The Human Epidermal Growth Factor Receptor 2 (HER2) in the Breast Cancer: from Measurement to Targeted Treatment

The human epidermal growth factor receptor 2 (HER2) in the breast cancer: from measurement to targeted treatment

Thèse

Daniela Furrer Soliz Urrutia

Doctorat en médecine expérimentale Philosophiæ doctor (Ph. D.)

Québec, Canada

The human epidermal growth factor receptor 2 (HER2) in breast cancer: from measurement to targeted treatment

Thèse

Daniela Furrer Soliz Urrutia

Sous la direction de :

Caroline Diorio, directrice de recherche

Résumé

La surexpression du récepteur 2 du facteur de croissance épidermique humain (HER2) et/ou l’amplification du gène HER2 sont des facteurs prédictifs du cancer du sein. Avec l’introduction du traitement ciblé au trastuzumab, l’évaluation fiable d’HER2 est devenue essentielle. Malheureusement, jusqu’à 50% des patientes HER2-positives développent une résistance envers ce médicament.

Les objectifs étaient : 1) déterminer la façon la plus fiable et économique pour évaluer le statut HER2 (cohorte de 521 cas consécutifs de cancer du sein); 2) examiner l’association entre deux polymorphismes d’HER2 (Ile655Val et Ala1170Pro), la consommation de tabac et d’alcool et la réponse au trastuzumab (cohorte de 236 patientes HER2-positives traitées au trastuzumab). De plus, dans une étude pilote, nous avons examiné l’association entre les patrons de méthylation d’ADN dans la tumeur et la réponse au trastuzumab (cohorte de 12 patientes HER2-positives traitées au trastuzumab).

Le statut HER2 a été évalué par immunohistochimie (IHC), hybridation fluorescente in situ (FISH) et essai TaqMan. Nous avons comparé le statut HER2 déterminé par FISH sur lame complète (LC, un tissu par lame) et par matrice tissulaire (TMA, 60 tissus par lame), ainsi que le statut HER2 évalué par IHC et FISH sur le bloc ayant servi pour le diagnostic (bloc diagnostique) et sur un bloc choisi aléatoirement (bloc aléatoire). Les informations cliniques ont été obtenues dans les dossiers médicaux, celles sur la consommation de tabac et d’alcool par des questionnaires validés. Le patron de méthylation d’ADN a été évalué en utilisant la micropuce Illumina Infinium HumanMethylation450 BeadChip.

La concordance générale entre le statut HER2 déterminé par FISH sur LC et TMA était de 98,2%, et celle entre les blocs diagnostiques et aléatoires était de 98,0% au FISH et de 93,6% à l’IHC. La consommation de tabac et l’allèle Val étaient associés à une moins bonne réponse, tandis que la consommation d’alcool était associée à une meilleure réponse. Le patron de méthylation dans les tumeurs de patientes atteintes d’un cancer du sein HER2- positif qui ont développé une résistance au trastuzumab diffère de celui des patientes qui répondent au traitement. Cependant, ces résultats semblent dépendre de la méthode bioinformatique d’analyse utilisée.

Nous concluons que l’évaluation d’HER2 par FISH sur TMA représente une méthode fiable et économique. Les taux de concordances obtenus par FISH, mais pas ceux observés à

iii

l’IHC, satisfont l’exigence du Collège des pathologistes américains d’au moins 95% de concordance entre les résultats obtenus avec la méthode de référence et la nouvelle méthode. Le tabagisme, la consommation d’alcool et le polymorphisme HER2 Ile655Val pourraient influencer la réponse au traitement au trastuzumab.

Abstract

The overexpression of the human epidermal growth factor receptor 2 (HER2) and/or HER2 gene amplification are predictive factors in breast cancer. Following the HER2-targeted treatment with trastuzumab, the reliable evaluation of HER2 has become essential. Unfortunately, up to 50% of HER2-positive breast cancer patients develop resistance towards this drug.

The objectives were: 1). To determine the most reliable and economical method to evaluate HER2 status (cohort of 521 consecutive breast cancer cases); 2). To examine the association between tobacco and alcohol consumption, and two HER2 polymorphisms (Ile655Val and Ala1170Pro), and the response to trastuzumab (cohort of 236 HER2-positive breast cancer patients treated with trastuzumab). Moreover, in a pilot study, we explored the association between genome-wide DNA methylation patterns in breast cancer tissues and the response to trastuzumab (cohort of 12 breast cancer patients treated with trastuzumab).

HER2 status was evaluated by immunohistochemistry (IHC), fluorescence in situ hybridization (FISH), and TaqMan assay. We compared HER2 status determined by FISH on whole tissue (WT, one tissue per slide) section and tissue microarray (TMA, 60 tissues per slide) section, and HER2 status evaluated by IHC and FISH on the block used for diagnostic (diagnostic block) and on a randomly chosen additional block (random block). Clinicopathological information were assessed by review of medical records, tobacco and alcohol consumption by an administered validated questionnaire. DNA methylation patterns were evaluated using the Illumina Infinium HumanMethylation450 BeadChip.

Overall concordance between HER2 status determined by FISH on WT and TMA sections was 98.2% and that between diagnostic and random blocks was 98.0% for FISH and 93.6% for IHC. Tobacco consumption and the Val allele were associated with a worse response, whereas alcohol consumption was associated with a better response. Methylation pattern in tumor tissues of HER2-positive breast cancer patients who acquired resistance to trastuzumab treatment differed from that of HER2-positive breast cancer patients who responded to trastuzumab treatment. However, this observation seemed to depend upon the method of bioinformatics analysis used.

We conclude that FISH performed on TMA section represents a reliable and economical method for the evaluation of HER2. Results obtained by FISH, but not those obtained by

IHC, fulfill the recommendations of the College of American Pathologists of concordance greater than 95% between the reference method and the new method. Tobacco use, alcohol consumption and Ile655Val HER2 polymorphism might influence the response to trastuzumab treatment.

Table of contents

Résumé ...... iii Abstract ...... v Table of contents ...... vii Tables index ...... xiii Figures index ...... xv List of abbreviations ...... xvi Acknowledgments ...... xxii Foreword ...... xxiv Chapter 1: Introduction ...... - 1 - Literature overview ...... - 2 - 1.1. Breast cancer intrinsic subtypes ...... - 2 - 1.1.1. Molecular classification of breast cancer ...... - 2 - 1.1.2. Molecular classification of breast cancer subtypes using immunohistochemical surrogates ...... - 5 - 1.1.3. HER2-enriched subtype vs. breast cancer clinically evaluated as HER2-positive ...... - 7 - 1.1.4. Breast cancer clinically evaluated as HER2-positive ...... - 8 - 1.2. Methods for the evaluation of HER2 status in breast cancer specimens...... - 9 - 1.2.1. Immunohistochemistry (IHC) and in situ hybridization (ISH) methods ...... - 9 - 1.2.2. Methodological worries related to HER2 testing ...... - 11 - 1.2.3. Concordance of HER2 status determined by IHC, FISH, CISH and SISH in the literature ...... - 13 - 1.3. HER2 biology...... - 14 - 1.3.1. HER2 structure and function ...... - 14 - 1.3.2. Consequences of constitutive HER2 receptor activation...... - 18 - 1.4. Treatment of HER2-positive breast cancer patients ...... - 19 - 1.4.1. Anti-HER2 agents ...... - 19 - 1.4.2. Endocrine therapy ...... - 24 - 1.4.3. Chemotherapeutic drugs ...... - 26 - 1.5. Resistance to trastuzumab ...... - 27 - 1.5.1. Molecular mechanisms of trastuzumab resistance ...... - 28 - 1.5.2. Genetic and epigenetic factors and survival of HER2-positive breast cancer patients ...... - 34 - 1.5.2.1. Single nucleotide polymorphisms (SNPs) ...... - 34 - 1.5.2.2. HER2 SNPs ...... - 34 - 1.5.2.3. HER2 SNPs and response to trastuzumab ...... - 35 -

vii

1.5.3. Epigenetics ...... - 35 - 1.5.3.1. DNA methylation ...... - 37 - 1.5.3.2. Distribution of CpG dinucleotides and CpG methylation in the human genome ...... - 39 - 1.5.3.3. Mechanisms of control of gene expression through DNA methylation ...... - 40 - 1.5.3.4. Aberrant DNA methylation in cancer ...... - 42 - 1.5.3.5. Link between DNA methylation and histone modifications ...... - 43 - 1.5.3.6. Link between DNA methylation and miRNAs ...... - 44 - 1.5.3.7. Methods for DNA methylation analysis ...... - 46 - 1.5.3.8. Aberrant methylation in breast cancer ...... - 50 - 1.5.3.9. DNA methylation signature in HER2-positive breast cancer ...... - 51 - 1.5.3.10. Genome-wide DNA methylation analysis and association with survival in breast cancer patients ...... - 52 - 1.5.3.11. Genome-wide DNA methylation analysis in breast cancer and response to chemotherapy- 53 - 1.5.3.12. DNA methylation in breast cancer tissues and trastuzumab response in HER2-positive breast cancer patients ...... - 53 - 1.5.3.13. Epigenetic therapies for treatment of breast cancer patients ...... - 54 - 1.5.3. Lifestyle factors and survival of HER2-positive breast cancer patients ...... - 55 - 1.5.3.1. Tobacco and alcohol consumption ...... - 55 - 1.5.3.2. Molecular and epidemiological ink between tobacco and alcohol exposure and HER2 ..... - 56 - Chapter 2: Advantages and disadvantages of technologies for HER2 testing in breast cancer specimens . - 58 - Résumé ...... - 59 - Abstract ...... - 60 - Introduction ...... - 61 - Principles of the methods of analysis, advantages and disadvantages of each described technique ...... - 62 - Southern blot ...... - 62 - Northern blot ...... - 63 - Enzyme-linked immunosorbent assay (ELISA) ...... - 63 - Western blot ...... - 65 - Polymerase chain reaction (PCR)-based assays ...... - 65 - Multiplex ligation-dependent probe amplification (MLPA) ...... - 67 - Immunohistochemistry (IHC) ...... - 68 - Fluorescence in situ hybridization (FISH) ...... - 70 - Bright-field in situ hybridization (BRISH) methods ...... - 73 - Chromogenic in situ hybridization (CISH) ...... - 73 -

viii

Silver-enhanced in situ hybridization (SISH) ...... - 76 - Gold-facilitated autometallographic in situ hybridization (GOLDFISH) ...... - 77 - Bright-field double in situ hybridization (BDISH) ...... - 77 - HER2 gene-protein assay ...... - 79 - mRNA in situ hybridization ...... - 80 - Instant-quality fluorescence in situ hybridization (IQFISH) ...... - 81 - Automated HER2 FISH assay ...... - 82 - Performance of assays used for the HER2 status determination in breast cancer in predicting response to anti-HER2 therapies ...... - 82 - Conclusion ...... - 83 - Acknowledgements ...... - 85 - Conflict of interest ...... - 85 - References ...... - 86 - Chapter 3: Contextualisation, hypothesis and objectives ...... - 106 - 3.1. Contextualisation ...... - 106 - 3.2. Hypothesis ...... - 106 - 3.3. Objectives ...... - 108 - Chapter 4: Tissue microarray is a reliable tool for the evaluation of HER2 amplification in breast cancer . - 110 - Résumé ...... - 111 - Abstract ...... - 112 - Introduction ...... - 113 - Materials and Methods ...... - 114 - Specimen collection and patient population ...... - 114 - Tissue microarray construction and processing ...... - 114 - Fluorescence in situ hybridization ...... - 115 - HER2 evaluation on TMA ...... - 116 - Statistical analysis ...... - 116 - Results ...... - 116 - Discussion ...... - 117 - Acknowledgements ...... - 120 - References ...... - 121 - Chapter 5: Concordance between immunohistochemistry and fluorescence in situ hybridization in the determination of human epidermal growth factor receptor 2 (HER2) status using tissue microarray in breast cancer specimens ...... - 129 -

Résumé ...... - 130 - Abstract ...... - 131 - 1. Introduction ...... - 132 - 2. Materials and methods ...... - 133 - 2.1. Specimen collection and patient population ...... - 133 - 2.2. Tissue microarray (TMA) construction ...... - 133 - 2.3. Immunohistochemistry ...... - 134 - 2.4. Fluorescence in situ hybridization ...... - 134 - 2.5. HER2 evaluation on TMA ...... - 134 - 2.6. Statistical analysis ...... - 135 - 3. Results ...... - 135 - 4. Discussion ...... - 137 - 5. Conclusions ...... - 140 - 6. Conflict of interest and sources of funding...... - 140 - 7. Acknowledgments ...... - 141 - References ...... - 142 - Chapter 6: Validation of a new classifier for the automated analysis of the human epidermal growth factor receptor 2 (HER2) gene amplification in breast cancer specimens ...... - 152 - Résumé ...... - 153 - Abstract ...... - 154 - Introduction ...... - 155 - Material and Methods ...... - 156 - Case selection ...... - 156 - Fluorescence In Situ Hybridization ...... - 156 - Manual scoring (reference method) ...... - 156 - Tile-sampling classifier ...... - 157 - Nuclei-sampling classifier ...... - 158 - Results ...... - 159 - Determination of the accuracy of the nuclei-sampling classifier on special specimens ...... - 160 - Determination of the accuracy of the nuclei-sampling classifier on equivocal specimens...... - 161 - Reproducibility of results ...... - 161 - Discussion ...... - 161 - Conclusions ...... - 167 -

List of abbreviations...... - 167 - Competing interests...... - 167 - Authors’ contributions ...... - 167 - Acknowledgements ...... - 167 - References ...... - 168 - Chapter 7: Evaluation of human epidermal growth factor receptor 2 (HER2) single nucleotide polymorphisms (SNPs) in normal and breast tumor tissues and their link with breast cancer prognostic factors ...... - 174 - Résumé ...... - 175 - Abstract ...... - 176 - Introduction ...... - 177 - Materials and Methods ...... - 178 - Study population and data collection...... - 178 - Ile655Val and Ala1170Pro polymorphisms ...... - 178 - Statistical analysis ...... - 179 - Results ...... - 179 - Discussion ...... - 180 - Conclusions ...... - 182 - Acknowledgments ...... - 183 - Conflict of interest statement ...... - 183 - Funding ...... - 183 - References ...... - 183 - Chapter 8: Association between tobacco and alcohol consumption, HER2 polymorphisms and response to trastuzumab in HER2-positive breast cancer patients ...... - 193 - Résumé ...... - 194 - Abstract ...... - 195 - Introduction ...... - 196 - Material and Methods ...... - 197 - Study population and data collection...... - 197 - Polymorphism substudy ...... - 198 - Study end points ...... - 198 - Statistical analysis ...... - 198 - Results ...... - 199 - HER2 polymorphism substudy ...... - 200 - Discussion ...... - 201 -

Conclusion ...... - 204 - Acknowledgments ...... - 205 - References ...... - 206 - Chapter 9: Association between genome-wide DNA methylation pattern and response to trastuzumab in HER2-positive breast cancer patients ...... - 216 - Résumé ...... - 217 - Abstract ...... - 218 - Introduction ...... - 219 - Material and Methods ...... - 220 - Study population ...... - 220 - DNA extraction ...... - 220 - Bioinformatic analysis ...... - 221 - Statistical analysis ...... - 223 - Results ...... - 223 - Discussion ...... - 225 - Acknowledgments ...... - 226 - References ...... - 227 - Chapter 10 : Discussion and conclusion ...... - 232 - 10.1. General discussion ...... - 232 - 10.2. Perspectives ...... - 243 - 10.3. Conclusion ...... - 245 - References ...... - 247 - Annexe I ...... - 299 -

xii

Tables index

Table 1.1. Surrogate definition of intrinsic subtypes of breast cancer as defined by the 2011 St Gallen International Expert Consensus Panel ...... - 6 - Table 2.1. Main characteristics of the described techniques ...... - 105 - Table 4.1. Interpretation criteria for fluorescence in situ hybridization according to the 2007 and the 2013 American Society of Clinical Oncology/College of American Pathologists scoring systems ...... - 124 - Table 4.2. Concordance of human epidermal growth factor receptor 2 (HER2) gene amplification status between fluorescence in situ hybridization performed on whole tissue sections (reference method) and on diagnostic tissue microarray (TMA) sections according to the 2013 American Society of Clinical Oncology/College of American Pathologists scoring system ...... - 125 - Table 4.3. Concordance of human epidermal growth factor receptor 2 (HER2) gene amplification status between fluorescence in situ hybridization performed on whole tissue sections (reference method) and on diagnostic tissue microarray (TMA) sections according to the 2013 American Society of Clinical Oncology/College of American Pathologists scoring system ...... - 126 - Table 4.4. Concordance of human epidermal growth factor receptor 2 (HER2) gene amplification status between fluorescence in situ hybridization performed on whole tissue sections (reference method) and on diagnostic tissue microarray (TMA) sections according to the number of informative cores using the 2007 American Society of Clinical Oncology/College of American Pathologists scoring system ...... - 127 - Table 4.5. Concordance of human epidermal growth factor receptor 2 (HER2) status determined by fluorescence in situ hybridization on whole tissue section (reference method) and on random TMA section according to the 2013 ASCO/CAP scoring system ...... - 128 - Table 5.1. Concordance of HER2 status determined by IHC and FISH on diagnostic TMA section according to the 2013 ASCO/CAP scoring system ...... - 146 - Table 5.2. Concordance of HER2 status determined by IHC and FISH on random TMA section according to the 2013 ASCO/CAP scoring system ...... - 147 - Table 5.3. Concordance of HER2 status determined by FISH on diagnostic TMA section and random TMA section according to the 2013 ASCO/CAP scoring system ...... - 148 - Table 5.4. Concordance of HER2 status determined by IHC on diagnostic TMA section and random TMA section according to the 2013 ASCO/CAP scoring system ...... - 149 - Table 5.A.1. Comparison of HER2 status determined by IHC and FISH on diagnostic TMA section according to the 2007 ASCO/CAP scoring system ...... - 150 - Table 5.A.2. Concordance of HER2 status according to the combined IHC and FISH between diagnostic TMA section and random TMA section according to the 2013 ASCO/CAP scoring system ...... - 151 - Table 6.1. Comparison of results obtained by different methods for non-amplified and amplified cases .... - 171 - Table 6.2. Comparison of results obtained by different methods for amplified cases without HSR ...... - 171 - Table 6.3. Comparison of results obtained by different methods for equivocal cases ...... - 171 - Table 7.1. Characteristics of the study population ...... - 189 - Table 7.2. Association of HER2 Ile655Val and Ala1170Pro polymorphisms in normal and tumor breast tissues with prognostic factors ...... - 190 - Table 7.3. Distribution of HER2 Ile655Val and Ala1170Pro polymorphisms in normal breast and tumor breast tissues ...... - 192 - Table 8.1. Baseline characteristics of the study population ...... - 211 -

xiii

Table 8.2. Unadjusted and adjusted hazard ratios for disease-free survival according to the tobacco consumption before breast cancer diagnosis and during trastuzumab treatment ...... - 212 - Table 8.3. Unadjusted and adjusted hazard ratios for disease-free survival according to the tobacco consumption before breast cancer diagnosis and during trastuzumab treatment, stratified by ER status .. - 213 - Table 8.4. Unadjusted and adjusted hazard ratios for disease-free survival according to the alcohol consumption before breast cancer diagnosis and during trastuzumab treatment ...... - 214 - Table 8.5. Unadjusted and adjusted hazard ratios for disease-free survival according to the HER2 polymorphisms ...... - 215 - Table 9.1. Baseline characteristics of cases and controls ...... - 231 - Table 10.1. Impact of the 2013 ASCO/CAP scoring criteria on the classification of cases used for the validation of the software programming algorithm (results obtained at manual counting) ...... - 236 - Table 10.2. Comparison of results obtained by different methods for non-amplified and amplified cases (HER2 gene amplification status according to the 2013 ASCO/CAP scoring criteria) ...... - 237 - Table S1. Concordance of HER2 status : IHC vs. SISH, FISH vs. SISH, CISH vs. SISH ...... - 299 -

xiv

Figures index

Figure 1.1. Immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) results of HER2. ... - 10 - Figure 1.2. Construction of Tissue microarray ...... - 13 - Figure 1.3. Signal transduction by the HER family members...... - 17 - Figure 1.4. Proposed mechanisms of action of trastuzumab ...... - 21 - Figure 1.5. Agents that target HER2 receptor...... - 22 - Figure 1.6. Proposed mechanisms of trastuzumab resistance ...... - 33 - Figure 1.7. Waddington’s classical epigenetic landscape...... - 36 - Figure 1.8. Conversion of cytosine to 5-methylcytosine by DNA methyltransferase (DNMT)...... - 38 - Figure 1.9. Proposed mechanisms of silencing of CpG islands (CGI) promoters by DNA methylation...... - 41 - Figure 1.10. Overview of the Infinium I and Infinium II assays ...... - 49 - Figure 6.1. Image analysis of fluorescence signals ...... - 172 - Figure 6.2. HER2 fluorescence in situ hybridization in amplified cases ...... - 173 -

List of abbreviations

ADCC antibody-dependent cellular cytotoxicity

AI aromatase inhibitor

Ala Alanine

ASCO/CAP American Society of Clinical Oncology/College of American Pathologists

ATP adenosine triphosphate

BDISH bright-filed double in situ hybridization

BMI body mass index

BER base excision repair

Cas9 CRISPR-associated protein 9

Cdk cyclin-dependent kinase

CEP17 chromosome enumeration probe 17

CGI CpG island

CI confidence interval

CIMP CpG island methylator phenotype

CISH chromogenic in situ hybridization

CRISPR clustered regularly interspaced short palindromic repeat ddNTP dideoxynucleotide triphosphates

DFS disease-free survival

DNMT DNA methyltransferase

ECD extracellular domain

xvi

EGF epidermal growth factor

EGFR epidermal growth factor receptor

EPG epigen

ER estrogen receptors

ERBB2 erb-b2 receptor tyrosine kinase 2

ERK extracellular signal-regulated kinase

FcgR Fc gamma receptor

FDA Food and Drug Administration

FDR false dicovery rate

FFPE formalin-fixed, paraffin-embedded

FGFR fibroblast growth factor receptor

FISH fluorescence in situ hybridization

GEP gene expression profiling

HAT histone acetyltransferase

HDAC histone deacetylase

H&E hematoxylin & eosin

HER1 human epidermal growth factor receptor 1

HER2 human epidermal growth factor receptor 2

HER3 human epidermal growth factor receptor 3

HER4 human epidermal growth factor receptor 4

HMT histone methyltransferase

xvii

HPLC high performance liquid chromatography

HR hazard ratio

IAP intracisternal A-particles

IGF-1R insulin-like growth factor I receptor

IgG immunoglobulin G

Ile Isoleucine

IHC immunohistochemistry

IR insulin receptor

LHRH luteinizing hormone-releasing hormone

LINE long interspersed nuclear elements

LVEF left ventricular ejection fraction

M methylated

MAPK mitogen-activated protein kinase

MBD methyl-CpG binding domain

MBP methyl-CpG-binding proteins

MCM2 mini-chromosome maintenance complex component 2

MGMT O6-methylguanine-DNA methyltransferase

MSP methylation-specific PCR

MUCA Mucin 4

NGS next-generation sequencing

NK Natural killer

xviii

NNK 4-(methylnitrosamino)-1-3-(3-pyridyl)-1-butanon

OS overall survival

PAM50 Prediction Analysis of Microarray 50

PCR polymerase chain reaction

PCR-RFLP polymerase chain reaction- restriction fragment length polymorphism

PFS progression-free survival

PI3K phosphatidylinositol 3’kinase

PR progesterone receptor

Pro Proline

PTEN phosphatase and tensin homolog deleted on chromosome 10 qRT-PCR quantitative reverse transcription polymerase chain reaction

RR response rate

RTK receptor tyrosine kinase

SAM S-adenosyl-methionine

SERD selective estrogen receptor down-regulator

SERM selective estrogen receptor modulator

SINE short interspersed nuclear elements

SISH silver-enhanced in situ hybridization

SNP single nucleotide polymorphism

TET ten-eleven translocation

TGFα transforming growth factor α

xix

TKI tyrosine kinase inhibitor

TMA tissue microarray

TME tumor microenvironment

TSG tumor suppressor gene

TSS transcription start site

U unmethylated

Val Valine

VEGF vascular endothelial growth factor

VEGFR vascular endothelial growth factor receptor

WGA whole genome amplification

WHO World Health Organization

WT whole tissue

To Jorge To Gabriel and Léa

“Life is not measured by the number of breaths we take, but by the moments that take our breath away.” Maya Angelou

xxi

Acknowledgments

Foremost, I would like to express my gratitude to my thesis supervisor, Dr. Caroline Diorio, for her continuous guidance and support throughout my Ph.D. studies. I am grateful for her availability, her wise advice, her sensibility and friendship. I would also thank her for encouraging me to pursue my ideas and suggested directions for research.

I would like to thank the pathologists of the Service de Pathologie of the St-Sacrement Hospital, including Drs. Chantal Caron, Anne Choquette, Michel Beauchemin, Mohamed Amin Hashem, Simon Jacob, Sophie Laberge, Mohib Morcos, Nathalie Mourad, Alexandre Odashiro, and Ion Popa. In particular, I would like to thank Dr. Simon Jacob for his support, his kindness, for enthusiastically sharing his broad knowledge of breast pathology and for always making time for me despite his very busy schedule.

I am grateful to the personnel of the Service de Pathologie of the St-Sacrement Hospital for their technical support. Special thanks to François Sanschagrin and Claudie Paquet for introducing me to the fascinating world of molecular diagnostics and for checking the FISH results.

I am grateful to Annick Michaud and Geneviève Ouellette for their great technical support during laboratory work. I would like to thank Sue-Ling Chang for her support with statistical analysis and for the English revision of the manuscripts. Many thanks to Isabelle Dumas for her help in the comprehension of epidemiological concepts and for her wise advice.

I am also grateful to the clinicians and to the personnel of the Clinique des Maladies du Sein Deschênes-Fabia, in particular to Drs. Louise Provencher and Julie Lemieux. I would also like to thank Christian Laflamme for his support during manuscript submission. I am also grateful to our great research nurses Danièle Audet and Lucie Tellier for their availability and kindness.

Many thanks also to Drs. Frédéric Barabé and Marc-André Côté for the critical review of one manuscript.

I would like to thank Caty Blanchette, Myrto Mondor and Eric Demers for their support in statistical analysis.

xxii

Many thanks to Frédéric Fournier and Arnaud Droit for performing the bioinformatics analysis using the R statistical programming environment. I am also grateful to Yvan Labrie and Marie-Christine Pouliot for introducing me to the GenomeStudio software and software for the analysis of pathways and networks.

I am grateful to the Fondation du cancer du sein du Québec and the Banque de tissus et de données of the Réseau de recherche sur le cancer of the FRQS, which is affiliated with the Canadian Tumor Repository Network, for providing clinical specimens.

I am grateful to all my friends at the Research Unit at St-Sacrement Hospital, including Carlotta Lunghi, Kaoutar Ennour-Idrissi, Danielle Larouche, Ludivine Soguel Alexander, Thyphavone Oudanonh, Sofia Laforest, Geneviève Larouche, and Elissar Issa for the many pleasant hours spent together and for creating a productive and enjoyable work environment.

I am grateful to Denis Guillette for his prompt informatics support. I would also like to thank the unit secretary, Ginette Desbiens, for her efficient work and her contagious good mood.

I am grateful to all women who participated in our studies.

I am indebted to the Fonds de recherche du Québec –Santé (FRQS) (doctoral fellowship), the Laval University Cancer Research Center (Bourse de distinction Luc Bélanger), the Fondation des Hôpitaux Enfant-Jésus – St-Sacrement and Hoffmann-La Roche Limited for financial support.

Special thanks to our friends, including Rossana, Vincent, Caroline, Gaëlle, Samuel, Aida, Richard, Chantal and Rumiana, who represent our extended family here in Quebec.

I am grateful to my extended family and my family-in-law for their constant love and support, and for their patience and comprehension. Special thanks to my mother Adriana for being a great example of strength and determination.

Special thanks go to my husband Jorge for being such an amazing companion and for sharing with me the joys and difficulties of this great Ph.D. adventure. A very special thanks to little Gabriel and Léa who allow us to re-discover and appreciate the very essence of this miraculous thing called life.

xxiii

Foreword

The research work presented in this Ph.D. thesis was carried out at the St-Sacrement Hospital in Quebec City. The aim of my research work was multifaceted. In the first part of my research, we aimed to identify the most reliable and economical method to assess HER2 status in breast cancer specimens. The results of this research question are presented in Chapters 4 and 5. With the exception of the writing of the grant research proposal, I participated to all aspects for the accomplishment of this project. I was responsible for the literature review, the laboratory work, the analysis of the results and the writing of the manuscripts. Drs. Simon Jacob and Chantal Caron, François Sanschagrin and Caroline Diorio participated in the conception of the study, and in the analysis and interpretation of results. Caroline Diorio supervised each of these stages. All co-authors provided constructive feedback and made specific suggestions to improve the final version of the manuscripts. Both articles have been published in the journal Anticancer Research. This part of the research was funded by the Fondation des Hôpitaux Enfant-Jésus – St- Sacrement and Hoffmann La Roche Limited. In addition, I also reviewed all existing methods for the evaluation of HER2 status in breast cancer specimens and I wrote a review article that has been published in the American Journal of Clinical Pathology (Chapter 2). All co- authors provided constructive feedback and made specific suggestions to improve the final version of the manuscript. In an attempt to further identify the best method to reliably determine HER2 status in breast cancer specimens, we validated a software programming algorithm that analyses fluorescent signals in single tumor cell nuclei within breast cancer tissue sections (Chapter 6). For this project, I participated in the conception of the study and I was responsible for the literature review, the laboratory work, the analysis of the results and the writing of the manuscript that has been published in the journal Diagnostic Pathology. Drs. Simon Jacob and Chantal Caron, François Sanschagrin and Caroline Diorio participated in the conception of the study, and in the analysis and interpretation of results. Caroline Diorio supervised each of these stages. All co-authors provided constructive feedback and made specific suggestions to improve the final version of the manuscript. This project was partly funded by the MetaSystems company.

In the second part of this thesis we aimed to investigate the factors that could affect the response to trastuzumab in HER2-positive breast cancer patients. We concentrated on several aspects, including HER2 polymorphisms, alcohol and tobacco consumption and genome-wide DNA methylation pattern in breast cancer tissues.

xxiv

As a first step, we analyzed the association between HER2 polymorphisms and breast cancer prognostic factors. For this part of the project, I wrote the manuscript that has been published in the review The Breast (Chapter 7). Annick Michaud performed the laboratory work. Christian Laflamme collected clinico-pathological information. Éric Demers performed statistical analysis. Caroline Diorio and Julie Lemieux supervised each of these stages. All co-authors provided constructive feedback and made specific suggestions to improve the final version of the manuscript. This project was funded by the Fondation des Hôpitaux Enfant-Jésus – St-Sacrement and Hoffmann La Roche Limited.

In a second step, I proposed to analyze the association between HER2 polymorphisms, alcohol and tobacco consumption and genome-wide DNA methylation pattern in breast cancer tissues and the response to trastuzumab in HER2-positive breast cancer patients (Chapters 8 and 9). For this part of the thesis, I was responsible for the literature review, the collection of clinicopathological information (histopathological data, information about treatment received, follow-up data), the laboratory work (only Chapter 9), the analysis of methylation data using GenomeStudio software and the analysis of pathways, conducting statistical analysis, and the writing of the manuscripts. Annick Michaud assessed HER2 genotyping in breast cancer tissues and normal breast tissues (Chapter 8). Frédéric Fournier and Arnaud Droit analysed methylation data using the statistical environment R and performed paired statistical analysis (Chapter 9). Caroline Diorio supervised each of these stages. All co-authors provided constructive feedback and made specific suggestions to improve the final version of the manuscript. The article on the association between HER2 polymorphisms, alcohol and tobacco intake and response to trastuzumab has been submitted to the journal Clinical Breast Cancer (24th May, 2017, under revision), while the manuscript on the association between DNA methylation pattern in breast cancer tissues and response to trastuzumab is finalized but is not yet submitted for publication. These projects were funded by the Fondation des Hôpitaux Enfant-Jésus – St-Sacrement and Hoffmann La Roche Limited.

Although it does not appear in this document, I also participated in another project, whose goal was to evaluate the association between the phosphorylated form of HER2 and the response to trastuzumab in HER2-negative breast cancer patients. For this project, I performed the literature review and I generated preliminary results. Moreover, I also contributed to the writing of research grant proposals based on these preliminary results.

xxv

The articles listed below represent my contribution to the field of molecular epidemiology. All these manuscripts were written during my doctoral studies. Articles 1, 2, 4, 5, 7, 8 and 9 are included in this Ph.D. thesis.

1. Furrer D., Jacob S., Caron S., Sanschagrin F., Provencher L., Diorio C. 2013. Validation of a new classifier for the automated analysis of the human epidermal growth factor receptor 2 (HER2) gene amplification in breast cancer specimens. Diagnostic pathology 8: 17.

2. Furrer D., Jacob S., Sanschagrin F., Diorio C. 2015. Advantages and disadvantages of technologies for HER2 testing in breast cancer specimens. American Journal of Clinical Pathology 144(5): 686-703.

3. Diorio C., Furrer D., Michaud A., Laberge S., Popa I., Jacob S., Provencher L., Hogue JC. 2016. Validation of EP1 antibody clone for estrogen receptor immunohistochemistry for breast cancer. Anticancer Research 36(1): 435-7.

4. Furrer D., Jacob S., Caron C., Sanschagrin F., Provencher L., Diorio C. 2016. Tissue microarray is a reliable tool for the evaluation of HER2 status in breast cancer specimens. Anticancer Research 36(9): 4661-6.

5. Furrer D., Côté M., Provencher L., Laflamme C., Barabé F., Jacob S., Michaud A., Lemieux J., Diorio C. 2016. Evaluation of human epidermal growth factor receptor 2 (HER2) single nucleotide polymorphisms (SNPs) in normal and breast tumor tissues and their link with breast cancer prognostic factors. The Breast 30: 191-196.

6. Yousef E.M., Furrer D., Laperrière D., Tahir M.R., Mader S., Diorio C., Gaboury L.A. 2017. MCM2: an alternative to Ki-67 for measuring breast cancer cell proliferation. Modern pathology 30(5): 682-697.

7. Furrer D., Jacob S., Caron C., Sanschagrin F., Provencher L., Diorio C. 2017. Concordance between immunohistochemistry and fluorescence in situ hybridization in the determination of human epidermal growth factor receptor 2 (HER2) status using tissue microarray in breast cancer specimens. Anticancer Research 37(6): 3323-3329.

xxvi

8. Furrer D., Jacob S., Michaud A., Provencher L., Lemieux J., Diorio C. 2017. Association between tobacco and alcohol consumption and HER2 polymorphisms and response to trastuzumab in HER2-positive breast cancer patients. Clinical Breast Cancer, submitted.

9. Furrer D., .Fournier F., Jacob S., Droit A., Diorio C. Association between genome-wide DNA methylation pattern and response to trastuzumab in HER2-positive breast cancer patients. In preparation.

10. Odashiro P.P., Orain M., Furrer D., Diorio C., Simonyan D., Moreira A., Joubert P. PD- L1 in thymomas and thymic carcinomas. In preparation.

xxvii

Chapter 1: Introduction

Over the past decades, microarray-based gene expression studies have highlighted the fact that breast cancer comprises a heterogeneous group of diseases in term of differentiation and proliferation, prognosis and treatment. These studies allowed the identification of breast cancer intrinsic subtypes (1-3). One of these subtypes is the so-called human epidermal growth factor receptor 2 (HER2)-enriched subtype. Clinically, HER2-positive breast cancer is characterized by the overexpression of the HER2 receptor and/or HER2 gene amplification (4). HER2-positive breast cancer patients have a particular worse prognosis. In addition, HER2-positive breast cancer patients are eligible to receive targeted treatment with trastuzumab, a monoclonal antibody specifically directed against the HER2 receptor (5).

Evaluation of the HER2 status in breast cancer specimens is crucial for defining patient management. Considering the clinical and economic implications of targeted anti-HER2 treatments, reliable HER2 test results are essential. False negative results would deny the patients access to the potential benefits of trastuzumab, whereas false positive results would expose patients to the potential cardiotoxic side effects of this expensive agent without experiencing any therapeutic advantages (6).

Immunohistochemistry (IHC) and in situ hybridization (ISH) techniques are the most commonly used techniques for the determination of HER2 status in breast cancer specimens (7, 8). As each technique has its own advantages and disadvantages, currently there is still no consensus on which method is the best for evaluating the HER2 status in breast cancer specimens (6). The first aim of this project, therefore, was to identify the most reliable and economical method to evaluate HER2 status in breast cancer specimens.

Trastuzumab treatment significantly improved survival in HER2-positive breast cancer patients. Two Cochrane systematic reviews reported that, when combined with standard chemotherapy, trastuzumab significantly improved overall survival (OS) and disease-free survival (DFS) in HER2-positive women with early and locally advanced breast cancer (9) and OS and progression-free survival (PFS) in metastatic HER2-positive breast cancer patients (10). .

Despite this noteworthy achievement, a subset of HER2-positive breast cancer patients does not respond to trastuzumab-based treatment (11). Although several molecular

- 1 -

mechanisms of trastuzumab resistance have been proposed in the literature (12), no clinically effective strategies to overcome trastuzumab resistance have been identified yet (11).

It has been reported that genetic and epigenetic factors can influence the efficacy of antineoplastic drugs in breast cancer patients (13-15). It has been shown that HER2 polymorphisms have an impact on HER2 function (16, 17). It has also been reported that methylation pattern in breast cancer tissues of HER2-positive breast cancer patients is heterogeneous (18). Given that trastuzumab response in HER2-positive breast cancer patients is also heterogeneous, we hypothesize that methylation pattern in tumor tissues of patients that respond to trastuzumab treatment is different to that of patients that develop resistance to this agent.

Evidence suggests that also lifestyle factors including tobacco and alcohol use might influence the survival of breast cancer patients (19-21). Interestingly, it has been reported the existence of molecular and epidemiological links between tobacco and ethanol exposures and HER2 (22-26). We hypothesize that HER2 polymorphisms, tobacco and alcohol consumption may represent factors influencing the efficacy of targeted therapies in HER2-positive breast cancer patients.

The second aim of our study was therefore to analyze the association of HER2 polymorphisms and methylation patterns in breast cancer tissues, and tobacco and alcohol consumption with the response to trastuzumab in HER2-positive breast cancer patients.

Literature overview

1.1. Breast cancer intrinsic subtypes

1.1.1. Molecular classification of breast cancer Breast cancer is the most frequently diagnosed cancer and the leading cause of cancer- related death amongst women worldwide (27). In 2015, it is estimated that in Canada 25,000 new cases of breast cancer have been diagnosed and 5,000 women have died from this malignancy (28).

It is now well recognized that breast cancer is a heterogeneous disease characterized by various pathological features, different treatment response, and substantial differences in survival (29). Based on the World Health Organization (WHO) classification, there are at

- 2 -

least 20 different histological types of breast cancer (30). A major drawback of this classification is that around 90% of breast cancers will eventually belong to the most common morphological subtypes, namely invasive ductal carcinomas or invasive lobular carcinomas, which represent approximatively 80% and 10% of invasive breast cancers, respectively. As patients with tumors showing similar histological appearance may have very different clinical outcomes, the histopathological classification presents limited prognostic and predictive values, and its clinical utility is relatively modest (31).

Given that histological diversity cannot entirely explain the difference in breast cancer behavior, at the beginning of the new century novel approaches, particularly gene expression profiling, have been used to refine breast cancer classification independently of histomorphological criteria. Perou and collaborators performed cDNA microarray analysis of RNA extracted from frozen tissue to characterize the gene expression of 65 breast cancer specimens obtained from 42 patients (1). The central hypothesis of this study was that tumors could be characterized on the up- or down-regulation of sets of genes, the so-called “intrinsic genes”. Intrinsic genes were defined as genes with significantly greater variability in expression between different tumors than between duplicate samples from the same tumor. This resulted in the selection of 496 genes representing an “intrinsic gene list” chosen from an initial list encompassing 8,102 genes that allowed classifying tumor specimens into subtypes. This seminal work showed for the first time that breast tumors could be classified into molecular subtypes distinguished by differences in their gene expression profiles.

A subsequent study that similarly analyzed the expression profiles of 115 independent breast tumor samples using RNA extracted from frozen tissue and an intrinsic set of 534 genes allowed a refined molecular classification of breast tumors as reported by Perou et al. (3). Using hierarchical clustering analyses tumors showing similar gene expression patterns were grouped together. This analysis allowed identifying two distinctive groups of breast cancers based on the expression of the estrogen receptors (ER) and ER-related genes (GATA3, X-box binding protein 1, HNF3A) (3): whereas one cluster expressed ER and ER-associated genes, the other one did not. The first group of tumors was defined as ER-positive, the second one as ER-negative. This finding was consistent with breast cancer cell biology, since the binding of estrogen to nuclear ER and the subsequent binding of the dimerized ER to estrogen responsive elements (EREs) located in the promoter of target genes regulate cellular growth and proliferation in several target tissues of the human body, including breast tissue (32).

- 3 -

Tumors grouped in the ER-positive cluster showed a gene expression profile similar to that of luminal cells found in the normal breast (3). For this reason, they were named as luminal tumors. Luminal tumors were subdivided into two categories: luminal A tumors, characterized by highest ER expression, and luminal B tumors that expressed ER-related genes at a lower level. Moreover, compared to luminal A tumors, luminal B tumors showed higher expression of proliferation/cell-cycle related genes or proteins, including MKI67 and AURKA, and lower expression of several luminal-related genes or proteins (e.g. progesterone receptor (PR) and FOXA) (33). In the ER-negative cluster, three subtypes were observed: the HER2-enriched tumors, the basal-like tumors, and the normal breast- like tumors. The HER2-enriched tumors were characterized by high expression of the human epidermal growth factor receptor 2 (HER2) gene and other genes associated with the HER2 pathway and/or HER2 amplicon located in the 17q12 chromosome. The basal- like tumors were named as such as they express cytokeratins (CK5/6, CK14, and CK17) that are usually expressed in normal breast myoepithelial cells, cells that underlie the breast luminal cells. The normal breast-like tumor group has been described to express genes that are usually expressed in adipose tissue and other non-epithelial cell types. These tumors were also shown to express basal epithelial genes. However, it has been postulated that this was likely an artificial category caused by poorly sampled tumor tissue (2).

As illustrated above, the intrinsic breast cancer subtypes were initially identified using unsupervised clustering analysis of gene expression using RNA isolated from frozen tissues. In order to improve applicability of gene expression profiling to routine pathology specimens, Parker and collaborators developed a quantitative reverse transcription polymerase chain reaction (qRT-PCR) assay to identify the intrinsic molecular subtypes using RNA extracted from more easily available formalin-fixed, paraffin-embedded (FFPE) tissues (34). They proposed an efficient 50-gene classifier, referred to as the Prediction Analysis of Microarray 50 (PAM50), that reanalyzed the previous five intrinsic subtypes defining the four major subtypes currently known: luminal A, luminal B, HER2-enriched and basal-like. The PAM50 classifier is an often used method to classify intrinsic subtypes (35). The prognostic value of the PAM50 classifier was validated in a test set of 761 node-negative breast cancer patients who did not receive any adjuvant therapy. Using statistical models, the risk of relapse (ROR) score was estimated for each test case. From this analysis it emerged that intrinsic subtypes, as defined with the PAM50 classifier, showed prognostic significance. Moreover, the results remained significant in multivariate analyses that incorporated standard parameters,

- 4 -

including ER status, histological grade, tumor size, and node status. In addition, in a set of 113 patients treated with neoadjuvant chemotherapy, the intrinsic subtype model predicted neoadjuvant chemotherapy (taxane and anthracycline regimen) efficacy with a negative predictive value for pathologic complete response of 97% (34).

Molecular understanding of breast cancer paved the way to personalized medicine. Since the initial description of the intrinsic subtypes, several studies have validated their prognostic and predictive significance (5, 36-43). It has emerged that patients with basal-like or HER2- enriched breast cancer showed the worst outcome, patients with luminal A tumors were associated with a favourable prognosis and patients with luminal B tumors showed intermediate outcome (3, 44). In addition, patients with luminal B tumors were less sensitive to endocrine therapy (the therapy for hormone receptor-positive breast cancer) than patients with luminal A tumors. However, patients with luminal B tumors showed a better response to neoadjuvant chemotherapy compared to patients with luminal A tumors, achieving higher pathologic complete response rate (5). Breast cancer patients with HER2-enriched tumors benefited from the treatment with an anti-HER2 agent, trastuzumab (33).

1.1.2. Molecular classification of breast cancer subtypes using immunohistochemical surrogates Given that methods for gene expression profiling present some limitations in the context of routine clinical pathology (unavailability of these techniques in many laboratories, specialized knowledge required to interpret results, requirement of fresh or frozen tissue), an alternative approach to identify breast cancer subtypes has been developed (33). Intrinsic breast cancer subtypes as primarily identified by gene expression profiling can, to a certain extent, be approximated using evaluation of molecular markers that can be assessed using routine clinical pathology methods, including immunohistochemistry (IHC) and in situ hybridization (ISH) techniques (37, 45-48). IHC allows the evaluation of protein expression in tissue sections by the utilization of antibodies directed against the specific protein (49). ISH methods allow the determination of the gene copy number within cell nuclei using a DNA probe specifically directed against the gene of interest (50).

This simplified clinicopathological classification was included in the 2011 St Gallen International Expert Consensus Panel in order to propose treatment strategies (51). In this classification breast cancer subtypes are defined on the basis of immunohistochemical analysis of ER, PR, the evaluation of overexpression and/or amplification of HER2, and the

- 5 -

assessment of the Ki-67 index, defined as the percentage of cancer cells positive for Ki-67 staining (Table 1.1.). Ki-67 is a cell proliferation marker used for the evaluation of cell proliferation in clinical specimens and it is considered one of the most important cell proliferation-related genes (52).

Table 1.1. Surrogate definition of intrinsic subtypes of breast cancer as defined by the 2011 St Gallen International Expert Consensus Panel

Intrinsic subtype (GEP) IHC and FISH classification (St. Gallen)

Luminal A “Luminal A” ER-positive and/or PR-negative, HER2-negative, Ki-67 low (< 14%)a Luminal B “Luminal B (HER2-negative)” ER and/or PR-positive, HER2-negative, Ki-67 ≥ 14% “Luminal B (HER2-positive)” ER and/or PR-positive, Any Ki-67, HER2 overexpressed or Amplified HER2-enriched “HER2-positive” HER2 overexpressed or amplified, ER and PR absent Basal-like “Triple negative” ER and PR absent, HER2-negative Abbreviations: GEP: gene expression profiling; IHC: immunohistochemistry; FISH: fluorescence in situ hybridization; ER: Estrogen Receptor; PR: Progesterone receptor; HER2: human epidermal growth factor receptor 2. a If a reliable Ki-67 measurement is not available, some alternative assessment of tumor proliferation such as tumor grade may be used to distinguish between “Luminal A” and “Luminal B” subtypes.

Adapted from Strategies for subtypes-dealing with the diversity of breast cancer: highlights of the St. Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2011 (51).

- 6 -

A Ki-67 cut-off value of 14% was recommended by the Panel with the aim of distinguishing “luminal A” from “luminal B (HER2-negative)” breast carcinomas (51). The differentiation between luminal A and luminal B tumors carries important therapeutic implications. Luminal A and luminal B tumors represent in fact clinically distinct subtypes in term of differentiation and proliferation, prognosis and treatment. Luminal A tumors tend to be more differentiated and show a lower proliferation rate compared to luminal B tumors. Moreover, luminal B patients show a worse prognosis compared to luminal A patients and are treated with chemotherapy in addition to hormone blockage (5).

Major drawbacks of Ki-67 are its lack of reproducibility, in addition to the difficulty to find an appropriate cut-off (10%, 14%, and 20%) (37, 53-55). Therefore, other markers for the evaluation of breast cancer proliferation have been proposed in the literature, including cyclins, topoisomerase II and the mini-chromosome maintenance complex component 2 (MCM2) (56, 57). A recent study has highlighted the clinical relevance of the evaluation of MCM2 in breast cancer specimens (58). MCM2 is a protein that belongs to the MCM complex. This complex functions as a replicative helicase (59, 60), an enzyme that unwinds DNA during chromosomal replication, a crucial step in the initiation of DNA replication (61). MCMs proteins are expressed throughout the G1 phase and have emerged as biomarkers of cell cycle state, as they allow distinguishing cycling cells from quiescent cells (60). Yousef and collaborators demonstrated that MCM2 assessment (using a cut-off of 40%) allowed a clear distinction between luminal A and luminal B/HER2-negative breast cancer specimens (58).

In these few pages, I have briefly highlighted the characteristics of the different breast cancer subtypes. However, since the molecule of interest of the present Ph.D. thesis is the transmembrane receptor HER2, in the following subsection I will focus on the HER2- enriched breast cancer subtype.

1.1.3. HER2-enriched subtype vs. breast cancer clinically evaluated as HER2- positive The HER2-enriched subtype is characterized by high expression of HER2-related and proliferation-related genes (e.g. ERBB/HER2 and GRB7) and shows low expression of luminal clusters (e.g. ESR1 and PGR) compared to luminal A and B tumors (62). In the absence of treatment, patients harboring tumors of the HER2-enriched subtype have a worse prognosis than those with luminal A tumors (5).

- 7 -

It is important to note that the simplified clinicopathological classification of breast cancer subtypes based on IHC and ISH results does not completely overlap with the HER2- enriched subtype based on gene expression profiling since only 50% of tumors belonging to the HER2-enriched subtype are HER2-positive and ER-negative (34, 63-65). It has been reported that 15% of HER2-enriched subtype are both HER2 and ER-positive, 16% are HER2-negative and ER-positive and 18% are negative for both receptors (64). It has been postulated that the subgroup of HER2-enriched tumors that are clinically evaluated as HER2-negative (around 30%) might present mutations of some downstream pathway component that mimic HER2 amplification (64).

1.1.4. Breast cancer clinically evaluated as HER2-positive The HER2 receptor is a 185 kDa transmembrane protein that is encoded by the HER2 (also known as erb-b2 receptor tyrosine kinase 2 [ERBB2]) gene, which is located on the long arm of chromosome 17 (17q12-21.32) (66). HER2 is normally expressed on cell membranes of epithelial cells of several organs like the breast and the skin, as well as gastrointestinal, respiratory, reproductive, and urinary tract (67). Whereas in normal breast epithelial cells HER2 is expressed at low levels (two copies of the HER2 gene and up to 20,000 HER2 receptors) (68), in HER2-positive breast cancer cells there is an increase in the number of HER2 gene copies (up to 25-50, termed gene amplification) and HER2 receptors (up to 40 to 100 fold increase, termed protein overexpression), resulting in up to 2 million receptors expressed at the tumor cell surface (69). Besides breast cancer, HER2 overexpression has also been reported in other types of tumors, including stomach, ovary, colon, bladder, lung, uterine cervix, head and neck, and oesophageal cancer as well as uterine serous endometrial carcinoma (70).

In human breast cancer, HER2 gene amplification and/or receptor overexpression, which occur in 15% to 20% of patients, are important markers for poor prognosis, including a more aggressive disease and shorter survival (4). Moreover, HER2-positive status is considered a predictive marker of response to HER2-targeted drugs, including trastuzumab and lapatinib (71). Patients are eligible for targeted anti-HER2 therapies when their breast cancer specimens overexpress the protein at the IHC and/or are HER2 gene amplified at ISH. Consequently, accurate testing of HER2 status is essential for the clinical management of breast cancer patients.

- 8 -

1.2. Methods for the evaluation of HER2 status in breast cancer specimens

1.2.1. Immunohistochemistry (IHC) and in situ hybridization (ISH) methods Currently, several diagnostic methods are approved for the evaluation of HER2 status in breast cancer specimens in routine clinical practice: immunohistochemistry (IHC) and in situ hybridization (ISH) techniques, most commonly fluorescent ISH (FISH) (7, 8).

IHC allows the evaluation of the HER2 protein expression in formalin-fixed, paraffin- embedded (FFPE) tissues using specific antibodies (8). By this method, it is then possible to estimate the number of cells showing membranous staining in the tissue section as well as the intensity of the staining (72). Membranous staining is scored on a semi-quantitative scale.

ISH techniques, instead, allow the quantification of HER2 gene copy number within tumor cell nuclei using a DNA probe coupled to a fluorescent, chromogenic, or silver detection system (i.e., FISH, chromogenic ISH [CISH], or silver-enhanced ISH [SISH]), or a combination of CISH and SISH systems (bright-field double ISH [BDISH]) (6). ISH is effectuated either as a single-color assay (HER2 probe only) to evaluate HER2 gene copies per nucleus or as a dual-color assay using differentially labeled HER2 and chromosome 17 centromere (chromosome enumeration probe 17, CEP17) probes simultaneously. The dual- color assay allows the determination of the HER2/CEP17 ratio (73). The HER2/CEP17 ratio is often regarded as a better reflection of the HER2 amplification status, as the latter may be influenced by abnormal chromosome 17 copy number (mainly polysomy) (74).

In 2007, the American Society of Clinical Oncology (ASCO)/ the College of American Pathologists (CAP) published guidelines for HER2 testing in breast cancer to improve reliability of HER2 test results (8). These guidelines included recommendations regarding specimen handling and testing requirements. The guidelines also included scoring criteria of immunohistochemical and ISH staining to classify breast cancer specimens into three categories: positive, equivocal and negative (Figure 1.1.). A case is considered positive at IHC when the HER2 expression is scored as 3+ (strong, complete, homogeneous membrane staining in >30% of tumor cells). A case is evaluated as positive (amplified) at ISH when the mean HER2 gene copy number is > 6 signals/nucleus or HER2 gene/chromosome 17 copy number ratio (HER2/CEP17) is > 2.2. Equivocal cases at IHC show a strong, complete membrane staining in ≤ 30% of tumor cells or weak/moderate

- 9 -

heterogeneous complete membrane staining in ≥ 10% of tumor cells (IHC score 2+). Equivocal cases at IHS show a mean HER2 gene copy number of ≥ 4 and ≤ 6 signals/nucleus or a HER2/CEP17 ratio between 1.8 and 2.2. A case is evaluated as negative at IHC when the HER2 expression is scored as 0/1+ (no staining/ weak or incomplete membrane staining in any percentage of tumor cells). A sample is evaluated as negative (non-amplified) at ISH when the mean HER2 gene copy number is < 4 signals/nucleus or the HER2/CEP17 ratio is < 1.8. The equivocal category requires additional testing with the alternative assay (IHC or ISH) for final determination. In 2013, the ASCO/CAP revised the guidelines published in 2007 to clarify the recommendations for HER2 testing in breast cancer specimens (7). The updated guidelines included new scoring criteria for IHC and ISH.

Figure 1.1. Immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) results of HER2.

A B C

D E F

Representative IHC HER2 staining are presented in the first row : A. Negative staining (IHC score 0/1+); B. Equivocal staining (IHC score 2+); C. Positive staining (IHC score 3+); Representative HER2 FISH results are presented in second row : D. Negative FISH result (non-amplified); E. Equivocal FISH result (equivocal); F. positive FISH result (amplified).

A detailed description of the IHC and the ISH methods for the evaluation of HER2 status in breast cancer specimens, including the commercially available diagnostic tests and antibody clones as well as scoring criteria for reporting IHC and ISH results according to the ASCO/CAP recommendations (2007 and 2013 ASCO/CAP scoring criteria) are presented in Chapter 2.

- 10 -

The updated ASCO/CAP guidelines also presented a new approach for the determination of HER2 status in breast cancer specimens. While the 2007 ASCO/CAP guidelines recommended performing HER2 testing on resection specimens and to retest when results were equivocal (8), the updated guidelines recommend effectuating an initial test (IHC or ISH) on core biopsy (7). If test results are equivocal or if there is an apparent histopathological discordance with the test result, reflex testing on tumor specimen section with an alternative assay (IHC or ISH) should be performed.

Although in clinical practice IHC and ISH are the most frequently used techniques for the evaluation of HER2, other methods have been developed over the years for the determination of HER2 status in breast cancer specimens. An exhaustive review of all existing methods to assess HER2 status in breast cancer, including a discussion on the advantages and disadvantages of each described technique, is presented in Chapter 2.

1.2.2. Methodological worries related to HER2 testing Although the FISH assay is considered a more objective and quantitative method compared to IHC, this technique presents several disadvantages (6). For example, manual counting of fluorescent signals is time-consuming. To overcome this limitation, several automated image analysis software have been developed for the enumeration of fluorescent signals (75, 76). Some of these software employ a programming algorithm through which the software quantifies fluorescent signals in images on the basis of square tiles of fixed dimensions (77). The major drawback for this method of analysis is that the size of the tile does not always correspond to the size of a single tumor cell nucleus. This method might therefore not completely reflect the biology of cells (78). In the literature, one study has analyzed the utility of an image analysis software (EIKONA3D, Alpha Tec Ltd) for the evaluation of HER2 amplification in single tumor cell nuclei in a cohort of 100 breast cancer cases (79). The authors found a very good concordance (100.0%) between the results obtained by manual scoring and those obtained with the image analysis software for non-amplified cases. However, the concordance between the two methods for amplified cases was moderate (74.1%). Therefore, better performing software programming algorithms that analyse fluorescent signals in single tumor cell nuclei within breast cancer tissue section are warranted.

The high cost of the FISH technique represents a further limitation of this method. Since tissue microarray (TMA) technology facilitates the simultaneous molecular characterization

- 11 -

of a large amount of specimens, this technique might represent an economical replacement for whole tissue section analysis for breast cancer specimens. This technique has already been extensively described elsewhere (80-82), therefore it is only briefly described here (Figure 1.2.). Tissue core cylinders obtained from several formalin-fixed, paraffin-embedded (FFPE) specimens are placed into a single, empty paraffin block using manual, automated or semiautomatic tissue microarrayers (82). Before taking tissue cores from the donor blocks, regions of interest are marked on hematoxylin & eosin (H&E) slides. The corresponding regions on the matching paraffin block (donor block) is then identified and marked. Tissue cores are removed from these areas of the donor block using hollow needles of the microarray instrument and placed at defined array coordinates in the paraffin recipient block. The diameter of tissue cores ranges from 0.6 to 2.0 mm. Location and details of each single core are then recorded to generate a clinical database. To date, only a few studies have analyzed the utility of TMA in the evaluation of HER2 gene amplification by FISH in breast cancer specimens (83-86). In these studies, the authors reported excellent agreement rates (ranging from 91% to 97%) between HER2 gene amplification status determined on whole tissue section and on TMA section. To date, no study evaluating this concordance using the 2013 ASCO/CAP scoring criteria has been conducted.

- 12 -

Figure 1.2. Construction of Tissue microarray

Cylindric tissue cores are removed from donor paraffin block and transferred into premade holes of an empty recipient paraffin block using a tissue microarrayer. For details, see text.

Source:http://apps.pathology.jhu.edu/blogs/pathology/tissue-microarrays.

1.2.3. Concordance of HER2 status determined by IHC, FISH, CISH and SISH in the literature Although several methods have been approved by the FDA for the evaluation of HER2 status in breast cancer specimens, measure variation between these techniques has been reported. Evaluation of the concordance of HER2 status determined by IHC and FISH has already been studied in numerous studies (for a systematic review and a meta-analysis see (87) and (88), respectively). The concordance between IHC and FISH was 96% for IHC- negative cases (IHC 0/1+) and 91% for IHC-positive cases (IHC 3+) cases. Among the IHC- equivocal cases (IHC 2+), 36% showed HER2 gene amplification and 64% were considered non-amplified (88). The concordance rate was influenced by several factors, including the antibody clone and the scoring criteria used (7, 8, 89). The majority of the studies that analyzed the concordance between IHC and FISH have been conducted on cohort of selected breast cancer specimens, where cases with equivocal immunostaining or with borderline HER2 gene amplification were overrepresented. Only a few studies have analyzed the concordance between IHC and FISH for all IHC categories in a cohort of consecutive breast cancer cases (72, 90-93).

- 13 -

Several recent studies have also evaluated the concordance between HER2 status determined by IHC, FISH, CISH and SISH. Concordance between CISH and IHC and FISH has been reviewed elsewhere (94). A detailed overview of the studies that have analyzed the concordance between IHC and SISH (single-color as well as dual-color), between FISH and SISH (single-color as well as dual-color) and between single-color CISH and dual-color SISH in breast cancer specimens is presented in Supplementary Table S1 of Annex I. Briefly, concordance between IHC and CISH ranged between 55.0% and 99.0%, whereas concordance between FISH and CISH ranged between 82.0% and 100.0%. Concordance between IHC and dual-color SISH ranged between 91.9% and 96.6%, whereas concordance between FISH and dual-color SISH ranged between 83.0% and 98.9%. Concordance between IHC and single-color SISH ranged between 98.3% and 99.3%, whereas concordance between FISH and single-color SISH was 97.0% (see Table S1 of Annex I).

1.3. HER2 biology

1.3.1. HER2 structure and function HER2 belongs to the epidermal growth factor receptor (EGFR) family. The EGFR family (also known as HER receptors family) is composed of four HER receptors: EGFR (also termed human epidermal growth factor receptor 1 [HER1]), HER2, human epidermal growth factor receptor 3 (HER3), and human epidermal growth factor receptor 4 (HER4) (95).

HER family members are transmembrane receptor tyrosine kinases. Tyrosine kinases are enzymes that carry out tyrosine phosphorylation, namely the transfer of the γ phosphate of adenosine triphosphate (ATP) to tyrosine residues on protein substrate (96).

HER receptors share a similar structure. They are composed of an extracellular domain (ECD), a transmembrane segment and an intracellular region (97). The ECD domain is divided into four parts: domains I and III, which play a role in ligand binding, and domains II and IV, which contain several cysteine residues that are important for disulfide bond formation (98). The transmembrane segment is composed of 19-25 amino acid residues. The intracellular region is composed of a juxtamembrane segment, a functional protein kinase domain (with the exception of HER3 that lacks tyrosine kinase activity (99) and must partner with another family member to be activated (100), and a C-terminal tail containing multiple phosphorylation sites required for propagation of downstream signaling (98). The

- 14 -

catalytic domain contains the ATP binding pocket, a conserved site essential to ATP binding (101).

HER receptors are activated by both homo- and heterodimerization, generally induced by ligand binding (102). This suggests that HER receptor family has evolved to provide a high degree of signal diversity (103). The cellular outcome produced by HER receptors activation depends on the signaling pathways that are induced, as well as their magnitude and duration, which are influenced by the composition of the dimer and the identity of the ligand (103).

Several growth factor ligands interact with the HER receptors (104). HER1 receptor is activated by six ligands: epidermal growth factor (EGF), epigen (EPG), transforming growth factor  (TGF), amphiregulin, heparin-binding EGF-like growth factor, betacellulin and epiregulin. HER3 and HER4 receptors bind neuregulins (neuregulin-1, neuregulin-2, neuregulin-3, and neuregulin-4). HER2 is a co-receptor for many ligands and is often transactivated by EGF-like ligands, inducing the formation of HER1-HER2 heterodimers. Neuregulins induces the formation of HER2-HER3 and HER2-HER4 heterodimers (104). However, no known ligand can promote HER2 homodimer formation, implying that no ligand can bind directly to HER2 (105).

The structural basis for receptor dimerization has been elucidated in recent years through crystallographic studies (106, 107). Dimerization is mediated by the dimerization arm, a region of the extracellular region of HER receptors. While in its inactivated state the dimerization arm of EGFR, HER3 and HER4 is hidden, ligand binding induces a receptor conformational change leading to exposure of the dimerization arm (106). In contrast to the other three HER receptors, the dimerization arm of the HER2 receptor is permanently partially exposed, thus permitting its dimerization even if the HER2 receptor lacks ligand- binding activity (107).

Interaction between the dimerization arms of two HER receptors promotes the formation of a stable receptor dimer in which the kinase regions of both receptors are closed enough to permit transphosphorylation of tyrosine residues, i.e. the transfer of a phosphate group by a protein kinase to a tyrosine residue in a different kinase molecule (108, 109). The first member of the dimer mediates the phosphorylation of the second, and the second dimer mediates the phosphorylation of the first (98).

- 15 -

The phosphorylation of specific tyrosine residues following HER receptor activation and the subsequent recruitment and activation of downstream signaling proteins leads to activation of downstream signaling pathways promoting cell proliferation, survival, migration, adhesion, angiogenesis and differentiation (110). The Phosphatidylinositol 3’-kinase (PI3K)- Akt pathway and the Ras/Raf/MEK/ERK pathway (also known as extracellular signal- regulated kinase/ mitogen-activated protein kinase (ERK/MAPK) pathway) are the two most important and most extensively studied downstream signaling pathways that are activated by the HER receptors (Figure 1.3.) (95, 111). These downstream signaling cascades control cell cycle, cell growth and survival, apoptosis, metabolism and angiogenesis (112, 113). Signaling from HER receptors is then terminated through the internalization of the activated receptors from the cell surface by endocytosis. Internalized receptors are then either recycled back to the plasma membrane (HER2, HER3, HER4) or degraded in lysosomes (HER1) (114, 115).

- 16 -

Figure 1.3. Signal transduction by the HER family members.

The HER receptors family is composed of four members: HER1, HER2, HER3, and HER4. Several growth factor ligands interact with HER1, HER3, and HER4. Phosphorylation of the tyrosine kinase domain by means of homodimerization or heterodimerization induces both cell proliferation and survival signaling. HER2 is the preferred dimerization partner for the other HER family members. The phosphorylated tyrosine residues on the intracellular domain of HER2 activate phosphoinositide 3- kinase (PI3-K)/Akt, driving cell survival. In addition, HER2 activation leads to activation of the RAS/RAF/mitogen-activated protein kinase (MAPK) pathway.

Adapted from Trastuzumab - Mechanism of action and use in clinical practice (116).

HER heterodimers produce more potent signal transduction than homodimers. This can be explained by the fact that heterodimerization provides additional phosphotyrosine residues necessary for the recruitment of effector proteins (103). Heterodimerization follows a strict hierarchical principle with HER2 representing the preferred dimerization and signaling partner for all other members of the HER family (117). HER2 seems to function mainly as a co-receptor, increasing the affinity of ligand binding to dimerized receptor complexes (118, 119). HER2 has the strongest catalytic kinase activity (117) and HER2-containing heterodimers produce intracellular signals that are significantly stronger than signals

- 17 -

generated from other HER heterodimers (120). The HER2-HER3 heterodimer in particular exhibits extremely potent mitogenic activity through the stimulation of the PI3K/Akt pathway, a master regulator of cell growth and survival (121). Furthermore, HER2 containing heterodimers have a slow rate of receptor internalization, which results in prolonged stimulation of downstream signaling pathways (103). HER2 can also be activated by complexing with other membrane receptors, such as Insulin-like growth factor I receptor (IGF-1R) (122).

1.3.2. Consequences of constitutive HER2 receptor activation Whereas in normal cells the activity of tyrosine kinases is a tightly controlled mechanism, in cancer cells alterations in tyrosine kinases (overexpression of receptor tyrosine kinase proteins, amplification or mutation in the corresponding gene, abnormal stimulation by autocrine growth factors loop or delayed degradation of activated receptor tyrosine kinase) lead to constitutive kinase activation and therefore to aberrant cellular growth and proliferation (109, 123). Constitutive activation of HER1, HER2, HER3, IGF-1R, Fibroblast growth factor receptor (FGFR), c-Met, Insulin Receptor (IR), Vascular Endothelial Growth Factor Receptor (VEGFR), Jak kinases and Src have been associated with human cancer (109, 124-128).

Several ways of aberrant activation of HER receptors have been described, including ligand binding, molecular structural alterations, lack of the phosphatase activity, or overexpression of the HER receptor (129).

In HER2-positive tumors, receptor overexpression has been identified as the mechanism of HER2 activation. The increased amount of cell surface HER2 receptors associated with HER2 overexpression leads to increased receptor-receptor interactions, provoking a sustained tyrosine phosphorylation of the kinase domain and therefore constant activation of the signaling pathways. HER2 overexpression also enhances HER2 heterodimerization with HER1 and HER3 (130) resulting in an increased activation of the downstream signaling pathways. It has also been shown that HER2 overexpression leads to enhanced HER1 membrane expression and HER1 signaling activity through interference with the endocytic regulation of HER1 (130-132). While HER1 undergoes endocytic degradation after ligand- mediated activation and homodimerization, HER1-HER2 heterodimers evade endocytic degradation in favor of the recycling pathway (133, 134), resulting in increased HER1 membrane expression and activity (131, 132, 135).

- 18 -

It has also been reported that HER2 overexpression enhances cell proliferation through the rapid degradation of the cyclin-dependent kinase (Cdk) inhibitor p27 and the upregulation of factors that promote cell cycle progression, including Cdk6 and cyclins D1 and E (136).

1.4. Treatment of HER2-positive breast cancer patients Current treatment strategies for breast cancer patients depend on several factors: histological type and grade, tumor size, lymph node status, cancer stage (early breast cancer or metastatic disease), menopausal status, patient age, ER status, and HER2 status.

Standard of care for HER2-positive breast cancer patients consists in mastectomy (segmental or total) and radiotherapy (when indicated), followed by systemic chemotherapy, adjuvant hormonal therapy (for patients with ER-positive breast cancer), and treatment with anti-HER2 agents.

1.4.1. Anti-HER2 agents FDA-approved anti-HER2 agents currently used in clinical settings in combination with chemotherapy can be subdivided into two classes (137). The first class comprises recombinant humanized monoclonal antibodies that bind the extracellular regions of HER2. Examples for this class of agents are trastuzumab and pertuzumab. The second class comprises small molecule tyrosine kinase inhibitors (TKIs), like lapatinib. They inhibit enzyme function of the intracellular catalytic domain of HER.

Trastuzumab (Herceptin®) is composed of the antigen-binding fragment (Fab) of the murine monoclonal antibody 4D5, directed against the ECD of HER2, spliced to the Fc fragment of human Immunoglobulin G (IgG) (138) (Figure 1.4.). In 1998, the FDA approved this drug for the treatment of HER2-positive metastatic breast cancer and in 2006 for the adjuvant treatment of HER2-positive early stage breast cancer. Although it has been shown that trastuzumab improves the outcome of HER2-positive breast cancer patients in both the metastatic and adjuvant settings, its mechanisms of action are not fully understood (139). Several possible mechanisms have been postulated, including inhibition of HER2 signaling through prevention of HER2 dimerization and receptor downregulation through stimulation of HER2 endocytosis (Figure 1.4.) (140). It has been also reported that trastuzumab inhibits the proteolytic cleavage of the HER2 ECD mediated by metalloproteases (141). The cleavage of the HER2 ECD leads to the formation of a truncated membrane-bound fragment (p95HER2) with constitutive tyrosine kinase activity (116). Trastuzumab produces a

- 19 -

cytostatic effect that is associated with G1 arrest via upregulation of p27, which binds and inhibit the Cyclin E-Cdk2 complex (142, 143). It has been shown that trastuzumab inhibits PI3K signaling by increasing the activity of phosphatase and tensin homolog deleted on chromosome 10 (PTEN) (144), a tumor suppressor gene that counteracts the PI3K pathway (145). Trastuzumab modulates the effects of different pro- and anti-angiogenic factors. In particular, it inhibits the production of VEGF (146). Moreover, in animal models trastuzumab induced the recruitment of Fc-competent immune effector cells, especially Natural killer (NK) cells, leading to tumor cell death (147). This mechanism, known as antibody-dependent cellular cytotoxicity (ADCC), occurs when the Fc gamma receptor (FcgR) on immune effector cells bind to Fc region of antibodies (148). The process involves binding of the antibody (trastuzumab) to the target antigen (HER2), recognition of the Fc fragment by immune effector cell via FcgR and activation of the immune effector cells (NK cells, macrophages). This drives the release of cytoplasmic granules containing perforin and granzymes, which ultimately leads to the destruction of the trastuzumab-coated HER2- positive breast cancer cells (149). The effects of trastuzumab on HER2 tyrosine phosphorylation are still a matter of debate. Whereas some studies suggest that trastuzumab blocks tyrosine phosphorylation of the receptor (16, 150), others postulate that the drug may induce HER2 phosphorylation (151, 152).

- 20 -

Figure 1.4. Proposed mechanisms of action of trastuzumab

A. Trastuzumab is composed of the antigen-binding fragment (Fab) of the murine monoclonal antibody 4D5, directed against the ECD of HER2, spliced to the Fc fragment of human Immunoglobulin G (IgG); B. Trastuzumab reduces shedding of the HER2 ECD, thereby reducing the formation of p95, a truncated membrane-bound fragment with constitutive tyrosine kinase activity. B. Trastuzumab inhibits HER2 signaling; C. Trastuzumab recruits Fc-competent immune effector cells favoring tumor-cell death; D. Trastuzumab enhances receptor downregulation through stimulation of HER2 endocytosis.

Adapted from Trastuzumab – Mechanism of action and use in clinical practice (116).

- 21 -

Trastuzumab and the other recombinant humanized monoclonal antibody, pertuzumab, bind to different epitopes in the ECD of HER2 (Figure 1.5.). Whereas trastuzumab binds to domain IV (153), pertuzumab binds near the center of domain II (154). Since their binding sites in the extracellular binding region of the protein are distinct, pertuzumab exerts its effects on HER2- positive tumor cells in a different way. In 2012, the FDA approved pertuzumab (Perjeta®) for use in combination with trastuzumab and docetaxel for treatment of patients with HER2- positive metastatic breast cancer (155). Pertuzumab sterically blocks dimerization of HER2 with EGFR and HER3 through binding to the dimerization arm of HER2, resulting in inhibition of signaling from HER2/EGFR and HER2/HER3 heterodimers (156). It has been shown that pertuzumab is more efficient than trastuzumab in the inhibition of HER2 heterodimerization (156, 157). It has been reported that pertuzumab reduced HER2 phosphorylation and inhibited cellular growth of HER2-positive xenografts tumors, probably through the disruption of dimerization of HER2 with other HER family members (156).

Figure 1.5. Agents that target HER2 receptor.

Monoclonal antibodies bind to epitopes located in the HER2 extracellular domain. Trastuzumab binds to subdomain IV, and pertuzumab to the dimerization arm of subdomain II. Lapatinib is an ATP-competitive inhibitor that acts by displacing ATP from binding its ATP binding pocket located in the kinase region.

Source : The evolving landscape of protein kinases in breast cancer : clinical implications (158).

- 22 -

In 2007, the FDA approved lapatinib (Tykerb/Tyverb) as combination therapy with the chemotherapy drug capecitabine for the treatment of patients with HER2-positive advanced breast cancer who have progressed on trastuzumab-based regimens. Lapatinib is an orally active dual TKI that targets both HER2 and HER1. It is an ATP-competitive inhibitor that acts by displacing ATP from binding its ATP binding pocket located in the kinase region (158). As a small molecule, lapatinib is able to cross the blood-brain barrier to provide effective therapeutic concentrations in cerebrospinal fluid. Contrary to monoclonal antibodies (that are unable to penetrate the blood-brain barrier due to their relatively high molecular weight) (159), this agent shows activity against central nervous system metastases (45, 160, 161). It has been observed that lapatinib blocked phosphorylation of HER2 and EGFR in HER2-positive breast cancer cell lines (152, 162).

In 2013, the FDA approved Trastuzumab-Emtansine (T-DM1, Kadcyla) for the treatment of metastatic HER2-positive breast cancer patients who have progressed during trastuzumab treatment (163). T-DM1 is an antibody-drug conjugate (ADC) comprising the chemotherapeutic agent emtansine conjugated to trastuzumab with a synthetic protease-cleavable linker. T-DM1 combines the action of trastuzumab with that of emtansine, a potent cytotoxic drug which blocks mitosis by inhibiting the assembly of microtubule in actively-dividing cells (164). Following the binding of T-DM1 to HER2 on the cell surface, the T-DM1/HER2 complex is internalized via endocytosis and undergoes proteolytic degradation, which ultimately leads to the intracellular release of the active compound (163). Combining these two agents facilitates anti-HER2 activity, in addition to selectively deliver the potent cytotoxic agent to tumor cells and to reduce therefore the systematic toxicity associated with chemotherapies (165).

Erlotinib (Tarceva) and gefitinib (Iressa™) are HER1 TKIs that have shown only minimal clinical benefit in HER2-positive metastatic breast cancer patients either as single agents or in combination with other agents (166-169). Other small TKIs that are currently in clinical development for the treatment of metastatic HER2-positive breast cancer patients are neratinib (HER1 and HER2 TKI) (170, 171) and afatinib (HER1, HER2 and HER4 TKI) (172). Moreover, alternative therapeutic approaches targeting HER2 downstream signaling pathways are currently being tested, including the mTOR inhibitors temsirolimus (173) and everolimus (174, 175).

Although anti-HER2 agents are generally well tolerated, trastuzumab administration has been associated with cardiac side effects, especially when used in combination with anthracyclines

- 23 -

(176). The reported incidence of trastuzumab-induced cardiotoxicity in clinical trials ranges from 2% to 4% of trastuzumab-treated patients (177). Cardiac dysfunction associated with trastuzumab ranges from asymptomatic drops in left ventricular ejection fraction (LVEF) to symptomatic heart failure (178). In most cases, decreases in LVEF are reversible (179).

The mechanisms underlying the cardiotoxic effects of trastuzumab remain unclear. It has been proposed that cardiotoxicity can be produced by the blockade of HER2 signaling in cardiomyocytes. Signaling through HER2-HER4 heterodimers plays a central role during the development of the embryonic heart and is essential for the contractile function in the adult heart (179). It has also been shown that mice carrying a conditional HER2 mutation developed a severe dilated cardiomyopathy (180). Furthermore, ADCC might contribute to trastuzumab- related cardiotoxicity (179). In addition, trastuzumab increases the formation of reactive oxygen species (ROS), exacerbating the damages to cardiac cellular membrane caused by free radicals produced by anthracyclines (181).

Aside from prior or concurrent exposure to anthracyclines, recognized risk factors for trastuzumab-related cardiac side effects include hypertension, and a LVEF < 55% at baseline (182).

Other HER2 inhibitors including pertuzumab, TDM-1 and lapatinib are less cardiotoxic than trastuzumab. Moreover, a combination of these agents with trastuzumab did not significantly increase cardiac adverse events (183).

Other most common side effects associated with the administration of HER2 inhibitors are diarrhea (pertuzumab, lapatinib), nausea (pertuzumab, TDM-1, lapatinib), thrombocytopenia (TDM-1), headache (lapatinib), dyspnea (lapatinib), dehydration (lapatinib), and rash (lapatinib) (182, 184, 185).

1.4.2. Endocrine therapy Since 15% of HER2-positive breast cancer patients are also considered ER-positive, they are also eligible for endocrine therapy (64). Drugs that inhibit or block estrogens are used for the treatment of ER-positive breast cancer patients (186). Current commonly used agents for adjuvant endocrine therapy can be divided into the following three classes: selective estrogen receptor modulators (SERMs), aromatase inhibitors (AIs), and selective estrogen receptor down-regulators (SERDs) (187).

- 24 -

SERMs are ER ligands that can modulate the transcriptional capabilities of the receptor in estrogen sensitive tissues. More precisely, they function as selective ER antagonists or agonists depending on the target tissue in which they operate (188). SERMs bind to ER (both ERα and ERβ), alter receptor conformation, and promote binding of coregulatory proteins that activate or repress transcriptional activation of estrogen target genes (189). Their tissue- specific agonist-antagonists activity can be explained by three distinct mechanisms: differential ER-expression within target tissue, differential ER conformation on ligand binding, and differential expression and binding to the ER coregulatory proteins (190).

The SERMs approved for clinical use are tamoxifen and raloxifene (Evista®) (191). Tamoxifen is currently used for the treatment of pre- and postmenopausal patients with ER-positive early breast cancer. Beneath estrogen blockade with tamoxifen, further strategies for the treatment of premenopausal ER-positive breast cancer patients include temporary or permanent interruption of ovarian estrogen synthesis by luteinizing hormone-releasing hormone (LHRH) agonists, or with oophorectomy, respectively (192).

The chemical structure of tamoxifen is very similar to that of estradiol. Tamoxifen, however, has a side chain that is important for its antagonistic action, since it hinders ER from adopting its active conformation (193). It functions through its active metabolites, 4-hydroxytamoxifen and endoxifen (194, 195). It acts as an estrogen antagonist in the mammary gland, but as an agonist in the uterus, endometrium and bone (189). Its agonist activity is associated with increased risk of endometrial cancer, ischaemic cerebrovascular and venous thromboembolic events, particularly with long-term use (190, 196).

Raloxifene acts as an estrogen antagonist in the breast and uterus, and as an estrogen agonist in bone, in the cardiovascular system and in the regulation of serum lipids (197). It was developed to avoid the uterotrophic effects of tamoxifen (197). Treatment with raloxifene has been approved for the prevention and treatment of osteoporosis in post-menopausal women (198). Moreover, as randomized clinical trials showed that treatment with raloxifene was associated with a decrease in breast cancer incidence in post-menopausal women (199, 200), raloxifene has been approved also for the prevention of breast cancer in high-risk postmenopausal women (198). Raloxifene significantly reduced breast cancer incidence in postmenopausal patients, with results comparable to that of tamoxifen, but with fewer of the serious side effects associated with tamoxifen treatment, including a lower risk of uterine cancer (201).

- 25 -

AIs inhibit aromatase cytochrome p450, an enzyme responsible for the peripheral conversion of androgens to estrogens. In postmenopausal women, this peripheral conversion (especially in adipose tissue and the adrenal glands) is the principal source of estrogen. The inhibition of aromatization of androgens by AIs lowers both serum and intra-mammary tissue levels of estrogens, thereby blocking tumor cell growth. Currently, three third-generation AIs have been approved by the FDA: anastrozole (Arimidex®) and letrozole (Femara®) that reversibly bind aromatase and exemestane (Aromasin®) that irreversibly binds aromatase (202).

The partial agonist effects of the SERMs and the development of cross-resistance between endocrine therapies with analogous modes of action have led to development of new hormonal therapies with novel mechanisms of action, such as SERDs (196). SERDs are pure ER antagonists. They both antagonize and, as their name suggests, degrade ER (203). Fulvestrant (Faslodex®) is a SERD that is indicated for the treatment of postmenopausal women with ER-positive, locally advanced or metastatic breast cancer. It is also active in patients who have progressed on adjuvant antiestrogen therapy such as tamoxifen or AIs, as they lack cross-resistance with tamoxifen (204). It competitively binds to ER with a binding affinity approximately 100 times greater than that of tamoxifen (203). It blocks dimerization and nuclear localisation of ER and it inhibits ER-mediated gene transcription (205). In addition, cellular ER expression in cells is reduced via several mechanisms, including stimulation of ER turnover, reduction of its half-life and stimulation of its degradation (205).

1.4.3. Chemotherapeutic drugs Despite the introduction of anti-HER2 therapies, chemotherapeutic agents remain a cornerstone of treatment for HER2-positive breast cancer patients. Chemotherapeutic drugs that are used for the treatment of HER2-positive breast cancer patients include taxane derivatives, carboplatin, cyclophosphamide, 5-Fluorouracil, and anthracycline agents. Chemotherapeutic agents target proliferating cells. They induce apoptosis through interference with DNA synthesis and replication (206).

The taxane derivatives [paclitaxel (taxol®) and docetaxel (taxotere®)] act as potent inhibitors of cell replication. They disrupt the microtubule network in cells that is essential for mitosis. Through binding to free tubulin, they produce dysfunctional microtubular bundles that inhibit mitosis. The accumulation of such bundles leads ultimately to apoptosis (207).

- 26 -

Carboplatin is a platinum-based drug. It undergoes activation inside cells and forms highly reactive platinum complexes that cause intra- and interstrand DNA crosslinks leading to inhibition of DNA replication (208).

Cyclophosphamide is an alkylating agent. It adds alkyl groups to DNA bases, resulting in DNA fragmentation via the action of repair enzymes in their attempt to replace the alkylated bases. Moreover, it leads to cross-linkage of guanine bases in DNA double-helix strands, which prevents DNA replication. It has also been reported that it produces mutations through the induction of nucleotide mispairing (209).

The antimetabolite drug 5-Fluorouracil is a pyrimidine analogue that can be incorporated into DNA or RNA in place of uracil or thymine, leading to damages and DNA and RNA malfunctions (210).

Epirubicin and doxorubicin (Adriamycin®) are anthracycline agents. They bind to DNA by intercalation between base pairs. They also inhibit the function of topoisomerase II. Topoisomerase II is an enzyme that plays an important role during DNA replication as it modulates the supercoiling of DNA through the formation of transient double stranded breaks in DNA, followed by enzymatic re-ligation of the cleaved DNAs. Anthracyclines enhance topoisomerase II-mediated DNA breaks by interfering with the enzymatic re-ligation of the cleaved DNA, leaving the double strand breaks. Moreover, they generate toxic oxygen-free radicals that cause DNA damages (211).

The major drawback of chemotherapeutic agents is that they affect rapidly diving cells, regardless of whether these cells are malignant or normal. Rapidly dividing normal cells include cells of the bone marrow and the intestine. Chemotherapeutic agents are therefore associated with several adverse effects, most commonly bone marrow suppression and gastrointestinal side effects (nausea, vomiting) (212).

1.5. Resistance to trastuzumab Two Cochrane systematic reviews reported that the administration of trastuzumab has led to significant improvement in survival of HER2-positive breast cancer patients when combined with standard chemotherapy in the adjuvant and metastatic settings (9, 10). Trastuzumab- containing regimens improved OS and DFS in HER2-positive women with early and locally advanced breast cancer (Hazard Ratio (HR) 0.66; 95% confidence interval (CI) 0.57 to 0.77, P < 0.00001; and HR 0.60; 95% CI 0.50 to 0.71, P < 0.00001, respectively) (9) and OS and

- 27 -

PFS in metastatic HER2-positive breast cancer patients (HR 0.82, 95% CI 0.71 to 0.94, P = 0.004; and HR 0.61, 95% CI 0.54 to 0.70, P < 0.00001, respectively) (10). Trastuzumab administration has therefore become a mainstay of clinical management of HER2-positive breast cancer patients. Despite this noteworthy achievement, a significant fraction of HER2- positive breast cancer patients treated with trastuzumab eventually relapse or develop progressive disease, suggesting that tumors possess or acquire mechanisms of resistance that allow escape from HER2 inhibition (11).

1.5.1. Molecular mechanisms of trastuzumab resistance Mechanisms of resistance to trastuzumab that have been proposed in the literature can be divided into five main categories: 1). Obstacles interfering with the binding of trastuzumab to HER2; 2). Upregulation of HER2 downstream signaling pathways; 3). Signaling through alternative pathways; 4). Defects in apoptosis and cell cycle control; and, 5). Influence of the tumor microenvironment (12, 213).

1). Obstacles interfering with the binding of trastuzumab to HER2

Several mechanisms that abrogate the binding of trastuzumab to HER2 have been associated with trastuzumab resistance (Figure 1.6.). A retrospective study has observed a strong association between the presence of p95HER2 and clinical resistance to trastuzumab treatment (214). As previously mentioned, p95HER2 is the truncated form of HER2, which lacks the ECD of the receptor, while retaining a functional HER2 kinase domain (215). As the ECD contains the trastuzumab binding site, trastuzumab cannot bind to HER2. In agreement with this observation, trastuzumab did not inhibit cell growth of p95HER2-positive MCF-7 cells. By contrast, lapatinib exerted several effects on p95HER2-positive MCF-7 cells, including inhibition of p95HER2 phosphorylation, inhibition of downstream phosphorylation of Akt and MAPK, and inhibition of cell growth (214). In addition, lapatinib inhibited growth of MCF-7p95HER2 xenografts tumors (214).

Mucin 4 (MUC4) is a large, membrane-associated glycoprotein (216), which may hinder trastuzumab binding to the HER2 receptor through epitope masking (217). Using the breast cancer cell line JIMT-1, which is a HER2-positive cell line established from a breast cancer patients showing HER2 gene amplification and primary resistance to trastuzumab (218), Nagy et al. reported that the presence of MUC4 was associated with decreased antibody-binding

- 28 -

capacity (219). Importantly, JIMT-1 resistance to trastuzumab could be reversed using RNA interference to knockdown MUC4 expression (219).

The CD44/hyaluronan polymer complex (CD44 is a transmembrane receptor for hyaluronan) (220) might mask HER2 epitopes and may therefore also prevent the binding of trastuzumab to its target. A preclinical study reported that treatment of JIMT-1 cells with an inhibitor of hyaluronan synthesis significantly decreased the hyaluronan levels in this cell line, leading to enhanced binding of trastuzumab to the HER2 receptor and its subsequent growth inhibitory effect (221).

2). Upregulation of HER2 downstream signaling pathways

Constitutive activation of the PI3K–Akt–mTOR signaling pathway drives aberrant cell growth and proliferation in a variety of tumor types, including breast cancer (222). In particular, upregulation of the PI3K pathway has been proposed as a mechanism for trastuzumab resistance (223). Persistent activation of PI3K occurs via two mechanisms: loss of function of the tumor suppressor gene PTEN, or activating mutations in the gene encoding the catalytic subunit of PI3K (PIK3CA) (224). The loss of function of PTEN, caused by mutation of PTEN itself or by transcriptional regulation, has been reported in up to 50% of breast cancers (225). Since PTEN normally inhibits the activation of PI3K, PTEN loss results in constitutive activation of the PI3K/Akt pathway (145). In a preclinical model, decreased levels of PTEN resulted in increased PI3K/Akt phosphorylation and signaling and prevented trastuzumab-induced growth arrest of HER2-positive breast cancer cells (144). Furthermore, response to trastuzumab- based therapy in HER2-positive patients showing diminished PTEN expression was worse compared to those patients showing normal PTEN expression levels (144, 225, 226). One study reported that among HER2-positive breast cancer patients who were refractory to trastuzumab-based therapy, those with activated PI3K (defined as either decreased PTEN expression or PI3KCA mutations) had a significantly shorter progression-free survival (PFS) than patients that did not show PI3K activation (227). Similarly, another study conducted in a cohort of HER2-overexpressing breast cancer patients reported that PI3KCA mutation and/or PTEN loss were predictive biomarkers of poor response to trastuzumab-based therapy (228).

- 29 -

3). Signaling through alternative pathways

Signaling through other RTKs can transactivate HER2 and represents therefore “escape” mechanism to circumvent the inhibitory effect of trastuzumab. Given the evidence that trastuzumab does not effectively inhibit HER2 dimerization, it has been proposed that overexpression of other HER family members, in particular HER3 and HER1, and formation of ligand-induced HER2 heterodimers may also contribute to trastuzumab resistance (229). As mentioned in a previous section, although HER2 is unable to bind ligands directly, it is the preferred dimerization partner of other HER members (95, 230). One study reported that the growth inhibitory effects of trastuzumab on HER2-positive breast cancer cells SKBR3 and BT474 were modulated by the co-expression of HER1 and the presence of its ligands, EGF and heregulin (231). BT474 cells that acquired trastuzumab resistance showed higher expression of HER1 levels, exhibited higher levels of HER1 phosphorylation and increased HER1-HER2 heterodimer formation compared to the parental trastuzumab-sensitive cells, suggesting increased HER1-mediated activation of HER2 (232). A study conducted in a cohort of 44 HER2-positive breast cancer patients who received trastuzumab-based neoadjuvant treatment identified HER1-coexpression as a potential independent predictor of poor pathological complete response (233). It has been reported that HER1 and HER3 expression was increased in HER2-positive breast cancer cell lines that had developed trastuzumab resistance after long-term exposure to this drug (234). Some clinical studies also suggest that HER3 expression may allow the identification of patients who will develop trastuzumab resistance (226, 235). In a cohort of HER2-positive metastatic breast cancer patients treated with trastuzumab, those patients whose breast cancer specimens showed high HER3 expression experienced shorter PFS compared to patients with low HER3 expression (235). A study conducted in a cohort of 125 metastatic HER2-positive breast cancer patients receiving trastuzumab-based treatment reported that HER3 expression was associated with worse PFS (226).

Apart from the HER family members, other RTKs that promote cellular growth, including IGF- 1R, c-Met, may also mediate trastuzumab resistance. Lu and collaborators have demonstrated that stable overexpression of IGF-1R prevented the trastuzumab-mediated growth arrest in the HER2-positive breast cancer cell line SKBR3 (236). Interestingly, higher levels of IGF-1R has been reported in HER2-positive breast cancer cells lines that acquired resistance to trastuzumab comparatively to the trastuzumab-sensitive parental cell line (122, 237). Moreover, IGF-1R induced phosphorylation of HER2 in the trastuzumab-resistant breast

- 30 -

cancer cells (122, 238-240). The translational significance of IGF-1R, however, still remains to be defined. Although in two clinical studies no association was observed between IGF-1R expression and trastuzumab response (241, 242), another study reported that elevated IGF- 1R expression was associated with a lower response rate to neoadjuvant trastuzumab-based treatment in HER2-positive breast cancer patients (243). Frequently overexpressed in HER2- positive breast cancer cell lines, the c-Met receptor may contribute to trastuzumab resistance (244). Whereas c-Met activation protected HER2-positive breast cancer cell lines against trastuzumab through inhibition of p27, inhibition of c-Met by RNA interference restored the sensitivity of these cell lines to trastuzumab-mediated growth inhibition (244).

In addition, increased levels of HER family ligands, including heregulin, EGF and TGF-α (245) and decreased activity of endogenous HER2 inhibitors such as mitogen-inducible gene 6 (MIG-6) (246) have been associated with trastuzumab resistance.

4). Defects in apoptosis and cell cycle control

Alterations in the normal apoptotic machinery have been proposed as a cause of resistance to trastuzumab. Survivin is a member of the inhibitor of apoptosis (IAP) protein family, which counteracts the activity of caspases, enzymes involved in programmed cell death (247). Increased levels of survivin and Mcl-1 (an anti-apoptotic BCL-2 family member) (248) were observed in trastuzumab-resistant cells, whereas treatment with a multikinase inhibitor that reduces expression levels of both survivin and Mcl-1 inhibited the growth of these cells (249). Inhibition of survivin using siRNA led to the suppression of cellular growth of the trastuzumab- unresponsive HER2-positive breast cancer cell line JIMT-1 upon trastuzumab treatment (250).

Alterations in factors that control progression through the cell cycle also may contribute to trastuzumab resistance. In a cohort of HER2-positive breast cancer patients receiving trastuzumab-based treatment, amplification and/or overexpression of cyclin E was associated with lower response to targeted therapy (251). As expected, inhibition of Cdk2 diminished growth of trastuzumab-resistant xenografts (251). Moreover, decreased levels of the Cdk inhibitor p27 was observed in a trastuzumab-resistant breast cancer cell line (252).

- 31 -

5). Influence of the tumor microenvironment

It is well recognized that the tumor microenvironment (TME) plays a critical role during the initiation and progression of carcinogenesis (253). TME consists in a variety of disparate cell types that can be divided into three main groups: cells of hematopoietic origin (T cells, B cells, NK cells, macrophages and neutrophils), cells of mesenchymal origin (fibroblasts, adipocytes and endothelial cells) and non-cellular component (extracellular matrix, including proteins, glycoproteins and proteoglycans) (254). The interaction between cancer cells and TME is complex: on the one hand tumor cells can change the nature of TME, on the other hand TME can influence tumor growth and formation of tumor metastasis (255).

Recent studies have reported that TME may contribute to the development of trastuzumab resistance either through the hindering of trastuzumab-induced ADCC or the activation of escape pathways (256-258).

As previously mentioned, FcgRs are glycoproteins that are expressed on innate immune effector cells (leukocytes), including NK cells, neutrophils, monocytes, and macrophages (259, 260). FcgRs leads to activation of the effector cells and therefore to the destruction of the trastuzumab-labeled breast cancer cells through ADCC (148). In mice lacking FcgRIII and, thus, deficient in NK cells and macrophages capable of binding the Fc region of trastuzumab, the therapeutic effect of trastuzumab was markedly reduced (261). In line with this observation, polymorphisms in the genes encoding FcgRIIa and FcgRIIa in breast cancer patients was associated with response to trastuzumab, where patients with VV and HH genotypes had better survival and response rates compared to patients with other genotypes (262). Another study reported that complete or partial remission of patients treated with neoadjuvant trastuzumab was associated with tumor infiltration of immune cells and higher in vitro ADCC activity in lysis assays (263).

It has been shown that adipocytes and preadipocytes blocked the trastuzumab-induced ADCC in HER2-positive breast cancer cells via the secretion of soluble factors (256). It has been demonstrated that differentiated adipocytes reduced trastuzumab-induced growth inhibition of HER2-positive breast cancer cells through the production of adipocytokines that stimulate the phosphorylation of Akt and therefore escape the HER2 signaling blockade (257). In an in vitro model of HER2-positive breast cancer cells co-cultured with conditional medium obtained from cancer-associated fibroblasts (CAFs) it has been demonstrated that CAFs induced

- 32 -

trastuzumab resistance through the activation, mainly via IL6, of multiple pathways, including NF-kB, JAK/STAT3 and PI3K/Akt (258).

Figure 1.6. Proposed mechanisms of trastuzumab resistance

c-Met)

Proposed mechanisms of trastuzumab resistance include: A. proteolytic generation of p95HER2, a constitutively kinase-active HER2 isoform lacking the trastuzumab-binding site; B. Physical blockade of binding of trastuzumab to HER2 by CD44/hyaluronan polymer complex or MUC4; C. Activation of downstream effectors or cross-talk pathways; D. Compensatory signaling by other cell surface receptors including EGFR/HER3 and other receptor tyrosine kinases.

Adapted from Primary trastuzumab resistance: new tricks for an old drug (264).

- 33 -

1.5.2. Genetic and epigenetic factors and survival of HER2-positive breast cancer patients Evidence suggests that genetic (i.e., single nucleotide polymorphism) and epigenetic factors (i.e., DNA methylation) could be associated with survival in breast cancer patients (see 1.5.2.1. and 1.5.3.10., respectively). Studies that analyzed the association between these factors and the response to trastuzumab in HER2-positive breast cancer patients are either few (single nucleotide polymorphism) or absent (DNA methylation).

1.5.2.1. Single nucleotide polymorphisms (SNPs) Single nucleotide polymorphisms (SNPs) are variations at a single position in a DNA sequence at a frequency of at least 1% among individuals in a population (265). This DNA variation leads to the substitution of one single nucleotide for another. Since non-synonymous SNPs can change the encoded amino acids, they have a potential impact on the protein function. It has been proposed that SNPs may influence promoter activity (and therefore gene expression), messenger RNA (mRNA) conformation and stability, and translational efficiency (266). They may therefore contribute to susceptibility to common diseases, including malignant transformation (267).

Several SNPs that are associated with breast cancer prognostic factors and survival of breast cancer patients have been reported in the literature (268-271); for a systematic review see (272). These SNPs are located on several genes including VEGF, RAD51B and Cyclin D1 (CCND1) (for the complete list, see (271). These genes are involved in cancer-related pathways such as tumor progression, DNA repair and cell cycle control among others (271).

1.5.2.2. HER2 SNPs Through the development of deep sequencing technologies, remarkable advances have been made in the identification of single polymorphic variants in HER2 gene (273). Recent studies revealed at least 306 polymorphic forms in this gene (274). Although the HER2 gene contains several SNPs, to date the most investigated variants are the Ile655Val (rs1136201) and the Ala1170Pro (rs1058808) polymorphisms (275, 276). The Ile655Val polymorphism, located in the transmembrane coding region at codon 655, results in the substitution of Isoleucine (Ile: ATC) with Valine (Val: GTC) (276). Results from preclinical studies suggest that Ile655Val SNP may alter the function of the HER2 gene. Computational models have indicated that the Ile to Val substitution at codon 655 promotes the formation of active HER2 dimers, resulting in enhanced activity of the tyrosine kinase domain (17). Moreover, an in vitro experiment has shown that HER2/Val-variant expressing cells exhibit a higher growth capacity and a lower apoptosis rate compared to HER2/Ile-variant expressing cells. Furthermore, only HER2/Val-

- 34 -

variant expressing cells developed tumors in nude mice. Cells carrying the HER2/Val-variant were more sensitive to trastuzumab treatment than those carrying the HER2/Ile-variant, as shown by the decrease in HER2 phosphorylation (16). The Ala1170Pro polymorphism, which has been identified in the coding region for the carboxyl-terminal regulatory domain a codon 1170, results in the substitution of Alanine (Ala:GCC) with Proline (Pro: CCC) (277). Its biological relevance, however, remains undefined (277).

The role of both HER2 SNPs as breast cancer prognostic factors is still under investigation. While some studies observed an association between Ile655Val and Ala1170Pro and prognostic factors (277-284), other studies did not (285, 286). Therefore, further studies that analyze these associations are needed.

One study has reported that a HER2 SNP (rs1810132) may be associated with increased breast cancer specific mortality, especially among women with low Native American ancestry (287).

1.5.2.3. HER2 SNPs and response to trastuzumab To date, only two studies have investigated the association between HER2 Ile655Val SNP [determined by polymerase chain reaction-restriction fragment length polymorphism (PCR- RFLP) assay] and the response to trastuzumab (16, 288). In the study conducted by Beauclair et al., they did not observe an association between HER2 genotype and response to trastuzumab in a cohort of 61 patients with advanced HER2-positive breast cancer (16). However, Han and collaborators found that Ile655Val SNP was associated with disease-free survival (DFS) among 212 HER2-positive breast cancer patients who received adjuvant trastuzumab treatment. In univariate analysis (multivariate analysis was not performed), Han et al. observed that patients with the Ile/Val or the Val/Val genotypes had a significantly better 5-year DFS compared to those with the Ile/Ile genotype (p=0.008) (288). Considering these contrasting results, further studies that evaluate this association are warranted.

1.5.3. Epigenetics The term “epigenetics” was coined by Conrad Waddington in 1942 to describe a branch of biology that studies “the causal interactions between genes and their products, which bring the phenotype into being” (289). He introduced this new branch following the observation that characteristics acquired in response to an environmental stimulus were inherited within a population. In his famous drawing of the “epigenetic landscape” (Figure 1.7.), he represented the developmental process as a series of “decisions” that the differentiating cell, depicted as a

- 35 -

ball rolling down valleys on an inclined surface, has to take during development, as a visual metaphor for the branching pathways of cell fate determination (290). Waddington's definition initially referred to the role of epigenetics in embryonic development. The definition of epigenetics, however, has broadened over time as it has become evident that it is implicated in a variety of physiological and pathophysiological processes (289). Currently, the term “epigenetics” refers to the study of heritable alterations in gene expression that occur independently of changes in the DNA sequence (291). Although the cells in an organism contain essentially the same genetic information, cell types and functions differ because of qualitative and quantitative differences in their gene expression (292). Not surprisingly, control of gene expression by epigenetic mechanisms plays a central role during cellular differentiation and development: most of these heritable alterations are established during differentiation and are stably maintained through cell division (293). Importantly, epigenetic modifications play also a role in several pathologies (294).

Figure 1.7. Waddington’s classical epigenetic landscape.

Conrad Waddington proposed the concept of cellular decision-making during development. At various points in this dynamic visual metaphor, the cell (represented by a ball) can take specific permitted trajectories, leading to different outcome or cell fates.

Source: Epigenetics: a landscape takes shape (289).

- 36 -

Epigenetic alterations include DNA methylation, covalent histone modifications, chromatin remodeling, and gene regulation by microRNAs (293). These modifications work synergistically to regulate the functioning of the genome and gene expression by altering the local structural dynamics of chromatin, mainly regulating its accessibility and compactness. The interplay of these modifications regulates the way the mammalian genome manifests itself in different cell types, developmental stages and disease states, including cancer (295).

1.5.3.1. DNA methylation DNA methylation is one of the most extensively studied epigenetic modifications in humans (296). It is an essential epigenetic mark that contributes to transcriptional regulation (297). In mammals, DNA methylation occurs mainly through the covalent attachment of a methyl group

(-CH3) on cytosine residues in CpG dinucleotides (298). A CpG dinucleotide is composed of a base cytosine (C) linked by a phosphate bond to the base guanine (G) in the DNA sequence nucleotide (299).

Cells have the ability to both methylate and demethylate DNA. DNA methylation is mediated by DNA methyltransferases (DNMTs), a family of enzymes that catalyze the transfer of the methyl group from the methyl donor S-adenosyl-methionine (SAM) to the fifth carbon of the cytosine ring within CpG dinucleotides at a CpG site, resulting in the formation of 5- methylcytosine (5mC) (Figure 1.8.) (300).

- 37 -

Figure 1.8. Conversion of cytosine to 5-methylcytosine by DNA methyltransferase (DNMT).

DNMT catalyses the transfer of a methyl group (CH3) from S-adenosylmethionine (SAM) to the 5-carbon position of cytosine.

Source: Epigenetics and gene expression (292).

DNMT1 is considered responsible for maintenance methylation (301), namely the copying of existing DNA methylation patterns of the parent strand onto the newly synthesized strand during DNA replication, as it preferentially methylates hemimethylated DNA in vitro (302) and it is maximally expressed during the S cell cycle phase (303). DNMT3A and DNMT3B have a preference for unmethylated CpG dinucleotides. They are responsible for de novo methylation by introducing new methyl group at CpG sites in which neither strand was previously methylated (304). DNMT3A and DNMT3B also participate with DNMT1 to ensure propagation of methylation pattern during DNA replication (305). DNA demethylation occurs either passively, when methylated CpGs fail to be reproduced during replication, or actively, through enzymatic removal of the methyl group from 5mC (292). The removal of 5mC through the formation of 5-hydroxymethylcytosine (5hmC) via the action of ten-eleven translocation (TET) proteins followed by subsequent deamination and the replacement of cytosine through the base excision repair (BER) has been proposed as a mechanism for active demethylation (306). DNA demethylation establishes a transcriptionally permissive state (307).

Consistent with their functional relevance, DNA methylation patterns are non-random, well- regulated and tissue-specific (308-310). DNA methylation patterns are established during early development through a highly controlled process that involves demethylation and de novo methylation (295, 311, 312). During pre-implantation growth, both paternal and maternal

- 38 -

genomes undergo a wave of demethylation, which erases most of the methylation patterns inherited from the gametes. Shortly after implantation, the embryo undergoes a wave of de novo methylation (313). De novo methylation also occurs during gametogenesis in both male and female germ cells and it plays an essential role in the establishment of genomic imprinting, an epigenetic process that results in differential modification of paternal and maternal alleles during gametogenesis and monoallelic expression of a small set of genes, known as imprinted genes, in the offspring (295, 312, 314). De novo methylation activity is present mainly in embryonic stem cells and embryonal carcinoma cells, early post-implantation embryos, and developing germ cells, whereas it is largely inhibited in differentiated somatic cells (313).

1.5.3.2. Distribution of CpG dinucleotides and CpG methylation in the human genome Intriguingly, it has been reported that the prevalence of CpG dinucleotides in human genome is much lower than expected based on the CG content (315). Indeed, CpG dinucleotide density is generally low in the bulk of the genome. CpGs dinucleotides are concentrated in specific regions known as CpG islands (CGIs), clusters of CpG dinucleotides in CG-rich regions approximately 0.5-2.0 kb in length (316). CGIs are found in approximately 50-70% of gene promoters, suggesting they may play a role in the regulation of gene expression (317). Most CGIs are located in the promoter of most housekeeping genes and in a proportion of tissue- specific genes and developmental regulatory genes. Although many CGIs are located in gene promoter regions, half of CGIS are also present within gene bodies (intragenic location) or between genes (intergenic location) (318). CpG dinucleotides are also located within “CGI shores”, regions located 2 kb from CGIs and within “GCI shelves”, regions located 2-4 kb from CGIs. “Open sea regions” are isolated GpG sites in the genome in regions without any enrichment of CpG content (319).

In healthy somatic cells, 70% of CpG dinucleotides are methylated (320). DNA methylation, however, is not distributed evenly throughout the genome. The vast majority of CpG dinucleotides at CpG poor regions dispersed throughout the genome are methylated. On the contrary, the majority of CGIs are unmethylated, allowing access to transcription factors and chromatin-associated protein for the expression of most housekeeping genes and other frequently expressed genes (321).

This difference in the distribution of methylated CpG dinucleotides throughout the genome could explain why the observed prevalence of CpG dinucleotides in the genome is lower than expected. This CpG dinucleotides shortage is mainly attributed to the hypermutability of

- 39 -

methylated CpGs to TpGs (322, 323). In contrast, CpGs within CGIs are often unmethylated (and therefore protected from mutation) and their frequencies are close to random expectation (324).

As mentioned before, DNA methylation is erased during early embryonic development and then re-established at the time of implantation (325, 326). Bimodal pattern of methylation (CGIs differ from bulk chromosomal DNA by their frequent lack of cytosine methylation) is achieved through two mechanisms: a wave of generalized de novo methylation and a mechanism that guarantees that CGIs remain unmethylated (327). The precise details of how CGIs are protected from methylation are not completely understood. It has been proposed that protection might be directed through the recognition of cis-acting regulatory elements located in CGIS (328-330) and mediated by an island-specific active demethylation (331).

DNA methylation is crucial for normal development and it plays an important role in various biological processes such as genomic imprinting (allele-specific expression of some genes) (332), X-chromosome inactivation (333), transposable elements silencing (334), stem cell differentiation (335), and embryonic development (336).

1.5.3.3. Mechanisms of control of gene expression through DNA methylation In general, GCIs methylation at promoter regions is associated with transcriptional silencing. How DNA methylation in CpG-rich promoter regions contributes to the repression of gene expression remains unknown. Several hypotheses have been proposed. It has been reported that methylation of CpGs can directly interfere with the binding of transcriptional factors to their target sites, therefore impeding transcription. Alternatively, it has also been proposed that methylation of CpGs attracts a family of methyl-CpG-binding proteins (MBPs) that specifically bind to methylated CpGs via their methyl-CpG binding domain (MBD). MBPs proteins, in turn, can attract repressors and histone deacetylases (HDACs) that lead to chromatin compaction and gene silencing (Figure 1.9.) (317).

- 40 -

Figure 1.9. Proposed mechanisms of silencing of CpG islands (CGI) promoters by DNA methylation.

MBPs

Methylated CpG Transcription Histone deacetylase factor Methyl-CpG-binding proteins

DNA methylation is associated with silencing of CGI promoters. This may be due to direct inhibition of transcription factor binding to their target sites by DNA methylation or may be mediated by methyl-CpG-binding proteins (MBPs), which in turn attract histone deacetylases (HDACs) leading to gene silencing.

Adapted from CpG islands and the regulation of transcription (317).

However, DNA methylation does not exclusively correlate with transcriptional repression. Recent genome-wide studies, in fact, have demonstrated that the location of methylated sites within the genome influences its relationship with gene control (307). While methylation in the immediate vicinity of the transcriptional start site (TSS) generally represses gene expression, methylation in the gene body (far from annotated TSS) may stimulate elongation and is therefore positively associated with gene expression (337, 338). Thus, it has been proposed that it is the initiation of transcription but not transcription elongation that is repressed by DNA methylation (307).

Whereas DNA methylation during normal development is associated with controlled gene silencing in a tissue-specific manner, aberrant pattern of DNA methylation have been reported for several diseases, including schizophrenia (339, 340), hypertension (341), diabetes mellitus 2 (342), and obesity (343). Abnormal methylation patterns have also been observed in cancer cells and are thought to play a role in the induction of abnormal gene expression in malignant cells (317, 344-346).

- 41 -

1.5.3.4. Aberrant DNA methylation in cancer It is widely accepted that tumorigenesis is a multistep process involving a series of genetic and epigenetic alterations (298, 321, 346, 347). Epigenetic hallmarks of cancer include wide areas of global hypomethylation along the genome, and localized areas of hypermethylation at CGIs within the gene promoter regions (348, 349).

In normal cells, methylation of repetitive regions like centromeres and satellite sequences [long interspersed nuclear elements (LINE), short interspersed nuclear elements (SINE) and intracisternal A-particles (IAP)] is important to ensure genomic stability (350). In many tumors loss of DNA methylation at these normally inactivated regions has been reported. As a consequence, these transposable elements are aberrantly reactivated and can integrate at random sites in the genome, favouring mutagenesis and genomic instability (351).

DNA hypermethylation in the promoter CGIs of tumor suppressor genes is thought to play a central role in carcinogenesis (318, 352). In tumor cells, CGIs hypermethylation has been identified as a mechanism of inactivation of several tumor suppressor genes, including p16INK4a, Rb, BRCA1, VHL and hMLH1 (353). The transcriptional silencing of tumor suppressor genes promotes cell proliferation, thereby providing a strong selective advantage for cancer cells (354).

DNA methylation changes occurring in CGI shores are gaining increasing attention. Interestingly, it has been reported that CpG sites located in CGI shores, in addition to those located within CGIs, display differential DNA methylation patterns between different tissues as well as between cancer tissues and their normal counterparts (355-357). Genome-wide methylation studies have reported that most of the differentially methylated regions between normal colon tissue and colon cancer tissue were located in CGI shores rather than in CGIs (355, 356).

The events leading to initiation of abnormal methylation patterns during tumorigenesis are not fully understood (293). As mentioned above, de novo methylation events occur during early development (358), suggesting that de novo methylation is particularly active at these stages. Recent reports, however, have reported that de novo methylation can also occur in adult somatic cells (359). Progressive methylation of a significant proportion of all CGIs has been observed in certain tissues during aging (360) or in cancer cells (361).

- 42 -

In addition, studies conducted on monozygotic twins suggest that external factors including lifestyle, diet and environmental exposure can affect methylation. While the genome of young twins is epigenetically similar, the methylation patterns of aged ones are clearly different (362). Of note, changes are particularly evident in monozygotic twins who had spent a long period of their lives apart, suggesting that environmental factors may influence methylation patterns (363, 364).

Moreover, recent findings suggest that genetic and epigenetic alterations cooperate during tumorigenesis (365). On the one hand, promoter hypermethylation in genes that encode for proteins that are responsible for DNA repair or for controlling cell cycle (for example O6- methylguanine-DNA methyltransferase (MGMT) and RASSf1A) can lead to accumulation of DNA damages and subsequent mutations in oncogenes and tumor suppressor genes (353). Conversely, genetic mutations in epigenetic regulators can lead to changes in the epigenome, including aberrant DNA methylation and histone modifications, which lead to abnormal gene expression and genomic instability (365). New evidence suggests that abnormal de novo methylation observed in cancer cells may be linked to histone modifications (366).

1.5.3.5. Link between DNA methylation and histone modifications Recent studies suggest that DNA methylation and histone modification are deeply intertwined. Both DNA methylation and histone modifications play critical roles in the regulation of gene expression during development in mammals (366).

The nucleosome is the fundamental unit of eukaryotic chromatin. Each nucleosome is composed of DNA wrapped around an octamer of histone proteins. These octamers consist of double units of H2A, H2B, H3, and H4 core histone proteins (367).

Each core histone protein has an N-terminal tail that extend from the core octamer between the coils of DNA (368). N-terminal tails are subject to a wide range of modifications, including acetylation, methylation, phosphorylation, and ubiquitylation (367). These modifications are reversible and are regulated by groups of enzymes, including histone acetyltransferases (HATs), histone deacetylases (HDACs), histone methyltransferases (HMTs), and histone demethylases (HDMTs) (367). These chemical modifications affect the access of regulatory factors and complexes to chromatin and have therefore an influence on gene expression. Tail modifications can modify the ability of non-histone proteins to bind to chromatin (369). Some tail modifications, like acetylation and phosphorylation, can also alter the charge of the tail and therefore the binding strength between the histone and the negatively charged DNA. This has

- 43 -

an impact on the condensation of chromatin (and influence therefore the exposure of DNA to regulatory proteins) (369). In general, acetylation of histone proteins (catalyzed by HATs) promotes a more relaxed chromatin structure and therefore gene expression, while deacetylation of histone proteins (catalyzed by HDACs) closes the chromatin (370). The impact of these modifications depends on the type and location of these modifications (366). For example, while tri-methylation of lysine 27 on histone H3 (H3K27m3), which is catalyzed by the Polycomb repressive complex 2 (Prc2) (371), is associated with gene repression, tri- methylation at lysine 4 on histone 3 (H3K4me3) is associated with gene activation (372).

It has been proposed that different histone modifications, located on one or more histone tails, form a “histone code” that is read by other proteins and that influence downstream events and therefore specify patterns of gene expression (373).

Current evidence indicates a biological relationship between DNA methylation and histone modifications (366). On the one hand, histone modifications, especially histone methylation, are essential for proper DNA methylation patterns (374). On the other hand, DNA methylation can influence histone modifications to produce changes in chromatin configuration (375).

Aberrant histone modifications in cancer may lead to tumorigenesis, as it deregulates the control of chromatin-based processes (376). Growing evidence suggests that aberrant histone modifications can contribute to cancer development via two mechanisms: through the alteration the gene expression, including abnormal regulation of oncogenes and tumor suppressor genes, or by affecting genome integrity and chromosome segregation (377).

1.5.3.6. Link between DNA methylation and miRNAs MicroRNAs (miRNAs) are approximately 22 nucleotides long non-coding RNAs which regulate gene expression in a sequence-specific manner. miRNAs are firstly transcribed by RNA polymerase II (Pol II) into primary miRNAs (pri-miRNAs). These long primary transcripts are then further processed within the nucleus to precursor miRNA (pre-miRNAs) by DROSHA, an RNAse III enzyme. Pre-miRNAs are then transported into the cytoplasm by Exportin-5. In the cytoplasm, pre-miRNAS are further processed to mature miRNAS by Dicer (378). miRNAs negatively regulate the expression of their target mRNAs through two mechanisms, depending on the extent of complementarity between the miRNA and the target. miRNAs that bind with nearly perfect complementarity to protein-coding mRNA sequences leads to degradation of the target. miRNAs that bind to their target with incomplete complementarity,

- 44 -

often in the 3′ UTR regions, leads to the translational suppression of the target gene (379). It has been estimated that more than 60% of human protein coding genes contain miRNA binding sites within the UTR regions (380). miRNAs exhibit a tissue-specific and developmental stage-specific pattern of expression. As such, miRNAs expression plays an important role in diverse biological processes, including cell proliferation, differentiation, and apoptosis (381).

Given the importance of miRNAs in the regulation of gene expression, it is not surprising that dysregulation of miRNAs expression contributes to the initiation and progression of cancer (382). Indeed, abnormal miRNA expression has been observed in several human cancers, including breast cancer (383).

Growing evidence suggest that miRNAs may function as a new class of oncogenes or tumor suppressor genes (TSGs). Those miRNAs whose expression is enhanced in cancerous tissues compared to normal tissues are considered oncogenes. These miRNAs are known as oncomirs. They promote tumor development by inhibiting expression of TSGs and genes that control cellular differentiation and apoptosis. Conversely, those miRNAs whose expression is decreased in cancerous tissues compared to normal tissues are regarded as TSGs. Tumor suppressor miRNAs prevent tumor development through the inhibition of expression of oncogenes and genes that control cellular differentiation and apoptosis (384).

Upregulated miRNAs in breast cancer are miR-21 (which negatively regulates the expression of PTEN), miR-155 (which targets caspase-3), miR-10b (which targets the homeobox D10 tumor suppressor signaling pathway) and miR-9 (which negatively regulates the expression of E-cadherin. Downregulated miRNAs in breast cancer are let-7 family (target gene: H-ras), miR- 145 (target gene: IRS-1), miR-205 (target gene: ZEB1 and ZEB2) (383).

Similar to abnormalities in oncogenes and TSGs, alterations in miRNAs expression can be caused by chromosomal deletion, amplification, mutation, epigenetic silencing and transcriptional dysregulation of pri-miRNA transcripts (385).

Recent studies suggest a link between miRNAs and other epigenetic mechanisms. On the one hand, it has been proposed that aberrant epigenetic regulation affecting DNA methylation and histone modification might be responsible for abnormal miRNA expression in malignancies (385, 386). On the other hand, it has been postulated that a specific group of miRNAs (the so-

- 45 -

called epi-miRNAs) can directly modulate the expression of epigenetic regulators, including DNMTs and HDACs, and therefore indirectly affect the expression of tumor suppressor genes (387).

1.5.3.7. Methods for DNA methylation analysis Several methods for DNA methylation analysis have been developed over the years (for a review see (388-392). The choice of method for DNA methylation analysis depends on several factors, including the aim of the study, the amount and quality of DNA samples, tissue source (i.e., FFPE tissues, fresh frozen tissues, cell lines), the availability of bio-informatics software for analysis and interpretation of data, and the costs (393).

Methods for DNA methylation analyses can be divided into two categories: methods for the global DNA methylation analysis and those for the gene-specific methylation analysis (394). Global genomic DNA methylation analysis provides a global picture of DNA methylation changes within the genome. It allows the quantitative determination of total level of DNA methylation in the genome, but it does not inform on the regional changes in DNA methylation. High performance liquid chromatography (HPLC) is the current “gold standard” assay for quantifying global DNA methylation. Although it is a very quantitative and reproducible technique, it requires relatively large amount of DNA (3-10 µg) (393, 394). Gene-specific methylation analysis identifies specific genes and regulatory regions that are differentially methylated (393). Gene-specific methylation analysis can be subdivided into “candidate gene” and “genome-wide” approaches (394). Candidate gene approach is suitable for studies focusing on a relatively small number of genes, while genome-wide approach is more appropriate for large-scale screening, especially when knowledge about the specific research question are limited (395).

Analysis of DNA methylation on a genome-wide scale can be assessed using either next- generation sequencing (NGS)-based methods or microarrays (393). In both approaches, DNA samples are first treated with sodium bisulfite, which converts unmethylated cytosines to uracil, while methylated cytosines remain unaffected. NGS-based methods, including whole-genome bisulfite sequencing (396) and pyrosequencing (397), allow comprehensive coverage of DNA methylation sites. Microarrays interrogate DNA methylation using thousands of oligonucleotide probes. Each of these probes targets a specific genomic location (398). While the genomic coverage of microarray-based platform is limited by both the number and specificity of probes, NGS-based technologies allow an exhaustive screening of the genome for epigenetically

- 46 -

altered loci (399). However, since sequencing-based DNA methylation assays are still rather expensive, methylation arrays are a widely used alternative for high-throughput DNA methylation analyses (400). Here, the genome-wide approach using DNA methylation microarrays, with particular focus on the Illumina 450K methylation array, will be described in details.

Bisulfite conversion

Current DNA sequencing approaches are not able to distinguish between methylated and unmethylated cytosine. To overcome this limitation, genomic DNA must be treated with sodium bisulfite prior to methylation analysis. Bisulfite treatment selectively deaminates unmethylated cytosine but not methylated cytosine to uracil, being methylated cytosine resistant to this conversion (401). PCR amplification of bisulfite-converted DNA results in the replacement of uracil with thymine (402). Bisulfite treatment causes therefore a primary sequence change in the DNA that allows the differentiation between methylated and unmethylated cytosine, since it induces C to T polymorphism at each unmethylated CpG site (394). Several commercially available kits, including EpiTect® Bisulfite kit (Qiagen) and EZ DNA MethylationTM kit (Zymo), provide fast and efficient conversion with minimal DNA loss (395).

DNA methylation arrays

Like NGS-based methods, DNA methylation arrays detect methylation status at single base resolution (403). The Infinium HumanMethylation450 BeadChip (Infinium Methylation 450K; Illumina, Inc., CA, USA) is currently the most widely used genome-wide technology to interrogate methylation in large-scale epigenome-wide association studies (EWAS) (404).

The Infinium Methylation 450K array analyses the methylation status of 485,577 CpGs in the human genome. It provides genome-wide coverage of 99 % RefSeq genes and 96 % CGIs. Probes on this bead array are distributed across all gene regions (promoter region, 5′ UTR, first exon, gene body and 3′ UTR), and all CGI regions, including shores and shelves (405).

Following treatment with sodium bisulfite, quality of bisulfite-converted DNA is assessed by methylation-specific PCR (MSP). Bisulfite-treated genomic DNA is then subjected to a whole genome amplification (WGA) step. The products are then enzymatically fragmented and hybridized to the probes attached to a solid microarray support. Base extension chemistry is then employed to incorporate a fluorescent labeled nucleotide (dideoxynucleotide

- 47 -

triphosphates, ddNTP) for detection onto each hybridized probe (398, 406). Each CpG locus on the array is targeted by a specific probe sequence attached to a bar-coded bead, with a median number of beads per locus being 14, which are randomly distributed across the array (407).

The Infinium Methylation 450K array is an extension of the previous version, the Infiunium HumanMethylation27 (Infinium HumanMethylation27K; Illumina, Inc., CA, USA), which interrogated 27,578 CpG sites associated with 14,495 protein-coding gene promoters (408). The general principle of the Infinium Methylation technique (both 27K and 450K arrays) is to assess cytosine methylation through quantitative genotyping of the C to T polymorphism generated at the CpG site by the bisulfite conversion (408). Whereas the 27K array uses only one type of assay chemistry (type I probe), the 450K array uses a combination of two distinct assay chemistries (type I and type II probes). The “hybrid” two-assay (type I and type II probes) design of the 450K array allows coverage of many more cytosines than the previous version.

The Infinium I and II probes are both 50 bases long but detect methylation levels through different mechanisms (Figure 1.10.). For the Infinium I technology, each CpG site is interrogated using two bead types, one for the methylated (M) and the other for the unmethylated (U) sequence. Basically, the two probes in each pair differ at the end-nucleotide that matches to the cytosine position of a CpG. The end-nucleotide of the probe can be either a guanine complementing the cytosine of the methylated CpG or an adenine complementing the thymine resulting from the bisulfite conversion of the unmethylated cytosine. The fluorescent-labeled single base extension generates the array signals. As both bead types (M and U) incorporate the same labeled nucleotide for the same target CpG, they produce the same color fluorescence. The β-value is the ratio of the methylated probe intensity and the overall intensity (sum of the methylated and unmethylated probe intensity) and it ranges between 0 (unmethylated) and 1 (methylated) (409). For the Infinium I assay, the β-value represents the ratio of the intensities from the two different probes in the same color [β value=M/(U+M)].

The Infinium II technology uses only one bead type for each locus and red (M) and green (U) fluorescent-labeled single base extension occurs differentially at thymine and cytosine of bisulfite converted DNA. For the Infinium II assay, the β-value represents the ratio of the intensity of the signal from the green channel (M) to the combined intensity obtained from red and green channels from the one position [β value=Green(M)/(Red(U)+Green (M))] (410).

- 48 -

Figure 1.10. Overview of the Infinium I and Infinium II assays

A. Infinium I assay: each CpG is interrogated using two bead types: methylated (M) and unmethylated (U). Both bead types will incorporate the same labeled nucleotide for the same target CpG, thereby producing the same color fluorescence; B. Infinium II assay: each target CpG is interrogated using a single bead type. Each locus is detected in two colors.

Source: Evaluation of the Infinium methylation 450K technology (411).

- 49 -

The greatest limitation of microarrays is that since the WGA reaction is compromised when DNA material is degraded (<1 kb) (412), high-quality DNA (for example DNA extracted from fresh-frozen tissues, blood samples or in vitro cultured cells) is required. DNA samples characterized by highly fragmented DNA, such as archival FFPE samples, are therefore less suitable for microarrays studies. To overcome this limitation, an FFPE restoration step characterized by DNA repair and ligation can be performed after bisulfite conversion and before the WGA step in the 450K assay (413). This additional step, therefore, allows the utilization of the 450K array for the interrogation of DNA methylation patterns in FFPE samples (413).

In 2015, Illumina has released a new methylation microarray, the MethylationEPIC (850K) array (414). Like the Infinium Methylation 450K array, the 850K array also uses a combination of the Infinium I and II assays (403). This new array contains 91.1% of the CpG sites included in the 450K array. The genomic coverage of the 850K array has been increased through the addition of 413,745 new CpG sites not present in the previous version (414). It has been shown that the 850K array can also be used for the determination of methylation status from FFPE samples (414).

1.5.3.8. Aberrant methylation in breast cancer As observed in other types of tumors, aberrant methylation has been reported to occur in breast cancer specimens or breast cancer cell lines (415-417). In addition to global DNA hypomethylation, which underlies chromosomal instability and abnormal gene expression patterns, hypermethylation of promoter regions of, for example, tumor suppressor genes has been observed in breast cancer (418). Besides the hypermethylation of tumor suppressor genes, genes involved in DNA repair, apoptosis, metabolism, cell cycle regulation, cell adherence, tissue invasion and metastasis, cellular homeostasis, cellular growth as well as genes encoding several epigenetic enzymes are frequently hypermethylated in breast cancer (416, 419).

Although it has been shown that the genome of a breast tumor is frequently hypomethylated compared to normal breast tissue (420), the number of genes reported to be hypomethylated in this type of malignancy is relatively small. This is probably associated with the fact that hypomethylated DNA is located in gene-poor regions of the genome (416).

- 50 -

1.5.3.9. DNA methylation signature in HER2-positive breast cancer Several studies have investigated aberrant methylation in HER2-positive breast cancer specimens (18, 421-427). A genome-wide analysis of DNA methylation conducted in a cohort of 17 HER2-positive breast cancer patients has shown that HER2-positive breast cancer tissues were characterized by extensive hypermethylation mostly affecting GCIs compared to normal breast tissues (423). Aberrant methylation was reported in CGIs of genes involved in multicellular development, differentiation, and transcription, in genes belonging to the PI3K and Wnt signaling pathways. The authors also reported hypomethylation within gene bodies of HER2, AKT3 and HIK1.

Interestingly, a recent study suggests that distinct breast cancer subtypes show different CpG methylation patterns on a genome-wide scale (18). In an attempt to identify DNA methylation- based signature in association with breast cancer subtypes as annotated by the PAM50 method, Stefansson and collaborators identified CpGs that were differentially methylated between normal breast and breast cancer tissues (18). In a test cohort of 40 breast tumors and 17 normal breast samples and in a validation cohort composed of 212 breast cancers tumors covering all molecular subtypes, they observed that whereas luminal B tumors were characterized by a distinctive DNA methylation pattern, with extensive DNA methylation of CGIs, HER2-enriched and luminal A tumors were more heterogeneous in term of their methylation pattern.

Using a genome-wide approach, Yamaguchi and collaborators analysed epigenetic alterations in cancer-related pathways in 24 HER2-positive breast cancer specimens (427). In 38% (9/24) of breast tumors, aberrant methylation in DKK3 and SFRP1 genes, two negative regulators of Wnt signaling, was observed. In addition, aberrant methylation in TP53 downstream genes was reported in 42% (10/24) of specimens.

In a cohort of 143 breast cancer patients, using a moderate throughput quantitative methylation analysis (MethyLight) (428), Fiegl and collaborators reported a higher prevalence of DNA methylation of three genes (PGR, HSD17B4 and CDH13) in HER2-positive breast cancer samples (422). These genes are either involved in hormonal regulation (PGR and HSD17B4) or are members of the cadherin family (CDH13).

Several studies have analyzed methylation levels using a methylation-specific PCR in breast cancer tissues. These studies have reported that compared to HER2-negative breast cancer

- 51 -

tissues, HER2-positive specimens were characterized by higher methylation levels in several genes, including RASSF1A, GSTP1, APC (425), RAR-beta2 (424), CDH1 (421).

Terada and collaborators analyzed the association between methylation levels (evaluated by RT-PCR) of CGIs promoters of 11 genes (LOC346978, 3OST2, GREM1, XT3, PCDH10, FLNc, THBD, COE2, CLDN3, F2R and AK5) and the HER2 gene amplification in a cohort of 63 breast cancer patients (HER2-positive and HER2-negative) (426). Within this cohort, 38% of patients showed HER2 gene amplification. The authors observed that HER2-amplified breast cancer tissues were more frequently methylated than HER2-negative specimens.

1.5.3.10. Genome-wide DNA methylation analysis and association with survival in breast cancer patients Some studies have highlighted the prognostic significance of DNA methylation in breast cancer patients using a genome-wide approach (429-433).

A study has shown that the presence of high degree of methylation (the CpG island methylator phenotype, or CIMP) in tumors of breast cancer patients allowed identifying a subset of patients with low metastatic risk and better overall survival in a cohort of 171 breast cancer patients (HER2-positive and HER2-negative) (429). Interestingly, CIMP loci were highly enriched for genes that make up the metastasis transcriptome, including LYN, MMP7, KLK10, and WNT6.

Another study has demonstrated that hypermethylation in gene clusters detected using methyl- CpG binding domain-based capture (MBDCap) sequencing (MBD-Seq) (434) was significantly correlated with overall survival and was positively associated with a poor survival in a cohort of 77 breast cancer patients (ER-positive and ER-negative) (432). Gene clusters represent a group of genes within the genome of an organism that encode for similar proteins which collectively share a generalized function and are often located within a few thousand base pairs of each other. In particular, metallothionein-1 (MT1) genes (MT1A, B, E, G, H, L, and X) were able to stratify patients with a statistical significance.

Using a genome-wide approach, a study conducted in a cohort of 176 patients with invasive breast cancer (HER2-positive and HER2-negative) reported that methylation level of 18 CpGs was associated with survival of patients. The methylation level of these CpGs was associated with the expression level of 26 genes, including IRF6, TBX5, CSNK1G2, MACF1, KCTD21 and EPN3 (430).

- 52 -

Following a genome-wide screen with the aim to identify genes silenced by promoter hypermethylation in breast cancer specimens, Jeschke and collaborators tested a subset of candidate hypermethylated genes for prognostic prediction in a cohort of 132 primary breast cancer patients (433). The authors reported that methylation of CD01, CKM, CRIP1, KL, and TAC1 was associated with poor overall survival of patients.

A study conducted in a cohort of 39 breast cancer patients (HER2-negative and HER2-positive) has shown that methylation of RECK, ACADL, SFRP2, ITR, UAP1L1 and UGT3A2 was associated with decreased disease-free survival (431).

1.5.3.11. Genome-wide DNA methylation analysis in breast cancer and response to chemotherapy Previous studies reported that genome-wide DNA methylation analysis allowed the identification of genes that were regulated by methylation and were also associated with chemotherapy response in breast cancer patients or breast cancer cell lines (435, 436). In a cohort of 157 breast cancer patients (33 patients in the discovery cohort and 124 patients in the validation cohort) treated with doxorubicin or combined treatment with 5-Fluorouracil and mitomycin C, Klajic et al. compared genome-wide DNA methylation pattern in breast cancer tissue before and after neoadjuvant therapy (436). They identified key cell-cycle regulatory genes that were differentially methylated before and after chemotherapy and were associated with treatment response to either doxorubicin or combined treatment with 5-Fluorouracil and mitomycin C. The authors concluded that the methylation patterns observed might represent potential predictive markers to chemotherapy sensitivity.

He et al. reported that breast cancer cell lines resistant to adriamycin or paclitaxel showed massive changes in DNA methylation compared to chemotherapy-sensitive parental cells (435). The authors observed enrichment in cell mobility and cellular surface interaction pathway.

1.5.3.12. DNA methylation in breast cancer tissues and trastuzumab response in HER2- positive breast cancer patients To date, no study has evaluated the association between DNA methylation in breast cancer tissues and trastuzumab response in HER2-positive breast cancer patients, nor using a genome-wide or a candidate-gene approach. Yet, only one in vitro study has investigated the association between epigenetic mechanisms regulating gene expression and response to trastuzumab in breast cancer cell lines (437). Ye and collaborators observed that epigenetic silencing of miR-375, a miRNA that was found to target IGF1R, was a key regulator of

- 53 -

trastuzumab responsiveness in HER2-positive breast cancer cell (437). Overexpression of miR-375 restored the sensitivity of breast cancer cells to the HER2 inhibitor.

1.5.3.13. Epigenetic therapies for treatment of breast cancer patients In the literature it has been proposed that, since epigenetic aberrations are potentially reversible, epigenetic therapy might be therapeutically relevant, as it might allow the re- expression of genes that play a role in the response to targeted therapies (438).

In the clinical settings, there are two classes of epigenetic drugs that are currently used: DNMT inhibitors and HDAC inhibitors (439). The DNMT inhibitors 5-azacytidine (azacytidine, Vidaza; Celgene) and 5-aza-2′-deoxycytidine (decitabine, Dacogen; SuperGen) have obtained FDA- approval in 2004 and 2006, respectively, for the treatment of patients with acute myeloid leukemia and with myelodysplastic syndromes (440). Both drugs used as monotherapy, however, showed limited activity against solid tumors and their use was associated with severe toxic side effects (419).

Both DNMT inhibitors are cytosine nucleoside analogs modified in position 5 of the pyrimidine ring that are incorporated into DNA (441). DNA methyltransferases methylate residues of both cytosines and cytosine analogs in the DNA. Cytidine analogs, however, prevents the resolution of the complex. As such, DNA methyltransferase remains covalently bound to DNA and its enzymatic activity is blocked (442). This hinders further methylation, and as consequence, the DNA of the daughter cells is not methylated (443).

The HDAC inhibitors suberoylanilide hydroxamic acid (SAHA, vorinostat; Merck) and romidepsin (depsipeptide, Istodax; Celgene) have obtained FDA-approval in 2006 and 2009, respectively, for the treatment of patients with progressive, persistent or recurrent cutaneous T-cell lymphoma (444). HDAC inhibitors impede histone deacetylation that accompanies the gene silencing state. More precisely, both vorinostat and romidespin inhibit HDAC activity through binding to the active site of the enzyme, thus preventing access of the substrate (445, 446). Although HDAC inhibitors are well tolerated and clinically effective against hematologic cancers, they show only poor anti-cancer activity against solid tumors when used as a monotherapy (447).

Preliminary results from Phase I and Phase II clinical trials suggest that HDAC inhibitors in combination with breast cancer therapies (chemotherapy, endocrine and anti-HER2 therapy)

- 54 -

might improve response of patients to standard therapy and/or to delay or reverse resistance to targeted therapies (344, 448).

A Phase I/II study of vorinostat combined with paclitaxel and trastuzumab reported that pathologic complete response was achieved in 54% (13/24) of locally advanced HER2-positive breast cancer patients (449). A Phase I study conducted in a cohort of 33 metastatic breast cancer patients who progressed on trastuzumab-based therapy reported promising activity of panobinostat, a pan-histone deacetylase inhibitor, combined with trastuzumab (33% of patients showed stable disease) (450).

A Phase II study of vorinostat in combination with tamoxifen conducted in a cohort of 43 metastatic breast cancer patients who had progressed on endocrine therapy (tamoxifen or aromatase inhibitors) showed encouraging results in reverting endocrine resistance (median response duration: 10.3 months) (451). Another Phase I/II study conducted in a cohort of 54 HER2-negative metastatic breast cancer patients reported benefits of combining chemotherapy (paclitaxel plus bevacizumab) and vorinostat (objective response observed in 55% of patients) (452).

An overview of the current ongoing clinical research studies that evaluate safety and efficacy of both DNMT and HDAC inhibitors in breast cancer patients has been illustrated by Falahi and collaborators (419). Results of these studies are awaiting publication.

1.5.3. Lifestyle factors and survival of HER2-positive breast cancer patients As shown in the previous sections, many studies have been conducted to understand trastuzumab resistance in HER2-positive breast cancer patients from a molecular point of view. But what is the impact of other aspects, for example lifestyle factors, on the clinical response toward trastuzumab in HER2-positive breast cancer patients?

1.5.3.1. Tobacco and alcohol consumption Several studies have highlighted that lifestyle factors such as tobacco and alcohol consumption can have an impact on survival of breast cancer patients (19-21). A meta-analysis of 9 studies (3,491 current smokers and 9,801 never smokers) that analysed the association between smoking at time of diagnosis and survival of breast cancer patients showed that breast cancer-specific mortality of breast cancer patients who smoked at time of diagnosis was 33% increased compared to never smokers (Hazard ratio (HR): 1.33; 95% CI: 1.12 to 1.58) (19). This observation has been confirmed in a recent study that analyzed the association

- 55 -

between breast-cancer specific survival in breast cancer patients and tobacco exposure before and after breast cancer diagnosis (21). In a cohort of 20,691 women diagnosed with early breast cancer between 1998 and 2008, Passarelli and collaborators observed that breast- cancer specific mortality was 25% increased in breast cancer patients that were active smokers one year before breast cancer diagnosis compared to never smokers (HR: 1.25; 95% IC: 1.13 to 1.37). In a subset of 4,562 patients for whom information about tobacco consumption before and after diagnosis was available, the authors reported that breast-cancer specific mortality was also increased (72%) in patients who continued to smoke after diagnosis compared to never smokers (HR: 1.72; 95% IC: 1.13 to 2.60). Furthermore, although not statistically significant, breast cancer-specific mortality decreased (33%) in patients who quit smoking after breast cancer diagnosis compared to those who continued to smoke (HR: 0.67; 95% IC: 0.38 to 1.19) (21). A study conducted in a cohort of 9,329 breast cancer patients diagnosed between 1990 and 2006 reported that alcohol consumption after breast cancer diagnosis did not increase risk of recurrence in breast cancer patients (HR: 1.03; 95% CI: 0.86 to 1.24) (20). Among postmenopausal women, however, the authors observed an association between alcohol consumption after breast cancer diagnosis and increased risk of recurrence (HR: 1.19; 95% CI: 1.01 to 1.40).

1.5.3.2. Molecular and epidemiological ink between tobacco and alcohol exposure and HER2 The existence of molecular and epidemiological links between tobacco and ethanol exposure and HER2 has been reported in the literature. It has been shown that 4-(methylnitrosamino)- 1-3-(3-pyridyl)-1-butanon (NNK), a potent tobacco-specific carcinogen (24), activates the ERK/MAPK signaling pathway in human normal mammary epithelial cells (26). Furthermore, a study that analyzed the association between tobacco consumption at time of breast cancer diagnosis and risk of recurrence in a cohort of 3,340 breast cancer patients diagnosed between 2001 and 2005 showed that recurrence risk was significantly increased in trastuzumab-naïve HER2-positive breast cancer patients (n=177) who smoked at time of diagnosis (HR: 3.64; 95% CI: 1.22 to 10.80) (25). A recent study that analysed the association between alcohol consumption before breast cancer diagnosis and risk of breast cancer by molecular subtype in a cohort of 105,972 women followed from 1980 to 2006 has highlighted that alcohol consumption might represent a risk factor for HER2-positive breast cancer (n=160 cases; HR: 1.16; 95% CI: 1.02 to 1.33) (22). Furthermore, it has been reported that the stimulatory effect of ethanol on the invasion capacity of breast cancer cell lines depends on the expression levels of HER2 (23).

- 56 -

Studies who specifically analyze the association between tobacco and alcohol consumption and survival in HER2-positive breast cancer patients treated with trastuzumab are warranted. To date only one study has specifically investigated the association between tobacco use and response to trastuzumab (453). In their study conducted in a cohort of 248 metastatic HER2- positive breast cancer patients treated with trastuzumab, diagnosed between 2004 and 2007 and for whom information about smoking habit was available, Santini and collaborators reported that response rate (RR) in smokers (formers and active) was not statistically different from that of never smokers (50.0% vs. 50.6%, p value= 0.736) (453). Yet, no study has examined the impact of alcohol consumption on response to trastuzumab in HER2-positive breast cancer patients.

- 57 -

Chapter 2: Advantages and disadvantages of technologies for HER2 testing in breast cancer specimens

American Journal of Clinical Pathology 2015; 144(5): 686-703

Daniela Furrer1,2,3, François Sanschagrin1,2,4,5, Simon Jacob1,2,4,5, Caroline Diorio1,2,3,4

1Centre de Recherche sur la cancer de l’Université Laval, Quebec City, QC G1V 0A6, Canada; 2Axe Oncologie, Centre de Recherche du Centre Hospitalier Universitaire de Québec, Hôpital du Saint-Sacrement, 1050 chemin Ste-Foy, Quebec City, QC G1S 4L8, Canada; 3Département de médecine sociale et préventive, Faculté de Médecine, Quebec City, QC G1V 0A6, Canada; 4Centre des Maladies du Sein Deschênes-Fabia, Hôpital du St-Sacrement, 1050 chemin Ste- Foy, Quebec City, QC G1S 4L8, Canada; 5Département de biologie moléculaire, de biochimie médicale et de pathologie, Faculté de Médecine, Quebec City, QC G1V 0A6, Canada

Corresponding author: Caroline Diorio, Axe Oncologie, Centre de Recherche du Centre Hospitalier Universitaire de Québec, Hôpital du Saint-Sacrement, 1050 chemin Ste-Foy, Quebec City, QC G1S 4L8, Canada

Keywords: Breast cancer; HER2 status; Trastuzumab; Lapatinib; HER2 inhibitors

- 58 -

Résumé Le facteur 2 de croissance épidermique humain (HER2) est un marqueur prédictif du cancer du sein. L’évaluation fiable du statut HER2 est essentielle afin de déterminer l’éligibilité des patientes atteintes d’un cancer du sein au traitement anti-HER2. Plusieurs méthodes existent pour l’évaluation d’HER2.

Dans cette revue, les principaux avantages et inconvénients des techniques qui ont été développées pour évaluer HER2 dans les spécimens de cancer du sein seront discutées.

Puisque chaque technique a ses propres avantages et inconvénients, à date aucun consensus n’a été atteint concernant la meilleure technique d’évaluation.

L’accent doit être mis sur la standardisation des procédures, l’exécution d’un contrôle de qualité et l’évaluation de la performance des méthodes déjà existantes. Le développement de nouvelles méthodes robustes et fiables doit également être encouragé. De plus, des grands essais cliniques sont nécessaires afin d’identifier la technique qui mieux prédit la réponse des patientes.

- 59 -

Abstract Background. Human epidermal growth factor receptor 2 (HER2) plays a central role as a prognostic and predictive marker in breast cancer specimens. Reliable HER2 evaluation is central to determine the eligibility of breast cancer patients to targeted anti-HER2 therapies such as trastuzumab and lapatinib. Presently, several methods exist for the determination of HER2 status at different levels (protein, RNA and DNA level).

Design. In this review we will discuss the main advantages and disadvantages of the techniques developed so far for the evaluation of HER2 status in breast cancer specimens.

Results. Each technique has its own advantages and disadvantages. It is therefore not surprising that no consensus has been reached so far on which technique is the best for the determination of HER2 status.

Conclusion. Currently, emphasis must be put on standardization of procedures, internal and external quality control assessment, and competency evaluation of already existing methods in order to ensure accurate, reliable, and clinically meaningful test results. Development of new robust and accurate diagnostic assays should also be encouraged. In addition, large clinical trials are warranted in order to identify the technique that the most reliable predict positive response to anti-HER2 drugs.

- 60 -

Introduction Human epidermal growth factor receptor 2 (HER2) is a transmembrane tyrosine kinase receptor belonging to the family of epidermal growth factor receptors (EGFR) (1). The protein is encoded by the HER2 (ERBB2) gene, which is located on the long arm of chromosome 17 (17q12-21.32) (2). HER2 is an essential mediator of cell proliferation and differentiation in the developing embryo and in adult tissues (3). Its inappropriate activation, however, is associated with the development of several malignancies, including breast, ovarian, gastric, colorectal, pancreatic and endometrial cancers (4). In human breast cancer, HER2 gene amplification and receptor overexpression, which occur in 15 to 20% of breast cancer patients, are important prognostic markers for poor prognosis, including a more aggressive disease and a shorter survival (5). Moreover, HER2-positive status is considered a predictive marker of response to HER2-targeted drugs, including trastuzumab and lapatinib (6). Trastuzumab (Herceptin®, Roche) is a recombinant humanized monoclonal antibody that specifically targets the extracellular domain of the HER2 protein (7). Trastuzumab improves the outcomes of HER2- positive breast cancer patients in both the metastatic (8, 9) and adjuvant settings (10, 11). The Food and Drug Administration (FDA) approved trastuzumab for the treatment of HER2-positive metastatic breast cancer in 1998 and as adjuvant treatment for HER2-positive early stage breast cancer in 2006. Lapatinib (Tykerb®/Tyverb®, GlaxoSmithKline) is a small molecule inhibitor of the intracellular tyrosine kinase domain of both HER2 and EGFR receptors (12). In 2007 Lapatinib has been approved by FDA as combination therapy with capecitabine for the treatment of patients with HER2-positive advanced breast cancer who have progressed on trastuzumab-based regimens (13). Given its prognostic, predictive and therefore therapeutic implications, an accurate evaluation of HER2 status is crucial for identification of patients who would most likely benefit from targeted anti-HER2 therapies.

Several techniques have been developed for the evaluation of HER2 status in breast cancer specimens in clinical practice, at the protein level, DNA level, and at RNA level. Currently, there are several FDA-approved methods to evaluate HER2 status, including immunohistochemical (IHC) determination of HER2 protein expression or assessment of HER2 gene amplification using in situ hybridization (ISH), most commonly fluorescent ISH (FISH) (14, 15). However, since each technique has its own advantages and disadvantages, there is still no consensus on which method is superior for assessing the HER2 status in breast cancer specimens. This review provides an overview of the techniques that have been developed and tested over the last few decades for the determination of HER2 status in breast

- 61 -

cancer specimens. The principle of each method will be briefly presented. The central objective of the present review article, however, is to highlight the main advantages and disadvantages of each described technique. Main characteristics of the presented techniques are summarized in Table 2.1., whereas each technique and its advantages and disadvantages are thoroughly depicted in the next section.

Principles of the methods of analysis, advantages and disadvantages of each described technique

Southern blot The Southern blot technique has been used to determine HER2 gene amplification in breast cancer samples (16-22). Following DNA extraction from breast cancer frozen tissues, DNA is digested through a restriction enzyme. Digested DNA fragments are then separated by gel electrophoresis on agarose gel. Following DNA denaturation, DNA specimens are transferred from the gel to a membrane and hybridized with a radioactive-labelled HER2 single stranded DNA (ssDNA) probe. Labelled HER2 ssDNA probe will hybridize with the HER2 ssDNA sequence on the basis of strand annealing between complementary ssDNA molecules. The labelled HER2 sequence is then visualized by autoradiography (18-20, 22). Autoradiograms are then scanned with a densitometer (23).

The HER2 gene copy number is then compared to that of a control gene (16), to control DNA extracted from blood (18) or to DNA extracted from normal breast tissues (19). Tumors that showed a more than twofold increase in copy number compared to control unamplified DNA are considered amplified (16).

Advantages. Since DNA is very stable, it is considerably less degraded in tissues compared to protein and mRNA (24).

Disadvantages. Although very reliable, this technique is not applicable in routine diagnostic settings, since it is time-consuming and requires large amount of DNA (25, 26). In addition, this technique does not allow the morphologic preservation of tissue and therefore the evaluation of histological features of tumor. Since non-amplified non-neoplastic cells present in tumor cannot be isolated from cancer cells, results obtained potentially underestimate the real HER2 gene amplification of the sample through a dilution effect (21).

- 62 -

Northern blot The Northern blot assay allows the detection of HER2 RNA in frozen breast cancer specimens (21). This method is very similar to the Southern blot, with the exception that RNA molecules are detected instead of DNA sequences. After extraction from homogenized tissue sample, total cellular RNA or mRNA is separated by size via electrophoresis in an agarose gel and transferred to a nitrocellulose membrane. The HER2 RNA is then visualized via hybridization with an isotopic-labelled complementary probe (27). The labelled HER2 sequence is then visualized by autoradiography. The relative optical density (OD) of bands is measured by densitometry scanning. Tumors are then divided into 4 RNA expression categories: 1+ (0.1 to 0.5 OD units); 2+ (> 0.5 to 1.0 OD units); 3+ (> 1.0 to 1.5 OD units); 4+ (> 1.5 OD units) (21).

Advantages. Northern blot reagents are not too expensive, which allows the running of many gels at low cost. Moreover, quantity and quality of RNA can be verified after gel electrophoresis (28).

Disadvantages. One of main disadvantages of the Northern blot technique is that RNA molecules are often degraded in tissues. Indeed, even a slight degradation of RNA can compromise the quality of data and therefore the ability to quantify gene expression. Similar to Southern blot, the Northern blot technique, in addition of being a labor-intensive technique, does not allow an exclusive evaluation of HER2 status in cancer cells, as the morphology of tissues is destroyed during the homogenization of tissue samples (21).

Enzyme-linked immunosorbent assay (ELISA) HER2 protein is composed of a cytoplasmic domain with tyrosine kinase activity, a transmembrane domain and an extracellular domain (ECD) (29). The HER2 ECD can be cleaved from the full-length HER2 receptor present on the breast cancer cell membrane by matrix metalloproteases (30) and released into the serum (31).

Enzyme-linked immunosorbent assay (ELISA) allows the detection and quantification of proteins in fluids or cell lysates (32). In breast cancer, manual and automated ELISA assays have been used for the determination of serum concentration of HER2 ECD breast cancer patients (33-37). HER2 ECD is detected using 2 monoclonal antibodies recognizing two distinct epitopes of the antigen. In the manual assay, HER2 ECD is immobilized using a 96- well plate coated with the first monoclonal antibody. The immobilized protein is then incubated with the second monoclonal antibody, which is labelled with horseradish peroxidase (HRP).

- 63 -

After application of the HRP substrate, detection is accomplished by assessing the colored end product with spectrophotometry which correlates to the HER2 ECD concentration in sample (37). In the automated assay, HER2 ECD is visualized through direct chemiluminescent technology, using antibodies that are labelled with chemiluminescent compounds (i.e., acridinium ester, fluorescein) (38). The measured chemiluminescence is directly proportional to the HER2 ECD concentration in sample (35).

Among commercially available ELISA assays, one automated (Immuno-1®, Siemens Healthcare Diagnostics) and one manual ELISA assay (Siemens Healthcare Diagnostics) have been approved by the FDA in 2000. Another automated platform (ADVIA CentaurTM, Siemens Healthcare Diagnostics) has also received FDA approval in 2003 (39).

Some studies suggest that HER2 ECD could be used as a biomarker for the monitoring of the disease course and the patient’s response to therapy. Circulating levels of HER2 ECD greater than 15 ng/ml (this reference cut off was derived from the sera of 242 healthy women (36)) in HER2-positive breast cancer patients may be associated with the progression of primary tumors to metastatic breast cancer (39). In both metastatic and early breast cancer patients, ECD levels might reflect the HER2 full-length protein expression, as elevated HER2 ECD levels in serum (≥15 ng/ml) have been correlated with higher scores at the HER2 IHC (34, 35, 40, 41). In metastatic breast cancer patients, high HER2 ECD serum concentrations (≥15 ng/ml) have been also associated with resistance to endocrine therapy and chemotherapy (42) and in both metastatic and early breast cancer patients with worse survival (43, 44). In metastatic patients treated with trastuzumab, decreased HER2 ECD serum levels (>20%) were predictive of response to treatment (45). Another study conducted on metastatic breast cancer patients, however, did not observe a clear association between the changes in ECD levels and response to trastuzumab therapy (46). In general, results regarding the associations between ECD circulating levels and prognostic and predictive factors are very variable, depending on which assay was used or on which cut off value was considered (for a recent review see (47). Based on these conflicting data, the clinical use of the ELISA assay for the determination of HER2 ECD in patient’s serum has not yet been widely implemented (47, 48).

Advantages. ELISA is a quick and simple assay in addition of being a less invasive (only blood samples are needed) and quantitative test (32, 47). Moreover, since HER2 ECD can be measured directly in serum, ELISA can be used to monitor the dynamic changes of HER2 status over the course of the disease progression or following treatment (47).

- 64 -

Disadvantages. Results obtained by ELISA might not be reliable if the serum samples are from patients receiving trastuzumab treatment, since trastuzumab still present in patient’s serum might compete with the two antibodies used in the assay (49).

Western blot Western blot has been used to evaluate HER2 protein expression in frozen and FFPE breast cancer tissue samples (21, 50, 51). Following protein extraction from tissues, sample proteins are separated by size using SDS-polyacrylamide gel electrophoresis. Separated proteins are then transferred to a nitrocellulose or a polyvinylidene fluoride (PVDF) membrane. Membrane is firstly incubated with a primary antibody directed against HER2, followed by the incubation with a secondary antibody (radioactive- or HRP-labeled) raised against the primary antibody host species. Radioactive-labeled bands are visualized by autoradiography (52). HER2 expression levels in individual tumors is then determined by densitometric scanning and expressed as HER2 units based on a laboratory standard (51). More commonly, nowadays signal from HRP-labeled target can be detected through enhanced chemiluminescence (50).

Advantages. High sensitivity (as little as 0.1 ng of protein can be detected in a sample) represents the major advantages of the Western blot technique (51).

Disadvantages. Since proteins are less stable than DNA, they are less well preserved in tissues than DNA (24). The Western blot technique presents the same disadvantage as the other blotting techniques, namely that it is a time-consuming method and the morphology of tissue samples is not preserved (21, 51, 53).

Polymerase chain reaction (PCR)-based assays PCR is a method for the detection of DNA samples through the exponential amplification of target DNA sequences. At first, double-stranded DNA is denatured into single-stranded DNA template. Oligonucleotide primers, i.e. short single-stranded sequences that match the DNA sequence at each end of the region to be amplified, are then annealed to the single-stranded DNA template. In the following step, DNA polymerase synthetizes a new DNA strand by adding deoxynucleotide triphosphates complementary to the bases of the single-stranded DNA template. As DNA doubles during each PCR cycle, this results in exponential accumulation of the targeted DNA fragment. The products of a PCR reaction are usually analyzed by agarose gel electrophoresis, which allows the detection of the presence (but not the quantification) of the target sequence and the length of the fragment.

- 65 -

Polymerase chain reaction (PCR)-based assays can evaluate changes in both HER2 gene copy number and expression. Quantitative PCR (qPCR) and reverse transcription PCR (RT- PCR) have been used to evaluate HER2 gene amplification and HER2 expression, respectively, in breast cancer specimens in both FFPE and frozen tissue (24, 25, 54-63). qPCR, also called Real time PCR, is a form of PCR that is used for the DNA quantification in samples. DNA amplification is monitored while the reaction proceeds through the implementation of either DNA-binding dyes or fluorescently-labelled sequence-specific primers. Fluorescence signal produced during the amplification process is detected using thermal cycler equipped with a detector to monitor the emitted fluorescence. As fluorescence signal increases with a growing amount of PCR products, qPCR allows quantifying the amount of DNA formed after each cycle. HER2 gene and the reference gene are simultaneously quantified. A ratio between HER2 and the reference gene ≥ 2.0 is regarded as HER2 amplification (25, 55, 58-60, 62).

RT-PCR, often denoted as real time RT-PCR or quantitative reverse transcriptase PCR (qRT- PCR), allows the quantification of mRNA in biological samples. Following RNA extraction from tissue samples, extracted RNA is reverse transcribed into complementary DNA (cDNA). cDNA is then measured by qPCR. The relative fold change in gene expression is usually calculated using the comparative ΔΔCT method (64). At first, the relative quantitation of HER2 gene expression is calculated comparing the target gene expression with that of one or several housekeeping genes. The relative HER2 gene expression measured in samples is then normalized to a calibrator obtained by mixing RNA from several normal breast tissues samples. Several cut off values have been used in the literature to define HER2 overexpression determined by RT-PCR (24, 25, 54, 56-58, 61). Importantly, the Oncotype DxTM (Genomic Health, Redwood City, CA) assay is based on RT-PCR-technology to analyse the expression of 21 genes involved in breast cancer biology, including HER2, estrogen receptor (ER) and progesterone receptor (PR). The test, performed on mRNA extracted from FFPE tumor tissues, is used to predict the likelihood of breast cancer recurrence in early-stage, node- negative, ER-positive breast cancer patients, in addition to predict their chemotherapy benefit (65).

Advantages. Real-time PCR allows a rapid and quantitative analysis of gene amplification (60, 66). Being real-time PCR a DNA-based technique, variations in tissue fixation and processing have little impact on the results. Moreover, since only small quantities of DNA fragments are

- 66 -

required, real-time PCR represents a suitable technique for the evaluation of HER2 gene amplification from DNA isolated from FFPE tissues. In addition, real-time PCR is an easy, quick and inexpensive technique that yields reliable results even in cases with low level amplification (55). Moreover, since real-time PCR is a quantitative method, it is less sensitive to interobserver variability (62). Real time RT-PCR presents several advantages, including a large dynamic range and an accurate quantification (67).

Disadvantages. Although the PCR technique is an easy and reproducible technique, PCR technology has not yet been approved as a diagnostic tool for the evaluation of HER2 amplification. The main reason is that PCR results are often associated with false negative results due to the dilution of amplified tumor cells with surrounding non-amplified stroma cells or non-invasive breast lesions (59, 63, 66). However, this effect can at least in part be resolved through laser-assisted microdissection (LAM), which allows the isolation of tumor cells from archival FFPE tissues (59). Some authors suggest that the tumor histological subtype might considerably influence the efficacy of this assay. The impact of surrounding stromal and non- malignant cells on test result might be in fact significantly bigger in tumors in which tumor cells are scattered throughout the stroma, such as diffuse lobular tumors (55). LAM might be therefore particularly important in this subgroup of breast carcinomas.

Moreover, since mRNA integrity can be damaged by several factors including tissue fixation and processing and storage time (24), the evaluation of HER2 status at the mRNA level by reverse-transcriptase PCR (RT-PCR) using FFPE can be problematic (60). Therefore, the widespread RNA fragmentation observed in archival FFPE tissues (but not in frozen tissues) (24) may limit the use of RT-PCR for the evaluation of HER2 status for clinical purposes (57, 63). Moreover, similar to qPCR, RT-PCR fails to detect equivocal cases at IHC/FISH and produces false-negative results (57).

Multiplex ligation-dependent probe amplification (MLPA) Multiplex ligation-dependent probe amplification (MLPA) is a recent developed PCR method used to detect copy number variation, such as gene amplifications and gene deletions (68). After DNA denaturation and fragmentation, genomic DNA is hybridized with sequence-specific MLPA probes. The MLPA kit commercially available for the analysis of the HER2 gene in FFPE tissue (P004 HER-2 kit, MRC Holland, Amsterdam, The Netherlands) contains 3 probes for the HER2 gene, 11 probes for chromosome 17, and 25 control probes located chromosomes other than chromosome 17 (26). Each MLPA probe consists of two halves. One half is

- 67 -

composed of a target-specific sequence flanked by a universal primer sequence. The other half consists of a target-specific sequence, a sequence showing a probe-specific length and a universal primer sequence. Since the two probe halves recognize adjacent DNA target sequences, they are joined through the action of a ligase. This produces a complete probe flanked by universal primer binding sites that can be amplified by PCR, whereas unbound probe halves cannot be amplified (68). As each complete probe is characterized by a unique length, the resulting PCR products can be separated and identified by capillary electrophoresis. As the amounts of probe is proportional to the target copy number, copy number variations such a deletions and amplifications of target sequence can be recognized by the relative peak heights (68). HER2 copy number is determined by comparing the mean ratio of the HER2 probe peaks with the reference probes located in stable regions of the genome. Mean values smaller than 1.5 indicate non-amplified cases, mean values between 1.5 and 2.0 low-level amplified cases and mean values bigger than 2.0 HER2-amplified cases (26).

Advantages. MLPA assay is a fast, accurate and inexpensive technique. Moreover, only small quantities of DNA extracted from paraffin embedded materials are required. Fragmentation of DNA does not have an impact on the reliability of results. In addition, this method yields quantitative results (26).

Disadvantages. Some probe signals may be affected by factors such as sample purity and small changes in experimental conditions. Therefore, copy number changes detected by MLPA should always be confirmed by other methods. Furthermore, tissue morphology is not conserved using this assay and tumor heterogeneity can be missed (69). Moreover, MLPA assay requires special equipment, including thermocycler and capillary sequencer (48).

Immunohistochemistry (IHC) Immunohistochemistry (IHC) assays involve the detection of specific antigens in formalin-fixed, paraffin embedded (FFPE) tissues using specific antibodies. Immunohistochemical analysis is a commonly used method for the evaluation of HER2 protein overexpression in FFPE breast cancer specimens (14, 15). After antigen retrieval, tissue sections are incubated with a primary antibody directed against the HER2 protein. A visualization reagent (secondary antibody directed against the primary antibody and dextran polymer conjugated with horseradish peroxidase (HRP)) is subsequently applied. Antigen is visualized by adding the enzyme- specific substrate 3,3-diaminobenzidine tetrahydrochloride (DAB). Through the action of the

- 68 -

enzyme, DAB is converted to a brownish reaction product, resulting in membranous staining (70, 71).

So far, several commercially available diagnostic tests for the determination of HER2 protein expression have been approved by the FDA for the identification of patients who might benefit from targeted anti-HER2 therapies: the HercepTestTM kit (Dako, Glostrup, Denmark), the PathwayTM kit (clone 4B5; Ventana Medical Systems, Inc., Tucson, AZ, USA), the InSiteTM HER2/neu kit (clone CB11, BioGenex Laboratories, Inc., San Ramon, CA, USA) and the Bond Oracle HER2 IHC System (Leica Biosystems Newcastle Ltd, UK).

Membranous staining is scored on a semi-quantitative scale. According to the American Society of Clinical Oncology (ASCO), the College of American Pathologists (CAP) and Canadian recommendations for HER2 testing in breast cancer published in 2007, HER2 expression is scored as 0 (no staining), 1+ (weak or incomplete membrane staining in any percentage of tumor cells), 2+ (strong, complete membrane staining in ≤ 30% of tumor cells or weak/moderate heterogeneous complete membrane staining in ≥ 10% of tumor cells) or 3+ (strong, complete, homogeneous membrane staining in > 30% of tumor cells) (15). In 2013, the ASCO/CAP updated the guidelines in order to clarify the recommendations published in 2007. According to the newest ASCO/CAP guidelines, HER2 expression is scored as 0 (no staining or weak/incomplete membrane staining in ≤ 10% of tumor cells), 1+ (weak, incomplete membrane staining in > 10% of tumor cells), 2+ (strong, complete membrane staining in ≤ 10% of tumor cells or weak/moderate and/or incomplete membrane staining in > 10% of tumor cells) or 3+ (strong, complete, homogeneous membrane staining in > 10% of tumor cells) (14). For both 2007 and 2013 ASCO/CAP guidelines, HER2 immunohistochemical status is considered negative if the immunohistochemical score is 0 or 1+, equivocal if the score is 2+, and positive if the score is 3+. Patients with a positive HER2 status at the IHC are eligible for targeted anti- HER2 therapy. The IHC 2+ group is regarded as borderline and confirmatory testing using a gene amplification test (an ISH assay) should be performed. Patients with a tumor equivocal at the IHC but HER2 amplified at the gene amplification test are also eligible for therapy with HER2 inhibitors, whereas patients with a tumor that scores 0 or 1+ at the IHC and that is not HER2 amplified are not eligible (14, 15).

Advantages. IHC is easy to perform and relatively inexpensive (72). Since staining results can be viewed using a conventional bright-field microscope, protein overexpression can be evaluated in the context of tissue morphology (73). Furthermore, since staining is permanent,

- 69 -

slides can be stored. Another advantage is that IHC detects the overexpression of the HER2 protein, which represents the direct target of trastuzumab and other anti-HER2 therapies.

Disadvantages. The ability to accurately determine HER2 protein expression status by IHC can be significantly affected by numerous factors (22, 74, 75), including warm/cold ischemic time (76), delay and duration of fixation and fixative used (77, 78). The method for antigen retrieval and the training of the staff also play an important role (14, 15, 79). Furthermore, since commercially available antibodies display different specificity and sensitivity, HER2 overexpression rates vary considerably depending on the antibody used (80-85). An important factor regards the interpretation of results. As the interpretation is based on semi-quantitative scoring, immuno-histochemical analysis is susceptible to considerable interobserver variability and therefore to substantial discrepancies in HER2 IHC results (86), especially for cases scoring 2+ (87).

Variability in staining results, however, has remarkably decreased after the introduction of the highly standardized HercepTest (48). Moreover, staining variability can be further reduced through the utilisation of fully automated staining systems such as Ventana BenchMark (for the PATHWAY HER-2/neu assay, clone 4B5) (88).

In order to reduce variability in scoring results, ASCO/CAP recommendations have been published in 2007 and updated in 2013 for the interpretation of HER2 overexpression (14, 15). Moreover, subjectivity in the interpretation of membrane staining can be reduced through the implementation of quantitative image analysis, including the automated cellular imaging system (ACISTM, ChromaVision Medical Systems, Inc., San Juan Capistrano, CA) (89, 90) and HER2-CONNECTTM (Visiopharm A/S, Hoersholm, Denmark) (91). ASCO/CAP guidelines recommend the use of quantitative image analysis, especially for cases showing a weak HER2 membrane staining (IHC score 1+ and 2+) (15).

Fluorescence in situ hybridization (FISH) The fluorescence in situ hybridization (FISH) is a cytogenetic technique that uses fluorescent- labelled probes to detect specific DNA sequences in tissue samples (92). In breast cancer specimens, FISH assay is used to quantify the HER2 gene copy number within tumor cell nuclei (92). FISH can be performed either as single-color (HER2 probe only) or as dual-color assay (using differentially labelled HER2 and chromosome 17 centromere probes simultaneously) (93).

- 70 -

Following enzymatic digestion of tissues, fluorescent-labelled probes are applied to tissue sections. After DNA denaturation, probes hybridize with their targets in tissue. The HER2 probe, labelled with an orange fluorophore, targets the HER2 gene locus on chromosome 17. The green fluorophore-labelled chromosome 17 centromere probe targets the alpha satellite DNA sequence located at the centromeric region of chromosome 17. Nuclei are counterstained with 4,6-diamino-2-phenylindole (DAPI). Fluorescent hybridization signals can be visualized using a fluorescence microscope equipped with 100 x oil immersion objectives and appropriate filters (SpectrumOrange for locus-specific probe HER2, SpectrumGreen for centromeric probe 17 and the UV filter for the DAPI nuclear counterstain) (94, 95).

Three FISH assay kits have been approved by the FDA so far for the evaluation of HER2 gene amplification in breast cancer specimens. The single-probe INFORMTM HER2 FIH DNA kit (Ventana Medical Systems, Inc., currently discontinued), the dual-probe PathVysionTM HER-2 DNA probe kit (Abbott Molecular, Des Plaines, IL) and the dual-probe HER2 FISH PharmDxTM kit (Dako, Glostrup, Denmark).

According to the ASCO/CAP and Canadian recommendations for HER2 testing in breast cancer published in 2007, an average HER2 gene copy number < 4 signals/nucleus or HER2 gene/chromosome 17 copy number ratio (HER2/CEP17) of < 1.8 is considered FISH-negative (non-amplified); an average HER2 gene copy number ≥ 4 and ≤ 6 signals/nucleus or HER2/CEP17 ratio between 1.8 and 2.2 is considered FISH-equivocal; and an average HER2 gene copy number > 6 signals/nucleus or HER2/CEP17 ratio ≥ 2.2 is considered FISH-positive (amplified) (15). According to the updated ASCO/CAP guidelines published in 2013, an average HER2 gene copy number < 4 signals/nucleus or HER2 gene/chromosome 17 copy number ratio (HER2/CEP17) of < 2.0 with an average HER2 copy number < 4 signals/nucleus is considered FISH-negative (non-amplified); an average HER2 gene copy number ≥ 4 and < 6 signals/nucleus or HER2/CEP17 ratio < 2.0 with an average HER2 copy number ≥ 4 and < 6 signals/nucleus is considered FISH-equivocal; and an average HER2 gene copy number ≥ 6 signals/nucleus or HER2/CEP17 ratio < 2.0 with an average HER2 copy number ≥ 6 signals/nucleus or HER2/CEP17 ratio ≥ 2.0 is considered FISH-positive (amplified) (14). A minimum of 20 tumor cell nuclei are counted in at least two invasive tumor areas. For equivocal FISH samples, results are confirmed by counting 20 additional cells or repeating the FISH test (14, 15).

- 71 -

Advantages. Similar to IHC, FISH is performed on FFPE tissues (92). FISH assay yields results that are considered more objective and quantitative than immunohistochemical scoring (92, 96). Although still a matter of debate, several investigators consider FISH as being more accurate and reliable than IHC in the determination of HER2 status in breast cancer specimens (81, 86, 97-99). Moreover, since DNA is more stable than protein, preanalytical factors have less impact on assay results compared with IHC (32). Another advantage is represented by the presence of internal controls, consisting of non-amplified signals in nonneoplastic cells (ductal epithelial cells and stromal cells) adjacent the tumor (93). The dual-probe system evaluates HER2 gene amplification as a ratio of the total HER2 signals to those of chromosome 17. The inclusion of the chromosome 17 probe enables the recognition of chromosome 17 polysomy and therefore allows the distinction between pseudoamplification due to polysomy from true HER2 gene amplification (100). Although some studies reports that chromosome 17 polysomy is a rare event (101-103) and therefore rarely affects HER2 results and although clinical relevance of chromosome 17 polysomy in breast cancer is not well established yet (63, 104, 105), some authors argue that inclusion of CEP17 probe is of central relevance for the accurate determination of HER2 amplification, particularly in cases with low-level amplification (106-108). In fact, chromosome 17 polysomy is reported as a frequent reason for discordance of results obtained with mono-color and dual-color FISH (109, 110).

Disadvantages. Although FISH has been established as a very robust method to evaluate HER2 gene amplification, this technique presents several disadvantages. FISH is 9 times more time-consuming (36 hours vs. 4 hours) (71) and three times more expensive compared to immunohistochemical analysis ($140 vs. $50) (111). Moreover, FISH testing requires costly equipment for signal detection and recognition. Expensive fluorescence microscope equipped with high magnification oil immersion objectives and multiband filters are indispensable, making thereby difficult to integrate FISH assay in every routine diagnostic laboratories (85). Furthermore, as fluorescence signals fade within a few weeks, hybridization results should ideally be recorded with an expensive CCD camera (85, 112). As morphological aspects of the tissues are difficult to evaluate under fluorescence, distinguishing of invasive breast cancer from breast carcinoma in situ - where HER2 gene amplification or protein overexpression has different clinical significance- is rather complicated and time-consuming (113). Therefore, FISH assay can be interpreted only by well-trained personnel (113, 114). Some authors suggest that since FISH analysis is performed at high magnification, intratumoral heterogeneity can be missed (115, 116).

- 72 -

Although FISH assay is considered a more objective and quantitative technique than IHC, reproducibility of FISH results depends on recognition of the invasive component (117). Moreover, discrepancies among observers occur mainly in samples showing low-level amplification (106). Another important aspect is that fluorescence signals counting is time- consuming. To overcome this problem, image analysis software for the automated evaluation of fluorescent signals have been developed. Several studies have demonstrated an excellent agreement between HER2/CEP17 ratios calculated through manual counting and those resulting from automated image analysis software (118-120). Some of these image analysis software have been approved by the FDA for the automated evaluation of HER2 gene amplification, including the Metafer (MetaSystems, Altussheim, Germany) and the Ariol HER2/neu FISH (Applied Imaging Corp, San Jose, CA). In addition to reduce the time required for the analysis of fluorescent signals, these software allow the storing of captured images and therefore the archiving of cases for future studies (118).

Bright-field in situ hybridization (BRISH) methods In order to overcome FISH limitations, alternative in situ methods, such as chromogenic in situ hybridization (CISH), silver-enhanced in situ hybridization (SISH), gold-facilitated autometallographic in situ hybridization (GOLDFISH) and bright-field double in situ hybridization (BDISH) have been developed to detect HER2 amplification. These recent techniques combine practical advantages of immunohistochemical analysis with the reproducibility of FISH method (121). Similar to FISH, these alternative ISH methods allow the quantification of HER2 and CEP17 signals and therefore the identification of HER2 gene amplification. However, since visualization is achieved using other reactions than fluorescence-labelled probe, signals can be analyzed using a standard bright-field microscope, allowing the simultaneous analysis of gene copy numbers and morphological features of tissue (121). Signals can be therefore exclusively analyzed in invasive compartment, excluding normal tissue elements and non-invasive carcinoma. Similar to immunohistochemical analysis and FISH, BRISH assays are performed on FFPE tissue sections (121).

In the following subsections, we will briefly describe each single BRISH technique, as well the distinctive advantages and disadvantages of each assay.

Chromogenic in situ hybridization (CISH) The chromogenic in situ hybridization (CISH) allows the visualization of target genes in tissue sections using peroxidase enxyme-labelled probes (121). Similar to FISH, CISH can be

- 73 -

performed either as single-color (HER2 probe only) or as dual-color (HER2 and chromosome 17 probes) assay. The single-color CISH assay (SPOT-Light® HER2 CISH kit, Life Technologies, Inc.) and the dual-color CISH assay (HER2 CISH PharmDx kit, DAKO) have been approved by the FDA in 2008 and 2011 respectively, for the evaluation of HER2 gene amplification in breast cancer specimens, either as primary test to evaluate HER2 status or to retest equivocal immunohistochemical results (IHC score 2+) (14).

With the monochromatic CISH, only the absolute HER2 copy number is scored (73, 122). Following heat-induced epitope retrieval and enzymatic digestion, a digoxigenin-labelled HER2 DNA probe is applied to the slide. After DNA denaturation and hybridization, the hybridized HER2 probe is visualized by sequential incubation with mouse antidigoxigenin and polymerized HRP-goat antimouse, followed by development with DAB as chromogen. HER2 gene copies are then recognizable as intranuclear brown chromogenic reaction product signals. Slides are counterstained with hematoxylin (96, 116, 122-124). Evaluation of chromosome 17 can be performed on a consecutive slide, using the same protocol as for the HER2 probe, with the exception that a biotin-labelled chromosome 17 centromeric (CEP17) probe will used instead of the digoxigenin-labelled HER2 DNA probe. The CEP17 probe is visualized through incubation with HRP-conjugated anti-biotin antibodies and DAB (96, 124).

HER2 signals are recognizable either as large peroxidase-positive intranuclear signal clusters or as numerous individual peroxidase-positive small signals (123). Cases with low-level amplification show six to 10 signals per nucleus in more than 50% of cancer cells, whereas high-level amplified cases are characterized by an average HER2 copy number > 10 or by large gene copy clusters in more than 50% of tumor cell nuclei (73, 106, 123, 125).

Similar to the dual-color FISH, the dual-color CISH assay allows the simultaneous visualization of the HER2 probe and the chromosome 17 centromere probe on the same slide (126, 127). The dual-color CISH assay is performed using sequential horseradish peroxidase (HRP) and alkaline phosphatase (AP) enzymatic reactions to detect digoxigenin-labelled HER2 probes and biotin or dinitrophenyl-labelled chromosome 17 probes (128, 129). HRP-conjugated antibodies are then visualized using a green chromogen (HER2 signals) and AP-conjugated antibodies using a red chromogen (CEP17 signals).

Alternatively, AP-conjugated antibodies and HRP-conjugated antibodies can be used to detect Texas Red-labelled HER2 DNA probe and fluorescein isothiocyanate (FITC)-labelled CEP17 probe, respectively. AP-conjugated anti-Texas Red antibodies are then developed using a red

- 74 -

chromogen (HER2 signals), whereas the HRP-conjugated anti-FITC antibodies are developed using a blue chromogen (CEP17 signals) (126, 127, 130). Red fluorescence signals produced by Texas Red are therefore converted to red chromogenic HER2 signals and green fluorescence signals produced by FITC are converted to blue chromogenic CEP17 signals. Slides are then counterstained with hematoxylin. Results obtained by dual-color CISH are reported as dual-color FISH (15, 113, 126, 128-130).

Advantages. Interpretation of results is performed using equipments already available in routine histopathology laboratories, such as a conventional bright-field microscope (123). Moreover, as results can be visualized using a standard bright-field microscope, CISH analysis allows the simultaneous evaluation of tissue morphology and copy number alterations, reducing the risk of analysing non-malignant compartment (126, 131). Similar to FISH, the CISH assay has the advantage of the built-in internal control (93). Another advantage over FISH is that tumor heterogeneity is promptly recognizable, even at 20x magnification (132). Furthermore, chromogenic signals are permanent (73). Slides can therefore be stored and used for re-evaluation or retrospective studies (126, 127). Comparatively to FISH, CISH assay is less expensive (98$/slide vs. 183$/slide) (48) and quicker (248 vs. 314 minutes) (96). As the CISH assay allows an easier identification of the invasive component compared to FISH, analysis of CISH signals is consequently less time consuming than FISH (96, 126). The inclusion of a CEP17 probe in the dual-color assay allows the calculation of the HER2/CEP17 ratio, enabling exclusion of chromosome 17 polysomy (124). Additionally, dual-color assay can be performed on a programmable automated slide stainer, improving the assay efficiency and reducing the risk of errors (128, 129).

Disadvantages. The main disadvantage of mono-color CISH is that the use of monochromatic signals does not allow the determination of the HER2/CEP17 ratio and therefore the detection of chromosome 17 polysomy, making thereby necessary to hybridize the control probe (chromosome 17) on an adjacent section. This additional staining is time-consuming and cost ineffective (126).The evaluation of HER2 gene copies number can be difficult in areas presenting high level amplification, as overlapping dots lead to formation of signal clusters (126).

Although relatively infrequent, technical problems including over- or underfixation of tissue samples, over- or underdigestion of tissue sections and high background debris can lead to erroneous results or loss of signal (73, 112, 122).

- 75 -

Silver-enhanced in situ hybridization (SISH) Silver-enhanced in situ hybridization (SISH) is an automated enzyme metallography procedure (133). Enzymatic metallography is a recent developed method for in situ hybridization, in which an enzyme reaction is used to selectively deposit metal from solution to produce a black, sharply defined staining (133). In the specific case of SISH, HRP-conjugated probes are used to deposit metallic silver at the reaction site (133).

All steps of the automated assay including deparaffinization and counterstaining are performed on the Ventana BenchMark® XT automated slide stainer (134-136). HER2 and chromosome 17 analysis is performed on sequential slides (134-136). After heat-induced epitope retrieval and proteolytic pre-treatment, slides are incubated with dinitrophenol (DNP)-labelled HER2 probe and with DNP-labelled CEP17 probe, respectively. Following DNA denaturation and hybridization, both HER2 and CEP17 probes are visualized by application of a rabbit anti-DNP antibody followed by a horseradish peroxidase (HRP)-labelled goat anti-rabbit antibody (134, 135, 137, 138). As mentioned above, probes are visualized through the process of enzyme metallography. This process is driven by the sequential addition of silver acetate, hydroquinone and hydrogen peroxide. Silver ions are reduced to metallic silver ions through the action of the reducing agent hydroquinone. This reaction is stimulated by hydrogen peroxide, which is a substrate for horseradish peroxidase. The silver precipitation is deposited in the nucleus and HER2 or CEP17 signals are recognizable as black dots within cell nuclei (135, 137, 139).

Similar to FISH, HER2 gene amplification status obtained by SISH is reported as HER2/CEP17 ratio, according to the ASCO/CAP guidelines (15, 115, 135-138, 140, 141).

Advantages. SISH is a fully automated bright-field in situ hybridization assay. In addition, since SISH protocol is automated, this assay is four times faster to perform than FISH (6 hours vs. 36 hours) (135). Moreover, staining results obtained with SISH show a higher signal resolution compared to those obtained with DAB-based methods (IHC, CISH) (133). Discrete black dots and signal clusters are easier to evaluate compared to other bright-field ISH techniques (142, 143). Furthermore, image analysis software (Ventana Image Analysis System, VIAS) for the automated quantification of HER2 gene and CEP17 copy number in sequential sections has been developed (121).

Disadvantages. To correct for chromosome 17 aneusomy, the hybridization of a further section is required for separate evaluation of chromosome 17 centromere copy number (142).

- 76 -

Moreover, SISH assay is a relatively new technique and requires a rather expensive automated slide stainer that may not be available in many routine pathology laboratories (128).

Gold-facilitated autometallographic in situ hybridization (GOLDFISH) The gold-facilitated autometallographic in situ hybridization (GOLDFISH) is a recently developed method that combines the use of streptavidin-labelled gold nanoparticles with gold- based autometallography. Autometallography is the process through which gold nanoparticles are enlarged and therefore visualized through bright-field microscopy by selective deposition of gold ions onto their surface (144, 145). GOLDFISH has been tested for the determination of HER2 gene amplification in breast cancer specimens (145, 146).

The GOLDFISH assay is based on the tyramide signal amplification principle (147). At first, tissue sections are hybridized with biotin-labelled probe. The probe signal is amplified through the addition of streptavidin-peroxidase and biotinylated tyramine. Tissue sections are then incubated with Streptavidin-Nanogold® (Nanoprobes, Inc., Yaphank, NY), 1.4 nm diameter gold particles covalently conjugated with Streptavidin, and GoldEnhance (Nanoprobes, Inc.), a reagent that catalyzes the deposition of gold ions in solution around the gold nanoparticles, increasing therefore their sizes (145). This produces a dense black, permanent signal visible with a conventional bright-field microscope (146).

Cells are considered non-amplified if up to two small nonconfluent black signals are identified within nuclei of invasive carcinoma. Amplified cases are characterized by large clusters of confluent black signals, whereas cases showing low level amplification are characterized by four to eight separately distinguishable black small signals (145).

Advantages. As amplification cases are immediately recognized by the presence of massive confluent deposit in cell nuclei, signal counting in amplified cases is not required, allowing a quicker examination of amplified (146).

Disadvantages. The signals obtained with the GOLDFIDH technique is less discrete compared with those obtained by SISH (142).

Bright-field double in situ hybridization (BDISH) Bright-field double in situ hybridization (BDISH) or dual-color in situ hybridization (Dual ISH) is a fully automated bright-field ISH assay for the simultaneous detection of HER2 and CEP17 signals on the same FFPE breast cancer tissue section (142, 148). This technique combines

- 77 -

the detection of HER2 gene copies through the deposition of metallic silver particles at reaction site, as in the mono-color SISH assay, with the visualization of CEP17 copies with fast red, similar to the CISH procedure (142, 149).

The whole assay is performed with the FDA-approved dual-color ISH assay (INFORM HER2 Dual ISH DNA probe cocktail assay, Ventana Medical Systems) using the BenchMark® XT automated slide stainer (Ventana) (142, 148, 149). Following antigen retrieval, DNP-labelled HER2 probe is applied to the slide. HER2 probe is visualized with a rabbit anti-DNP antibody followed by a HRP-labelled goat antirabbit antibody. HRP enzyme is then developed with a silver precipitate. HER2 signals are visualized as discrete black spots in the nuclei (142, 148, 149).

Sequentially, DNP-labelled CEP17 probe is applied to the slide. CEP17 probe is visualized by the incubation with a rabbit anti-DNP antibody, followed by an AP-labelled goat anti-rabbit antibody. AP enzyme is finally developed by fast red reagent. The CEP17 signals are visualized as red spots in the nuclei. CEP17 red spots are slightly larger than the HER2 black spots. Slides are then counterstained with hematoxylin (142, 148).

Similar to FISH and SISH, HER2 gene amplification status determined by BDISH is reported as HER2/CEP17 ratios, as recommended by the ASCO/CAP guidelines (15).

Advantages. The BDISH assay allows the simultaneous visualization of both HER2 and CEN17 targets on the same slide. This is very relevant especially for cases displaying chromosome 17 aneusomy or intratumoral heterogeneity (142). Furthermore, since the assay procedure is completely automated, reproducibility of results is increased, as risk of human errors is diminished (142, 143). Moreover, in analogy to the SISH assay, HER2 signals are visualized as discrete black dots that are particularly easy to count (142, 143). As the HER2 signals and CEP17 signals differ in color and size, both signals can be distinguished from each other even they co-localize within cell nuclei (142). Moreover, since the assay is completely automated, results are available within 6 hours (143). Similar to other bright-field ISH techniques, HER2 amplification status can be examined with a conventional bright-field microscope in the context of tissue morphology. Moreover, slides can be permanently archived (142).

Disadvantages. BDISH assay presents the same disadvantages as CISH and SISH.

- 78 -

HER2 gene-protein assay These recently developed assays allow the simultaneous evaluation of HER2 protein expression and HER2 gene amplification on a single FFPE breast cancer tissue sections. Several combinations have been described so far in the literature.

Downs-Kelly and collaborators combined IHC with mono-color SISH (EnzMet GenePro)

(150). HER2 gene amplification assessment was performed using a HRP-based silver deposition method, followed by HER2 immunohistochemical staining using AP and fast red substrate. They observed an excellent correlation between the results obtained with the combination gene-protein assay and those obtained by IHC alone.

Ni and collaborators combined HER2 immunohistochemical staining with mono-color CISH (151). Similar to Downs-Kelly et al., HER2 immunohistochemical staining was achieved using an AP-based fast red system. Subsequently, CISH assay was performed using digoxigenin- labelled HER2 probe, which was visualized via anti mouse anti-digoxigenin antibody and an antimouse HRP polymer conjugate, followed by incubation with DAB. The combined protocol was performed partially on an automated slide stainer.

Reisenbichler and collaborators (152) tested a HER2 gene-protein assay consisting in the sequential performance of mono-color CISH followed by HER2 immunohistochemical staining. The digoxigenin-labelled HER2 probe was visualized via anti mouse anti-digoxigenin antibody, antimouse HRP polymer conjugate, followed by incubation with DAB. HER2 protein expression was evaluated with an HRP-based DAB system. They observed 100% agreement between the gene amplification results obtained when CISH was performed alone compared to the sequential protocol. Compared with separately performed immunohistochemical staining, agreement with immunohistochemical staining obtained with the combined protocol was observed in 78% of cases.

Nitta and collaborators (153) evaluated the performance of an automated tricolor HER2 gene protein assay, consisting of a HER2 immunohistochemical staining followed by dual-color ISH. Analysis of HER2 gene amplification was performed using the FDA-approved dual-color ISH assay (INFORM HER2 Dual ISH DNA probe cocktail assay, Ventana Medical Systems). DNP- labelled HER2 probe was detected with anti-DNP antibody, HRP-conjugated antibody and metallic silver deposition. HER2 signal was visualized as discrete black dot. Digoxin (DIG)- labeled CEP17 probe was visualized with anti-DIG antibody, AP-conjugated antibody and fast

- 79 -

red. CEP17 signal appeared as red dot. HER2 protein expression was evaluated using DAB as chromogen. Both staining were performed using an automated slide stainer. The overall concordance between the results obtained with the combined tests and the separate IHC and dual-color ISH was excellent (concordance rate ranged between 97.8% and 99.5% for IHC and concordance oscillated between 96% and 97.7% for dual-color ISH).

Analysis of the HER2 gene-protein assay is performed according to the ASCO/CAP guidelines used in the individual assay (15, 150-153).

Advantages. The HER2 gene-protein assays combine the advantages of IHC with those of BRISH methods. Whereas IHC represents an effective and inexpensive method to detect tumor heterogeneity, ISH assays allow quantitative analysis of gene amplification (153). These assays facilitate the interpretation of results by conventional bright-field microscopy and improve the identification of discordance between protein overexpression and gene amplification (150). Since the protein-gene assay allows the simultaneous evaluation of HER2 protein overexpression and HER2 gene amplification at the single cell level, this method could be particularly advantageous in cases with equivocal results using one single technique or in cases showing heterogeneous gene amplification (152). This is clinically relevant, since intratumoral heterogeneity has been identified as a major cause of discordance between HER2 IHC and HER2 FISH assay results in breast cancer specimens (153).

Disadvantages. ISH assays are expensive and costly automated slide stainers are required. mRNA in situ hybridization This new method allows the evaluation of gene expression in FFPE tissues, using specific single stranded DNA probes that hybridize to target gene mRNA. HistoSonda (Cenbimo, Lugo, Spain) is the kit presently available for the determination of HER2 gene expression in FFPE breast cancer samples using the mRNA in situ hybridization technique (154). Following protein digestion, tissue sections are hybridized with digoxigenin-labelled single stranded DNA probe targeting HER2 mRNA (154). After hybridization, tissues sections are incubated with an anti-digoxigenin antibody and a HRP-labelled secondary antibody. Hybridized probes are subsequently visualized using DAB as a chromogen.

The interpretation of results is carried out in a binary fashion. A case is considered positive for HER2 expression when a brownish coloration of cell cytoplasm is observed. On the contrary, the absence of staining is interpreted as a negative result (154).

- 80 -

Advantages. mRNA in situ hybridization is a simple, relative inexpensive and fast technique. As the hybridization step is very quickly (1 hour), the procedure takes 3 hours from deparaffinization to counterstaining. The assay can be performed on an automated IHC stainer (154). Furthermore, this method is less toxic compared to other in situ hybridization techniques, since the use of the toxic formamide as hybridization buffer has been suppressed. Detection of chromogenic signals can be performed using a conventional bright-field microscope. In addition, this method is highly reproducible. Since single stranded DNA fragments have a length that is up to 30 times greater than commercial oligoprobes, the assay specificity is increased, as the likelihood that the DNA probe hybridizes with other than the target sequence is remarkably reduced.

Disadvantages. mRNA is significantly less stable than DNA in fixed tissues and this could affect the reliability of the test (24).

Instant-quality fluorescence in situ hybridization (IQFISH) In analogy to conventional FISH, instant-quality fluorescence in situ hybridization (IQFISH) allows the quantitative determination of HER2 gene amplification in breast cancer specimens (155, 156). The HER2 IQFISH pharmDX kit (Dako, Glostrup, Denmark) has received FDA approval in 2013. The IQFISH technique is performed in the same way as manual FISH (i.e., pre-treatment, pepsin digestion, denaturation, hybridization, stringent washing and visualization). The main difference lays in the introduction of a new hybridization buffer (IQFISH buffer) which dramatically reduces the time required for the hybridization step (1 hour vs. 16 hours) (155, 156). The utilization of this new buffer therefore considerably decreases the total assay time necessary to perform the IQFISH compared to conventional FISH (3½ hours vs. 16-20 hours).

IQFISH analysis is performed according to the same guidelines as for FISH (15, 155, 156).

Advantages. Since the IQFISH hybridization buffer considerably decreases the assay time, this new method allows the determination of HER2 gene amplification status in one day (instead of 2 days as for the traditional FISH) (155, 156). This technique is therefore particularly relevant when HER2 status determination is urgent, such as in cases where neoadjuvant therapy is indicated (155). Another important aspect concerns safety. While hybridization buffers currently used contain the toxic formamide, the IQFISH buffer is non-toxic (155).

- 81 -

Disadvantages. IQFISH has the majority of disadvantages of manual FISH (i.e., high reagent cost, disruption of tissue morphology, etc.).

Automated HER2 FISH assay This new assay allows the fully automation of the FISH technique for the determination of HER2 gene amplification (157, 158). The automated HER2 FISH assay is performed with the Leica HER2 FISH System and the Leica BOND-MAX stainer (Leica Biosystems Newcatsle Ldt, UK), using the FDA-approved PathVysionTM HER2 DNA Probe kit, which contains dual HER2 and chromosome 17 FISH probes. So far, two studies have analyzed the performance of fully automated HER2 FISH assay in breast cancer specimens (157, 158). Both studies observed high level of concordance between results obtained with manual and automated FISH (concordance rate ranging between 96 and 100%) (157, 158).

Counting and analysis of automated HER2 FISH assay is performed in the same manner as manual FISH (15, 157, 158).

Advantages. Compared to manual FISH, automated FISH is less expensive, since less human intervention is required (157). Moreover, automated FISH enables faster processing of samples and recording of results (157). Furthermore, the automation of the technique might also reduce laboratory errors and therefore reduce inter-observer and inter-laboratory variability, as well as improve results reproducibility (157, 158). Yoon and collaborators observed that nuclear membranes was less digested using the automated FISH compared to manual FISH, resulting in better conservation of nuclear material and therefore easier distinction of the invasive component in the slide (158).

Disadvantages. Especially in breast cancer specimens with abundant stroma, the automated FISH assay produced a foggier background compared to manual FISH (158). This technique and manual FISH present several common disadvantages (i.e., expensive and time- consuming methods).

Performance of assays used for the HER2 status determination in breast cancer in predicting response to anti-HER2 therapies To date, only few studies have compared the predictive performance of different assays used for the HER2 evaluation in breast cancer specimens (8, 9, 159, 160). Results obtained from these studies are often quite different. Main reasons for inconsistencies between results could be attributed to the different sensitivity of the applied antibodies, in the lack of standardized

- 82 -

staining evaluation, in the different definition of HER2 overexpression and HER2 gene amplification (especially in studies published before the publication of the 2007 ASCO/CAP scoring criteria), and the type of treatment received.

One study reported that FISH performed better compared to IHC (clones 4D5 and CB11) in the prediction of trastuzumab response in a cohort of 114 HER2-overexpressing metastatic breast cancer patients (9). It is however relevant to note that in this early study, tumors showing IHC 2+ were also considered as HER2-overexpressing, independently from the HER2 gene amplification status. When only 3+ cases are considered, the response rates to trastuzumab were very similar to those observed in the HER2 gene amplified group.

In a cohort of metastatic breast cancer patients treated with trastuzumab monotherapy, it was reported that centrally determined IHC (using HercepTest) was slightly better compared to central FISH in predicting response to trastuzumab (159). Sensitivity in determining response to trastuzumab was 100% for HercepTest compared to 84.2% for FISH.

Among 95 metastatic breast cancer patients treated with trastuzumab in combination with paclitaxel, Seidman and colleagues reported that the strongest association between clinical response rate and the assays used for the determination of HER2 (IHC: HercepTest, TAB250, CB11, and Pab1; FISH) occurred with FISH and CB11 and TAB250 antibodies (8).

In a cohort of 52 consecutive HER2-positive metastatic breast cancer patients treated with trastuzumab combined with chemotherapy, patients with tumors with HER2 FISH ratio < 3.0 showed a shorter progression free survival, whereas patients with tumors that did not overexpress HER2 (IHC score 1+) had a shorter overall survival (160).

Conclusion Assessment of the HER2 status in breast cancer specimens is a key component for defining prognosis and patient management. Considering the clinical and economic implications of targeted anti-HER2 treatments, accurate HER2 test results are essential. False negative results would deny the access of patient to the potential benefits of targeted treatment, whereas false positive results would expose patients to the potential cardiotoxic side effects of these expensive drugs without experiencing any therapeutic advantages.

The optimal method for assessing HER2 status in breast cancer specimens has remained controversial, since each technique is characterized by its own advantages and

- 83 -

disadvantages. IHC is an easy and relatively inexpensive technique. However, the reliability of immunohistochemical results is influenced by a variety of pre-analytical, analytical and post- analytical factors. FISH is considered a more robust technique than IHC, but high cost of reagents and the requirement of expensive laboratory equipment for the evaluation of fluorescent signals that fade over time represent major limitations of this technique. Bright-field ISH assays allow the simultaneous evaluation of HER2 gene amplification and histopathological morphology using a conventional bright field microscope. However, these techniques are relatively new and require sometimes expensive automated slide stainers. Although some publications consider FISH a more accurate and reliable test than IHC for the determination of HER2 status in breast cancer specimens, ASCO⁄CAP guidelines (14, 15) do not define a true gold standard method. They stress instead that standardization of laboratory procedures and test interpretation, as well as internal and external quality control assessment are key elements to provide reliable test results. ASCO/CAP guidelines recommend the use of IHC, FISH or bright-field ISH assays for the evaluation of HER2 status in breast cancer specimens. The introduction of new test methodology is encouraged, but the ASCO/CAP panel recommends that the new diagnostic test has to be previously compared with a reference test in order to confirm the clinical benefit of the new test for the patient (14, 15).

Molecular biology techniques including Northern blot, Southern blot and Western blot are very useful for the comprehension of the biology of HER2. However, they are rather impractical in the clinical setting. New technologies including mRNA in situ hybridization or high-throughput technologies such as PCR-based assays and MLPA are promising and potential less costly compared to IHC and ISH assays. Their implementation in clinical routine, however, still needs to be thoroughly examined. Moreover, guidelines regarding the standardization and interpretation of diagnostic tests in addition to conventional IHC and ISH assays are needed (161).

As a general remark it should be kept in mind that although the ideal diagnostic test is 100% accurate and 100% precise, in practice, every diagnostic test is prone to error (161).

In conclusion, since the optimal approach to testing HER2 status in breast cancer is still matter of debate, stringent standardization of existing tests and development of new robust and reliable diagnostic test are warranted. As further concluding remarks it should also be noted that the superiority of one assay over another in the evaluation of HER2 status in breast cancer specimens can be unequivocally determined only by analyzing the clinical response to anti-

- 84 -

HER2 therapies as reference. As the few researches that have been conducted so far comparing the predictive value of HER2 assays toward anti-HER2 therapies have shown conflictive results, more studies that analyse this intriguing subject should be performed.

Acknowledgements DF received doctoral fellowships from the Fonds de recherche du Québec - Santé and the Laval University Cancer Research Center. CD is a Junior Investigator of the Canadian Cancer Society.

Conflict of interest Authors have no conflict of interest to declare.

- 85 -

References 1. Yarden Y, Sliwkowski MX. Untangling the ErbB signalling network. Nature reviews Molecular cell biology. 2001;2(2):127-37.

2. Popescu NC, King CR, Kraus MH. Localization of the human erbB-2 gene on normal and rearranged chromosomes 17 to bands q12-21.32. Genomics. 1989;4(3):362-6.

3. Olayioye MA, Neve RM, Lane HA, Hynes NE. The ErbB signaling network: receptor heterodimerization in development and cancer. The EMBO journal. 2000;19(13):3159-67.

4. English DP, Roque DM, Santin AD. HER2 expression beyond breast cancer: therapeutic implications for gynecologic malignancies. Molecular diagnosis & therapy. 2013;17(2):85-99.

5. Soerjomataram I, Louwman MW, Ribot JG, Roukema JA, Coebergh JW. An overview of prognostic factors for long-term survivors of breast cancer. Breast cancer research and treatment. 2008;107(3):309-30.

6. Esteva FJ, Yu D, Hung MC, Hortobagyi GN. Molecular predictors of response to trastuzumab and lapatinib in breast cancer. Nature reviews Clinical oncology. 2010;7(2):98- 107.

7. Nahta R, Hung MC, Esteva FJ. The HER-2-targeting antibodies trastuzumab and pertuzumab synergistically inhibit the survival of breast cancer cells. Cancer research. 2004;64(7):2343-6.

8. Seidman AD, Fornier MN, Esteva FJ, et al. Weekly trastuzumab and paclitaxel therapy for metastatic breast cancer with analysis of efficacy by HER2 immunophenotype and gene amplification. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2001;19(10):2587-95.

9. Vogel CL, Cobleigh MA, Tripathy D, et al. Efficacy and safety of trastuzumab as a single agent in first-line treatment of HER2-overexpressing metastatic breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2002;20(3):719- 26.

- 86 -

10. Gianni L, Dafni U, Gelber RD, et al. Treatment with trastuzumab for 1 year after adjuvant chemotherapy in patients with HER2-positive early breast cancer: a 4-year follow-up of a randomised controlled trial. The lancet oncology. 2011;12(3):236-44.

11. Piccart-Gebhart MJ, Procter M, Leyland-Jones B, et al. Trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer. The New England journal of medicine. 2005;353(16):1659-72.

12. Xia W, Mullin RJ, Keith BR, et al. Anti-tumor activity of GW572016: a dual tyrosine kinase inhibitor blocks EGF activation of EGFR/erbB2 and downstream Erk1/2 and AKT pathways. Oncogene. 2002;21(41):6255-63.

13. Ryan Q, Ibrahim A, Cohen MH, et al. FDA drug approval summary: lapatinib in combination with capecitabine for previously treated metastatic breast cancer that overexpresses HER-2. The oncologist. 2008;13(10):1114-9.

14. Wolff AC, Hammond ME, Hicks DG, et al. Recommendations for human epidermal growth factor receptor 2 testing in breast cancer: American Society of Clinical Oncology/College of American Pathologists clinical practice guideline update. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2013;31(31):3997-4013.

15. Wolff AC, Hammond ME, Schwartz JN, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2007;25(1):118-45.

16. Andrulis IL, Bull SB, Blackstein ME, et al. neu/erbB-2 amplification identifies a poor- prognosis group of women with node-negative breast cancer. Toronto Breast Cancer Study Group. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 1998;16(4):1340-9.

17. Berger MS, Locher GW, Saurer S, et al. Correlation of c-erbB-2 gene amplification and protein expression in human breast carcinoma with nodal status and nuclear grading. Cancer research. 1988;48(5):1238-43.

- 87 -

18. Jukkola A, Bloigu R, Soini Y, Savolainen ER, Holli K, Blanco G. c-erbB-2 positivity is a factor for poor prognosis in breast cancer and poor response to hormonal or chemotherapy treatment in advanced disease. European journal of cancer. 2001;37(3):347-54.

19. Scorilas A, Yotis J, Pateras C, Trangas T, Talieri M. Predictive value of c-erbB-2 and cathepsin-D for Greek breast cancer patients using univariate and multivariate analysis. Clinical cancer research : an official journal of the American Association for Cancer Research. 1999;5(4):815-21.

20. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science. 1987;235(4785):177-82.

21. Slamon DJ, Godolphin W, Jones LA, et al. Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science. 1989;244(4905):707-12.

22. Tsuda H, Tani Y, Hasegawa T, Fukutomi T. Concordance in judgments among c-erbB- 2 (HER2/neu) overexpression detected by two immunohistochemical tests and gene amplification detected by Southern blot hybridization in breast carcinoma. Pathology international. 2001;51(1):26-32.

23. Miyada CG, Wallace RB. Oligonucleotide hybridization techniques. Methods in enzymology. 1987;154:94-107.

24. Cronin M, Pho M, Dutta D, et al. Measurement of gene expression in archival paraffin- embedded tissues: development and performance of a 92-gene reverse transcriptase- polymerase chain reaction assay. The American journal of pathology. 2004;164(1):35-42.

25. Gjerdrum LM, Sorensen BS, Kjeldsen E, Sorensen FB, Nexo E, Hamilton-Dutoit S. Real-time quantitative PCR of microdissected paraffin-embedded breast carcinoma: an alternative method for HER-2/neu analysis. The Journal of molecular diagnostics : JMD. 2004;6(1):42-51.

26. Moelans CB, de Weger RA, van Blokland MT, et al. HER-2/neu amplification testing in breast cancer by multiplex ligation-dependent probe amplification in comparison with immunohistochemistry and in situ hybridization. Cellular oncology : the official journal of the International Society for Cellular Oncology. 2009;31(1):1-10.

- 88 -

27. Slamon DJ, deKernion JB, Verma IM, Cline MJ. Expression of cellular oncogenes in human malignancies. Science. 1984;224(4646):256-62.

28. Streit S, Michalski CW, Erkan M, Kleeff J, Friess H. Northern blot analysis for detection and quantification of RNA in pancreatic cancer cells and tissues. Nature protocols. 2009;4(1):37-43.

29. Hudis CA. Trastuzumab--mechanism of action and use in clinical practice. The New England journal of medicine. 2007;357(1):39-51.

30. Codony-Servat J, Albanell J, Lopez-Talavera JC, Arribas J, Baselga J. Cleavage of the HER2 ectodomain is a pervanadate-activable process that is inhibited by the tissue inhibitor of metalloproteases-1 in breast cancer cells. Cancer research. 1999;59(6):1196-201.

31. Sanderson MP, Dempsey PJ, Dunbar AJ. Control of ErbB signaling through metalloprotease mediated ectodomain shedding of EGF-like factors. Growth factors. 2006;24(2):121-36.

32. Yeh IT. Measuring HER-2 in breast cancer. Immunohistochemistry, FISH, or ELISA? American journal of clinical pathology. 2002;117 Suppl:S26-35.

33. Bramwell VH, Doig GS, Tuck AB, et al. Changes over time of extracellular domain of HER2 (ECD/HER2) serum levels have prognostic value in metastatic breast cancer. Breast cancer research and treatment. 2009;114(3):503-11.

34. Fornier MN, Seidman AD, Schwartz MK, et al. Serum HER2 extracellular domain in metastatic breast cancer patients treated with weekly trastuzumab and paclitaxel: association with HER2 status by immunohistochemistry and fluorescence in situ hybridization and with response rate. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2005;16(2):234-9.

35. Ludovini V, Gori S, Colozza M, et al. Evaluation of serum HER2 extracellular domain in early breast cancer patients: correlation with clinicopathological parameters and survival. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2008;19(5):883-90.

- 89 -

36. Schwartz MK, Smith C, Schwartz DC, Dnistrian A, Neiman I. Monitoring therapy by serum HER-2/neu. The International journal of biological markers. 2000;15(4):324-9.

37. Sias PE, Kotts CE, Vetterlein D, Shepard M, Wong WL. ELISA for quantitation of the extracellular domain of p185HER2 in biological fluids. Journal of immunological methods. 1990;132(1):73-80.

38. Esteva FJ, Cheli CD, Fritsche H, et al. Clinical utility of serum HER2/neu in monitoring and prediction of progression-free survival in metastatic breast cancer patients treated with trastuzumab-based therapies. Breast cancer research : BCR. 2005;7(4):R436-43.

39. Carney WP, Leitzel K, Ali S, Neumann R, Lipton A. HER-2/neu diagnostics in breast cancer. Breast cancer research : BCR. 2007;9(3):207.

40. Molina R, Jo J, Filella X, et al. C-erbB-2 oncoprotein in the sera and tissue of patients with breast cancer. Utility in prognosis. Anticancer research. 1996;16(4B):2295-300.

41. Sorensen PD, Jakobsen EH, Langkjer ST, et al. Serum HER-2 concentrations for monitoring women with breast cancer in a routine oncology setting. Clinical chemistry and laboratory medicine : CCLM / FESCC. 2009;47(9):1117-23.

42. Lipton A, Ali SM, Leitzel K, et al. Elevated serum Her-2/neu level predicts decreased response to hormone therapy in metastatic breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2002;20(6):1467-72.

43. Kong Y, Dai S, Xie X, et al. High serum HER2 extracellular domain levels: correlation with a worse disease-free survival and overall survival in primary operable breast cancer patients. Journal of cancer research and clinical oncology. 2012;138(2):275-84.

44. Sandri MT, Johansson H, Colleoni M, et al. Serum levels of HER2 ECD can determine the response rate to low dose oral cyclophosphamide and methotrexate in patients with advanced stage breast carcinoma. Anticancer research. 2004;24(2C):1261-6.

45. Ali SM, Carney WP, Esteva FJ, et al. Serum HER-2/neu and relative resistance to trastuzumab-based therapy in patients with metastatic breast cancer. Cancer. 2008;113(6):1294-301.

- 90 -

46. Lennon S, Barton C, Banken L, et al. Utility of serum HER2 extracellular domain assessment in clinical decision making: pooled analysis of four trials of trastuzumab in metastatic breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2009;27(10):1685-93.

47. Lam L, McAndrew N, Yee M, Fu T, Tchou JC, Zhang H. Challenges in the clinical utility of the serum test for HER2 ECD. Biochimica et biophysica acta. 2012;1826(1):199-208.

48. Moelans CB, de Weger RA, Van der Wall E, van Diest PJ. Current technologies for HER2 testing in breast cancer. Critical reviews in oncology/hematology. 2011;80(3):380-92.

49. Hayashi N, Nakamura S, Tokuda Y, et al. Serum HER2 levels determined by two methods in patients with metastatic breast cancer. International journal of clinical oncology. 2012;17(1):55-62.

50. Becker KF, Schott C, Hipp S, et al. Quantitative protein analysis from formalin-fixed tissues: implications for translational clinical research and nanoscale molecular diagnosis. The Journal of pathology. 2007;211(3):370-8.

51. Borg A, Tandon AK, Sigurdsson H, et al. HER-2/neu amplification predicts poor survival in node-positive breast cancer. Cancer research. 1990;50(14):4332-7.

52. Tandon AK, Clark GM, Chamness GC, Ullrich A, McGuire WL. HER-2/neu oncogene protein and prognosis in breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 1989;7(8):1120-8.

53. Ross JS, Fletcher JA. The HER-2/neu Oncogene in Breast Cancer: Prognostic Factor, Predictive Factor, and Target for Therapy. The oncologist. 1998;3(4):237-52.

54. Baehner FL, Achacoso N, Maddala T, et al. Human epidermal growth factor receptor 2 assessment in a case-control study: comparison of fluorescence in situ hybridization and quantitative reverse transcription polymerase chain reaction performed by central laboratories. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2010;28(28):4300-6.

55. Benohr P, Henkel V, Speer R, et al. Her-2/neu expression in breast cancer--A comparison of different diagnostic methods. Anticancer research. 2005;25(3B):1895-900.

- 91 -

56. Capizzi E, Gruppioni E, Grigioni AD, et al. Real time RT-PCR approach for the evaluation of ERBB2 overexpression in breast cancer archival samples: a comparative study with FISH, SISH, and immunohistochemistry. Diagnostic molecular pathology : the American journal of surgical pathology, part B. 2008;17(4):220-6.

57. Dabbs DJ, Klein ME, Mohsin SK, Tubbs RR, Shuai Y, Bhargava R. High false-negative rate of HER2 quantitative reverse transcription polymerase chain reaction of the Oncotype DX test: an independent quality assurance study. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2011;29(32):4279-85.

58. Lehmann-Che J, Amira-Bouhidel F, Turpin E, et al. Immunohistochemical and molecular analyses of HER2 status in breast cancers are highly concordant and complementary approaches. British journal of cancer. 2011;104(11):1739-46.

59. Merkelbach-Bruse S, Wardelmann E, Behrens P, Losen I, Buettner R, Friedrichs N. Current diagnostic methods of HER-2/neu detection in breast cancer with special regard to real-time PCR. The American journal of surgical pathology. 2003;27(12):1565-70.

60. Nistor A, Watson PH, Pettigrew N, Tabiti K, Dawson A, Myal Y. Real-time PCR complements immunohistochemistry in the determination of HER-2/neu status in breast cancer. BMC clinical pathology. 2006;6:2.

61. Noske A, Loibl S, Darb-Esfahani S, et al. Comparison of different approaches for assessment of HER2 expression on protein and mRNA level: prediction of chemotherapy response in the neoadjuvant GeparTrio trial (NCT00544765). Breast cancer research and treatment. 2011;126(1):109-17.

62. Susini T, Bussani C, Marini G, et al. Preoperative assessment of HER-2/neu status in breast carcinoma: the role of quantitative real-time PCR on core-biopsy specimens. Gynecologic oncology. 2010;116(2):234-9.

63. Vanden Bempt I, Van Loo P, Drijkoningen M, et al. Polysomy 17 in breast cancer: clinicopathologic significance and impact on HER-2 testing. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2008;26(30):4869-74.

64. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods. 2001;25(4):402-8.

- 92 -

65. Cronin M, Sangli C, Liu ML, et al. Analytical validation of the Oncotype DX genomic diagnostic test for recurrence prognosis and therapeutic response prediction in node-negative, estrogen receptor-positive breast cancer. Clinical chemistry. 2007;53(6):1084-91.

66. Jacquemier J, Spyratos F, Esterni B, et al. SISH/CISH or qPCR as alternative techniques to FISH for determination of HER2 amplification status on breast tumors core needle biopsies: a multicenter experience based on 840 cases. BMC cancer. 2013;13:351.

67. Huggett J, Dheda K, Bustin S, Zumla A. Real-time RT-PCR normalisation; strategies and considerations. Genes and immunity. 2005;6(4):279-84.

68. Sellner LN, Taylor GR. MLPA and MAPH: new techniques for detection of gene deletions. Human mutation. 2004;23(5):413-9.

69. Sapino A, Goia M, Recupero D, Marchio C. Current Challenges for HER2 Testing in Diagnostic Pathology: State of the Art and Controversial Issues. Frontiers in oncology. 2013;3:129.

70. Varshney D, Zhou YY, Geller SA, Alsabeh R. Determination of HER-2 status and chromosome 17 polysomy in breast carcinomas comparing HercepTest and PathVysion FISH assay. American journal of clinical pathology. 2004;121(1):70-7.

71. Yaziji H, Goldstein LC, Barry TS, et al. HER-2 testing in breast cancer using parallel tissue-based methods. JAMA : the journal of the American Medical Association. 2004;291(16):1972-7.

72. Hanna W, Kahn HJ, Trudeau M. Evaluation of HER-2/neu (erbB-2) status in breast cancer: from bench to bedside. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 1999;12(8):827-34.

73. Gupta D, Middleton LP, Whitaker MJ, Abrams J. Comparison of fluorescence and chromogenic in situ hybridization for detection of HER-2/neu oncogene in breast cancer. American journal of clinical pathology. 2003;119(3):381-7.

74. Jacobs TW, Gown AM, Yaziji H, Barnes MJ, Schnitt SJ. Specificity of HercepTest in determining HER-2/neu status of breast cancers using the United States Food and Drug

- 93 -

Administration-approved scoring system. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 1999;17(7):1983-7.

75. Ross JS, Fletcher JA. HER-2/neu (c-erb-B2) gene and protein in breast cancer. American journal of clinical pathology. 1999;112(1 Suppl 1):S53-67.

76. Yildiz-Aktas IZ, Dabbs DJ, Bhargava R. The effect of cold ischemic time on the immunohistochemical evaluation of estrogen receptor, progesterone receptor, and HER2 expression in invasive breast carcinoma. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2012;25(8):1098-105.

77. Middleton LP, Price KM, Puig P, et al. Implementation of American Society of Clinical Oncology/College of American Pathologists HER2 Guideline Recommendations in a tertiary care facility increases HER2 immunohistochemistry and fluorescence in situ hybridization concordance and decreases the number of inconclusive cases. Archives of pathology & laboratory medicine. 2009;133(5):775-80.

78. Penault-Llorca F, Adelaide J, Houvenaeghel G, Hassoun J, Birnbaum D, Jacquemier J. Optimization of immunohistochemical detection of ERBB2 in human breast cancer: impact of fixation. The Journal of pathology. 1994;173(1):65-75.

79. Hicks DG, Schiffhauer L. Standardized assessment of the HER2 status in breast cancer by immunohistochemistry. Lab Med. 2011;42:459-67.

80. Lebeau A, Deimling D, Kaltz C, et al. Her-2/neu analysis in archival tissue samples of human breast cancer: comparison of immunohistochemistry and fluorescence in situ hybridization. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2001;19(2):354-63.

81. Press MF, Hung G, Godolphin W, Slamon DJ. Sensitivity of HER-2/neu antibodies in archival tissue samples: potential source of error in immunohistochemical studies of oncogene expression. Cancer research. 1994;54(10):2771-7.

82. Sapino A, Coccorullo Z, Cassoni P, et al. Which breast carcinomas need HER-2/neu gene study after immunohistochemical analysis? Results of combined use of antibodies against different c-erbB2 protein domains. Histopathology. 2003;43(4):354-62.

- 94 -

83. Thomson TA, Hayes MM, Spinelli JJ, et al. HER-2/neu in breast cancer: interobserver variability and performance of immunohistochemistry with 4 antibodies compared with fluorescent in situ hybridization. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2001;14(11):1079-86.

84. Tsuda H, Sasano H, Akiyama F, et al. Evaluation of interobserver agreement in scoring immunohistochemical results of HER-2/neu (c-erbB-2) expression detected by HercepTest, Nichirei polyclonal antibody, CB11 and TAB250 in breast carcinoma. Pathology international. 2002;52(2):126-34.

85. Zhao J, Wu R, Au A, Marquez A, Yu Y, Shi Z. Determination of HER2 gene amplification by chromogenic in situ hybridization (CISH) in archival breast carcinoma. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2002;15(6):657-65.

86. Tubbs RR, Pettay JD, Roche PC, Stoler MH, Jenkins RB, Grogan TM. Discrepancies in clinical laboratory testing of eligibility for trastuzumab therapy: apparent immunohistochemical false-positives do not get the message. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2001;19(10):2714-21.

87. Hoang MP, Sahin AA, Ordonez NG, Sneige N. HER-2/neu gene amplification compared with HER-2/neu protein overexpression and interobserver reproducibility in invasive breast carcinoma. American journal of clinical pathology. 2000;113(6):852-9.

88. Bankfalvi A, Boecker W, Reiner A. Comparison of automated and manual determination of HER2 status in breast cancer for diagnostic use: a comparative methodological study using the Ventana BenchMark automated staining system and manual tests. International journal of oncology. 2004;25(4):929-35.

89. Bloom K, Harrington D. Enhanced accuracy and reliability of HER-2/neu immunohistochemical scoring using digital microscopy. American journal of clinical pathology. 2004;121(5):620-30.

90. Tawfik OW, Kimler BF, Davis M, et al. Comparison of immunohistochemistry by automated cellular imaging system (ACIS) versus fluorescence in-situ hybridization in the evaluation of HER-2/neu expression in primary breast carcinoma. Histopathology. 2006;48(3):258-67.

- 95 -

91. Brugmann A, Eld M, Lelkaitis G, et al. Digital image analysis of membrane connectivity is a robust measure of HER2 immunostains. Breast cancer research and treatment. 2012;132(1):41-9.

92. Hicks DG, Tubbs RR. Assessment of the HER2 status in breast cancer by fluorescence in situ hybridization: a technical review with interpretive guidelines. Human pathology. 2005;36(3):250-61.

93. Ross JS, Slodkowska EA, Symmans WF, Pusztai L, Ravdin PM, Hortobagyi GN. The HER-2 receptor and breast cancer: ten years of targeted anti-HER-2 therapy and personalized medicine. The oncologist. 2009;14(4):320-68.

94. Kovacs A, Stenman G. HER2-testing in 538 consecutive breast cancer cases using FISH and immunohistochemistry. Pathology, research and practice. 2010;206(1):39-42.

95. Varga Z, Noske A, Ramach C, Padberg B, Moch H. Assessment of HER2 status in breast cancer: overall positivity rate and accuracy by fluorescence in situ hybridization and immunohistochemistry in a single institution over 12 years: a quality control study. BMC cancer. 2013;13:615.

96. Bhargava R, Lal P, Chen B. Chromogenic in situ hybridization for the detection of HER- 2/neu gene amplification in breast cancer with an emphasis on tumors with borderline and low- level amplification: does it measure up to fluorescence in situ hybridization? American journal of clinical pathology. 2005;123(2):237-43.

97. Bartlett JM, Going JJ, Mallon EA, et al. Evaluating HER2 amplification and overexpression in breast cancer. The Journal of pathology. 2001;195(4):422-8.

98. Paik S, Bryant J, Tan-Chiu E, et al. Real-world performance of HER2 testing--National Surgical Adjuvant Breast and Bowel Project experience. Journal of the National Cancer Institute. 2002;94(11):852-4.

99. Persons DL, Borelli KA, Hsu PH. Quantitation of HER-2/neu and c-myc gene amplification in breast carcinoma using fluorescence in situ hybridization. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 1997;10(7):720-7.

- 96 -

100. Hicks DG, Kulkarni S. HER2+ breast cancer: review of biologic relevance and optimal use of diagnostic tools. American journal of clinical pathology. 2008;129(2):263-73.

101. Marchio C, Lambros MB, Gugliotta P, et al. Does chromosome 17 centromere copy number predict polysomy in breast cancer? A fluorescence in situ hybridization and microarray-based CGH analysis. The Journal of pathology. 2009;219(1):16-24.

102. Moelans CB, de Weger RA, van Diest PJ. Absence of chromosome 17 polysomy in breast cancer: analysis by CEP17 chromogenic in situ hybridization and multiplex ligation- dependent probe amplification. Breast cancer research and treatment. 2010;120(1):1-7.

103. Yeh IT, Martin MA, Robetorye RS, et al. Clinical validation of an array CGH test for HER2 status in breast cancer reveals that polysomy 17 is a rare event. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2009;22(9):1169-75.

104. Hanna WM, Ruschoff J, Bilous M, et al. HER2 in situ hybridization in breast cancer: clinical implications of polysomy 17 and genetic heterogeneity. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2014;27(1):4-18.

105. Wang S, Hossein Saboorian M, Frenkel EP, et al. Aneusomy 17 in breast cancer: its role in HER-2/neu protein expression and implication for clinical assessment of HER-2/neu status. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2002;15(2):137-45.

106. Hanna WM, Kwok K. Chromogenic in-situ hybridization: a viable alternative to fluorescence in-situ hybridization in the HER2 testing algorithm. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2006;19(4):481-7.

107. Lal P, Salazar PA, Ladanyi M, Chen B. Impact of polysomy 17 on HER-2/neu immunohistochemistry in breast carcinomas without HER-2/neu gene amplification. The Journal of molecular diagnostics : JMD. 2003;5(3):155-9.

108. McCormick SR, Lillemoe TJ, Beneke J, Schrauth J, Reinartz J. HER2 assessment by immunohistochemical analysis and fluorescence in situ hybridization: comparison of HercepTest and PathVysion commercial assays. American journal of clinical pathology. 2002;117(6):935-43.

- 97 -

109. Dal Lago L, Durbecq V, Desmedt C, et al. Correction for chromosome-17 is critical for the determination of true Her-2/neu gene amplification status in breast cancer. Molecular cancer therapeutics. 2006;5(10):2572-9.

110. Vranic S, Teruya B, Repertinger S, Ulmer P, Hagenkord J, Gatalica Z. Assessment of HER2 gene status in breast carcinomas with polysomy of chromosome 17. Cancer. 2011;117(1):48-53.

111. Ridolfi RL, Jamehdor MR, Arber JM. HER-2/neu testing in breast carcinoma: a combined immunohistochemical and fluorescence in situ hybridization approach. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2000;13(8):866-73.

112. van de Vijver M, Bilous M, Hanna W, et al. Chromogenic in situ hybridisation for the assessment of HER2 status in breast cancer: an international validation ring study. Breast cancer research : BCR. 2007;9(5):R68.

113. Pedersen M, Rasmussen BB. The correlation between dual-color chromogenic in situ hybridization and fluorescence in situ hybridization in assessing HER2 gene amplification in breast cancer. Diagnostic molecular pathology : the American journal of surgical pathology, part B. 2009;18(2):96-102.

114. Dowsett M, Hanna WM, Kockx M, et al. Standardization of HER2 testing: results of an international proficiency-testing ring study. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2007;20(5):584-91.

115. Bilous M, Morey AL, Armes JE, et al. Assessing HER2 amplification in breast cancer: findings from the Australian In Situ Hybridization Program. Breast cancer research and treatment. 2012;134(2):617-24.

116. Kiyose S, Igarashi H, Nagura K, et al. Chromogenic in situ hybridization (CISH) to detect HER2 gene amplification in breast and gastric cancer: comparison with immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH). Pathology international. 2012;62(11):728-34.

- 98 -

117. Mansfield AS, Sukov WR, Eckel-Passow JE, et al. Comparison of fluorescence in situ hybridization (FISH) and dual-ISH (DISH) in the determination of HER2 status in breast cancer. American journal of clinical pathology. 2013;139(2):144-50.

118. Furrer D, Jacob S, Caron C, Sanschagrin F, Provencher L, Diorio C. Validation of a new classifier for the automated analysis of the human epidermal growth factor receptor 2 (HER2) gene amplification in breast cancer specimens. Diagnostic pathology. 2013;8:17.

119. Stevens R, Almanaseer I, Gonzalez M, et al. Analysis of HER2 gene amplification using an automated fluorescence in situ hybridization signal enumeration system. The Journal of molecular diagnostics : JMD. 2007;9(2):144-50.

120. Tubbs RR, Pettay JD, Swain E, et al. Automation of manual components and image quantification of direct dual label fluorescence in situ hybridization (FISH) for HER2 gene amplification: A feasibility study. Applied immunohistochemistry & molecular morphology : AIMM / official publication of the Society for Applied Immunohistochemistry. 2006;14(4):436- 40.

121. Penault-Llorca F, Bilous M, Dowsett M, et al. Emerging technologies for assessing HER2 amplification. American journal of clinical pathology. 2009;132(4):539-48.

122. Gong Y, Gilcrease M, Sneige N. Reliability of chromogenic in situ hybridization for detecting HER-2 gene status in breast cancer: comparison with fluorescence in situ hybridization and assessment of interobserver reproducibility. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2005;18(8):1015-21.

123. Isola J, Tanner M, Forsyth A, Cooke TG, Watters AD, Bartlett JM. Interlaboratory comparison of HER-2 oncogene amplification as detected by chromogenic and fluorescence in situ hybridization. Clinical cancer research : an official journal of the American Association for Cancer Research. 2004;10(14):4793-8.

124. Vocaturo A, Novelli F, Benevolo M, et al. Chromogenic in situ hybridization to detect HER-2/neu gene amplification in histological and ThinPrep-processed breast cancer fine- needle aspirates: a sensitive and practical method in the trastuzumab era. The oncologist. 2006;11(8):878-86.

- 99 -

125. Tanner M, Gancberg D, Di Leo A, et al. Chromogenic in situ hybridization: a practical alternative for fluorescence in situ hybridization to detect HER-2/neu oncogene amplification in archival breast cancer samples. The American journal of pathology. 2000;157(5):1467-72.

126. Garcia-Caballero T, Grabau D, Green AR, et al. Determination of HER2 amplification in primary breast cancer using dual-colour chromogenic in situ hybridization is comparable to fluorescence in situ hybridization: a European multicentre study involving 168 specimens. Histopathology. 2010;56(4):472-80.

127. Mollerup J, Henriksen U, Muller S, Schonau A. Dual color chromogenic in situ hybridization for determination of HER2 status in breast cancer: a large comparative study to current state of the art fluorescence in situ hybridization. BMC clinical pathology. 2012;12:3.

128. Hwang CC, Pintye M, Chang LC, et al. Dual-colour chromogenic in-situ hybridization is a potential alternative to fluorescence in-situ hybridization in HER2 testing. Histopathology. 2011;59(5):984-92.

129. Laakso M, Tanner M, Isola J. Dual-colour chromogenic in situ hybridization for testing of HER-2 oncogene amplification in archival breast tumours. The Journal of pathology. 2006;210(1):3-9.

130. Kato M, Watabe K, Hamasaki T, et al. Association of low serum adiponectin levels with erosive esophagitis in men: an analysis of 2405 subjects undergoing physical check-ups. Journal of gastroenterology. 2011;46(12):1361-7.

131. Loring P, Cummins R, O'Grady A, Kay EW. HER2 positivity in breast carcinoma: a comparison of chromogenic in situ hybridization with fluorescence in situ hybridization in tissue microarrays, with targeted evaluation of intratumoral heterogeneity by in situ hybridization. Applied immunohistochemistry & molecular morphology : AIMM / official publication of the Society for Applied Immunohistochemistry. 2005;13(2):194-200.

132. Lambros MB, Natrajan R, Reis-Filho JS. Chromogenic and fluorescent in situ hybridization in breast cancer. Human pathology. 2007;38(8):1105-22.

133. Powell RD, Pettay JD, Powell WC, et al. Metallographic in situ hybridization. Human pathology. 2007;38(8):1145-59.

- 100 -

134. Bartlett JM, Campbell FM, Ibrahim M, et al. Chromogenic in situ hybridization: a multicenter study comparing silver in situ hybridization with FISH. American journal of clinical pathology. 2009;132(4):514-20.

135. Francis GD, Jones MA, Beadle GF, Stein SR. Bright-field in situ hybridization for HER2 gene amplification in breast cancer using tissue microarrays: correlation between chromogenic (CISH) and automated silver-enhanced (SISH) methods with patient outcome. Diagnostic molecular pathology : the American journal of surgical pathology, part B. 2009;18(2):88-95.

136. Fritzsche FR, Bode PK, Moch H, Kristiansen G, Varga Z, Bode B. Determination of the Her-2/neu gene amplification status in cytologic breast cancer specimens using automated silver-enhanced in-situ hybridization (SISH). The American journal of surgical pathology. 2010;34(8):1180-5.

137. Dietel M, Ellis IO, Hofler H, et al. Comparison of automated silver enhanced in situ hybridisation (SISH) and fluorescence ISH (FISH) for the validation of HER2 gene status in breast carcinoma according to the guidelines of the American Society of Clinical Oncology and the College of American Pathologists. Virchows Archiv : an international journal of pathology. 2007;451(1):19-25.

138. Shousha S, Peston D, Amo-Takyi B, Morgan M, Jasani B. Evaluation of automated silver-enhanced in situ hybridization (SISH) for detection of HER2 gene amplification in breast carcinoma excision and core biopsy specimens. Histopathology. 2009;54(2):248-53.

139. Papouchado BG, Myles J, Lloyd RV, et al. Silver in situ hybridization (SISH) for determination of HER2 gene status in breast carcinoma: comparison with FISH and assessment of interobserver reproducibility. The American journal of surgical pathology. 2010;34(6):767-76.

140. Dekker TJ, Borg ST, Hooijer GK, et al. Determining sensitivity and specificity of HER2 testing in breast cancer using a tissue micro-array approach. Breast cancer research : BCR. 2012;14(3):R93.

141. Park K, Han S, Kim JY, Kim HJ, Kwon JE, Gwak G. Silver-Enhanced In Situ Hybridization as an Alternative to Fluorescence In Situ Hybridization for Assaying HER2 Amplification in Clinical Breast Cancer. Journal of breast cancer. 2011;14(4):276-82.

- 101 -

142. Nitta H, Hauss-Wegrzyniak B, Lehrkamp M, et al. Development of automated brightfield double in situ hybridization (BDISH) application for HER2 gene and chromosome 17 centromere (CEN 17) for breast carcinomas and an assay performance comparison to manual dual color HER2 fluorescence in situ hybridization (FISH). Diagnostic pathology. 2008;3:41.

143. Schiavon BN, Jasani B, de Brot L, et al. Evaluation of reliability of FISH versus brightfield dual-probe in situ hybridization (BDISH) for frontline assessment of HER2 status in breast cancer samples in a community setting: influence of poor tissue preservation. The American journal of surgical pathology. 2012;36(10):1489-96.

144. Hainfeld JF, Powell RD. New frontiers in gold labeling. The journal of histochemistry and cytochemistry : official journal of the Histochemistry Society. 2000;48(4):471-80.

145. Tubbs R, Pettay J, Skacel M, et al. Gold-facilitated in situ hybridization: a bright-field autometallographic alternative to fluorescence in situ hybridization for detection of Her-2/neu gene amplification. The American journal of pathology. 2002;160(5):1589-95.

146. Tubbs R, Skacel M, Pettay J, et al. Interobserver interpretative reproducibility of GOLDFISH, a first generation gold-facilitated autometallographic bright field in situ hybridization assay for HER-2/neu amplification in invasive mammary carcinoma. The American journal of surgical pathology. 2002;26(7):908-13.

147. Tubbs R, Pettay J, Hicks D, et al. Novel bright field molecular morphology methods for detection of HER2 gene amplification. Journal of molecular histology. 2004;35(6):589-94.

148. Brugmann A, Lelkaitis G, Nielsen S, Jensen KG, Jensen V. Testing HER2 in breast cancer: a comparative study on BRISH, FISH, and IHC. Applied immunohistochemistry & molecular morphology : AIMM / official publication of the Society for Applied Immunohistochemistry. 2011;19(3):203-11.

149. Bartlett JM, Campbell FM, Ibrahim M, et al. A UK NEQAS ISH multicenter ring study using the Ventana HER2 dual-color ISH assay. American journal of clinical pathology. 2011;135(1):157-62.

150. Downs-Kelly E, Pettay J, Hicks D, et al. Analytical validation and interobserver reproducibility of EnzMet GenePro: a second-generation bright-field metallography assay for

- 102 -

concomitant detection of HER2 gene status and protein expression in invasive carcinoma of the breast. The American journal of surgical pathology. 2005;29(11):1505-11.

151. Ni R, Mulligan AM, Have C, O'Malley FP. PGDS, a novel technique combining chromogenic in situ hybridization and immunohistochemistry for the assessment of ErbB2 (HER2/neu) status in breast cancer. Applied immunohistochemistry & molecular morphology : AIMM / official publication of the Society for Applied Immunohistochemistry. 2007;15(3):316- 24.

152. Reisenbichler ES, Horton D, Rasco M, Andea A, Hameed O. Evaluation of dual immunohistochemistry and chromogenic in situ hybridization for HER2 on a single section. American journal of clinical pathology. 2012;137(1):102-10.

153. Nitta H, Kelly BD, Padilla M, et al. A gene-protein assay for human epidermal growth factor receptor 2 (HER2): brightfield tricolor visualization of HER2 protein, the HER2 gene, and chromosome 17 centromere (CEN17) in formalin-fixed, paraffin-embedded breast cancer tissue sections. Diagnostic pathology. 2012;7:60.

154. Bernet L, Martinez Benaclocha M, Castera C, et al. mRNA in situ hybridization (HistoSonda): a new diagnostic tool for HER2-status in breast cancer-a multicentric Spanish study. Diagnostic molecular pathology : the American journal of surgical pathology, part B. 2012;21(2):84-92.

155. Franchet C, Filleron T, Cayre A, et al. Instant-quality fluorescence in-situ hybridization as a new tool for HER2 testing in breast cancer: a comparative study. Histopathology. 2014;64(2):274-83.

156. Matthiesen SH, Hansen CM. Fast and non-toxic in situ hybridization without blocking of repetitive sequences. PloS one. 2012;7(7):e40675.

157. Ohlschlegel C, Kradolfer D, Hell M, Jochum W. Comparison of automated and manual FISH for evaluation of HER2 gene status on breast carcinoma core biopsies. BMC clinical pathology. 2013;13:13.

158. Yoon N, Do IG, Cho EY. Analysis of HER2 status in breast carcinoma by fully automated HER2 fluorescence in situ hybridization (FISH): comparison of two

- 103 -

immunohistochemical tests and manual FISH. APMIS : acta pathologica, microbiologica, et immunologica Scandinavica. 2013.

159. Hofmann M, Stoss O, Gaiser T, et al. Central HER2 IHC and FISH analysis in a trastuzumab (Herceptin) phase II monotherapy study: assessment of test sensitivity and impact of chromosome 17 polysomy. Journal of clinical pathology. 2008;61(1):89-94.

160. Kim JW, Kim JH, Im SA, et al. HER2/CEP17 ratio and HER2 immunohistochemistry predict clinical outcome after first-line trastuzumab plus taxane chemotherapy in patients with HER2 fluorescence in situ hybridization-positive metastatic breast cancer. Cancer chemotherapy and pharmacology. 2013;72(1):109-15.

161. Bartlett JM, Starczynski J. Quantitative reverse transcriptase polymerase chain reaction and the Oncotype DX test for assessment of human epidermal growth factor receptor 2 status: time to reflect again? Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2011;29(32):4219-21.

- 104 -

Table 2.1. Main characteristics of the described techniques

Method of analysis Target Specimen type FDA approval Southern blot DNA Frozen tissue No Northern blot mRNA Frozen tissue No ELISA Protein Blood sample Yes Western blot Protein Frozen tissue and FFPE tissue No PCR-based assays qPCR: DNA Frozen tissue and FFPE tissue No RT-PCR: mRNA MLPA DNA FFPE tissue No IHC Protein FFPE tissue Yes FISH DNA FFPE tissue Yes CISH DNA FFPE tissue Yes SISH DNA FFPE tissue No GOLDFISH DNA FFPE tissue No BDISH DNA FFPE tissue Yes HER2-gene protein assay Protein and DNA FFPE tissue No mRNA in situ hybridization mRNA FFPE tissue No IQFISH DNA FFPE tissue Yes Automated HER2 FISH assay DNA FFPE tissue No

Abbreviations: FDA: Food and Drug Administration; ELISA: Enzyme-linked immunosorbent assay; PCR: Polymerase-chain reaction; qPCR: quantitative PCR; RT-PCR: reverse transcription PCR; MLPA: Multiplex ligation-dependent probe amplification; IHC: Immunohistochemistry; FFPE: Formalin-fixed, paraffin embedded; FISH: Fluorescence in situ hybridization; CISH: Chromogenic in situ hybridization; SISH: Silver-enhanced in situ hybridization; GOLDFISH: Gold-facilitated autometallographic in situ hybridization; BDISH: Bright-field double in situ hybridization; HER2: Human epidermal growth factor receptor 2; IQFISH: Instant-quality fluorescence in situ hybridization.

- 105 -

Chapter 3: Contextualisation, hypothesis and objectives

3.1. Contextualisation HER2 is a molecular predictive marker in breast cancer. With the introduction of targeted therapies directed against this receptor, the accurate evaluation of HER2 status has become crucial for identification of patients who would most likely benefit from treatment with HER2 inhibitors. In 2009, a public health scandal struck the province of Quebec, since some HER2 results were erroneously classified. Erroneous HER2 status classification has an enormous impact on the clinical management of breast cancer patients: while false-negative results would deny the access of patients to the potential benefits of targeted treatment, false- positive results would expose patients to the potential cardiotoxic side effects of these expensive drugs without experiencing any therapeutic advantages. Part of this study was therefore established with the aim to improve accuracy of HER2 testing in breast cancer specimens.

Although the implementation of targeted anti-HER2 therapies such as trastuzumab has considerably increased the survival of HER2-positive breast cancer patients, a subgroup of HER2-positive breast cancer patients experience tumor recurrence and progression. The aim of this study was also therefore to identify genetic, epigenetic and lifestyle factors that could influence trastuzumab response in HER2-positive breast cancer patients.

3.2. Hypothesis As each technique has its own advantages and disadvantages, currently there is still no consensus on which method is the best for determining the HER2 status in breast cancer specimens. Although FISH is considered a more quantitative technique compared to IHC, the FISH assay is three times more expensive than immunohistochemical analysis. Since the TMA technique reduces the reagent amount required for staining, we propose that the evaluation of HER2 gene amplification on TMA section represents an economical and reliable method for the evaluation of HER2 status in breast cancer specimens. Considering that both IHC and FISH are two FDA-approved techniques for the evaluation of HER2 status in breast cancer specimens, we anticipate high concordance rate between HER2 status determined by IHC and FISH on TMA tissue section.

- 106 -

There have been major changes regarding the specimen type to use for the determination of HER2 status in breast cancer specimens between the 2007 and the 2013 ASCO/CAP guidelines. Whereas the 2007 ASCO/CAP guidelines recommended performing HER2 testing on resection specimens, the 2013 ASCO/CAP guidelines recommend performing an initial HER2 test on core biopsy. We postulate that HER2 status obtained using the paraffin block used for diagnostic purposes is similar to that obtained using a randomly chosen additional block (a proxy of a random core biopsy).

It has been reported that HER2 SNPs, especially Ile655Val SNP (the Ile to Val substitution at codon 655 promotes the activity of the tyrosine kinase domain), have an impact on HER2 function. We hypothesize that Ile655Val and Ala1170Pro SNPs are associated with breast cancer prognostic factors. We also postulate that Ile655Val SNP might influence response to trastuzumab.

Methylation patterns in the breast cancer tissues of HER2-positive breast cancer patients are heterogeneous. Since trastuzumab response in HER2-positive breast cancer patients is also heterogeneous, we hypothesize that methylation pattern in tumor tissues of patients that respond to trastuzumab treatment is different to that of patients that develop resistance to this drug. The identification of genes that are differentially methylated between these two groups could lead to the identification of genes that play a role in the response to trastuzumab treatment and therefore to the identification of novel biomarkers with therapeutic significance.

It has been observed that lifestyle factors including tobacco consumption and to a certain extent alcohol use may influence survival of breast cancer patients. Evidence reported in the literature suggest the existence of molecular and epidemiological links between tobacco and ethanol exposure and HER2. Considered that NNK, a tobacco-specific carcinogen, activates the ERK/MAPK signaling pathway in human normal mammary epithelial cells and that recurrence risk was significantly increased in trastuzumab-naïve HER2-positive breast cancer patients who smoked at time of diagnosis, we postulate that tobacco consumption might negatively impact the response to trastuzumab in HER2-positive breast cancer patients. Given that alcohol consumption has been reported to be a risk factor for HER2- positive breast cancer and that the stimulatory effect of ethanol on the invasion capacity of breast cancer cell lines depends on the expression levels of HER2, we propose that alcohol use might also negatively influence trastuzumab response.

- 107 -

3.3. Objectives The aim of the work presented in this thesis was twofold. First, we aimed to identify the most reliable method for the evaluation of HER2 status in breast cancer specimens. Second, we aimed to identify factors that could have an impact on trastuzumab response. Among all factors that could influence patient’s response to trastuzumab, we decided to focus on HER2 polymorphisms and on methylation pattern, two aspects that are feasible to examine in a clinical context and that are easily measurable in FFPE breast cancer specimens. Since lifestyle factors might influence survival of breast cancer patients, we decided to evaluate the impact of tobacco and alcohol consumption on response to trastuzumab.

Specifically, the study objectives included:

1. To compare HER2 gene amplification status determined by FISH on whole tissue section and on TMA section in breast cancer specimens. We also compared concordance between HER2 gene amplification from the whole tissue section and a TMA section from a randomly chosen additional block (a proxy of the core biopsy). Results of this objective are presented in Chapter 4.

2. The initial goal of objective 2 was to examine the concordance between HER2 status determined by IHC, FISH and a dual-color ISH assay in a representative cohort of breast cancer patients. Although we sent our TMA sections to several laboratories with dual-color ISH assay (INFORM HER2 Dual ISH DNA probe cocktail assay) experience, we were unfortunately unable to obtain evaluable HER2 status using this type of assay. In the summer of 2013, we sent the TMA sections to a laboratory in Montreal. Staining of our TMA slides was clearly unsatisfactory (red CEP17 signals were completely absent). Our TMA slides were then sent to the Applications Support Lab of Ventana Medical Systems (Tucson, Arizona, USA). Although satisfactory staining on validation slides was achieved, staining was not evaluable in the majority of our specimens (CEP17 signals were very pale). Therefore, in the present work, we exclusively analysed the concordance between HER2 status determined by IHC and FISH on TMA sections. We also evaluated concordance between HER2 status determined by IHC and FISH on TMA sections from the block used for diagnostic and a randomly chosen additional block (a proxy of the core biopsy). Results of this research are presented in Chapter 5.

- 108 -

3. To validate the clinical performance of a new software programming algorithm that analyses fluorescent signals in single tumor cell nuclei within breast cancer tissue sections. Results of this validation study are presented in Chapter 6.

4. To evaluate the association between HER2 polymorphisms and breast cancer prognostic factors. Results of this research question are presented in Chapter 7.

5. To analyze the association between HER2 polymorphisms, tobacco and alcohol consumption and the response to trastuzumab treatment in HER2-positive breast cancer patients. Results of this objective are presented in Chapter 8.

6. To analyze the association between methylation patterns and the response to trastuzumab in HER2-positive breast cancer patients. Results of this research question are presented in Chapter 9.

- 109 -

Chapter 4: Tissue microarray is a reliable tool for the evaluation of HER2 amplification in breast cancer

Anticancer Research 2016; 36(9): 4661-6

Daniela Furrer1,2,3, Simon Jacob4,5,6, Chantal Caron4,5, François Sanschagrin2,4, Louise Provencher2,4,7 and Caroline Diorio1,2,3,4

1Cancer Research Centre at Laval University; 2Oncology Axis, CHU of Quebec Research Center; and Departments of 3Social and Preventive Medicine, 6Molecular Biology, Medical Biochemistry and Pathology, and 7Surgery, Faculty of Medicine, Laval University, Quebec City, QC, Canada; 4Deschênes-Fabia Center for Breast Diseases and 5Pathology Service, Saint-Sacrement Hospital, Quebec City, QC, Canada

Correspondence to: Associate Professor Caroline Diorio, Ph.D., Département de médecine sociale et préventive, Université Laval, Axe oncologie, Centre de recherche du CHU de Québec, Centre des maladies du sein Deschênes-Fabia, Hôpital du Saint-Sacrement, 1050, chemin Ste-Foy, local J0-16, Québec (Qc) G1S 4L8, Canada. Tel.: +1 4186827511, extension 84726, Fax.: +1 4186827949. E-mail: [email protected]

Keywords: Breast cancer, trastuzumab, TMA, ASCO/CAP guidelines, HER2 status, FISH.

- 110 -

Résumé Nous avons examiné une méthode économique pour évaluer l’amplification du gène qui code pour le récepteur 2 du facteur de croissance épidermique humain (HER2) dans des échantillons de cancer du sein.

Nous avons comparé l’amplification d’HER2 déterminée par hybridation fluorescente in situ (FISH) sur lame complète (LC) à partir du bloc ayant servi pour le diagnostic et sur des sections de matrice tissulaire (tissue microarray, TMA) dans une cohorte consécutive de 521 patientes atteintes d’un cancer du sein. Dans un sous-groupe de 116 patientes, nous avons examiné la concordance du statut HER2 déterminé sur lame complète et une section TMA obtenue à partir d’un bloc qui a été choisi aléatoirement (un proxy de la biopsie).

La concordance générale entre l’amplification du gène HER2 obtenue sur LC et sur section TMA était de 98,2%, et entre LC et la section TMA du bloc additionnel était de 99,0%.

Les taux de concordance élevés appuient l’utilisation des sections TMA pour l’évaluation de l’amplification d’HER2 dans le cancer du sein et suggèrent que le FISH peut être utilisé pour déterminer le statut HER2 sur biopsie.

- 111 -

Abstract Aim: We examined an economical method for evaluating the amplification of the human epidermal growth factor receptor 2 (HER2) gene in breast cancer specimens.

Materials and Methods: We compared HER2 amplification determined by fluorescence in situ hybridization (FISH) on whole tissue (WT) blocks used for diagnostic and on tissue microarray (TMA) sections for a cohort of 521 consecutive patients with breast cancer. In a subset of 116 patients, we examined HER2 concordance from the WT section and a TMA section from a randomly chosen additional block (a proxy of the core biopsy).

Results: Overall concordance for HER2 amplification between WT and TMA sections was 98.2%, and between sections from WT and from the additional block was 99.0%.

Conclusion: The high concordance rates support the use of TMA for the evaluation of HER2 amplification in breast cancer and suggest that FISH can be used to assess HER2 using core biopsies.

- 112 -

Introduction The human epidermal growth factor receptor 2 (HER2) gene is located on chromosome 17 and encodes a transmembrane tyrosine kinase receptor protein (1). HER2 gene amplification and receptor overexpression, which occur in 15 to 20% of patients with breast cancer, are important markers for poor prognosis, including a more aggressive disease and a shorter survival (2). Moreover, HER2-positive status is considered a predictive marker of response to anti-HER2 therapies (3). Given its prognostic, predictive and therefore therapeutic implications, accurate diagnostic assessment of HER2 is essential for reliable identification of patients eligible for HER2-targeted drugs.

Currently, there are several Food and Drug (FDA)-approved methods for evaluating HER2 status, including immunohistochemical (IHC) evaluation of HER2 protein expression and assessment of HER2 gene amplification by in situ hybridization (ISH), most commonly fluorescent ISH (FISH) (4, 5). Since both IHC and FISH present advantages and disadvantages, there is still no consensus on which method is superior for assessing the HER2 status in breast cancer specimens (6). In 2013, the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) updated the guidelines to clarify the recommendations for HER2 testing in breast cancer specimens published in 2007 (4). In addition to report new scoring criteria, the updated guidelines recommend performing an initial test of a core biopsy. If test results are equivocal, reflex testing on tumor specimen section with an alternative assay should be carried out. In addition, repeat testing should be performed if there is an apparent histopathological discordance with the test result. The 2007 ASCO/CAP guidelines recommended to perform HER2 testing on resection specimens and to retest when results were equivocal (5). Although FISH is a reliable method for the evaluation of HER2 status in breast cancer for diagnostic purposes or in the framework of quality control, it is a very expensive technique (7).

Tissue microarray (TMA) allows for simultaneous molecular analysis of large numbers of samples by means of arranging tissue cores from multiple specimens into an empty paraffin block (8, 9). A low degree of representativity compared to whole tissue sections, however, represents the major limitation of this technology. The purpose of this project was to determine the most reliable and economical method for evaluating HER2 status in breast cancer specimens. In a cohort of 521 consecutive patients with breast cancer, we aimed to evaluate if FISH analysis performed on TMA section provides comparable results to those obtained using whole tissue sections used for diagnostic

- 113 -

purposes (one section per slide). Furthermore, we wanted to assess whether FISH analysis carried out on a randomly chosen additional block (a proxy of a random core biopsy) yielded similar results. In a subset of 116 breast cancer specimens, we aimed to analyze the concordance between HER2 gene amplification status obtained from the same whole tissue section used for diagnostics and that obtained from the randomly chosen additional block using TMA.

Materials and Methods

Specimen collection and patient population Specimens were obtained from mastectomies and segmental resections that were performed at the Centre des Maladies du Sein Deschênes-Fabia at the Saint-Sacrement Hospital in Québec, Québec, Canada, between February 2011 and April 2012. Samples were fixed with 10% neutral buffered formalin, embedded in paraffin, cut into 4-μm tissue sections, stained with hematoxylin and eosin (H&E), and used for routine pathological evaluation. The study population consisted of 554 consecutive cases of invasive breast carcinoma, with a tumor size on histological slides of at least ̴ 1 cm, which did not receive chemotherapy prior to surgery. All eligible women provided their written informed consent for use of their tissue. Ethical approval of the study was obtained in 2010 from the Research Ethics Committee of the Centre de Recherche du CHU de Québec (#DR-002-1286).

Tissue microarray construction and processing TMAs were constructed as previously described (10). For the 554 consecutive breast cancer specimens, the most representative tissue block from each case was chosen by the pathologist for HER2 evaluation by IHC and FISH. Subsequently, two tumoral regions showing the strongest IHC staining were delineated on the IHC slide by the pathologist. Within each delimited tumor zone, pathologists indicated where two 0.6 mm tissue cylinders were to be punched. The four tissue cylinders were punched using a manual arraying instrument (Beecher Instruments, Silver Spring, MD, USA) and were inserted into empty recipient paraffin blocks. We call these TMA blocks 'diagnostic TMA'. Among the 554 cases present on the diagnostic TMA, we identified 279 cases that had at least two additional paraffin blocks presenting the same histological type as the paraffin block that was used for the construction of the diagnostic TMA and having enough tumor tissue to extract four tissue cylinders. Among these 279 cases, 100 cases were randomly chosen. For each of these randomly selected cases, one paraffin block was randomly chosen among all eligible

- 114 -

paraffin blocks. Suitable blocks were identified using H&E section and two tumoral regions were identified without previous IHC staining on the H&E section by two trained technologists and verified by a pathologist. Four 0.6 mm tissue cylinders were then punched from these regions. Moreover, we performed an oversampling of HER2-positive cases: 16 cases that were scored as HER2-positive on diagnostic TMA section but that had not been selected at random were added to the 100 cases. Therefore, a total of 116 cases were inserted into empty recipient paraffin blocks. We call these TMA blocks 'random TMA' in order to differentiate it from the diagnostic TMA. On each array block, breast cancer cell lines (MCF-7, MDA-231 and SKBR-3) were included in duplicate and served as negative and positive controls. Four-micrometer TMA sections were processed by FISH following the same protocol as for the whole tissue section used for diagnostics. One section from each TMA block was stained with H&E for reference histology.

Fluorescence in situ hybridization HER2 gene copy number was evaluated using the FDA-approved PathVysion™ HER2 DNA Probe kit (Abbott Molecular, Des Plaines, IL, USA/Inter Medico, Markham, Canada). Fluorescent signals were analyzed with an epifluorescence microscope Axio Imager M1 (Zeiss, Göttingen, Germany), equipped with a triple-band filter (4’,6-diamidino-2- phenylindole (DAPI)/green/orange). Automated analysis of fluorescence signals was performed using the FDA-approved MetaSystemsTM image analysis system, equipped with Metafer software with extended focus/tile sampling (MetaSystems, Altussheim, Germany) (11). After selection of 5 to 10 non-overlapping fields of infiltrating carcinoma, field images were automatically captured and analyzed by the software. HER2 gene amplification was reported according to the 2007 and 2013 ASCO/CAP guidelines for the evaluation of HER2 status in breast cancer specimens (Table 4.1.) (4, 5). For equivocal cases, manual counting was performed from at least 60 nonoverlapping tumor cells from two distinct tumor areas. Equivocal cases were counted by two independent technologists and reviewed by the pathologist of the study. In addition, manual counting was performed in 40 nonoverlapping tumor cells when the average HER2 copy number per tile was ≥4.0 and ≤6.0 at the automated image analysis. Normal breast epithelial cells, lymphocytes and fibroblasts were used as internal control. Slides were analyzed by trained technologists, and all results were validated by breast pathologists with experience in FISH interpretation.

- 115 -

HER2 evaluation on TMA A core was considered unsatisfactory for analysis if it was absent, it contained no tumor tissue, or if tumor tissue occupied less than 10% of the total core area. Fluorescent signals on TMA slides were evaluated in the same way as for the whole tissue section. Fluorescent signals were first analyzed using an FDA-approved automated image analysis system (MetaSystems) (11). Slides were scanned at low magnification (×5) to generate a position list corresponding to each core in order to link the core location to subsequent high- resolution (×40) FISH images. Automated image analysis was followed by manual counting or visual verification using the same criteria as for whole tissue sections. Each informative core was scored separately in a blind fashion. Average FISH results of informative cores were considered. All results were validated by breast pathologists. HER2 FISH results were reported using the same scoring criteria as for whole tissue section analysis (4, 5).

Statistical analysis Only cases with at least one informative core were included in the analysis. FISH results on both whole tissue and TMA sections were available for 521 cases out of the 554 total cases (94.0%). FISH results on both whole tissue and random TMA sections were available for 103 cases out of the 116 selected cases (88.8%). In order to evaluate the agreement between FISH results obtained by the different methods (whole tissue section vs. diagnostic TMA section and whole tissue section vs. random TMA section), positive, negative and overall concordance were calculated. Positive concordance (sensitivity) was calculated as the number of samples positive for both methods divided by the number of samples positive by the reference FISH assay. Negative agreement (specificity) was calculated similarly. Overall concordance was defined as the combination of sensitivity and specificity, as recommended by the ASCO/CAP guidelines (4, 5). In addition, the level of concordance was measured using Cohen’s kappa test. All analyses were performed using SAS software (version 9.1.3, SAS Institute Inc., Cary, NC, USA).

Results Concordance of the FISH HER2 results between whole tissue and diagnostic TMA sections are summarized in Table 4.2. Using the 2013 ASCO⁄CAP scoring system, the overall agreement was 98.2% (kappa value = 0.81). The positive agreement and the negative agreements were 95.5%, and 98.6%, respectively. The one case that was evaluated as non- amplified on the whole tissue section but amplified on diagnostic TMA section showed

- 116 -

borderline amplification (mean HER2/CEP17 ratio 2.27 on diagnostic TMA section and HER2/CEP17 ratio of 1.80 on whole tissue section). Similarly, the three cases that were scored as amplified on whole tissue sections but non-amplified on diagnostic TMA sections showed borderline amplification (HER2/CEP17 ratios ranging from 2.11 to 2.31 on whole tissue sections, mean HER2/CEP17 ratios ranging from 1.37 to 1.85 on diagnostic TMA sections).

We also assessed the concordance rates between HER2 gene amplification status determined by FISH on whole tissue sections and on TMA sections according to the number of informative cores per case (one or two vs. three or four cores). The overall, positive, and negative agreements for cases with one or two informative cores were 98.6%, 95.8% and 99.2%, respectively. For cases with three or four evaluable cores, the overall, positive and negative agreements were 97.9%, 95.3% and 98.4%, respectively.

Similar concordance rates between the two methods were observed when the 2007 ASCO/CAP guidelines were applied (Table 4.3. and 4.4.).

Concordance of HER2 status determined by FISH on whole tissue sections and random TMA sections is summarized in Table 4.5. The overall agreement was 99.0% (kappa value=0.94). The positive agreement and the negative agreements were 97.3%, and 100.0%, respectively. The case that was considered amplified on the whole tissue section and equivocal on random TMA section had a HER2/CEP17 ratio of 2.50 on the whole tissue section, while on random TMA section it had a mean HER2/CEP17 ratio of 1.67 and an average HER2 gene copy number of 4.53.

Discussion We observed a high concordance rate between the HER2 gene amplification status assessed by FISH on whole tissue sections and on diagnostic TMA section, independently from the ASCO/CAP guideline used for the classification of cases. Using the 2013 ASCO/CAP scoring criteria, 98.2% of cases were correctly classified, whereas overall concordance was 99.8% when the oldest scoring criteria were applied. Similar studies performed on breast cancer specimens reported concordance rates ranging from 91% to 97% between the HER2 gene amplification status obtained by FISH on whole tissue sections and TMA sections using 2007 ASCO/CAP scoring criteria (12-15). In our hands, no case considered amplified on whole tissue section was evaluated as non-amplified on

- 117 -

diagnostic TMA section, when HER2 gene amplification status was scored according to the 2007 ASCO/CAP scoring criteria (4.5% using the newest scoring criteria). Gancberg and collaborators reported similar results (12), whereas another study observed 13% of non- amplified cases on TMA section among those considered amplified on whole tissue section (15).

In analogy to others (14, 16, 17), we noticed that the overall agreement rate between the two methods was higher with three- or four-core analysis compared to one- or two-core analysis when 2007 ASCO/CAP scoring criteria were used (100.0% vs. 99.4%). However, when the 2013 ASCO/CAP guidelines were used, we obtained opposite results (three or four cores: 98.0%, one or two cores: 98.6%).

The overall concordance between HER2 gene amplification status between the two methods was also high when HER2 status determined on TMA section was considered the reference method (95.6% using the newest guidelines and 98.6% using the 2007 ASCO/CAP scoring criteria), further confirming the reliability of this method in the determination of HER2 gene amplification in breast cancer specimens.

The purpose of our project was also to compare HER2 gene amplification status obtained from the same whole tissue section used for diagnostics to those obtained from the randomly chosen additional block using TMA. The updated ASCO/CAP guidelines recommend performing an initial test (IHC or ISH) in core biopsy and to retest specimens using an alternative assay when results are equivocal. Since FISH was not routinely performed in core biopsy for the 554 consecutive breast cancer specimens, we performed FISH for 116 selected breast cancer specimens on an additional paraffin block, randomly chosen among all paraffin blocks of the same specimen that presented the same histological type as the routine diagnostic paraffin block, i.e. the paraffin block that was used to make the diagnosis. Since the additional block and the additional cores were randomly chosen, we consider this block a proxy of the core biopsy performed by the radiologist under ultrasound examination. We observed an excellent overall concordance (99.0%, kappa value of 0.94) between the HER2 gene amplification status observed on whole tissue and on random TMA sections. These results highlight that reliable HER2 gene amplification status can be achieved even when the tumor area is randomly chosen.

Similar studies have analyzed concordance of HER2 status (determined on whole tissue section) between different blocks from different tumor foci (18, 19). Compared to our

- 118 -

results, two studies have reported slightly lower concordance rates in HER2 gene amplification determined on different blocks, ranging from 90.3% to 94%. These studies, however, analyzed their results using the 2007 ASCO/CAP guidelines (19) or other criteria for HER2 gene amplification (18). Analogous studies that analyzed the concordance between HER2 gene amplification determined on the needle core biopsy and subsequent excisional biopsy (whole tissue section) of the same tumor observed comparatively to our study lower agreement rates ranging between 86% and 92% (20, 21).

TMA technology has several advantages over the traditional method. Variability between batches is considerably reduced, since specimens can be simultaneously processed using identical conditions (8, 9). Furthermore, the TMA technique significantly reduces the reagents and technical time required for staining (22, 23). This is particularly attractive for expensive and time-consuming techniques such as FISH. The amount of tissue needed is also reduced (23).

Although TMA represents a useful tool for rapid and efficient examination of large numbers of tumor tissues, this technique presents some weaknesses, including the need for training of highly qualified technicians and core losses (8, 24). Moreover, it has been criticized that TMA may not accurately represent histopathological characteristics of the whole tissue section. Nonetheless, it has been shown that punching of multiple cores from different representative tumor regions reduces sampling error (15) and provides reliable results even for heterogeneous tumors (25, 26).

In conclusion, our results suggest that FISH is a robust technique and that it can be used for the determination of HER2 status in core biopsies. The high concordance rates between HER2 gene amplification status determined on whole tissue and on random TMA sections (99.0%) suggest that reliable results can be achieved even when tumor area is randomly chosen. Moreover, the observed agreement rates fulfill the ASCO/CAP recommendation of concordance greater than 95% for clearly amplified and non-amplified cases (4, 5). Furthermore, the high concordance rates between whole tissue and diagnostic TMA sections justify the implementation of TMA for the determination of HER2 amplification on surgical specimens in the framework of quality control.

- 119 -

Acknowledgements The Authors express special thanks to Drs. Anne Choquette, Michel Beauchemin, Mohamed Amin Hashem, Sophie Laberge, Mohib Morcos, Nathalie Mourad, Alexandre Odashiro, and Ion Popa. We are grateful to the personnel of the Service de Pathologie, especially to Céline Plourde, for their precious technical support.

DF received doctoral fellowships from the Fonds de recherche du Québec - Santé (FRQS) and the Laval University Cancer Research Center. CD is a recipient of the Canadian Breast Cancer Foundation-Canadian Cancer Society Capacity Development award (award #703003) and the FRQS Research Scholar.

This study was supported by the Fondation des Hôpitaux Enfant-Jésus – St- Sacrement. Clinical specimens were provided by the Fondation du cancer du sein du Québec and the Banque de tissus et de données of the Réseau de recherche sur le cancer of the FRQS, which is affiliated with the Canadian Tumour Repository Network. The Authors also acknowledge Hoffmann-La Roche Limited for its support.

Declaration of interests: None to declare.

Authors’ contributions: CD, SJ and CC designed the research study. DF, SJ, CC and FS performed the research under the supervision of CD. DF analyzed the data and drafted the manuscript. CD, SJ, CC, FS and LP critically revised the manuscript and approved the final version of the manuscript.

- 120 -

References 1 Yarden Y and Sliwkowski MX: Untangling the erbb signalling network. Nat Rev Mol Cell Biol 2(2): 127-137, 2001.

2 Soerjomataram I, Louwman MW, Ribot JG, Roukema JA and Coebergh JW: An overview of prognostic factors for long-term survivors of breast cancer. Breast Cancer Res Treat 107(3): 309-330, 2008.

3 Esteva FJ, Yu D, Hung MC and Hortobagyi GN: Molecular predictors of response to trastuzumab and lapatinib in breast cancer. Nat Rev Clin Oncol 7(2): 98-107, 2010.

4 Wolff AC, Hammond ME, Hicks DG, Dowsett M, McShane LM, Allison KH, Allred DC, Bartlett JM, Bilous M, Fitzgibbons P, Hanna W, Jenkins RB, Mangu PB, Paik S, Perez EA, Press MF, Spears PA, Vance GH, Viale G, Hayes DF, American Society of Clinical O and College of American P: Recommendations for human epidermal growth factor receptor 2 testing in breast cancer: American society of clinical oncology/college of american pathologists clinical practice guideline update. J Clin Oncol 31(31): 3997-4013, 2013.

5 Wolff AC, Hammond ME, Schwartz JN, Hagerty KL, Allred DC, Cote RJ, Dowsett M, Fitzgibbons PL, Hanna WM, Langer A, McShane LM, Paik S, Pegram MD, Perez EA, Press MF, Rhodes A, Sturgeon C, Taube SE, Tubbs R, Vance GH, van de Vijver M, Wheeler TM, Hayes DF and American Society of Clinical Oncology/College of American P: American society of clinical oncology/college of american pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. Arch Pathol Lab Med 131(1): 18-43, 2007.

6 Furrer D, Sanschagrin F, Jacob S and Diorio C: Advantages and disadvantages of technologies for her2 testing in breast cancer specimens. Am J Clin Pathol 144(5): 686-703, 2015.

7 Ross JS: Point: Fluorescence in situ hybridization is the preferred approach over immunohistochemistry for determining her2 status. Clin Chem 57(7): 980-982, 2011.

8 Camp RL, Charette LA and Rimm DL: Validation of tissue microarray technology in breast carcinoma. Lab Invest 80(12): 1943-1949, 2000.

- 121 -

9 Gillett CE, Springall RJ, Barnes DM and Hanby AM: Multiple tissue core arrays in histopathology research: A validation study. J Pathol 192(4): 549-553, 2000.

10 Kononen J, Bubendorf L, Kallioniemi A, Barlund M, Schraml P, Leighton S, Torhorst J, Mihatsch MJ, Sauter G and Kallioniemi OP: Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med 4(7): 844-847, 1998.

11 Tubbs RR, Pettay JD, Swain E, Roche PC, Powell W, Hicks DG and Grogan T: Automation of manual components and image quantification of direct dual label fluorescence in situ hybridization (fish) for her2 gene amplification: A feasibility study. Appl Immunohistochem Mol Morphol 14(4): 436-440, 2006.

12 Gancberg D, Di Leo A, Rouas G, Jarvinen T, Verhest A, Isola J, Piccart MJ and Larsimont D: Reliability of the tissue microarray based fish for evaluation of the her-2 oncogene in breast carcinoma. J Clin Pathol 55(4): 315-317, 2002.

13 Graham AD, Faratian D, Rae F and Thomas JS: Tissue microarray technology in the routine assessment of her-2 status in invasive breast cancer: A prospective study of the use of immunohistochemistry and fluorescence in situ hybridization. Histopathology 52(7): 847- 855, 2008.

14 O'Grady A, Flahavan CM, Kay EW, Barrett HL and Leader MB: Her-2 analysis in tissue microarrays of archival human breast cancer: Comparison of immunohistochemistry and fluorescence in situ hybridization. Appl Immunohistochem Mol Morphol 11(2): 177-182, 2003.

15 Zhang D, Salto-Tellez M, Putti TC, Do E and Koay ES: Reliability of tissue microarrays in detecting protein expression and gene amplification in breast cancer. Mod Pathol 16(1): 79-84, 2003.

16 Fons G, Hasibuan SM, van der Velden J and ten Kate FJ: Validation of tissue microarray technology in endometrioid cancer of the endometrium. J Clin Pathol 60(5): 500-503, 2007.

17 Hoos A, Urist MJ, Stojadinovic A, Mastorides S, Dudas ME, Leung DH, Kuo D, Brennan MF, Lewis JJ and Cordon-Cardo C: Validation of tissue microarrays for immunohistochemical profiling of cancer specimens using the example of human fibroblastic tumors. Am J Pathol 158(4): 1245-1251, 2001.

- 122 -

18 Buggi F, Folli S, Curcio A, Casadei-Giunchi D, Rocca A, Pietri E, Medri L and Serra L: Multicentric/multifocal breast cancer with a single histotype: Is the biological characterization of all individual foci justified? Ann Oncol 23(8): 2042-2046, 2012.

19 Choi Y, Kim EJ, Seol H, Lee HE, Jang MJ, Kim SM, Kim JH, Kim SW, Choe G and Park SY: The hormone receptor, human epidermal growth factor receptor 2, and molecular subtype status of individual tumor foci in multifocal/multicentric invasive ductal carcinoma of breast. Hum Pathol 43(1): 48-55, 2012.

20 Apple SK, Lowe AC, Rao PN, Shintaku IP and Moatamed NA: Comparison of fluorescent in situ hybridization her-2/neu results on core needle biopsy and excisional biopsy in primary breast cancer. Mod Pathol 22(9): 1151-1159, 2009.

21 D'Alfonso T, Liu YF, Monni S, Rosen PP and Shin SJ: Accurately assessing her-2/neu status in needle core biopsies of breast cancer patients in the era of neoadjuvant therapy: Emerging questions and considerations addressed. Am J Surg Pathol 34(4): 575-581, 2010.

22 Jourdan F, Sebbagh N, Comperat E, Mourra N, Flahault A, Olschwang S, Duval A, Hamelin R and Flejou JF: Tissue microarray technology: Validation in colorectal carcinoma and analysis of p53, hmlh1, and hmsh2 immunohistochemical expression. Virchows Arch 443(2): 115-121, 2003.

23 Khouja MH, Baekelandt M, Sarab A, Nesland JM and Holm R: Limitations of tissue microarrays compared with whole-tissue sections in survival analysis. Oncol Lett 1(5): 827- 831, 2010.

24 Conway C, Dobson L, O'Grady A, Kay E, Costello S and O'Shea D: Virtual microscopy as an enabler of automated/quantitative assessment of protein expression in tmas. Histochem Cell Biol 130(3): 447-463, 2008.

25 Giltnane JM and Rimm DL: Technology insight: Identification of biomarkers with tissue microarray technology. Nat Clin Pract Oncol 1(2): 104-111, 2004.

26 Gomaa W, Ke Y, Fujii H and Helliwell T: Tissue microarray of head and neck squamous carcinoma: Validation of the methodology for the study of cutaneous fatty acid-binding protein, vascular endothelial growth factor, involucrin and ki-67. Virchows Arch 447(4): 701- 709, 2005.

- 123 -

Table 4.1. Interpretation criteria for fluorescence in situ hybridization according to the 2007 and the 2013 American Society of Clinical Oncology/College of American Pathologists scoring systems

Scoring system Result category Interpretation criteria 2007 Non-amplified HER2/CEP17 ratio <1.8 Equivocal HER2/CEP17 ratio 1.8-2.2 Amplified HER2/CEP17 ratio > 2.2 2013 Non-amplified Average HER2 gene copy number <4 signals/nucleus or HER2/CEP17 ratio of <2.0 with an average HER2 gene copy number <4 signals/nucleus Equivocal Average HER2 gene copy number ≥4 and < 6 signals/nucleus or a ratio <2.0 with an average HER2 gene copy number ≥4 and <6 signals/nucleus Amplified Average HER2 gene copy number ≥6 signals/nucleus or a ratio <2.0 with an average HER2 gene copy number ≥6 signals/nucleus or a ratio ≥2.0 HER2/CEP17: human epidermal growth factor receptor 2/ chromosome 17 centromere.

- 124 -

Table 4.2. Concordance of human epidermal growth factor receptor 2 (HER2) gene amplification status between fluorescence in situ hybridization performed on whole tissue sections (reference method) and on diagnostic tissue microarray (TMA) sections according to the 2013 American Society of Clinical Oncology/College of American Pathologists scoring system

Whole tissue section (reference), n Diagnostic TMA section Non-amplified Equivocal Amplified Total Non-amplified, n 424 19 3 446 Equivocal, n 5 5 0 10 Amplified, n 1 0 64 65 Total 430 24 67 521 Overall agreement 98.2% Positive agreement (sensitivity) 95.5% Negative agreement (specificity) 98.6% k Value 0.81* *When equivocal cases were excluded: k value=0.96.

- 125 -

Table 4.3. Concordance of human epidermal growth factor receptor 2 (HER2) gene amplification status between fluorescence in situ hybridization performed on whole tissue sections (reference method) and on diagnostic tissue microarray (TMA) sections according to the 2013 American Society of Clinical Oncology/College of American Pathologists scoring system

Whole tissue section (reference) Diagnostic TMA section Non-amplified Equivocal Amplified Total Non-amplified 451 3 0 454 Equivocal 0 2 1 3 Amplified 0 4 60 64 Total 451 9 61 521 Overall agreement 99.8% Positive agreement (sensitivity) 98.4% Negative agreement (specificity) 100.0% k Value 0.93* *When equivocal cases were excluded: k value=1.00.

- 126 -

Table 4.4. Concordance of human epidermal growth factor receptor 2 (HER2) gene amplification status between fluorescence in situ hybridization performed on whole tissue sections (reference method) and on diagnostic tissue microarray (TMA) sections according to the number of informative cores using the 2007 American Society of Clinical Oncology/College of American Pathologists scoring system

Number of assessable Number Overall agreement Positive agreement Negative agreement cores of cases n/N % n/N % n/N % 1 or 2 162 156/157 99.4 21/22 95.4 135/135 100.0 3 or 4 359 355/355 100.0 39/39 100.0 316/316 100.0 N=521 n/N: Number of cases on TMA section/number of cases on whole tissue section.

- 127 -

Table 4.5. Concordance of human epidermal growth factor receptor 2 (HER2) status determined by fluorescence in situ hybridization on whole tissue section (reference method) and on random TMA section according to the 2013 ASCO/CAP scoring system

Whole tissue section (reference) Random TMA Non-amplified Equivocal Amplified Total section Non-amplified 63 2 0 65 Equivocal 0 1 1 2 Amplified 0 0 36 36 Total 63 3 37 103 Overall agreement 99.0% Positive agreement (sensitivity) 97.3% Negative agreement (specificity) 100.0% k Value 0.94* *When equivocal cases were excluded: k value=1.00

- 128 -

Chapter 5: Concordance between immunohistochemistry and fluorescence in situ hybridization in the determination of human epidermal growth factor receptor 2 (HER2) status using tissue microarray in breast cancer specimens

Anticancer Research 2017; 37(6): 3323-3329

Daniela Furrer1,2,3, Simon Jacob4,5,6, Chantal Caron4,5, François Sanschagrin2,4, Louise Provencher2,4,7, Caroline Diorio1,2,3,4

1Centre de Recherche sur la cancer de l’Université Laval, 1050 chemin Ste-Foy, Quebec City, QC G1S 4L8, Canada; 2Axe Oncologie, Centre de Recherche du Centre Hospitalier Universitaire de Québec, Hôpital du Saint-Sacrement, 1050 chemin Ste-Foy, Quebec City, QC G1S 4L8, Canada; 3Département de médecine sociale et préventive, Faculté de Médecine, Université Laval, 1050 Avenue de la Médecine, Quebec City, QC G1V 0A6, Canada; 4Centre des Maladies du Sein Deschênes-Fabia, Hôpital du St-Sacrement, 1050 chemin Ste-Foy, Quebec City, QC G1S 4L8, Canada; 5Service de Pathologie, Hôpital du Saint-Sacrement, 1050 chemin Ste-Foy, Quebec City, QC G1S 4L8, Canada; 6 Département de biologie moléculaire, de biochimie médicale et de pathologie, Faculté de Médecine, Université Laval, 1050 Avenue de la Médecine, Quebec City, QC G1V 0A6, Canada; 7Département de chirurgie, Faculté de Médecine, Université Laval, 1050 Avenue de la Médecine, Quebec City, QC G1V 0A6, Canada

Keywords : Breast neoplasms; HER2 status; IHC; FISH; TMA; multiple blocks; ASCO/CAP guidelines; biopsy; trastuzumab

- 129 -

Résumé HER2 est un marqueur prédictif du cancer du sein. Les méthodes couramment utilisées pour la détermination d’HER2 dans les spécimens du cancer du sein sont l’immunohistochimie (IHC) et les techniques d’hybridation in situ (ISH).

Dans une cohorte de 498 patientes consécutives atteintes d’un cancer du sein, nous avons examiné la concordance entre le statut HER2 déterminé par IHC et FISH sur des sections de matrice tissulaire (tissue microarray, TMA). De plus, dans un sous-groupe de 116 spécimens, nous avons examiné la concordance entre le statut HER2 obtenu avec le bloc ayant servi au diagnostic et avec un bloc additionnel (un proxy de la biopsie).

La concordance générale entre l’IHC et le FISH sur TMA était de 93,8%. La concordance générale entre le statut HER2 déterminé sur le bloc diagnostique et sur le bloc additionnel était de 93,6% pour l’IHC et de 98,0% pour le FISH.

Puisque nous avons observé des cas discordants entre les deux méthodes, nous suggérons que les deux méthodes devraient être utilisées pour évaluer le statut HER2 dans les spécimens de cancer du sein. Le taux de concordance plus petit observé entre le bloc diagnostique et le bloc additionnel à l’IHC suggère que le marquage à l’IHC entre les différentes régions de la tumeur est plus variable comparativement au marquage obtenu par FISH.

- 130 -

Abstract HER2 is a predictive marker in breast cancer. Commonly used methods for the determination of HER2 in breast cancer specimens are immunohistochemistry (IHC) and in situ hybridization (ISH) techniques. In a cohort of 498 consecutive breast cancer patients, we examined the concordance between HER2 status determined by IHC and FISH on tissue microarray (TMA) section. Moreover, in a subset of 116 specimens, we examined the concordance between HER2 status from the routine diagnostic paraffin block and a randomly chosen additional block (a proxy of the core biopsy). Overall concordance between IHC and FISH on TMA section was 93.8%. Overall concordance between HER2 status determined on diagnostic block and on additional block was 93.6% for IHC and 98.0% for FISH. Since we observed some discordant cases between the two methods, we suggest that both methods should be used to assess HER2 status in breast cancer specimens. The lower concordance rate between diagnostic and additional blocks using IHC compared to FISH suggests a greater variability of IHC staining across tumor regions than FISH results.

- 131 -

1. Introduction Human epidermal growth factor receptor 2 (HER2) is a transmembrane tyrosine kinase receptor belonging to the family of epidermal growth factor receptor (EGFR) 1. The protein is encoded by the HER2 (ERBB2) gene, which is located on the long arm of chromosome 17 (17q12-21.32) 2. HER2 gene amplification and protein overexpression, which occur in 15 to 20% of breast cancer patients, are important markers for poor prognosis, including a more aggressive disease and a shorter survival 3. Moreover, HER2-positive status is considered a predictive marker of response to HER2-targeted drugs, including trastuzumab and lapatinib 4. Given its prognostic, predictive and therefore therapeutic implications, an accurate evaluation of HER2 status is crucial for identification of patients who would most likely benefit from targeted anti-HER2 therapies.

Currently, there are several Food and Drug (FDA)-approved methods for the evaluation of HER2 status in breast cancer specimens, including immunohistochemical (IHC) determination of HER2 protein expression or assessment of HER2 gene amplification using in situ hybridization (ISH), most commonly fluorescent ISH (FISH) 5,6. Patients are eligible for targeted anti-HER2 therapies when their breast cancer specimens overexpress the protein at the IHC and/or are HER2 gene amplified at ISH. Since both IHC and FISH present advantages and disadvantages, there is still no consensus on which method is superior for assessing the HER2 status in breast cancer specimens 7. In 2013, the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) updated the guidelines to clarify the recommendations for HER2 testing in breast cancer specimens published in 2007 5. In particular, the 2013 ASCO/CAP guidelines include new scoring criteria for IHC and FISH. Moreover, the updated guidelines recommend performing an initial test (IHC or ISH) on core biopsy. If test results are equivocal, reflex testing on tumor specimen section with an alternative assay (IHC or ISH) should be carried out. In addition, repeat testing should be performed if there is an apparent histopathologic discordance with the test result 5. The 2007 ASCO/CAP guidelines recommended to perform HER2 testing on resection specimens and to retest when results were equivocal 6.

Several publications have demonstrated a very good concordance between the results obtained by IHC and FISH for the determination of HER2 status in breast cancer specimens 8,9. Other studies, however, have reported up to 13% of FISH positive cases among those scoring negative by IHC 10-15. Considering this discordance, it would be therefore interesting to perform both IHC and FISH on every specimen. Since the vast majority of studies

- 132 -

published so far analyzed the concordance between IHC and FISH according to the 2007 ASCO/CAP criteria 6, it is pertinent to analyse the concordance between these two FDA- approved techniques according to the recent ASCO/CAP scoring criteria 5. Moreover, we want to evaluate the new approach by which HER2 status is determined at first on core biopsy as recommended by the updated ASCO/CAP guidelines 5.

Tissue microarray (TMA) allows for molecular characterization of large amount of specimens by means of arranging tissue cores from multiple samples into an empty paraffin block. Large amount of specimens can therefore be processed under identical conditions and analysed simultaneously 16,17. Goal of this study is to analyze the concordance of HER2 status determined by IHC and FISH in 498 consecutive breast cancer specimens on TMA sections and to evaluate the impact of the new ASCO/CAP scoring guidelines on the concordance results. Moreover, in a subset of 116 breast cancer specimens, we aim to evaluate the concordance between HER2 status from the routine diagnostic paraffin block and a randomly chosen additional block (a proxy of the core biopsy) using TMA.

2. Materials and methods

2.1. Specimen collection and patient population Study population has already been described elsewhere 18. Briefly, formalin-fixed, paraffin- embedded breast cancer tissues from 554 consecutive patients with invasive breast carcinoma were used. Patients who received neoadjuvant chemotherapy or with a tumor size smaller than 1 cm on histological slides were excluded from the study. All samples were collected at the Centre des Maladies du Sein Deschênes-Fabia at the Saint-Sacrement Hospital in Québec, Québec, Canada, between February 2011 and April 2012. ASCO/CAP recommendations regarding formalin fixation time and time to fixative were followed 5. Written informed consent was obtained from all participants. Ethical approval of the study was obtained from the Research Ethics Committee of the Centre de Recherche du CHU de Québec (Permit Number: DR-002-1286).

2.2. Tissue microarray (TMA) construction TMAs were constructed as previously described 18. Briefly, for the 554 consecutive breast cancer specimens, the most representative tissue block from each case was selected by the pathologist for HER2 assessment by IHC and FISH. Two tumoral regions showing the strongest IHC staining were delineated on the IHC slide by the pathologist. Four 0.6 mm

- 133 -

tissue cylinders were punched using a manual arraying instrument within these delineated regions (Beecher Instruments, Silver Spring, MD, USA) and were inserted into empty recipient paraffin blocks. We call these TMA blocks “diagnostic TMA”. Among the 554 cases present on the diagnostic TMA, cases that had at least two additional paraffin blocks showing the same histological type as the paraffin block that was used for the construction of the diagnostic TMA were selected. Among these cases, 100 cases were randomly chosen. Suitable blocks were identified using H&E section and two tumoral regions were delineated without previous IHC staining on the H&E section by two trained technologists and verified by a pathologist. Four 0.6 mm tissue cylinders were then punched from these regions. Since we also performed an oversampling of HER2 positive cases (16 cases that were scored as HER2 positive on diagnostic TMA section but that had not been randomly selected were added to the 100 cases), a total of 116 cases were inserted into an empty paraffin block. We call these TMA blocks “random TMA”. Consecutive 4-μm sections were processed by IHC and FISH. From each TMA block one section was stained with H&E for reference histology.

2.3. Immunohistochemistry HER2 protein expression was performed using the FDA-approved HercepTest™ kit (DAKO Diagnostics, Glostrup, Denmark) on an automated immunostaining system (Autostainer, DAKO), according to the manufacturer’s instructions.

2.4. Fluorescence in situ hybridization HER2 gene copy number was evaluated using the FDA-approved PathVysion™ HER2 DNA Probe kit (Abbott Molecular, Des Plaines, IL, USA/ Inter Medico, Markham, Canada), according to the manufacturer’s instructions.

2.5. HER2 evaluation on TMA A core was considered satisfactory for analysis if tumor tissue occupied > 10% of the core area.

Immunohistochemical staining was analyzed visually. Cytoplasmic staining was ignored and only invasive tumor was scored.

Fluorescent signals were evaluated as previously described 18. Briefly, fluorescent signals were evaluated with an epifluorescence microscope Axio Imager M1 (Zeiss, Göttingen,

- 134 -

Germany). Automated analysis of fluorescence signals was effectuated using the FDA- approved MetaSystemsTM image analysis system 19. For cores showing equivocal results, automated image analysis was followed by manual counting in at least 60 nonoverlapping tumor cells. Moreover, manual counting was performed in 40 nonoverlapping tumor cells in the following situations: the average HER2 copy number per tile was ≥ 4.0 and ≤ 6.0 at the automated image analysis; automated counting of signals from the invasive cancer cells was impeded (low cellular density or high stromal density); polysomy or monosomy of chromosome 17 as defined by Tubbs and collaborators 20 was suspected. In addition, all cores with ratio > 1.5 and < 3.0 at the automated analysis were visually verified.

Each informative core was evaluated separately in a blind fashion. Cores were analyzed by trained technologists, and all results were validated by breast pathologists. Average results of informative cores were considered. IHC and FISH results were reported according to the 2013 ASCO/CAP guidelines 5.

2.6. Statistical analysis Only cases with at least one informative core for IHC or FISH were included in the analysis. The concordance between IHC and FISH was analyzed by comparing average immunohistochemical staining and average gene amplification results. Averages of HER2 status by IHC and FISH on diagnostic TMA sections were available for 498 cases out of the 554 total cases (89.9%). Averages of HER2 status by IHC on diagnostic TMA and random TMA were available for 106 cases out of the 116 selected cases (91.4%), whereas averages of HER2 status by FISH on diagnostic TMA and random TMA were available for 87.9% of cases (102/116). Averages of HER2 status by IHC and FISH on random TMA sections were available for 99 cases out of the 116 selected cases (85.3%).

To evaluate the agreement between the two methods, positive, negative and overall concordance were calculated as previously published 18 following the recommendation of the 2013 ASCO/CAP guidelines 5. The level of agreement was also measured using the Cohen’s kappa test (agreement adjusted by chance). All analyses were performed using SAS software (version 9.1.3, SAS Institute, Inc., Cary, NC, USA).

3. Results Concordance of HER2 status determined by IHC and FISH on diagnostic TMA section is summarized in Table 5.1. The overall agreement between the two methods was 93.8%

- 135 -

(kappa value=0.60). The positive agreement and the negative agreements were 69.2%, and 92.9%, respectively. Among the 423 cases considered FISH non-amplified, protein overexpression was observed in one (0.2%) case. Five (7.7%) of the 65 cases that showed gene amplification were negative at IHC.

We also assessed the concordance rates between HER2 status determined by IHC and FISH on TMA section according to the number of informative cores per case (one or two vs. three or four cores). The overall, positive and negative agreements for cases with one or two informative cores were lower (85.7%, 58.3% and 90.3%, respectively) than for cases with three or four evaluable cores (91.9%, 75.6% and 94.3%, respectively).

Table 5.2. shows concordance rate of HER2 status determined by IHC and FISH on random TMA section. The overall agreement was 83.5% (kappa value=0.67). The positive agreement and the negative agreements were 71.4%, and 90.3%, respectively. Among the 62 cases considered FISH non-amplified, none was evaluated as 3+ for protein overexpression. Of the 35 cases that showed gene amplification, two (5.7%) were negative at IHC.

Concordance of HER2 status determined by FISH on diagnostic TMA section and random TMA section is summarized in Table 5.3. In this analysis, we observed an overall agreement of 98.0%. The positive and the negative agreements were 97.2%, and 98.5%, respectively. Of the 66 cases considered FISH non-amplified on diagnostic TMA section, one (1.5%) was evaluated as equivocal on random TMA section. One (2.8%) of the 36 cases that showed gene amplification on diagnostic TMA section was equivocal for HER2 gene amplification on random TMA section.

Table 5.4. presents the concordance of HER2 status determined by IHC on diagnostic TMA section and random TMA section. The overall agreement between the two methods was 93.6% (kappa value=0.86). The positive agreement and the negative agreements were 92.9%, and 93.9%, respectively. Among the 66 cases considered negative for protein overexpression on diagnostic TMA section, four (6.1%) were evaluated as equivocal on random TMA section. Two of the 28 cases evaluated as positive for protein overexpression on diagnostic TMA section were considered equivocal on random TMA section.

- 136 -

4. Discussion We have observed a moderate concordance rate between the HER2 status determined by IHC and FISH on TMA section in 498 consecutive breast cancer specimens. In our hands, we observed 93.8% concordance (kappa value of 0.60) between the two techniques when 2013 ASCO/CAP scoring criteria were used. Similar studies performed on breast cancer specimens using TMA section reported concordance rates ranging from 78.1% to 98.0% between the HER2 status obtained by IHC and FISH 21-27. The concordance results observed in those studies were evaluated using the 2007 ASCO/CAP scoring criteria 6. When we analysed our data using the 2007 ASCO/CAP scoring criteria, overall concordance was 91.5% (kappa value = 0.66, Table 5.A.1), within the concordance range reported in the above mentioned studies. Among the 431 cases considered FISH non-amplified, one (0.2%) was evaluated as positive for protein expression. Six (9.4%) of the 64 cases that showed gene amplification were negative at IHC.

This is the first study that analyzed the concordance between IHC and FISH on TMA section in breast cancer specimens according to the 2007 and 2013 ASCO/CAP criteria. To date, only one study analyzed the concordance between IHC and FISH on whole tissue section in 189 non-consecutive breast cancer specimens using both ASCO/CAP scoring criteria 28. Similar to Garbar and collaborators, we observed a decrease in the percentage of cases evaluated as FISH amplified and IHC negative when the 2013 ASCO/CAP scoring criteria were used. However, Garbar and al. reported one case considered FISH non-amplified and IHC positive using the 2013 ASCO/CAP scoring criteria vs. no cases observed using the 2007 ASCO/CAP scoring criteria, whereas in our study the utilization of the recent ASCO/CAP scoring criteria did not change the percentage of these cases.

We observed 7.7% of amplified cases among those considered negative at IHC (9.4% when 2007 ASCO/CAP scoring criteria were used). Similar studies comparing the concordance between IHC and FISH on TMA section reported between 1.24% and 11.5% of IHC negative cases being FISH amplified 21-23,25,27. In our study, one non-amplified case was considered HER2 overexpressing with IHC, representing thus 0.23% of the IHC positive cases (same proportion when 2007 ASCO/CAP guidelines were used). Similar studies have reported discordance rates for this category ranging from 0% to 23.5% 21-23,25. We remarked that this discordant case showed polysomy of chromosome 17. In agreement to previous studies 21,26,27, we postulate that the presence of chromosome 17 aneusomy could at least in part explain the inconsistence between results.

- 137 -

Goal of our project was also to evaluate the impact of the new ASCO/CAP guidelines on the evaluation of HER2 status. The updated guidelines recommend performing an initial test (IHC or ISH) in core biopsy and specimens should be retested with an alternative assay when results are equivocal 5. Since IHC and FISH were not routinely performed in core biopsy for the 554 consecutive breast cancer specimens, for 116 selected breast cancer specimens we performed IHC and FISH on an additional paraffin block, randomly chosen among all paraffin blocks of the same specimen that presented the same histological type as the routine diagnostic paraffin block, i.e. the paraffin block that was used to make the diagnosis. Since this additional block has been randomly chosen, we consider this block a proxy of the core biopsy performed by the radiologist under ultrasound control (which is a random selection of a tumor area). We then compared IHC and FISH results obtained on random TMA section (which simulate the HER2 status that we would obtain when IHC and FISH would have been performed in core biopsy) with IHC and FISH results obtained on diagnostic TMA section (which represent results obtained on excisional breast cancer specimens). We observed an excellent overall concordance (98.0%, kappa value 0.94) between the HER2 gene amplification status observed on diagnostic TMA section and on random TMA section. The overall concordance rate between HER2 protein overexpression obtained on diagnostic TMA section and random TMA section, however, was 93.6%, and therefore does not fulfill the ASCO/CAP suggestion of concordance greater than 95% for clearly negative and positive cases 5. Overall concordance rates between IHC and FISH performed on random TMA section and on diagnostic TMA section were also both lower than 95% (83.5% vs. 93.8%).

In a similar study conducted in a cohort of 139 breast cancer cases for which HER2 status was determined in more than one block of a single tumor focus, Bethune et al. analysed the HER2 concordance between the different blocks (both determined on whole tissue section) and reported 96.4% concordance rate 29. Results were considered concordant if the final HER2 status of all blocks was the same, regardless of whether HER2 status was determined by IHC or FISH. Selected cases of our cohort are comparable to those of the mentioned study, since blocks were from different tumor foci for only one patient of the subset. When HER2 status was defined according to the combined IHC and FISH results, we observed 93.7% concordance rate between diagnostic block and randomly selected block (Table 5.A.2). This difference might be explained by the fact that Bethune et al. used the former ASCO/CAP scoring guidelines. Similar to Bethune et al., we observed that the majority of

- 138 -

discordant cases (seven out of nine) displayed an equivocal HER2 immunostaining. Unlike our study, Bethune et al. did not differentiate whether concordance rate between the blocks varied according to which test has been used to determine HER2 status (IHC and/or FISH), as in their study HER2 status has been determined by IHC, and FISH has been mainly performed on specimens with equivocal immunohistochemical staining. In our hands, HER2 concordance rate between the different blocks was higher when HER2 status was determined by FISH than by IHC. A study performed in a cohort of breast cancer specimens showing equivocal immunostaining (evaluated on whole tissue section according to the scoring guidelines of the HercepTest) reported that 68% of cases had a different score when the immunohistochemical staining was performed on additional blocks from the same breast tumor specimen or from axillary lymph node metastasis 30. This discordance rate in immunological staining between different blocks is substantially higher compared to that we observed in our study. This difference may be explained by the fact that Lewis et al. analyzed the concordance for IHC scoring exclusively in cases that were originally evaluated as equivocal at IHC. Being the interpretation of immunostaining based on semi-quantitative scoring, it has been reported that interobserver variability and therefore discrepancies in HER2 IHC results were particularly elevated for cases scoring 2+ 31.

Analogous studies performed on multifocal and multicentric breast cancer specimens have analyzed concordance of HER2 status (determined on whole tissue section) between different blocks from different tumor foci 32-34. Compared to our results, two studies have observed slightly lower concordance rate in HER2 gene amplification determined on different blocks, ranging from 90.3% to 94% 33,34. These studies, however, reported their results using the 2007 ASCO/CAP guidelines 34 or other criteria for HER2 gene amplification 33. Another study reported 93.5% concordance rate in HER2 status determined by IHC and FISH from different blocks, where the largest focus displayed the most positive result in 98.4% of cases 32. This latter study, however, did not compare IHC and FISH results separately.

In contrast to our study, an analogous study that compared HER2 status determined on the needle core biopsy and subsequent excisional biopsy (whole tissue section) of the same tumor by IHC and FISH observed a higher concordance rate when HER2 status was evaluated by IHC (98% vs 92%) 35.

- 139 -

TMA technology has several advantages over the traditional method, including batch variability reduction and decrease in reagent and technical time required for staining 16,17,36,37. The amount of tissue needed is reduced as well 37. In addition, visual assessment of immunohistochemical staining might be easier on TMA slides comparatively to whole tissue sections as the evaluator can compare staining intensities from different specimens on the same TMA slide 38.

Although TMAs represent a useful tool for rapid and efficient examination of large numbers of tumor tissues, this technique presents some limitations, including the training of high qualified technicians and core losses 16,38. Furthermore, it has been criticized that TMA section may not accurately represent histopathological characteristics of the whole tissue section. However, we and others have demonstrated that even one to two 0.6 mm cores per case can reliably reproduce results achieved on whole tissue sections 18,39.

5. Conclusions In conclusion, whereas HER2 gene amplification seems to be very constant in different tumor regions, we observed greater variability regarding HER2 IHC staining. Indeed, we observed a very good concordance (98%) between HER2 gene amplification status determined on diagnostic TMA section and on random TMA section. Concordance between HER2 protein overexpression observed on diagnostic TMA section and on random TMA section instead was 93.6%, lower than the 95% concordance suggested by the ASCO/CAP. The recent guidelines for HER2 testing in breast cancer developed by the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) recommend effectuating an initial test with IHC or ISH in core biopsy specimen. If test results are equivocal, reflex testing on tumor specimen section with an alternative assay (IHC or ISH) should be performed. The guidelines do not recommend retesting when the assay result is clearly negative or positive. Nonetheless, based on our observation, we suggest that HER2 status determined on core biopsy by IHC should be interpreted carefully and a confirmatory retest using ISH methods in core biopsy or on tumor specimen section should be considered. In our hands, the FISH method allows a reliable evaluation of HER2 status on core biopsy.

6. Conflict of interest and sources of funding This work was supported by Hoffmann-La Roche Limited. Funding source had no role in study design; data collection and analysis, decision to publish; or preparation of the manuscript.

- 140 -

7. Acknowledgments Special thanks to Drs. Anne Choquette, Michel Beauchemin, Mohamed Amin Hashem, Sophie Laberge, Mohib Morcos, Nathalie Mourad, Alexandre Odashiro, and Ion Popa. We are grateful to the personnel of the Service de Pathologie, especially to Céline Plourde, for their precious technical support. DF received doctoral fellowships from the Fonds de recherche du Québec - Santé (FRQS) and the Laval University Cancer Research Center. CD is a recipient of the Canadian Breast Cancer Foundation-Canadian Cancer Society Capacity Development award (award #703003) and the FRQS Research Scholar. This study was supported by the Fondation des Hôpitaux Enfant-Jésus – St-Sacrement. Clinical specimens were provided by the Fondation du cancer du sein du Québec and the Banque de tissus et de données of the Réseau de recherche sur le cancer of the FRQS, which is affiliated with the Canadian Tumour Repository Network.

Authors’ contribution: CD, SJ and CC designed the research study. DF, SJ, CC and FS performed the research. DF analysed the data and drafted the manuscript. CD, SJ, CC, FS and LP critically revised the manuscript and approved the final version of the manuscript.

- 141 -

References 1. Yarden Y, Sliwkowski MX. Untangling the ErbB signalling network. Nat Rev Mol Cell Biol. 2001;2(2):127-137.

2. Popescu NC, Mark GE. Localization of the pKs gene, a raf related sequence on human chromosomes X and 7. Oncogene. 1989;4(4):517-519.

3. Soerjomataram I, Louwman MW, Ribot JG, et al. An overview of prognostic factors for long-term survivors of breast cancer. Breast Cancer Res Treat. 2008;107(3):309-330.

4. Esteva FJ, Yu D, Hung MC, et al. Molecular predictors of response to trastuzumab and lapatinib in breast cancer. Nat Rev Clin Oncol. 2010;7(2):98-107.

5. Wolff AC, Hammond ME, Hicks DG, et al. Recommendations for human epidermal growth factor receptor 2 testing in breast cancer: American Society of Clinical Oncology/College of American Pathologists clinical practice guideline update. J Clin Oncol. 2013;31(31):3997-4013.

6. Wolff AC, Hammond ME, Schwartz JN, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. Arch Pathol Lab Med. 2007;131(1):18-43.

7. Furrer D, Sanschagrin F, Jacob S, et al. Advantages and disadvantages of technologies for HER2 testing in breast cancer specimens. American journal of clinical pathology. 2015;144(5):686-703.

8. Ahmed SS, Iqbal J, Thike AA, et al. HER2/neu revisited: quality and interpretive issues. J Clin Pathol. 2011;64(2):120-124.

9. Arnould L, Roger P, Macgrogan G, et al. Accuracy of HER2 status determination on breast core-needle biopsies (immunohistochemistry, FISH, CISH and SISH vs FISH). Mod Pathol. 2012;25(5):675-682.

10. Bahreini F, Soltanian AR, Mehdipour P. A meta-analysis on concordance between immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) to detect HER2 gene overexpression in breast cancer. Breast Cancer. 2014.

- 142 -

11. Kovacs A, Stenman G. HER2-testing in 538 consecutive breast cancer cases using FISH and immunohistochemistry. Pathol Res Pract. 2010;206(1):39-42.

12. Lal P, Salazar PA, Hudis CA, et al. HER-2 testing in breast cancer using immunohistochemical analysis and fluorescence in situ hybridization: a single-institution experience of 2,279 cases and comparison of dual-color and single-color scoring. American journal of clinical pathology. 2004;121(5):631-636.

13. Martin V, Camponovo A, Ghisletta M, et al. Internal Quality Assurance Program for ERBB2 (HER2) Testing Improves the Selection of Breast Cancer Patients for Treatment with Trastuzumab. Patholog Res Int. 2012;2012:261857.

14. Owens MA, Horten BC, Da Silva MM. HER2 amplification ratios by fluorescence in situ hybridization and correlation with immunohistochemistry in a cohort of 6556 breast cancer tissues. Clin Breast Cancer. 2004;5(1):63-69.

15. Prati R, Apple SK, He J, et al. Histopathologic characteristics predicting HER-2/neu amplification in breast cancer. Breast J. 2005;11(6):433-439.

16. Camp RL, Charette LA, Rimm DL. Validation of tissue microarray technology in breast carcinoma. Lab Invest. 2000;80(12):1943-1949.

17. Gillett CE, Springall RJ, Barnes DM, et al. Multiple tissue core arrays in histopathology research: a validation study. J Pathol. 2000;192(4):549-553.

18. Furrer D, Jacob S, Caron C, et al. Tissue Microarray Is a Reliable Tool for the Evaluation of HER2 Amplification in Breast Cancer. Anticancer Res. 2016;36(9):4661-4666.

19. Tubbs RR, Pettay JD, Swain E, et al. Automation of manual components and image quantification of direct dual label fluorescence in situ hybridization (FISH) for HER2 gene amplification: A feasibility study. Appl Immunohistochem Mol Morphol. 2006;14(4):436-440.

20. Tubbs RR, Hicks DG, Cook J, et al. Fluorescence in situ hybridization (FISH) as primary methodology for the assessment of HER2 Status in adenocarcinoma of the breast: a single institution experience. Diagn Mol Pathol. 2007;16(4):207-210.

21. Egervari K, Szollosi Z, Nemes Z. Tissue microarray technology in breast cancer HER2 diagnostics. Pathol Res Pract. 2007;203(3):169-177.

- 143 -

22. Hwang CC, Pintye M, Chang LC, et al. Dual-colour chromogenic in-situ hybridization is a potential alternative to fluorescence in-situ hybridization in HER2 testing. Histopathology. 2011;59(5):984-992.

23. Lee Y, Ryu Y, Jeong H, et al. Effectiveness of silver-enhanced in situ hybridization for evaluating HER2 gene status in invasive breast carcinoma: a comparative study. Arch Med Res. 2012;43(2):139-144.

24. O'Grady A, Flahavan CM, Kay EW, et al. HER-2 analysis in tissue microarrays of archival human breast cancer: comparison of immunohistochemistry and fluorescence in situ hybridization. Appl Immunohistochem Mol Morphol. 2003;11(2):177-182.

25. Park K, Han S, Kim JY, et al. Silver-Enhanced In Situ Hybridization as an Alternative to Fluorescence In Situ Hybridization for Assaying HER2 Amplification in Clinical Breast Cancer. J Breast Cancer. 2011;14(4):276-282.

26. Saez A, Andreu FJ, Segui MA, et al. HER-2 gene amplification by chromogenic in situ hybridisation (CISH) compared with fluorescence in situ hybridisation (FISH) in breast cancer-A study of two hundred cases. Breast. 2006;15(4):519-527.

27. Schiavon BN, Jasani B, de Brot L, et al. Evaluation of reliability of FISH versus brightfield dual-probe in situ hybridization (BDISH) for frontline assessment of HER2 status in breast cancer samples in a community setting: influence of poor tissue preservation. Am J Surg Pathol. 2012;36(10):1489-1496.

28. Garbar C, Savoye AM, Mascaux C, et al. The human epidermal growth factor receptor 2 screening tests for breast cancer suggested by the new updated recommendation of the american society of clinical oncology/college of american pathologists will involve a rise of the in-situ hybridization tests for the European laboratories of pathology. ISRN Oncol. 2014;2014:793695.

29. Bethune GC, Mullen JB, Chang MC. Detecting intratumoral heterogeneity in routine breast-HER2 testing: low yield of testing multiple blocks. Ann Diagn Pathol. 2015.

30. Lewis JT, Ketterling RP, Halling KC, et al. Analysis of intratumoral heterogeneity and amplification status in breast carcinomas with equivocal (2+) HER-2 immunostaining. American journal of clinical pathology. 2005;124(2):273-281.

- 144 -

31. Hoang MP, Sahin AA, Ordonez NG, et al. HER-2/neu gene amplification compared with HER-2/neu protein overexpression and interobserver reproducibility in invasive breast carcinoma. American journal of clinical pathology. 2000;113(6):852-859.

32. Bethune GC, Mullen JB, Chang MC. HER2 testing of multifocal invasive breast carcinoma: how many blocks are enough? American journal of clinical pathology. 2013;140(4):588-592.

33. Buggi F, Folli S, Curcio A, et al. Multicentric/multifocal breast cancer with a single histotype: is the biological characterization of all individual foci justified? Ann Oncol. 2012;23(8):2042-2046.

34. Choi Y, Kim EJ, Seol H, et al. The hormone receptor, human epidermal growth factor receptor 2, and molecular subtype status of individual tumor foci in multifocal/multicentric invasive ductal carcinoma of breast. Hum Pathol. 2012;43(1):48-55.

35. Apple SK, Lowe AC, Rao PN, et al. Comparison of fluorescent in situ hybridization HER-2/neu results on core needle biopsy and excisional biopsy in primary breast cancer. Mod Pathol. 2009;22(9):1151-1159.

36. Jourdan F, Sebbagh N, Comperat E, et al. Tissue microarray technology: validation in colorectal carcinoma and analysis of p53, hMLH1, and hMSH2 immunohistochemical expression. Virchows Arch. 2003;443(2):115-121.

37. Khouja MH, Baekelandt M, Sarab A, et al. Limitations of tissue microarrays compared with whole tissue sections in survival analysis. Oncol Lett. 2010;1(5):827-831.

38. Conway C, Dobson L, O'Grady A, et al. Virtual microscopy as an enabler of automated/quantitative assessment of protein expression in TMAs. Histochem Cell Biol. 2008;130(3):447-463.

39. Gancberg D, Di Leo A, Rouas G, et al. Reliability of the tissue microarray based FISH for evaluation of the HER-2 oncogene in breast carcinoma. J Clin Pathol. 2002;55(4):315- 317.

- 145 -

Table 5.1. Concordance of HER2 status determined by IHC and FISH on diagnostic TMA section according to the 2013 ASCO/CAP scoring system

FISH TMA IHC TMA Non- Equivocal Amplified Total amplified 0-1+ 393 8 5 406 2+ 29 2 15 46 3+ 1 0 45 46 Total 423 10 65 498 Abbreviations: IHC: immunohistochemistry; FISH: fluorescence in situ hybridization; TMA: Tissue microarray; ASCO/CAP: American Society of Clinical Oncology/College of American Pathologists.

Overall agreement (%): 93.8% Sensitivity (%) : 69.2% Specificity (%) : 92.9% kappa value : 0.60; kappa value when equivocal cases are excluded: 0.93

- 146 -

Table 5.2. Concordance of HER2 status determined by IHC and FISH on random TMA section according to the 2013 ASCO/CAP scoring system

FISH random TMA section IHC random TMA Non- Equivocal Amplified Total section amplified 0/1+ 56 1 2 59 2+ 6 1 8 15 3+ 0 0 25 25 Total 62 2 35 99 Abbreviations: IHC: immunohistochemistry; FISH: fluorescence in situ hybridization; TMA: Tissue microarray; ASCO/CAP: American Society of Clinical Oncology/College of American Pathologists.

Overall agreement (%): 83.5% Sensitivity (%) : 71.4% Specificity (%) : 90.3% kappa value : 0.67; kappa value when equivocal cases are excluded: 0.94

- 147 -

Table 5.3. Concordance of HER2 status determined by FISH on diagnostic TMA section and random TMA section according to the 2013 ASCO/CAP scoring system

FISH diagnostic TMA section FISH random TMA Non- Equivocal Amplified Total section amplified Non-amplified 65 0 0 65 Equivocal 1 0 1 2 Amplified 0 0 35 35 Total 66 0 36 102 Abbreviations: FISH: fluorescence in situ hybridization; TMA: Tissue microarray; ASCO/CAP: American Society of Clinical Oncology/College of American Pathologists.

Overall agreement (%): 98.0% Sensitivity (%) : 97.2% Specificity (%) : 98.5% kappa value when equivocal cases are excluded: 1.00.

- 148 -

Table 5.4. Concordance of HER2 status determined by IHC on diagnostic TMA section and random TMA section according to the 2013 ASCO/CAP scoring system

IHC diagnostic TMA section IHC random TMA 0/1+ 2+ 3+ Total section 0/1+ 62 2 0 64 2+ 4 10 2 16 3+ 0 0 26 26 Total 66 12 28 106 Abbreviations: IHC: immunohistochemistry; TMA: Tissue microarray; ASCO/CAP: American Society of Clinical Oncology/College of American Pathologists.

Overall agreement (%): 93.6% Sensitivity (%) : 92.9% Specificity (%) : 93.9% kappa value : 0.86; kappa value when equivocal cases are excluded: 1.00.

- 149 -

Table 5.A.1. Comparison of HER2 status determined by IHC and FISH on diagnostic TMA section according to the 2007 ASCO/CAP scoring system

FISH TMA IHC TMA Non- Equivocal Amplified Total amplified 0/1+ 412 2 6 420 2+ 18 1 17 36 3+ 1 0 41 42 Total 431 3 64 498 Abbreviations: IHC: immunohistochemistry; FISH: fluorescence in situ hybridization; TMA: Tissue microarray; ASCO/CAP: American Society of Clinical Oncology/College of American Pathologists.

Overall agreement (%): 91.5% Sensitivity (%) : 64.1% Specificity (%) : 95.6% kappa value : 0.66; kappa value when equivocal cases are excluded: 0.91.

- 150 -

Table 5.A.2. Concordance of HER2 status according to the combined IHC and FISH between diagnostic TMA section and random TMA section according to the 2013 ASCO/CAP scoring system

IHC and FISH diagnostic TMA section IHC and FISH Negative Equivocal Positive Total random TMA section Negative 54 2 0 56 Equivocal 4 2 2 8 Positive 0 0 35 35 Total 58 4 37 99 Abbreviations: IHC: immunohistochemistry; FISH: fluorescence in situ hybridization; TMA: Tissue microarray; ASCO/CAP: American Society of Clinical Oncology/College of American Pathologists.

Overall agreement (%): 93.7% Sensitivity (%) : 94.6% Specificity (%) : 93.1% kappa value : 0.85, kappa value when equivocal cases are excluded: 1.00.

- 151 -

Chapter 6: Validation of a new classifier for the automated analysis of the human epidermal growth factor receptor 2 (HER2) gene amplification in breast cancer specimens

Diagnostic pathology 2013; 8: 17

Daniela Furrer1,2, Simon Jacob1,2,3,4, Chantal Caron2,3,4, François Sanschagrin1,2,3,4, Louise Provencher1,3,5, Caroline Diorio1,3,6

1Unité de recherche en santé des populations at Centre de recherche du CHU de Québec, 2Service de Pathologie and 3Centre des Maladie du Sein Deschênes-Fabia at St-Sacrement Hospital, 1050 chemin Ste-Foy, Quebec City, QC, Canada G1S 4L8, 4Département de biologie médicale, 5Département de chirurgie and 6Département de médecine sociale et préventive, Faculté de Médecine, Université Laval, 1050 avenue de la Médecine, Quebec City, QC, Canada G1V 0A6

Author for correspondence:

Caroline Diorio, Ph.D., Unité de recherche en santé des populations, Centre de recherche du CHU de Québec, Hôpital du St-Sacrement, 1050 chemin Ste-Foy, Quebec City, QC, Canada G1S 4L8, Phone :418.682.7511, ext. 4726, Fax : 418.682.7949, E-mail: [email protected]

Keywords: Fluorescence in situ hybridization (FISH), trastuzumab, HER2, image analysis, tile-sampling analysis, nuclei-sampling analysis, nuclei analysis, accuracy, breast cancer

- 152 -

Résumé L’hybridation fluorescente in situ (FISH) est une technique fréquemment utilisée pour la détermination du statut du récepteur 2 du facteur de croissance épidermique humain (HER2) dans le cancer du sein. Des logiciels d’analyse d’image automatique ont été développés afin de quantifier les signaux. Certains de ces logiciels utilisent ce qu’on appelle un «classificateur par tuile», un algorithme à travers lequel le logiciel quantifie les signaux fluorescents des images sur la base de tuiles carrées. Puisque la taille de la tuile ne correspond pas toujours à celle d’un seul noyau de cellule cancéreuse, un nouveau classificateur capable de déterminer le statut HER2 par noyaux, le «classificateur par noyau», a été développé.

Nous avons évalué la performance du «classificateur par noyau» pour la détermination de l’amplification du gène HER2. Dans une cohorte de 64 spécimens de cancer du sein, nous avons comparé les résultats obtenus par décompte manuel et ceux obtenus avec le nouveau classificateur.

La concordance générale entre le décompte manuel et l’analyse d’image par noyau automatique était de 98,4% et après une étape de correction humaine facultative, la concordance était de 100%.

Nous concluons que le classificateur par noyau représente un outil fiable pour l’analyse quantitative automatique des signaux fluorescents HER2 dans les noyaux des spécimens de cancer du sein.

- 153 -

Abstract Amplification of the human epidermal growth factor receptor 2 (HER2) is a prognostic marker for poor clinical outcome and a predictive marker for therapeutic response to targeted therapies in breast cancer patients. With the introduction of anti-HER2 therapies, accurate assessment of HER2 status has become essential. Fluorescence in situ hybridization (FISH) is a widely used technique for the determination of HER2 status in breast cancer. However, the manual signal enumeration is time-consuming. Therefore, several companies like MetaSystems have developed automated image analysis software. Some of these signal enumeration software employ the so-called “tile-sampling classifier”, a programming algorithm through which the software quantifies fluorescent signals in images on the basis of square tiles of fixed dimensions. Considering that the size of tile does not always correspond to the size of a single tumor cell nucleus, some users argue that this analysis method might not completely reflect the biology of cells. For that reason, MetaSystems has developed a new classifier which is able to recognize nuclei within tissue sections in order to determine the HER2 amplification status on nuclei basis. We call this new programming algorithm “nuclei-sampling classifier”. In this study, we evaluated the accuracy of the “nuclei- sampling classifier” in determining HER2 gene amplification by FISH in nuclei of breast cancer cells. To this aim, we randomly selected from our cohort 64 breast cancer specimens (32 non-amplified and 32 amplified) and we compared results obtained through manual scoring and through this new classifier. The new classifier automatically recognized individual nuclei. The automated analysis was followed by an optional human correction, during which the user interacted with the software in order to improve the selection of cell nuclei automatically selected. Overall concordance between manual scoring and automated nuclei-sampling analysis was 98.4% (100% for non-amplified cases and 96.9% for amplified cases). However, after human correction, concordance between the two methods was 100%. We conclude that the nuclei-based classifier is a new available tool for automated quantitative HER2 FISH signals analysis in nuclei in breast cancer specimen and it can be used for clinical purposes.

- 154 -

Introduction The human epidermal growth factor receptor 2 (HER2) gene is located on chromosome 17 and encodes a transmembrane tyrosine kinase receptor protein [1]. HER2 gene amplification and receptor overexpression, which occur in 15% to 20% of human breast cancers, are important prognostic markers for poor prognosis, including a more aggressive disease and a shorter survival [2]. Moreover, HER2-positive status is a predictive marker of response to trastuzumab therapy in both metastatic and adjuvant settings [3, 4]. An accurate evaluation of HER2 status is therefore crucial for identification of patients who would most likely benefit from targeted anti-HER2 therapies. Currently, there are several Food and Drug Administration (FDA)-approved methods to evaluate HER2 status, such as immunohistochemical (IHC) assessment of HER2 protein expression or evaluation of HER2 gene amplification using in situ hybridization (ISH), most commonly, fluorescent ISH (FISH) [5, 6]. FISH assay is considered to be one of the reference methods for HER2 evaluation in breast cancer, as it accurately predicts response to trastuzumab therapy [7]. Patients are eligible for trastuzumab therapy when their breast cancer specimens are positive at IHC (i.e. 3+) and/or amplified at FISH (ratio> 2.2). However, patients whose tumor specimen is equivocal at FISH (ratio between 1.8 and 2.2) but whose ratio is ≥ 2.0 represent also potential candidates for targeted treatment.

The classical evaluation method for gene amplification, the manual signal enumeration by visual estimation, is a rather time-consuming analysis. Therefore, several companies have developed automated signal quantification systems, which operate through a computer with scanning and image analysis software like Metafer 4 produced by MetaSystems [8, 9]. The programming algorithm (the so called “classifier”) currently used by the latter determines the ratio between the average copy number for HER2 to average copy number for chromosome 17 (CEP17) on the basis of equi-sized square tiles [10]. This programming algorithm is defined as “tile-sampling classifier”. However, as the size of tile does not always correspond to the size of a single tumor cell nucleus, some postulate that results obtained might therefore not completely reflect the biology of single cells. To encounter this statement, MetaSystems have recently developed a new programming algorithm, the “nuclei-sampling classifier”, which is able to automatically quantify fluorescent signals in nuclei within tissue sections. In this study, we have compared results obtained with the reference method, the manual scoring, with those obtained with the new nuclei-sampling classifier from

- 155 -

MetaSystems in 64 clearly non-amplified (n=32) and clearly amplified (n=32) breast cancer specimens.

Material and Methods

Case selection A total of 4641 invasive breast cancer cases were identified among all specimens that were examined by FISH between 2009 and 2012 at the Service de Pathologie at

CHU-Hôpital du Saint-Sacrement, Québec, Canada. Of these examined cases, 3802 were non-amplified, 636 were amplified and 203 were equivocal. After Ethical Review Board approval, we randomly selected 32 clearly non-amplified and 32 clearly amplified cases among all non-amplified (n= 3802) and amplified (n= 636) cases from our cohort and 32 cases among all equivocal cases (n= 203).

Fluorescence In Situ Hybridization HER2 gene copy number was evaluated using the FDA-approved PathVysion™ HER2 DNA Probe kit (Abbott Molecular, Des Plaines, IL/ Inter Medico, Markham, Canada), according to manufacturer’s directions. Briefly, deparaffinized 4-μm sections were immersed in 0.2 N HCl for 20 minutes. Slides were placed in pre-treatment solution at 98°C for 30 minutes and then subjected to protease digestion at 37°C for 5 minutes. Slides were hybridized with PathVysion HER2 DNA probe mixture containing a HER2 DNA probe (labeled with Spectrum Orange) and a CEP17 DNA probe (labeled with Spectrum Green). The CEP17 DNA probe allows for a correction of HER2 gene copy number to the number of copies of chromosome 17. Slide glass coverslips were applied and sealed with rubber cement. Slides were then denatured at 74°C for 2 minutes and hybridized overnight at 37°C in a humidified hybridization chamber (ThermoBrite™, Abbott Molecular/ DAKO, Glostrup, Denmark). On the following day, slides were washed in a post-hybridization buffer at 73.5°C for 2 minutes and dried in the dark. Nuclei were subsequently counterstained with 10 μL of 4’,6-diamidino- 2-phenylindole (DAPI). Slides were stored in the dark at 4°C until signal enumeration.

Manual scoring (reference method) Analysis of fluorescent signals were performed with an epifluorescence microscope Axio Imager M1 (Zeiss, Göttingen, Germany), equipped with a triple-band filter (DAPI/green/orange). Slides were visually scored according to the protocol described in PathVysion HER2 package insert. Briefly, slides were first analyzed at low magnification

- 156 -

using a DAPI filter to identify areas of invasive carcinoma showing optimal tissue digestion and non-overlapping nuclei within tumor areas selected by a pathologist. Slides were analyzed by trained technologists, and results were validated by two breast pathologists with experience in FISH interpretation (SJ or CC). Results were reported according to the ASCO/CAP and Canadian HER2 scoring criteria [5, 6]. Specimens with a HER2/CEP17 ratio of > 2.2 were considered amplified, those with a ratio between 1.8 and 2.2 were considered as equivocal, while specimens with a ratio of <1.8 were considered non-amplified [5, 6]. Average copy numbers of HER2 and CEP17 from at least 20 randomly selected nuclei, each with at least one HER2 and one CEP17 signal, from different areas of the invasive carcinoma were counted and HER2/CEP17 ratio calculated. For equivocal cases, ratios were calculated from at least 60 tumor cells. Equivocal cases were counted by two independent trained technologists and reviewed by the pathologists of this study.

Tile-sampling classifier Automated fluorescence signal analysis was performed through the FDA-approved MetaSystemsTM image analysis system. The capture station is composed of a scanner, an automated fluorescence microscope, a M4+ CCD camera with Mercury Lamp HBO 100 and a computer with the microscope scanning and analysis software Metafer 4 with

“tile-sampling” method (MetaSystems, Altussheim, Germany). The Local Area Network (LAN)-connected capture station worked on PC Intel Core2 Duo CPU E8500 3.16 GHz, 3.25 GB of RAM server, with 1280x1024 pixel size. After selection of 5 to 10 non-overlapping fields of infiltrating carcinoma by trained technologists within tumor areas identified by a pathologist, fields’ images were automatically captured and analyzed by the software. In analogy to manual scoring, representative images with optimal tissue digestion and non- overlapping nuclei were selected. All images were captured at 400x microscope magnification. Minimal integration time was 0.04 seconds. Image size was 1088 x 880 pixels x 8 bits (between 40 and 90 MB, in MetaSystems format .TRN, with 255 gray levels). A classifier is a programming algorithm which defines how images are captured and analyzed by the software. The classifier used for this type of analysis, the “tile-sampling classifier”, permitted extensive tumor sampling by placing non-overlapping equi-sized square tiles in counterstain images (DAPI image). Square tiles were 71 pixels in size and were generally on the order of the size of one or two nuclei. Classifier placed tiles in regions of images where cellular material was the highest, in order to include as much cellular material and as little empty space as possible in tiles. Classifier recognized these regions through the DAPI

- 157 -

filter. Aim of this strategy was to maximize the total fluorescence intensity covered (Figure 6.1a) [10]. Captured images underwent image processing through several filters (Gaussian smoothing filter, TopHat filter and Laplace) and application of a counterstain mask [10]. Image analysis software carried out FISH spot counting through analysis of fluorescence signals present within tiles using several signal colour channels. HER2 spots were recognized through Spectrum Orange filter and were defined as object with an area of 0.05µm2, a distance of 0.8µm between objects and an intensity of 33% after image processing. CEP17 spots were recognized through Spectrum Green filter and were defined as object with an area of 0.18µm2, a distance of 0.5µm between the objects and an intensity of 30% after image processing. Single tiles were rejected by the classifier when less than 40% of the surface was occupied with nuclei, or when they contained only one fluorescent signal (only one orange or only one green fluorescent signal). Cases were rejected when they contained less than 32 tiles. Time consumption for analysis by tile-sampling classifier on a local station was between 3 and 5 minutes. Image analysis software calculated signal ratio by dividing the average HER2 and CEP17 spot number per tile [10]. Results were reported using the same scoring criteria as for manual scoring (non-amplified, equivocal, amplified).

Nuclei-sampling classifier The same images produced during the tile-sampling analysis were analyzed with a new Metafer 4 classifier (MetaSystems). The new classifier was tested on the analysis station, which was composed of a PC Intel Core2 Duo CPU E4600 2.4GHz, 1GB of RAM server with 1280x1024 pixel size. Analysis station was LAN-connected with capture station. Size of images was identical as for the tile-sampling analysis, as we used the same images (1088 x 880 pixels x 8 bits). This new classifier automatically recognized individual cells in counterstain image through segmentation of nuclei, i.e. the outlining of individual’s cell nucleus. Nuclei were recognized by the classifier when they had an object area between 12µm2 and 400µm2 with certain roundness in DAPI images. Color of nuclei outlines informed the user of which nuclei have been considered for analysis. Nuclei that were automatically segmented and considered in analysis showed green outlines, while pre- segmented nuclei (nuclei that were automatically segmented but not considered in the analysis) were represented by a white outline (Figure 6.1b). The software recognized appropriate nuclei for analysis on the basis of size and shape of nuclei and on quality of fluorescent signals. In this paper, we defined this method as the “nuclei-sampling analysis”

- 158 -

to differentiate it from the “tile-sampling analysis”. The nuclei-sampling analysis was performed blinded from results obtained by manual scoring and tile-sampling analysis. This automated nucleus segmentation was followed by an optional human correction, in which the user interacted with interface in order to improve selection of cell nuclei automatically selected by the software. This optional interactive phase required an interactive touch screen (e.g. the WACOM® DTU-2231, 1920 x 1080 pixels, MetaSystems). The interaction was performed using a mouse or an interactive pen display. The user utilized the following interaction options: addition of nuclei (not yet considered during the automated segmentation) to be analyzed, selection of pre-segmented nuclei (nuclei that were recognised by the software but not considered in the analysis), deletion of automatically selected nuclei, partition of overlapping nuclei, or connection of separated objects (for instance parts of the same nucleus that were accidentally separated during the automated segmentation) (Figure 6.1c). Any of these operations led to automatic updates of signal ratio results. The software carried out FISH spot counting for both the automated and the human corrected nuclei-sampling analyses in the same way as for the tile-sampling classifier. Again, results for both automated and human corrected nuclei-sampling analyses were reported using the same scoring criteria as for manual scoring (non-amplified, equivocal, amplified).

Results Validation of the nuclei-sampling classifier in non-amplified and amplified breast cancer specimens

In order to validate the new classifier, we examined the concordance of results obtained by manual scoring and by nuclei-sampling analysis in 32 clearly non-amplified and 32 clearly amplified cases, chosen randomly in our cohort of breast cancer patients. Our selection criteria fulfill the requirements of the ASCO/CAP and Canadian recommendations for HER2 testing in breast cancer for the validation of a new test. Indeed, these guidelines recommend that a new test has to be compared with a reference test in at least 25 samples, ideally by using 50% cases that are clearly positive and 50% cases that are clearly negative [5, 6]. Table 6.1. shows the comparison between results obtained by manual scoring, the reference method, and the tile-sampling analysis and nuclei-sampling analysis in these 64 breast cancer specimens. For the nuclei-sampling analysis, concordance between reference method (manual scoring) and results obtained with the automated analysis (that is before human correction) was 100% for non-amplified cases and 96.9% for amplified cases. Overall

- 159 -

concordance rate was 98.4%. One case out of 32 considered as amplified by manual scoring was considered as borderline equivocal (ratio of 2.2) with the nuclei-sampling analysis. However, after human correction, concordance between manual scoring and nuclei- sampling analysis was 100% for both the non-amplified and the amplified cases. Concordance between manual scoring and tile-sampling analysis was 100% for both the non-amplified and the amplified cases.

Determination of the accuracy of the nuclei-sampling classifier on special specimens In a closer analysis of the randomly selected amplified cases, we observed that all examined cases were amplified with Homogeneously Staining Region (HSR). HSR are large clusters of HER2 fluorescence signals indicating the presence of HER2 gene amplification in tandem repeats (Figure 6.2a.). In samples in which HSR occurred and individual signals could not be counted, the software adopted a different spot-counting analysis, which evaluated the signal area in the HER2 channel instead of individual spot counts [10]. As in the population of this study approximately 4% of amplified cases do not show HSR (Figure 6.2b.), we examined the accuracy of the new classifier also on these less common cases. In our cohort, we identified 28 amplified cases without HSR. Although cases were recognized as amplified without HSR on the basis of results obtained through tile-sampling analysis, manual scoring represented the reference method also for these cases. Results for amplified cases without HSR are summarized in Table 6.2. Of these 28 amplified cases without HSR, 21 cases were classified as amplified with the automated nuclei-sampling analysis, and this number increased to 24 after human correction. Concordance between manual scoring and nuclei- sampling analysis was 75% (21/28). Of the 7 discordant cases that showed a ratio ≤ 2.2 with the automated nuclei-sampling analysis, 4 had a ratio between 2.0 and 2.2, whereas 3 had a ratio smaller than 2.0. After human correction, the concordance between the manual scoring and the nuclei-sampling analysis was 86% (24/28). Of the 4 discordant cases that showed a ratio ≤ 2.2 with nuclei-sampling analysis after human correction, 3 cases showed a ratio between 2.0 and 2.2, whereas 1 case showed a ratio smaller than 2.0. Concordance between tile-sampling analysis and manual scoring method was 100%. This result is not surprising, as the samples without HSR were identified on the basis of results obtained through tile-sampling analysis.

- 160 -

Determination of the accuracy of the nuclei-sampling classifier on equivocal specimens To further evaluate the accuracy of the new classifier, we examined 32 equivocal cases. However, only 29 cases have been analyzed, since in 3 cases, images were of very poor quality. Equivocal cases constitute approximately 5% of our cohort. Results for equivocal cases are summarized in Table 6.3. Of these 29 equivocal cases, 9 cases were classified as equivocal, 19 as non-amplified and one case was classified as amplified with tile- sampling analysis. With nuclei-sampling analysis after human correction, 17 cases were classified as equivocal, 11 as non-amplified and one case was classified as amplified. Concordance between manual scoring and tile-sampling analysis was 31% (9/29), whereas the concordance between manual scoring and nuclei-sampling analysis after human correction was 59% (17/29). Among the 15 cases whose ratio was ≥ 2.0 at the manual scoring, 3 (20%) were also ≥ 2.0 at the tile-sampling method, whereas 6 out of these 15 cases (40%) were ≥ 2.0 at the nuclei-sampling method after human correction. Among the 14 cases whose ratio was < 2.0 at the manual scoring, 13 (93%) were also < 2.0 at the tile- sampling method, whereas 12 out of these 14 cases (86%) were < 2.0 at the nuclei-sampling method after human correction.

Reproducibility of results All cases analyzed in this study (clearly non-amplified, clearly amplified, equivocal and amplified cases without HSR) were assessed blindly by a second independent observer. The results obtained by the two observers were similar (data not shown).

Discussion Our results showed an excellent concordance between manual scoring, our reference method, and nuclei-sampling analysis for clearly non-amplified and clearly amplified cases. Indeed, the concordance of results for non-amplified cases was 100%, both for the automated and the human corrected nuclei-sampling analyses. For amplified cases, the concordance between the two methods was 96.9% for the automated nuclei-sampling analysis and rose to 100% following human correction. These concordance rates with manual scoring results fulfill the ASCO/CAP requirements of concordance greater than 95% for clearly amplified and non-amplified cases [6].

Our results are consistent to those obtained by Theodosiou and collaborators [11]. In their study, they examined the utility of an image analysis software (EIKONA3D, Alpha Tec Ltd)

- 161 -

for the evaluation of HER2 amplification in nuclei in 100 breast cancer cases from two institutions. Similar to the analysis software present here, the user had the possibility to manually correct the results obtained through the automated nuclei segmentation. They found a very good overall concordance (92.8%) between the results obtained by manual scoring by an expert and those obtained with the image analysis software. Similar to our results, the concordance for non-amplified cases was 100%, whereas the concordance for amplified cases was lower, 74.1% [11].

In this work, we validated the new Metafer 4 classifier in 64 breast cancer specimens (32 non-amplified and 32 amplified cases), chosen randomly among eligible clearly non- amplified and amplified cases of our cohort, as required from the ASCO/CAP and the Canadian guidelines for HER2 testing in breast cancer for validation of a new test. Accordingly to these recommendations, a new test has to be compared with the reference test in at least 25 samples, ideally by using 50% cases unequivocally positive and 50% cases unequivocally negative [5, 6]. The new classifier evaluated here was able to recognize cell nuclei on the image and therefore to calculate HER2 FISH ratio on nucleus basis. Moreover, this new classifier allowed the user to interact with the software during an optional interactive phase, in order to improve the selection of cells automatically selected by the software. In this study, we defined this method as “nuclei-sampling analysis” to differentiate it from “tile-sampling analysis”, which was performed with the Metafer 4 classifier currently used. This classifier, in fact, calculated HER2 FISH ratio on the basis of equi-sized tiles placed by the software on images. In order to validate this new classifier, we compared results obtained through manual scoring of slides, considered as the reference method, with those obtained through nuclei-sampling analysis. Moreover, we analyzed the accuracy of both the automated and the human corrected nuclei-sampling analyses.

As all randomly selected amplified cases analyzed in this study were amplified with HSR, we decided to evaluate the accuracy of this new classifier on the less common amplified cases without HSR. These cases represented indeed about 4% of all amplified cases in our cohort of breast cancer patients. For the 28 cases without HSR that we have analyzed, concordance between manual scoring and automated nuclei-based method was 75%. After human correction, concordance between the two methods rose to 86%. Considering that patients whose specimen is equivocal at FISH (ratio between 1.8 and 2.2) but whose ratio is ≥ 2.0 represent also potential candidates for trastuzumab treatment, 4 patients out of 7 discordant cases at the automated nuclei-sampling analysis, and 3 cases out of 4 discordant

- 162 -

cases at the nuclei-sampling analysis after human correction would therefore be eligible to receive a targeted treatment. We noticed that some discordant cases were polysomic or monosomic (4 out of 7 discordant cases) and we postulate that this aneuploidy status could explain the discordance. It has been reported that biological variance reduces sampling efficiency [12]. Indeed, higher biological variance associated with aneuploidy status could have had an impact on the spot counting by the software and this could explain the discrepancy with results obtained by manual scoring. Moreover, quality of the images of some discordant cases (2 out of 7 discordant cases) was poor (cell nuclei were blurred in the image), which could also be an additional explanation for this discrepancy.

With the aim to further analyse the accuracy of the new classifier, we also examined equivocal cases, which represent about 5% of our cohort population. Overall concordance between manual scoring and tile-sampling method was 31%, whereas concordance between manual scoring and nuclei-sampling method after human correction was 59%. If equivocal cases were splitted in those with ratio ≥ 2.0 and those with ratio < 2.0, we noticed that twice as many cases were correctly classified with a ratio ≥ 2.0 using nuclei-sampling method after human correction as compared to tile-sampling method. So even if concordance between manual scoring and nuclei-sampling method was not optimal, these results suggest that nuclei-sampling method is more reliable than tile-sampling method for the identification of patients who could potentially benefit from targeted anti-HER2 therapies. Similar to the amplified cases without HSR, we also noticed that some discordant cases were aneuploid (4 out of 12 discordant cases). Also, in 2 out of 12 discordant cases the quality of images was poor (cell nuclei were blurred in images).

Tile-sampling method has been developed by MetaSystems and other companies in order to overcome the difficulties that are frequently encountered when fluorescent signals are enumerated via automated image analysis software. Firstly, reliable separation of overlapping nuclei in tissue sections is very difficult, especially in dense packed tissues like breast cancer. Secondly, it is arduous for image analysis software to automatically distinguish distinct cell populations (normal and tumor cells) present in analyzed fields. To overcome these difficulties, the Metafer 4 software places non-overlapping tiles of equal size on images in order to cover the majority of nuclear material and therefore quantify fluorescent signals. Moreover, a ratio estimation algorithm was introduced with the aim to improve the accuracy of results of the automated analysis in samples in which distinct cell populations are present [10, 13].

- 163 -

Although the tile-sampling analysis is in general well performing [8, 9], the nuclei-based analysis offers some advantages compared to the tile-based analysis. Firstly, the way in which the new classifier selects nuclei for analysis coincides better to what the user does when the user is analyzing a sample. Whereas the nuclei-sampling analysis recognizes cell nuclei, the tile-sampling analysis places equi-sized tiles on the image. In addition, as the size of tile does not always correspond to the size of a single nucleus, nuclei are often truncated during tile-sampling. As a consequence, one single tile may contain signals from multiple nuclei or only part of a nucleus. This can be disadvantageous especially in cases of chromosome 17 monosomy or polysomy, where exact number of CEP17 signal per cell is relevant. Secondly, the nuclei-based analysis offers the advantage that the user can improve the selection of cell nuclei that have been automatically selected by the software through active interaction with the software. During the interactive phase, the user can add nuclei that were not considered during the automated selection, delete unsuitable nuclei, divide overlapping nuclei or connect separated pieces of the same nucleus. This optional, interactive phase requires additional time, in average 7 minutes for equivocal cases and amplified cases without HSR and 4 minutes for non-amplified cases and amplified cases with HSR, but it is very helpful and effective especially in difficult cases, for instance in samples with abundant stroma or intermixed normal cells. In fact, we observed a better concordance between results obtained by the reference method and those obtained with the nuclei-based analysis after the interactive phase, compared to results obtained with the automated analysis. Theodosiou and collaborators observed similar results using a similar method. In their hand, manual correction required up to 5 minutes for each case (non- amplified and amplified cases) and it was particularly useful in cases with low image quality [11].We noticed that among all functions that the user could choose during the interactive phase, the delete function was the most effective one. In fact, when discordant cases were evaluated blindly by a second independent observer who used exclusively the delete function, results obtained by the two observers were similar (data not shown). We may therefore conclude that the delete function is very effective in improving the results obtained with the automated nuclei-based analysis. Moreover, the time necessary for human correction can additionally be reduced if only the delete function is used during the interactive phase (6 minutes in average for equivocal cases and amplified cases without HSR). The automated nuclei-sampling analysis required between 3 and 5 minutes per case, depending on the cellularity of images. Our image analysis software is slower compared to others, for example Matlab, which required 3.5 seconds for analysis of a single image on

- 164 -

local server [14]. As we analyzed between 5 and 10 images for each case, Matlab would have taken between 17.5 and 35 seconds to evaluate a case.

In this study, the reference method was represented by the manual scoring of specimen. A closer examination of nuclei automatically selected by the software during nuclei-based analysis allowed us to observe how the user can also be biased when the user is analysing a case. In particular, human eyes have a tendency to pay more attention to those cell nuclei in which more fluorescent signals are present. One could therefore argue that human brain considers those nuclei more attractive and preferentially chooses them during the manual signal enumeration. Nuclei-based analysis, on the contrary, selects nuclei on the basis of the shape of cell nuclei and on quality of fluorescent signals and is therefore more “neutral” in the choice of the nuclei. Therefore, eligible nuclei that have less fluorescent signals (and may be judged as less attractive by human brain) are also taken into account for analysis from software. Opinions on this topic are divergent. Whereas some underline that software do not always select the most appropriate nuclei for analysis [11], others claim that results obtained with automated analysis are more accurate especially in amplified and borderline cases, as manual analysis of HER2 signals can only be estimated when probe signals cluster closely together [15]. Another advantage of image analysis system over manual scoring is that storing of captured images allows archiving of cases for future study.

Some limitations are associated with the new classifier. Accuracy of the new classifier to recognize nuclei is markedly reduced in images with dense packing of cells or in images in which DAPI counterstain is blurred. As mentioned above, results obtained with any quantitative image analysis software depend tremendously on the fields chosen by the observer for analysis. If the fields chosen are not representative of the sample, results obtained by quantitative image analysis can be rather different from those obtained through manual scoring. This issue is common to all diagnostic algorithms. Reliable sampling procedure is prerequisite for diagnostic accuracy in virtual microscopy [12, 16].

Standardization of images capture is a central point in the development of a diagnostic algorithm in virtual microscopy [17]. In our study, optimal specification for the capture of images from FISH HER2 slides hybridized with PathVysion™ HER2 DNA Probe kit (image size, size of tiles, identification criteria for HER2 and CEP17 spots, segmentation criteria for nuclei, filtering) has been previously established using 400 slides (personal communication, Ulrich Klingbeil, MetaSystems). Quality of captured images is in general excellent, since

- 165 -

quality and intensity of fluorescence signals are reproducible and background is very low. Dissimilar to other algorithms used in object-related diagnosis [16], fluorescent spot identification is less problematic. In contrary to other structures within tissue that are difficult to be recognized, fluorescent spots are easily identified by the classifier, as they are mostly of the same size and intensity, except for HSR cases. However, spot identification is also reliable in amplified cases with HSR (where spot dimensions can be more variable), since the software adopted a different spot-counting analysis (evaluation of the signal area in the HER2 channel instead of individual spot counts) [10].

Tissue-based diagnosis has been subjected to remarkable changes following the introduction of new technologies. For instance, technological advances in tissue-based diagnosis allow the implementation of digitized images into routine clinical pathology. Virtual pathology has several advantages compared to conventional microscopy. For example, virtual pathology allows archiving of virtual images, promotes continuing education as well as interactive remote consultation between pathologists [18]. Moreover, it has been reported that analysis of digitized slides gives results as accurate as that obtained through conventional microscopy [19, 20]. However, one critical point is whether the diagnostic information contained in the virtual slides reliably reflect the real whole slide. In this context, the adopted sampling procedure plays a central role [12]. This is an important point to consider, when the efficacy of virtual diagnostic algorithms are compared [12]. Both the tile- sampling classifier and the nuclei-sampling classifier are based on a stratified and passive sampling method as defined in Kayser et al. [12]. However, whereas the tile-sampling classifier recognizes nuclear material through the DAPI filter (and put square tiles on the image, where the DAPI coloration is the strongest), the nuclei-sampling classifier recognizes single nuclei within tissue on the basis of nuclei characteristics, such as nuclei size and roundness. Spot recognition and spot counting is effectuated in same way for both methods.

In our clinical context, pathologists share virtual images and results via a LAN platform. This form of information sharing represents one of the first steps towards the so called “Grid technology”. A Grid is an open and dynamic communication system consisting of connected nodes (i.e. servers) that are linked together via Internet connections and share certain communication rules in using open standards [21]. The Grid technology will also have an impact on the quality in tissue-based diagnosis as such implementation will require appropriate standardization of legal, medical and technological aspects associated with virtual pathology [17].

- 166 -

Conclusions In summary, we observed an excellent concordance between results obtained by manual scoring and those by nuclei-sampling analysis in 32 clearly non-amplified and 32 clearly amplified breast cancer specimens. However, accurate determination of HER2 amplification in equivocal cases (ratio between 1.8 and 2.2) remains a challenge [5, 6]. Manual assessment of these cases, therefore, remains the standard procedure. We conclude that the new image analysis software Metafer 4 classifier is a reliable tool to evaluate the unequivocal status of HER2 in breast cancer specimens and it is ready to be implemented in clinics, as it offers several advantages compared to the Metafer 4 classifier currently used. Moreover, although more time-consuming, human correction after the completion of the automated nuclei-sampling analysis is recommended, especially in particular cases (like those with abundant stroma), as this operation leads to an improvement of the results obtained during the automated analysis.

List of abbreviations HER2: human epidermal growth factor receptor 2; CEP17: chromosome 17 centromere; FISH: Fluorescence in situ hybridization; ASCO/CAP: American Society of Clinical Oncology/College of American Pathologists; HSR: Homogeneously Staining Region.

Competing interests The new Metafer 4 software version and the interactive touch screen were kindly provided by MetaSystems.

Authors’ contributions DF participated in the conception of the study, data collection, analysis and interpretation of results and wrote the manuscript. SJ, CC, FS and CD participated in the conception of the study, data collection, analysis and interpretation of results and reviewed the manuscript. LP participated in the analysis and interpretation of the results and editing of the manuscript. All authors read and approved the final manuscript.

Acknowledgements This project was supported by grants from Hoffmann-La Roche Limited and La Fondation des Hôpitaux Enfant-Jésus – Saint-Sacrement. CD is a Junior Investigator of the Canadian Research Society (2011-700657).

- 167 -

References 1. Rubin I, Yarden Y: The basic biology of HER2. Ann Oncol 2001, 12 Suppl 1:S3-8.

2. Soerjomataram I, Louwman MW, Ribot JG, Roukema JA, Coebergh JW: An overview of prognostic factors for long-term survivors of breast cancer. Breast Cancer Res Treat 2008, 107(3):309-330.

3. Gianni L, Dafni U, Gelber RD, Azambuja E, Muehlbauer S, Goldhirsch A, Untch M, Smith I, Baselga J, Jackisch C et al: Treatment with trastuzumab for 1 year after adjuvant chemotherapy in patients with HER2-positive early breast cancer: a 4-year follow-up of a randomised controlled trial. Lancet Oncol 2011, 12(3):236-244.

4. Piccart-Gebhart MJ, Procter M, Leyland-Jones B, Goldhirsch A, Untch M, Smith I, Gianni L, Baselga J, Bell R, Jackisch C et al: Trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer. N Engl J Med 2005, 353(16):1659-1672.

5. Hanna W, O'Malley F P, Barnes P, Berendt R, Gaboury L, Magliocco A, Pettigrew N, Robertson S, Sengupta S, Tetu B et al: Updated recommendations from the Canadian National Consensus Meeting on HER2/neu testing in breast cancer. Curr Oncol 2007, 14(4):149-153.

6. Wolff AC, Hammond ME, Schwartz JN, Hagerty KL, Allred DC, Cote RJ, Dowsett M, Fitzgibbons PL, Hanna WM, Langer A et al: American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol 2007, 25(1):118-145.

7. Ross JS, Linette GP, Stec J, Clark E, Ayers M, Leschly N, Symmans WF, Hortobagyi GN, Pusztai L: Breast cancer biomarkers and molecular medicine. Expert Rev Mol Diagn 2003, 3(5):573-585.

8. Stevens R, Almanaseer I, Gonzalez M, Caglar D, Knudson RA, Ketterling RP, Schrock DS, Seemayer TA, Bridge JA: Analysis of HER2 gene amplification using an automated fluorescence in situ hybridization signal enumeration system. J Mol Diagn 2007, 9(2):144-150.

9. Tubbs RR, Pettay JD, Swain E, Roche PC, Powell W, Hicks DG, Grogan T: Automation of manual components and image quantification of direct dual label fluorescence

- 168 -

in situ hybridization (FISH) for HER2 gene amplification: A feasibility study. Appl Immunohistochem Mol Morphol 2006, 14(4):436-440.

10. Lörch T, Piper J, Tomisek JD: "Tile sampling": a new method for the automated quantitative analysis of samples with high cell density and its application to Her2 scanning. Proceedings of Quantitative Molecular Cytogenetics 2002.

11. Theodosiou Z, Kasampalidis IN, Karayannopoulou G, Kostopoulos I, Bobos M, Bevilacqua G, Aretini P, Starita A, Lyroudia K, Pitas I: Evaluation of FISH image analysis system on assessing HER2 amplification in breast carcinoma cases. Breast 2008, 17(1):80- 84.

12. Kayser K, Schultz H, Goldmann T, Gortler J, Kayser G, Vollmer E: Theory of sampling and its application in tissue based diagnosis. Diagn Pathol 2009, 4:6.

13. Piper J, Lörch T, Poole I, Tomisek JD: Computing the Her2:CEP-17 ratio of tumor cells in breast cancer tissue sections by analysis of the FISH spot counts of a tiles sampling. Proceedings of Quantitative Molecular Cytogenetics 2002.

14. Markiewicz T: Using MATLAB software with Tomcat server and Java platform for remote image analysis in pathology. Diagn Pathol 2011, 6(Suppl 1):S18.

15. Ellis CM, Dyson MJ, Stephenson TJ, Maltby EL: HER2 amplification status in breast cancer: a comparison between immunohistochemical staining and fluorescence in situ hybridisation using manual and automated quantitative image analysis scoring techniques. J Clin Pathol 2005, 58(7):710-714.

16. Kayser K, Radziszowski D, Bzdyl P, Sommer R, Kayser G: Towards an automated virtual slide screening: theoretical considerations and practical experiences of automated tissue-based virtual diagnosis to be implemented in the Internet. Diagn Pathol 2006, 1:10.

17. Kayser K, Gortler J, Goldmann T, Vollmer E, Hufnagl P, Kayser G: Image standards in tissue-based diagnosis (diagnostic surgical pathology). Diagn Pathol 2008, 3:17.

18. Pantanowitz L, Valenstein PN, Evans AJ, Kaplan KJ, Pfeifer JD, Wilbur DC, Collins LC, Colgan TJ: Review of the current state of whole slide imaging in pathology. J Pathol Inform 2011, 2:36.

- 169 -

19. Kayser K, Kayser G: Virtual Microscopy and Automated Diagnosis. In Virtual Microscopy and Virtual Slides in Teaching, Diagnosis and Research 2005:Edited by: Gu J, Ogilvie R. Boca Raton: Taylor & Francis.

20. Florell SR, Smoller BR, Boucher KM, Grossman D, Harris RM, Bowen GM, Leachman SA: Sampling of melanocytic nevi for research purposes: a prospective, pilot study to determine effect on diagnosis. J Am Acad Dermatol 2008, 59(5):814-821.

21. Gortler J, Berghoff M, Kayser G, Kayser K: Grid technology in tissue-based diagnosis: fundamentals and potential developments. Diagn Pathol 2006, 1:23.

- 170 -

Table 6.1. Comparison of results obtained by different methods for non-amplified and amplified cases (n=64) Tile-sampling analysis Nuclei-sampling analysis Automated analysis After human correction Manual scoring Nam Eq Am Nam Eq Am Nam Eq Am Nam 32 0 0 32 0 0 32 0 0 Am 0 0 32 0 1 31 0 0 32 Nam: non-amplified; Eq: equivocal; Am: amplified.

Table 6.2. Comparison of results obtained by different methods for amplified cases without HSR (n=28) Tile-sampling analysis Nuclei-sampling analysis Automated analysis After human correction Manual < 2.0 2.0 - 2.2 > 2.2 < 2.0 2.0 - 2.2 > 2.2 < 2.0 2.0 - 2.2 > 2.2 scoring > 2.2 0 0 28 3 4 21 1 3 24 HSR = Homogeneously Staining Region.

Table 6.3. Comparison of results obtained by different methods for equivocal cases (n=29) Tile-sampling analysis Nuclei-sampling analysis, after human correction Manual < 1.8 ≥ 1.8 - < 2.0 2.0 – 2.2 > 2.2 < 1.8 ≥ 1.8 - < 2.0 2.0 – 2.2 > 2.2 scoring ≥ 1.8 - < 2.0 12 1 1 0 7 5 1 1 2.0 – 2.2 7 5 2 1 4 5 6 0

- 171 -

Figure 6.1. Image analysis of fluorescence signals

- 172 -

Figure 6.2. HER2 fluorescence in situ hybridization in amplified cases

Figure legends:

Figure 6.1. Image analysis of fluorescence signals. a). Automated quantitative image analysis through the tile-sampling classifier. Non-overlapping square tile were placed on DAPI-counterstain image in order to maximize the nuclear material covered. Each tile may contain a single nucleus or portions of one or more nuclei. The software effectuated the spot count in each tile. b). Quantitative image analysis through the automated nuclei-sampling classifier. The classifier automatically recognized individual nuclei in counterstain image (same image as in Figure 6.1a). Nuclei that were automatically recognized and considered in the analysis showed green outlines, while nuclei that were automatically recognized but not considered in the analysis had a white outline. The software effectuated spot count in each nucleus. c). After completion of the automated image analysis, the user could improve the selection of automatically selected nuclei via interaction with the software. In this particular case, some nuclei that were not yet considered during the automatic segmentation (Figure 6.1b) were selected by the user, some automatically selected nuclei were deleted and some overlapping nuclei were divided (yellow line between two overlapping nuclei).

Figure 6.2. HER2 fluorescence in situ hybridization in amplified cases. a). Amplified case showing Homogeneously Staining Region (HSR). HSR are large clusters of HER2 fluorescence signals indicating the presence of HER2 gene amplification in tandem repeats. b). Amplified case without HSR. HER2 signals are identifiable as single spots.

- 173 -

Chapter 7: Evaluation of human epidermal growth factor receptor 2 (HER2) single nucleotide polymorphisms (SNPs) in normal and breast tumor tissues and their link with breast cancer prognostic factors

The Breast 2016 ; 30: 191-196.

Daniela Furrer1,2,4, Julie Lemieux1,2,3,4, Marc-André Côté3,4, Louise Provencher1,2,3,4, Christian Laflamme1,2,3, Frédéric Barabé4,5, Simon Jacob1,2,4, Annick Michaud1,2, Caroline Diorio1,2,3,4

1Centre de Recherche sur le cancer de l’Université Laval, 1050 Avenue de la Médecine, Québec (QC), G1V 0A6, Canada, 2Axe Oncologie, Centre de Recherche du CHU de Québec-Université Laval, 1050 Avenue de la Médecine, Québec (QC), G1V 0A6, Canada; 3Centre des maladies du sein Deschênes-Fabia, Hôpital du Saint-Sacrement, 1050, chemin Ste-Foy, Québec (Qc) G1S 4L8, Canada; 4Faculté de Médecine, Université Laval, 1050 Avenue de la Médecine, Québec (QC), G1V 0A6, Canada; 5Axe Maladies infectieuses et immunitaires, Centre de Recherche du CHU de Québec-Université Laval, 1050 Avenue de la Médecine, Québec (QC), G1V 0A6, Canada

Corresponding Author: Caroline Diorio, Ph.D., Associate professor, Département de médecine sociale et préventive, Université Laval, Axe oncologie, Centre de recherche du CHU de Québec, Centre des maladies du sein Deschênes-Fabia, Hôpital du Saint- Sacrement, 1050, chemin Ste-Foy, local J0-16, Québec (Qc) G1S 4L8, Tel.: 418-682-7511 poste 84726, Fax.: 418-682-7949, e-mail adresse: [email protected]

Keywords: breast neoplasms; HER2 gene; gene polymorphism; cancer progression

- 174 -

Résumé L’amplification du gène qui code pour le récepteur 2 du facteur de croissance épidermique humain (HER2) est associée à une mauvaise survie chez les patientes atteintes d’un cancer du sein. Le gène HER2 contient plusieurs polymorphismes, donc les plus étudiés sont Ile655Val et Ala1170Pro.

Nous avons évalué l’association entre ces deux polymorphismes d’HER2 dans les tissus normaux du sein et dans les tissus cancéreux du sein et les facteurs pronostiques du cancer du sein dans une cohorte de 73 femmes atteintes d’un cancer du sein HER2-positif. Les polymorphismes d’HER2 ont été mesurés en utilisant l’essai TaqMan.

Le polymorphisme Ala1170Pro dans le tissu normal et le tissu cancéreux du sein était associé à l’âge au diagnostic, à la taille tumorale et à l’envahissement ganglionnaire. Aucune association entre le polymorphisme Ile655Val et les facteurs pronostiques n’a été observée. Cependant, nous avons observé des différences significatives dans la distribution du génotypage Ile655Val et Ala1170Pro entre les tissus normaux et les tissus cancéreux du sein.

Nous concluons que seul le polymorphisme Ala1170Pro est associé à des facteurs pronostiques chez les patientes atteintes d’un cancer du sein HER2-positif et que les deux polymorphismes pourraient jouer un rôle dans la carcinogénèse chez ces patientes.

- 175 -

Abstract Amplification of the human epidermal growth factor receptor 2 (HER2) gene is associated with worse prognosis and decreased overall survival in breast cancer patients. The HER2 gene contains several polymorphisms; two of the best-characterized HER2 polymorphisms are Ile655Val and Ala1170Pro. The aim of this study was to evaluate the association between these two HER2 polymorphisms in normal breast and breast cancer tissues and known breast cancer prognostic factors in a retrospective cohort study of 73 women with non-metastatic HER2-positive breast cancer. HER2 polymorphisms were assessed in breast cancer tissue and normal breast tissue using TaqMan assay. Ala1170Pro polymorphism in normal breast tissue was associated with age at diagnosis (p=0.007), tumor size (p=0.004) and lymphovascular invasion (p=0.06). Similar significant associations in cancer tissues were observed. No association between the Ile655Val polymorphism and prognostic factors were observed. However, we found significant differences in the distribution of Ile655Val (p=0.03) and Ala1170Pro (p=0.01) genotypes between normal breast and breast tumor tissues. This study demonstrates that only the Ala1170Pro polymorphism is associated with prognostic factors in HER2-positive breast cancer patients. Moreover, our results suggest that both HER2 polymorphisms could play a significant role in carcinogenesis in non-metastatic HER2-positive breast cancer women.

- 176 -

Introduction The human epidermal growth factor receptor 2 (HER2) gene is located on chromosome 17 and encodes a transmembrane tyrosine kinase receptor protein involved in the regulation of cell growth, differentiation and survival (1). Alterations in this gene, including gene amplification, are associated with enhanced activation of intracellular signaling pathways, such as MAPK and PI3K/Akt, leading to uncontrolled cell proliferation (2). In human breast cancer, HER2 gene amplification and receptor overexpression, which occur in 15% to 20% of patients, are important markers for poor prognosis, including a more aggressive disease and a shorter survival (3-5).

Although the HER2 gene contains several single nucleotide polymorphisms (SNPs), the most investigated variants are the Ile655Val and the Ala1170Pro polymorphisms (6, 7). The Ile655Val polymorphism, located in the transmembrane coding region at codon 655, results in the substitution of Isoleucine (Ile:ATC) for Valine (Val:GTC) (7). A preclinical study suggested that the Val substitution might have functional consequences, as it predisposes the HER2 receptor to assume an active conformation resulting in enhanced activity of the tyrosine kinase domain and increased tumorigenic potential of breast cancer cells (8). The Ala1170Pro polymorphism, which has been identified in the coding region of the carboxyl- terminal regulatory domain at codon 1170, leads to the substitution of Alanine (Ala:GCC) for Proline (Pro:CCC) (9).

The role of both polymorphisms as prognostic factors and as markers of breast cancer risk is still controversial (9-14). To date, only a few studies have investigated the clinical and biological relevance of these polymorphisms in breast cancer (9, 10, 12). Some studies have found that the Ile655Val and Ala1170Pro polymorphisms are associated with breast cancer prognostics factors (9, 13, 15-20). Val allele carriers have been shown to have significantly higher serum levels of HER2 and lower survival rates compared to Ile allele carriers, suggesting that the Ile655Val polymorphism could be considered a potential biomarker of poor prognosis in breast cancer (20). Concerning the Ala1170Pro polymorphism, Tommasi et al. showed that in breast cancer the Pro allele was associated with positive estrogen receptor status (9).

Allelic imbalance, defined as the loss or gain of one allele (21), has been reported for the Ile655Val polymorphism in a cohort of patients with ovarian (n=8) and breast (n=27) cancers (HER2-negative and HER2-positive) (22). As a change from the Ile/Val genotype in normal

- 177 -

tissue to the Val/Val genotype in cancer tissue was observed, and considering that the Val/Val genotype was overrepresented in some HER2-positive breast tumors, the authors suggested that this genotype might potentially promote growth and survival of breast cancer cells.

The aim of this study was to analyze the association of Ile655Val and Ala1170Pro polymorphisms in normal breast and breast cancer tissues with known breast cancer prognostic factors in a cohort of 73 non-metastatic HER2-positive breast cancer patients.

Materials and Methods

Study population and data collection The study population consisted of 73 women with non-metastatic HER2-positive breast cancer diagnosed (or chemotherapy completed) between July 1st, 2005 and January 1st, 2010. The study took place at the Centre des maladies du sein Deschênes-Fabia, a specialized breast center in Quebec City, Canada. HER2 status was evaluated on mastectomy and segmental resection specimens by immunohistochemistry (IHC) (HercepTest, Glostrup, Denmark) and fluorescence in situ hybridization (PathVysionTM HER2 DNA Probe Kit, Abbott Molecular, Des Plaines, IL/ Inter Medico, Markham, Canada). HER2 results were validated by the study pathologist (SJ) and reported according to the 2013 ASCO/CAP scoring criteria (23). Estrogen and progesterone receptor status was determined by IHC by the study pathologist (SJ), and tumors were considered positive if ≥ 1% of tumor cells showed clear nuclear staining (24). Information on tumor characteristics and prognostic factors were collected from medical records. Informed consent was obtained from all individual participants included in the study. Ethical approval for the study was obtained from the Research Ethics Committee of the CHU of Quebec. Since only HER2- positive breast cancer patients are eligible to receive targeted anti-HER2 treatment (25), we decided to focus our analysis in order to gain knowledge on HER2 function for this specific group of breast cancer patients.

Ile655Val and Ala1170Pro polymorphisms DNA was extracted from normal breast and breast tumor tissues (both formalin-fixed, paraffin-embedded tissues) using the Qiagen DNA Mini Kit (Qiagen, Mississauga, Ontario, Canada) (26). Normal tissues were located more than 1 cm from any lesion as selected by the study pathologist (SJ). For 19 patients, DNA was also extracted from blood using the

- 178 -

same protocol as for breast tissues. DNA samples were then blindly genotyped for two SNPs [Ile655Val (rs1136201) and Ala1170Pro (rs1058808)] located in HER2 gene by TaqMan SNP Genotyping Assays (Life Technologies, Burlington, Ontario, Canada). Each batch also included negative (no DNA) and positive controls to ensure accuracy of genotyping. In this study, concordance of genotyping between normal breast tissue and blood was 100% for both SNPs. Deviation from Hardy-Weinberg equilibrium was assessed for each SNP; p- value was 0.39 for Ile655Val SNP and 0.03 for Ala1170Pro SNP. The linkage disequilibrium strength was evaluated with Lewontin’s D’ statistic for pair-wise SNPs and was 0.02 which is similar to another study (10).

Statistical analysis Fisher’s exact test was used to evaluate the association between the presence of HER2 polymorphisms and breast cancer prognostic factors. The Bowker’s test of symmetry was applied to compare differences in the distribution of Ile655Val and Ala1170Pro polymorphisms between normal breast and breast tumor tissues.

Results Clinicopathological characteristics of the study population are presented in Table 7.1. Mean patient age at diagnosis was 54 years (SD ± 4.5); 37% of patients were younger than 50 years. Forty patients (54.8%) had a body mass index higher than 25 Kg/m2. Forty-eight patients (65.8%) had grade III tumors. Lymphovascular invasion was observed in 39 women (53.4%), whereas positive lymph nodes were found in 49 patients (67.1%). Expression of estrogen and progesterone receptors was observed in 45 (61.6%) and 31 women (42.5%), respectively. Nine patients (12.3%) had tumors larger than 5 cm at diagnosis.

The associations of HER2 Ile655Val and Ala1170Pro polymorphisms in normal breast and breast tumor tissues with prognostic factors are shown in Table 7.2. Whereas no significant association between Ile655Val polymorphism and prognostic factors was found, either in normal or in tumor tissues, we observed a significant association between Ala1170Pro polymorphism and some prognostic factors in both normal and cancer tissues, including age at the diagnosis, lymphovascular invasion, and tumor size. In younger women (≤ 50 years), there were more patients showing homo- or heterogeneous Pro genotypes (Pro/Ala or Pro/Pro genotype) than the homozygous genotype Ala/Ala, whether in normal (p=0.007) or cancer (p=0.01) tissues. Regarding tumor size, we observed that all patients having large tumors (> 5cm) at diagnosis had Pro/Ala or Pro/Pro genotypes; no homozygous genotype

- 179 -

Ala/Ala was observed in our cohort for tumors larger than 5 cm (p=0.004 and p=0.002 in normal and cancer tissues, respectively). Compared to the patients without lymphovascular invasion, a greater proportion of patients with lymphovascular invasion showed the Pro/Ala or the Pro/Pro genotypes, both in normal (p=0.06) and tumor (p=0.06) tissues, although not statistically significant. Finally, a trend was observed between Ala1170Pro polymorphism and positive estrogen receptor status in cancer tissues (p=0.14).

Genotypes from normal breast and breast tumor tissues for the Ile655Val polymorphism and for the Ala1170Pro polymorphism were available for 71 and 69 patients, respectively. Genotype distributions in normal breast and breast tumor tissues are shown in Table 7.3. The distribution of the Ile655Val SNP in normal breast tissue was 52 patients with Ile/Ile (73.2%), 16 patients with Ile/Val (22.6%), and 3 patients with Val/Val (4.2%). In breast tumor tissues, some genotype changes were observed compared to normal breast tissues. Whereas the proportion of patients with Ile/Ile remained almost the same (73.2% in normal breast tissues vs. 77.5% in breast cancer tissues), we observed a decrease in the proportion of patients with Ile/Val (22.6% in normal breast tissues vs. 7.0% in breast cancer tissues) and an increase in the proportion of patients with Val/Val (4.2% in normal breast tissues vs. 15.5% in breast cancer tissues). Globally, the genotype distribution of the Ile655Val polymorphism was significantly different between the normal breast and the breast tumor tissues (p=0.03). Similar results were observed for the Ala1170Pro polymorphism. The distribution of the Ala1170Pro SNP in normal breast tissue was 32 patients with Ala/Ala (46.4%), 23 patients with Pro/Ala (33.3%) and 14 patients with Pro/Pro (20.3%) compared to 35 patients with Ala/Ala (50.7%), 9 patients with Pro/Ala (13.1%) and 25 patients with Pro/Pro (36.2%) in breast cancer tissue. This difference in the genotype distribution for the Ala1170Pro SNP between normal and tumor tissues was significant (p=0.01), and we mainly observed changes from the Pro/Ala to Pro/Pro genotypes.

Discussion None of the prognostic factors tested in this study correlated with the Ile655Val polymorphism. Age at diagnosis and tumor size were associated with the Ala1170Pro polymorphism in both normal breast and breast cancer tissues, whereas a trend was observed for lymphovascular invasion in both tissues and for positive estrogen receptor status in breast cancer tissue. We also observed a genotype change at codons 655 and 1170 between normal breast and breast tumor tissues.

- 180 -

In this sample from the French Canadian population, the genotype distribution of the Val allele was similar to that of other Caucasian populations (27, 28).

Some SNPs have been reported to influence the development and progression of several cancer types, including breast, prostate, colorectal, and lung cancers (29-32). In breast cancer, the association between Ile655Val polymorphism and prognostic factors has been evaluated in some studies (9, 15-20, 33-35). In contrast to our results, Val containing genotypes (Ile/Val or Val/Val) have been associated with localised breast cancer (17), smaller tumor size (9, 15), a more advanced cancer stage (16, 19, 20), higher histological grade (19), positive lymph node status (19), progesterone-positive receptor status (33) and young age at diagnosis (9). Notably, the Val/Val genotype was more common in patients with invasive disease (9, 13), and has been shown to be associated with lymphovascular invasion (19), positive lymph node status (18, 20) and reduced survival (20). The fact that our population consists exclusively of HER2-positive breast cancer patients may partly explain these discordances. In line with our observations, other studies did not find a significant association between the Ile655Val polymorphism and prognostic factors (35, 36).

To the best of our knowledge, only two studies have evaluated the association between Ala1170Pro polymorphism and breast cancer prognostic factors (9, 37). Similar to our results, Tommasi and collaborators observed that the Pro containing genotypes (Pro/Ala or Pro/Pro) were more commonly found in estrogen receptor-positive breast cancers (9). It has been postulated that this observation could be explained by the fact that HER2 gene has a modulating role in the steroid pathway (9, 38). However, Su et al. in a cohort of 303 primary breast cancer patients (HER2 negative and HER2 positive) did not observe a significant association between Ala1170Pro polymorphism and estrogen receptor status and other breast cancer prognostic factors (37). The same authors reported an association between the Ala/Ala and Ala/Pro genotypes and HER2 protein overexpression, whereas they did not observe an association between Ile655Val SNP and receptor overexpression. Similarly, Cresti and collaborators did not observe an association between Ile655Val SNP and HER2 expression. However, they reported that Ala/Pro and Pro/Pro genotypes were associated with HER2 overexpression (39).

In previous studies, genotype changes at codons 655 and 1170 of the HER2 gene have been evaluated using an unpaired approach, i.e. the genotype of breast cancer patients has been compared to that of normal healthy subjects (9, 12, 15, 16, 18, 27, 35, 40, 41).

- 181 -

Genotyping was also mainly performed using peripheral whole blood. In the majority of these studies, the genotype and allele frequency of both HER2 polymorphisms were not significantly different between breast cancer cases and normal subjects (9, 12, 15, 16, 18, 27, 40, 41). However, in a meta-analysis of case-control studies including 10,642 breast cancer patients and 11,259 controls, the Val allele frequency (determined mostly from blood samples) was significantly higher in cancer cases than in healthy subjects (p=0.047) (14). Similarly, a recent study found that the Ile/Val genotype was significantly higher in breast cancer patients compared to healthy age-matched controls (p<0.009) in a Turkish population (35). These conflicting reports can be explained by several factors such as selection bias, sample size and ethnic background.

Although our sample size was small, we observed a significant difference in the HER2 SNPs genotype distribution between normal breast and tumor breast tissues in the same patient. This suggests that HER2 genotype can change during carcinogenesis. Because genotyping was determined using normal breast tissue, we cannot completely exclude measurement errors. However, we observed an excellent genotype concordance between blood and normal breast tissue. Moreover, if information bias had occurred, this would have decreased the strength of the association. Furthermore, our results are in line with the results reported in previous similar studies (22, 39, 42). Studies that examined genotype changes in Ile655Val and Ala1170Pro SNPs between normal breast and breast tumor tissues in both HER2-positive and HER2-negative patients using a paired approach have observed that loss of heterozygosity (LOH) was more frequent in HER2-positive breast cancer patients (22, 39, 42). It has been proposed that LOH observed among HER2-positive patients was caused by mono-allelic amplification of HER2 (42). In agreement with these results, we also observed an allelic imbalance at codons 655 and 1170 with an overrepresentation of the Val/Val and Pro/Pro genotypes in tumor tissues.

Our study shows that the wild-type allele at codons 655 and 1170 tends to be lost during carcinogenesis. This genotype change between normal breast and tumor breast tissues suggests that HER2 polymorphisms may play a role in breast cancer development.

Conclusions Our work suggests that HER2 polymorphisms could play a significant role in carcinogenesis in non-metastatic HER2-positive breast cancer women. The significant differences observed in the genotype distribution of HER2 SNPs between normal breast and tumor breast tissues

- 182 -

suggest that HER2 polymorphisms may play a role in breast carcinogenesis in a Caucasian population. However, only Ala1170Pro polymorphism was associated with some breast cancer prognostic factors in non-metastatic HER2-positive breast cancer patients either in normal breast or breast cancer tissues. A larger sample is needed to further evaluate these associations.

Acknowledgments Clinical specimens were provided by the Fondation du cancer du sein du Québec and the Banque de tissus et de données of the Réseau de recherche sur le cancer of the FRQS, which is affiliated with the Canadian Tumour Repository Network. DF received doctoral fellowships from the Fonds de recherche du Québec - Santé and the Laval University Cancer Research Center. CD is a recipient of the Canadian Breast Cancer Foundation-Canadian Cancer Society Capacity Development award (award #703003) and the FRQS Research Scholar. JL was a Clinical Research Scholar from the Fonds de la recherche en santé du Québec at the time of the study.

Conflict of interest statement: None to declare.

Funding: This project was supported by la Fondation des Hôpitaux Enfant-Jésus – Saint- Sacrement. Part of this study was funded by an unrestricted grant from Hoffmann-La Roche Limited.

Ethical approval: Ethical approval for the study was obtained from the Research ethics committee of the CHU of Quebec.

References 1. Rubin I, Yarden Y. The basic biology of HER2. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2001;12 Suppl 1:S3-8.

2. Yarden Y, Sliwkowski MX. Untangling the ErbB signalling network. Nature reviews Molecular cell biology. 2001;2(2):127-37.

3. Konecny G, Pauletti G, Pegram M, Untch M, Dandekar S, Aguilar Z, et al. Quantitative association between HER-2/neu and steroid hormone receptors in hormone receptor-positive primary breast cancer. Journal of the National Cancer Institute. 2003;95(2):142-53.

- 183 -

4. Quenel N, Wafflart J, Bonichon F, de Mascarel I, Trojani M, Durand M, et al. The prognostic value of c-erbB2 in primary breast carcinomas: a study on 942 cases. Breast cancer research and treatment. 1995;35(3):283-91.

5. Ross JS, Slodkowska EA, Symmans WF, Pusztai L, Ravdin PM, Hortobagyi GN. The HER-2 receptor and breast cancer: ten years of targeted anti-HER-2 therapy and personalized medicine. Oncologist. 2009;14(4):320-68.

6. Han W, Kang D, Lee JE, Park IA, Choi JY, Lee KM, et al. A haplotype analysis of HER-2 gene polymorphisms: association with breast cancer risk, HER-2 protein expression in the tumor, and disease recurrence in Korea. Clin Cancer Res. 2005;11(13):4775-8.

7. Papewalis J, Nikitin A, Rajewsky MF. G to A polymorphism at amino acid codon 655 of the human erbB-2/HER2 gene. Nucleic Acids Res. 1991;19(19):5452.

8. Fleishman SJ, Schlessinger J, Ben-Tal N. A putative molecular-activation switch in the transmembrane domain of erbB2. Proceedings of the National Academy of Sciences of the United States of America. 2002;99(25):15937-40.

9. Tommasi S, Fedele V, Lacalamita R, Bruno M, Schittulli F, Ginzinger D, et al. 655Val and 1170Pro ERBB2 SNPs in familial breast cancer risk and BRCA1 alterations. Cell Oncol. 2007;29(3):241-8.

10. Benusiglio PR, Lesueur F, Luccarini C, Conroy DM, Shah M, Easton DF, et al. Common ERBB2 polymorphisms and risk of breast cancer in a white British population: A case control study. Breast cancer research : BCR. 2005;7(2):204-2099.

11. Dahabreh IJ, Murray S. Lack of replication for the association between HER2 I655V polymorphism and breast cancer risk: a systematic review and meta-analysis. Cancer epidemiology. 2011;35(6):503-9.

12. Frank B, Hemminki K, Wirtenberger M, Bermejo JL, Bugert P, Klaes R, et al. The rare ERBB2 variant Ile654Val is associated with an increased familial breast cancer risk. Carcinogenesis. 2005;26(3):643-7.

- 184 -

13. Millikan RC, Hummer AJ, Wolff MS, Hishida A, Begg CB. HER2 codon 655 polymorphism and breast cancer: results from kin-cohort and case-control analyses. Breast Cancer Res Treat. 2005;89(3):309-12.

14. Tao W, Wang C, Han R, Jiang H. HER2 codon 655 polymorphism and breast cancer risk: a meta-analysis. Breast Cancer Res Treat. 2009;114(2):371-6.

15. Kallel I, Kharrat N, Al-fadhly S, Rebai M, Khabir A, Boudawara TS, et al. HER2 polymorphisms and breast cancer in Tunisian women. Genet Test Mol Biomarkers. 2010;14(1):29-35.

16. Lee SC, Hou MF, Hsieh PC, Wu SH, Hou LA, Ma H, et al. A case-control study of the HER2 Ile655Val polymorphism and risk of breast cancer in Taiwan. Clin Biochem. 2008;41(3):121-5.

17. McKean-Cowdin R, Kolonel LN, Press MF, Pike MC, Henderson BE. Germ-line HER- 2 variant and breast cancer risk by stage of disease. Cancer Res. 2001;61(23):8393-4.

18. Naidu R, Yip CH, Taib NA. Polymorphisms of HER2 Ile655Val and cyclin D1 (CCND1) G870A are not associated with breast cancer risk but polymorphic allele of HER2 is associated with nodal metastases. Neoplasma. 2008;55(2):87-95.

19. Papadopoulou E, Simopoulos K, Tripsianis G, Tentes I, Anagnostopoulos K, Sivridis E, et al. Allelic imbalance of HER-2 codon 655 polymorphism among different religious/ethnic populations of northern Greece and its association with the development and the malignant phenotype of breast cancer. Neoplasma. 2007;54(5):365-73.

20. Papadopoulou E, Tripsianis G, Anagnostopoulos K, Tentes I, Kakolyris S, Galazios G, et al. The influence of serum HER-2 levels and HER-2 codon 655 polymorphism on breast cancer outcome. Neoplasma. 2008;55(2):113-21.

21. Wagner JR, Ge B, Pokholok D, Gunderson KL, Pastinen T, Blanchette M. Computational analysis of whole-genome differential allelic expression data in human. PLoS computational biology. 2010;6(7):e1000849.

- 185 -

22. Puputti M, Sihto H, Isola J, Butzow R, Joensuu H, Nupponen NN. Allelic imbalance of HER2 variant in sporadic breast and ovarian cancer. Cancer Genet Cytogenet. 2006;167(1):32-8.

23. Wolff AC, Hammond ME, Hicks DG, Dowsett M, McShane LM, Allison KH, et al. Recommendations for human epidermal growth factor receptor 2 testing in breast cancer: American Society of Clinical Oncology/College of American Pathologists clinical practice guideline update. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2013;31(31):3997-4013.

24. Hammond ME, Hayes DF, Dowsett M, Allred DC, Hagerty KL, Badve S, et al. American Society of Clinical Oncology/College Of American Pathologists guideline recommendations for immunohistochemical testing of estrogen and progesterone receptors in breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2010;28(16):2784-95.

25. Gianni L, Dafni U, Gelber RD, Azambuja E, Muehlbauer S, Goldhirsch A, et al. Treatment with trastuzumab for 1 year after adjuvant chemotherapy in patients with HER2- positive early breast cancer: a 4-year follow-up of a randomised controlled trial. Lancet Oncol. 2011;12(3):236-44.

26. Gagnon JF, Sanschagrin F, Jacob S, Tremblay AA, Provencher L, Robert J, et al. Quantitative DNA methylation analysis of laser capture microdissected formalin-fixed and paraffin-embedded tissues. Exp Mol Pathol. 2010;88(1):184-9.

27. Keshava C, McCanlies EC, Keshava N, Wolff MS, Weston A. Distribution of HER2(V655) genotypes in breast cancer cases and controls in the United States. Cancer Lett. 2001;173(1):37-41.

28. Millikan R, Eaton A, Worley K, Biscocho L, Hodgson E, Huang WY, et al. HER2 codon 655 polymorphism and risk of breast cancer in African Americans and whites. Breast Cancer Res Treat. 2003;79(3):355-64.

29. Cheng I, Plummer SJ, Neslund-Dudas C, Klein EA, Casey G, Rybicki BA, et al. Prostate cancer susceptibility variants confer increased risk of disease progression. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for

- 186 -

Cancer Research, cosponsored by the American Society of Preventive Oncology. 2010;19(9):2124-32.

30. Cowper-Sal lari R, Zhang X, Wright JB, Bailey SD, Cole MD, Eeckhoute J, et al. Breast cancer risk-associated SNPs modulate the affinity of chromatin for FOXA1 and alter gene expression. Nature genetics. 2012;44(11):1191-8.

31. Shiels MS, Engels EA, Shi J, Landi MT, Albanes D, Chatterjee N, et al. Genetic variation in innate immunity and inflammation pathways associated with lung cancer risk. Cancer. 2012;118(22):5630-6.

32. Yao L, Tak YG, Berman BP, Farnham PJ. Functional annotation of colon cancer risk SNPs. Nature communications. 2014;5:5114.

33. Han X, Diao L, Xu Y, Xue W, Ouyang T, Li J, et al. Association between the HER2 Ile655Val polymorphism and response to trastuzumab in women with operable primary breast cancer. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2014;25(6):1158-64.

34. Millikan RC, Player JS, Decotret AR, Tse CK, Keku T. Polymorphisms in DNA repair genes, medical exposure to ionizing radiation, and breast cancer risk. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology. 2005;14(10):2326-34.

35. Ozturk O, Canbay E, Kahraman OT, Fatih Seyhan M, Aydogan F, Celik V, et al. HER2 Ile655Val and PTEN IVS4 polymorphisms in patients with breast cancer. Mol Biol Rep. 2013;40(2):1813-8.

36. Watrowski R, Castillo-Tong DC, Wolf A, Schuster E, Fischer MB, Speiser P, et al. HER2 Codon 655 (Ile/Val) Polymorphism and Breast Cancer in Austrian Women. Anticancer Res. 2015;35(12):6667-70.

37. Su Y, Jiang Y, Sun S, Yin H, Shan M, Tao W, et al. Effects of HER2 genetic polymorphisms on its protein expression in breast cancer. Cancer epidemiology. 2015;39(6):1123-7.

- 187 -

38. Chung YL, Sheu ML, Yang SC, Lin CH, Yen SH. Resistance to tamoxifen-induced apoptosis is associated with direct interaction between Her2/neu and cell membrane estrogen receptor in breast cancer. Int J Cancer. 2002;97(3):306-12.

39. Cresti N, Lee J, Rourke E, Televantou D, Jamieson D, Verrill M, et al. Genetic variants in the HER2 gene: Influence on HER2 overexpression and loss of heterozygosity in breast cancer. Eur J Cancer. 2016;55:27-37.

40. Siddig A, Mohamed AO, Kamal H, Awad S, Hassan AH, Zilahi E, et al. HER-2/neu Ile655Val polymorphism and the risk of breast cancer. Ann N Y Acad Sci. 2008;1138:84-94.

41. Zubor P, Kajo K, Stanclova A, Szunyogh N, Galo S, Dussan CA, et al. Human epithelial growth factor receptor 2[Ile655Val] polymorphism and risk of breast fibroadenoma. Eur J Cancer Prev. 2008;17(1):33-8.

42. Benz CC, Fedele V, Xu F, Ylstra B, Ginzinger D, Yu M, et al. Altered promoter usage characterizes monoallelic transcription arising with ERBB2 amplification in human breast cancers. Genes Chromosomes Cancer. 2006;45(11):983-94.

- 188 -

Table 7.1. Characteristics of the study population (n=73)

N % Age (years) ≤ 50 27 37.0 > 50 46 63.0 Body mass index (kg/m2)a ≤ 25kg/m2 32 43.8 > 25 kg/m2 40 54.8 Grade I/II 25 34.2 III 48 65.8 Lymphovascular invasion No 34 46.6 Yes 39 53.4 Lymph node status Negative 24 32.9 Positive 49 67.1 Estrogen receptor status Negative 28 38.4 Positive 45 61.6 Progesterone receptor status 42 57.5 Negative 31 42.5 Positive Tumor size ≤ 5 cm 64 87.7 > 5 cm 9 12.3 a Non-adding numbers indicate missing data

- 189 -

Table 7.2. Association of HER2 Ile655Val and Ala1170Pro polymorphisms in normal and tumor breast tissues with prognostic factors

Normal tissues Cancer tissues Prognostic factors Ile655Val polymorphism Ile/Ile Ile/Val or P-value Ile/Ile Ile/Val or P-value Val/Val Val/Val Age (years) 0.60 0.56 ≤ 50 18 (24.7)a 9 (12.3) 19 (26.4) 7 (9.7) > 50 34 (46.6) 12 (16.4) 37 (51.4) 9 (12.5) Body mass index (kg/m2) b 0.29 0.39 ≤ 25 kg/m2 20 (28.2) 11 (15.5) 22 (31.4) 9 (12.9) > 25 kg/m2 31 (43.7) 9 (12.7) 32 (45.7) 7 (10.0) Grade 0.59 1.00 I/II 19 (26.0) 6 (8.2) 19 (26.4) 5 (7.0) III 33 (45.2) 15 (20.6) 37 (51.4) 11 (15.3) Lymphovascular invasion 0.44 1.00 No 26 (35.6) 8 (11.0) 26 (36.1) 7 (9.7) Yes 26 (35.6) 13 (17.8) 30 (41.7) 9 (12.5) Lymph node status 0.79 0.77 Negative 18 (24.7) 6 (8.2) 18 (25.0) 6 (8.3) Positive 34 (46.6) 15 (20.6) 38 (52.8) 10 (13.9) Estrogen receptor status 0.59 0.38 Negative 20 (27.4) 6 (8.2) 22 (31.0) 4 (5.6) Positive 32 (43.8) 15 (20.6) 33 (46.5) 12 (16.9) Progesterone receptor 0.44 0.40 status 29 (39.7) 9 (12.3) 32 (44.5) 7 (9.7) Negative 23 (31.5) 12 (16.5) 24 (33.3) 9 (12.5) Positive Tumor size 0.26 0.36 ≤ 5 cm 47 (64.4) 17 (23.3) 51 (70.8) 13 (18.1) > 5 cm 5 (6.9) 4 (5.5) 5 (6.9) 3 (4.2) Ala1170Pro polymorphism Ala/Ala Pro/Ala or P-value Ala/Ala Pro/Ala P-value Pro/Pro or Pro/Pro Age (years) 0.007 0.01 ≤ 50 6 (8.2) 21 (28.8) 8 (11.4) 18 (25.7) > 50 26 (35.6) 20 (27.4) 28 (40.0) 16 (22.9) Body mass index (kg/m2) b 0.82 1.00 ≤ 25 kg/m2 13 (18.3) 19 (26.8) 14 (20.6) 15 (22.1) > 25 kg/m2 17 (24.0) 22 (31.0) 20 (29.4) 19 (27.9) Grade 0.80 1.00 I/II 10 (13.7) 15 (20.6) 12 (17.1) 12 (17.1) III 22 (30.1) 26 (35.6) 24 (34.3) 22 (31.4) Lymphovascular invasion 0.06 0.06 No 19 (26.0) 15 (20.6) 20 (28.6) 11 (15.7) Yes 13 (17.8) 26 (35.6) 16 (22.9) 23 (32.9) Lymph node status 1.00 1.00 Negative 11 (15.1) 13 (17.8) 12 (17.1) 11 (15.7) Positive 21 (28.8) 28 (38.4) 24 (34.3) 23 (32.9)

- 190 -

Table 7.2 (continued) Normal tissues Cancer tissues Prognostic factors Ala1170Pro polymorphism Ala/Ala Pro/Ala or P-value Ala/Ala Pro/Ala P-value Pro/Pro or Pro/Pro Estrogen receptor status 0.46 0.14 Negative 13 (17.8) 13 (17.8) 16 (22.9) 9 (12.9) Positive 19 (26.1) 28 (38.3) 20 (28.5) 25 (35.7) Progesterone receptor 0.35 0.23 status 19 (26.0) 19 (26.0) 22 (31.4) 15 (21.4) Negative 13 (17.9) 22 (30.1) 14 (20.0) 19 (27.2) Positive Tumor size 0.004 0.002 ≤ 5 cm 32 (43.8) 32(43.8) 36 (51.4) 26 (37.1) > 5 cm 0 (0.0) 9 (12.3) 0 (0.0) 8 (11.4) a N (%); b Non-adding numbers indicate missing data.

- 191 -

Table 7.3. Distribution of HER2 Ile655Val and Ala1170Pro polymorphisms in normal breast and tumor breast tissues

Breast cancer tissue Normal breast tissue Ile655Val polymorphism Ile/Ile Ile/Val Val/Val Total P-value Ile/Ile 50 2 0 52 (73.2) 0.03 Ile/Val 5 3 8 16 (22.6) Val/Val 0 0 3 3 (4.2) Total 55 (77.5) 5 (7.0) 11 (15.5) 71 (100.0) Ala1170Pro polymorphism Ala/Ala Pro/Ala Pro/Pro Total P-value Ala/Ala 32 0 0 32 (46.4) 0.01 Pro/Ala 3 7 13 23 (33.3) Pro/Pro 0 2 12 14 (20.3) Total 35 (50.7) 9 (13.1) 25 (36.2) 69 (100.0)

- 192 -

Chapter 8: Association between tobacco and alcohol consumption, HER2 polymorphisms and response to trastuzumab in HER2-positive breast cancer patients

Submitted to Clinical Breast Cancer

Daniela Furrer a,b,d, Simon Jacob a,b,c,d, Annick Michaud a,b, Louise Provencher a,b,c,d, Julie Lemieux a,b,c,d, Caroline Diorio a,b,c,d a Centre de Recherche sur le cancer de l'Université Laval, 1050 Avenue de la Médecine, Québec, QC, G1V 0A6, Canada ; b Axe Oncologie, Centre de Recherche du CHU de Québec-Université Laval, 1050 Avenue de la Médecine, Québec, QC, G1V 0A6, Canada; c Centre des maladies du sein Deschênes-Fabia, Hôpital du Saint-Sacrement, 1050, chemin Ste-Foy, Québec, QC, G1S 4L8, Canada; d Faculté de Médecine, Université Laval, 1050 Avenue de la Médecine, Québec, QC, G1V 0A6, Canada

Corresponding author :

Caroline Diorio, Axe Oncologie, Centre de Recherche du Centre Hospitalier Universitaire de Québec, Hôpital du Saint-Sacrement, 1050 chemin Ste-Foy, Quebec City, QC G1S 4L8, Canada; Phone: +1-418-682-7511 ext. 84726; Fax: +1-418-682-7949

Keywords : breast neoplasms, HER2 gene, SNP, smoking, tobacco use, NNK, alcohol use, alcohol drinking, ethanol, resveratrol, HER2 inhibitors

- 193 -

Résumé L’administration du trastuzumab a beaucoup amélioré la survie des patientes atteintes d’un cancer HER2-positif. Cependant, il est de plus en plus reconnu que le développement de la résistance au trastuzumab représente un obstacle clinique majeur. Nous avons évalué l’association entre la consommation de tabac et d’alcool, les polymorphismes d’HER2 et la survie sans récidive (SSR) chez les patients atteintes d’un cancer du sein HER2-positif. Dans une cohorte de 236 patientes HER2-positives non-métastatiques traitées au trastuzumab, les données cliniques ont été recueillies dans les dossiers médicaux, la consommation de tabac et alcool par des questionnaires validés et les polymorphismes d’HER2 (Ile655Val et Ala1170Pro) ont été évalués par essai TaqMan. Des modèles multivariés de risques proportionnels de Cox ont été utilisés. Comparé aux non-fumeuses, la SRR était moins bonne chez les patientes qui fumaient avant le diagnostic d’un cancer du sein (Hazard ratio (HR): 2,63 (95% IC 1,48 – 4,68); valeur p = 0,001), et cette association était plus forte chez les patientes qui fumaient > 20 cigarettes/jour avant le diagnostic ou qui étaient fumeuses depuis plus de 20 ans (HR: 3,65 (95% IC 1,35 – 9,89), valeur p= 0,01, et HR: 2,06 (95% IC 1,01 – 4,22), valeur p= 0,05, respectivement). La consommation de tabac pendant le traitement au trastuzumab était associée à la SSR, mais seulement parmi les patientes avec une tumeur négative pour les récepteurs des oestrogènes. Comparé aux non-buveuses, les patientes qui consommaient de l’alcool avant le diagnostic du cancer du avaient une meilleure SSR (HR: 0,56 (95% IC 0,33-0,94), valeur p= 0,03). Aucune association n’a été observée entre la consommation d’alcool pendant le traitement et la SSR. Concernant les polymorphismes d’HER2, comparé aux porteuses Ile/Ile, les porteuses Ile/Val et Val/Val avaient une moins bonne SSR (HR: 4,96 (95% IC 1,45 – 17,02, valeur p= 0,01). Nos résultats suggèrent que la consommation de tabac et d’alcool ainsi que le polymorphisme HER2 Ile655Val pourraient influencer la réponse au trastuzumab. Ces résultats doivent être confirmés dans une étude de cohorte plus grande.

- 194 -

Abstract The administration of trastuzumab has significantly improved the survival of HER2-positive breast cancer patients. However, resistance to trastuzumab has been increasingly recognized as a major clinical obstacle. The objective of this study was to evaluate the association between HER2 polymorphisms, tobacco and alcohol consumption and disease- free survival (DFS) in HER2-positive breast cancer patients. In a cohort of 236 non- metastatic HER2-positive trastuzumab-treated breast cancer patients, clinicopathological information were obtained by review of medical records, tobacco and alcohol consumption by an administered validated questionnaire, and HER2 polymorphisms (Ile655Val and Ala1170Pro) by using TaqMan assay. Cox proportional-hazards multivariate models were used. Compared to non-smokers, patients who smoked before breast cancer diagnosis showed a worse DFS (hazard ratio (HR): 2.63 (95% CI 1.48 – 4.68); p = 0.001). Compared to non-smokers, patients who smoked > 20 cigarettes/day or who spent > two decades smoking before breast cancer diagnosis showed a worse DFS (HR: 3.65 (95% CI 1.35 – 9.89), p= 0.01, and HR: 3.19 (95% CI 1.55 – 6.56), p= 0.002, respectively). Compared to non-smokers, patients who smoked during trastuzumab treatment showed a worse DFS (HR: 4.49 (95% CI 1.26 – 16.00); p= 0.02), but only among patients with estrogen receptor- negative tumors. Compared to non-drinkers, patients who consumed alcohol before breast cancer diagnosis were statistically significantly associated with a better DFS (HR: 0.56 (95% CI 0.33-0.94), p= 0.03). However, this effect was limited to wine consumption. Concerning HER2 polymorphisms, compared to Ile/Ile carriers, HER2 Ile/Val and Val/Val carriers were associated with worse DFS (HR: 4.96 (95% CI 1.45 – 17.02), p=0.01). Our results suggest that tobacco and alcohol consumption as well as HER2 Ile655Val polymorphism could influence trastuzumab response. These results need to be confirmed in a larger cohort study.

- 195 -

Introduction The human epidermal growth factor receptor 2 (HER2) is a receptor tyrosine kinase that activates several growth-promoting signaling pathways including PI3K/AKT and Ras/MAPK (1). HER2 gene amplification and receptor overexpression occur in 15 to 20% of breast cancer patients and are associated with poor prognosis (2). In addition, HER2-positive status is regarded as a predictive marker of response to anti- HER2 therapies (3). The administration of trastuzumab has led to significant improvement in disease-free and overall survival of HER2-positive breast cancer patients when combined with standard chemotherapy in the adjuvant and metastatic settings (4, 5). Despite this noteworthy achievement, primary and acquired resistance to trastuzumab represents a major obstacle in the clinical management of HER2-positive breast cancer patients (6). Therefore, there is an urgent need to identify factors that could influence trastuzumab response in HER2- positive breast cancer patients.

There is increasing evidence that lifestyle factors including tobacco and alcohol consumption might have an impact on survival of breast cancer patients (7-9). Notably, associations between tobacco and ethanol exposure and HER2 have been reported in the literature. It has been shown that 4-(methylnitrosamino)-1-3-(3-pyridyl)-1-butanon (NNK), a potent tobacco-specific carcinogen (10), activates the ERK/MAPK signaling pathway in human normal mammary epithelial cells (11). Furthermore, a study that analyzed the association between tobacco consumption at time of breast cancer diagnosis and risk of recurrence in a cohort of 3,340 breast cancer patients showed that recurrence risk was significantly increased in trastuzumab-naïve HER2-positive breast cancer patients (n=177) who smoked at time of diagnosis (12). The only study that analyzed the association between tobacco use and response to trastuzumab reported that response rate in smokers (former and active) in a cohort of 248 metastatic trastuzumab-treated HER2-positive breast cancer patients was not statistically different from that of never smokers (13). Concerning alcohol consumption, a recent study that analysed the association between alcohol use before breast cancer diagnosis and risk of breast cancer by molecular subtype in a cohort of 105,972 women has highlighted that alcohol consumption might represent a risk factor for HER2-positive breast cancer (14). Furthermore, it has been reported that the stimulatory effect of ethanol on the invasion capacity of breast cancer cell lines depends on the expression levels of HER2 (15). No study has evaluated the impact of alcohol consumption on trastuzumab response in HER2-positive breast cancer patients yet.

- 196 -

In addition to lifestyle factors, it has been reported that genetic factors can influence the efficacy of antineoplastic drugs in breast cancer patients (16, 17). The HER2 gene contains several single nucleotide polymorphisms (SNPs) (18). Two of the most investigated HER2 SNPs are Ile655Val and Ala1170Pro (19, 20). The Ile655Val polymorphism leads to the substitution of Isoleucine for Valine (20). A preclinical study reported that the Val substitution might have functional consequences, as it predisposes the HER2 receptor to assume an active conformation leading to enhanced activity of the tyrosine kinase domain and increased tumorigenic potential of breast cancer cells (21). The Ala1170Pro polymorphism results in the substitution of Alanine for Proline (22). The biological relevance of this polymorphism, however, remains undefined (22). To date, only two studies have investigated the association between HER2 Ile655Val SNP and the response to trastuzumab (23, 24). Whereas Beauclair and collaborators did not observe an association between HER2 genotype and response to trastuzumab in a cohort of 61 patients with advanced HER2-positive breast cancer (23), Han et al. found that Ile655Val SNP was associated with better disease-free survival (DFS) among 212 HER2-positive breast cancer patients who received adjuvant trastuzumab treatment (24).

Thus, we hypothesized that the risk of recurrence among trastuzumab-treated HER2- positive breast cancer patients is increased by tobacco and alcohol consumption and affected by HER2 polymorphisms. The goal of our study was to evaluate the association between tobacco and alcohol consumption and HER2 polymorphisms with the risk of recurrence in a cohort of non-metastatic HER2-positive breast cancer patients treated with trastuzumab.

Material and Methods

Study population and data collection The study population consisted of 237 women with non-metastatic HER2-positive breast cancer diagnosed (or chemotherapy completed) between July 1st, 2005 and January 1st, 2010 and treated with trastuzumab. One patient was excluded due to loss of follow-up during treatment. Therefore, the study population consisted of 236 patients. The study took place at the Centre des Maladies du Sein Deschênes-Fabia (CMSDF), a specialized breast center in Quebec City, Canada. Clinicopathological information was extracted from a registry held by CMSDF. Informed written consent was obtained from all participants. Ethical approval for

- 197 -

the main study and substudy was obtained from the Research Committee of the CHU of Quebec (#DR-002-1227 and DR-002-1265).

Smoking status, including the average number of cigarettes smoked per day and the number of years spent smoking, was assessed at the time of the breast cancer diagnosis (n=236) and during trastuzumab treatment (n=131) using a validated self-administered questionnaire. For 211 out of 236 patients, alcohol consumption before breast cancer diagnosis (yes/no) was documented at the time of the breast cancer diagnosis. Moreover, 128 patients completed a validated self-administered questionnaire that investigated the type (wine, beer and spirits) and quantity of alcohol intake per week before breast cancer diagnosis and during trastuzumab treatment. A graduated beverage-specific frequency questionnaire was employed (25).

Polymorphism substudy Seventy-three patients were included in the substudy. Ile655Val and Ala1170Pro polymorphisms were measured in breast tumor tissues as previously reported (26). Briefly, DNA was extracted from formalin-fixed, paraffin-embedded (FFPE) breast tumor tissues using the Qiagen DNA Mini Kit (Qiagen, Mississauga, Ontario, Canada) (27). DNA samples were blindly genotyped for two SNPs [Ile655Val (rs1136201) and Ala1170Pro (rs1058808)] located in HER2 gene by TaqMan SNP Genotyping Assays (Life Technologies, Burlington, Ontario, Canada). Each batch also included negative (no DNA) and positive controls to ensure genotyping accuracy. Deviation from Hardy-Weinberg equilibrium was evaluated for each SNP; p-value was 0.39 for Ile655Val SNP and 0.03 for Ala1170Pro SNP. The linkage disequilibrium strength was assessed with Lewontin's D' statistic for pairwise SNPs and was 0.02, which is similar to another study (28).

Study end points For all analyses, the primary end point was disease-free survival (DFS), defined as time from date of mastectomy to first recurrence (locoregional or distant). Patients who were alive at the last follow-up were censored at the last follow-up date (August 18, 2016), and patients who died without experiencing breast cancer recurrence were censored at the time of death. Six patients died from causes other than breast cancer.

Statistical analysis

- 198 -

Descriptive statistics were used for baseline characteristics. Univariate and multivariate Cox proportional hazard regressions were used to estimate hazard ratios (HR) and corresponding 95% confidence intervals (CI) for the effect of tobacco use, alcohol consumption or HER2 SNPs on survival (DFS). The Cox proportional hazards assumption was tested both numerically and graphically. Prognostic factors associated with DFS with a p-value <0.20 in univariate analyses were included in multivariate analyses to account for potential confounders. Adjustment variables included age at diagnosis (≤ 55 (reference), > 55 years), body mass index (< 20, ≥ 20 and < 25 (reference), ≥ 25 kg/m2), stage (I (reference), II, III), adjuvant endocrine therapy (yes (reference), no), and radiotherapy (yes (reference), no). One patient for whom adjustment variables were unknown was excluded from the multivariate analysis. Given that MAPK signaling is reportedly significantly hyperactivated in ER-negative tumors overexpressing HER2 29, we performed stratified analysis according to ER status. All statistical tests were two-sided, and p-values <0.05 were considered statistically significant. All analyses were performed using the SAS software (version 9.1.3; SAS Institute Inc., Cary, NC, USA).

Results Over a median follow-up period of 7.35 years, 66 patients out of 236 (28.0%) experienced recurrence. Baseline characteristics of the study population are presented in Table 1. Approximately 35.0% of patients were younger than 50 years. One hundred thirty patients (55.1%) had a body mass index higher than 25 kg/m2, 63.1% of patients had grade III tumors, and positive lymph node status was observed in 53.4% of patients. Tumors bigger than 5 cm were observed in 5.5% of patients, and 29.7% of patients had stage III tumors. Positive ER and PR status was observed in 66.1% and 47.0% of patients, respectively. Most of the patients had radiotherapy (85.2%) and 63.6% of patients received hormonotherapy. Before their diagnosis, 16% of patients were smokers and 58% of patients consumed alcohol.

Association between tobacco use and DFS is presented in Table 2. Compared to non- smokers, the adjusted hazard ratio for DFS was 2.63 (95% CI, 1.48 to 4.68, p=0.001) for patients who smoked before breast cancer diagnosis. Average number of cigarettes smoked per day and number of years spent smoking before breast cancer diagnosis were associated with breast cancer recurrence, but only in the extreme category (HR: 3.65, 95% CI, 1.35 to 9.89, p=0.01 for > 20 cigarettes per day, and HR: 3.19, 95% CI, 1.55 to 6.56, p=0.002 for > 20 years spent smoking, respectively). For the 131 patients for whom we had information

- 199 -

about smoking status during trastuzumab treatment, we did not observe a statistically significant association between smoking status during trastuzumab treatment and breast cancer recurrence (HR: 2.15, 95% CI, 0.92 to 5.04, p=0.08).

When we performed a stratified analysis according to the ER status, associations between tobacco use and breast cancer recurrence were only observed in the ER-negative subgroup (Table 3). Among patients with ER-negative tumors, the adjusted hazard ratio for DFS was 3.73 (95% CI, 1.56 to 8.89, p=0.003) for patients who smoked before breast cancer diagnosis and was 4.49 (95% CI, 1.26 to 16.00, p=0.02) for patients who smoked during trastuzumab treatment when compared to non-smokers. In the ER-positive subgroup, smoking status neither before breast cancer diagnosis nor during trastuzumab treatment were associated with DFS (HR: 1.91; 95% CI, 0.84 to 4.31, p = 0.12, and HR: 1.33; 95% CI, 0.27 to 6.47, p=0.72, respectively).

Association between alcohol consumption and DFS is presented in Table 4. Compared to non-drinkers, the adjusted hazard ratio for DFS was 0.56 (95% CI, 0.33 to 0.94, p=0.03) for patients who consumed alcohol before breast cancer diagnosis. Further analyses on alcohol type consumption revealed that wine consumption before breast cancer diagnosis was associated with a reduced risk of breast cancer recurrence compared to control (no wine consumption) (adjusted HR: 0.42; 95% CI, 0.18 to 0.95, p=0.04). However, beer consumption before breast cancer diagnosis was not significantly associated with a higher risk of breast cancer recurrence compared to control (no beer consumption) (adjusted HR: 1.60; 95% CI, 0.47 to 5.47, p=0.46). For the 128 patients for whom we had information about alcohol use during trastuzumab treatment, we did not observe a significant association between DFS and consumption of alcohol (adjusted HR: 0.68, 95% CI, 0.30 to 1.56, p=0.36), wine (HR: 0.55; 95% CI, 0.23 to 1.33, p=0.18) or beer (HR: 1.98; 95% CI, 0.53 to 7.33, p=0.31). We could not generate results regarding the association between spirits consumption neither before breast cancer diagnostic nor during trastuzumab treatment and risk of cancer recurrence, since only a few women consumed spirits. Moreover, there is no data on the association between alcohol consumption and DFS according ER status due to insufficient number of patients.

HER2 polymorphism substudy HER2 genotype was determined in breast cancer tissue of 73 breast cancer patients. As previously reported, Ile655Val and Ala1170Pro genotypes measured in breast cancer

- 200 -

tissues were available for 71 and 69 breast cancer patients, respectively (26). Briefly, 77.5% (50/71) of patients were homozygous for the Ile/Ile genotype, 7.0% (5/71) were heterozygous for the Ile/Val genotype, and 15.5% (11/71) were homozygous for the Val/Val genotype. Regarding the Ala1170Pro polymorphism, 50.7% (35/69) of patients were homozygous for the Ala/Ala genotype, 13.1% (9/69) were heterozygous for the Ala/Pro genotype, and 36.2% (25/69) were homozygous for the Pro/Pro genotype.

Association between Ile655Val polymorphism and DFS is presented in Table 5. Compared to patients with Ile/Ile genotype, those with the Val/Ile or the Val/Val genotypes showed worse DFS after adjustment for potential confounders (HR 4.96, 95% CI 1.45 to 17.02, p = 0.01). Regarding the Ala1170Pro polymorphism, compared to patients with the Ala/Ala genotype, those with the Ala/Pro or the Pro/Pro genotypes showed non-significant better DFS even after adjustment for confounding factors (HR 0.70, 95% CI 0.23 – 2.09, p = 0.77).

Discussion In a cohort of non-metastatic HER2-positive breast cancer patients we observed that tobacco use before breast cancer diagnosis or during trastuzumab treatment was associated with breast cancer recurrence. Moreover, our results suggest that heavy exposure to tobacco (more than two decades spent smoking or > 20 cigarettes smoked per day) before breast cancer diagnosis increase risk of recurrence. To the best of our knowledge, only one study has investigated the association between tobacco consumption and trastuzumab response (13). In a cohort of 248 metastatic trastuzumab-treated HER2- positive breast cancer patients, Santini and collaborators reported that tobacco exposure before breast cancer diagnosis was not associated with targeted treatment response (13). Furthermore, they reported that the number of cigarettes smoked before breast cancer diagnosis was not associated with patient outcome. This inconsistence with our results could be explained by the differences in study population (our cohort was composed of early HER2-positive breast cancer patients, whereas the cohort in the Santini study was composed of metastatic HER2-positive breast cancer patients) and in the analysis performed (our models were adjusted for potential confounders, whereas Santini and collaborators performed univariate models). Moreover, Santini et al. did not consider the numbers of years spent smoking before breast cancer diagnosis, nor did they perform stratified analysis according to ER status. It has been reported that MAPK signaling was significantly hyperactivated in ER-negative tumors overexpressing HER2 (29). In agreement with these preclinical observations, we observed that the association between tobacco

- 201 -

exposure during trastuzumab treatment and worse outcome was stronger in the ER- negative subgroup.

In terms of alcohol, consumption before breast cancer diagnosis was associated with a better response towards trastuzumab. However, we did not observe an association between alcohol consumption during trastuzumab treatment and breast cancer recurrence. To date, no other study evaluated the association between alcohol intake and response to trastuzumab in HER2-positive breast cancer patients. Given that the observed results were in contrast with our initial hypothesis that alcohol exposure augmented the risk of relapse in HER2-positive trastuzumab-treated breast cancer patients and considering that some preclinical data suggest that resveratrol, a polyphenol contained in the skin of grapes and in red wine (30), may be protective against some type of cancers including breast cancer (31- 33), we decided to further analyse the association by evaluating the effect of alcohol type consumption with DFS. Although our sample size was small, our results suggest that wine consumption before breast cancer diagnosis is associated reduced risk of recurrence. Although not statistically significant, it seems that beer consumption could be associated with a higher risk of breast cancer recurrence. Similar results, although not statistically significant, were observed when the type of alcohol consumed during trastuzumab treatment was considered. These results are in line with a recent preclinical study, in which it has been shown that the addition of resveratrol to HER2-positive breast cancer cell lines treated with trastuzumab increased the cytotoxicity of the anti-HER2 agent compared to HER2-positive breast cancer cell lines treated with trastuzumab alone (34).

In our subgroup study, we observed that patients carrying the Ile/Val or Val/Val genotypes showed significantly worse DFS compared to patients carrying the Ile/Ile genotype, after adjustment for potential confounders. Two previous similar studies that analyzed the association between HER2 Ile655Val polymorphism and trastuzumab response in HER2- positive breast cancer patients reported conflicting results (23, 24). Beauclair and collaborators have shown that Ile655Val polymorphism measured in blood was not associated with response to trastuzumab-based therapy in a cohort of 61 patients with advanced HER2-positive breast cancer patients in univariate analysis (23). In a cohort of 212 trastuzumab-treated HER2-positive breast cancer patients, Han and collaborators observed that patients carrying the Ile/Val or Val/Val genotypes measured in blood showed significantly better DFS compared to those carrying Ile/Ile genotype in univariate analysis (24). Several elements could explain the discrepancies between the results reported in our

- 202 -

study and those published in the two previously studies. Firstly, whereas Han and Beauclair measured Ile655Val genotype using genomic DNA extracted from buffy coat fractions (23, 24), we analysed Ile655Val genotype using DNA extracted from FFPE breast cancer tissues. We have previously reported significant differences in the genotype distribution of Ile655Val and Ala1170Pro polymorphisms between normal breast tissue and breast cancer tissues (26). Of note, we observed an excellent concordance between genotype measured from blood and normal breast tissue (26). Similar to Beauclair et al., we did not observe a significant association between Ile655Val polymorphism and response to trastuzumab, when Ile655Val polymorphism was measured in normal breast tissue (data not shown). Secondly, both previous studies exclusively analysed the association between Ile655Val polymorphism and response to trastuzumab by univariate analysis, thus not taking into account potential confounding factors (23, 24). Moreover, characteristics of the study population differed among the studies. Whereas our cohort consisted of early breast patients, the study population of Beauclair et al. consisted of metastatic breast cancer patients (23).

In a preclinical model, it has been proposed that substitution of a Val for an Ile residue at codon 655 predisposes the HER2 receptor to adopt an active conformation that results in increased activity of the tyrosine kinase domain and enhanced tumorigenic potential of breast cancer cells (21). Beauclair et al. analysed the association between Ile655Val polymorphism and response to trastuzumab using fibroblast cells that were transfected to stably express HER2 protein carrying the Val or the Ile genotype (23). The authors showed that fibroblast cells that expressed HER2 protein carrying the Val genotype exhibited a higher growth capacity and a lower apoptosis rate compared to fibroblast cells expressing HER2 protein carrying the Ile genotype. The authors also observed that inhibition of HER2 phosphorylation was stronger in cells expressing HER2 carrying the Val genotype compared to those expressing HER2 carrying the Ile genotype. In addition, they observed that only fibroblast cells that expressed HER2 protein carrying the Val genotype developed tumors in nude mice. The authors concluded that fibroblast cells that expressed HER2 protein carrying the Val genotype showed a more aggressive phenotype and were more sensitive to trastuzumab compared to those expressing HER2 protein carrying the Ile genotype (23). Previous studies that analyzed the effects of trastuzumab on HER2 phosphorylation in breast cancer cell lines reported conflicting results. Whereas one study confirmed that trastuzumab reduced HER2 phosphorylation (35), others have shown that trastuzumab

- 203 -

increased HER2 phosphorylation (36, 37). It has also been reported that trastuzumab did not block dimerization with other HER family members (38, 39). Of note, it has been shown that signaling through alternative pathways, including other HER family members, can contribute to the development of trastuzumab resistance in preclinical models (40). Since Beauclair et al. transfected fibroblast cells, a cell line known to express undetectable levels of HER family members, it can be postulated that this model does not reproduce the molecular context of the HER2-positive breast cancer cell surface. In our opinion, our results concordant with HER2 biology. It can be argued that since Val substitution leads to enhanced activity of the tyrosine kinase domain, that the effect of trastuzumab on HER2 phosphorylation have not been completely unraveled yet and that trastuzumab does not block dimerization with other HER family members, HER2-positive breast cancer cells carrying the Val allele might be more prone to develop resistance toward trastuzumab in the presence of alternative signaling pathways.

Our study adds knowledge to the few others that have investigated the association between tobacco consumption and HER2 polymorphisms and trastuzumab response in trastuzumab- treated HER2-positive breast cancer patients in addition to provide information on the association between alcohol use and trastuzumab response in this class of patient. However, our study presents some weaknesses, including the small sample size and its retrospective design. Moreover, although we performed multivariate analysis, we cannot completely exclude the possibility of residual confounding. In addition, it is well known that the measurements of tobacco and alcohol consumption through validated self-administered questionnaires are subjected to error, but if any, it would result in non-differential misclassification, and would have underestimated the true associations.

Conclusion Although the sample size of our study population was small, our results suggest that lifestyle factors such as tobacco and alcohol consumption and genetic factors like HER2 polymorphism could influence efficacy of trastuzumab treatment in HER2-positive breast cancer patients. These results need to be confirmed in a larger cohort, using detailed questionnaires about tobacco use and type of alcohol consumed at several times following a diagnosis of breast cancer. If confirmed in a larger prospective cohort study, these results might contribute to developing lifestyle habits recommendations for trastuzumab-treated HER2-positive breast cancer patients.

- 204 -

- 205 -

References 1. Yarden Y, Sliwkowski MX. Untangling the ErbB signalling network. Nature reviews Molecular cell biology. 2001;2(2):127-37.

2. Soerjomataram I, Louwman MW, Ribot JG, Roukema JA, Coebergh JW. An overview of prognostic factors for long-term survivors of breast cancer. Breast cancer research and treatment. 2008;107(3):309-30.

3. Esteva FJ, Yu D, Hung MC, Hortobagyi GN. Molecular predictors of response to trastuzumab and lapatinib in breast cancer. Nature reviews Clinical oncology. 2010;7(2):98- 107.

4. Gianni L, Dafni U, Gelber RD, Azambuja E, Muehlbauer S, Goldhirsch A, et al. Treatment with trastuzumab for 1 year after adjuvant chemotherapy in patients with HER2- positive early breast cancer: a 4-year follow-up of a randomised controlled trial. The Lancet Oncology. 2011;12(3):236-44.

5. Seidman AD, Fornier MN, Esteva FJ, Tan L, Kaptain S, Bach A, et al. Weekly trastuzumab and paclitaxel therapy for metastatic breast cancer with analysis of efficacy by HER2 immunophenotype and gene amplification. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2001;19(10):2587-95.

6. Vu T, Claret FX. Trastuzumab: updated mechanisms of action and resistance in breast cancer. Front Oncol. 2012;2:62.

7. Berube S, Lemieux J, Moore L, Maunsell E, Brisson J. Smoking at time of diagnosis and breast cancer-specific survival: new findings and systematic review with meta-analysis. Breast cancer research : BCR. 2014;16(2):R42.

8. Kwan ML, Chen WY, Flatt SW, Weltzien EK, Nechuta SJ, Poole EM, et al. Postdiagnosis alcohol consumption and breast cancer prognosis in the after breast cancer pooling project. Cancer Epidemiol Biomarkers Prev. 2013;22(1):32-41.

9. Passarelli MN, Newcomb PA, Hampton JM, Trentham-Dietz A, Titus LJ, Egan KM, et al. Cigarette Smoking Before and After Breast Cancer Diagnosis: Mortality From Breast Cancer and Smoking-Related Diseases. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2016;34(12):1315-22.

- 206 -

10. Sanner T, Grimsrud TK. Nicotine: Carcinogenicity and Effects on Response to Cancer Treatment - A Review. Front Oncol. 2015;5:196.

11. Siriwardhana N, Choudhary S, Wang HC. Precancerous model of human breast epithelial cells induced by NNK for prevention. Breast cancer research and treatment. 2008;109(3):427-41.

12. Seibold P, Vrieling A, Heinz J, Obi N, Sinn HP, Flesch-Janys D, et al. Pre-diagnostic smoking behaviour and poorer prognosis in a German breast cancer patient cohort - Differential effects by tumour subtype, NAT2 status, BMI and alcohol intake. Cancer Epidemiol. 2014;38(4):419-26.

13. Santini D, Vincenzi B, Adamo V, Addeo R, Fusco V, Russo A, et al. Cigarette smoking habit does not reduce the benefit from first line trastuzumab-based treatment in advanced breast cancer patients. Oncol Rep. 2011;25(6):1545-8.

14. Hirko KA, Chen WY, Willett WC, Rosner BA, Hankinson SE, Beck AH, et al. Alcohol consumption and risk of breast cancer by molecular subtype: Prospective analysis of the nurses' health study after 26 years of follow-up. Int J Cancer. 2016;138(5):1094-101.

15. Ma C, Lin H, Leonard SS, Shi X, Ye J, Luo J. Overexpression of ErbB2 enhances ethanol-stimulated intracellular signaling and invasion of human mammary epithelial and breast cancer cells in vitro. Oncogene. 2003;22(34):5281-90.

16. Caskey CT. Using genetic diagnosis to determine individual therapeutic utility. Annu Rev Med. 2010;61:1-15.

17. Moreno-Munoz D, de la Haba-Rodriguez JR, Conde F, Lopez-Sanchez LM, Valverde A, Hernandez V, et al. Genetic variants in the renin-angiotensin system predict response to bevacizumab in cancer patients. Eur J Clin Invest. 2015;45(12):1325-32.

18. Alaoui-Jamali MA, Morand GB, da Silva SD. ErbB polymorphisms: insights and implications for response to targeted cancer therapeutics. Front Genet. 2015;6:17.

19. Han W, Kang D, Lee JE, Park IA, Choi JY, Lee KM, et al. A haplotype analysis of HER-2 gene polymorphisms: association with breast cancer risk, HER-2 protein expression in the tumor, and disease recurrence in Korea. Clin Cancer Res. 2005;11(13):4775-8.

- 207 -

20. Papewalis J, Nikitin AY, Rajewsky MF. G to A polymorphism at amino acid codon 655 of the human erbB-2/HER2 gene. Nucleic acids research. 1991;19(19):5452.

21. Fleishman SJ, Schlessinger J, Ben-Tal N. A putative molecular-activation switch in the transmembrane domain of erbB2. Proc Natl Acad Sci U S A. 2002;99(25):15937-40.

22. Tommasi S, Fedele V, Lacalamita R, Bruno M, Schittulli F, Ginzinger D, et al. 655Val and 1170Pro ERBB2 SNPs in familial breast cancer risk and BRCA1 alterations. Cell Oncol. 2007;29(3):241-8.

23. Beauclair S, Formento P, Fischel JL, Lescaut W, Largillier R, Chamorey E, et al. Role of the HER2 [Ile655Val] genetic polymorphism in tumorogenesis and in the risk of trastuzumab-related cardiotoxicity. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2007;18(8):1335-41.

24. Han X, Diao L, Xu Y, Xue W, Ouyang T, Li J, et al. Association between the HER2 Ile655Val polymorphism and response to trastuzumab in women with operable primary breast cancer. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2014;25(6):1158-64.

25. Dawson DA. Methodological issues in measuring alcohol use. Alcohol Res Health. 2003;27(1):18-29.

26. Furrer D, Lemieux J, Cote MA, Provencher L, Laflamme C, Barabe F, et al. Evaluation of human epidermal growth factor receptor 2 (HER2) single nucleotide polymorphisms (SNPs) in normal and breast tumor tissues and their link with breast cancer prognostic factors. Breast. 2016;30:191-6.

27. Gagnon JF, Sanschagrin F, Jacob S, Tremblay AA, Provencher L, Robert J, et al. Quantitative DNA methylation analysis of laser capture microdissected formalin-fixed and paraffin-embedded tissues. Exp Mol Pathol. 2010;88(1):184-9.

28. Benusiglio PR, Lesueur F, Luccarini C, Conroy DM, Shah M, Easton DF, et al. Common ERBB2 polymorphisms and risk of breast cancer in a white British population: A case control study. Breast cancer research : BCR. 2005;7(2):204-2099.

- 208 -

29. Creighton CJ, Hilger AM, Murthy S, Rae JM, Chinnaiyan AM, El-Ashry D. Activation of mitogen-activated protein kinase in estrogen receptor alpha-positive breast cancer cells in vitro induces an in vivo molecular phenotype of estrogen receptor alpha-negative human breast tumors. Cancer research. 2006;66(7):3903-11.

30. Brakenhielm E, Cao R, Cao Y. Suppression of angiogenesis, tumor growth, and wound healing by resveratrol, a natural compound in red wine and grapes. FASEB J. 2001;15(10):1798-800.

31. Aggarwal BB, Bhardwaj A, Aggarwal RS, Seeram NP, Shishodia S, Takada Y. Role of resveratrol in prevention and therapy of cancer: preclinical and clinical studies. Anticancer Res. 2004;24(5A):2783-840.

32. Athar M, Back JH, Tang X, Kim KH, Kopelovich L, Bickers DR, et al. Resveratrol: a review of preclinical studies for human cancer prevention. Toxicol Appl Pharmacol. 2007;224(3):274-83.

33. Stewart JR, Artime MC, O'Brian CA. Resveratrol: a candidate nutritional substance for prostate cancer prevention. J Nutr. 2003;133(7 Suppl):2440S-3S.

34. Abdel-Latif GA, Al-Abd AM, Tadros MG, Al-Abbasi FA, Khalifa AE, Abdel-Naim AB. The chemomodulatory effects of resveratrol and didox on herceptin cytotoxicity in breast cancer cell lines. Sci Rep. 2015;5:12054.

35. Kumar R, Shepard HM, Mendelsohn J. Regulation of phosphorylation of the c-erbB- 2/HER2 gene product by a monoclonal antibody and serum growth factor(s) in human mammary carcinoma cells. Mol Cell Biol. 1991;11(2):979-86.

36. Gijsen M, King P, Perera T, Parker PJ, Harris AL, Larijani B, et al. HER2 phosphorylation is maintained by a PKB negative feedback loop in response to anti-HER2 herceptin in breast cancer. PLoS Biol. 2010;8(12):e1000563.

37. Scaltriti M, Verma C, Guzman M, Jimenez J, Parra JL, Pedersen K, et al. Lapatinib, a HER2 tyrosine kinase inhibitor, induces stabilization and accumulation of HER2 and potentiates trastuzumab-dependent cell cytotoxicity. Oncogene. 2009;28(6):803-14.

- 209 -

38. Agus DB, Akita RW, Fox WD, Lewis GD, Higgins B, Pisacane PI, et al. Targeting ligand-activated ErbB2 signaling inhibits breast and prostate tumor growth. Cancer Cell. 2002;2(2):127-37.

39. Ghosh R, Narasanna A, Wang SE, Liu S, Chakrabarty A, Balko JM, et al. Trastuzumab has preferential activity against breast cancers driven by HER2 homodimers. Cancer research. 2011;71(5):1871-82.

40. Pohlmann PR, Mayer IA, Mernaugh R. Resistance to Trastuzumab in Breast Cancer. Clin Cancer Res. 2009;15(24):7479-91.

- 210 -

Table 8.1. Baseline characteristics of the study population

N % Age ≤50 82 34.8 >50 154 65.2 Body mass index (kg/m2) >15-<20 18 7.6 ≥20-<25 88 37.3 ≥25 130 55.1 Grade I/II 86 34.4 III 149 63.1 Unknown 1 2.5 Lymph node status Negative 110 46.6 Positive 126 53.4 Tumor size (cm) ≤5 222 94.1 >5 13 5.5 Unknown 1 0.4 Stage I 60 25.4 II 106 44.9 III 70 29.7 Estrogen receptor status Negative 79 33.5 Positive 156 66.1 Unknown 1 0.4 Progesterone receptor status Negative 124 52.6 Positive 111 47.0 Unknown 1 0.4 Hormonotherapy No 85 36.0 Yes 150 63.6 Unknown 1 0.4 Radiotherapy No 35 14.8 Yes 201 85.2

- 211 -

Table 8.2. Unadjusted and adjusted hazard ratios for disease-free survival according to the tobacco consumption before breast cancer diagnosis and during trastuzumab treatment

Factor Without recurrence With recurrence HR (95% CI) p-value Adjusted HR Adjusted (95% CI)* p-value* Before breast cancer (n=170) (n=66) diagnosis No 150 (88.2%) 48 (72.7%) 1.00 (reference) 1.00 (reference) Yes 20 (11.8%) 18 (27.3%) 2.58 (1.46 – 4.55) 0.001 2.63 (1.48 – 4.68) 0.001 Cigarettes/day (n=154) (n=57)

0 134 (87.0%) 44 (77.2%) 1.00 (reference) 1.00 (reference) 1-20 17 (11.0%) 8 (14.0%) 1.28 (0.60 – 2.73) 0.52 1.32 (0.62 – 2.85) 0.47 > 20 3 (2.0%) 5 (8.8%) 4.27 (1.67 – 10.95) 0.003 3.65 (1.35 – 9.89) 0.01 Smoking duration (n=137) (n=52) (years) 0 117 (85.4%) 35 (67.3%) 1.00 (reference) 1.00 (reference) 1-20 9 (6.6%) 5 (9.6%) 2.04 (0.80 – 5.20) 0.14 2.15 (0.81 – 5.67) 0.13 > 20 11 (8.0%) 12 (23.1%) 2.86 (1.48 – 5.54) 0.002 3.19 (1.55 – 6.56) 0.002 During trastuzumab (n=96) (n=35) treatment No 86 (89.6%) 28 (80.0%) 1.00 (reference) 1.00 (reference) Yes 10 (10.4%) 7 (20.0%) 2.03 (0.88 – 4.68) 0.09 2.15 (0.92 – 5.04) 0.08 HR: hazard ratio; CI: confidence interval. * Multivariable HRs and 95% CIs are adjusted for age at diagnosis, body mass index, stage, hormonotherapy and radiotherapy.

- 212 -

Table 8.3. Unadjusted and adjusted hazard ratios for disease-free survival according to the tobacco consumption before breast cancer diagnosis and during trastuzumab treatment, stratified by ER status

Factor Without recurrence With recurrence HR (95% CI) p-value Adjusted HR Adjusted (95% CI)* p-value* ER-negative patients Before breast cancer (n=50) (n=29) diagnosis No 46 (92.0%) 20 (69.0%) 1.00 (reference) 1.00 (reference) Yes 4 (8.0%) 9 (31.0%) 3.73 (1.64 – 8.48) 0.002 3.73 (1.56 – 8.89) 0.003 During trastuzumab (n=33) (n=16) treatment No 30 (91.0%) 11 (68.8%) 1.00 (reference) 1.00 (reference) Yes 3 (9.0%) 5 (31.2%) 3.92 (1.29 – 11.85) 0.01 4.49 (1.26 – 16.00) 0.02

ER-positive patients Before breast cancer (n=120) (n=36) diagnosis No 104 (87.0%) 27 (75.0%) 1.00 (reference) 1.00 (reference) Yes 16 (13.0%) 9 (25.0%) 1.90 (0.89 – 4.04) 0.09 1.91 (0.84 – 4.31) 0.12 During trastuzumab (n=63) (n=19) treatment No 56 (89.0%) 17 (89.5%) 1.00 (reference) 1.00 (reference) Yes 7 (11.0%) 2 (10.5%) 0.96 (0.22 – 4.14) 0.95 1.33 (0.27 – 6.47) 0.72 HR: hazard ratio; CI: confidence interval. * Multivariable HRs and 95% CIs are adjusted for age at diagnosis, body mass index, stage, hormonotherapy and radiotherapy.

- 213 -

Table 8.4. Unadjusted and adjusted hazard ratios for disease-free survival according to the alcohol consumption before breast cancer diagnosis and during trastuzumab treatment

Factor Without With recurrence HR (95% CI) p-value Adjusted HR Adjusted recurrence (95% CI)* p-value* Before breast cancer (n=150) (n=61) diagnosis No 56 (37.3%) 33 (54.1%) 1.00 (reference) 1.00 (reference) Yes 94 (62.7%) 28 (45.9%) 0.55 (0.33 – 0.92) 0.02 0.56 (0.33 – 0.94) 0.03

During trastuzumab (n=95) (n=33) treatment 0-2 drinks per week 64 (67.4%) 23 (69.7%) 1.00 (reference) 1.00 (reference) > 2 drinks per week 31 (32.6%) 10 (30.3%) 0.91 (0.43 – 1.91) 0.79 0.68 (0.30 – 1.56) 0.36 HR: hazard ratio; CI: confidence interval. * Multivariable HRs and 95% CIs are adjusted for age at diagnosis, body mass index, stage, hormonotherapy and radiotherapy.

- 214 -

Table 8.5. Unadjusted and adjusted hazard ratios for disease-free survival according to the HER2 polymorphisms

Polymorphism Without With recurrence HR (95% CI) p-value Adjusted HR Adjusted recurrence (95% CI)* p-value* Ile655Val (n=53) (n=18) Ile/Ile 42 (79.3%) 13 (72.2%) 1.00 (reference) 1.00 (reference) Ile/Val or Val/Val 11 (20.7%) 5 (27.6%) 2.00 (0.70 – 5.81) 0.20 4.96 (1.45-17.02) 0.01

Ala1170Pro (N=51) (n=18) Ala/Ala 26 (51.0%) 10 (55.5%) 1.00 (reference) 1.00 (reference) Ala/Pro or Pro/Pro 25 (49.0%) 8 (44.5%) 0.84 (0.33-2.14) 0.71 0.70 (0.23-2.09) 0.77

HR: hazard ratio; CI: confidence interval. * Multivariable HRs and 95% CIs are adjusted for age at diagnosis, body mass index, stage, hormonotherapy and radiotherapy.

- 215 -

Chapter 9: Association between genome-wide DNA methylation pattern and response to trastuzumab in HER2-positive breast cancer patients

Daniela Furrer a,b,c, Frédéric Fournier a,b,d, Simon Jacob a,b,c,e,f, Arnaud Droit a,b,d, Caroline Diorio a,b,c,e,f

Manuscript in preparation. a Centre de Recherche sur le cancer de l'Université Laval, 1050 Avenue de la Médecine, Québec, QC, G1V 0A6, Canada; b Axe Oncologie, Centre de Recherche du CHU de Québec-Université Laval, 1050 Avenue de la Médecine, Québec, QC, G1V 0A6, Canada; c Département de médecine sociale et préventive, Faculté de Médecine, Université Laval, 1050 Avenue de la Médecine, Quebec City, QC G1V 0A6, Canada; d Département de médecine moléculaire, Faculté de Médecine, Université Laval, 1050 Avenue de la Médecine, Quebec City, QC G1V 0A6, Canada ; e Département de biologie moléculaire, de biochimie médicale et de pathologie, Faculté de Médecine, Université Laval, 1050 Avenue de la Médecine, Quebec City, QC G1V 0A6, Canada ; f Centre des maladies du sein Deschênes-Fabia, Hôpital du Saint-Sacrement, 1050, chemin Ste-Foy, Québec, QC, G1S 4L8, Canada

Keywords: breast neoplasms; epigentic; HER2 inhibitors; treatment response

- 216 -

Résumé L’administration du trastuzumab a conduit à l’amélioration de la survie des patientes atteintes d’un cancer du sein HER2-positif. Le développement de la résistance au trastuzumab, cependant, est de plus en plus reconnu comme un obstacle clinique majeur. Des données récentes indiquent que des mécanismes épigénétiques pourraient être associés à la résistance acquise aux thérapies antitumorales. Le but de cette étude était d’explorer l’association entre le patron de méthylation de l’ADN sur l’ensemble du génome dans le tissu du cancer du sein et la réponse au trastuzumab.

Dans cette étude, le patron de méthylation de l’ADN a été évalué dans les tissus de cancer du sein de patientes atteintes d’un cancer du sein HER2-positif traitées au trastuzumab qui ont acquis une résistance au traitement (groupe des cas, n=6) et comparé au patron des patientes HER2-positives traitées au trastuzumab qui n’ont pas développé une résistance au traitement (groupe témoin, n=6) en utilisant la micropuce Illumina Infinium HumanMethylation450. Les patientes ont été appariées pour plusieurs facteurs. Les analyses bio-informatiques ont été effectuées en utilisant les logiciels Illumina GenomeStudio et R (version 3.2.2) et l’ensemble d’outils Bioconductor minfi afin d’identifier les sites différentiellement méthylés (differentially methylated probes, DMPs) et les régions différentiellement méthylées (differentially methylated regions, DMRs) entre les cas et les témoins.

En utilisant le logiciel GenomeStudio, 1,662 DMPs entre les cas et les témoins ont été identifiés. L’ensemble des gènes différentiellement méthylés était enrichi en fonctions moléculaires et cellulaires associées à la différentiation et la croissance cellulaire, au développement cellulaire, ainsi qu’à la mort et la survie cellulaire. Cependant, ces résultats, par contre, n’ont pas pu être confirmés avec R.

Nous concluons que l’absence de résultats statistiquement significatifs pourrait être expliquée par la petite taille de l’échantillon. Un échantillon de plus grande taille permettra d’évaluer ultérieurement cette association.

- 217 -

Abstract Background. The administration of trastuzumab has led to significant improvement in survival of HER2-positive breast cancer patients in the adjuvant and metastatic settings. Trastuzumab resistance, however, has been increasingly recognized as a major obstacle in the clinical management of this subgroup of patients. Recent evidence suggests that epigenetic mechanisms might be associated with acquired resistance to cancer therapies. Aim of this study was to explore the association between genome-wide DNA methylation pattern in breast cancer tissue and the response to trastuzumab.

Patients and methods. DNA methylation pattern was assessed in breast cancer tissues of trastuzumab-treated HER2-positive breast cancer patients who acquired resistance to treatment (case group, n=6) and compared to that of trastuzumab-treated HER2-positive breast cancer patients who did not develop resistance (control group, n=6) using the Illumina Infinium HumanMethylation450 BeadChip. Patients were matched for several factors. Bioinformatics analyses were performed using the Illumina GenomeStudio software and the R statistical environment to identify differentially methylated probes (DMPs) and differentially methylated regions (DMRs) between case and control samples.

Results. Using the GenomeStudio software, 1,662 DMPs between cases and controls were identified. The differentially methylated set of genes was enriched in molecular and cellular functions associated cellular growth and differentiation, cellular development, as well as cellular death and survival. However, these results could not be confirmed when bioinformatics analyses were performed using the R statistical environment.

Conclusions. Lack of statistical results could be explained by the small sample size. A larger sample is needed to further evaluate this association.

- 218 -

Introduction The human epidermal growth factor receptor 2 (HER2) is a transmembrane tyrosine kinase receptor (1). HER2 gene amplification and receptor overexpression, which occur in approximatively 15-20% of breast cancer patients, are prognostic and predictive factors in invasive breast cancer patients (2). Targeted treatment with trastuzumab, a monoclonal antibody directed against the extracellular region of HER2, has significantly improved the disease-free and overall survival rates of metastatic and early stage HER2-positive breast cancer patients (3, 4). Despite these successes, resistance to trastuzumab, both primary and acquired, emerged as a major clinical problem in the treatment of HER2-positive breast cancer patients (5). Although several molecular mechanisms of trastuzumab resistance have been proposed in preclinical models, no clinically applicable strategy to overcome trastuzumab resistance has been identified yet (6). Therefore, there is an urgent need to identify reliable predictive molecular marker of trastuzumab resistance with the ultimate goal to develop targeted drugs able to overcome resistance.

Evidence suggests that epigenetic regulatory mechanisms might play a role in acquiring resistance to cancer therapies (7). Previous studies indicated that genome-wide DNA methylation analysis allowed the identification of genes that were regulated by methylation and were also associated to chemotherapy response in breast cancer patients or breast cancer cell lines (8, 9). Moreover, it has been reported that epigenetic silencing of a specific miRNA was associated with trastuzumab resistance in a HER2-positive trastuzumab- resistant breast cancer cell line (10). In addition, it has been reported that epigenetic therapy, combined with chemotherapy and trastuzumab, increased clinical response in HER2- positive breast cancer patients who progressed on trastuzumab-based therapies (11). A genome-wide study has reported that the DNA methylation pattern in tumors tissues of HER2-positive breast cancer patients was heterogeneous (12).

Given that epigenetic mechanisms might play a role in controlling the response to trastuzumab, that HER2-positive breast cancer is characterized by a heterogeneous methylation pattern and that response to trastuzumab in HER2-positive breast cancer patients is also heterogeneous, we hypothesize that DNA methylation pattern in tumor tissues of patients that respond to trastuzumab treatment is different to that of patients that develop resistance against this drug. The aim of our pilot study was to analyze the association between DNA methylation patterns in breast cancer specimens and response

- 219 -

to trastuzumab in a cohort of 12 trastuzumab-treated, non-metastatic HER2-positive breast cancer patients.

Material and Methods

Study population The study population consisted of 12 women (6 cases and 6 controls) selected among 106 trastuzumab-treated patients with non-metastatic, HER2-positive breast cancer diagnosed between July 1st, 2005 and December 31th, 2010 at the Centre des maladies du sein Deschênes- Fabia, a specialized breast center in Quebec City, Canada. Information on tumor characteristics and prognostic factors and follow-up information were collected from medical records. Clinical endpoint in this study was disease free survival (DFS). All recurrences (locoregional recurrence, recurrence in the contralateral breast, and distant breast cancer recurrence) were considered as events, whereas death (from any cause) before recurrence and loss of follow-up were considered censoring events.

Over a mean follow-up period of 6.22 years, 22 patients out of 106 experienced recurrence. Eight cases were randomly selected among all patients who developed recurrence during follow-up and six had a sufficient amount of primary breast cancer tissue available for DNA extraction (see below). Of note, baseline characteristics of the six selected cases were comparable to the total population of cases (n=22) for all characteristics, with the exception of tumor grade (proportion of grade III tumors among the selected cases 67% vs. 41% for the total population of cases). For each case, one control was selected from the 85 patients who had not developed recurrence and was alive at the date of the case’s recurrence and for which enough material was available for subsequent DNA extraction. Controls were matched to cases for the following factors: age at diagnosis (with 5-year age categories), estrogen receptor (ER) status, year of diagnosis (with 2-year categories), and menopausal status. All patients provided written informed consent. Ethical approval of the study was obtained from the Research Ethics Committee of the Centre de Recherche du CHU de Québec (# 2016-2802).

DNA extraction For each patient, formalin-fixed, paraffin-embedded (FFPE) blocks of primary breast cancer tissue were selected. In order to assure that DNA methylation was analyzed to the greatest possible extent in breast cancer tissue and reduce contamination with other cell types

- 220 -

(lymphocytes, adipocytes, fibroblasts), we constructed tissue microarray (TMA) for each patient. For each block, representative regions of the tumor were demarcated on the H&E slide for TMA construction, as previously described (13). Briefly, tissue cores of 1 mm in diameter were punched from the original block within these delineated regions and transferred to a recipient paraffin block using a manual arraying instrument (Beecher Instruments, Silver Spring, MD, USA). From each TMA block, one section was stained with H&E to verify cellular composition of the cores. Cores were removed from TMA block if they contained no tumor tissue or if tumor tissue occupied < 70% of the core area before proceeding to DNA extraction. H&E sections were prepared from different levels of the TMA blocks: at the beginning of the TMA block, at regular intervals (every ten serial sections 10µm thick), and after the last section (approximately the 34th section). In average, tumor tissue occupied 83.3% (82.5% for the case group and 84.2% for the control group) of the core in the first cut section and 72.3% (72.0% for the case group and 72.5% for the control group) of the core in the last section. Genomic DNA was then extracted from tissue cores using the GeneJET FFPE DNA Purification kit (ThermoScientific) with minor modifications to the manufacturer’s instruction in which samples were incubated with Digestion buffer during six minutes and incubated with Proteinase K solution during 180 minutes.

DNA samples were processed at McGill University and Génome Québec Innovation Center (Montreal, Canada). Bisulfite conversion was performed using the EZ DNA MethylationTM kit (ZymoResearch) following the manufacturer’s instruction. Quality control of converted DNA was assessed by real-time PCR (Illumina FFPE QC kit, Illumina, Inc., CA, USA). Converted DNA samples were then restored using an FFPE restoration kit (Infinium HD FFPE DNA restore kit, Illumina) following the manufacturer’s instruction in order to repair degraded FFPE DNA samples (14). Restored DNA was then hybridized on the Infinium HumanMethylation450 Beadchip (Illumina) according to manufacturer’s instructions. This Beadchip interrogates 482,421 CpG sites, 3091 non-CpG sites and 65 random SNPs and covers 21,231 RefSeq genes (15).

Bioinformatic analysis Analysis of methylation data was performed using two methods, namely using the Illumina GenomeStudio software and the R statistical environment.

Analysis of methylation data using the GenomeStudio software:

- 221 -

Methylation data were visualized and analysed using the GenomeStudio software version 2011.1 (Illumina, Inc.) and the Methylation Module (1.9.0). Intensities were normalized using Illumina’s internal control probes and background-subtracted.

The β-value is the ratio between methylated probe intensity and overall probe intensities (sum of methylated and unmethylated probe intensities) (16). For quality control (QC), methylation measures with a detection p-value > 0.05 and samples with CpG coverage <95% were removed. Moreover, probes located on sex chromosomes (Y and X), having polymorphism located within or near the probe or corresponding to CpG not annotated in the human genome build 37 were filtered out from analysis. In addition, we focused our analysis on those probes located in proximal promoters, i.e. the sum of all probes located within 200 bp or 1,500 bp upstream of the transcription start site, in the 5’-untranslated region and first exon (17). Furthermore, only probes that showed 2-fold changes in β-value between cases and controls were considered.

Differential methylation was computed using the Illumina custom model to identify CpG sites that were differentially methylated between controls and cases. Multiple testing correction using the False Discovery Rate (FDR) (cutoff < 5%) was performed. Differentially methylated CpGs between cases and controls were identified through the DiffScore, defined as the differential methylation score calculated from beta-values and differences in beta-values (delta beta). CpG sites associated with a significant Diffscore value ≥ |22| (correlated p-value ≤ 0.01) were considered as significant.

Pathways and functions analyses

Identification of overrepresented pathways and functions were performed using QIAGEN’s Ingenuity Pathway Analysis (IPA, QIAGEN Redwood City, www.qiagen.com/ingenuity) software with default setting. Gene lists were uploaded using gene symbols and submitted for IPA Core Analysis. IPA calculates a p-value that represents the statistical significance of association between the genes and the networks using the Fisher's exact test. P-values < 0.05 were considered significant.

- 222 -

Analysis of methylation using the R statistical environment:

Raw intensity data from GenomeStudio were imported into the R statistical programming environment (version 3.2.2) using the Bioconductor package minfi (version 1.16.0). Initial quality control was assessed using the detectionP function in minfi, whereby the quality of each sample was evaluated using the internal control probes located on the BeadChip array. Any probes with the following criteria were filtered out before analysis: detection P > 0.05, probes located on sex chromosomes (Y and X), having polymorphism located within or near the probe, and cross-reactive probes (18). As a result, DNA methylation for 411,416 probes (85%) were used for further analysis. Probes were normalized using the PreprocessFunnorm function in minfi. M values (logit transformed β-values) were used for statistical analyses, whereas β-values were used for the visualization of the methylation levels.

Differentially methylated probes (DMPs) were identified using LIMMA taking into account the matching factors between cases and controls (i.e. age at diagnosis, ER status, year of diagnosis, and menopausal status). Multiple testing correction was performed using FDR estimation (cutoff < 5%).

Differentially methylated regions (DMRs) were identified using the bumphunter function in minfi with inclusion threshold set to 0.99. Multiple testing correction was performed using FDR estimation (cutoff < 5%). Difference in global methylation levels between cases and controls was assessed using a Wilcoxon signed rank test for paired samples on median beta values of cases and controls.

Statistical analysis Baseline characteristics between cases and controls were compared using Fisher’s exact test for categorical variables and using the Mann-Whitney test for continuous variables.

Results Genome-wide DNA methylation data were obtained from six trastuzumab-treated HER2- positive breast cancer patients who experienced a recurrence during follow-up and six frequency-matched controls. Baseline characteristics of cases and controls are summarized in Table 1. Baseline characteristics for both groups were comparable in terms of clinicopathological characteristics (tumor grade, lymph node status, and tumor size) and

- 223 -

treatment received. Compared to controls, a higher proportion of cases (50%) had a body mass index >25 (although not statistically significant).

Before any statistical analyses were conducted, we evaluated the data for the presence of substantial confounding batch effects due to separate chips (19). We did not observe any batch effect in our data. All samples passed quality-control tests and were therefore retained in the analysis.

Analysis of DNA methylation data performed using the Illumina custom model allowed the identification of 1,662 probes (737 genes) that were differentially methylated between cases and controls: 1,222 probes (460 genes) were significantly hypermethylated and 440 probes (277 genes) were significantly hypomethylated in tumor tissues of cases compared to those of controls after multiple testing. In order to investigate which genes exhibiting altered DNA methylation from our dataset were present in defined canonical pathways, we used the “Core analyses” included in Ingenuity Pathway Analysis. The differentially hypermethylated set of genes was significantly enriched in molecular and cellular functions such as cellular growth and proliferation as well as cellular death and survival and in canonical pathways associated with molecular mechanisms of cancer, and embryonic stem cell pluripotency. The differentially hypomethylated set was significantly enriched in functions including cellular growth and proliferation, and cellular development and in canonical pathways implicated in cell cycle regulation and embryonic stem cell pluripotency.

Interrogation of DNA methylation data using R environment allowed the identification of 23,153 probes that were differentially methylated between cases and controls, before correction for multiple testing was applied. The most differentially methylated probes were localized on the following genes: FAM19A5, cyclin dependent kinase 18 (Cdk18), and Dock2. After correction for multiple testing, however, no differentially methylated probes were identified. The function bumphunter in minfi allowed the identification of 246 regions that were differentially methylated between cases and controls. After correction for multiple testing, however, no regions were differentially methylated between the two groups. In addition, global methylation levels between cases and controls were not statistically different: median beta value of cases was 0.461 and median beta value of controls was 0.452 (p-value: 0.844).

- 224 -

Discussion In a cohort of 12 HER2-positive breast cancer patients treated with trastuzumab, interrogation of DNA methylation using the Infinium HumanMethylation450 Beadchip and the GenomeStudio software, allowed the identification of probes located in the promoter region that were differentially methylated between cases and controls. Interestingly, the differentially methylated set of genes was significantly enriched in molecular and cellular functions associated with cancer, including cellular growth and proliferation, cell death and survival, and cellular development.

Given that the Illumina GenomeStudio software does not allow to perform a matched analysis of the DNA methylation differences between cases and controls, we also analysed the methylation data using the more flexible R statistical environment. Indeed, the LIMMA package allows the identification of DMPs taking into account of the matching factors between the individuals of the two groups. Of note, among the eight probes that were identified as the most differentially methylated between cases and controls, we observed genes that are associated with human cancer, including FAM19A5 (probe localized in first exon), cyclin-dependent kinase 18 (Cdk18), and Dock2 (both probes localized in gene body). FAM19A5 (also known as TAFA5) is a member of the TAFA family (20). Somatic mutations in this gene have been associated with pancreatic cancer (21). In addition, germline amplification of the chromosomal region harboring this gene is frequently observed in gliomas (22). It has been recently reported that Cdk18 plays an important role as regulator of genome stability, since siRNA-mediated depletion of Cdk18 leads to increased DNA damage and chromosomal instabilities (23). It has been shown that Dock2, a member of the CDM protein family that plays a central role in lymphocyte migration (24), regulates cellular proliferation in tumor cells, including B cell lymphoma and leukemia cells (25, 26). However, after correction for multiple testing, none of the differentially methylated probes retained statistical significance. Similarly, we did not identify regions that were significantly differentially methylated between cases and controls after correction for multiple testing.

This is the first study that analyzed the association between DNA methylation pattern and response to trastuzumab in HER2-positive breast cancer patients. This study was conducted in a small cohort of 12 HER2-positive breast cancer patients. It. has been reported that false- negative rate is often high when sample size is small (27). Moreover, recommendations for an epigenetic study suggest to analyse approximately 10 subjects for each dependent variable to obtain a statistical power of 80% and a FDR of 5% (28, 29). Therefore, it is

- 225 -

possible that the lack of statistical results observed in our pilot study when the R statistical environment was used could be associated to the small sample size. Therefore, it might be pertinent to repeat this study in a larger cohort of HER2-positive breast cancer patients.

Evaluation of methylation patterns in breast cancer specimens present some advantages. Since DNA methylation can be stably detected in tissues (DNA is more stable compared to mRNA), DNA methylation changes may be more advantageous than gene expression for the identification of tissue-based biomarkers (30). Analysis of DNA methylation pattern may also lead to the identification of novel biomarkers of trastuzumab resistance. This might promote the development of new targeted drugs that could be administered to trastuzumab- resistant HER2-positive breast cancer patients.

Acknowledgments DF received doctoral fellowships from the Fonds de recherché du Québec – Santé (FRQS) and the Laval University Cancer Research Center. CD is a recipient of the Canadian Breast Cancer Foundation-Canadian Cancer Society Development award (award #703003) and the FRQS Research Scholar.

This study was supported by the Fondation des Hôpitaux Enfant Jésus – St-Sacrement. Clinical specimens were provided by the Fondation du cancer du sein du Québec and the Banque de tissus et de données of the Réseau de recherche sur le cancer of the FRQS, which is affiliated with the Canadian Tumour Repository Network.

- 226 -

References 1. Yarden Y, Sliwkowski MX. Untangling the ErbB signalling network. Nature reviews Molecular cell biology. 2001;2(2):127-37.

3. Romond EH, Perez EA, Bryant J, Suman VJ, Geyer CE, Jr., Davidson NE, et al. Trastuzumab plus adjuvant chemotherapy for operable HER2-positive breast cancer. N Engl J Med. 2005;353(16):1673-84.

4. Vogel CL, Cobleigh MA, Tripathy D, Gutheil JC, Harris LN, Fehrenbacher L, et al. Efficacy and safety of trastuzumab as a single agent in first-line treatment of HER2- overexpressing metastatic breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2002;20(3):719-26.

5. Nahta R, Yu D, Hung MC, Hortobagyi GN, Esteva FJ. Mechanisms of disease: understanding resistance to HER2-targeted therapy in human breast cancer. Nat Clin Pract Oncol. 2006;3(5):269-80.

6. Zhang S, Huang WC, Li P, Guo H, Poh SB, Brady SW, et al. Combating trastuzumab resistance by targeting SRC, a common node downstream of multiple resistance pathways. Nat Med. 2011;17(4):461-9.

7. Lackner MR, Wilson TR, Settleman J. Mechanisms of acquired resistance to targeted cancer therapies. Future Oncol. 2012;8(8):999-1014.

8. He DX, Gu F, Gao F, Hao JJ, Gong D, Gu XT, et al. Genome-wide profiles of methylation, microRNAs, and gene expression in chemoresistant breast cancer. Sci Rep. 2016;6:24706.

9. Klajic J, Busato F, Edvardsen H, Touleimat N, Fleischer T, Bukholm I, et al. DNA methylation status of key cell-cycle regulators such as CDKNA2/p16 and CCNA1 correlates with treatment response to doxorubicin and 5-fluorouracil in locally advanced breast tumors. Clin Cancer Res. 2014;20(24):6357-66.

- 227 -

10. Ye XM, Zhu HY, Bai WD, Wang T, Wang L, Chen Y, et al. Epigenetic silencing of miR-375 induces trastuzumab resistance in HER2-positive breast cancer by targeting IGF1R. BMC Cancer. 2014;14:134.

11. Tu Y, Hershman DL, Bhalla K, Fiskus W, Pellegrino CM, Andreopoulou E, et al. A phase I-II study of the histone deacetylase inhibitor vorinostat plus sequential weekly paclitaxel and doxorubicin-cyclophosphamide in locally advanced breast cancer. Breast cancer research and treatment. 2014;146(1):145-52.

12. Stefansson OA, Moran S, Gomez A, Sayols S, Arribas-Jorba C, Sandoval J, et al. A DNA methylation-based definition of biologically distinct breast cancer subtypes. Mol Oncol. 2015;9(3):555-68.

13. Furrer D, Jacob S, Caron C, Sanschagrin F, Provencher L, Diorio C. Tissue Microarray Is a Reliable Tool for the Evaluation of HER2 Amplification in Breast Cancer. Anticancer Res. 2016;36(9):4661-6.

14. Moran S, Vizoso M, Martinez-Cardus A, Gomez A, Matias-Guiu X, Chiavenna SM, et al. Validation of DNA methylation profiling in formalin-fixed paraffin-embedded samples using the Infinium HumanMethylation450 Microarray. Epigenetics. 2014;9(6):829-33.

15. Bibikova M, Barnes B, Tsan C, Ho V, Klotzle B, Le JM, et al. High density DNA methylation array with single CpG site resolution. Genomics. 2011;98(4):288-95.

16. Du P, Zhang X, Huang CC, Jafari N, Kibbe WA, Hou L, et al. Comparison of Beta- value and M-value methods for quantifying methylation levels by microarray analysis. BMC Bioinformatics. 2010;11:587.

17. Sandoval J, Heyn H, Moran S, Serra-Musach J, Pujana MA, Bibikova M, et al. Validation of a DNA methylation microarray for 450,000 CpG sites in the human genome. Epigenetics. 2011;6(6):692-702.

18. Chen YA, Lemire M, Choufani S, Butcher DT, Grafodatskaya D, Zanke BW, et al. Discovery of cross-reactive probes and polymorphic CpGs in the Illumina Infinium HumanMethylation450 microarray. Epigenetics. 2013;8(2):203-9.

- 228 -

19. Leek JT, Scharpf RB, Bravo HC, Simcha D, Langmead B, Johnson WE, et al. Tackling the widespread and critical impact of batch effects in high-throughput data. Nat Rev Genet. 2010;11(10):733-9.

20. Tom Tang Y, Emtage P, Funk WD, Hu T, Arterburn M, Park EE, et al. TAFA: a novel secreted family with conserved cysteine residues and restricted expression in the brain. Genomics. 2004;83(4):727-34.

21. Narayanan R. Phenome-genome association studies of pancreatic cancer: new targets for therapy and diagnosis. Cancer Genomics Proteomics. 2015;12(1):9-19.

22. Diaz de Stahl T, Hartmann C, de Bustos C, Piotrowski A, Benetkiewicz M, Mantripragada KK, et al. Chromosome 22 tiling-path array-CGH analysis identifies germline- and tumor-specific aberrations in patients with glioblastoma multiforme. Genes Chromosomes Cancer. 2005;44(2):161-9.

23. Barone G, Staples CJ, Ganesh A, Patterson KW, Bryne DP, Myers KN, et al. Human CDK18 promotes replication stress signaling and genome stability. Nucleic Acids Res. 2016;44(18):8772-85.

24. Reif K, Cyster J. The CDM protein DOCK2 in lymphocyte migration. Trends Cell Biol. 2002;12(8):368-73.

25. Wang L, Nishihara H, Kimura T, Kato Y, Tanino M, Nishio M, et al. DOCK2 regulates cell proliferation through Rac and ERK activation in B cell lymphoma. Biochem Biophys Res Commun. 2010;395(1):111-5.

26. Wu M, Hamaker M, Li L, Small D, Duffield AS. DOCK2 interacts with FLT3 and modulates the survival of FLT3-expressing leukemia cells. Leukemia. 2016.

27. Bock C. Analysing and interpreting DNA methylation data. Nat Rev Genet. 2012;13(10):705-19.

28. Baumgartel K, Zelazny J, Timcheck T, Snyder C, Bell M, Conley YP. Molecular genomic research designs. Annual review of nursing research. 2011;29:1-26.

29. Liu P, Hwang JT. Quick calculation for sample size while controlling false discovery rate with application to microarray analysis. Bioinformatics. 2007;23(6):739-46.

- 229 -

30. Duffy MJ, Napieralski R, Martens JW, Span PN, Spyratos F, Sweep FC, et al. Methylated genes as new cancer biomarkers. Eur J Cancer. 2009;45(3):335-46.

- 230 -

Table 9.1. Baseline characteristics of cases and controls

Cases Controls p-value n (%) n (%) Factor Age (years) 51.3 ± 6.1 51.2 ± 5.6 0.87 Follow-up time (years) 3.8 ± 2.8 7.2 ± 2.1 0.04 Grade I/II 2 (33%) 1 (17%) 1.00 III 4 (67%) 5 (83%) Lymph node status Negative 0 (0%) 1 (17%) 1.00 Positive 6 (100%) 5 (83%) Tumor size ≤5 5 (83%) 6 (100%) 1.00 >5 1 (17%) 0 (0%) Estrogen receptor status Negative 2 (33%) 2 (33%) 1.00 Positive 4 (67%) 4 (67%) Progesterone receptor status Negative 3 (50%) 3 (50%) 1.00 Positive 3 (50%) 3 (50%) Menopausal status Pre 2 (33%) 2 (33%) 1.00 Post 4 (67%) 4 (67%) Body mass index (kg/m2) 23.8 ± 2.6 23.9 ± 1.6 0.94 Radiotherapy No 1 (17%) 0 (0%) 1.00 Yes 5 (83%) 6 (100%) Endocrine Therapy No 2 (33%) 2 (33%) 1.00 Yes 4 (67%) 4 (67%) Chemotherapy No 0 (0%) 0 (0%) 1.00 Yes 6 (100%) 6 (100%) Trastuzumab treatment completed No 0 (0%) 0 (0%) 1.00 Yes 6 (100%) 6 (100%)

- 231 -

Chapter 10 : Discussion and conclusion

10.1. General discussion HER2 gene amplification and receptor overexpression are considered prognostic and predictive markers in breast cancer. Considering its prognostic, predictive and therefore therapeutic implications, reliable assessment of HER2 is essential for the identification of patients who will benefit from targeted anti-HER2 therapies. IHC and FISH are the most commonly used techniques for the determination of HER2 status in breast cancer specimens. Since both methods present advantages and disadvantages, there is still no consensus on what is the best method for the determination of HER2 status in breast cancer specimens. One goal of the present Ph.D. thesis was to determine the most reliable and economic method to evaluate HER2 status in breast cancer specimens.

In a cohort of 521 consecutive breast cancer patients, we compared HER2 gene amplification status determined by FISH on whole tissue section and on TMA sections (Article 2, Chapter 4). Independently from the ASCO/CAP scoring criteria used for the classification of cases, we observed an excellent concordance rate between HER2 status obtained on whole tissue section and on TMA sections: according to the newest ASCO/CAP scoring criteria, we observed 98.2% concordance and when the 2007 ASCO/CAP scoring criteria were used, the concordance was 99.8%.

Previous studies reported concordance rates ranging from 91% to 97% between HER2 FISH results obtained on whole tissue sections and TMA sections using the 2007 ASCO/CAP scoring criteria (83-86). Similar to others, we observed that even one to two 0.6 mm cores per case can reliably reproduce HER2 status obtained on whole tissue section (85, 454, 455).

Our results have important clinical implications since they suggest that TMA represents a reliable technique for the determination of HER2 gene amplification in breast cancer specimens. Given that FISH is a very expensive technique and that TMA technology considerably reduces the amount of reagents used, this method presents economic advantages. This might be particular pertinent for the determination of HER2 amplification on surgical specimens in the framework of quality control of HER2 testing.

- 232 -

Our study is the first that analyzed this concordance using the 2013 ASCO/CAP scoring criteria on TMA sections.

We also evaluated the concordance between HER2 status determined by IHC and FISH in a cohort of 498 breast cancer patients using TMA section (Article 3, Chapter 5). When the 2013 ASCO/CAP scoring criteria were applied, the overall concordance between the two methods was 93.8%. Using the 2007 ASCO/CAP scoring criteria, we observed 91.5% concordance rate, which was within the concordance range reported in previous similar studies performed on breast cancer specimens evaluated on TMA section using the 2007 ASCO/CAP scoring criteria (78.1% to 98.0%) (85, 456-461).

Several explanations have been proposed in the literature to explain the discordance between HER2 IHC and FISH results in breast cancer specimens. It has been reported that pre-analytical factors, including warm/cold ischemic time, delay and duration of fixation, and fixative used, can have an impact on the detection of protein expression and can therefore significantly and adversely affect the accurate evaluation of HER2 overexpression (462, 463). Other factors including the antibody used (464) or the interobserver variability in the interpretation of staining (465), can also influence the immunohistochemical test result. Given that ASCO/CAP recommendations regarding formalin fixation time and time to fixative (7) have been strictly followed, that batch variability is reduced when staining is performed on TMA slides and that immunohistochemical staining have been evaluated by the same persons, we are confident that these factors play only a marginal role in our study.

The presence of chromosome 17 polysomy, defined as the presence of at least three chromosome 17 centromere signals per nuclei (466), is regarded as a further explanation for discordant test results. Although gene amplification is regarded as the primary genetic mechanism underlying HER2 overexpression, it has been reported that increased HER2 gene copies secondary to chromosome 17 polysomy also contributes to increased HER2 protein levels (467). Several studies correlated chromosome 17 polysomy with increased IHC score in specimens lacking HER2 gene amplification (72, 467-469). In line with these observations, the non-amplified case that overexpressed the receptor observed in our cohort showed chromosome 17 polysomy.

This is the first study that evaluated concordance between HER2 IHC and FISH results using the 2013 ASCO/CAP scoring criteria on TMA section. Contrary to the majority of previous similar studies, we evaluated concordance between the two methods in a cohort of

- 233 -

consecutive breast cancer specimens. Since our cohort is composed of unselected breast cancer cases, our cohort might better reflect the full diagnostic spectrum observed in routine pathology practice.

In 2013, the ASCO/CAP updated the guidelines to clarify the recommendations for HER2 testing in breast cancer specimens published in 2007. There have been major changes regarding the specimen type to use for the HER2 status evaluation in breast cancer specimens: whereas the 2007 ASCO/CAP guidelines recommended to perform HER2 testing on resection specimens, the newest version of the guidelines recommend to perform an initial test in core biopsy. In order to test the impact of these changes, we evaluated the concordance between IHC and FISH results obtained on random TMA section (which represent results that would be obtained if IHC and FISH would be performed in core biopsy) to those obtained on diagnostic TMA section (which represent results obtained on excisional breast cancer specimens) in a subset of 116 breast cancer specimens from the cohort of 498 breast cancer consecutive cases (Article 3, Chapter 5). In our hands, HER2 concordance rate between the diagnostic TMA and random TMA was higher when HER2 status was determined by FISH than by IHC (98.0% concordance using FISH and 93.6% using IHC).

Compared to our results, previous studies that analyzed concordance of HER2 gene amplification status between different blocks or between needle core biopsy and subsequent excisional biopsy reported lower concordance rates ranging from 86% to 94% (470-473). These slightly different results can be partly explained by the fact that the authors analyzed their results using the 2007 ASCO/CAP scoring criteria (472) or other criteria for HER2 gene amplification (471). A similar study that analyzed concordance of HER2 status defined according to the combined IHC and FISH results between different blocks reported 96.4% concordance rate (474). Compared to Bethune et al., we observed lower concordance rate (93.7%) between HER2 status obtained using diagnostic block and randomly selected block when combined IHC and FISH results were considered (Chapter 5). In contrast to our observations, an analogous study that compared HER2 status determined on needle core biopsy and subsequent excisional biopsy of the same tumor by IHC and FISH observed a higher concordance rate when HER2 status was evaluated by IHC compared to FISH (98% vs 92%) (470). The observed differences between our results and the previously reported results might be explained by the fact that the authors used the former ASCO/CAP scoring criteria.

- 234 -

Our study is the first that evaluated concordance between HER2 status obtained from different blocks (diagnostic vs. random block) from the same tumor by FISH or FISH and IHC combined using the 2013 ASCO/CAP scoring criteria.

Although the FISH assay is considered a more objective and quantitative method compared to IHC, manual counting of fluorescent signals is time consuming (6). Several automated image analysis software have been developed for the enumeration of fluorescent signals (75, 76). While these software considerably reduce the time needed to evaluate HER2 gene amplification in breast cancer specimens, the major drawback of this analysis method is that fluorescent signals in images are quantified on the basis of square tiles of fixed dimensions (77), whose size does not always correspond to that of a single tumor cell nucleus.

In a cohort of 64 breast cancer specimens, we therefore validated the clinical performance of a software programming algorithm that analyses fluorescent signals in single tumor cell nuclei within breast cancer tissue section (Article 4, Chapter 6). We observed an excellent overall concordance between results obtained at manual scoring and at nuclei-sampling analysis (98.4% and 100% concordance at automated analysis and after human correction, respectively).

One previous study has analyzed the performance of another image analysis software for the evaluation of HER2 amplification in single tumor cell nuclei in a cohort of 100 breast cancer cases (79). Whereas the authors found a very good concordance (100.0%) between the results obtained by manual scoring and those obtained with the image analysis software for non-amplified cases, the concordance between the two methods for the amplified cases was moderate (74.1%).

Our results suggest that this new software programming algorithm reliably analyses fluorescent signals in single tumor cell nuclei within breast cancer tissue section.

This validation study has been conducted before the newest ASCO/CAP guidelines were published in 2013. Given that the 2013 ASCO/CAP guidelines also reported new scoring criteria for the evaluation of HER2 status in breast cancer specimens (Article 1, Chapter 2), we evaluated the impact of these new scoring criteria on our validation results. The impact of the newest scoring criteria on the HER2 status determined by manual counting of the cases used for validation are presented in Table 10.1. Of the 32 cases that were considered non-amplified at the manual scoring using the former ASCO/CAP scoring criteria, 29

- 235 -

remained non-amplified using the 2013 ASCO/CAP scoring criteria, whereas three cases were considered equivocal. Of the 32 cases that were considered amplified according to the 2007 ASCO/CAP scoring criteria, all cases remained amplified using the newest scoring criteria.

Table 10.1. Impact of the 2013 ASCO/CAP scoring criteria on the classification of cases used for the validation of the software programming algorithm (results obtained at manual counting)

(n=64) 2007 ASCO/CAP scoring criteria 2013 ASCO/CAP Non-amplified Amplified Total scoring criteria Non-amplified 29 0 29 Equivocal 3 0 3 Amplified 0 32 32 Total 32 32 64 Abbreviations: ASCO/CAP: American Society of Clinical Oncology/College of American Pathologists.

Given that the ASCO/CAP recommend that a new test has to be compared to a reference test using clearly negative and clearly positive samples, the three cases initially evaluated as non-amplified at manual scoring according to the 2007 ASCO/CAP scoring criteria that became equivocal using the newest scoring criteria were excluded from the concordance analysis. Table 10.2. shows the comparison between results obtained by manual scoring, the reference method, and the tile-sampling and nuclei-sampling analysis according to the 2013 ASCO/CAP scoring criteria in the 61 breast cancer specimens considered for the analysis. Concordance between manual scoring and nuclei-sampling analysis was 100.0% with the automated analysis as well as after human correction, and this for both non- amplified and amplified cases. Concordance between manual scoring and tile-sampling analysis was also 100.0% for both non-amplified and amplified cases.

- 236 -

Table 10.2. Comparison of results obtained by different methods for non- amplified and amplified cases (HER2 gene amplification status according to the 2013 ASCO/CAP scoring criteria) (n=61) Tile-sampling analysis Nuclei-sampling analysis

Manual Automated analysis After human scoring correction NAm Eq Am NAm Eq Am NAm Eq Am NAm 29 0 0 29 0 0 29 0 0 Am 0 0 32 0 0 32 0 0 32 Abbreviations: HER2: human epidermal growth factor receptor 2; ASCO/CAP: American Society of Clinical Oncology/College of American Pathologists; Nam: non-amplified; Eq: equivocal; Am: amplified.

We can therefore conclude that this software programming algorithm represents a reliable instrument for the automated quantitative HER2 FISH signals analysis in nuclei in breast cancer specimens also when HER2 gene amplification status is evaluated using the newest ASCO/CAP scoring criteria.

In order to further analyze the function of HER2 protein and its prognostic role in breast cancer patients, we evaluated the distribution of two HER2 gene polymorphisms, Ile655Val and Ala1170Pro, in normal breast and breast tumor tissues in a cohort of 73 non-metastatic HER2-positive breast cancer patients (Article 5, Chapter 7). We observed significant changes in the distribution of Ile655Val and Ala1170Pro genotypes between normal breast and breast tumor tissues. Notably, we found an allelic imbalance at codons 655 and 1170 with an overrepresentation of the Val/Val and Pro/Pro genotypes in tumor tissues. Previous similar studies that examined genotype changes in Ile655Val and Ala1170Pro SNPs between normal breast and breast tumor tissues in both HER2-negative and HER2-positive breast cancer patients reported analogous results (475-477). They observed that loss of heterozygosity (LOH) was more frequent in HER2-positive breast cancer patients.

Within the same cohort, we also evaluated the association between two and known breast cancer prognostic factors (Article 5, Chapter 7). Whereas Ile665Val polymorphism was not associated with any breast cancer prognostic factors, Ala1170Pro polymorphism was associated with age at diagnosis, tumor size and lymphovascular invasion.

- 237 -

Previous studies that analyzed these associations in breast cancer patients reported conflicting results. Whereas some authors observed a significant association between Ile655Val polymorphism and several breast cancer prognostic factors, including age at diagnosis (277), tumor size (283), and lymph node status (282, 284), others did not (285, 286), which is in line with our observations.

Some authors reported a significant association between Ala1170Pro polymorphisms and HER2 overexpression (476, 478). While one study observed a significant association between Ala1170Pro polymorphism and ER-positive status (277), another study did not (478).

These previously published studies have analyzed the association between Ile655Val and Ala1170Pro polymorphisms and breast cancer prognostic factors and/or genotype changes between normal breast and breast cancer tissues in cohorts comprised of HER2-positive and HER2-negative breast cancer patients. Our study is the first that evaluated these associations exclusively in a cohort of HER2-positive breast cancer patient. This study, therefore, broadens knowledge on HER2 function for this subgroup of patients that is eligible to receive anti-HER2 agents (479). However, we cannot completely exclude that our results are due to chance since sample size of our study population was small and residual confounding cannot be excluded.

Untreated HER2-positive breast cancer is one of the most aggressive breast cancer phenotypes (5). The prognosis of HER2-positive breast cancer patients, however, has been transformed with the introduction of HER2-targeting drugs, including trastuzumab. Although trastuzumab has led to significant improvement in disease-free and overall survival in HER2- positive breast cancer patients, resistance to trastuzumab has been increasingly recognized as a major obstacle in the clinical management of HER2-positive breast cancer patients. The second goal of my project was therefore to identify factors that could have an impact on trastuzumab response.

Since it has been reported that HER2 SNPs, especially Ile655Val SNP, have an impact on HER2 function (the Ile to Val substitution at codon 655 promotes the activity of the tyrosine kinase domain) (17), we analyzed the association between Ile655Val and Ala1170Pro SNPs and response to trastuzumab (Article 6, Chapter 8).

- 238 -

In a subgroup of 73 trastuzumab-treated HER2-positive breast cancer patients, we observed that patients carrying the Ile/Val or Val/Val genotypes showed significantly worse DFS compared to those carrying the Ile/Ile genotype, after adjustment for potential confounders.

Two previous similar studies that analyzed the association between HER2 Ile655Val polymorphism and trastuzumab response in HER2-positive breast cancer patients reported conflicting results (15, 16). Although Beauclair and collaborators have shown that Ile655Val polymorphism was not associated with response to trastuzumab-based therapy among advanced cases (15), Han et al. observed that patients carrying the Ile/Val or Val/Val genotypes showed significantly better DFS compared to those carrying the Ile/Ile genotype (16).

Several elements could explain the discrepancies between the results reported in our study and those published in the two previously studies. On the one hand, both previous studies measured Ile65Val genotype using DNA extracted from the buffy coat fraction. In our study, we evaluated the association between Ile655Val polymorphism measured in FFPE breast cancer tissues and response to trastuzumab. This difference in study design is not negligible, since we have previously demonstrated a difference in the genotype distribution of HER2 SNPs between normal breast and breast cancer tissues (480). Indeed, similar to Beauclair et al., we did not observe a significant association between Ile655Val polymorphism and response to trastuzumab when the polymorphism was measured in normal breast tissue. Also, we performed multivariate analysis, while both previous studies only conducted univariate analysis. However, we cannot completely exclude that the observed results are due to chance, since sample size of our study population was small and residual confounding cannot be excluded.

We did not observe a significant association between Ala1170Pro polymorphism and trastuzumab response. To date, no study has evaluated the association between HER2 Ala1170Pro polymorphism and trastuzumab response in HER2-positive breast cancer patients.

Previous studies have reported that lifestyle factors such as tobacco and alcohol consumption can have an impact on survival of breast cancer patients (19-21). Given that molecular and epidemiological links between tobacco and alcohol exposure and HER2 have been reported in the literature (22, 23, 453), we analyzed the association between smoking habits and alcohol consumption before breast cancer diagnosis and during trastuzumab

- 239 -

treatment and risk of recurrence in a cohort of 236 trastuzumab-treated HER2-positive breast cancer patients (Article 6, Chapter 8).

We observed that tobacco consumption before breast cancer diagnosis was significantly associated with worse outcome. In addition, we observed that heavy exposure to tobacco (more than two decades spent smoking and > 20 cigarettes smoked per day) before breast cancer diagnosis was also associated with worse outcome. For the subgroup of patients (n=131) for whom we had information about smoking habits during trastuzumab treatment, we noticed that tobacco exposure during trastuzumab treatment was associated with a worse response to trastuzumab, but only in the ER-negative subgroup. Yet, only one study has investigated the association between tobacco consumption before breast cancer diagnosis and trastuzumab response (453). In their study performed in a cohort of 248 metastatic trastuzumab-treated HER2-positive breast cancer patients, Santini and collaborators reported that tobacco exposure before breast cancer diagnosis was not associated with response rate in univariate analysis (453). This inconsistence with our results could be explained by several factors, including the differences in cohort characteristics and in the analyses performed.

We observed that alcohol consumption before breast cancer diagnosis was associated with a better response to trastuzumab. However, we did not observe a statistically significant association between alcohol consumption during trastuzumab treatment and response to the anti-HER2 drug in the subgroup of 128 patients for whom we had this information.

We also analyzed the association between alcohol type consumption and response to trastuzumab. Our results suggest that wine consumption before breast cancer diagnosis might be associated with a reduced risk of recurrence. However, beer consumption was not significantly associated with a higher risk of breast cancer recurrence. Although not statistically significant, we observed similar trends when wine and beer consumption during trastuzumab treatment was considered.

To date, no previous studies have analysed the association between alcohol consumption before breast cancer diagnosis or during trastuzumab treatment and the risk of recurrence among HER2-positive breast cancer patients that receive adjuvant trastuzumab treatment.

These results have important potential clinical implications. If confirmed in a larger cohort study, these findings could be used to raise awareness among HER2-positive breast cancer

- 240 -

patients that lifestyle habits, including tobacco and alcohol consumption, might influence anti-HER2 treatment efficacy.

In clinical practice, biomarkers that reliably predict the development of trastuzumab resistance in HER2-positive breast cancer patients have not yet been identified. In an attempt to identify molecular markers that allow identification of patients that develop resistance to trastuzumab, we explored the association between DNA methylation pattern in breast cancer specimens and risk of recurrence in a cohort of 12 trastuzumab-treated HER2-positive breast cancer patients (six patients who acquired resistance to treatment (the case group) and six paired who did not develop resistance (the control group) (Article 7, Chapter 9).

In our hands, analysis of methylation data using the GenomeStudio software allowed the identification of probes that were differentially methylated between cases and controls. However, when the R statistical programming environment was used, we did not observe differentially methylated probes or regions between cases and controls after correction for multiple testing.

Our study is the first that analyzed the association between DNA methylation pattern and response to trastuzumab in HER2-positive breast cancer patients.

This study was conducted in a small cohort of 12 HER2-positive breast cancer patients. Bock et al. has reported that the false-negative rate is often high when sample size is small (481). In addition, subject size recommendations for an epigenetic study follow similar guidelines to a global gene expression analysis of approximately 10 subjects for each dependent variable to obtain a statistical power of 80% and a FDR of 5% (482, 483). Therefore, it is possible that the lack of statistical results observed in our pilot study could be attributable to a too small sample size. Therefore, it might be pertinent to repeat this study in a larger cohort of HER2-positive breast cancer patients.

Another limitation of this study is the single time point measurement. Indeed, DNA methylation patterns were evaluated in mastectomy specimens. To detect methylation changes over time, an alternative would be to investigate methylation changes in circulating DNA. Indeed, it has been demonstrated that methylation changes can be detected in circulating DNA (484). This approach would enable real-time monitoring of changes in DNA

- 241 -

methylation patterns during targeted treatment, in order to analyse the association between changes in methylation patterns and response to trastuzumab.

In addition, given that DNA methylation patterns have been measured on FFPE tissues, misclassification of DNA patterns cannot be excluded. It has indeed been reported that age of the tumor block might negatively affect the reproducibility of differentially methylated loci detected in DNA from matched FFPE and fresh frozen samples (485).

Evaluation of methylation patterns in breast cancer specimens present some advantages. Given that DNA methylation can be stably detected in tissues (DNA is more stable compared to mRNA), DNA methylation changes may be more advantageous than gene expression for the identification of tissue-based biomarkers (486). Analysis of methylation pattern may also lead to the identification of novel biomarkers of trastuzumab resistance. This might promote the development of new targeted drugs that could be administered to trastuzumab-resistant HER2-positive breast cancer patients. Indeed, it has been shown that the combination of other targeted drugs with trastuzumab in trastuzumab-resistant HER2-positive breast cancer patients, including other HER2 inhibitors (487) or inhibitors of HER2 downstream signaling pathways like PI3K/AKT/mTOR (488), could improve their survival.

Considering that epigenetic aberrations are potentially reversible, it has been proposed that epigenetic therapy might be therapeutically useful since it might allow the re-expression of genes that play a role in the response to targeted therapies (438). A preliminary clinical study that analyzed the efficacy of a combination of trastuzumab therapy with epigenetic drugs in breast cancer patients who progressed on trastuzumab-based regimen reported promising results (449, 450). A major drawback of epigenetic therapy is that these drugs are non- specific with respect to their target genes and cells (489). As such, they can modify the DNA methylation patterns with unintended consequences (490). For this reason, these drugs cannot be administered for prolonged periods of time. However, growing understanding of the histone code and the protein complexes involved in epigenetic regulation as well as the identification of unique domains and residues essential to the catalytic mechanisms used by these enzymes will help the design of epigenetic therapies that will target only specific epigenetic modifications (490).

Because HDACs and HATs are part of macromolecular protein complexes, targeting them can also lead to unintended consequences. The lack of specificity is due to the fact that epigenetic modifications are not stand alone processes, with synergistic interactions

- 242 -

between and within marks increasing the complexity of regulatory control. The inhibitors designed to target HMTs and HDMs mainly target cofactors and/or cofactor binding sites, leading to considerable non-specificity given the vast array of enzymes using the same cofactors to catalyze multiple processes. Thus, a more comprehensive structural analysis is needed to identify unique domains and residues critical to the catalytic mechanisms used by these enzymes. Increased understanding of the histone code and the macromolecular protein complexes involved in epigenetic regulation will help refine the targeting of epigenetic therapies. This will involve the design of so-called smart drugs that will target only specific epigenetic modifications, either alone or in combination with other marks.

10.2. Perspectives In our study conducted in a cohort of 498 consecutive breast cancer patients, we observed a good concordance rate (93.8%) between HER2 status determined by IHC and FISH. Currently, there is still no consensus on which technique is the best for HER2 status assessment in breast cancer specimens. As a general observation, it is pertinent to note that the superiority of one assay over another in the determination of HER2 status in breast cancer tissues can be reliably determined only by evaluating the clinical response to anti- HER2 agents as reference. Therefore, it would be interesting to analyze the predictive value of both IHC and FISH to trastuzumab treatment in a larger cohort of HER2-positive trastuzumab-treated breast cancer patients for whom both IHC and FISH test results are available.

In a cohort of 73 non-metastatic HER2-positive breast cancer patients treated with trastuzumab, we found that Ala1170Pro SNP was associated with some breast cancer prognostic factors. We also observed significant differences in the distribution of Ile655Val and Ala1170Pro genotypes between normal breast and breast cancer tissues. In addition, we observed a significant association between Ile655Val SNP and an impaired response to trastuzumab. Although exciting, these results require confirmation in a larger cohort of trastuzumab-treated, HER2-positive breast cancer patients.

In a cohort of 12 HER2-positive breast cancer patients who received adjuvant trastuzumab treatment, we observed some CpG sites differentially methylated between patients who acquired resistance to trastuzumab treatment and patients who responded to trastuzumab treatment. These results need to be confirmed in a larger cohort study.

- 243 -

The development of deep sequencing technologies allowed remarkable advances in the identification of single polymorphic variants in HER2 gene (273). The impact of those SNPs leading to amino acid substitutions within the HER2 receptor on the HER2 receptor signaling and breast cancer cell susceptibility to HER2 inhibitors, however, still needs to be addressed. Recent techniques that allow the manipulation of gene expression in intact cells, such as the clustered regularly interspaced short palindromic repeat/CRISPR-associated protein 9 (CRISPR/Cas9) system (491), can be used to develop cell models expressing wild- type versus polymorphic variants for functional studies. The CRISPR/Cas9 system could therefore be adopted to evaluate the impact of HER2 SNPs on receptor activity and response to anti-HER2 agents.

In our hands, in a cohort of 236 trastuzumab-treated HER2-positive breast cancer patients, smoking at time of diagnosis and, for a subset of patients, smoking during trastuzumab treatment were associated with a higher risk of breast cancer recurrence. In addition, in the same cohort, we observed that alcohol consumption before breast cancer diagnosis was associated with a lower risk of breast cancer recurrence. These observations need to be confirmed in a larger cohort of HER2-positive breast cancer patients who received adjuvant trastuzumab treatment. In addition, questionnaires that more precisely investigate smoking habits, including average number of cigarettes smoked per day as well as frequency of smoking, might be useful to further analyse the association between smoking exposure (before breast cancer diagnosis as well as during trastuzumab treatment) and response to anti-HER2 therapy. Furthermore, since the harmful tobacco-specific nitrosamine NNK has been detected in the aerosol of electronic cigarettes and given that electronic cigarettes are more and more popular (492), it would be interesting to ask specific questions about electronic cigarette use in the questionnaire. The same questionnaire could be proposed to patients at different time points, i.e., before, during and after trastuzumab treatment. Moreover, since our results suggest that the effects of alcohol intake on the response to trastuzumab depend on the alcohol type consumed, more precise questionnaires that investigate alcohol consumption are necessary. These questionnaires should contain detailed questions about the type of alcohol consumed (i.e., red wine, white wine, beer, liquor) as well as questions about alcohol consumption frequency per week (number and size of drinks). Similar as for the tobacco study, the same questionnaire that investigates alcohol consumption could be proposed to the patients at different time points, i.e., before, during and after trastuzumab treatment. Furthermore, it would be interesting to measure

- 244 -

biological markers of tobacco and alcohol consumption, including nicotine and alcohol levels in saliva or blood (493, 494), before and during trastuzumab treatment. Finally, this study that investigates the association between tobacco and alcohol consumption and response to trastuzumab should be conducted ideally in a large prospective cohort study, in order to take into account all potential confounding factors.

10.3. Conclusion This present Ph.D. research project has been able to provide novel insights into the optimal strategy for the evaluation of HER2 status in breast cancer specimens as well as factors that might influence patient response to trastuzumab.

This could be summarized as follows:

1) TMA represents a reliable and economical tool for the evaluation of HER2 gene amplification in breast cancer specimens.

2) Since HER2 status determined by IHC and FISH was discordant in some cases, we suggest that both techniques should be used to assess HER2 in breast cancer specimens.

3) FISH can be reliably used to evaluate HER2 status in core biopsies. Our results suggest that the solely IHC utilisation may not reliably determine HER2 in core biopsies.

4) The nuclei-based classifier developed by MetaSystems is a reliable tool for the automated quantitative analysis of HER2 FISH fluorescent signals in nuclei in breast cancer specimens.

5) Ala1170Pro polymorphism, but not Ile655Val polymorphism, was associated with some breast cancer prognostic factors in non-metastatic HER2-positive breast cancer patients. Furthermore, since the distribution of Ile655Val and Ala1170Pro genotypes between normal breast and breast cancer tissues differed, both HER2 polymorphisms may play a role in carcinogenesis in this subgroup of breast cancer patients.

6) Ile655Val polymorphism, but not Ala1170Pro polymorphism, in tumor tissues of HER2- positive breast cancer patients was associated with a higher risk of recurrence in trastuzumab-treated, non-metastatic HER2-positive breast cancer patients.

7) Methylation pattern in tumor tissues of HER2-positive breast cancer patients who acquired resistance to trastuzumab treatment differed from that of HER2-positive breast

- 245 -

cancer patients who responded to trastuzumab treatment. However, this observation seemed to depend upon the method of bioinformatics analysis used.

8) Smoking before breast cancer diagnosis and during trastuzumab treatment was associated with a higher risk of breast cancer recurrence in HER2-positive breast cancer patients treated with trastuzumab. We also observed that alcohol consumption, in particular wine, before breast cancer diagnosis was associated with a reduced risk of breast cancer recurrence.

As a concluding remark, it should be noted that although in clinical practice HER2-positive breast cancer is often considered a single disease entity, it is now increasingly evident from previously published studies and our observations that this subgroup of breast cancer patients is clinically and biologically heterogeneous (34, 495, 496). Further clinical research that reliably identifies biomarkers of trastuzumab resistance and factors that might influence response is urgently needed. This new findings will lead to the implementation of a highly personalized analysis of breast cancer specimens as well as recommendations regarding life habits that might hopefully improve treatment response in HER2-positive breast cancer patients.

- 246 -

References

1. Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, et al. Molecular portraits of human breast tumours. Nature. 2000;406(6797):747-52.

2. Sorlie T. Molecular classification of breast tumors: toward improved diagnostics and treatments. Methods Mol Biol. 2007;360:91-114.

3. Sorlie T, Tibshirani R, Parker J, Hastie T, Marron JS, Nobel A, et al. Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci U S A. 2003;100(14):8418-23.

4. Soerjomataram I, Louwman MW, Ribot JG, Roukema JA, Coebergh JW. An overview of prognostic factors for long-term survivors of breast cancer. Breast cancer research and treatment. 2008;107(3):309-30.

5. Yersal O, Barutca S. Biological subtypes of breast cancer: Prognostic and therapeutic implications. World J Clin Oncol. 2014;5(3):412-24.

6. Furrer D, Sanschagrin F, Jacob S, Diorio C. Advantages and disadvantages of technologies for HER2 testing in breast cancer specimens. American journal of clinical pathology. 2015;144(5):686-703.

7. Wolff AC, Hammond ME, Hicks DG, Dowsett M, McShane LM, Allison KH, et al. Recommendations for human epidermal growth factor receptor 2 testing in breast cancer: American Society of Clinical Oncology/College of American Pathologists clinical practice guideline update. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2013;31(31):3997-4013.

8. Wolff AC, Hammond ME, Schwartz JN, Hagerty KL, Allred DC, Cote RJ, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. Archives of pathology & laboratory medicine. 2007;131(1):18-43.

9. Moja L, Tagliabue L, Balduzzi S, Parmelli E, Pistotti V, Guarneri V, et al. Trastuzumab containing regimens for early breast cancer. Cochrane Database Syst Rev. 2012(4):CD006243.

- 247 -

10. Balduzzi S, Mantarro S, Guarneri V, Tagliabue L, Pistotti V, Moja L, et al. Trastuzumab-containing regimens for metastatic breast cancer. Cochrane Database Syst Rev. 2014(6):CD006242.

11. Vu T, Claret FX. Trastuzumab: updated mechanisms of action and resistance in breast cancer. Front Oncol. 2012;2:62.

12. Pohlmann PR, Mayer IA, Mernaugh R. Resistance to Trastuzumab in Breast Cancer. Clin Cancer Res. 2009;15(24):7479-91.

13. Caskey CT. Using genetic diagnosis to determine individual therapeutic utility. Annu Rev Med. 2010;61:1-15.

14. Lackner MR, Wilson TR, Settleman J. Mechanisms of acquired resistance to targeted cancer therapies. Future Oncol. 2012;8(8):999-1014.

15. Moreno-Munoz D, de la Haba-Rodriguez JR, Conde F, Lopez-Sanchez LM, Valverde A, Hernandez V, et al. Genetic variants in the renin-angiotensin system predict response to bevacizumab in cancer patients. Eur J Clin Invest. 2015;45(12):1325-32.

16. Beauclair S, Formento P, Fischel JL, Lescaut W, Largillier R, Chamorey E, et al. Role of the HER2 [Ile655Val] genetic polymorphism in tumorogenesis and in the risk of trastuzumab-related cardiotoxicity. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2007;18(8):1335-41.

17. Fleishman SJ, Schlessinger J, Ben-Tal N. A putative molecular-activation switch in the transmembrane domain of erbB2. Proc Natl Acad Sci U S A. 2002;99(25):15937-40.

18. Stefansson OA, Moran S, Gomez A, Sayols S, Arribas-Jorba C, Sandoval J, et al. A DNA methylation-based definition of biologically distinct breast cancer subtypes. Mol Oncol. 2015;9(3):555-68.

19. Berube S, Lemieux J, Moore L, Maunsell E, Brisson J. Smoking at time of diagnosis and breast cancer-specific survival: new findings and systematic review with meta-analysis. Breast cancer research : BCR. 2014;16(2):R42.

- 248 -

20. Kwan ML, Chen WY, Flatt SW, Weltzien EK, Nechuta SJ, Poole EM, et al. Postdiagnosis alcohol consumption and breast cancer prognosis in the after breast cancer pooling project. Cancer Epidemiol Biomarkers Prev. 2013;22(1):32-41.

21. Passarelli MN, Newcomb PA, Hampton JM, Trentham-Dietz A, Titus LJ, Egan KM, et al. Cigarette Smoking Before and After Breast Cancer Diagnosis: Mortality From Breast Cancer and Smoking-Related Diseases. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2016;34(12):1315-22.

22. Hirko KA, Chen WY, Willett WC, Rosner BA, Hankinson SE, Beck AH, et al. Alcohol consumption and risk of breast cancer by molecular subtype: Prospective analysis of the nurses' health study after 26 years of follow-up. Int J Cancer. 2016;138(5):1094-101.

23. Ma C, Lin H, Leonard SS, Shi X, Ye J, Luo J. Overexpression of ErbB2 enhances ethanol-stimulated intracellular signaling and invasion of human mammary epithelial and breast cancer cells in vitro. Oncogene. 2003;22(34):5281-90.

24. Sanner T, Grimsrud TK. Nicotine: Carcinogenicity and Effects on Response to Cancer Treatment - A Review. Front Oncol. 2015;5:196.

25. Seibold P, Vrieling A, Heinz J, Obi N, Sinn HP, Flesch-Janys D, et al. Pre-diagnostic smoking behaviour and poorer prognosis in a German breast cancer patient cohort - Differential effects by tumour subtype, NAT2 status, BMI and alcohol intake. Cancer Epidemiol. 2014;38(4):419-26.

26. Siriwardhana N, Choudhary S, Wang HC. Precancerous model of human breast epithelial cells induced by NNK for prevention. Breast cancer research and treatment. 2008;109(3):427-41.

27. Hortobagyi GN, de la Garza Salazar J, Pritchard K, Amadori D, Haidinger R, Hudis CA, et al. The global breast cancer burden: variations in epidemiology and survival. Clinical breast cancer. 2005;6(5):391-401.

28. Homepage Canadian Cancer Society CCS.

29. Polyak K. Heterogeneity in breast cancer. J Clin Invest. 2011;121(10):3786-8.

- 249 -

30. Sinn HP, Kreipe H. A Brief Overview of the WHO Classification of Breast Tumors, 4th Edition, Focusing on Issues and Updates from the 3rd Edition. Breast Care (Basel). 2013;8(2):149-54.

31. Viale G. The current state of breast cancer classification. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2012;23 Suppl 10:x207-10.

32. Bjornstrom L, Sjoberg M. Mechanisms of estrogen receptor signaling: convergence of genomic and nongenomic actions on target genes. Mol Endocrinol. 2005;19(4):833-42.

33. Prat A, Pineda E, Adamo B, Galvan P, Fernandez A, Gaba L, et al. Clinical implications of the intrinsic molecular subtypes of breast cancer. Breast. 2015;24 Suppl 2:S26-35.

34. Parker JS, Mullins M, Cheang MC, Leung S, Voduc D, Vickery T, et al. Supervised risk predictor of breast cancer based on intrinsic subtypes. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2009;27(8):1160-7.

35. Sweeney C, Bernard PS, Factor RE, Kwan ML, Habel LA, Quesenberry CP, Jr., et al. Intrinsic subtypes from PAM50 gene expression assay in a population-based breast cancer cohort: differences by age, race, and tumor characteristics. Cancer Epidemiol Biomarkers Prev. 2014;23(5):714-24.

36. Bertucci F, Finetti P, Rougemont J, Charafe-Jauffret E, Cervera N, Tarpin C, et al. Gene expression profiling identifies molecular subtypes of inflammatory breast cancer. Cancer research. 2005;65(6):2170-8.

37. Cheang MC, Chia SK, Voduc D, Gao D, Leung S, Snider J, et al. Ki67 index, HER2 status, and prognosis of patients with luminal B breast cancer. J Natl Cancer Inst. 2009;101(10):736-50.

38. Rouzier R, Perou CM, Symmans WF, Ibrahim N, Cristofanilli M, Anderson K, et al. Breast cancer molecular subtypes respond differently to preoperative chemotherapy. Clin Cancer Res. 2005;11(16):5678-85.

- 250 -

39. Sotiriou C, Neo SY, McShane LM, Korn EL, Long PM, Jazaeri A, et al. Breast cancer classification and prognosis based on gene expression profiles from a population-based study. Proc Natl Acad Sci U S A. 2003;100(18):10393-8.

40. van de Vijver MJ, He YD, van't Veer LJ, Dai H, Hart AA, Voskuil DW, et al. A gene- expression signature as a predictor of survival in breast cancer. The New England journal of medicine. 2002;347(25):1999-2009.

41. Wang ZC, Lin M, Wei LJ, Li C, Miron A, Lodeiro G, et al. Loss of heterozygosity and its correlation with expression profiles in subclasses of invasive breast cancers. Cancer research. 2004;64(1):64-71.

42. Yu K, Lee CH, Tan PH, Tan P. Conservation of breast cancer molecular subtypes and transcriptional patterns of tumor progression across distinct ethnic populations. Clin Cancer Res. 2004;10(16):5508-17.

43. Zhao H, Langerod A, Ji Y, Nowels KW, Nesland JM, Tibshirani R, et al. Different gene expression patterns in invasive lobular and ductal carcinomas of the breast. Mol Biol Cell. 2004;15(6):2523-36.

44. Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A. 2001;98(19):10869-74.

45. Blows FM, Driver KE, Schmidt MK, Broeks A, van Leeuwen FE, Wesseling J, et al. Subtyping of breast cancer by immunohistochemistry to investigate a relationship between subtype and short and long term survival: a collaborative analysis of data for 10,159 cases from 12 studies. PLoS Med. 2010;7(5):e1000279.

46. 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(21):2492-502.

47. Hugh J, Hanson J, Cheang MC, Nielsen TO, Perou CM, Dumontet C, et al. Breast cancer subtypes and response to docetaxel in node-positive breast cancer: use of an immunohistochemical definition in the BCIRG 001 trial. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2009;27(8):1168-76.

- 251 -

48. 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(16):5367-74.

49. Duraiyan J, Govindarajan R, Kaliyappan K, Palanisamy M. Applications of immunohistochemistry. J Pharm Bioallied Sci. 2012;4(Suppl 2):S307-9.

50. Bishop R. Applications of fluorescence in situ hybridization (FISH) in detecting genetic aberrations of medical significance. Bioscience Horizons. 2010;3(1):85-95.

51. Goldhirsch A, Wood WC, Coates AS, Gelber RD, Thurlimann B, Senn HJ, et al. Strategies for subtypes--dealing with the diversity of breast cancer: highlights of the St. Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2011. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2011;22(8):1736-47.

52. Inwald EC, Klinkhammer-Schalke M, Hofstadter F, Zeman F, Koller M, Gerstenhauer M, et al. Ki-67 is a prognostic parameter in breast cancer patients: results of a large population-based cohort of a cancer registry. Breast cancer research and treatment. 2013;139(2):539-52.

53. Bevilacqua P, Verderio P, Barbareschi M, Bonoldi E, Boracchi P, Dalla Palma P, et al. Lack of prognostic significance of the monoclonal antibody Ki-S1, a novel marker of proliferative activity, in node-negative breast carcinoma. Breast cancer research and treatment. 1996;37(2):123-33.

54. Bustreo S, Osella-Abate S, Cassoni P, Donadio M, Airoldi M, Pedani F, et al. Optimal Ki67 cut-off for luminal breast cancer prognostic evaluation: a large case series study with a long-term follow-up. Breast cancer research and treatment. 2016;157(2):363-71.

55. Colozza M, Sidoni A, Piccart-Gebhart M. Value of Ki67 in breast cancer: the debate is still open. The Lancet Oncology. 2010;11(5):414-5.

56. Beresford MJ, Wilson GD, Makris A. Measuring proliferation in breast cancer: practicalities and applications. Breast cancer research : BCR. 2006;8(6):216.

- 252 -

57. van Diest PJ, van der Wall E, Baak JP. Prognostic value of proliferation in invasive breast cancer: a review. Journal of clinical pathology. 2004;57(7):675-81.

58. Yousef EM, Furrer, D., Laperriere, D., Tahir M.R., Mader, S., Diorio, C., Gaboury, L.A. MCM2: an alternative to Ki-67 for measuring breast cancer cell proliferation. Modern pathology. 2017;In press.

59. Fragkos M, Ganier O, Coulombe P, Mechali M. DNA replication origin activation in space and time. Nature reviews Molecular cell biology. 2015;16(6):360-74.

60. Gonzalez MA, Tachibana KE, Laskey RA, Coleman N. Control of DNA replication and its potential clinical exploitation. Nat Rev Cancer. 2005;5(2):135-41.

61. Shin JH, Kelman Z. The replicative helicases of bacteria, archaea, and eukarya can unwind RNA-DNA hybrid substrates. The Journal of biological chemistry. 2006;281(37):26914-21.

62. Kittaneh M, Montero AJ, Gluck S. Molecular profiling for breast cancer: a comprehensive review. Biomark Cancer. 2013;5:61-70.

63. de Ronde JJ, Hannemann J, Halfwerk H, Mulder L, Straver ME, Vrancken Peeters MJ, et al. Concordance of clinical and molecular breast cancer subtyping in the context of preoperative chemotherapy response. Breast cancer research and treatment. 2010;119(1):119-26.

64. Prat A, Perou CM. Deconstructing the molecular portraits of breast cancer. Mol Oncol. 2011;5(1):5-23.

65. Weigelt B, Mackay A, A'Hern R, Natrajan R, Tan DS, Dowsett M, et al. Breast cancer molecular profiling with single sample predictors: a retrospective analysis. The Lancet Oncology. 2010;11(4):339-49.

66. Popescu NC, King CR, Kraus MH. Localization of the human erbB-2 gene on normal and rearranged chromosomes 17 to bands q12-21.32. Genomics. 1989;4(3):362-6.

67. Press MF, Cordon-Cardo C, Slamon DJ. Expression of the HER-2/neu proto- oncogene in normal human adult and fetal tissues. Oncogene. 1990;5(7):953-62.

- 253 -

68. Ross JS, Fletcher JA, Bloom KJ, Linette GP, Stec J, Clark E, et al. HER-2/neu testing in breast cancer. American journal of clinical pathology. 2003;120 Suppl:S53-71.

69. Kallioniemi OP, Kallioniemi A, Kurisu W, Thor A, Chen LC, Smith HS, et al. ERBB2 amplification in breast cancer analyzed by fluorescence in situ hybridization. Proc Natl Acad Sci U S A. 1992;89(12):5321-5.

70. Iqbal N, Iqbal N. Human Epidermal Growth Factor Receptor 2 (HER2) in Cancers: Overexpression and Therapeutic Implications. Mol Biol Int. 2014;2014:852748.

71. Esteva FJ, Yu D, Hung MC, Hortobagyi GN. Molecular predictors of response to trastuzumab and lapatinib in breast cancer. Nature reviews Clinical oncology. 2010;7(2):98- 107.

72. Varshney D, Zhou YY, Geller SA, Alsabeh R. Determination of HER-2 status and chromosome 17 polysomy in breast carcinomas comparing HercepTest and PathVysion FISH assay. American journal of clinical pathology. 2004;121(1):70-7.

73. Ross JS, Slodkowska EA, Symmans WF, Pusztai L, Ravdin PM, Hortobagyi GN. The HER-2 receptor and breast cancer: ten years of targeted anti-HER-2 therapy and personalized medicine. Oncologist. 2009;14(4):320-68.

74. Tse CH, Hwang HC, Goldstein LC, Kandalaft PL, Wiley JC, Kussick SJ, et al. Determining true HER2 gene status in breast cancers with polysomy by using alternative chromosome 17 reference genes: implications for anti-HER2 targeted therapy. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2011;29(31):4168-74.

75. Stevens R, Almanaseer I, Gonzalez M, Caglar D, Knudson RA, Ketterling RP, et al. Analysis of HER2 gene amplification using an automated fluorescence in situ hybridization signal enumeration system. J Mol Diagn. 2007;9(2):144-50.

76. Tubbs RR, Pettay JD, Swain E, Roche PC, Powell W, Hicks DG, et al. Automation of manual components and image quantification of direct dual label fluorescence in situ hybridization (FISH) for HER2 gene amplification: A feasibility study. Appl Immunohistochem Mol Morphol. 2006;14(4):436-40.

- 254 -

77. Lörch T, Piper J, Tmisek JD. "Tile sampling": a new method for the automated quantitative analysis of samples with high cell density and its application to Her2 scanning. Proceedings of Quantitative Molecular Cytogenetics. 2002.

78. Furrer D, Jacob S, Caron C, Sanschagrin F, Provencher L, Diorio C. Validation of a new classifier for the automated analysis of the human epidermal growth factor receptor 2 (HER2) gene amplification in breast cancer specimens. Diagn Pathol. 2013;8:17.

79. Theodosiou Z, Kasampalidis IN, Karayannopoulou G, Kostopoulos I, Bobos M, Bevilacqua G, et al. Evaluation of FISH image analysis system on assessing HER2 amplification in breast carcinoma cases. Breast. 2008;17(1):80-4.

80. Kallioniemi OP, Wagner U, Kononen J, Sauter G. Tissue microarray technology for high-throughput molecular profiling of cancer. Human molecular genetics. 2001;10(7):657- 62.

81. Kampf C, Olsson I, Ryberg U, Sjostedt E, Ponten F. Production of tissue microarrays, immunohistochemistry staining and digitalization within the human protein atlas. J Vis Exp. 2012(63).

82. Vogel U. Overview on Techniques to Construct Tissue Arrays with Special Emphasis on Tissue Microarrays. Microarrays (Basel). 2014;3(2):103-36.

83. Gancberg D, Di Leo A, Rouas G, Jarvinen T, Verhest A, Isola J, et al. Reliability of the tissue microarray based FISH for evaluation of the HER-2 oncogene in breast carcinoma. Journal of clinical pathology. 2002;55(4):315-7.

84. Graham AD, Faratian D, Rae F, Thomas JS. Tissue microarray technology in the routine assessment of HER-2 status in invasive breast cancer: a prospective study of the use of immunohistochemistry and fluorescence in situ hybridization. Histopathology. 2008;52(7):847-55.

85. O'Grady A, Flahavan CM, Kay EW, Barrett HL, Leader MB. HER-2 analysis in tissue microarrays of archival human breast cancer: comparison of immunohistochemistry and fluorescence in situ hybridization. Appl Immunohistochem Mol Morphol. 2003;11(2):177-82.

- 255 -

86. Zhang D, Salto-Tellez M, Putti TC, Do E, Koay ES. Reliability of tissue microarrays in detecting protein expression and gene amplification in breast cancer. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2003;16(1):79-84.

87. Cuadros M, Villegas R. Systematic review of HER2 breast cancer testing. Appl Immunohistochem Mol Morphol. 2009;17(1):1-7.

88. Bahreini F, Soltanian AR, Mehdipour P. A meta-analysis on concordance between immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) to detect HER2 gene overexpression in breast cancer. Breast cancer. 2015;22(6):615-25.

89. Hicks DG, Schiffhauer L. Standardized assessment of the HER2 status in breast cancer by immunohistochemistry. Lab Med. 2011;42:459-67.

90. Hoang MP, Sahin AA, Ordonez NG, Sneige N. HER-2/neu gene amplification compared with HER-2/neu protein overexpression and interobserver reproducibility in invasive breast carcinoma. American journal of clinical pathology. 2000;113(6):852-9.

91. Kovacs A, Stenman G. HER2-testing in 538 consecutive breast cancer cases using FISH and immunohistochemistry. Pathology, research and practice. 2010;206(1):39-42.

92. Lal P, Salazar PA, Hudis CA, Ladanyi M, Chen B. HER-2 testing in breast cancer using immunohistochemical analysis and fluorescence in situ hybridization: a single- institution experience of 2,279 cases and comparison of dual-color and single-color scoring. American journal of clinical pathology. 2004;121(5):631-6.

93. Owens MA, Horten BC, Da Silva MM. HER2 amplification ratios by fluorescence in situ hybridization and correlation with immunohistochemistry in a cohort of 6556 breast cancer tissues. Clinical breast cancer. 2004;5(1):63-9.

94. Rosa FE, Santos RM, Rogatto SR, Domingues MA. Chromogenic in situ hybridization compared with other approaches to evaluate HER2/neu status in breast carcinomas. Braz J Med Biol Res. 2013;46(3):207-16.

95. Yarden Y, Sliwkowski MX. Untangling the ErbB signalling network. Nature reviews Molecular cell biology. 2001;2(2):127-37.

- 256 -

96. Hubbard SR, Till JH. Protein tyrosine kinase structure and function. Annu Rev Biochem. 2000;69:373-98.

97. Carpenter G. Receptors for epidermal growth factor and other polypeptide mitogens. Annu Rev Biochem. 1987;56:881-914.

98. Roskoski R, Jr. ErbB/HER protein-tyrosine kinases: Structures and small molecule inhibitors. Pharmacol Res. 2014;87:42-59.

99. Sierke SL, Cheng K, Kim HH, Koland JG. Biochemical characterization of the protein tyrosine kinase homology domain of the ErbB3 (HER3) receptor protein. Biochem J. 1997;322 ( Pt 3):757-63.

100. Kim HH, Vijapurkar U, Hellyer NJ, Bravo D, Koland JG. Signal transduction by epidermal growth factor and heregulin via the kinase-deficient ErbB3 protein. Biochem J. 1998;334 ( Pt 1):189-95.

101. Carrera AC, Alexandrov K, Roberts TM. The conserved lysine of the catalytic domain of protein kinases is actively involved in the phosphotransfer reaction and not required for anchoring ATP. Proc Natl Acad Sci U S A. 1993;90(2):442-6.

102. Lemmon MA, Schlessinger J. Cell signaling by receptor tyrosine kinases. Cell. 2010;141(7):1117-34.

103. Zaczek A, Brandt B, Bielawski KP. The diverse signaling network of EGFR, HER2, HER3 and HER4 tyrosine kinase receptors and the consequences for therapeutic approaches. Histol Histopathol. 2005;20(3):1005-15.

104. Wilson KJ, Gilmore JL, Foley J, Lemmon MA, Riese DJ, 2nd. Functional selectivity of EGF family peptide growth factors: implications for cancer. Pharmacol Ther. 2009;122(1):1-8.

105. Rubin I, Yarden Y. The basic biology of HER2. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2001;12 Suppl 1:S3-8.

106. Burgess AW, Cho HS, Eigenbrot C, Ferguson KM, Garrett TP, Leahy DJ, et al. An open-and-shut case? Recent insights into the activation of EGF/ErbB receptors. Mol Cell. 2003;12(3):541-52.

- 257 -

107. Sliwkowski MX. Ready to partner. Nat Struct Biol. 2003;10(3):158-9.

108. Chen H, Xu CF, Ma J, Eliseenkova AV, Li W, Pollock PM, et al. A crystallographic snapshot of tyrosine trans-phosphorylation in action. Proc Natl Acad Sci U S A. 2008;105(50):19660-5.

109. Paul MK, Mukhopadhyay AK. Tyrosine kinase - Role and significance in Cancer. Int J Med Sci. 2004;1(2):101-15.

110. Moasser MM. The oncogene HER2: its signaling and transforming functions and its role in human cancer pathogenesis. Oncogene. 2007;26(45):6469-87.

111. Nguyen B, Keane MM, Johnston PG. The biology of growth regulation in normal and malignant breast epithelium: from bench to clinic. Crit Rev Oncol Hematol. 1995;20(3):223- 36.

112. Hemmings BA, Restuccia DF. PI3K-PKB/Akt pathway. Cold Spring Harb Perspect Biol. 2012;4(9):a011189.

113. Santarpia L, Lippman SM, El-Naggar AK. Targeting the MAPK-RAS-RAF signaling pathway in cancer therapy. Expert Opin Ther Targets. 2012;16(1):103-19.

114. Baulida J, Kraus MH, Alimandi M, Di Fiore PP, Carpenter G. All ErbB receptors other than the epidermal growth factor receptor are endocytosis impaired. The Journal of biological chemistry. 1996;271(9):5251-7.

115. Bertelsen V, Stang E. The Mysterious Ways of ErbB2/HER2 Trafficking. Membranes (Basel). 2014;4(3):424-46.

116. Hudis CA. Trastuzumab--mechanism of action and use in clinical practice. The New England journal of medicine. 2007;357(1):39-51.

117. Tzahar E, Waterman H, Chen X, Levkowitz G, Karunagaran D, Lavi S, et al. A hierarchical network of interreceptor interactions determines signal transduction by Neu differentiation factor/neuregulin and epidermal growth factor. Mol Cell Biol. 1996;16(10):5276-87.

- 258 -

118. Atalay G, Cardoso F, Awada A, Piccart MJ. Novel therapeutic strategies targeting the epidermal growth factor receptor (EGFR) family and its downstream effectors in breast cancer. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2003;14(9):1346-63.

119. Graus-Porta D, Beerli RR, Daly JM, Hynes NE. ErbB-2, the preferred heterodimerization partner of all ErbB receptors, is a mediator of lateral signaling. EMBO J. 1997;16(7):1647-55.

120. Gutierrez C, Schiff R. HER2: biology, detection, and clinical implications. Archives of pathology & laboratory medicine. 2011;135(1):55-62.

121. Way TD, Lin JK. Role of HER2/HER3 co-receptor in breast carcinogenesis. Future Oncol. 2005;1(6):841-9.

122. Nahta R, Yuan LX, Zhang B, Kobayashi R, Esteva FJ. Insulin-like growth factor-I receptor/human epidermal growth factor receptor 2 heterodimerization contributes to trastuzumab resistance of breast cancer cells. Cancer research. 2005;65(23):11118-28.

123. Bononi A, Agnoletto C, De Marchi E, Marchi S, Patergnani S, Bonora M, et al. Protein kinases and phosphatases in the control of cell fate. Enzyme Res. 2011;2011:329098.

124. Arteaga CL. Epidermal growth factor receptor dependence in human tumors: more than just expression? . Oncologist. 2002;7(Suppl 4):31-9.

125. Brooks AN, Kilgour E, Smith PD. Molecular pathways: fibroblast growth factor signaling: a new therapeutic opportunity in cancer. Clin Cancer Res. 2012;18(7):1855-62.

126. Nishikawa R, Ji XD, Harmon RC, Lazar CS, Gill GN, Cavenee WK, et al. A mutant epidermal growth factor receptor common in human glioma confers enhanced tumorigenicity. Proceedings of the National Academy of Sciences of the United States of America. 1994;91(16):7727-31.

127. Ocana A, Vera-Badillo F, Seruga B, Templeton A, Pandiella A, Amir E. HER3 overexpression and survival in solid tumors: a meta-analysis. J Natl Cancer Inst. 2013;105(4):266-73.

- 259 -

128. Roskoski R, Jr. Vascular endothelial growth factor (VEGF) signaling in tumor progression. Crit Rev Oncol Hematol. 2007;62(3):179-213.

129. Stern DF. Tyrosine kinase signalling in breast cancer: ErbB family receptor tyrosine kinases. Breast cancer research : BCR. 2000;2(3):176-83.

130. Hendriks BS, Opresko LK, Wiley HS, Lauffenburger D. Quantitative analysis of HER2-mediated effects on HER2 and epidermal growth factor receptor endocytosis: distribution of homo- and heterodimers depends on relative HER2 levels. The Journal of biological chemistry. 2003;278(26):23343-51.

131. Huang G, Chantry A, Epstein RJ. Overexpression of ErbB2 impairs ligand-dependent downregulation of epidermal growth factor receptors via a post-transcriptional mechanism. J Cell Biochem. 1999;74(1):23-30.

132. Wang Z, Zhang L, Yeung TK, Chen X. Endocytosis deficiency of epidermal growth factor (EGF) receptor-ErbB2 heterodimers in response to EGF stimulation. Mol Biol Cell. 1999;10(5):1621-36.

133. Lenferink AE, Pinkas-Kramarski R, van de Poll ML, van Vugt MJ, Klapper LN, Tzahar E, et al. Differential endocytic routing of homo- and hetero-dimeric ErbB tyrosine kinases confers signaling superiority to receptor heterodimers. EMBO J. 1998;17(12):3385-97.

134. Waterman H, Yarden Y. Molecular mechanisms underlying endocytosis and sorting of ErbB receptor tyrosine kinases. FEBS Lett. 2001;490(3):142-52.

135. Hendriks BS, Wiley HS, Lauffenburger D. HER2-mediated effects on EGFR endosomal sorting: analysis of biophysical mechanisms. Biophys J. 2003;85(4):2732-45.

136. Timms JF, White SL, O'Hare MJ, Waterfield MD. Effects of ErbB-2 overexpression on mitogenic signalling and cell cycle progression in human breast luminal epithelial cells. Oncogene. 2002;21(43):6573-86.

137. Arteaga CL, Sliwkowski MX, Osborne CK, Perez EA, Puglisi F, Gianni L. Treatment of HER2-positive breast cancer: current status and future perspectives. Nature reviews Clinical oncology. 2011;9(1):16-32.

- 260 -

138. Baselga J, Norton L, Albanell J, Kim YM, Mendelsohn J. Recombinant humanized anti-HER2 antibody (Herceptin) enhances the antitumor activity of paclitaxel and doxorubicin against HER2/neu overexpressing human breast cancer xenografts. Cancer research. 1998;58(13):2825-31.

139. Nahta R. Molecular Mechanisms of Trastuzumab-Based Treatment in HER2- Overexpressing Breast Cancer. ISRN oncology. 2012;2012:428062.

140. Sliwkowski MX, Lofgren JA, Lewis GD, Hotaling TE, Fendly BM, Fox JA. Nonclinical studies addressing the mechanism of action of trastuzumab (Herceptin). Semin Oncol. 1999;26(4 Suppl 12):60-70.

141. Baselga J, Albanell J. Mechanism of action of anti-HER2 monoclonal antibodies. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2001;12 Suppl 1:S35-41.

142. Baselga J, Albanell J, Molina MA, Arribas J. Mechanism of action of trastuzumab and scientific update. Semin Oncol. 2001;28(5 Suppl 16):4-11.

143. Lane HA, Motoyama AB, Beuvink I, Hynes NE. Modulation of p27/Cdk2 complex formation through 4D5-mediated inhibition of HER2 receptor signaling. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2001;12 Suppl 1:S21-2.

144. Nagata Y, Lan KH, Zhou X, Tan M, Esteva FJ, Sahin AA, et al. PTEN activation contributes to tumor inhibition by trastuzumab, and loss of PTEN predicts trastuzumab resistance in patients. Cancer Cell. 2004;6(2):117-27.

145. Georgescu MM. PTEN Tumor Suppressor Network in PI3K-Akt Pathway Control. Genes Cancer. 2010;1(12):1170-7.

146. Wen XF, Yang G, Mao W, Thornton A, Liu J, Bast RC, Jr., et al. HER2 signaling modulates the equilibrium between pro- and antiangiogenic factors via distinct pathways: implications for HER2-targeted antibody therapy. Oncogene. 2006;25(52):6986-96.

- 261 -

147. Shi Y, Fan X, Meng W, Deng H, Zhang N, An Z. Engagement of immune effector cells by trastuzumab induces HER2/ERBB2 downregulation in cancer cells through STAT1 activation. Breast cancer research : BCR. 2014;16(2):R33.

148. Hirvinen M, Heiskanen R, Oksanen M, Pesonen S, Liikanen I, Joensuu T, et al. Fc- gamma receptor polymorphisms as predictive and prognostic factors in patients receiving oncolytic adenovirus treatment. J Transl Med. 2013;11:193.

149. Varchetta S, Gibelli N, Oliviero B, Nardini E, Gennari R, Gatti G, et al. Elements related to heterogeneity of antibody-dependent cell cytotoxicity in patients under trastuzumab therapy for primary operable breast cancer overexpressing Her2. Cancer research. 2007;67(24):11991-9.

150. Kumar R, Shepard HM, Mendelsohn J. Regulation of phosphorylation of the c-erbB- 2/HER2 gene product by a monoclonal antibody and serum growth factor(s) in human mammary carcinoma cells. Mol Cell Biol. 1991;11(2):979-86.

151. Gijsen M, King P, Perera T, Parker PJ, Harris AL, Larijani B, et al. HER2 phosphorylation is maintained by a PKB negative feedback loop in response to anti-HER2 herceptin in breast cancer. PLoS Biol. 2010;8(12):e1000563.

152. Scaltriti M, Verma C, Guzman M, Jimenez J, Parra JL, Pedersen K, et al. Lapatinib, a HER2 tyrosine kinase inhibitor, induces stabilization and accumulation of HER2 and potentiates trastuzumab-dependent cell cytotoxicity. Oncogene. 2009;28(6):803-14.

153. Cho HS, Mason K, Ramyar KX, Stanley AM, Gabelli SB, Denney DW, Jr., et al. Structure of the extracellular region of HER2 alone and in complex with the Herceptin Fab. Nature. 2003;421(6924):756-60.

154. Franklin MC, Carey KD, Vajdos FF, Leahy DJ, de Vos AM, Sliwkowski MX. Insights into ErbB signaling from the structure of the ErbB2-pertuzumab complex. Cancer Cell. 2004;5(4):317-28.

155. Blumenthal GM, Scher NS, Cortazar P, Chattopadhyay S, Tang S, Song P, et al. First FDA approval of dual anti-HER2 regimen: pertuzumab in combination with trastuzumab and docetaxel for HER2-positive metastatic breast cancer. Clin Cancer Res. 2013;19(18):4911-6.

- 262 -

156. Agus DB, Akita RW, Fox WD, Lewis GD, Higgins B, Pisacane PI, et al. Targeting ligand-activated ErbB2 signaling inhibits breast and prostate tumor growth. Cancer Cell. 2002;2(2):127-37.

157. Ghosh R, Narasanna A, Wang SE, Liu S, Chakrabarty A, Balko JM, et al. Trastuzumab has preferential activity against breast cancers driven by HER2 homodimers. Cancer research. 2011;71(5):1871-82.

158. Ocana A, Amir E, Seruga B, Martin M, Pandiella A. The evolving landscape of protein kinases in breast cancer: clinical implications. Cancer Treat Rev. 2013;39(1):68-76.

159. Grossi PM, Ochiai H, Archer GE, McLendon RE, Zalutsky MR, Friedman AH, et al. Efficacy of intracerebral microinfusion of trastuzumab in an athymic rat model of intracerebral metastatic breast cancer. Clin Cancer Res. 2003;9(15):5514-20.

160. Lin NU, Carey LA, Liu MC, Younger J, Come SE, Ewend M, et al. Phase II trial of lapatinib for brain metastases in patients with human epidermal growth factor receptor 2- positive breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2008;26(12):1993-9.

161. Lin NU, Dieras V, Paul D, Lossignol D, Christodoulou C, Stemmler HJ, et al. Multicenter phase II study of lapatinib in patients with brain metastases from HER2-positive breast cancer. Clin Cancer Res. 2009;15(4):1452-9.

162. Nahta R, Yuan LX, Du Y, Esteva FJ. Lapatinib induces apoptosis in trastuzumab- resistant breast cancer cells: effects on insulin-like growth factor I signaling. Mol Cancer Ther. 2007;6(2):667-74.

163. Poon KA, Flagella K, Beyer J, Tibbitts J, Kaur S, Saad O, et al. Preclinical safety profile of trastuzumab emtansine (T-DM1): mechanism of action of its cytotoxic component retained with improved tolerability. Toxicol Appl Pharmacol. 2013;273(2):298-313.

164. Maher M. Current and Emerging Treatment Regimens for HER2-Positive Breast Cancer. P T. 2014;39(3):206-12.

165. Rosen LS, Ashurst HL, Chap L. Targeting signal transduction pathways in metastatic breast cancer: a comprehensive review. The oncologist. 2010;15(3):216-35.

- 263 -

166. Britten CD, Finn RS, Bosserman LD, Wong SG, Press MF, Malik M, et al. A phase I/II trial of trastuzumab plus erlotinib in metastatic HER2-positive breast cancer: a dual ErbB targeted approach. Clinical breast cancer. 2009;9(1):16-22.

167. Dickler MN, Cobleigh MA, Miller KD, Klein PM, Winer EP. Efficacy and safety of erlotinib in patients with locally advanced or metastatic breast cancer. Breast cancer research and treatment. 2009;115(1):115-21.

168. Emlet DR, Brown KA, Kociban DL, Pollice AA, Smith CA, Ong BB, et al. Response to trastuzumab, erlotinib, and bevacizumab, alone and in combination, is correlated with the level of human epidermal growth factor receptor-2 expression in human breast cancer cell lines. Mol Cancer Ther. 2007;6(10):2664-74.

169. Somlo G, Martel CL, Lau SK, Frankel P, Ruel C, Gu L, et al. A phase I/II prospective, single arm trial of gefitinib, trastuzumab, and docetaxel in patients with stage IV HER-2 positive metastatic breast cancer. Breast cancer research and treatment. 2012;131(3):899- 906.

170. Hickish T, Wheatley D, Lin N, Carey L, Houston S, Mendelson D, et al. Use of BIBW 2992, a novel irreversible EGFR/HER2 tyrosine kinase inhibitor (TKI), to treat patients with HER2-positive metastatic breast cancer after failure of treatment with trastuzumab. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2009;27(15_suppl):1023.

171. Wong KK, Fracasso PM, Bukowski RM, Lynch TJ, Munster PN, Shapiro GI, et al. A phase I study with neratinib (HKI-272), an irreversible pan ErbB receptor tyrosine kinase inhibitor, in patients with solid tumors. Clin Cancer Res. 2009;15(7):2552-8.

172. Geyer CE, Forster J, Lindquist D, Chan S, Romieu CG, Pienkowski T, et al. Lapatinib plus capecitabine for HER2-positive advanced breast cancer. The New England journal of medicine. 2006;355(26):2733-43.

173. Chan S, Scheulen ME, Johnston S, Mross K, Cardoso F, Dittrich C, et al. Phase II study of temsirolimus (CCI-779), a novel inhibitor of mTOR, in heavily pretreated patients with locally advanced or metastatic breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2005;23(23):5314-22.

- 264 -

174. Ellard SL, Clemons M, Gelmon KA, Norris B, Kennecke H, Chia S, et al. Randomized phase II study comparing two schedules of everolimus in patients with recurrent/metastatic breast cancer: NCIC Clinical Trials Group IND.163. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2009;27(27):4536-41.

175. Fleming GF, Ma CX, Huo D, Sattar H, Tretiakova M, Lin L, et al. Phase II trial of temsirolimus in patients with metastatic breast cancer. Breast cancer research and treatment. 2012;136(2):355-63.

176. Verma S, Ewer MS. Is cardiotoxicity being adequately assessed in current trials of cytotoxic and targeted agents in breast cancer? Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2011;22(5):1011-8.

177. Slamon D, Eiermann W, Robert N, Pienkowski T, Martin M, Press M, et al. Adjuvant trastuzumab in HER2-positive breast cancer. The New England journal of medicine. 2011;365(14):1273-83.

178. Keefe DL. Trastuzumab-associated cardiotoxicity. Cancer. 2002;95(7):1592-600.

179. Hansel TT, Kropshofer H, Singer T, Mitchell JA, George AJ. The safety and side effects of monoclonal antibodies. Nat Rev Drug Discov. 2010;9(4):325-38.

180. Ozcelik C, Erdmann B, Pilz B, Wettschureck N, Britsch S, Hubner N, et al. Conditional mutation of the ErbB2 (HER2) receptor in cardiomyocytes leads to dilated cardiomyopathy. Proc Natl Acad Sci U S A. 2002;99(13):8880-5.

181. Minotti G, Menna P, Salvatorelli E, Cairo G, Gianni L. Anthracyclines: molecular advances and pharmacologic developments in antitumor activity and cardiotoxicity. Pharmacol Rev. 2004;56(2):185-229.

182. Perez EA. Cardiac toxicity of ErbB2-targeted therapies: what do we know? Clinical breast cancer. 2008;8 Suppl 3:S114-20.

183. Sendur MA, Aksoy S, Altundag K. Cardiotoxicity of novel HER2-targeted therapies. Curr Med Res Opin. 2013;29(8):1015-24.

184. Moy B, Goss PE. Lapatinib-associated toxicity and practical management recommendations. The oncologist. 2007;12(7):756-65.

- 265 -

185. Wong DJ, Hurvitz SA. Recent advances in the development of anti-HER2 antibodies and antibody-drug conjugates. Ann Transl Med. 2014;2(12):122.

186. Burstein HJ, Temin S, Anderson H, Buchholz TA, Davidson NE, Gelmon KE, et al. Adjuvant endocrine therapy for women with hormone receptor-positive breast cancer: american society of clinical oncology clinical practice guideline focused update. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2014;32(21):2255-69.

187. Chang J, Fan W. Endocrine therapy resistance: current status, possible mechanisms and overcoming strategies. Anticancer Agents Med Chem. 2013;13(3):464-75.

188. Lerner LJ, Jordan VC. Development of antiestrogens and their use in breast cancer: eighth Cain memorial award lecture. Cancer research. 1990;50(14):4177-89.

189. Osborne CK, Zhao H, Fuqua SA. Selective estrogen receptor modulators: structure, function, and clinical use. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2000;18(17):3172-86.

190. Riggs BL, Hartmann LC. Selective estrogen-receptor modulators -- mechanisms of action and application to clinical practice. The New England journal of medicine. 2003;348(7):618-29.

191. Vogel VG. The NSABP Study of Tamoxifen and Raloxifene (STAR) trial. Expert Rev Anticancer Ther. 2009;9(1):51-60.

192. Hubalek M, Brantner C, Marth C. Adjuvant endocrine therapy of premenopausal women with early breast cancer: an overview. Wien Med Wochenschr. 2010;160(7-8):167- 73.

193. Goodsell DS. The molecular perspective: tamoxifen and the estrogen receptor. The oncologist. 2002;7(2):163-4.

194. Goetz MP, Knox SK, Suman VJ, Rae JM, Safgren SL, Ames MM, et al. The impact of cytochrome P450 2D6 metabolism in women receiving adjuvant tamoxifen. Breast cancer research and treatment. 2007;101(1):113-21.

- 266 -

195. Ruddy KJ, Desantis SD, Gelman RS, Wu AH, Punglia RS, Mayer EL, et al. Personalized medicine in breast cancer: tamoxifen, endoxifen, and CYP2D6 in clinical practice. Breast cancer research and treatment. 2013;141(3):421-7.

196. Osborne CK, Wakeling A, Nicholson RI. Fulvestrant: an oestrogen receptor antagonist with a novel mechanism of action. Br J Cancer. 2004;90 Suppl 1:S2-6.

197. Muchmore DB. Raloxifene: A selective estrogen receptor modulator (SERM) with multiple target system effects. The oncologist. 2000;5(5):388-92.

198. Gennari L, Merlotti D, Paola VD, Nuti R. Raloxifene in breast cancer prevention. Expert Opin Drug Saf. 2008;7(3):259-70.

199. Cummings SR, Eckert S, Krueger KA, Grady D, Powles TJ, Cauley JA, et al. The effect of raloxifene on risk of breast cancer in postmenopausal women: results from the MORE randomized trial. Multiple Outcomes of Raloxifene Evaluation. JAMA. 1999;281(23):2189-97.

200. Martino S, Cauley JA, Barrett-Connor E, Powles TJ, Mershon J, Disch D, et al. Continuing outcomes relevant to Evista: breast cancer incidence in postmenopausal osteoporotic women in a randomized trial of raloxifene. J Natl Cancer Inst. 2004;96(23):1751-61.

201. Vogel VG, Costantino JP, Wickerham DL, Cronin WM, Cecchini RS, Atkins JN, et al. Effects of tamoxifen vs raloxifene on the risk of developing invasive breast cancer and other disease outcomes: the NSABP Study of Tamoxifen and Raloxifene (STAR) P-2 trial. JAMA. 2006;295(23):2727-41.

202. Hong S, Didwania A, Olopade O, Ganschow P. The expanding use of third- generation aromatase inhibitors: what the general internist needs to know. J Gen Intern Med. 2009;24 Suppl 2:S383-8.

203. Ciruelos E, Pascual T, Arroyo Vozmediano ML, Blanco M, Manso L, Parrilla L, et al. The therapeutic role of fulvestrant in the management of patients with hormone receptor- positive breast cancer. Breast. 2014;23(3):201-8.

- 267 -

204. Lai A, Kahraman M, Govek S, Nagasawa J, Bonnefous C, Julien J, et al. Identification of GDC-0810 (ARN-810), an Orally Bioavailable Selective Estrogen Receptor Degrader (SERD) that Demonstrates Robust Activity in Tamoxifen-Resistant Breast Cancer Xenografts. J Med Chem. 2015;58(12):4888-904.

205. Howell A. Pure oestrogen antagonists for the treatment of advanced breast cancer. Endocr Relat Cancer. 2006;13(3):689-706.

206. Bracci L, Schiavoni G, Sistigu A, Belardelli F. Immune-based mechanisms of cytotoxic chemotherapy: implications for the design of novel and rationale-based combined treatments against cancer. Cell Death Differ. 2014;21(1):15-25.

207. Hidaka M, Koga T, Kiyota H, Horiguchi T, Shi QW, Hirose K, et al. Relationship between the structures of taxane derivatives and their microtubule polymerization activity. Biosci Biotechnol Biochem. 2012;76(2):349-52.

208. Sousa GF, Wlodarczyk, S.R., Monteiro, G. Carboplatin: molecular mechanisms of action associated with chemoresistance. Braz J Pharm Sci. 2014;50(4):693-701.

209. Emadi A, Jones RJ, Brodsky RA. Cyclophosphamide and cancer: golden anniversary. Nature reviews Clinical oncology. 2009;6(11):638-47.

210. Zhang N, Yin Y, Xu SJ, Chen WS. 5-Fluorouracil: mechanisms of resistance and reversal strategies. Molecules. 2008;13(8):1551-69.

211. Yang F, Teves SS, Kemp CJ, Henikoff S. Doxorubicin, DNA torsion, and chromatin dynamics. Biochim Biophys Acta. 2014;1845(1):84-9.

212. Basu S, Ma R, Boyle PJ, Mikulla B, Bradley M, Smith B, et al. Apoptosis of human carcinoma cells in the presence of potential anti-cancer drugs: III. Treatment of Colo-205 and SKBR3 cells with: cis -platin, Tamoxifen, Melphalan, Betulinic acid, L-PDMP, L-PPMP, and GD3 ganglioside. Glycoconj J. 2004;20(9):563-77.

213. Rexer BN, Arteaga CL. Intrinsic and acquired resistance to HER2-targeted therapies in HER2 gene-amplified breast cancer: mechanisms and clinical implications. Crit Rev Oncog. 2012;17(1):1-16.

- 268 -

214. Scaltriti M, Rojo F, Ocana A, Anido J, Guzman M, Cortes J, et al. Expression of p95HER2, a truncated form of the HER2 receptor, and response to anti-HER2 therapies in breast cancer. J Natl Cancer Inst. 2007;99(8):628-38.

215. Arribas J, Baselga J, Pedersen K, Parra-Palau JL. p95HER2 and breast cancer. Cancer research. 2011;71(5):1515-9.

216. Carraway KL, Price-Schiavi SA, Komatsu M, Jepson S, Perez A, Carraway CA. Muc4/sialomucin complex in the mammary gland and breast cancer. J Mammary Gland Biol Neoplasia. 2001;6(3):323-37.

217. Price-Schiavi SA, Jepson S, Li P, Arango M, Rudland PS, Yee L, et al. Rat Muc4 (sialomucin complex) reduces binding of anti-ErbB2 antibodies to tumor cell surfaces, a potential mechanism for herceptin resistance. Int J Cancer. 2002;99(6):783-91.

218. Tanner M, Kapanen AI, Junttila T, Raheem O, Grenman S, Elo J, et al. Characterization of a novel cell line established from a patient with Herceptin-resistant breast cancer. Mol Cancer Ther. 2004;3(12):1585-92.

219. Nagy P, Friedlander E, Tanner M, Kapanen AI, Carraway KL, Isola J, et al. Decreased accessibility and lack of activation of ErbB2 in JIMT-1, a herceptin-resistant, MUC4-expressing breast cancer cell line. Cancer research. 2005;65(2):473-82.

220. Ponta H, Sherman L, Herrlich PA. CD44: from adhesion molecules to signalling regulators. Nature reviews Molecular cell biology. 2003;4(1):33-45.

221. Palyi-Krekk Z, Barok M, Isola J, Tammi M, Szollosi J, Nagy P. Hyaluronan-induced masking of ErbB2 and CD44-enhanced trastuzumab internalisation in trastuzumab resistant breast cancer. Eur J Cancer. 2007;43(16):2423-33.

222. Porta C, Paglino C, Mosca A. Targeting PI3K/Akt/mTOR Signaling in Cancer. Front Oncol. 2014;4:64.

223. Brown KK, Toker A. The phosphoinositide 3-kinase pathway and therapy resistance in cancer. F1000Prime Rep. 2015;7:13.

224. Hennessy BT, Smith DL, Ram PT, Lu Y, Mills GB. Exploiting the PI3K/AKT pathway for cancer drug discovery. Nat Rev Drug Discov. 2005;4(12):988-1004.

- 269 -

225. Pandolfi PP. Breast cancer--loss of PTEN predicts resistance to treatment. The New England journal of medicine. 2004;351(22):2337-8.

226. Park YH, Jung HA, Choi MK, Chang W, Choi YL, Do IG, et al. Role of HER3 expression and PTEN loss in patients with HER2-overexpressing metastatic breast cancer (MBC) who received taxane plus trastuzumab treatment. Br J Cancer. 2014;110(2):384-91.

227. Berns K, Horlings HM, Hennessy BT, Madiredjo M, Hijmans EM, Beelen K, et al. A functional genetic approach identifies the PI3K pathway as a major determinant of trastuzumab resistance in breast cancer. Cancer Cell. 2007;12(4):395-402.

228. Esteva FJ, Guo H, Zhang S, Santa-Maria C, Stone S, Lanchbury JS, et al. PTEN, PIK3CA, p-AKT, and p-p70S6K status: association with trastuzumab response and survival in patients with HER2-positive metastatic breast cancer. The American journal of pathology. 2010;177(4):1647-56.

229. Sergina NV, Rausch M, Wang D, Blair J, Hann B, Shokat KM, et al. Escape from HER-family tyrosine kinase inhibitor therapy by the kinase-inactive HER3. Nature. 2007;445(7126):437-41.

230. Olayioye MA. Update on HER-2 as a target for cancer therapy: intracellular signaling pathways of ErbB2/HER-2 and family members. Breast cancer research : BCR. 2001;3(6):385-9.

231. Diermeier S, Horvath G, Knuechel-Clarke R, Hofstaedter F, Szollosi J, Brockhoff G. Epidermal growth factor receptor coexpression modulates susceptibility to Herceptin in HER2/neu overexpressing breast cancer cells via specific erbB-receptor interaction and activation. Exp Cell Res. 2005;304(2):604-19.

232. Ritter CA, Perez-Torres M, Rinehart C, Guix M, Dugger T, Engelman JA, et al. Human breast cancer cells selected for resistance to trastuzumab in vivo overexpress epidermal growth factor receptor and ErbB ligands and remain dependent on the ErbB receptor network. Clin Cancer Res. 2007;13(16):4909-19.

233. Yonemori K, Tsuta K, Shimizu C, Hatanaka Y, Hirakawa A, Ono M, et al. Immunohistochemical expression of HER1, HER3, and HER4 in HER2-positive breast

- 270 -

cancer patients treated with trastuzumab-containing neoadjuvant chemotherapy. J Surg Oncol. 2010;101(3):222-7.

234. Narayan M, Wilken JA, Harris LN, Baron AT, Kimbler KD, Maihle NJ. Trastuzumab- induced HER reprogramming in "resistant" breast carcinoma cells. Cancer research. 2009;69(6):2191-4.

235. Lipton A, Goodman L, Leitzel K, Cook J, Sperinde J, Haddad M, et al. HER3, p95HER2, and HER2 protein expression levels define multiple subtypes of HER2-positive metastatic breast cancer. Breast cancer research and treatment. 2013;141(1):43-53.

236. Lu Y, Zi X, Zhao Y, Mascarenhas D, Pollak M. Insulin-like growth factor-I receptor signaling and resistance to trastuzumab (Herceptin). J Natl Cancer Inst. 2001;93(24):1852- 7.

237. Jerome L, Alami N, Belanger S, Page V, Yu Q, Paterson J, et al. Recombinant human insulin-like growth factor binding protein 3 inhibits growth of human epidermal growth factor receptor-2-overexpressing breast tumors and potentiates herceptin activity in vivo. Cancer research. 2006;66(14):7245-52.

238. Chakraborty AK, Liang K, DiGiovanna MP. Co-targeting insulin-like growth factor I receptor and HER2: dramatic effects of HER2 inhibitors on nonoverexpressing breast cancer. Cancer research. 2008;68(5):1538-45.

239. Huang X, Gao L, Wang S, McManaman JL, Thor AD, Yang X, et al. Heterotrimerization of the growth factor receptors erbB2, erbB3, and insulin-like growth factor-i receptor in breast cancer cells resistant to herceptin. Cancer research. 2010;70(3):1204-14.

240. Nahta R. Deciphering the role of insulin-like growth factor-I receptor in trastuzumab resistance. Chemother Res Pract. 2012;2012:648965.

241. Kostler WJ, Hudelist G, Rabitsch W, Czerwenka K, Muller R, Singer CF, et al. Insulin- like growth factor-1 receptor (IGF-1R) expression does not predict for resistance to trastuzumab-based treatment in patients with Her-2/neu overexpressing metastatic breast cancer. J Cancer Res Clin Oncol. 2006;132(1):9-18.

- 271 -

242. Smith BL, Chin D, Maltzman W, Crosby K, Hortobagyi GN, Bacus SS. The efficacy of Herceptin therapies is influenced by the expression of other erbB receptors, their ligands and the activation of downstream signalling proteins. Br J Cancer. 2004;91(6):1190-4.

243. Harris LN, You F, Schnitt SJ, Witkiewicz A, Lu X, Sgroi D, et al. Predictors of resistance to preoperative trastuzumab and vinorelbine for HER2-positive early breast cancer. Clin Cancer Res. 2007;13(4):1198-207.

244. Shattuck DL, Miller JK, Carraway KL, 3rd, Sweeney C. Met receptor contributes to trastuzumab resistance of Her2-overexpressing breast cancer cells. Cancer research. 2008;68(5):1471-7.

245. Nahta R. Pharmacological strategies to overcome HER2 cross-talk and Trastuzumab resistance. Curr Med Chem. 2012;19(7):1065-75.

246. Anastasi S, Sala G, Huiping C, Caprini E, Russo G, Iacovelli S, et al. Loss of RALT/MIG-6 expression in ERBB2-amplified breast carcinomas enhances ErbB-2 oncogenic potency and favors resistance to Herceptin. Oncogene. 2005;24(28):4540-8.

247. Deveraux QL, Reed JC. IAP family proteins--suppressors of apoptosis. Genes Dev. 1999;13(3):239-52.

248. Perciavalle RM, Stewart DP, Koss B, Lynch J, Milasta S, Bathina M, et al. Anti- apoptotic MCL-1 localizes to the mitochondrial matrix and couples mitochondrial fusion to respiration. Nat Cell Biol. 2012;14(6):575-83.

249. Valabrega G, Capellero S, Cavalloni G, Zaccarello G, Petrelli A, Migliardi G, et al. HER2-positive breast cancer cells resistant to trastuzumab and lapatinib lose reliance upon HER2 and are sensitive to the multitargeted kinase inhibitor sorafenib. Breast cancer research and treatment. 2011;130(1):29-40.

250. Oliveras-Ferraros C, Vazquez-Martin A, Cufi S, Torres-Garcia VZ, Sauri-Nadal T, Barco SD, et al. Inhibitor of Apoptosis (IAP) survivin is indispensable for survival of HER2 gene-amplified breast cancer cells with primary resistance to HER1/2-targeted therapies. Biochem Biophys Res Commun. 2011;407(2):412-9.

- 272 -

251. Scaltriti M, Eichhorn PJ, Cortes J, Prudkin L, Aura C, Jimenez J, et al. Cyclin E amplification/overexpression is a mechanism of trastuzumab resistance in HER2+ breast cancer patients. Proc Natl Acad Sci U S A. 2011;108(9):3761-6.

252. Nahta R, Takahashi T, Ueno NT, Hung MC, Esteva FJ. P27(kip1) down-regulation is associated with trastuzumab resistance in breast cancer cells. Cancer research. 2004;64(11):3981-6.

253. Quail DF, Joyce JA. Microenvironmental regulation of tumor progression and metastasis. Nature medicine. 2013;19(11):1423-37.

254. Pattabiraman DR, Weinberg RA. Tackling the cancer stem cells - what challenges do they pose? Nat Rev Drug Discov. 2014;13(7):497-512.

255. Chen F, Zhuang X, Lin L, Yu P, Wang Y, Shi Y, et al. New horizons in tumor microenvironment biology: challenges and opportunities. BMC Med. 2015;13:45.

256. Duong MN, Cleret A, Matera EL, Chettab K, Mathe D, Valsesia-Wittmann S, et al. Adipose cells promote resistance of breast cancer cells to trastuzumab-mediated antibody- dependent cellular cytotoxicity. Breast cancer research : BCR. 2015;17:57.

257. Griner SE, Wang KJ, Joshi JP, Nahta R. Mechanisms of Adipocytokine-Mediated Trastuzumab Resistance in HER2-Positive Breast Cancer Cell Lines. Curr Pharmacogenomics Person Med. 2013;11(1):31-41.

258. Mao Y, Zhang Y, Qu Q, Zhao M, Lou Y, Liu J, et al. Cancer-associated fibroblasts induce trastuzumab resistance in HER2 positive breast cancer cells. Mol Biosyst. 2015;11(4):1029-40.

259. Hogarth PM. Fc receptors are major mediators of antibody based inflammation in autoimmunity. Curr Opin Immunol. 2002;14(6):798-802.

260. Nimmerjahn F, Ravetch JV. Fcgamma receptors as regulators of immune responses. Nat Rev Immunol. 2008;8(1):34-47.

261. Clynes RA, Towers TL, Presta LG, Ravetch JV. Inhibitory Fc receptors modulate in vivo cytotoxicity against tumor targets. Nature medicine. 2000;6(4):443-6.

- 273 -

262. Musolino A, Naldi N, Bortesi B, Pezzuolo D, Capelletti M, Missale G, et al. Immunoglobulin G fragment C receptor polymorphisms and clinical efficacy of trastuzumab- based therapy in patients with HER-2/neu-positive metastatic breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2008;26(11):1789-96.

263. Gennari R, Menard S, Fagnoni F, Ponchio L, Scelsi M, Tagliabue E, et al. Pilot study of the mechanism of action of preoperative trastuzumab in patients with primary operable breast tumors overexpressing HER2. Clin Cancer Res. 2004;10(17):5650-5.

264. Wilken JA, Maihle NJ. Primary trastuzumab resistance: new tricks for an old drug. Ann N Y Acad Sci. 2010;1210:53-65.

265. Erichsen HC, Chanock SJ. SNPs in cancer research and treatment. Br J Cancer. 2004;90(4):747-51.

266. Shastry BS. SNPs: impact on gene function and phenotype. Methods Mol Biol. 2009;578:3-22.

267. Chin L, Hahn WC, Getz G, Meyerson M. Making sense of cancer genomic data. Genes Dev. 2011;25(6):534-55.

268. Bayraktar S, Thompson PA, Yoo SY, Do KA, Sahin AA, Arun BK, et al. The relationship between eight GWAS-identified single-nucleotide polymorphisms and primary breast cancer outcomes. The oncologist. 2013;18(5):493-500.

269. Fasching PA, Pharoah PD, Cox A, Nevanlinna H, Bojesen SE, Karn T, et al. The role of genetic breast cancer susceptibility variants as prognostic factors. Human molecular genetics. 2012;21(17):3926-39.

270. Lu Q, Jiang K, Li Q, Ji YJ, Chen WL, Xue XH. Polymorphisms in the MTHFR gene are associated with breast cancer risk and prognosis in a Chinese population. Tumour Biol. 2015;36(5):3757-62.

271. Pirie A, Guo Q, Kraft P, Canisius S, Eccles DM, Rahman N, et al. Common germline polymorphisms associated with breast cancer-specific survival. Breast cancer research : BCR. 2015;17:58.

- 274 -

272. Sapkota Y. Germline DNA variations in breast cancer predisposition and prognosis: a systematic review of the literature. Cytogenet Genome Res. 2014;144(2):77-91.

273. Alaoui-Jamali MA, Morand GB, da Silva SD. ErbB polymorphisms: insights and implications for response to targeted cancer therapeutics. Front Genet. 2015;6:17.

274. http://www.ncbi.nlm.nih.gov/snp dd.

275. Han W, Kang D, Lee JE, Park IA, Choi JY, Lee KM, et al. A haplotype analysis of HER-2 gene polymorphisms: association with breast cancer risk, HER-2 protein expression in the tumor, and disease recurrence in Korea. Clin Cancer Res. 2005;11(13):4775-8.

276. Papewalis J, Nikitin AY, Rajewsky MF. G to A polymorphism at amino acid codon 655 of the human erbB-2/HER2 gene. Nucleic acids research. 1991;19(19):5452.

277. Tommasi S, Fedele V, Lacalamita R, Bruno M, Schittulli F, Ginzinger D, et al. 655Val and 1170Pro ERBB2 SNPs in familial breast cancer risk and BRCA1 alterations. Cell Oncol. 2007;29(3):241-8.

278. Kallel I, Kharrat N, Al-fadhly S, Rebai M, Khabir A, Boudawara TS, et al. HER2 polymorphisms and breast cancer in Tunisian women. Genet Test Mol Biomarkers. 2010;14(1):29-35.

279. Lee SC, Hou MF, Hsieh PC, Wu SH, Hou LA, Ma H, et al. A case-control study of the HER2 Ile655Val polymorphism and risk of breast cancer in Taiwan. Clin Biochem. 2008;41(3):121-5.

280. McKean-Cowdin R, Kolonel LN, Press MF, Pike MC, Henderson BE. Germ-line HER- 2 variant and breast cancer risk by stage of disease. Cancer research. 2001;61(23):8393-4.

281. Millikan RC, Hummer AJ, Wolff MS, Hishida A, Begg CB. HER2 codon 655 polymorphism and breast cancer: results from kin-cohort and case-control analyses. Breast cancer research and treatment. 2005;89(3):309-12.

282. Naidu R, Yip CH, Taib NA. Polymorphisms of HER2 Ile655Val and cyclin D1 (CCND1) G870A are not associated with breast cancer risk but polymorphic allele of HER2 is associated with nodal metastases. Neoplasma. 2008;55(2):87-95.

- 275 -

283. Papadopoulou E, Simopoulos K, Tripsianis G, Tentes I, Anagnostopoulos K, Sivridis E, et al. Allelic imbalance of HER-2 codon 655 polymorphism among different religious/ethnic populations of northern Greece and its association with the development and the malignant phenotype of breast cancer. Neoplasma. 2007;54(5):365-73.

284. Papadopoulou E, Tripsianis G, Anagnostopoulos K, Tentes I, Kakolyris S, Galazios G, et al. The influence of serum HER-2 levels and HER-2 codon 655 polymorphism on breast cancer outcome. Neoplasma. 2008;55(2):113-21.

285. Ozturk O, Canbay E, Kahraman OT, Fatih Seyhan M, Aydogan F, Celik V, et al. HER2 Ile655Val and PTEN IVS4 polymorphisms in patients with breast cancer. Mol Biol Rep. 2013;40(2):1813-8.

286. Watrowski R, Castillo-Tong DC, Wolf A, Schuster E, Fischer MB, Speiser P, et al. HER2 Codon 655 (Ile/Val) Polymorphism and Breast Cancer in Austrian Women. Anticancer Res. 2015;35(12):6667-70.

287. Slattery ML, John EM, Stern MC, Herrick J, Lundgreen A, Giuliano AR, et al. Associations with growth factor genes (FGF1, FGF2, PDGFB, FGFR2, NRG2, EGF, ERBB2) with breast cancer risk and survival: the Breast Cancer Health Disparities Study. Breast cancer research and treatment. 2013;140(3):587-601.

288. Han X, Diao L, Xu Y, Xue W, Ouyang T, Li J, et al. Association between the HER2 Ile655Val polymorphism and response to trastuzumab in women with operable primary breast cancer. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2014;25(6):1158-64.

289. Goldberg AD, Allis CD, Bernstein E. Epigenetics: a landscape takes shape. Cell. 2007;128(4):635-8.

290. Bhattacharya S, Zhang Q, Andersen ME. A deterministic map of Waddington's epigenetic landscape for cell fate specification. BMC Syst Biol. 2011;5:85.

291. Berger SL, Kouzarides T, Shiekhattar R, Shilatifard A. An operational definition of epigenetics. Genes Dev. 2009;23(7):781-3.

- 276 -

292. Gibney ER, Nolan CM. Epigenetics and gene expression. Heredity (Edinb). 2010;105(1):4-13.

293. Sharma S, Kelly TK, Jones PA. Epigenetics in cancer. Carcinogenesis. 2010;31(1):27-36.

294. Brookes E, Shi Y. Diverse epigenetic mechanisms of human disease. Annu Rev Genet. 2014;48:237-68.

295. Li E. Chromatin modification and epigenetic reprogramming in mammalian development. Nat Rev Genet. 2002;3(9):662-73.

296. van Dongen J, Ehli EA, Slieker RC, Bartels M, Weber ZM, Davies GE, et al. Epigenetic variation in monozygotic twins: a genome-wide analysis of DNA methylation in buccal cells. Genes (Basel). 2014;5(2):347-65.

297. Reddington JP, Pennings S, Meehan RR. Non-canonical functions of the DNA methylome in gene regulation. Biochem J. 2013;451(1):13-23.

298. Jin B, Li Y, Robertson KD. DNA methylation: superior or subordinate in the epigenetic hierarchy? Genes Cancer. 2011;2(6):607-17.

299. Wajed SA, Laird PW, DeMeester TR. DNA methylation: an alternative pathway to cancer. Ann Surg. 2001;234(1):10-20.

300. Ngo TT, Yoo J, Dai Q, Zhang Q, He C, Aksimentiev A, et al. Effects of cytosine modifications on DNA flexibility and nucleosome mechanical stability. Nat Commun. 2016;7:10813.

301. Probst AV, Dunleavy E, Almouzni G. Epigenetic inheritance during the cell cycle. Nature reviews Molecular cell biology. 2009;10(3):192-206.

302. Vilkaitis G, Suetake I, Klimasauskas S, Tajima S. Processive methylation of hemimethylated CpG sites by mouse Dnmt1 DNA methyltransferase. The Journal of biological chemistry. 2005;280(1):64-72.

303. Szyf M, Bozovic V, Tanigawa G. Growth regulation of mouse DNA methyltransferase gene expression. The Journal of biological chemistry. 1991;266(16):10027-30.

- 277 -

304. Hattori N, Abe T, Hattori N, Suzuki M, Matsuyama T, Yoshida S, et al. Preference of DNA methyltransferases for CpG islands in mouse embryonic stem cells. Genome Res. 2004;14(9):1733-40.

305. Weber M, Schubeler D. Genomic patterns of DNA methylation: targets and function of an epigenetic mark. Curr Opin Cell Biol. 2007;19(3):273-80.

306. Pastor WA, Aravind L, Rao A. TETonic shift: biological roles of TET proteins in DNA demethylation and transcription. Nature reviews Molecular cell biology. 2013;14(6):341-56.

307. Jones PA. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat Rev Genet. 2012;13(7):484-92.

308. Eckhardt F, Lewin J, Cortese R, Rakyan VK, Attwood J, Burger M, et al. DNA methylation profiling of human chromosomes 6, 20 and 22. Nat Genet. 2006;38(12):1378- 85.

309. Illingworth R, Kerr A, Desousa D, Jorgensen H, Ellis P, Stalker J, et al. A novel CpG island set identifies tissue-specific methylation at developmental gene loci. PLoS Biol. 2008;6(1):e22.

310. Rauch TA, Wu X, Zhong X, Riggs AD, Pfeifer GP. A human B cell methylome at 100- base pair resolution. Proc Natl Acad Sci U S A. 2009;106(3):671-8.

311. Bird AP, Wolffe AP. Methylation-induced repression--belts, braces, and chromatin. Cell. 1999;99(5):451-4.

312. Reik W, Dean W, Walter J. Epigenetic reprogramming in mammalian development. Science. 2001;293(5532):1089-93.

313. Chen T, Ueda Y, Dodge JE, Wang Z, Li E. Establishment and maintenance of genomic methylation patterns in mouse embryonic stem cells by Dnmt3a and Dnmt3b. Mol Cell Biol. 2003;23(16):5594-605.

314. Jaenisch R. DNA methylation and imprinting: why bother? Trends Genet. 1997;13(8):323-9.

- 278 -

315. Saxonov S, Berg P, Brutlag DL. A genome-wide analysis of CpG dinucleotides in the human genome distinguishes two distinct classes of promoters. Proc Natl Acad Sci U S A. 2006;103(5):1412-7.

316. Zampieri M, Ciccarone F, Calabrese R, Franceschi C, Burkle A, Caiafa P. Reconfiguration of DNA methylation in aging. Mech Ageing Dev. 2015;151:60-70.

317. Deaton AM, Bird A. CpG islands and the regulation of transcription. Genes Dev. 2011;25(10):1010-22.

318. Jones PA. The DNA methylation paradox. Trends Genet. 1999;15(1):34-7.

319. Rechache NS, Wang Y, Stevenson HS, Killian JK, Edelman DC, Merino M, et al. DNA methylation profiling identifies global methylation differences and markers of adrenocortical tumors. J Clin Endocrinol Metab. 2012;97(6):E1004-13.

320. Strichman-Almashanu LZ, Lee RS, Onyango PO, Perlman E, Flam F, Frieman MB, et al. A genome-wide screen for normally methylated human CpG islands that can identify novel imprinted genes. Genome Res. 2002;12(4):543-54.

321. Jones PA, Baylin SB. The fundamental role of epigenetic events in cancer. Nat Rev Genet. 2002;3(6):415-28.

322. Antequera F. Structure, function and evolution of CpG island promoters. Cell Mol Life Sci. 2003;60(8):1647-58.

323. Bird AP. DNA methylation and the frequency of CpG in animal DNA. Nucleic Acids Res. 1980;8(7):1499-504.

324. Han L, Su B, Li WH, Zhao Z. CpG island density and its correlations with genomic features in mammalian genomes. Genome Biol. 2008;9(5):R79.

325. Kafri T, Ariel M, Brandeis M, Shemer R, Urven L, McCarrey J, et al. Developmental pattern of gene-specific DNA methylation in the mouse embryo and germ line. Genes Dev. 1992;6(5):705-14.

- 279 -

326. Monk M, Boubelik M, Lehnert S. Temporal and regional changes in DNA methylation in the embryonic, extraembryonic and germ cell lineages during mouse embryo development. Development. 1987;99(3):371-82.

327. Okano M, Bell DW, Haber DA, Li E. DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell. 1999;99(3):247- 57.

328. Brandeis M, Frank D, Keshet I, Siegfried Z, Mendelsohn M, Nemes A, et al. Sp1 elements protect a CpG island from de novo methylation. Nature. 1994;371(6496):435-8.

329. Macleod D, Charlton J, Mullins J, Bird AP. Sp1 sites in the mouse aprt gene promoter are required to prevent methylation of the CpG island. Genes Dev. 1994;8(19):2282-92.

330. Siegfried Z, Eden S, Mendelsohn M, Feng X, Tsuberi BZ, Cedar H. DNA methylation represses transcription in vivo. Nat Genet. 1999;22(2):203-6.

331. Frank D, Keshet I, Shani M, Levine A, Razin A, Cedar H. Demethylation of CpG islands in embryonic cells. Nature. 1991;351(6323):239-41.

332. Barlow DP. Genomic imprinting: a mammalian epigenetic discovery model. Annu Rev Genet. 2011;45:379-403.

333. Sharp AJ, Stathaki E, Migliavacca E, Brahmachary M, Montgomery SB, Dupre Y, et al. DNA methylation profiles of human active and inactive X chromosomes. Genome Res. 2011;21(10):1592-600.

334. Bestor TH. The host defence function of genomic methylation patterns. Novartis Found Symp. 1998;214:187-95; discussion 95-9, 228-32.

335. Meissner A. Epigenetic modifications in pluripotent and differentiated cells. Nat Biotechnol. 2010;28(10):1079-88.

336. Reik W. Stability and flexibility of epigenetic gene regulation in mammalian development. Nature. 2007;447(7143):425-32.

337. Jjingo D, Conley AB, Yi SV, Lunyak VV, Jordan IK. On the presence and role of human gene-body DNA methylation. Oncotarget. 2012;3(4):462-74.

- 280 -

338. Maunakea AK, Nagarajan RP, Bilenky M, Ballinger TJ, D'Souza C, Fouse SD, et al. Conserved role of intragenic DNA methylation in regulating alternative promoters. Nature. 2010;466(7303):253-7.

339. Connor CM, Akbarian S. DNA methylation changes in schizophrenia and bipolar disorder. Epigenetics. 2008;3(2):55-8.

340. Nishioka M, Bundo M, Kasai K, Iwamoto K. DNA methylation in schizophrenia: progress and challenges of epigenetic studies. Genome Med. 2012;4(12):96.

341. Alexeeff SE, Baccarelli AA, Halonen J, Coull BA, Wright RO, Tarantini L, et al. Association between blood pressure and DNA methylation of retrotransposons and pro- inflammatory genes. Int J Epidemiol. 2013;42(1):270-80.

342. Dayeh T, Volkov P, Salo S, Hall E, Nilsson E, Olsson AH, et al. Genome-wide DNA methylation analysis of human pancreatic islets from type 2 diabetic and non-diabetic donors identifies candidate genes that influence insulin secretion. PLoS Genet. 2014;10(3):e1004160.

343. Xu X, Zeng H, Xiao D, Zhou H, Liu Z. Genome wide association study of obesity. Zhong Nan Da Xue Xue Bao Yi Xue Ban. 2013;38(1):95-100.

344. Baylin SB, Jones PA. A decade of exploring the cancer epigenome - biological and translational implications. Nat Rev Cancer. 2011;11(10):726-34.

345. Feinberg AP. Phenotypic plasticity and the epigenetics of human disease. Nature. 2007;447(7143):433-40.

346. Jones PA, Baylin SB. The epigenomics of cancer. Cell. 2007;128(4):683-92.

347. Heichman KA, Warren JD. DNA methylation biomarkers and their utility for solid cancer diagnostics. Clin Chem Lab Med. 2012;50(10):1707-21.

348. Baylin SB, Hoppener JW, de Bustros A, Steenbergh PH, Lips CJ, Nelkin BD. DNA methylation patterns of the calcitonin gene in human lung cancers and lymphomas. Cancer research. 1986;46(6):2917-22.

- 281 -

349. Feinberg AP, Vogelstein B. Hypomethylation distinguishes genes of some human cancers from their normal counterparts. Nature. 1983;301(5895):89-92.

350. Konkel MK, Batzer MA. A mobile threat to genome stability: The impact of non-LTR retrotransposons upon the human genome. Semin Cancer Biol. 2010;20(4):211-21.

351. Kulis M, Esteller M. DNA methylation and cancer. Adv Genet. 2010;70:27-56.

352. Baylin SB, Herman JG, Graff JR, Vertino PM, Issa JP. Alterations in DNA methylation: a fundamental aspect of neoplasia. Adv Cancer Res. 1998;72:141-96.

353. Esteller M. CpG island hypermethylation and tumor suppressor genes: a booming present, a brighter future. Oncogene. 2002;21(35):5427-40.

354. Vogelstein B, Papadopoulos N, Velculescu VE, Zhou S, Diaz LA, Jr., Kinzler KW. Cancer genome landscapes. Science. 2013;339(6127):1546-58.

355. Doi A, Park IH, Wen B, Murakami P, Aryee MJ, Irizarry R, et al. Differential methylation of tissue- and cancer-specific CpG island shores distinguishes human induced pluripotent stem cells, embryonic stem cells and fibroblasts. Nat Genet. 2009;41(12):1350- 3.

356. Irizarry RA, Ladd-Acosta C, Wen B, Wu Z, Montano C, Onyango P, et al. The human colon cancer methylome shows similar hypo- and hypermethylation at conserved tissue- specific CpG island shores. Nat Genet. 2009;41(2):178-86.

357. Lin IH, Chen DT, Chang YF, Lee YL, Su CH, Cheng C, et al. Hierarchical clustering of breast cancer methylomes revealed differentially methylated and expressed breast cancer genes. PLoS One. 2015;10(2):e0118453.

358. Jaenisch R, Harbers K, Jahner D, Stewart C, Stuhlmann H. DNA methylation, retroviruses, and embryogenesis. J Cell Biochem. 1982;20(4):331-6.

359. Bird A. DNA methylation patterns and epigenetic memory. Genes Dev. 2002;16(1):6- 21.

360. Issa JP. CpG-island methylation in aging and cancer. Curr Top Microbiol Immunol. 2000;249:101-18.

- 282 -

361. Baylin SB, Herman JG. DNA hypermethylation in tumorigenesis: epigenetics joins genetics. Trends Genet. 2000;16(4):168-74.

362. Fraga MF, Ballestar E, Paz MF, Ropero S, Setien F, Ballestar ML, et al. Epigenetic differences arise during the lifetime of monozygotic twins. Proc Natl Acad Sci U S A. 2005;102(30):10604-9.

363. Feil R, Fraga MF. Epigenetics and the environment: emerging patterns and implications. Nat Rev Genet. 2012;13(2):97-109.

364. Poulsen P, Esteller M, Vaag A, Fraga MF. The epigenetic basis of twin discordance in age-related diseases. Pediatr Res. 2007;61(5 Pt 2):38R-42R.

365. You JS, Jones PA. Cancer genetics and epigenetics: two sides of the same coin? Cancer Cell. 2012;22(1):9-20.

366. Cedar H, Bergman Y. Linking DNA methylation and histone modification: patterns and paradigms. Nat Rev Genet. 2009;10(5):295-304.

367. Zhang Y, Reinberg D. Transcription regulation by histone methylation: interplay between different covalent modifications of the core histone tails. Genes Dev. 2001;15(18):2343-60.

368. Morales V, Richard-Foy H. Role of histone N-terminal tails and their acetylation in nucleosome dynamics. Mol Cell Biol. 2000;20(19):7230-7.

369. Mersfelder EL, Parthun MR. The tale beyond the tail: histone core domain modifications and the regulation of chromatin structure. Nucleic Acids Res. 2006;34(9):2653-62.

370. Xu WS, Parmigiani RB, Marks PA. Histone deacetylase inhibitors: molecular mechanisms of action. Oncogene. 2007;26(37):5541-52.

371. Cao R, Wang L, Wang H, Xia L, Erdjument-Bromage H, Tempst P, et al. Role of histone H3 lysine 27 methylation in Polycomb-group silencing. Science. 2002;298(5595):1039-43.

- 283 -

372. Dong X, Weng Z. The correlation between histone modifications and gene expression. Epigenomics. 2013;5(2):113-6.

373. Turner BM. Cellular memory and the histone code. Cell. 2002;111(3):285-91.

374. Geiman TM, Robertson KD. Chromatin remodeling, histone modifications, and DNA methylation-how does it all fit together? J Cell Biochem. 2002;87(2):117-25.

375. Soppe WJ, Jasencakova Z, Houben A, Kakutani T, Meister A, Huang MS, et al. DNA methylation controls histone H3 lysine 9 methylation and heterochromatin assembly in Arabidopsis. EMBO J. 2002;21(23):6549-59.

376. Sawan C, Herceg Z. Histone modifications and cancer. Adv Genet. 2010;70:57-85.

377. Bannister AJ, Kouzarides T. Regulation of chromatin by histone modifications. Cell Res. 2011;21(3):381-95.

378. Wahid F, Shehzad A, Khan T, Kim YY. MicroRNAs: synthesis, mechanism, function, and recent clinical trials. Biochim Biophys Acta. 2010;1803(11):1231-43.

379. Pillai RS. MicroRNA function: multiple mechanisms for a tiny RNA? RNA. 2005;11(12):1753-61.

380. Davis BN, Hata A. Regulation of MicroRNA Biogenesis: A miRiad of mechanisms. Cell Commun Signal. 2009;7:18.

381. Croce CM, Calin GA. miRNAs, cancer, and stem cell division. Cell. 2005;122(1):6-7.

382. Shenouda SK, Alahari SK. MicroRNA function in cancer: oncogene or a tumor suppressor? Cancer Metastasis Rev. 2009;28(3-4):369-78.

383. Wang W, Luo YP. MicroRNAs in breast cancer: oncogene and tumor suppressors with clinical potential. J Zhejiang Univ Sci B. 2015;16(1):18-31.

384. Zhang B, Pan X, Cobb GP, Anderson TA. microRNAs as oncogenes and tumor suppressors. Dev Biol. 2007;302(1):1-12.

385. Suzuki HI, Katsura A, Matsuyama H, Miyazono K. MicroRNA regulons in tumor microenvironment. Oncogene. 2015;34(24):3085-94.

- 284 -

386. Fabbri M, Calin GA. Epigenetics and miRNAs in human cancer. Adv Genet. 2010;70:87-99.

387. Sato F, Tsuchiya S, Meltzer SJ, Shimizu K. MicroRNAs and epigenetics. FEBS J. 2011;278(10):1598-609.

388. Brena RM, Huang TH, Plass C. Quantitative assessment of DNA methylation: Potential applications for disease diagnosis, classification, and prognosis in clinical settings. J Mol Med (Berl). 2006;84(5):365-77.

389. Callinan PA, Feinberg AP. The emerging science of epigenomics. Human molecular genetics. 2006;15 Spec No 1:R95-101.

390. Laird PW. The power and the promise of DNA methylation markers. Nat Rev Cancer. 2003;3(4):253-66.

391. Lieb JD, Beck S, Bulyk ML, Farnham P, Hattori N, Henikoff S, et al. Applying whole- genome studies of epigenetic regulation to study human disease. Cytogenet Genome Res. 2006;114(1):1-15.

392. Ushijima T. Detection and interpretation of altered methylation patterns in cancer cells. Nat Rev Cancer. 2005;5(3):223-31.

393. Kurdyukov S, Bullock M. DNA Methylation Analysis: Choosing the Right Method. Biology (Basel). 2016;5(1).

394. Shen L, Waterland RA. Methods of DNA methylation analysis. Curr Opin Clin Nutr Metab Care. 2007;10(5):576-81.

395. Sant KE, Nahar MS, Dolinoy DC. DNA methylation screening and analysis. Methods Mol Biol. 2012;889:385-406.

396. Taylor KH, Kramer RS, Davis JW, Guo J, Duff DJ, Xu D, et al. Ultradeep bisulfite sequencing analysis of DNA methylation patterns in multiple gene promoters by 454 sequencing. Cancer research. 2007;67(18):8511-8.

397. Tost J, Gut IG. DNA methylation analysis by pyrosequencing. Nat Protoc. 2007;2(9):2265-75.

- 285 -

398. Bibikova M, Barnes B, Tsan C, Ho V, Klotzle B, Le JM, et al. High density DNA methylation array with single CpG site resolution. Genomics. 2011;98(4):288-95.

399. Teh AL, Pan H, Lin X, Lim YI, Patro CP, Cheong CY, et al. Comparison of Methyl- capture Sequencing vs. Infinium 450K methylation array for methylome analysis in clinical samples. Epigenetics. 2016;11(1):36-48.

400. Maksimovic J, Gordon L, Oshlack A. SWAN: Subset-quantile within array normalization for illumina infinium HumanMethylation450 BeadChips. Genome Biol. 2012;13(6):R44.

401. Clark SJ, Harrison J, Paul CL, Frommer M. High sensitivity mapping of methylated cytosines. Nucleic Acids Res. 1994;22(15):2990-7.

402. Wang RY, Gehrke CW, Ehrlich M. Comparison of bisulfite modification of 5- methyldeoxycytidine and deoxycytidine residues. Nucleic Acids Res. 1980;8(20):4777-90.

403. Maksimovic J, Phipson B, Oshlack A. A cross-package Bioconductor workflow for analysing methylation array data. F1000Res. 2016;5:1281.

404. Morris TJ, Beck S. Analysis pipelines and packages for Infinium HumanMethylation450 BeadChip (450k) data. Methods. 2015;72:3-8.

405. Chen Y, Breeze CE, Zhen S, Beck S, Teschendorff AE. Tissue-independent and tissue-specific patterns of DNA methylation alteration in cancer. Epigenetics Chromatin. 2016;9:10.

406. Steemers FJ, Chang W, Lee G, Barker DL, Shen R, Gunderson KL. Whole-genome genotyping with the single-base extension assay. Nat Methods. 2006;3(1):31-3.

407. Wright ML, Dozmorov MG, Wolen AR, Jackson-Cook C, Starkweather AR, Lyon DE, et al. Establishing an analytic pipeline for genome-wide DNA methylation. Clin Epigenetics. 2016;8:45.

408. Bibikova M, Le J, Barnes B, Saedinia-Melnyk S, Zhou L, Shen R, et al. Genome- wide DNA methylation profiling using Infinium(R) assay. Epigenomics. 2009;1(1):177-200.

- 286 -

409. Du P, Zhang X, Huang CC, Jafari N, Kibbe WA, Hou L, et al. Comparison of Beta- value and M-value methods for quantifying methylation levels by microarray analysis. BMC Bioinformatics. 2010;11:587.

410. Chen YA, Lemire M, Choufani S, Butcher DT, Grafodatskaya D, Zanke BW, et al. Discovery of cross-reactive probes and polymorphic CpGs in the Illumina Infinium HumanMethylation450 microarray. Epigenetics. 2013;8(2):203-9.

411. Dedeurwaerder S, Defrance M, Calonne E, Denis H, Sotiriou C, Fuks F. Evaluation of the Infinium Methylation 450K technology. Epigenomics. 2011;3(6):771-84.

412. Thirlwell C, Eymard M, Feber A, Teschendorff A, Pearce K, Lechner M, et al. Genome-wide DNA methylation analysis of archival formalin-fixed paraffin-embedded tissue using the Illumina Infinium HumanMethylation27 BeadChip. Methods. 2010;52(3):248-54.

413. Moran S, Vizoso M, Martinez-Cardus A, Gomez A, Matias-Guiu X, Chiavenna SM, et al. Validation of DNA methylation profiling in formalin-fixed paraffin-embedded samples using the Infinium HumanMethylation450 Microarray. Epigenetics. 2014;9(6):829-33.

414. Moran S, Arribas C, Esteller M. Validation of a DNA methylation microarray for 850,000 CpG sites of the human genome enriched in enhancer sequences. Epigenomics. 2016;8(3):389-99.

415. Esteller M. Aberrant DNA methylation as a cancer-inducing mechanism. Annu Rev Pharmacol Toxicol. 2005;45:629-56.

416. Jovanovic J, Ronneberg JA, Tost J, Kristensen V. The epigenetics of breast cancer. Mol Oncol. 2010;4(3):242-54.

417. Karsli-Ceppioglu S, Dagdemir A, Judes G, Ngollo M, Penault-Llorca F, Pajon A, et al. Epigenetic mechanisms of breast cancer: an update of the current knowledge. Epigenomics. 2014;6(6):651-64.

418. Locke WJ, Clark SJ. Epigenome remodelling in breast cancer: insights from an early in vitro model of carcinogenesis. Breast cancer research : BCR. 2012;14(6):215.

- 287 -

419. Falahi F, van Kruchten M, Martinet N, Hospers GA, Rots MG. Current and upcoming approaches to exploit the reversibility of epigenetic mutations in breast cancer. Breast cancer research : BCR. 2014;16(4):412.

420. Haffner MC, Chaux A, Meeker AK, Esopi DM, Gerber J, Pellakuru LG, et al. Global 5-hydroxymethylcytosine content is significantly reduced in tissue stem/progenitor cell compartments and in human cancers. Oncotarget. 2011;2(8):627-37.

421. Bertolo C, Guerrero D, Vicente F, Cordoba A, Esteller M, Ropero S, et al. Differences and molecular immunohistochemical parameters in the subtypes of infiltrating ductal breast cancer. American journal of clinical pathology. 2008;130(3):414-24.

422. Fiegl H, Millinger S, Goebel G, Muller-Holzner E, Marth C, Laird PW, et al. Breast cancer DNA methylation profiles in cancer cells and tumor stroma: association with HER- 2/neu status in primary breast cancer. Cancer research. 2006;66(1):29-33.

423. Lindqvist BM, Wingren S, Motlagh PB, Nilsson TK. Whole genome DNA methylation signature of HER2-positive breast cancer. Epigenetics. 2014;9(8):1149-62.

424. Shinozaki M, Hoon DS, Giuliano AE, Hansen NM, Wang HJ, Turner R, et al. Distinct hypermethylation profile of primary breast cancer is associated with sentinel lymph node metastasis. Clin Cancer Res. 2005;11(6):2156-62.

425. Sunami E, Shinozaki M, Sim MS, Nguyen SL, Vu AT, Giuliano AE, et al. Estrogen receptor and HER2/neu status affect epigenetic differences of tumor-related genes in primary breast tumors. Breast cancer research : BCR. 2008;10(3):R46.

426. Terada K, Okochi-Takada E, Akashi-Tanaka S, Miyamoto K, Taniyama K, Tsuda H, et al. Association between frequent CpG island methylation and HER2 amplification in human breast cancers. Carcinogenesis. 2009;30(3):466-71.

427. Yamaguchi T, Mukai H, Yamashita S, Fujii S, Ushijima T. Comprehensive DNA Methylation and Extensive Mutation Analyses of HER2-Positive Breast Cancer. Oncology. 2015;88(6):377-84.

428. Zuo T, Tycko B, Liu TM, Lin JJ, Huang TH. Methods in DNA methylation profiling. Epigenomics. 2009;1(2):331-45.

- 288 -

429. Fang F, Turcan S, Rimner A, Kaufman A, Giri D, Morris LG, et al. Breast cancer methylomes establish an epigenomic foundation for metastasis. Sci Transl Med. 2011;3(75):75ra25.

430. Fleischer T, Frigessi A, Johnson KC, Edvardsen H, Touleimat N, Klajic J, et al. Genome-wide DNA methylation profiles in progression to in situ and invasive carcinoma of the breast with impact on gene transcription and prognosis. Genome Biol. 2014;15(8):435.

431. Hill VK, Ricketts C, Bieche I, Vacher S, Gentle D, Lewis C, et al. Genome-wide DNA methylation profiling of CpG islands in breast cancer identifies novel genes associated with tumorigenicity. Cancer research. 2011;71(8):2988-99.

432. Jadhav RR, Ye Z, Huang RL, Liu J, Hsu PY, Huang YW, et al. Genome-wide DNA methylation analysis reveals estrogen-mediated epigenetic repression of metallothionein-1 gene cluster in breast cancer. Clin Epigenetics. 2015;7:13.

433. Jeschke J, Van Neste L, Glockner SC, Dhir M, Calmon MF, Deregowski V, et al. Biomarkers for detection and prognosis of breast cancer identified by a functional hypermethylome screen. Epigenetics. 2012;7(7):701-9.

434. Brinkman AB, Simmer F, Ma K, Kaan A, Zhu J, Stunnenberg HG. Whole-genome DNA methylation profiling using MethylCap-seq. Methods. 2010;52(3):232-6.

435. He DX, Gu F, Gao F, Hao JJ, Gong D, Gu XT, et al. Genome-wide profiles of methylation, microRNAs, and gene expression in chemoresistant breast cancer. Sci Rep. 2016;6:24706.

436. Klajic J, Busato F, Edvardsen H, Touleimat N, Fleischer T, Bukholm I, et al. DNA methylation status of key cell-cycle regulators such as CDKNA2/p16 and CCNA1 correlates with treatment response to doxorubicin and 5-fluorouracil in locally advanced breast tumors. Clin Cancer Res. 2014;20(24):6357-66.

437. Ye XM, Zhu HY, Bai WD, Wang T, Wang L, Chen Y, et al. Epigenetic silencing of miR-375 induces trastuzumab resistance in HER2-positive breast cancer by targeting IGF1R. BMC Cancer. 2014;14:134.

- 289 -

438. Pouliot MC, Labrie Y, Diorio C, Durocher F. The Role of Methylation in Breast Cancer Susceptibility and Treatment. Anticancer Res. 2015;35(9):4569-74.

439. Nervi C, De Marinis E, Codacci-Pisanelli G. Epigenetic treatment of solid tumours: a review of clinical trials. Clin Epigenetics. 2015;7:127.

440. Gnyszka A, Jastrzebski Z, Flis S. DNA methyltransferase inhibitors and their emerging role in epigenetic therapy of cancer. Anticancer Res. 2013;33(8):2989-96.

441. Goffin J, Eisenhauer E. DNA methyltransferase inhibitors-state of the art. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2002;13(11):1699-716.

442. Santi DV, Norment A, Garrett CE. Covalent bond formation between a DNA-cytosine methyltransferase and DNA containing 5-azacytosine. Proc Natl Acad Sci U S A. 1984;81(22):6993-7.

443. Lyko F, Brown R. DNA methyltransferase inhibitors and the development of epigenetic cancer therapies. J Natl Cancer Inst. 2005;97(20):1498-506.

444. Mann BS, Johnson JR, Cohen MH, Justice R, Pazdur R. FDA approval summary: vorinostat for treatment of advanced primary cutaneous T-cell lymphoma. The oncologist. 2007;12(10):1247-52.

445. Richon VM. Cancer biology: mechanism of antitumour action of vorinostat (suberoylanilide hydroxamic acid), a novel histone deacetylase inhibitor. Br J Cancer. 2006;95 (Suppl 1):S2-S6.

446. VanderMolen KM, McCulloch W, Pearce CJ, Oberlies NH. Romidepsin (Istodax, NSC 630176, FR901228, FK228, depsipeptide): a natural product recently approved for cutaneous T-cell lymphoma. J Antibiot (Tokyo). 2011;64(8):525-31.

447. Kim HJ, Bae SC. Histone deacetylase inhibitors: molecular mechanisms of action and clinical trials as anti-cancer drugs. Am J Transl Res. 2011;3(2):166-79.

448. Lustberg MB, Ramaswamy B. Epigenetic Therapy in Breast Cancer. Curr Breast Cancer Rep. 2011;3(1):34-43.

- 290 -

449. Tu Y, Hershman DL, Bhalla K, Fiskus W, Pellegrino CM, Andreopoulou E, et al. A phase I-II study of the histone deacetylase inhibitor vorinostat plus sequential weekly paclitaxel and doxorubicin-cyclophosphamide in locally advanced breast cancer. Breast cancer research and treatment. 2014;146(1):145-52.

450. Conte P, Campone M, Pronzato P, Amadori D, Frank R, Schuetz F, et al. Phase I trial of panobinostat (LBH589) in combination with trastuzumab in pretreated HER2-positive metastatic breast cancer (mBC): Preliminary safety and tolerability results. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2009;27(15_suppl):1081.

451. Munster PN, Lacevic M, Thomas S, Christian C, Ismail-Khan R, Melisko M, et al. Phase II trial of the histone deacetylase inhibitor, vorinostat, to restore hormone sensitivity to the antiestrogen tamoxifen in patients with advanced breast cancer who progressed on prior hormone therapy. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2009;27(15_suppl):1075.

452. Ramaswamy B, Fiskus W, Cohen B, Pellegrino C, Hershman DL, Chuang E, et al. Phase I-II study of vorinostat plus paclitaxel and bevacizumab in metastatic breast cancer: evidence for vorinostat-induced tubulin acetylation and Hsp90 inhibition in vivo. Breast cancer research and treatment. 2012;132(3):1063-72.

453. Santini D, Vincenzi B, Adamo V, Addeo R, Fusco V, Russo A, et al. Cigarette smoking habit does not reduce the benefit from first line trastuzumab-based treatment in advanced breast cancer patients. Oncol Rep. 2011;25(6):1545-8.

454. Fons G, Hasibuan SM, van der Velden J, ten Kate FJ. Validation of tissue microarray technology in endometrioid cancer of the endometrium. Journal of clinical pathology. 2007;60(5):500-3.

455. Hoos A, Urist MJ, Stojadinovic A, Mastorides S, Dudas ME, Leung DH, et al. Validation of tissue microarrays for immunohistochemical profiling of cancer specimens using the example of human fibroblastic tumors. The American journal of pathology. 2001;158(4):1245-51.

456. Egervari K, Szollosi Z, Nemes Z. Tissue microarray technology in breast cancer HER2 diagnostics. Pathology, research and practice. 2007;203(3):169-77.

- 291 -

457. Hwang CC, Pintye M, Chang LC, Chen HY, Yeh KY, Chein HP, et al. Dual-colour chromogenic in-situ hybridization is a potential alternative to fluorescence in-situ hybridization in HER2 testing. Histopathology. 2011;59(5):984-92.

458. Lee Y, Ryu Y, Jeong H, Chang H, Kim Y, Kim A. Effectiveness of silver-enhanced in situ hybridization for evaluating HER2 gene status in invasive breast carcinoma: a comparative study. Archives of medical research. 2012;43(2):139-44.

459. Park K, Han S, Kim JY, Kim HJ, Kwon JE, Gwak G. Silver-Enhanced In Situ Hybridization as an Alternative to Fluorescence In Situ Hybridization for Assaying HER2 Amplification in Clinical Breast Cancer. Journal of breast cancer. 2011;14(4):276-82.

460. Saez A, Andreu FJ, Segui MA, Bare ML, Fernandez S, Dinares C, et al. HER-2 gene amplification by chromogenic in situ hybridisation (CISH) compared with fluorescence in situ hybridisation (FISH) in breast cancer-A study of two hundred cases. Breast. 2006;15(4):519- 27.

461. Schiavon BN, Jasani B, de Brot L, Vassallo J, Damascena A, Cirullo-Neto J, et al. Evaluation of reliability of FISH versus brightfield dual-probe in situ hybridization (BDISH) for frontline assessment of HER2 status in breast cancer samples in a community setting: influence of poor tissue preservation. The American journal of surgical pathology. 2012;36(10):1489-96.

462. Middleton LP, Price KM, Puig P, Heydon LJ, Tarco E, Sneige N, et al. Implementation of American Society of Clinical Oncology/College of American Pathologists HER2 Guideline Recommendations in a tertiary care facility increases HER2 immunohistochemistry and fluorescence in situ hybridization concordance and decreases the number of inconclusive cases. Archives of pathology & laboratory medicine. 2009;133(5):775-80.

463. Yildiz-Aktas IZ, Dabbs DJ, Bhargava R. The effect of cold ischemic time on the immunohistochemical evaluation of estrogen receptor, progesterone receptor, and HER2 expression in invasive breast carcinoma. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2012;25(8):1098-105.

464. Lebeau A, Deimling D, Kaltz C, Sendelhofert A, Iff A, Luthardt B, et al. Her-2/neu analysis in archival tissue samples of human breast cancer: comparison of

- 292 -

immunohistochemistry and fluorescence in situ hybridization. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2001;19(2):354-63.

465. Tubbs RR, Pettay JD, Roche PC, Stoler MH, Jenkins RB, Grogan TM. Discrepancies in clinical laboratory testing of eligibility for trastuzumab therapy: apparent immunohistochemical false-positives do not get the message. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2001;19(10):2714-21.

466. Tubbs RR, Hicks DG, Cook J, Downs-Kelly E, Pettay J, Hartke MB, et al. Fluorescence in situ hybridization (FISH) as primary methodology for the assessment of HER2 Status in adenocarcinoma of the breast: a single institution experience. Diagnostic molecular pathology : the American journal of surgical pathology, part B. 2007;16(4):207- 10.

467. Lal P, Salazar PA, Ladanyi M, Chen B. Impact of polysomy 17 on HER-2/neu immunohistochemistry in breast carcinomas without HER-2/neu gene amplification. J Mol Diagn. 2003;5(3):155-9.

468. Hyun CL, Lee HE, Kim KS, Kim SW, Kim JH, Choe G, et al. The effect of chromosome 17 polysomy on HER-2/neu status in breast cancer. Journal of clinical pathology. 2008;61(3):317-21.

469. Ma Y, Lespagnard L, Durbecq V, Paesmans M, Desmedt C, Gomez-Galdon M, et al. Polysomy 17 in HER-2/neu status elaboration in breast cancer: effect on daily practice. Clin Cancer Res. 2005;11(12):4393-9.

470. Apple SK, Lowe AC, Rao PN, Shintaku IP, Moatamed NA. Comparison of fluorescent in situ hybridization HER-2/neu results on core needle biopsy and excisional biopsy in primary breast cancer. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2009;22(9):1151-9.

471. Buggi F, Folli S, Curcio A, Casadei-Giunchi D, Rocca A, Pietri E, et al. Multicentric/multifocal breast cancer with a single histotype: is the biological characterization of all individual foci justified? Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2012;23(8):2042-6.

- 293 -

472. Choi Y, Kim EJ, Seol H, Lee HE, Jang MJ, Kim SM, et al. The hormone receptor, human epidermal growth factor receptor 2, and molecular subtype status of individual tumor foci in multifocal/multicentric invasive ductal carcinoma of breast. Human pathology. 2012;43(1):48-55.

473. D'Alfonso T, Liu YF, Monni S, Rosen PP, Shin SJ. Accurately assessing her-2/neu status in needle core biopsies of breast cancer patients in the era of neoadjuvant therapy: emerging questions and considerations addressed. The American journal of surgical pathology. 2010;34(4):575-81.

474. Bethune GC, Veldhuijzen van Zanten D, MacIntosh RF, Rayson D, Younis T, Thompson K, et al. Impact of the 2013 American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 (HER2) testing of invasive breast carcinoma: a focus on tumours assessed as 'equivocal' for HER2 gene amplification by fluorescence in-situ hybridization. Histopathology. 2015.

475. Benz CC, Fedele V, Xu F, Ylstra B, Ginzinger D, Yu M, et al. Altered promoter usage characterizes monoallelic transcription arising with ERBB2 amplification in human breast cancers. Genes Chromosomes Cancer. 2006;45(11):983-94.

476. Cresti N, Lee J, Rourke E, Televantou D, Jamieson D, Verrill M, et al. Genetic variants in the HER2 gene: Influence on HER2 overexpression and loss of heterozygosity in breast cancer. Eur J Cancer. 2016;55:27-37.

477. Puputti M, Sihto H, Isola J, Butzow R, Joensuu H, Nupponen NN. Allelic imbalance of HER2 variant in sporadic breast and ovarian cancer. Cancer Genet Cytogenet. 2006;167(1):32-8.

478. Su Y, Jiang Y, Sun S, Yin H, Shan M, Tao W, et al. Effects of HER2 genetic polymorphisms on its protein expression in breast cancer. Cancer Epidemiol. 2015;39(6):1123-7.

479. Gianni L, Dafni U, Gelber RD, Azambuja E, Muehlbauer S, Goldhirsch A, et al. Treatment with trastuzumab for 1 year after adjuvant chemotherapy in patients with HER2- positive early breast cancer: a 4-year follow-up of a randomised controlled trial. The Lancet Oncology. 2011;12(3):236-44.

- 294 -

480. Furrer D, Lemieux J, Cote MA, Provencher L, Laflamme C, Barabe F, et al. Evaluation of human epidermal growth factor receptor 2 (HER2) single nucleotide polymorphisms (SNPs) in normal and breast tumor tissues and their link with breast cancer prognostic factors. Breast. 2016;30:191-6.

481. Bock C. Analysing and interpreting DNA methylation data. Nat Rev Genet. 2012;13(10):705-19.

482. Baumgartel K, Zelazny J, Timcheck T, Snyder C, Bell M, Conley YP. Molecular genomic research designs. Annual review of nursing research. 2011;29:1-26.

483. Liu P, Hwang JT. Quick calculation for sample size while controlling false discovery rate with application to microarray analysis. Bioinformatics. 2007;23(6):739-46.

484. Warton K, Mahon KL, Samimi G. Methylated circulating tumor DNA in blood: power in cancer prognosis and response. Endocr Relat Cancer. 2016;23(3):R157-71.

485. Dumenil TD, Wockner LF, Bettington M, McKeone DM, Klein K, Bowdler LM, et al. Genome-wide DNA methylation analysis of formalin-fixed paraffin embedded colorectal cancer tissue. Genes Chromosomes Cancer. 2014;53(7):537-48.

486. Duffy MJ, Napieralski R, Martens JW, Span PN, Spyratos F, Sweep FC, et al. Methylated genes as new cancer biomarkers. Eur J Cancer. 2009;45(3):335-46.

487. Mayer IA. Treatment of HER2-positive metastatic breast cancer following initial progression. Clinical breast cancer. 2009;9 Suppl 2:S50-7.

488. Wilks ST. Potential of overcoming resistance to HER2-targeted therapies through the PI3K/Akt/mTOR pathway. Breast. 2015;24(5):548-55.

489. Kanwal R, Gupta S. Epigenetics and cancer. J Appl Physiol (1985). 2010;109(2):598- 605.

490. Bojang P, Jr., Ramos KS. The promise and failures of epigenetic therapies for cancer treatment. Cancer Treat Rev. 2014;40(1):153-69.

- 295 -

491. D'Agostino Y, D'Aniello S. Molecular basis, applications and challenges of CRISPR/Cas9: a continuously evolving tool for genome editing. Brief Funct Genomics. 2017.

492. Flora JW, Meruva N, Huang CB, Wilkinson CT, Ballentine R, Smith DC, et al. Characterization of potential impurities and degradation products in electronic cigarette formulations and aerosols. Regul Toxicol Pharmacol. 2016;74:1-11.

493. Feyerabend C, Higenbottam T, Russell MA. Nicotine concentrations in urine and saliva of smokers and non-smokers. Br Med J (Clin Res Ed). 1982;284(6321):1002-4.

494. Thokala MR, Dorankula SP, Muddana K, Velidandla SR. Alcohol saliva strip test. J Clin Diagn Res. 2014;8(3):307-8.

495. Montemurro F, Di Cosimo S, Arpino G. Human epidermal growth factor receptor 2 (HER2)-positive and hormone receptor-positive breast cancer: new insights into molecular interactions and clinical implications. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2013;24(11):2715-24.

496. Prat A, Baselga J. Dual human epidermal growth factor receptor 2 (HER2) blockade and hormonal therapy for the treatment of primary HER2-positive breast cancer: one more step toward chemotherapy-free therapy. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2013;31(14):1703-6.

497. Dietel M, Ellis IO, Hofler H, Kreipe H, Moch H, Dankof A, et al. Comparison of automated silver enhanced in situ hybridisation (SISH) and fluorescence ISH (FISH) for the validation of HER2 gene status in breast carcinoma according to the guidelines of the American Society of Clinical Oncology and the College of American Pathologists. Virchows Archiv : an international journal of pathology. 2007;451(1):19-25.

498. Capizzi E, Gruppioni E, Grigioni AD, Gabusi E, Grassigli A, Grigioni WF, et al. Real time RT-PCR approach for the evaluation of ERBB2 overexpression in breast cancer archival samples: a comparative study with FISH, SISH, and immunohistochemistry. Diagnostic molecular pathology : the American journal of surgical pathology, part B. 2008;17(4):220-6.

- 296 -

499. Francis GD, Jones MA, Beadle GF, Stein SR. Bright-field in situ hybridization for HER2 gene amplification in breast cancer using tissue microarrays: correlation between chromogenic (CISH) and automated silver-enhanced (SISH) methods with patient outcome. Diagnostic molecular pathology : the American journal of surgical pathology, part B. 2009;18(2):88-95.

500. Shousha S, Peston D, Amo-Takyi B, Morgan M, Jasani B. Evaluation of automated silver-enhanced in situ hybridization (SISH) for detection of HER2 gene amplification in breast carcinoma excision and core biopsy specimens. Histopathology. 2009;54(2):248-53.

501. Fritzsche FR, Bode PK, Moch H, Kristiansen G, Varga Z, Bode B. Determination of the Her-2/neu gene amplification status in cytologic breast cancer specimens using automated silver-enhanced in-situ hybridization (SISH). The American journal of surgical pathology. 2010;34(8):1180-5.

502. Papouchado BG, Myles J, Lloyd RV, Stoler M, Oliveira AM, Downs-Kelly E, et al. Silver in situ hybridization (SISH) for determination of HER2 gene status in breast carcinoma: comparison with FISH and assessment of interobserver reproducibility. The American journal of surgical pathology. 2010;34(6):767-76.

503. Koh YW, Lee HJ, Lee JW, Kang J, Gong G. Dual-color silver-enhanced in situ hybridization for assessing HER2 gene amplification in breast cancer. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2011;24(6):794-800.

504. Carbone A, Botti G, Gloghini A, Simone G, Truini M, Curcio MP, et al. Delineation of HER2 gene status in breast carcinoma by silver in situ hybridization is reproducible among laboratories and pathologists. J Mol Diagn. 2008;10(6):527-36.

505. Tanner M, Gancberg D, Di Leo A, Larsimont D, Rouas G, Piccart MJ, et al. Chromogenic in situ hybridization: a practical alternative for fluorescence in situ hybridization to detect HER-2/neu oncogene amplification in archival breast cancer samples. The American journal of pathology. 2000;157(5):1467-72.

506. Brugmann A, Lelkaitis G, Nielsen S, Jensen KG, Jensen V. Testing HER2 in breast cancer: a comparative study on BRISH, FISH, and IHC. Appl Immunohistochem Mol Morphol. 2011;19(3):203-11.

- 297 -

507. Kang J, G.Y. K, Lee H, Gong G. Comparison of Silver-Enhanced in situ Hybridization and Fluorescence in situ Hybridization for HER2 Gene Status in Breast Carcinomas. Journal of breast cancer. 2009;12(4):235-40.

508. Mansfield AS, Sukov WR, Eckel-Passow JE, Sakai Y, Walsh FJ, Lonzo M, et al. Comparison of fluorescence in situ hybridization (FISH) and dual-ISH (DISH) in the determination of HER2 status in breast cancer. American journal of clinical pathology. 2013;139(2):144-50.

509. Dekker TJ, Borg ST, Hooijer GK, Meijer SL, Wesseling J, Boers JE, et al. Determining sensitivity and specificity of HER2 testing in breast cancer using a tissue microarray approach. Breast cancer research : BCR. 2012;14(3):R93.

510. Jang JS, Jang EJ, Park J. HER2 expression in breast cancer: Comparisons of fluorescence in situ hybridization and silver in situ hybridization, and immunohistochemical staining using monoclonal antibody and polyclonal antibody. Basic and Applied Pathology. 2010;3:115-20.

511. Bartlett JM, Campbell FM, Ibrahim M, Wencyk P, Ellis I, Kay E, et al. Chromogenic in situ hybridization: a multicenter study comparing silver in situ hybridization with FISH. American journal of clinical pathology. 2009;132(4):514-20.

512. Ellis IO, Bartlett J, Dowsett M, Humphreys S, Jasani B, Miller K, et al. Best Practice No 176: Updated recommendations for HER2 testing in the UK. Journal of clinical pathology. 2004;57(3):233-7.

513. Unal B, Karaveli FS, Pestereli HE, Erdogan G. Determination of HER2 gene amplification in breast cancer using dual-color silver enhanced in situ hybridization (dc- SISH) and comparison with fluorescence ISH (FISH). Asian Pac J Cancer Prev. 2013;14(10):6131-4.

514. Jacquemier J, Spyratos F, Esterni B, Mozziconacci MJ, Antoine M, Arnould L, et al. SISH/CISH or qPCR as alternative techniques to FISH for determination of HER2 amplification status on breast tumors core needle biopsies: a multicenter experience based on 840 cases. BMC Cancer. 2013;13:351.

- 298 -

Annexe I

Table S1. Concordance of HER2 status : IHC vs. SISH, FISH vs. SISH, CISH vs. SISH

Study Cohort and selection of cases Methods of Interpretation criteria Overall Comment analyses and concordance kit/antibody used Dietel et al., One hundred and ten FISH (PathVysion); FISH and SISH: 2007 FISH – SISH: Discrepancies 2007 consecutive breast cancer Dual-color SISH ASCO/CAP scoring 96.0%, k were mostly (497) specimens diagnosed in 2005 (INFORM HER2 criteria. value = 0.754. seen in tumors and 2006 at the DNA, testing for the showing Interdisciplinary Breast Cancer HER2 gene and intratumoral Centre of the Charité-University chromosome 17 heterogeneity of Hospital Berlin (Berlin, were performed on HER2 gene Germany). sequential sections). amplification.

Capizzi et Eighty-three breast cancer FISH (PathVysion); FISH and SISH: 2007 FISH – SISH: al., 2008 specimens were selected from Dual-color SISH ASCO/CAP scoring 87.0%. (498) cases diagnosed between (INFORM HER2 criteria. January 2001 and January DNA). 2005 at the Pathology Unit of the University Hospital Bologna (Bologna, Italy). Cases were selected according to HER2 expression (assessed by HercepTest) to obtain a series of tumors from all IHC categories.

- 299 -

Francis et Five hundred and eighty-nine Single-color CISH Single-color CISH and Single-color Discordant al., 2009 breast cancer patients selected (Invitrogen); single-color SISH: CISH – cases between (499) from a database of cases Single-color mean HER2 gene copy single-color single-color diagnosed between 1990 and (INFORM HER2 number < 2.5: non- SISH: 96.0% CISH/SISH and 2002 at Pathology Queensland DNA) and dual-color amplified, mean HER2 dual-color SISH of the Princess Alexandra SISH (INFORM gene copy number > Single-color showed Hospital (Queensland, HER2 DNA, testing 4.0 and < 6.0: CISH – dual- polysomy of Australia) (selection criteria not for the HER2 gene equivocal; mean HER2 color SISH: chromosome specified). and chromosome 17 gene copy number > 95.6%. 17. were performed on 6.0 and <10.0: low- sequential sections) level amplification, mean HER2 gene copy HER2 status was number ≥10.0: high- determined on TMA level amplification slides (0.6 mm cores, two cores per Dual-color SISH: 2007 patient). ASCO/CAP scoring criteria. Shousha et Sixty-five breast cancer FISH (Ventana); FISH and SISH: 2007 FISH – SISH: al., 2009 specimens were selected on Dual-color SISH ASCO/CAP scoring 94.0%. (500) the basis of their IHC staining (INFORM HER2 criteria. to represent all IHC scoring DNA). categories. Cases were diagnosed at the Cardiff University School of Medicine (Cardiff, UK) (timeframe not indicated).

- 300 -

Fritzsche et Seventy-one breast cancer FISH (PathVysion); FISH and SISH: 2007 FISH – SISH: al., 2010 specimens with known FISH Dual-color SISH ASCO/CAP scoring 92.3%. (501) status diagnosed at the (INFORM HER2 criteria. Division of Cytology of the DNA). Institute of Surgical Pathology of the University Hospital Zurich (Zurich, Switzerland): forty- eight consecutive specimens diagnosed in 2009 and twenty- three HER2 amplified specimens selected from a cohort of breast cancer patients diagnosed between 2006 and 2008.

Papouchado Two hundred and twenty-eight FISH (PathVysion); FISH and SISH: two FISH – SISH: et al., 2010 breast cancer biopsies (archival SISH (INFORM different scoring 96.6% when (502) needle core biopsies or HER2 DNA). systems were used for FDA- excisional biopsies) diagnosed analysis: approved at the Department of Anatomic 1). FDA-approved scoring Pathology at the Cleveland scoring criteria criteria were Clinic Foundation (Cleveland, (amplification when used; OH) (timeframe and selection mean HER2 gene copy 98.9% when criteria not specified). number > 2.0); 2). 2007 2007 ASCO/CAP scoring ASCO/CAP criteria. guidelines were used.

- 301 -

Koh et al., One hundred and forty-six FISH (PathVysion); FISH, single- and dual- FISH – dual- 2011 invasive breast cancer Single-color color SISH: 2007 color SISH: (503) specimens diagnosed at the SISH (INFORM ASCO/CAP scoring 97.0%, k Asan Medical Center (Seoul, HER2 DNA); criteria. value = 0.91. South Korea) during 2003 and Dual-color SISH 2004. (INFORM HER2 Single-color DNA, performed on SISH – dual- sequential sections). color SISH: 97.0%, k HER2 status was value = 0.93. determined on TMA slides (1.0 mm cores, two cores per patient). Carbone et Eighty-nine breast cancer IHC (clone 4B5); FISH and SISH: 2007 IHC – SISH: Discrepant al., 2008 specimens from five Italian FISH (PathVysion); ASCO/CAP scoring 96.6%. cases showed (504) National Cancer Institutes, Single-color CISH criteria; challenging diagnosed between 2000 and (Zymed); CISH: FISH – SISH: situations, 2001. Cases were selected Dual-color SISH As defined by Tanner 95.1%. including according to HER2 expression (INFORM HER2 et al., 2000 (505) polysomy of (assessed with clone 4B5) to DNA). SISH – CISH: chromosome obtain a series of tumors from 97.8%. 17, HER2 gene all IHC categories. deletion and intratumor heterogeneity.

- 302 -

Brügmann Two hundred and eighteen FISH (Dako); FISH and SISH: 2007 FISH – SISH: et al., 2011 consecutive primary breast Dual-color SISH ASCO/CAP scoring 98.5%. (506) cancer specimens diagnosed at (INFORM HER2 criteria. In statistical Department of Pathology, DNA). analysis, however, Aarhus University Hospital FDA-approved scoring (Aarhus, Denmark) (timeframe HER2 status was criteria were used not indicated). determined on TMA (amplification when slides (1.5 mm HER2/CEP17 ratio ≥ cores, number of 2.0). cores per patient not indicated). Kang et al., One hundred and sixty-five IHC (HercepTest); IHC, FISH and SISH: IHC – SISH 2009 consecutive breast cancer FISH (PathVysion); 2007 ASCO/CAP (excluding (507) specimens diagnosed at the Dual-color SISH scoring criteria. equivocal Asan Medical Center (Seoul, (INFORM HER2 cases): South Korea) during 2003 and DNA). 91.9%, k 2004. value = 0.78 HER2 status was determined on TMA FISH – SISH: slides (1.0 mm 98.2%; k cores, two cores per value = 0.94. patient), except for IHC that was performed on whole slide section.

- 303 -

Mansfield et Two hundred and fifty-one FISH (PathVysion); FISH and SISH: 2007 FISH – SISH: al., 2013 samples selected from 7’872 Dual-color SISH ASCO/CAP scoring 83.0%, k (508) breast cancer specimens (INFORM HER2 criteria. value = 0.58. diagnosed at Mayo Clinic DNA). Probe (Rochester, MN) between D17S122 was used January 2010 and December in samples 2011. Difficult cases (cases exhibiting polysomy showing monosomy or of chromosome 17 polysomy of chromosome 17, (≥ 6 CEP17 HER2 gene deletion, and signals). equivocal cases) were overrepresented (201/251 were difficult cases). Dekker et One thousand and two hundred IHC (HercepTest, IHC: 2007 ASCO/CAP IHC – mono- al., 2012 and ten breast cancer clones Sp3, 4B5); scoring criteria. color SISH: (509) specimens from six hospitals in Single-color and Mono-color SISH: Using The Netherlands diagnosed dual-color SISH mean HER2 gene copy HercepTest: between 2006 and 2008. (Ventana) number < 6.0: non- 98.3%; amplified; mean HER2 Using Sp3 HER2 status was gene copy number ≥ clone: 99.3%; determined on TMA 6.0: amplified. Using 4B5 slides (0.6 mm Dual-color SISH: 2007 clone: 99.3%. cores, three cores ASCO/CAP scoring per patient). criteria. Dual-color – mono-color SISH: 99.2%.

- 304 -

Jang et al., One hundred and ten (not FISH (PathVysion); FISH and SISH: 2007 FISH – SISH: Discrepant 2010 specified if consecutive cases) Dual-color SISH ASCO/CAP scoring 98.8%, k cases showed (510) breast cancer specimens, (INFORM HER2 criteria. value = 0.96. SISH signals diagnosed between January DNA). Positive that were and December 2003 at the agreement: difficult to Department of Pathology of the HER2 status was 94.7%; interpret. Kyungpook National University determined on TMA negative Hospital (Kyungpook, South slides (3.0 mm agreement: Korea). cores, one core per 98.5%. patient). Park et al., Two hundred and fifty-seven IHC (HercepTest); IHC: 2007 ASCO/CAP FISH – SISH: Half of 2011 consecutive primary breast FISH (PathVysion); scoring criteria; 96.5%, k discordant (459) cancer specimens diagnosed at Single-color CISH CISH: : mean HER2 value = 0.903 cases between the Inje University Paik Hospital (Zymed); gene copy number > single-color (Seoul, South Korea)(timeframe Double-color SISH 4.0 or when a large SISH – CISH: CISH and dual- not indicated). (INFORM HER2 copy cluster was seen 96.2% (k color SISH DNA). in more than 50% of value not showed cancer cell nuclei: indicated). polysomy of HER2 status was amplified; chromosome determined on TMA FISH and SISH: 2007 17. slides (2.0 mm ASCO/CAP scoring cores, three cores criteria. per patient).

- 305 -

Bartlett et Commercially available TMA FISH (PathVysion); Evaluation of HER2 FISH – SISH: No indication al., 2009 blocks containing two cores Dual-color SISH status for both FISH 96.0% (range given about (511) from 45 breast cancer (INFORM HER2 and SISH performed 88.9-100%). diameter of specimens. DNA). according to the cores nor about recommendations for selection SISH performed in seven HER2 testing in the UK criteria of laboratories and FISH in one (512): a case is cases. central laboratory. considered non- amplified when ratio is < 2.0 and amplified when ratio is ≥ 2.0. Unal et al., Forty invasive ductal carcinoma FISH (Kreatech); FISH, dual-color SISH: FISH – SISH: 2013 showing equivocal Dual-color SISH when HER2/CEP17 92.3%. (513) immunohistochemical staining (Ventana). ratio < 2.2: non- (2+) evaluated at the amplified; when Department Pathology of HER2/CEP17 ratio ≥ Antalya, Turkey. 2.2: amplified.

Jacquemier Eight hundred and forty breast FISH (PathVysion); FISH, dual-color SISH: FISH – SISH: et al., 2013 cancer specimens, from 15 Dual-color SISH 2007 ASCO/CAP 97.0%. (514) centers in France. Each center (Ventana). scoring criteria. provided a mean of 56 cases, according to the following IHC distribution: 38% of cases were IHC 0, 22% IHC 1+, 13% were 2+ and 27% were 3+.

Of the 840 cases, FISH and dual-color SISH results were available for 498 cases.

- 306 -

Lee et al., Five hundred and forty-three FISH (PathVysion); FISH, dual-color SISH: FISH – SISH: Number of 2012 breast cancer specimens Dual-color SISH 2007 ASCO/CAP 96.7% (k cores per (458) diagnosed between 1992 and (Ventana). scoring criteria. value = 0.92). patient not 2004 at Korea University Guro specified. Hospital, Seoul, South Korea. HER2 status was determined on TMA slides (2.0 mm cores). Abbreviations: HER2: human epidermal growth factor receptor 2; IHC: immunohistochemistry; SISH: silver-enhanced in situ hybridization; FISH: fluorescence in situ hybridization; CISH: chromogenic in situ hybridization; ASCO/CAP: American Society of Clinical Oncology/College of American Pathologists; TMA: Tissue microarray.

- 307 -