Cancer Genetics and Cytogenetics 203 (2010) 222e229

EGFR expression and gene copy number in triple-negativebreast carcinoma

Berrak Gumuskayaa,b,*, Murat Alpera, Sema Hucumenoglua, Kadri Altundagc,Aysegul Unerb,c, Gulnur Gulerb,d

a2nd Pathology Department, Ankara Diskapi Yildirim Beyazit Education and Research Hospital, Irfan Bastug cad. Diskapi, Ankara, 06110, TurkeybPathology Department of Hacettepe University, Hacettepe University Hospitals, 06100 Sihhiye/Ankara, Turkey

cHacettepe University Institute of Oncology, Hacettepe University Hospitals, 06100 Sihhiye/Ankara, Turkeyd2nd Pathology Department of Ataturk Education and Research Hospital, Bilkent yolu no:2 Ankara, Turkey

Received 6 August 2009; received in revised form 5 June 2010; accepted 3 July 2010

Abstract Most basal-like breast carcinomas are estrogen

This study was p

Congress.

* Corresponding a

66 90.

E-mail address: b

0165-4608/$ - see fro

doi:10.1016/j.cancerg

receptor negative, progesterone receptor negative,and cerb-B2/HER-2/neu negativedthe so-called triple-negative breast carcinomasdwith highepidermal growth factor receptor (EGFR) expression, which makes EGFR a target of treatment.We evaluated EGFR expression by immunohistochemistry (IHC) with two different clones(EGFR.31G7 and EGFR.25) and gene copy number by fluorescence in situ hybridization (FISH)with Locus specific identifier EGFR/CEP 7 dual probe in 62 triple-negative breast carcinomas.Any complete or incomplete membranous and/or cytoplasmic expression was regarded as IHC posi-tive. Cases showing gene amplification (a ratio of EGFR gene to chromosome 7 of $2 or 15 copiesper cell in $10% of cells) and high polysomy ($4 copies in $40% of cells) were considered FISHpositive. We detected EGFR.31G7 positivity in 38 of 62 cases (61.4%), which was composed of 12of 62 (19.4%) cytoplasmic, 14 of 62 (22.6%) incomplete membranous, and 12 of 62 (19.4%)complete membranous staining. Among 38 of 49 (77.6%) EGFR.25-positive cases, 7 of 49(14.3%) exhibited cytoplasmic, 10 of 49 (20.4%) exhibited incomplete membranous, and 21 of49 (42.9%) exhibited complete membranous staining pattern. Ten of 62 (16.1%) FISH-positivecases were identified; 1 of 62 (1.6%) showed amplification, and the rest showed high polysomy.All FISH-positive cases were also found to be IHC positive (P 5 0.01) by both EGFR clones.The amplified case displayed strong complete membranous staining with both clones. Amongthe high polysomic cases; 4 of 9 (44.4%) incomplete membranous, 4 of 9 (44.4%) completemembranous and 1 of 9 (11.1%) cytoplasmic expression of EGFR.31G7, and 6 of 8 (75%) completemembranous and 2 of 6 (25%) cytoplasmic expression of EGFR.25 were detected. Here, we reportthat membranous EGFR expression is associated with increased gene copy number (P 5 0.035 forEGFR.31G7 and P 5 0.026 for EGFR.25 clone). Because the markers to predict anti-EGFR treat-ment response in other system tumors such as EGFR mutation and amplification seem to be rareevents in breast cancer, membranous staining pattern of EGFR might be the best way to decidethe patient eligibility for anti-EGFR therapy. � 2010 Elsevier Inc. All rights reserved.

1. Introduction

The estimated rate of new breast cancer cases amongwomen is 27% [1]. Gene expression profiling of breastcancers revealed prognostically important molecular

resented at the 18th Turkish National Pathology

uthor. Tel.: þ90 312 596 27 14; fax: þ90 312 318

errakg206@yahoo.com (B. Gumuskaya).

nt matter � 2010 Elsevier Inc. All rights reserved.

encyto.2010.07.118

subtypes. One of them, basal-like breast carcinoma, hasa worse overall and disease-free survival compared to otherbreast tumor subtypes [2]. Basal-like breast tumorscomprise 10e20% of all breast cancers; they result in poorprognosis and occur at a higher prevalence among premen-opausal and young patients [3]. Several studies have soughtto discover the morphological and immunohistochemicalcharacteristics of this group. Basal-like breast carcinomasare morphologically characterized by the presence ofmedullary-like features, central scar, tumor necrosis,spindle cells or squamous metaplasia, high mitotic activity,

223B. Gumuskaya et al. / Cancer Genetics and Cytogenetics 203 (2010) 222e229

and high nuclearecytoplasmic ratio [4]. Regarding the im-munophenotypical features of basal-like breast tumors, it isreported that a panel of markers including estrogen receptor(ER), progesterone receptor (PR), HER-2, epidermal growthfactor receptor (EGFR), andCK5/6 can detect themwith highspecifity [5]. The highest concordant immunohistochemicalprofile is reported as ER negative, HER-2 negative, EGFRpositive, CK5/6 positive, Vimentin positive, and CK8/18positive [6]. Cheang et al. reported that a panel of markerscomposed of ER, PR, HER-2, EGFR, and CK5/6 can identifybasal-like breast carcinomas with 100% sensitivity and 76%specificity [7,8]. The majority of ER-negative, PR-negative,and HER-2-negative (the so-called triple-negative breasttumors, TNBC) are basal-like breast tumors that can beshown by EGFR and/or CK5/6 positivity [9].

EGFR is a molecular therapeutic target in colorectalcancer, non-small-cell lung cancer (NSCLC), andsquamous-cell carcinoma of the head and neck [10].Patients who have EGFR gene mutations and increasedcopy numbers are sensitive to anti-EGFR therapies,including tyrosine kinase inhibitors and monoclonal anti-bodies [10,11]. Monoclonal antibodies prevent ligandbinding and interrupt the signaling cascade via extracellulardomain of EGFR. Tyrosine kinase inhibitors repress down-stream effects of EGFR ligand binding via intracellulardomain of EGFR.

Most of the TNBCs express EGFR, which is a strongcandidate for a therapeutic challenge [5,8]. The first studiesfound the EGFR expression associated with gene amplifica-tion by chromogen in situ hybridization method in meta-plastic breast carcinoma, believed to be one of thephenotypes of basal-like breast cancers [12,13]. Althougha subsequent study on metaplastic breast carcinomas re-vealed 4%amplification by fluorescence in situ hybridization(FISH), in a recent study, EGFR gene amplification was notobserved in any case by the same method [14,15].

Toyama et al. reported high EGFR gene copy number byreal-time polymerase chain reaction (PCR) in TNBCs, butthis method does not take chromosome 17 copy changesinto account [16]. These authors did not identify an associ-ation between EGFR protein expression and gene copynumber.

We undertook the present study to determine whetherthe overexpression of EGFR is associated with gene ampli-fication in TNBCs by means of dual-probe FISH, whichalso verifies chromosome 17 copy changes.

2. Materials and methods

2.1. Patients and tissues

Sixty-two breast carcinomas that were negative for ER,PR, and HER-2/neu diagnosed at Hacettepe University’spathology department were included in the study. Allhematoxylin and eosine and immunohistochemically

stained sections of these tumors were reviewed to confirmmorphological and immunohistochemical parameters [17].

2.2. Tissue microarray

Tissue microarrays (Advanced Tissue Arrayer, ATA100,Chemicon International, Temecula, CA) were constructedof 62 TNBC specimens. The presence of tumor tissue onthe arrayed samples was verified via analysis of hematox-ylin and eosinestained sections. Each case was representedin duplicate on the tissue microarrayer. One core from thecentral and another from the peripheral part of the tumorwere represented in the array.

2.3. Immunohistochemistry

The unstained whole sections cut at 5-mm thickness wereprepared from blocks for immunostaining. Cases werestained with two different clones of EGFR antibody;EGFR.31G7 and EGFR.25. Immunohistochemical stainingof EGFR was performed by the standard streptavidinebiotin complex method with two antibodies raised againstEGFR (mouse monoclonal, 31G7, 1:20, Zymed [SouthSan Francisco, CA], and mouse monoclonal, EGFR 25,1:50, Novocastra [Newcastle, United Kingdom]). In brief,sections were deparaffinized and rehydrated, and endoge-nous peroxidase activity was blocked with a 0.3% solutionof hydrogen peroxidase in phosphate-buffered saline (0.01mol/L, pH 7.5) at room temperature for 10 minutes. Thesections of EGFR.31G7 were treated with trypsin inphosphate-buffered solution for 10 minutes at 37�C forepitope retrieval. Sections of EGFR.25 were immersed in0.01 mol/L sodium citrate buffer (pH 6.0) for 3 minutes.Then primary antibody was allowed to react at room temper-ature for 40 minutes in EGFR.31G7 dilutions of 1:20 and for60 minutes in EGFR.25 dilutions of 1:50. After washing inphosphate-buffered saline, secondary antibody was appliedfor 10 minutes, followed by streptavidineperoxidasecomplex (ScyTek Laboratories, Logan, UT). Peroxidasewas visualized by diaminobenzidine tetrahydrochloride con-taining 0.3% H2O2. After rinsing in deionized water andcounterstaining in Harris hematoxylin, the slides were dehy-drated and mounted. A colon cancer case and normalplacenta were used as control tissue for EGFR.31G7 andEGFR.25.

EGFR staining pattern was evaluated as completemembranous, incomplete membranous, cytoplasmic, andnegative (Fig. 1). Cases were considered EGFR positivewhen the tumor cells showed any cytoplasm and/ormembrane staining. While deciding the staining pattern,cases showing any incomplete or complete membranousstaining with or without cytoplasmic reaction wereincluded in incomplete and complete membranous stainingpattern groups, respectively. The cases with pure cyto-plasmic staining were included in cytoplasmic stainingpattern group.

Fig. 1. Representative cases. (A) Intense complete membranous staining pattern (the case resembling amplification). (B) Incomplete membranous staining

pattern. (C) Cytoplasmic staining pattern. (D) A negative case (IHC, original magnification �400) with the EGFR.31G7 clone. (E,F) Yellow arrows show

epidermal growth factor receptor (EGFR) genes marked by red florescent probe; white arrows show chromosome 7 centromeres marked by green florescent

probe. (E) Tight EGFR gene clusters (red arrows) in an amplified diploid tumor cell population on tissue core (fluorescence in situ hybridization, original

magnification �1,000). Tumor cell nucleus of a high polysomic case, which resembles the 5 copy number ratio of red EGFR gene vs. green chromosome

7 signals (inset). (F) Two red and two green signals in a disomic tumor cell.

224 B. Gumuskaya et al. / Cancer Genetics and Cytogenetics 203 (2010) 222e229

2.4. FISH

Fluorescence in situ hybridization (FISH) was per-formed by hybridizing slides with fluorescent-labeleddual-colored probes (LSI EGFR/CEP7; Abbott Molecular,Des Plaines, IL) according to the manufacturer’s instruc-tions. The Abbott dual-color probe consists of the LSIEGFR SpectrumOrange probe containing the entire EGFRgene, and the CEP7 SpectrumGreen probe hybridizing tothe alpha satellite DNA located at the centromere of chro-mosome 7. The cases were evaluated by the CytoVision

program. Signals were scored in 20 to 60 nonoverlappingtumor cell nuclei with a fluorescence microscope (OlympusBX51). In each nucleus, the number of EGFR and chromo-some 7 copies were independently assessed, and the meancopy number for EGFR and chromosome 7 was calculated.

Patients were classified into six FISH strata according tothe previously described criteria of Cappuzzo et al. [18], asfollows: (1) disomy (#2 copies in O90% of cells); (2) lowtrisomy (#2 copies in $40% of cells, 3 copies in 10e40%of the cells, $4 copies in !10% of cells); (3) high trisomy

225B. Gumuskaya et al. / Cancer Genetics and Cytogenetics 203 (2010) 222e229

(#2 copies in $40% of cells, 3 copies in $40% of cells,$4 copies in !10% of cells); (4) low polysomy ($4copies in 10e40% of cells); (5) high polysomy ($4 copiesin $40% of cells); and (6) gene amplification (defined bypresence of tight EGFR gene clusters and a ratio of EGFRgene to chromosome of$2 or$15 copies of EGFR per cellin$10% of analyzed cells). Cases showing gene amplifica-tion or high polysomy were considered to be FISH positive.

2.5. Statistical analysis

Statistical analysis was carried out by SPSS software,version 10 (SPSS, Chicago, IL). The chi-square test andthe Fisher exact test were used to examine the correlationbetween immunohistochemical expression of EGFR andgene copy number of EGFR detected by FISH. Statisticalsignificance was set at P ! 0.05.

3. Results

3.1. Study group

All patients were women, with a mean age of 50.97 years(range, 25e83 years). We detected 62 cases, including 55cases (88.8%) of invasive ductal carcinomaenot otherwisespecified, 2 cases (3.2%) ofmetaplastic breast cancer, 2 cases(3.2%) of mixed (ductal and lobular), and 3 cases (4.8%) ofother pathologic subtypes (lobular, mucinous, and apocrine).All of the tumor tissues were revised for hormone receptorstatus and were found to be ER, PR, and cerb-B2 negative.Mean tumor diameter was 2.81 cm. According to themodified Bloom-Richardson grading system, we found 53grade III cases (85.5%) and 9 grade II cases (14.5%) [17].

3.2. EGFR expression level and patterns

During the immunohistochemical process, 3 tissues forEGFR.31G7 clone testing and 13 tissues for EGFR.25clone testing were lost. As a result, 59 cases could be eval-uated with the EGFR.31G7 clone, and 49 cases could beevaluated with the EGFR.25 clone.

Thirty-eight of 62 (61.4%) EGFR.31G7-clone-positivecases were detected. Fourteen of 62 cases (22.6%) showedincomplete membranous, 12 of 62 (19.4%) showed completemembranous, and 12 of 62 (19.4%) showed cytoplasmic

Table 1

Staining patterns of the two EGFR clones (P 5 0.001)

EGFR.31G7 clone

Negative Inc

EGFR.25 clone Negative 10 (20.4%) 0

Incomplete membranous 2 (4.1%) 2 (4

Complete membranous 2 (4.1%) 8 (1

Cytoplasmic 4 (8.2%) 1 (2

Abbreviation: EGFR, epidermal growth factor receptor.

staining pattern. There were 24 of 62 (38.7%) immunohisto-chemically EGFR.31G7 negative cases.

Among 38 of 49 (77.6%) EGFR.25-clone-positive cases,7 of 49 (14.3%) exhibited cytoplasmic, 10 of 49 (20.4%)incomplete membranous, and 21 of 49 (42.9%) completemembranous staining pattern. Eleven of 49 (22.4%) wereimmunohistochemically negative with the EGFR.25 clone.

3.3. Comparison of different EGFR clones

Among the 18 of 49 EGFR.31G7-negative cases (36.7%),10 of 49 (20.4%) showed negative reaction, 4 of 49 (8.2%)showed cytoplasmic, 2 of 49 (4.1%) showed incompletemembranous, and 2 of 49 (4.1%) showed complete membra-nous staining pattern with the EGFR.25 clone.

Of 49 cases, 1 (2%) exhibited negative and 2 (4.1%) ex-hibited positive cytoplasmic reaction; and 3 (6.1%) ex-hibited incomplete membranous and 6 (12.2%) exhibitedcomplete membranous reaction with the EGFR.25 clone.Twelve of 49 samples (24.5%) stained cytoplasmic withthe EGFR.31G7 clone.

Among 11 of 49 cases (22.4%) that were designated anincomplete membranous staining pattern with theEGFR.31G7 clone, 1 of 49 (2%) showed cytoplasmic, 2of 49 (4.1%) showed incomplete membranous, and 8 of49 (16.3%) showed complete membranous staining pattern.

Five of 49 cases (10.2%) represented complete membra-nous staining pattern with both of the EGFR clones.However, 3 of 49 cases (6.1%) demonstrated incompletepattern with the EGFR.25 and complete pattern with theEGFR.31G7 clone.

The results of the EGFR antibodies is in concordancewith each other (P 5 0.001). Table 1 compares the stainingpatterns of the two different clones.

According to any cytoplasm and/or membrane staining,29 of 49 cases (59.2%) were considered immunohisto-chemically positive with both clones. Likewise, 9 of 49cases (18.4%) displayed negative reaction with both anti-bodies. Nine of 49 cases (18.4%), which were immunohis-tochemically positive with the EGFR.25 clone, exhibitednegative reaction with the EGFR.31G7. And two cases,which were immunohistochemically positive with theEGFR.31G7 clone, exhibited negative reaction with theEGFR.25 clone. We found that the positivity or the nega-tivity of the reaction produced by one of the antibodies isin concordance with the other (P 5 0.001).

omplete membranous Complete membranous Cytoplasmic

0 1 (2.0%)

.1%) 3 (6.1%) 3 (6.1%)

6.3%) 5 (10.2%) 6 (12.2%)

.0%) 0 2 (4.2%)

Table 2

EGFR protein expression pattern and FISH resultsa

Clone

Negative,

n (%)

Incomplete

membranous,

n (%)

Complete

membranous,

n (%)

Cytoplasmic,

n (%)

EGFR.31G7

FISH

negative

21 (35.6) 10 (16.9) 7 (11.9) 11 (18.6)

FISH

positive

0 4 (6.8) 5 (8.5) 1 (1.7)

EGFR.25

FISH

negative

9 (19.1) 10 (21.3) 14 (29.8) 5 (10.6)

FISH

positive

0 0 7 (14.9) 2 (4.3)

Abbreviations: EGFR, epidermal growth factor receptor; FISH, fluores-

cence in situ hybridization.a P 5 0.03 for EGFR.31G7 and P 5 0.013 for EGFR.25.

226 B. Gumuskaya et al. / Cancer Genetics and Cytogenetics 203 (2010) 222e229

The percentage of the immunohistochemically positivecases was 63.3% with the EGFR.31G7 clone and was77.6% with the EGFR.25 clone. Although 24.5% of thecases displayed a cytoplasmic staining pattern, 22.4% ofthe cases displayed incomplete membranous and 16.3%complete membranous staining patterns by the EGFR.31G7clone; 14.3% of the cases exhibited cytoplasmic, 20.4%incomplete membranous, and 42.9% complete membra-nous patterns by the EGFR.25 clone.

3.4. Comparison of EGFR FISH withimmunohistochemistry

Of 62 tissue samples placed on tissue microarrayer, threewere lost during the FISH process. According to Cappuzzoet al., FISH results were as follows: 10 FISH-positive cases(16.1%) were assigned; one case (1.7%) showed amplifica-tion, and the rest showed high polysomy [18]. Of all 49FISH-negative cases (83.1%), diploidy was detected in 29(49.2%), low polysomy in 19 (32.2%), and high trisomyin 1 (1.7%). Representative cases displaying EGFR geneamplification and disomy are shown in Figures 1E and 1F.

The association of EGFR.31G7 and EGFR.25 expres-sion patterns and the FISH results are listed in Table 2.All FISH-positive cases were also immunohistochemistry(IHC) positive (P 5 0.01). The one with amplification

Table 3

FISH and IHC results with the EGFR.31G7 clone

EGFR.31G7 clone IHC pattern, n (%)

EGFR FISH Negative Incomplete membranous

Diploid 15 (25.4) 5 (8.5)

Low polysomy 5 (8.5) 5 (8.5)

High trisomy 1 (1.7)

High polysomy 4 (6.8)

Amplification

Total 21 (35.6) 14 (23.8)

Abbreviations: FISH, fluorescence in situ hybridization; IHC, immunohistoc

showed strong complete membranous staining with boththe EGFR.31G7 and EGFR.25 clones, and was grade IIIinvasive ductal carcinoma.

Among the 8 high-polysomic cases, 6 (75%) revealedcomplete membranous and 2 (25%) cytoplasmic stainingpattern with the EGFR.25 clone. Four cases (44.4%) withhigh polysomy displayed incomplete membranous stainingpattern, and 4 cases (44.4%) displayed complete membra-nous staining pattern with the EGFR.31G7 clone (P 50.009). However, 1 of the 9 FISH-positive cases (11.1%)showed cytoplasmic expression of EGFR.31G7 clone.Of the 9 high polysomic cases, 6 (10.2%) were invasiveductal carcinoma, and there was 1 case each (1.7%) ofmetaplastic carcinoma, apocrine carcinoma, and myoepi-thelial carcinoma.

In FISH-negative EGFR.31G7-clone-expressing cases,14 (50%) showed diploidy and 14 (50%) showed low pol-ysomy. Among the 14 FISH-negative/EGFR.31G7-positive diploid cases, 7 (50%) displayed cytoplasmic, 5(35.7%) incomplete membranous, and 2 (14.3%) completemembranous staining patterns. There were 19 diploid cases,in which 5 (26.3%) displayed cytoplasmic, 8 (42.1%)incomplete membranous, and 6 (31.6%) complete membra-nous staining patterns with the EGFR.25 clone.

Of the 12 cases exhibiting complete membranous stain-ing pattern with the EGFR.31G7 clone, 2 (16.7%) werediploid, 5 (41.7%) were low polysomic, 4 (33.3%) werehigh polysomic, and 1 (8.3%) was amplified (P 5 0.035).Twenty-one cases were found to be completely membra-nous by the EGFR.25 clone. Among these, 6 cases(28.6%) were diploid, 8 (38%) low polysomic, 6 (28.6%)high polysomic, and 1 case (4.8%) amplified (P 5 0.026).

Incomplete membranous staining was determined in 14cases with the EGFR.31G7 clone and in 10 cases withthe EGFR.25 clone. Among the former, 5 cases (35.7%)were diploid, 5 (35.7%) low polysomic, and 4 (28.6%) highpolysomic. With EGFR.25, 8 cases (80%) were diploid and2 (20%) were low polysomic. Tables 3 and 4 illustrate FISHand IHC results of EGFR.31G7 and EGFR.25. In caseswith cytoplasmic expression with the EGFR.31G7 clone,6 cases (54.5%) were diploid, 4 (36.4%) low polysomic,and 1 (9.1%) high polysomic. In cases with cytoplasmicexpression with the EGFR.25 clone, 5 cases (71.4%) werediploid and 2 (28.6%) were high polysomic. There were 21

Complete membranous Cytoplasmic Total, n (%)

2 (3.4) 7 (11.9) 29 (49.2)

5 (8.5) 4 (6.8) 19 (32.2)

1 (1.7)

4 (6.8) 1 (1.7) 9 (15.3)

1 (1.7) 0 1 (1.7)

12 (20.3) 12 (20.3) 59 (100)

hemistry; EGFR, epidermal growth factor receptor.

Table 4

FISH and IHC results with the EGFR.25 clone

EGFR.25 clone IHC pattern, n (%)

EGFR FISH Negative Incomplete membranous Complete membranous Cytoplasmic Total, n (%)

Diploid 4 (8.5) 8 (17) 6 (12.8) 5 (10.6) 23 (48.9)

Low polysomy 4 (8.5) 2 (4.3) 8 (17) 14 (29.8)

High trisomy 1 (2.1) 1 (2.1)

High polysomy 6 (12.8) 2 (4.3) 8 (17)

Amplification 1 (2.1) 1 (2.1)

Total 9 (19.1) 10 (21.3) 21 (44.7) 7 (14.9) 47 (100)

Abbreviations: FISH, fluorescence in situ hybridization; IHC, immunohistochemistry; EGFR, epidermal growth factor receptor.

227B. Gumuskaya et al. / Cancer Genetics and Cytogenetics 203 (2010) 222e229

of 59 negative cases (35.6%) of EGFR.31G7 clone, and 9of 47 negative cases (19.1%) of EGFR.25 clone. Fifteencases (71.4%) of the former showed diploidy, 5 (23.8%)displayed low polysomy, and 1 (4.8%) showed hightrisomy. Among the EGFR.25-clone-negative cases, 4 dip-loidic cases (44.4%), 4 low polysomic cases (44.4%), and 1high trisomic case (2.1%) were detected. High polysomyand low polysomy were found in two metaplastic carci-nomas. One of the metaplastic carcinomas showed anincomplete membranous staining pattern, and the othershowed cytoplasmic staining pattern with the EGFR.31G7clone. Both of them displayed complete membranous stain-ing pattern with the EGFR.25 clone.

Seven of the nine FISH-positive cases had incompleteand/or complete membranous staining with both antibodies.Among 9 of 47 FISH-positive cases, 7 of 47 of them had anincomplete and/or complete membranous staining patternwith EGFR.25. In addition, in 10 of 59 of the FISH-positive cases, 9 were IHC positive with the EGFR.31G7clone. Eleven of the 38 FISH-negative cases had membra-nous staining with both antibodies. In Table 5, the stainingpatterns of the FISH-positive and -negative cases with eachand both of the antibodies are listed.

Among FISH-negative cases, 57.1% of cases were posi-tive with the EGFR.31G7 clone and 76.3% of cases werepositive with the EGFR.25 clone. All FISH-positive caseswere EGFR positive for both clones. We found that theEGFR expression pattern correlates with FISH status,regardless of EGFR clone (P 5 0.03 for EGFR.31G7 andP 5 0.013 for EGFR.25).

Table 5

Staining patterns of FISH-positive and -negative cases with each and both of the

EGFR.25

Staining pattern FISH negative FISH posit

Negative 9/38 0

Cytoplasmic 5/38 2/9

Incomplete membranous 10/38 0

Complete membranous 14/38 7/9

Complete and/or incomplete membranous 24/38 7/9

IHC positive (any cytoplasmic or membranous) 29/38 9/9

Abbreviations: FISH, fluorescence in situ hybridization; EGFR, epidermal gr

4. Discussion

Anti-EGFR therapies have shown to provide longersurvival times in selected cases of colon carcinoma,NSCLC, and squamous-cell carcinoma of the head andneck [10].

Among breast carcinomas, TNBCs, most of which over-lap with basal-like breast tumors, show significant EGFRoverexpression. This group of tumors has poor prognosisand occurs at a higher prevalence among premenopausaland young patients [3]. Because most of these cases areER negative, PR negative, and HER-2/neu negative, currenthormone therapies and targeted therapies against HER-2/neu are not beneficial. Therefore, anti-EGFR therapies arepotentially promising in this particular group of cases.However, tests for predicting response to anti-EGFR thera-pies are still controversial.

Presence of activating EGFR mutations is an effectivepredictor of response to anti-EGFR therapy [19e21].Approximately 90% of EGFR mutations in NSCLC consistsof short in-frame deletions of exon 19 and a point mutation inexon 21 at nucleotide 2573 (L858R mutation). In severalrecent prospective clinical trials of EGFR tyrosine kinaseinhibitors, response rates of EGFR-mutated NSCLC casesranged 75e78% [22e24]. Half of the EGFR-mutated casesshow an increase in EGFR gene copy number; conversely,75% of the cases with increased EGFR gene copy numberdisplay mutations [25]. One study revealed that in bron-chioalveolar carcinoma, increased EGFR gene copy numberdetected by FISH is associated with improved survival whenmanagement includes gefitinib therapy [26,27]. In NSCLC,

antibodies

EGFR.31G7 Both antibodies

ive FISH negative FISH positive FISH negative FISH positive

21/49 0 8/38 0

11/49 1/10 1/38 1/9

10/49 4/10 2/38 0

7/49 5/10 1/38 4/9

17/49 9/10 11/38 7/9

28/49 10/10 22/38 9/9

owth factor receptor; IHC, immunohistochemistry.

228 B. Gumuskaya et al. / Cancer Genetics and Cytogenetics 203 (2010) 222e229

EGFR FISH was reported to be a predictive factor forselection of patients for the EGFR monoclonal antibodyinhibitor cetuximab plus chemotherapy [27].

Bhargava et al. detected no EGFR gene mutation ininvasive breast carcinomas [12]. The screened EGFR muta-tions were the ones that detected most frequently in lungadenocarcinomas. Gilbert et al. evaluated EGFR exons18, 19, and 21 and KIT exons 9, 11, 13, and 17 for muta-tions in 38 metaplastic breast carcinoma cases, and theyfound that activating EGFR mutations were absent [14].In our study, activating mutations were not investigated.

IHC alone is limited in defining the group of patientswho will benefit from anti-EGFR therapy [28]. However,in one report, the protein expression by IHC in combinationwith the EGFR gene copy number status was determined tobe an effective predictor of gefitinib therapy in patientswith NSCLC [26]. It was claimed that EGFR FISH andIHC alone or in combination was capable of predictingthe outcome, independent of the presence of EGFR muta-tion, KRAS mutation, or pAKT expression. AlthoughEGFR-negative/FISH-negative cases showed poor outcomedespite presence of EGFR mutations and absence of KRASmutations or pAKT expression, EGFR-positive/FISH-positive cases showed excellent outcome regardless ofEGFR mutation, KRAS mutation, or pAKT expression.By means of combination of EGFR FISH and IHC nega-tivity, they were able to identify ~30% of non-small-celllung cancer patients who had no clinical benefit from gefi-tinib treatment [26].

There are three previous studies in the literature whereEGFR gene copy number of metaplastic breast carcinomasthat exhibit many characteristics of basal-like breast carci-nomas was investigated. One study revealed high EGFRcopy number detected by chromogenic in situ hybridization(CISH) [13]. However, the other two studies detected lowamplification rate by FISH, in concordance with our study.Bhargava et al. reported EGFR gene amplification in 6% ofthe breast tumors by CISH [12]. Toyama et al. reportedincreased EGFR gene copy number by real-time reversetranscription (RT)-PCR in 21% of TNBC and found nosignificant correlations between EGFR protein expressionand increased EGFR gene copy number [16]. In our study,the EGFR expression was associated with increased genecopy number (P 5 0.03). In real-time RT-PCR and CISHmethods, the quantity of chromosome 7 is not taken intoaccount. Previous studies on metaplastic breast carcinomasby CISH revealed high EGFR gene copy number like thestudy done by real-time RT-PCR method. However, thestudies like ours, in which FISH was used, revealed thatan increase in EGFR gene copy number correlates with pol-ysomy. Consequently, dual-probe tests, including chromo-some 7 copy number, are appropriate for detecting EGFRgene status. To our knowledge, this report is novel in thatit revealed the EGFR gene status of TNBC by calculatingthe ratio of number of gene copies to number of chromo-some 7 copies per each tissue.

Rare EGFR gene amplification (1.7%) was found in ourstudy, as was reported in a study in metaplastic breast carci-nomas [14]. An increase in gene copy number was mostlyfound associated with chromosome 7 polysomy. Half of theFISH-negative cases showing EGFR.31G7 were diploid,and 65.5% of the FISH-negative and EGFR.25-positivecases were diploid. Sixty percent of the FISH-negativeand EGFR.31G7-positive cases demonstrated a membra-nous staining pattern. The 88.9% of the FISH-negativeand EGFR.25-positive cases showed a membranous stain-ing pattern. An alternative splicing of EGFR RNA causingan in-frame deletion that may be expressed at the cellmembrane has been recently reported in glial tumors[29]. It seems that posttranslational events similar to thismay be responsible for overexpression of EGFR withoutgene copy gain in some TNBCs. Future studies investi-gating details of the molecular mechanisms of EGFRexpression in TNBCs can provide valuable information topredict the response to anti-EGFR treatment.

In the other hand, it is well known that different clonesof an antibody can provide different immunostainingresults. In a similar study that compared immunostainingof different HER-2 clones with FISH results, the authorsused two kinds of monoclonal antibodies: CB11and 4D5.Whereas CB11 antibody was positive in 25 of the 86 cases(29.1%), 4D5 antibody was positive in only 15 (17.4%). Allcases positive for CB11 or 4D5 were HercepTest positive.In three FISH-positive cases, the IHC showed no overex-pression with any of the antibodies [30]. In our study,among FISH-positive cases, although one case showedcytoplasmic staining, four showed incomplete membranousstaining and five showed complete membranous stainingpatterns with the EGFR.31G7 clone. Two cases displayedcytoplasmic staining patterns, and seven cases displayedcomplete membranous staining patterns with the EGFR.25clone. When only the complete staining pattern is evalu-ated, the EGFR.25 clone can predict 77.8% of FISH-positive cases (P 5 0.013), and the EGFR.31G7 clonecan predict only 50%. If complete and incomplete membra-nous staining patterns are both taken into account, thepredictive rate of EGFR.31G7 clone would be 90% (P 50.003). However, the false-positive ratio of membranousexpression by IHC, to predict FISH status, would be 65%(17 of 26) for EGFR.31.G7 and 77% for EGFR.25.

Trials of anti-EGFR therapy are progressing. Our resultsfound that EGFR immunohistochemical staining patterns(especially membranous and incomplete membranouspatterns) are associated with increased gene copy numberin TNBCs.

When the absence or lack of activating mutations inbreast tumors in the literature is taken into consideration,detecting an increase in gene copy number by FISH mighthelp estimate response to anti-EGFR therapy. However,increase in gene copy number is rare and amplification iseven rarer in breast cancer. Membranous staining of EGFRby IHC alone might be actually the best way to decide

229B. Gumuskaya et al. / Cancer Genetics and Cytogenetics 203 (2010) 222e229

patient eligibility for anti-EGFR therapy in breast cancer.However, we need future studies that test or generate addi-tional EGFR-specific antibodies.

Acknowledgments

We are grateful to A. Gunay, who helped prepare thetissue microarray blocks; C. Karakoc for performing theIHC staining and FISH; and Lokman Kale for archivalassistance. This study is supported by a research and educa-tion grant from Committee of Educational Coordination ofAnkara Diskapi Hospital.

References

[1] Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics,

2009. CA Cancer J Clin 2009;59:225e49.[2] Rakha EA, El-Rehim DA, Paish C, Green AR, Lee AH, Robertson JF,

et al. Basal phenotype identifies a poor prognostic subgroup of breast

cancer of clinical importance. Eur J Cancer 2006;42:3149e56.

[3] Carey LA, Perou CM, Livasy CA, Dressler LG, Cowan D,

Conway K, et al. Race, breast cancer subtypes, and survival in the

Carolina Breast Cancer Study. JAMA 2006;295:2492e502.

[4] Fulford LG, Easton DF, Reis-Filho JS, Sofronis A, Gillett CE,

Lakhani SR, et al. Specific morphological features predictive for

the basal phenotype in grade 3 invasive ductal carcinoma of breast.

Histopathology 2006;49:22e34.

[5] Siziopikou KP, Cobleigh M. The basal subtype of breast carcinomas

may represent the group of breast tumors that could benefit from

EGFR-targeted therapies. Breast 2007;16:104e7.

[6] Livasy CA, Karaca G, Nanda R, Tretiakova MS, Olopade OI,

Moore DT, et al. Phenotypic evaluation of the basal-like subtype of

invasive breast carcinoma. Mod Pathol 2006;19:264e71.

[7] Cheang MC, Voduc D, Bajdik C, Leung S, McKinney S, Chia SK,

et al. Basal-like breast cancer defined by five biomarkers has superior

prognostic value than triple-negative phenotype. Clin Cancer Res

2008;14:1368e76.

[8] Nielsen TO, Hsu FD, Jensen K, Cheang M, Karaca G, Hu Z, et al.

Immunohistochemical and clinical characterization of the basal-like

subtype of invasive breast carcinoma. Clin Cancer Res 2004;10:

5367e74.

[9] Rakha EA, Tan DS, Foulkes WD, Ellis IO, Tutt A, Nielsen TO, et al.

Are triple-negative tumours and basal-like breast cancer synony-

mous? Breast Cancer Res 2007;9:404.

[10] Vokes EE, Chu E. Anti-EGFR therapies: clinical experience in colo-

rectal, lung, and head and neck cancers. Oncology (Williston Park)

2006;20:15e25.

[11] Hirsch FR, Varella-Garcia M, Bunn PA Jr, Franklin WA,

Dziadziuszko R, Thatcher N, et al. Molecular predictors of outcome

with gefitinib in a phase III placebo-controlled study in advanced

non-small-cell lung cancer. J Clin Oncol 2006;24:5034e42.

[12] Bhargava R, Gerald WL, Li AR, Pan Q, Lal P, Ladanyi M, et al.

EGFR gene amplification in breast cancer: correlation with epidermal

growth factor receptor mRNA and protein expression and HER-2

status and absence of EGFR-activating mutations. Mod Pathol

2005;18:1027e33.

[13] Reis-Filho JS, Pinheiro C, Lambros MB, Milanezi F, Carvalho S,

Savage K, et al. EGFR amplification and lack of activating mutations

in metaplastic breast carcinomas. J Pathol 2006;209:445e53.

[14] Gilbert JA, Goetz MP, Reynolds CA, Ingle JN, Giordano KF,

Suman VJ, et al. Molecular analysis of metaplastic breast carcinoma:

high EGFR copy number via aneusomy. Mol Cancer Ther 2008;7:

944e51.

[15] Gwin K, Lezon-Geyda K, Harris L, Tavassoli FA. Chromosome 7

aneusomy in metaplastic breast carcinomas with chondroid, squa-

mous, and spindle-cell differentiation. Int J Surg Pathol (In press).

[16] Toyama T, Yamashita H, Kondo N, Okuda K, Takahashi S, Sasaki H,

et al. Frequently increased epidermal growth factor receptor (EGFR)

copy numbers and decreased BRCA1 mRNA expression in Japanese

triple-negative breast cancers. BMC Cancer 2008;8:309.

[17] Tavassoli FA, Devilee P. Pathology and genetics of the breast and

female genital organs. In: World Health Organization classification

of tumours. Lyon: IARC Press, 2003.

[18] Cappuzzo F, Hirsch FR, Rossi E, Bartolini S, Ceresoli GL, Bemis L,

et al. Epidermal growth factor receptor gene and protein and gefitinib

sensitivity in non-small-cell lung cancer. J Natl Cancer Inst 2005;97:

643e55.

[19] Lynch TJ, Bell DW, Sordella R, Gurubhagavatula S, Okimoto RA,

Brannigan BW, et al. Activating mutations in the epidermal growth

factor receptor underlying responsiveness of non-small-cell lung

cancer to gefitinib. N Engl J Med 2004;350:2129e39.

[20] Paez JG, Janne PA, Lee JC, Tracy S, Greulich H, Gabriel S, et al.

EGFR mutations in lung cancer: correlation with clinical response

to gefitinib therapy. Science 2004;304:1497e500.

[21] Pao W, Miller V, Zakowski M, Doherty J, Politi K, Sarkaria I, et al.

EGF receptor gene mutations are common in lung cancers from

‘‘never smokers’’ and are associated with sensitivity of tumors to ge-

fitinib and erlotinib. Proc Natl Acad Sci U S A 2004;101:13306e11.

[22] Inoue A, Suzuki T, Fukuhara T, Maemondo M, Kimura Y,

Morikawa N, et al. Prospective phase II study of gefitinib for

chemotherapy-naive patients with advanced non-small-cell lung

cancer with epidermal growth factor receptor gene mutations. J Clin

Oncol 2006;24:3340e6.

[23] Sunaga N, Tomizawa Y, Yanagitani N, Iijima H, Kaira K, Shimizu K,

et al. Phase II prospective study of the efficacy of gefitinib for the

treatment of stage III/IV nonesmall cell lung cancer with EGFR

mutations, irrespective of previous chemotherapy. Lung Cancer

2007;56:383e9.

[24] Sutani A, Nagai Y, Udagawa K, Uchida Y, Koyama N, Murayama Y,

et al. Gefitinib for non-small-cell lung cancer patients with epidermal

growth factor receptor gene mutations screened by peptide nucleic

acidelocked nucleic acid PCR clamp. Br J Cancer 2006;95:1483e9.

[25] Li AR, Chitale D, Riely GJ, Pao W, Miller VA, Zakowski MF, et al.

EGFR mutations in lung adenocarcinomas: clinical testing experi-

ence and relationship to EGFR gene copy number and immunohisto-

chemical expression. J Mol Diagn 2008;10:242e8.

[26] Hirsch FR, Varella-Garcia M, Cappuzzo F, McCoy J, Bemis L,

Xavier AC, et al. Combination of EGFR gene copy number and

protein expression predicts outcome for advanced non-small-cell lung

cancer patients treated with gefitinib. Ann Oncol 2007;18:752e60.

[27] Hirsch FR, Herbst RS, Olsen C, Chansky K, Crowley J, Kelly K,

et al. Increased EGFR gene copy number detected by fluorescent in

situ hybridization predicts outcome in non-small-cell lung cancer

patients treated with cetuximab and chemotherapy. J Clin Oncol

2008;26:3351e7.

[28] Zakowski MF, Hussain S, Pao W, Ladanyi M, Ginsberg MS,

Heelan R, et al. Morphologic features of adenocarcinoma of the lung

predictive of response to the epidermal growth factor receptor kinase

inhibitors erlotinib and gefitinib. Arch Pathol Lab Med 2009;133:

470e7.[29] Ekstrand AJ, Sugawa N, James CD, Collins VP. Amplified and

rearranged epidermal growth factor receptor genes in human glio-

blastomas reveal deletions of sequences encoding portions of the

N- and/or C-terminal tails. Proc Natl Acad Sci U S A 1992;89:

4309e13.

[30] Gouvea AP, Milanezi F, Olson SJ, Leitao D, Schmitt FC, Gobbi H.

Selecting antibodies to detect HER-2 overexpression by immunohis-

tochemistry in invasive mammary carcinomas. Appl Immunohisto-

chem Mol Morphol 2006;14:103e8.