as 4-hydroxy-2-nonenal (4-HNE) [17] and reactive oxygen species [18–22]. Moreover,oxidative stress can modulate cell fate by stimulating either cellproliferation or cell death at mild or high concentrations, re-spectively [23]. Thus, overexpression of HER2 in breast cancermay favor tumors’ proliferation by oxidative stress pathways.Studies regarding oxidative stress and HER2 overexpression inbreast cancer patients are still scarce. Cadenas et al. found arelationship between HER2 and thioredoxin reductase 1 ex-pression in breast tumors associated with poor disease progno-sis [24]. The glutathione pathway in HER2 overexpressingbreast tumor was evaluated by Perguin and collaborators[25]. They found a negative link between HER2 expressionand the glutathione pathway. Tsai et al. did not find anydifference in MnSOD expression in HER2 breast cancer cellscompared with normal breast cells [26]. Contrarily, in vitrotransfection of MnSOD in breast cancer cells could effectivelysuppress HER2 expression at the transcriptional level [27].

The findings point out to an association between the oxi-dative stress pathways and HER2. However, most of thesedata are based mainly on in vitro studies or in the expressionof oxidative stress-related proteins in tumoral tissue. AlthoughHER2 is overexpressed in several tumors types and has beenwidely studied as a target for therapeutic research [28], thesystemic oxidant/antioxidant parameters associated with

HER2 overexpression in humans remains unknown. The in-vestigation of oxidative and the antioxidant profile in plasmafrom women diagnosed with HER2-enriched tumors usinghigh sensitivity and multiple parameters approaches may helpto understand, at least in part, why HER2 overexpressionimply in poor prognosis, new tumor development, or meta-static disease. To reach these goals, we characterized HER2overexpression in breast tumors and associated it with bio-chemical markers regarding systemic oxidative stressprofiling.

Subjects and methods

This study was approved by Research and Ethics NationalCouncil (CAAE 0009.0.268.000-07), and patients providedsigned informed consent.

Patient selection and study design

To determine a significant number of patients for this study,the following formula was applied:

n0 ¼ Z2⋅p 1−pð ÞÇD

n ¼ n0

1þ n0N

N0=number scaled (384.16), Z =confidence level (1.96),P=probability (50 %); D =margin of error (5 %); n =samplesize, and N =population size (54).

Instituto Nacional do Câncer (INCA-Brazil) estimates anincidence of 54 breast cancer cases for 100,000 women. Thus,in a region of 100,000 women, a sample size of 48 patients isrequired. In this study, a total of 954 women diagnosed withbreast cancer were screened. The application of inclusion andexclusion criteria (described below) resulted in a cohort of 52eligible patients in a period ranging from January 2009 toDecember 2011. Women were divided into 52 healthy age-matched controls and 52 breast cancer women (BC). BCgroup was further subdivided into 30 negative for HER2overexpression (HER-) and 22 positive for HER2 proteinoverexpression (HER+) determined by immunohistochemicalstaining (IHC) employing streptavidin biotin method con-firmed by fluorescence in situ hybridization (FISH) whenIHC scoring was ≥2. All antibodies were purchased fromDako (Dako, Germany) and used accordingly to datasheetrecommendations, following the current guidelines for immu-nohistochemistry performance. Excluded subjects from thisstudy were smokers, regular alcohol consumers, antioxidantsupplement users, pregnancy, lactation, excessive physicalexercises practitioners, hormone replacement therapy users,and with another chronic disease. All patients were naturallypostmenopausal women or induced postmenopausal women.

Tumor Biol.

Patients’ characterization included age at diagnosis, TNM(Tumor, Node, and Metastasis) classification, histologicaltumor grade, hormonal receptors (estrogen receptors(ER), progesterone receptors (PR)). No previous chemo-therapy treatment was established before blood collec-tion, and TNM0 patients were excluded from this study.Heparinized blood of all participants was collected atDepartment of General Pathology, Londrina State Uni-versity, PR-Brazil. Blood was centrifuged for red bloodcells (RBCs) and plasma obtainment. All analysis employingRBCs were performed at the collect day and plasma aliquotswere stored in −86 °C (Indrel Ultra Freezer) to furtheranalysis.

Cellular cultures

Human breast carcinoma HCC1954 (ATCC CRL2338™) andMCF-7 (ATCC HTB-22™) cell lines were maintained inRPMI-1640 media supplemented with FBS. To the experi-ments, cells were plated into six-well plates (2×106) wells andtrypsinized after 24 h of culture. The total content of the wells(cells+supernatant) was collected and frozen at −86 °C forin vitro oxidative stress analysis.

Western blotting analysis

Total protein content was determined as described by Brad-ford [29]. Aliquots containing 30 μg of protein were separatedon a 10 % SDS-PAGE gel and transferred to nitrocellulosemembrane. Membranes were blocked with 5 % milk andincubated with mouse anti-human extracellular domain ofHER2/ErbB2 primary antibody (HER2-ECD, CellSignalling, USA) diluted to 1:1,000 at 4 °C overnight, follow-ed by incubation with secondary anti-mouse antibody (diluted1:2,000) at room temperature during 2 h. Membrane antibodybinding was detected using chemiluminescent reagents andcaptured on a film after 3 min exposure. Bands were scannedand quantified in NIH Image J software [30].

Total thiol content as indicative of the Redox statusof HCC1954 and MCF-7 cell lines

Total thiol content was determined as previous described bySedlak and Lindsay [31], with some modifications. Aliquotsof 100 μL were deproteinized by adding 250 μL of trichloricacetic acid. After 10 min at room temperature, tubes werecentrifuged at 1,400×g during 15 min. The supernatant wascollected and added to 1 mL of 0.4 M TRIS buffer, pH8.9.After, 100 μL of DTNB was added to the mixture and thesample read by a spectrophotometer at 412 nm. The resultswere expressed as micromolars of thiol.

Immunohistochemical characterization of breast tissue

Tumor tissue sections were histologically analyzed for diag-nosis by a pathologist. Samples were considered as positivefor ER/PR when at least 10 % of the tumor cells nuclei werestained. HER2 were considered overexpressed when a strongmembrane staining (3+) was detected or FISH analysis am-plification of HER2 in samples with moderate (2+) membranestaining was detected. FISH analysis was performed in tissuesections using a Dako commercial kit. For immunohistochem-ical assay, the reaction was performed on 5 μm-thick paraffin-embedded sections from tumors by the labeled streptavidinbiotin method using commercial kits with microwave accen-tuation. For each case, negative controls were performed onserial sections, whose were analyzed and classified accord-ingly positive or negative results.

Determination of lipid peroxidation by high sensitivechemiluminescence (CL)

To evaluate oxidative stress in cell lines, RBCs, and plasmasamples, a chemiluminescence-based method was employed[32, 33]. For determination of the lipid peroxidation ofHCC1954 and MCF-7 cell lines, 500 μL of cells+supernatantwere added to 480 μL of 30 mM phosphate-KCl 120 mMbuffer, pH 7.4, 37 °C and reaction was initiated by the addic-tion of 20μL of tert-butyl 3 mM solution. RBCwas diluted 1,200× in 10 mM monobasic phosphate buffer, at 37 °C, andreaction started with addition of 10 μL of 3 mM tert-butylhydroperoxide. Plasmatic chemiluminescence reaction wasinitiated by the addiction of 10 μL of 3 mM tert-butyl hydro-peroxide in 125 μL of plasma and 865 μL of 30 mMphosphate-KCl 120 mM buffer, pH 7.4, 37 °C. Readings wereperformed in Glomax luminometer (TD 20/20 Turner De-signers) for 40 min, one reading/second. Data were treatedand analyzed in Origin 8.0 Software, and results wereexpressed in relative light units (RLU).

Malondialdehyde (MDA) levels

High-performance liquid chromatography (HPLC) determina-tions were made using an equipment HPLC-20AT Shimadzuequipped with a LC20AT pump and SPDM20A UV–diodearray absorbance detector employing a C18 reverse phasecolumn. For preparation of standard solution of MDA,10 mL of 0.1 M HCl were added in 10 mL of 1,1,3,3-tetraetoxipropane (TEP), and this solution was maintainedfor 5 min in boiling water, following ice bath to completesynthesis of MDA. A 160 μL of plasma samples or standardsolution reacted with 100 μL of 0.5 M perchloric acid. Sam-ples were centrifuged for 5 min, 5,000×g , 4 °C. The 180 μL ofsupernatant was recovered to react with 100 μL of thiobarbi-turic acid for 30 min, 95 °C, and transferred to ice bath to stop

Tumor Biol.

reaction. A 100 μL of 1 M NaH2PO4, pH 7.0, were added tostabilize sample pH. Furthermore, samples were centrifugedfor 10 min, 5,000×g at 4 °C. Mobile phase was constituted of65 % 50 mM KH2PO4 buffer, pH 7.0, and 35 % methanolHPLC grade. Readings were executed at 535 nm during12 min with isocratic flow of 0.8 mL/min, and results wereexpressed in nM of MDA [33].

Free 8-isoprostanes F2 levels in plasma

8-Isoprostane F2 plasmatic levels were quantified with a com-petitive immunoenzymatic kit (Cayman Chemical, USA),after alkaline hydrolysis of isoprostanes esters. Superna-tants were added to the microplate and analysis per-formed by ELISA. Sample concentrations were deter-mined when compared with recombinant standard curvein pictograms per milliliter.

Evaluation of NO levels

Nitric oxide (NO) concentration in plasma sample was esti-mated by measuring nitrite as previous described [34, 35].Aliquots of 60 μL of plasma samples were deproteinized byadding 50 μL of 75 mM ZnSO4 solution (Merck). Sampleswere mixed, centrifuged at 9,500×g for 2 min, 25 °C, anddeproteinized with 55 mM NaOH (Merck). Samples werecentrifuged at 9,500×g for 5 min, 25 °C, and supernatantsrecovered and diluted 5:1 in 45 g/L glycine, pH 9.7 (Merck).Cadmium granules (Fluka) were activated in 5 mM CuSO4 in15 g/L glycine-NaOH buffer, pH 9.7 (Merck) during 5 mi-nutes. Activated granules were added to samples and incubat-ed for 10 minutes. Griess reagent (Sigma) was added tosupernatants. A calibration curve was prepared by dilutionof NaNO2 (Merck) in distilled sterile water. The absorbancewas performed at 550 nm using a standard microplate reader(Multiskan EX, LabSystems, Minnesota USA). Results wereexpressed in micromolars of nitrite.

Protein carbonyls

To evaluate protein carbonyls levels, 100 μL of plasma sam-ple reacted with 1 mL of 10 mM dinitrophenilhidrazine and2.5 M HCl were employed to precipitate proteins. After 1 h,precipitation with 1.25 mL of 20% trichloroacetic acid (TCA)was performed following a 20 min ice bath and 15 minutes of1,500×g centrifugation. A 1.25 mL of 10 % TCAwas addedto pellet. After 20 min of ice bath, samples were centrifugedfor 15 min, 1,500×g . Pellets were rinsed with ethanol anddiluted in 6 M guanidine [36]. Readings were performed inspectrophotometer (UV-Shimadzu 355–390 nm) and resultswere expressed in nanomolars per milliliter per milligram totalproteins based onmolar extinction coefficient of 22M–1 cm–1.

Total radical antioxidant parameter (TRAP) thoughtchemiluminescence assay

2, 2′Azo-bis, 2 amidinopropane (ABAP), a potent free radicalgenerator, decomposes itself and emits photons in this pro-cess. The action of ABAP is neutralized as long as antioxi-dants are capable of inhibiting its function. Initially, ABAPemission (900 μL of 0.1 M glycine buffer, pH 8.6, 50 μL ofluminol and 50 μL of ABAP) was measured as a pro-emissionstandard. An antioxidant standard solution (25 μM Trolox–6-Hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid) wasadded in order to neutralize ABAP autoxidation (830 μL of0.1 M glycine buffer, pH 8.6, 70 μL of Trolox, 50 μL ofluminol, and 50 μL ABAP ). Subsequently, plasma samplesdiluted 1:50 (830 μL of 0.1 M glycine buffer, pH 8.6, 70 μL ofsample, 50 μL of luminol and 50 μL of ABAP) were used.Readings were performed in a Glomax luminometer (TurnerDesigns TD 20/20) during 30 min, 5 readings/s. Results wereexpressed as nanomolar sample equivalents of Trolox [33, 37].

Superoxide dismutase (SOD) activity determination

RBCs were hemolysed in distillated water in a proportion of1:20. Runs of 5, 10, and 20 μL of samples were measured. Toeach run, distillated water, 1 M TRIS (tris–hydroxymethyl–aminomethane) buffer and pyrogallol (1.2 mg/mL) wereadded. The inhibition of pyrogallol auto-oxidation was mea-sured at 420 nm in spectrophotometer (Shimadzu UV-1650PC) during 6 min kinetic. The results were expressed as SODunities per milliliter accordingly to recommendations for SODactivity over pyrogallol oxidation [38].

Catalase activity determination

RBCs were diluted in distillated water in a proportion of 1:80.Then, 10 μL of sample were incubated in a system containing1 M TRIS buffer and 200 mM hydrogen peroxide. Kinetic ofabsorbance disappearance was monitored in spectrophometerat 240 nm (Shimadzu UV-1650 PC). The results wereexpressed in absorbance values minute per milliliter [39].

Erythrocytic reduced glutathione (GSH) levels

RBCs were hemolysed at a ratio of 1:10 in distillated water,and then 1.25 mL of EDTA (ethylenediamine tetra acetic acid)and 250 μL of 50 % TCAwere added. After 15 min, sampleswere centrifuged at 2,400×g for 15 min. Next, 1 mL ofsupernatant was added to 2 mL of 0.4 M TRIS buffer,pH 8.9. Finally, 50 μL of DTNB (5, 5′–dithiobis–2–nitrobenzoic acid) was added to react with GSH. A standard curvewas performed in order to determine GSH concentration insamples. The absorbance was read at 412 nm, and results wereexpressed in nanomolars [31].

Tumor Biol.

Statistical analysis

All analyses were conducted in triplicate sets. Statistical anal-ysis were performed using GRAPHPAD PRISM version 5.0(GRAPHPAD Software, San Diego, CA), Microsoft OfficeExcel 2007 and OriginLab 8.0 software. Results wereexpressed as arithmetic means and standard error of means(SEM). Differences among groups were assessed by two-wayanalysis of variance (ANOVA) to lipid peroxidation curves,with Bonferroni’s test as post hoc, chi-square, and Fisher’sexact test for clinicapathological data, Mann–Whitney, orStudent’s unpaired t test to others parameters. Data werechecked in GraphPad Software to eliminate significant out-liers (p <0.05). Comparisons performed were: Control×breastcancer group (BC) and HER positive×HER negative group.Differences were considered statistically significant when p <0.05. * marked differences between control and BC group; #marked differences between HER- and HER+ group.

Results

To investigate the oxidative status induced by HER2 amplifi-cation, we performed an in vitro analysis by comparing breastcancer cell lineages known as HER2-negative (MCF7) andHER2-amplified (HCC159). Figure 1a shows Western blot-ting analysis for HER2 status of cell lines. Figure 1b showsthat HER2-positive cells present an enhanced basal status oflipid peroxidation associated with reduced antioxidant capac-ity (1c), confirming that HER2 amplification is able to modifythe oxidative stress in breast cancer cells.

Clinicopathological data are summarized in Table 1. Allpatients displayed infiltrative ductal carcinoma. There was aprevalence of ER and PR presence in BC (ER+=81.5 %, ER−=18.5 %; p <0.001), HER- (ER +=90 %,ER−=10 %; p <0.001) and HER+ (ER+=73 %, ER−=27 %; p <0.001) groups. Advanced disease (AD=stageIII/ IV) was prevalent in BC (AD=80.5 %, early disease

(ED=stage I/II)=19.5 %; p <0.001), HER− (AD=75 %,ED=25 %; p <0.001), and HER+ (AD=86 %, ED=14 %; p <0.001) groups. Besides, histological tumorgrade analysis classified most BC (grade 1=14 %, grade2=62 %, grade 3=24 %; p <0.001), HER2− (grade 1=10 %, grade 2=70 %, grade 3=20 %; p <0.001) andHER2+ (grade 1=18 %, grade 2=55 %, grade 3=27 %;p <0.001) patients in grade 2.

To assess plasmatic and RBC oxidative stress, tert-butyl-induced CL was performed applying two-way ANOVA statis-tics to evaluate if curves are different, Bonferroni’s test to checkwhich curve’s points are different, and curve’s Student’s t test toanalyze mean curves difference. RBC oxidative stress profilewas significantly higher in BC group than in control (p <0.0001) independently of HER2 status, and no points of thecurve was different (Fig. 2a, b). Integration of area under thecurve (AUC) was increased only for BC group (ctr=722,200±58,990 AUC; BC=1,097,000±228,200 AUC; p <0.05). Statis-tical analyses of RBC lipid peroxidation evaluation are present-ed in Fig. 2b.

Figure 2c shows significantly higher levels of plasma lipidhydroperoxides in BC (p <0.0005) and HER- (p <0.0001).Bonferroni’s test did not show differences in points of thecurves and AUC (Fig. 2c, d). MDA levels were decreasedonly in HER+ group (HER−=73.81±7.997 nM; HER+=65.88±21.11 nM, p =0.03; Fig. 3a). No alterations were foundin levels of 8-isoprostanes F2 (Fig. 3b), protein carbonylscontent (Fig. 3c), and nitrite (Fig. 3d).

BC group showed a significant reduction in TRAP levels(ctr=366.6±17.36 nM Trolox/mgx dL−1; BC=200.7±15.81 nM Trolox/mgx dL−1; p <0.0001) independently ofHER2 status (HER-=202.1±20.99 nM Trolox/mgx dL−1;HER+=198.4±23.29 nM Trolox/mgx dL−1) (Fig. 4a).

SOD evaluation showed significant augment in HER+

group (HER-=3.234±0.1661 USOD/mL; HER+=5.145±0.8173 USOD/mL; p =0.03) (Fig. 4b). Catalase activity waslowered in BC (ctr=544.5±11.20 Vabs/min/mL; BC=504.7±8.977 Vabs/min/mL; p =0.007) independently of HER2 status

b) c)

HER2 – ECD (200 kDa)

MCF-7 HCC159 a)Fig. 1 Oxidative profiling ofHER2-positive cells. a Westernblotting analysis for HER2 status ofcell lines, b lipid peroxidation, andc antioxidant content fromHCC159 and MCF7 cells. Asteriskindicates statistical differencebetweeen groups (p<0.05)

Tumor Biol.

(Fig. 4c). GSH levels displayed tendency to decrease in BC(ctr=16.62±1.813 nM; BC=14.70±1.428 nM; p =0.08) andsignificantly augment in HER+ (HER-=9.510±1.687 nM;HER+=17.24±2.457 nM; p =0.04) (Fig. 4d).

Discussion

In the present study, we report several pathological alterationsthat occur due to overexpression of HER2 in women withbreast cancer. Our data revealed that cells overexpressingHER2 present enhanced oxidative stress when compared withthose HER2-negative. In patients, our findings revealed thepresence of specific oxidative changes according to HER2status. HER2-dependent events included attenuation of lipidperoxidation, reduction in MDA levels, enhanced SOD activ-ity, and increased erythrocyte GSH level. Other alterationssuch as TRAP consumption and reduction in catalase activitywere not dependent on HER2 status and may reflect changesthat occurred during oxidative metabolism that originatedfrom other signaling pathways.

As previous described by our laboratory, breast cancer wom-en displayed iron overload as well as increased ferritin levels[40]. High levels of plasma iron can induce oxidative stressresponses by generating hydroxyl radicals through Fenton’sreaction [41] that lead to lipid peroxidation [42]; we observedincreased lipid peroxidation in the plasma of the HER− patientsand in the RBCs of all BC patients. Another important findingfrom our study was the characterization of different oxidativeprofiles according to HER2 overexpression. The lipid peroxi-dation reaction involves initiation, propagation, and terminationsteps [43]. Here, initiation was characterized by the ascending

part of the curve and was dependent on the antioxidant contentof the sample. High antioxidant content is able to retard theinitiation process and decreased AUC is indicative of reducedlipid hydroperoxides [44]. In this study, we showed that lipidperoxidation was increased in RBCs of BC patients, indepen-dent of their HER2 status, and pre-formed lipid hydroperoxidesincreased in the absence of HER2 overexpression. Likewise,Badid et al. [45] demonstrated higher susceptibility to oxidationof low density lipoprotein (LDL) in obese breast cancer pa-tients. Increased concentration of 4-hydroxy-2-nonenal, a prod-uct and marker of lipid peroxidation, have been also reportedduring breast carcinogenesis assessed on breast tumor samples[46]. According to the hypothesis of Gago-Dominguez andcolleagues [47], products derived from lipid peroxidation mayrepresent a protective mechanism against breast cancer. Thus,HER2 overexpression seems to protect against lipid peroxida-tion, because we further detected reduced MDA levels duringHER2 overexpression.

Although we have previously demonstrated increased NOlevels and high carbonyl content in advanced breast cancer[48], that report did not investigate any differences in theseparameters with respect to HER2 status. To delineate the system-ic antioxidant profile and its relationship to HER2 status, weanalyzed the non-enzymatic (TRAP and GSH levels) and enzy-matic antioxidant (SOD and catalase activities) defense path-ways. TRAP was decreased in BC patients, independent ofHER2 overexpression, suggesting that low-molecular-weightantioxidants are consumed as a defense mechanism against theoxidative processes caused by cancer-induced metabolic chang-es. Enhanced TRAP consumption has been described in ad-vanced breast cancer patients [48], but the role of HER2 statusin antioxidant consumption has not been reported.

Table 1 Clinicopathological dataof breast cancer patients (BC),breast cancer women negative forHER2 (HER−) and positive forHER2 overexpression (HER+)

Individual groups were comparedusing Qui-square and Fisher’sexact test

ER estrogen receptor, PR proges-terone receptor, TNM tumor,node, metastasis classification

*p <0.05 indicates statisticalsignificance)

BC HER− HER+

Age at diagnosis (years) 33–72 39–72 33–63

Histologycal type Infiltrative ductal carcinoma 100 % 100 % 100 %

Hormone receptors ER+ 81.5 %* 90 %* 73 %*

ER− 18.5 % 10 % 27 %

PR+ 82 %* 87 %* 77 %*

PR− 18 % 13 % 23 %

TNM I/II 19.5 % 25 % 14 %

III/IV 80.5 %* 75 %* 86 %*

Metastatic site (stage IV) Axilar 3 % 6 % 0 %

Brain 5 % 6 % 5 %

Bones 30 % 37 % 24 %

Contralateral breast 16 % 19 % 14 %

Liver 19 % 7 % 28.5 %

Lungs 27 % 25 % 28.5 %

Histological grade 1 14 % 10 % 18 %

2 62 %* 70 %* 55 %*

3 24 % 20 % 27 %

Tumor Biol.

HER2 overexpression increased SOD activity, indicatingthat either HER2 signaling leads to enhanced superoxideanion production or directly increases SOD catalytic activity/expression. Enhanced superoxide dismutase activity is indic-ative of high rates of hydrogen peroxide production, which

trigger catalase activity [23]. Also, our previous study dem-onstrated decreased catalase activity in advanced breast cancerpatients [48]. After HER2 activation, Ras/ERK and PI3K/ Aktare the most relevant induced pathways [49]. Downregulationof HER2 expression suppresses cellular surviving and

d)

Statistical Analysis CTR X BC HER- X HER+

Two-way ANOVA P<0.0001 No significance

Bonferroni’s Test No points No points

Curves Student’s t Test P= 0.0598 No significance

AUC unpaired Student’s t Test

No significance No significance

a) b)

c)Statistical Analysis CTR X BC HER- X HER+

Two-way ANOVA P<0.0005 P<0.0001

Bonferroni’s Test No points No points

Curves Student’s t Test P= 0.0460 P<0.0001

AUC unpaired Student’s t Test

No significance No significance

Fig. 2 Oxidative stress evaluated by high-sensitivity chemiluminescencein breast cancer patients (BC), breast cancer women negative for HER2(HER−), and positive for HER2 overexpression (HER+). a Profile of redblood cells (RBC) lipid peroxidation, b statistical significance of RBClipid peroxidation curves, c profile of plasmatic lipid peroxidation, and dstatistical significance of plasmatic lipid peroxidation curves are shown.

RLU =relative light unities, AUC =area under the curve. Data areexpressed as mean of values. Asterisk indicates a statistically significantdifference between control and BRCA; number sign indicates a statisti-cally significant difference between HER− and HER+. P <0.05 wasadopted as significant

Fig. 3 Pro-oxidative profile ofbreast cancer patients (BC), breastcancer women negative for HER2(HER−) and positive for HER2overexpression (HER+). amalondialdehyde (MDA) levelsmeasured by HPLC, b 8-isoprostanes F2 levels, c nitrite asestimative of nitric oxide (NO),and d protein carbonyls levels.Individual distributions of valuesand means were evaluated by (b ,c , d) Student’s unpaired t test or(a) Mann–Whitney U test.Number sign indicates astatistically significant differencebetween HER− and HER+. P<0.05 was adopted as significant

Tumor Biol.

proliferation via PI3K/ Akt and ERK/MAPK [22]. Currently,hydrogen peroxide has been recognized as a positive mediatorof Phospho-ERK1/2, PI3K/Akt, and p38 MAPK pathways[18]. Besides, Akt activity is stimulated by low concentrationsof ROS [50]. PI3K/Akt and Ras/ ERK play an important rolemediating oncogenic signaling function of HER2+ tumorcells, and their activation is known to augment cellsurvival [5, 6] and promote cell migration and prolifer-ation [7, 8, 10, 49]. Inhibition of PI3K/Akt pathway, ina model of non-small cell lung cancer, can upregulateSOD, glutathione peroxidase, and catalase activity [51].Also, activation of PI3K/Akt can participate in the in-duction of heme oxygenase-1 antioxidant enzyme [52].Here, we showed that catalase activity was reduced inBC patients independent of their HER2 status, eventhough HER2+ tumors displayed increased SOD activi-ty. This finding suggests that HER2 signaling in breastcancer patients may trigger SOD activity independent ofcatalase activity.

Another mechanism by which hydrogen peroxide is re-moved, other than by catalase activity, is the GSH system.GSH in the presence of glutathione peroxidase can reducehydrogen peroxide to water [53]. The induction of pathwaysmediated by Jun amino-terminal kinase, an important anti-apoptotic pathway induced by the HER family [3], can induceGSH production [54]. As we demonstrated here, HER2 over-expression prevented GSH consumption and tended to in-crease gamma-glutamyl-transpeptidase (GGT, data notshown), probably by a mechanism that prevents excess hy-drogen peroxide production. In agreement with our results,Dogan and colleagues [13] showed that blockage of HER2

decreases SOD, catalase, and GSH, leading to increased oxi-dative stress in vitro. Besides, Zeglinksi et al. demonstratedthat inhibition of HER2 increases NADPH oxidase activityleading to ROS production and activates ASK-1 and p38/JNK-associated pathways culminating in apoptosis [16];therefore, HER2 overexpression signaling may act asan antioxidant.

In conclusion, this study provides new data regarding theinvolvement of HER2 overexpression and oxidative stress inwomen with breast cancer, according to their HER2 status.Here, we reported that, in breast cancer patients, HER2 over-expression attenuates oxidative stress by preventing plasmalipid peroxidation and MDA formation, and increasing SODand GSH levels. Based on our findings, we observed that theattenuation of oxidative stress only occurs in vivo, since ourin vitro data show that HER2 amplification promotes theoxidative stress signaling. Once patients bearing HER2 tu-mors presented reduced oxidative profile, we believe that theattenuation of oxidative stress must be triggered by somemediators produced during the imbalance of metabolic andimmune responses in breast cancer patients. Therefore, theseresults help to better understand the effect of HER2 overex-pression on oxidative signaling and may enable the develop-ment of new strategies for anti-HER2 therapy.

Acknowledgments The authors would like to thank Jesus AntônioVargas for excellent technical assistance, all of the participating womenfor making the study possible, and CAPES, CNPq, and Fundação Arau-cária for providing financial support.

Conflicts of interest None

Fig. 4 Antioxidant parameters inbreast cancer patients (BC), breastcancer women negative for HER2(HER−), and positive for HER2overexpression (HER+). aAntioxidant profile determined bytotal radical antioxidant parameter(TRAP) corrected by uric acidlevels, b superoxide dismutaseactivity, c catalase activity, and dreduced glutathione levels.Individual distributions of valuesand means were evaluated byStudent’s unpaired t test. Asteriskindicates a statistically significantdifference between control andBRCA; number sign indicates astatistically significant differencebetween HER− and HER+. P<0.05 was adopted as significant.Triangle indicates p valuebetween 0.1 and 0.05

Tumor Biol.

References

1. World Health Organization (WHO) 2011 <http://www.who.int/topics/cancer/en>

2. Gutierrez C, Schiff R. HER2 biology, detection, and clinical impli-cations. Arch Pathol Lab Med. 2011;135:55–62.

3. Yarden Y, Sliwkowski MX. Untangling the Erbb signalling network.Nat Rev Mol Cell Biol. 2001;2:127–37.

4. Aranda V, Haire T, Nolan ME, et al. MuthuswamyPar6–aPKC un-couples ErbB2 induced disruption of polarized epithelial organiza-tion from proliferation control. Nat Cell Biol. 2006;8:1235–45.

5. Kim AH, Khursigara G, Sun X, Franke TF, Chao MV. Akt phos-phorylates and negatively regulates apoptosis signal-regulating ki-nase. Mol Cell Biol. 2001;21:893–901.

6. Tang ED, Nunez G, Barri FG, Guan KL. Negative regulation of theforkhead transcription factor FKHR by Akt. J Biol Chem. 1999;274:16741–6.

7. Dittmar T, Husemann A, Schewe Y, Nofer J, Niggemann B, ZänkerKS, Brandt BH. Induction of cancer cell migration by epidermalgrowth factor is initiated by specific phosphorylation of tyrosine1248 of c-erbB-2 receptor via EGFR1. FASEB J. 2002; 1823–25.

8. Zhou BP, Liao Y, XiaW, Spohn B, LeeMH, HungMC. Cytoplasmiclocalization of p21Cip1/WAF1 by Akt-induced phosphorylation inHER-2/neu overexpressing cells. Nat Cell Biol. 2001;3:245–52.

9. Dan HC, Sun M, Yang L, et al. Phosphatidylinositol 3-kinase/Aktpathway regulates tuberous sclerosis tumor suppressor complex byphosphorylation of tuberin. J Biol Chem. 2002;277:35364–70.

10. HigashiyamaM, Doi O, Kodama K, et al. MDM2 gene amplificationand expression in non small cell lung cancer: immunohistochemicalexpression of its protein is a favourable prognostic marker in patientswithout p53 protein accumulation. Brit J Cancer. 1997;9:1302–8.

11. Azambuja E, Bedard PL, Suter T, Piccart-Gebhart M. Cardiac toxic-ity with anti-HER-2 therapies-what have we learned so far? TargOncol. 2009;4:77–88.

12. Aird KM, Allensworth JL, Batinic-Haberle I, Lyerly HK, DewhirstMW, Devi GR. ErbB1/2 tyrosine kinase inhibitor mediates oxidativestress-induced apoptosis in inflammatory breast cancer cells. BreastCancer Res Treat. 2012;132(1):109–19.

13. Dogan I, CumaogluA, AriciogluA, Ekmekci A. Inhibition of ErbB2 byHerceptin reduces viability and survival, induces apoptosis and oxidativestress in Calu-3 cell line. Mol Cell Biochem. 2011;347:41–51.

14. Timolati F, Ott D, Pentassuglia L, GiraudMN, Perriard JC, Suter TM,et al. Neuregulin-1 beta attenuates doxorubicin-induced alterations ofexcitation–contraction coupling and reduces oxidative stress in adultrat cardiomyocytes. J Mol Cell Cardiol. 2006;41:845–54.

15. Gordon LI, BurkeMA, Singh AT, Prachand S, Lieberman ED, Sun L,et al. Blockade of the erbB2 receptor induces cardiomyocyte deaththrough mitochondrial and reactive oxygen species-dependent path-ways. J Biol Chem. 2009;284(4):2080–7.

16. Zeglinski M, Ludke A, Jassal DS, Singal PK. Trastuzumab-inducedcardiac dysfunction: a ‘dual-hit’. Exp Clin Cardiol. 2011;16(3):70–4.

17. Forman HJ, Fukuto JM, Miller T, Zhang H, Rinna A, Levy S. Thechemistry of cell signaling by reactive oxygen and nitrogen speciesand 4-hydroxynonenal. Arch Biochem Biophys. 2008;2:183–95.

18. Angeloni C, Motori E, Fabbri D, Malaguti M, Leoncini E, LorenziniA, et al. H2O2 preconditioning modulates phase II enzymes throughp38 MAPK and PI3K/Akt activation. Am J Physiol. 2011;300(6):H2196–205. doi:10.1152/ajpheart.00934.2010.

19. Chang CY, Chan HL, Lin HY, Way TD, Kao MC, Song MZ, et al.Rhein induces apoptosis in human breast cancer cells. Evid BasedComplement Alternat Med. 2011;2012:952504.

20. Kuo HP, Chuang TC, Yeh MH, et al. Growth suppression of HER2-overexpressing breast cancer cells by berberine via modulation of theHER2/PI3K/Akt signaling pathway. J Agric Food Chem.2011;59(15):8216–24.

21. Seo HS, Choi HS, Choi HS, et al. Phytoestrogens induce apoptosisvia extrinsic pathway, inhibiting nuclear factor-{kappa}B signalingin HER2-overexpressing breast cancer cells. Anticancer Res.2011;31(10):3301–13.

22. Shin-Kang S, Ramsauer VP, Lightner J et al. Tocotrienols inhibitAKT and ERK activation and suppress pancreatic cancer cell prolif-eration by suppressing the ErbB2 pathway. Free Radic Biol Med.2011;15;51(6):1164–74.

23. Halliwell B. Oxidative stress and cancer: have we moved forward?Biochem J. 2007;401:1–11.

24. Cadenas C, Franckenstein D, Schmidt M, Gehrmann M, Hermes M,Geppert B, et al. Research aortic thioredoxin reductase 1 andthioredoxin interacting protein in prognosis of breast cancer. BreastCancer Res. 2010;12:R44.

25. Perquin M, Oster T, Maul A, Froment N, Untereiner M, Bagrel D.The glutathione-related detoxification system is increased in humanbreast cancer in correlation with clinical and histopathological fea-tures. J Cancer Res Clin Oncol. 2001;127:368–74.

26. Tsai SM, Hou MF, Wu SH, Hu BW, Yang SF, Chen WT, et al.Expression of manganese superoxide dismutase in patients withbreast cancer. Kaohsiung J Med Sci. 2011;27:167–72.

27. Chuang TC, Liu JY, Lin CT, Tang YT, Yeh MH, Chang SC, et al.Human manganese superoxide dismutase suppresses HER2/neu-mediated breast cancer malignancy. FEBS Lett. 2007;581:4443–9.

28. Hyne NE, Lane HA. Erbb receptors and cancer: the complexity oftargeted inhibitors. Nat Rev. 2005;5:341–54.

29. Panis C, Pizzatti L, Herrera ACSA, Cecchini R, Abdelhay E. Putativecirculating markers of the early and advanced stages of breast canceridentified by high-resolution label-free proteomics. Cancer Lett.2013;330:57–66.

30. Bradford MM. Rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dyebinding. Anal Biochem. 1976;72:248–54.

31. Sedlak J, Lindsay RH. Estimation of total, protein-bound, and non-protein sulfhydryl groups in tissue with Ellman’s reagent. AnalBiochem. 1968;25:192–205.

32. Herrera ACSA, Panis C, Victorino VJ et al. Molecular subtype isdeterminant on inflammatory status and immunological profile frominvasive breast cancer patients. Cancer Immunol Immunother. 2012;doi:10.1007/s00262-012-1283-8.

33. Victorino VJ, Panis C, Campos FC et al. Decreased oxidant profileand increased antioxidant capacity in naturally postmenopausalwomen. AGE 2012. doi:10.1007/s11357-012-9431-9.

34. Panis C, Lemos LGT, Victorino VJ et al. Immunological effects ofTaxol and Adryamicin in breast cancer patients. Cancer ImmunolImmunother 2012; doi:10.1007/s00262-011-1117-0.

35. Panis C, Mazzuco TL, Costa CZF, Victorino VJ, Tatakihara VLH,Yamauchi LM, et al. Trypanosoma cruzi: effect of the absence of 5-lipoxygenase (5-LO)-derived leukotrienes on levels of cytokines,nitric oxide and iNOS expression in cardiac tissue in the acute phaseof infection in mice. Exp Parasitol. 2011;127:58–65.

36. Reznick AZ, Packer L. Oxidative damage to proteins: spectrophoto-metric method for carbonyl assay. Methods Enzymol.1994;233:357–63.

37. Repetto M, Reides C, Carretero MLG, Costa M, Griemberg G,Llesuy S. Oxidative stress in blood of HIV infected patients. ClinChim Acta. 1996;225:107–17.

38. Marklund S, Marklund G. Involvement of the superoxide anionradical in the autoxidation of pyrogallol and a convenient assay forsuperoxide dismutase. Eur J Biochem. 1974;47:474–96.

39. Aebi H. Catalase in vitro. Methods Enzymol. 1984;105:121–6.40. Panis C, Victorino VJ, Herreira ACSA et al. Differential oxidative

status and immune characterization of the early and advanced stagesof human breast cancer. Breast Cancer Res Tr. 2011; doi:10.1007/s10549-011-1851-1.

41. Halliwell B, Gutteridge JMC. Free radicals in biology and medicine. In:

Tumor Biol.

oxygen is a toxic gas – an introduction to oxygen toxicity and reactivespecies. 4th edn. New York: Oxford University 2007; pp 1–28.

42. Jian J, Yang Q, Dai J, Eckard J, Axelrod D, Smith J, et al. Effects ofiron deficiency and iron overload on angiogenesis and oxidativestress—a potential dual role for iron in breast cancer. Free Rad BiolMed. 2011;50:841–7.

43. Zwart LL, Meerman JHN, Commandeur JNM, Vermeulen NP.Biomarkers of free radical damage applications in experimentalanimals and in humans. Free Rad Biol Med. 1999;26:202–26.

44. Colado-Simão AN, Suzukawa AA, Casado MF, Oliveira RD,Cecchini R. Genistein abrogates pre-hemolytic and oxidative stressdamage induced by 2,2′-azobis (Amidinopropane). Life Sci.2006;11:1202–10.

45. Badid N, Ahmed FZB, Merzouk H, Belbraouet S, Mokhtari N,Merzouk SA, et al. Oxidant/antioxidant status, lipids and hormonalprofile in overweight women with breast cancer. Pathol Oncol Res.2010;16:159–67.

46. Karihtala P, Kauppila S, Puistola U, Jukkola-Vuorinen A. Divergentbehaviour of oxidative stress markers 8-hydroxydeoxyguanosine (8-OHdG) and 4-hydroxy-2-nonenal (HNE) in breast carcinogenesis.Histopathology. 2011;58:854–62.

47. Gago-Dominguez M, Jiang X, Castelao JE. Lipid peroxidation, ox-idative stress genes and dietary factors in breast cancer protection: ahypothesis. Breast Cancer Res. 2007;9:201.

48. Panis C, Herreira ACSA, Victorino VJ et al. Oxidative stress and

hematological profiles of advanced breast cancer patients subjected topaclitaxel or doxorubicin chemotherapy. Breast Cancer Res Tr. 2011;doi:10.1007/s10549-011-1693-x.

49. Henson ES, Johnston JB, LosM, Gibson SB. Clinical activities of theepidermal growth factor receptor family inhibitors in breast cancer.Biologics: Targets Ther. 2007;1(3):229–39.

50. Dong-Yun S, Yu-Ru D, Shan-Lin L, Ya-Dong Z, Lian W. Redoxstress regulates cell proliferation and apoptosis of human hepa-toma through Akt protein phosphorylation. FEBS Lett.2003;542(1–3):60–4.

51. Akca H, Demiray A, Aslan M, Acikbas I, Tokgun O. Tumoursuppressor PTEN enhanced enzyme activity of GPx, SOD and cat-alase by suppression of PI3K/AKTpathway in non-small cell lungcancer cell lines. J Enzyme Inhib Med Chem. 2013;28(3):539–44.

52. Dal-Cim T, Molz S, Egea J, Parada E, Romero A, Budni J, et al.Guanosine protects human neuroblastoma SH-SY5Y cells againstmitochondrial oxidative stress by inducing heme oxigenase-1 viaPI3K/Akt/GSK-3b pathway. Neurochem Int. 2012;61:397–404.

53. AruomaOI. Free radicals, oxidative stress, and antioxidants in humanhealth and disease. JAOCS. 1998;75:199–212.

54. Zou X, Zhihui F, Li Yet al. Stimulation of GSH synthesis to preventoxidative stress-induced apoptosis by hydroxytyrosol in human ret-inal pigment epithelial cells: activation of Nrf2 and JNK-p62/SQSTM1 pathways. J Nutr Biochem 2011; doi:10.1016/j.jnutbio.2011.05.006.

Tumor Biol.