ARTICLE IN PRESS

Pathology – Research and Practice 206 (2010) 39–42

Contents lists available at ScienceDirect

Pathology – Research and Practice

0344-03

doi:10.1

� Corr

E-m

journal homepage: www.elsevier.de/prp

Original Article

HER2-testing in 538 consecutive breast cancer cases using FISH andimmunohistochemistry

Aniko Kov�acs �, Goran Stenman

Department of Pathology, Sahlgrenska University Hospital, Gothenburg SE-413 45, Sweden

a r t i c l e i n f o

Article history:

Received 15 April 2009

Received in revised form

29 June 2009

Accepted 5 August 2009

Keywords:

HER2

Breast cancer

Immunohistochemistry

HercepTest

FISH

38/$ - see front matter & 2009 Elsevier Gmb

016/j.prp.2009.08.003

esponding author. Tel.: +46 313426162; fax:

ail address: aniko.kovacs@vgregion.se (A. Kov

a b s t r a c t

HER2 is amplified and overexpressed in up to 20% of all breast carcinomas, and accurate assessment of

the HER2-status is essential for the appropriate use of anti-HER2 therapies, including trastuzumab.

Here, we have evaluated the accuracy of two commercially available and commonly used HER2

antibodies (antibody A0485 and HercepTest), followed by FISH-test, in 538 consecutive breast cancers,

to identify a reliable procedure for the determination of HER2-status. Immunohistochemical (IHC)

analysis revealed that 282 cases (52.2%) were positive (2+/3+) for antibody A0485, while only 179 cases

(33.3%) were positive (2+/3+) for HercepTest (81.2% concordance between A0485 and HercepTest). FISH-

analysis of the 103 IHC discordant cases revealed 89.3% concordance with HercepTest. The results of the

present study, which is the largest in Scandinavia, revealed a significantly higher concordance rate

between HercepTest and HER2 copy number status than between antibody A0485 and HER2 copy

number status. Eleven cases (11% of 103 discordant cases) with HER2-amplification were positive for

antibody A0485 and false negative for HercepTest. Our findings indicate that HercepTest is a reliable

and accurate IHC-assay and that in the absence of clinical outcome data for IHC-positive/FISH-negative

and IHC-negative/FISH-positive patients, FISH-analysis should be performed on all IHC 2+ and 3+ cases.

& 2009 Elsevier GmbH. All rights reserved.

Introduction

The ERBB2 oncogene (a.k.a. HER2 or NEU) is located onchromosome 17q21 and encodes a 185-kD transmembraneglycoprotein, which belongs to the human epidermal growthfactor receptor family of tyrosine kinases [7]. The gene wasoriginally identified as a mutated transforming gene in a ratneuroblastoma and was shortly thereafter shown to be amplifiedin a human mammary carcinoma cell line [8,21]. Subsequentstudies have demonstrated that HER2 is amplified and over-expressed in up to 20% of breast cancers and that these cancershave a worse biological behavior and prognosis [7,22]. In addition,experimental evidence, together with studies of human breastcancer samples, has shown that amplification of HER2 is an earlyevent in mammary tumorigenesis [7], making HER2 an attractivetherapeutic target. This knowledge has formed the basis for thedevelopment and use of anti-HER2 antibodies (trastuzumab) inthe clinical management of breast cancer patients. Several studieshave shown that trastuzumab (Herceptin; Genentechs, Inc., SouthSan Francisco, CA) in combination with chemotherapy prolongsthe survival of patients with metastatic, HER2-overexpressing

H. All rights reserved.

+46 31827194.�acs).

breast cancer [3,5]. Recent studies have also demonstrateddramatic improvements in disease-free survival when trastuzu-mab is used in an adjuvant therapy setting in combination with orfollowing chemotherapy [17].

The mechanism of action of trastuzumab is still uncertain [16].There is currently only limited evidence supporting the hypoth-esis that trastuzumab downregulates HER2 or inhibits HER2signaling. Instead, several studies have shown that trastuzumab,for example, has antiangiogenic effects, inhibits proteolyticcleavage and shedding of HER2, and induces antibody-dependentcellular toxicity against tumor cells [16].

Accurate assessment of the HER2-status in breast cancer isessential for the appropriate use of trastuzumab. HER2-testing isroutinely performed using immunohistochemical staining of theHER2 protein and/or FISH analysis to determine the genecopy number. However, in a recent guideline for HER2 testing inbreast cancer compiled by the American Society of ClinicalOncology/College of American Pathologists [29], it is concludedthat approximately 20% of current HER2 testing may beinaccurate. One decisive factor in this respect is the choice ofHER2 antibody used for IHC-testing. Here, we have evaluated theaccuracy of two commonly used antibodies, that is the polyclonalantibody A0485 (Dako, Carpenteria, CA) and HercepTest (Dako)followed by FISH-analysis, in 538 consecutive breast cancercases to identify a reliable procedure for the determination ofHER2-status.

ARTICLE IN PRESS

A. Kov�acs, G. Stenman / Pathology – Research and Practice 206 (2010) 39–4240

Materials and methods

Tumor samples

The study material comprised 538 consecutive, unselected in-house cases of primary breast cancers diagnosed at the Depart-ment of Pathology, Sahlgrenska University Hospital, Gothenburg,Sweden, between March 2003 and March 2004. The tumors wereclassified according to histology and graded according to theBloom–Richardson–Elston system. Among the 538 cases, therewere 387 invasive ductal carcinomas (71.9%), 82 invasive lobularcarcinomas (15.2%), 16 invasive tubular carcinomas (3%), 29 ductalcancers in situ (DCIS) (5.4%), 8 invasive mucinous carcinomas(1.5%), 7 invasive medullary carcinomas (1.3%), 6 invasivepapillary carcinomas (1.1%), 2 invasive apocrine carcinomas(0.4%), and 1 invasive metaplastic carcinoma (0.2%). The 538tumors were obtained from 532 female patients (3 patients hadbilateral and 3 had bifocal cancers). There were 163 grade IIItumors (32%), 222 grade II tumors (43.6%), 124 grade I tumors(24.3%), and 29 grade III DCIS tumors (5.4%). Lymph nodemetastases were found in 190 cases (37.3%). Patient age rangedfrom 29 to 93 years, with a mean of 61 years.

Immunohistochemistry (IHC)

Four-mm-thick formalin-fixed paraffin sections were usedfor IHC-staining. For antigen retrieval, sections were boiled in10 mmol/L citrate buffer supplied in the HercepTest kit (Dako).The staining procedure, positive and negative controls, as well asthe scoring methods were carried out as recommended by thesupplier.

Immunostaining with the HER2-specific polyclonal antibodyA0485 (Dako) was performed in the daily routine using theTechMateTM 500Plus automated immunostaining system (Dako).Slides were microwave-treated for 30 min in tissue retrievalsolution (Dako), and thereafter incubated for 40 min at roomtemperature with antibody A0485 at a 1:500 dilution. Withintervening washing steps in phosphate-buffered saline, slideswere subsequently incubated for 30 min at room temperaturein an anti-rabbit-specific labeled polymer (Dako), followed by10 min at 371 in a solution containing 3% hydrogen peroxide and3.30-diaminobenzidine. Slides were finally counterstained withhematoxylin.

HER2-staining was scored according to the guidelines of themanufacturer. Scores 0 and 1+ were considered as negative forHER2 overexpression, and 2+ and 3+ as positive.

Slides were also routinely stained with the following mono-clonal antibodies: Estrogen Receptor Alpha (1:30), ProgesteroneReceptor (1:400), Ki-67 (1:100), and p53 (1:400) (Dako).

During the course of this study, we participated in twodifferent IHC quality control programs for HER2 immunocyto-chemistry, i.e., Nordic immunohistochemical Quality Control

Table 1Score-distribution of the 103 IHC-discordant, amplified, and non-amplified cases.

No. of cases IHC-results

Antibody A 0485 HercepTest

34 1+ and 2+a Negative

49 2+ Negative

8 2+ and 3+a Negative

12 3+ Negative

a Heterogeneous staining pattern interpreted as HER2-positive.

(NordiQC) and National External Quality Assessment Services(NEQAS) for Laboratory Medicine, UK. The quality of ourimmunostainings was approved by both programs.

Fluorescence in situ hybridization (FISH)

FISH analysis was performed using the FDA-approved Path-Vysions HER-2 DNA Probe Kit (Vysis Inc., Downers Grove, IL)according to the manufacturers’ recommended protocol. Theprobe cocktail, including the LSI HER-2/neu probe (labeled withSpectrumOrange) and the CEP 17 probe (labeled with Spectrum-Green), together with the tissue sections, was denatured at 73 1Cfor 5 min and hybridized for 20 h at 37 1C using a HYBriteTM

instrument (Vysis). The nuclei were subsequently counterstainedin blue with 40,60-diamidino-20-phenylindole dihydrochloride(DAPI). Slides were examined using a Zeiss Axioskop 2 mot plusfluorescence microscope equipped with the appropriate filtercombinations. Fluorescence signals were digitalized, processed,and analyzed using the CytoVisionTM image analysis system(Applied Imaging International Ltd., Newcastle-Upon-Tyne, UK).Between 40 and 100 nuclei were scored from each case. Signalratios (HER2:CEP17) of o2 were classified as non-amplified andZ2 as amplified.

Results

HER2 overexpression was demonstrated in 52.2% of the tumors(282/538) using antibody A0485 and in 33.3% (179/538) usingHercepTest. The number of concordant cases showing the samescore with both antibodies (0, 1+, 2+ or 3+) was 274 (50.9%), andthe number of concordant cases showing either positivity (scores2+ and 3+) or negativity (scores 0 and 1+) was 437 (81.2%). All the103 discordant cases were positive for antibody A0485 andnegative for HercepTest (Table 1), including 96 invasivecarcinomas (93.2%), of which 29 (30.2%) had lymph nodemetastases, and 7 were DCIS (6.7%).

Antibody A0485 often produced a stronger but more unevenand heterogeneous membrane staining compared to HercepTest.This sometimes made it more difficult to differentiate between 1+and 2+ using this antibody. In contrast, slides stained withHercepTest only rarely resulted in an ambiguous staining pattern.

FISH analysis was performed on all 103 discordant cases. HER2

amplification was detected in 11 of these cases (10.7%) (Table 1).All amplified tumors stained positive with antibody A0485(six had 2+ and five had 3+) and were negative for HercepTest(Fig. 1). In addition, there was one case with polysomy 17 that waspositive with antibody A0485 (2+) and negative for HercepTest.89.3% of the cases were thus false positive using antibody A 0485,and 10.7% of the cases were false negative using HercepTest.

FISH-results

Amplified Non-amplified Polysomy 17

0 34 0

6 42 1

0 8 0

5 7 0

ARTICLE IN PRESS

Fig. 1. Invasive mucinous carcinoma. (A) positive staining with antibody A0485 (2+ and 3+), (B) negative staining with HercepTest, and (C) amplification of HER2 with FISH-

test (red signals – HER2; green signals – CEP17).

Table 2Percentage of false negative IHC-cases (IHC-negative and FISH-positive) in the

present study and in 10 previously published investigations.

Percentage of

IHC-/FISH+ cases (%)

Number of cases (%) Reference

1.1 561 Lal et al. [10]

2 173 Kobayashi et al. [9]

2.3 215 Lottner et al. [13]

2.8 2963 Yaziji et al. [26]

2.9 303 Birner et al. [2]

4 102 Hammock et al. [6]

9.9 199 Prati et al. [19]

10 529 Mass et al. [14]

10 2279 Lal et al. [11]

11.5 6556 Owens et al. [18]

10.7 538 Present study

A. Kov�acs, G. Stenman / Pathology – Research and Practice 206 (2010) 39–42 41

Discussion

In this study, we evaluated the accuracy of two commerciallyavailable and commonly used HER2 antibodies (antibody A0485and the FDA-approved HercepTest), including FISH analysis ofdiscordant cases, in a series of 538 consecutive breast cancer casesto identify a reliable procedure for the determination of HER2-status. This study, which, to our knowledge, is the largest inScandinavia, revealed a significantly higher concordance ratebetween HercepTest and HER2 copy number status than betweenantibody A0485 and HER2 copy number status.

Because IHC- and FISH-assays measure related but distincttargets, that is protein expression versus gene copy number, onemight expect that they may show some disagreement [28].For example, HER2 overexpression without gene amplificationhas been detected in 3–10% of breast cancers [10]. These caseshave mainly been attributed to false positive immunostaining orchromosome 17 polysomy. Conversely, gene amplification withoutHER2 protein overexpression has been reported in 1–12% of breastcancers (Table 2) [2,6,9,10,11,13,14,18,19,26]. The correspondingfigure in the present study was 10.7%.

In cases with HER2 amplification and negative IHC-staining,the lack of HER2-staining could be due to the type of fixation used,the duration of the fixation (insufficient tissue preservation maylead to protein degradation and reduced sensitivity of IHC), or themethod used for epitope retrieval [23,25]. HER2 may also be

altered as a result of rearrangements of genetically unstable HER2

amplicons. If such a rearrangement affects a part of the HER2 thatencodes the epitope recognized by the antibody, the IHC-stainingmay be negative.

In cases without HER2 amplification, HER2 may be over-expressed due to elevated transcriptional activity induced by, forexample, the transcription factors AP-2 and YY1 [1] or by othertranscriptional and posttranscriptional modifications. An alter-native mechanism for overexpression of HER2 in the absence ofgene amplification is polysomy 17 [27]. In two large studies,polysomy 17 was found in almost half of the non-amplifiedtumors that were IHC 3+ [10,24]. In the present study, however,we only found one tumor with polysomy 17 among 92 HER2-positive (A0485)/HercepTest-negative, non-amplified tumors(Table 1).

A plausible explanation for the high percentage of HER2-positivity observed with antibody A0485 (52.2% compared to33.3% with HercepTest) in our study is that this polyclonalantibody has higher sensitivity than other commercially availableantibodies. Thus, Field et al. [4] found that 53% of breast cancerswere HER2-positive with antibody A0485, whereas only 21% werepositive with HercepTest. In contrast, Lebeau et al. [12] observedHER2 overexpression in 42% of the cases with both antibodyA0485 and HercepTest. The antibodies most commonly used forthe detection of HER2 protein expression (e.g. CB11/Novocastra,TAB250/Zymed, A8010/Oncor, A0485/Dako, and HercepTest/Dako)recognize different epitopes located in either the extra- or intra-cellular domains of the HER2 receptor protein. Shedding of theextracellular domain may result in negative immunostaining ifantibodies directed against this domain are used [12]. However,both antibodies used in this study are directed against epitopes inthe intracellular domain of HER2, and hence this phenomenoncannot explain the differences between the two antibodies. Thedifferences observed may instead at least partly be explainedby the fact that several large tumors contained necrotic partsdue to insufficient vascularization or preoperative neoadjuvantchemotherapy, leading to partial degradation of the HER2 protein,which may only be detected by the more sensitive A0485antibody.

In summary, we have found a significantly higher concordancerate between HercepTest and HER2 copy number status thanbetween antibody A0485 and HER2 copy number status in a seriesof 538 breast cancers. Eleven cases with HER2-amplification

ARTICLE IN PRESS

A. Kov�acs, G. Stenman / Pathology – Research and Practice 206 (2010) 39–4242

(10.7% of 103 discordant cases) were positive for antibody A0485and false negative for HercepTest. Our findings are in line with arecent consensus report by the American Society of ClinicalOncology/College of American Pathologists concluding that avail-able data do not clearly demonstrate the superiority of either IHCor FISH as a predictor of benefit from anti-HER2 therapy [29].Moreover, reliable information about how IHC-positive/FISH-negative and IHC-negative/FISH-positive breast cancer patientsrespond to anti-HER2 therapy is not yet available. In the absenceof such information, we suggest that FISH-analysis should beperformed on all IHC 2+ and 3+ cases [20]. This conclusion is alsoin line with recent studies demonstrating that FISH-test is morestrongly correlated with responsiveness to trastuzumab than IHC[15,20].

Acknowledgements

We thank Barbro Wedell for her technical assistance and UlricPedersen for his help in preparing the illustration. The authorsdeclare that they have received financial support for this studyfrom Roche AB, Sweden and Dako, Sweden. This publication onlyreflects the authors’ opinions and has not been submitted toRoche or Dako for approval.

References

[1] D.Y. Begon, L. Delacroix, D. Vernimmen, P. Jackers, R. Winkler, Yin Yang, 1cooperates with activator protein 2 to stimulate ERBB2 gene expression inmammary cancer cells, J. Biol. Chem. 280 (2005) 24428–24434.

[2] P. Birner, G. Oberhuber, J. Stani, C. Reithofer, H. Samonigg, H. Hausmaninger, E.Kubista, W. Kwasny, D. Kandioler-Eckerberger, M. Gnant, R. Jakesz, Evaluationof the United States Food and Drug Administration-approved Scoring and TestSystem of HER-2 protein expression in breast cancer, Clin. Cancer Res. 7(2001) 1669–1675.

[3] H.J. Burstein, A. Keshaviah, A.D. Baron, R.D. Hart, R. Lambert-Falls, P.K.Marcom, R. Gelman, E.P. Winer, Trastuzumab plus vinorelbine or taxanechemotherapy for HER2-overexpressing metastatic breast cancer: thetrastuzumab and vinorelbine or taxane study, Cancer 110 (2007) 965–972.

[4] A.S. Field, N.L. Chamberlain, Tran, A.L. Morey, Suggestions for HER-2/neutesting in breast carcinoma, based on a comparison of immunohistochemistryand fluorescence in situ hybridisation, Pathology 33 (2001) 278–282.

[5] A.M. Gonzalez-Angulo, G.N. Hortob�agyi, F.J. Esteva, Adjuvant therapy withTrastuzumab for HER-2/neu-positive breast cancer, Oncologist 11 (2006)857–867.

[6] L. Hammock, M. Lewis, C. Phillips, C. Cohen, Strong HER-2/neu proteinoverexpression by immunohistochemistry often does not predict oncogeneamplification by fluorescence in situ hybridization, Hum. Pathol. 34 (2003)1043–1047.

[7] N.E. Hynes, D.F. Stern, The biology of erbB-2/neu/HER-2 and its role in cancer,Biochim. Biophys. Acta 1198 (1994) 165–184.

[8] C.R. King, M.H. Kraus, S.A. Aaronson, Amplification of a novel c-erbB-relatedgene in a human mammary carcinoma, Science 229 (1985) 974–976.

[9] M. Kobayashi, A. Ooi, Y. Oda, I. Nakanishi, Protein overexpression and geneamplification of c-erbB-2 in breast carcinomas: a comparative study ofimmunohistochemistry and fluorescence in situ hybridization of formalin-fixed, paraffin-embedded tissues, Hum. Pathol. 33 (2002) 21–28.

[10] P. Lal, P.A. Salazar, M. Ladanyi, B. Chen, Impact of polysomy 17 on HER-2/neuimmunohistochemistry in breast carcinomas without HER-2/neu geneamplification, J. Mol. Diagn. 5 (2003) 155–159.

[11] P. Lal, P.A. Salazar, C.A. Hudis, M. Ladanyi, B. Chen, HER-2 testing in breastcancer using immunohistochemical analysis and fluorescence in situhybridisation: a single-institution experience of 2279 cases and comparisonof dual-color and single-color scoring, Am. J. Clin. Pathol. 121 (2004) 631–636.

[12] A. Lebeau, D. Deimling, C. Kaltz, A. Sendelhofert, A. Iff, B. Luthardt, M. Untch,U. Lohrs, HER-2/neu analysis in archival tissue samples of human breastcancer: comparison of immunohistochemistry and fluorescence in situhybridisation, J. Clin. Oncol. 19 (2001) 354–363.

[13] C. Lottner, S. Schwarz, S. Diermeier, A. Hartmann, R. Knuechel, F. Hofstaedter,G. Brockhoff, Simultaneous detection of HER2/neu gene amplification andprotein overexpression in paraffin-embedded breast cancer, J. Pathol. 205(2005) 577–584.

[14] R.D. Mass, C. Sanders, K. Charlene, L. Johnson, The concordance between theclinical trial assay (CTA) and fluorescence in situ hybridization (FISH) in theHerceptin Pivotal Trials, Proc. Am. Soc. Clin. Oncol. 1 (2000) 75.

[15] R.D. Mass, M.F. Press, S. Anderson, M.A. Cobleigh, C.L. Vogel, N. Dybdal, G.Leiberman, D.J. Slamon, Evaluation of clinical outcomes according to HER2detection by fluorescence in situ hybridization in women with metastaticbreast cancer treated with trastuzumab, Clin. Breast Cancer 6 (2005)240–246.

[16] M.M. Moasser, Targeting the function of the HER2 oncogene in human cancertherapeutics, Oncogene 26 (2007) 6577–6592.

[17] R. Nishimura, Y. Okumura, N. Arima, Trastuzumab monotherapy versuscombination therapy for treating recurrent breast cancer: time to progressionand survival, Breast Cancer 15 (2008) 57–64.

[18] M.A. Owens, B.C. Horten, M.M. Da Silva, HER2 amplification ratios byfluorescence in situ hybridization and correlation with immunohistochem-istry in a cohort of 6556 breast cancer tissues, Clin. Breast Cancer 5 (2004)63–69.

[19] R. Prati, S.K. Apple, J. He, J.A. Gornbein, H.R. Chan, Histopathologiccharacteristics predicting HER-2/neu amplification in breast cancer, Breast J.11 (2005) 433–439.

[20] G. Sauter, J. Lee, J.M.S. Bartlett, D.J. Slamon, M.F. Press, Guidelines for humanepidermal growth factor receptor 2 testing: biologic and methodologicconsiderations, J. Clin. Oncol. 27 (2009) 1323–1333.

[21] C. Shih, L.C. Padhy, M. Murray, R.A. Weinberg, Transforming genes ofcarcinomas and neuroblastomas introduced into mouse fibroblasts, Nature290 (1981) 261–264.

[22] D.J. Slamon, G.M. Clark, S.G. Wong, W.J. Levin, A. Ullrich, W.L. McGuire,Human breast cancer: correlation of relapse and survival with amplificationof the HER-2/neu oncogene, Science 235 (1987) 177–182.

[23] I. Van den Bempt, V. Vanhentenrijk, M. Drijkoningen, I. Wlodarska, P.Vandenberghe, C. De Wolf-Peeters, Real-time reverse transcription-PCR andfluorescence in-situ hybridization are complementary to understand themechanisms involved in HER-2/neu overexpression in human breastcarcinomas, Histopathology 46 (2005) 431–441.

[24] D. Varshney, Y.Y. Zhou, S.A. Geller, R. Alsabeh, Determination of HER-2 statusand chromosome 17 polysomy in breast carcinomas comparing HercepTestand PathVysion FISH assay, Am. J. Clin. Pathol. 121 (2004) 70–77.

[25] J.M. Vera-Rom �an, L.A. Rubio-Martinez, Comparative assays for the HER-2/neuoncogene status in breast cancer, Arch. Pathol. Lab. Med. 128 (2004) 627–633.

[26] H. Yaziji, L. Goldstein, T.S. Barry, R. Werling, H. Hwang, G.K. Ellis, J.R. Gralow,R.B. Livingstone, A.M. Gown, HER-2 testing in breast cancer using paralleltissue-based methods, JAMA 291 (2004) 1972–1977.

[27] H. Yaziji, A.M. Gown, Accuracy and precision in HER2/neu testing in breastcancer: are we there yet?, Hum. Pathol. 35 (2004) 143–146.

[28] S. Wang, M.H. Saboorian, E. Frenkel, L. Hynan, S.T. Gokaslan, R. Ashfaq,Laboratory assessment of the status of Her-2/neu protein and oncogene inbreast cancer specimens: comparisons of immunohistochemistry assay withfluorescence in situ hybridisation assays, J. Clin. Pathol. 53 (2000) 374–381.

[29] A.C. Wolff, M.E. Hammond, J.N. Schwartz, K.L. Hagerty, D.C. Allred, D.F. Hayes,American Society of Clinical Oncology/College of American PathologistsGuideline Recommendations for Human Epidermal Growth Factor Receptor 2Testing in Breast Cancer, J. Clin. Oncol. 25 (2007) 118–145.