“Omics” of HER2-Positive Breast Cancer

Abstract

HER2/neu amplification/overexpression is the only somatic mutation widely considered to be a marker of disease outcome and response to treatment in breast cancer. Pathologists have made large efforts to achieve accuracy in characterizing HER2/neu status. The introduction of transtuzumab contributed to development of additional measures to identify sensitive and resistant subclasses of HER2/neu-positive tumors. In this article, we describe the latest advances in HER2/neu status diagnostic assessment and the most relevant research emerging from “Omics” (genomics, epigenetics, transcriptomics, and proteomics) studies on HER2/neu-positive breast cancer. A large quantity of biomarkers from different studies highlighted HER2/neu-positive specific proliferation, cell cycle arrest, and apoptosis mechanisms, as well as immunological and metabolic behavior. Major driver genes of tumor progression have had a candidate status (GRB7, MYC, CCND1, EGFR, etc.), even though the main role for HER2/neu is largely recognized. Nonetheless, existing omics data and HER2/neu-positive molecular profiles seem to suggest that few proteogenomic alterations in HER2, EGFR, and PI3K networks could significantly affect the effectiveness of transtuzumab. The systematic search of molecular alterations in and across these pathways can help to select the most appropriate drug for a given patient based on in-depth understanding of complexity in tumor biology.

Introduction

Breast cancer (BC) represents a highly heterogeneous group of tumors at both the clinical and molecular level (Kennecke, 2010; Minafra, 2012; Voduc, 2010). Human epidermal receptor protein-2 (HER2/neu; c-erb-2) positive breast cancer is one of four molecular breast carcinoma subtypes (luminal A, luminal B, basal like, HER2/neu positive), first defined by Perou and colleagues (2000), then corroborated by others (Sørlie, 2001; Wirapati, 2008), and is characterized by HER2 gene overexpression in tumor cells, and by a clinical aggressive growth and poor prognosis. HER2 trans-membrane molecule, coded by a gene located on 17q21.1, belongs to the tyrosine-kinase receptor superfamily and is expressed in a variety of normal epithelia, including breast duct epithelium.

It is widely accepted that HER2 amplification or overexpression, which occurs in 15%–20% of BC cases, is the only somatic mutation considered a marker of disease outcome, as well as a predictive marker of response to treatment. As prognostic marker, it is closely related to a poor prognosis in both node-positive and- negative patients (Gruver, 2001), and is associated with the absence of hormonal receptor expression, poorly differentiated histologic subtypes, DNA aneuploidy, and high cell proliferation. Nonetheless, the relationship between the HER2/neu gene status and other relevant prognostic factors (i.e., tumor size, patient age, lymph node status, and survival) remains unclear. Moreover, HER2-positive (HER2+) tumors are generally resistant to cytoxan- or tamoxifen-based regimens, while they are sensitive to anthracycline-based chemotherapy and to transtuzumab (Herceptin®) treatment. Of the biological targeted therapies, transtuzumab is one of the first developed monoclonal receptor-targeting antibodies, belonging to a series of similar molecules, some of which are clinically approved (Lapatinib, Tykerb®), while others are still under experimentation (Pertuzumab, Omnitarg®) (Barginear, 2012; Smith, 2012). In this article, we describe the latest advances in HER2/neu status diagnostic assessment and the most relevant research data from recent genomic, epigenetic, transcriptomic, and proteomic studies.

HER2/neu Diagnostic Assessment

The high implication of correct HER2/neu diagnostic assessment in BC therapeutic decisions is of primary importance in clinical oncology. HER2/neu oncogene evaluation provides important prognostic information and helps clinicians to identify patients with primary or advanced metastatic cancer who are the most likely to benefit from Herceptin-targeted therapy. For this reason, this review cannot discuss HER2+ omics profiles without pointing out pathologists' efforts to correctly characterize HER2 status (Goddard, 2012).

Quantitative reverse transcription PCR (qRT-PCR), immunohistochemistry (IHC), and fluorescence in situ hybridization (FISH) have been, over the last decades, the most common techniques used to detect HER2/neu amplification. Currently, all these approaches are significantly hampered by tumor DNA dilution in nontumor DNA derived from benign breast tissue and inflammatory cells. At present, IHC and FISH methods are used in conjunction, according to FDA guidelines. However, the American Society of Clinical Oncologists (ASCO) and the College of American Pathologists (CAP) have recently developed new guidelines for the laboratory testing of HER2 status in BC (Wolff, 2007). Accordingly, the IHC assay is used as an initial screening assay to clearly identify negative or positive cases, then the FISH assay is used to validate the results or assign IHC equivocal scores. There is great concordance between HER2 protein overexpression and gene dosage detected by IHC and FISH, respectively, but biological and laboratory gaps still remain. Indeed, despite HER2/neu gene amplification corresponds to HER2 protein overexpression in 90% of cases, IHC and FISH discordance is well documented by several authors (Pauletti, 1996; Wang, 2002). Pauletti et al. (1996) demonstrated that at least 3% of BCs show protein overexpression in the absence of concomitant gene amplification, implying a mechanism other than gene amplification, or chromosome 17 polysomy in the enhancing of protein production.

Nevertheless, numerous technical adjustments have been suggested to reduce false-negative/positive results. Furthermore, in 2009 ASCO/CAP societies expanded its HER2 testing guideline to assay genetic heterogeneity. According to Gown, et al. (2008), the quality of tissue fixation, the choice of antibody, and the threshold determination for reporting positive results are the main parameters ensuring the laboratory accuracy of HER2 status assessment. Furthermore, imaging analysis using dedicated software and other kinds of scoring normalization, consisting of subtraction of the non-neoplastic breast epithelium signal score from that of the tumor (Gown, 2008) significantly improve the rate of success in HER2 determination. Several commercial kits have been developed to address this issue. Among them, the FDA-approved HercepTest^™ (DAKO, Glostrup, Denmark), an IHC testing kit for high level accuracy and reproducibility of HER2 IHC testing, has produced significant numbers of false positive results, nonvalidated by FISH. On the other hand, other commercial kits based on genomic assays, HER2+ not restricted, including the MapQuant Dx, an Affymetrix microarray based assay kit (IPSOGEN, Marseille, France), the FDA-approved Oncotype DX Breast Cancer Assay (Genomic Health Inc, Redwood City, CA), and the MammaPrint (Agendia, Irvine, CA), are already clinically used despite ongoing validation studies (Reis-Filho, 2011). Furthermore, the Gene expression Grade Index (GGI), a valuable tool able to assign histologic grade into low- and high-risk and to predict response to chemotherapy, based on the gene expression signature, has been also generated (Reis-Filho, 2011).

In addition, the qRT-PCR technique, predominantly used to detect HER2/neu mRNA in peripheral blood and bone marrow samples, is specific to gene amplification (Tubbs, 2001). Although, RNA integrity is not completely preserved in formalin-fixed paraffin-embedded (FFPE) tissue, the qRT-PCR evaluation of HER2/neu in breast tissue has been also proposed both in fresh/frozen and in FFPE tissue. However, to our knowledge, the IHC/FISH tests are the only two recommended methodologies for the diagnostic characterization of HER/neu status.

Genomics of HER2+ Tumors

Since the first BC classification into molecular subtypes by Perou and colleagues (2000), today's wide-scale genome analysis has led to a higher level of breast tumor characterization, and this allows us to question the HER2/neu positive BC genomic profile.

A lot of effort was put to reach high and accurate standard levels in HER2 assessment. More recently, the introduction of transtuzumab allowed the identification of sensitive and resistant subclasses of HER2/neu-positive breast tumors. Among treatment-resistant patients, some demonstrate primary resistance to therapy, while others show acquired resistance. Many women receiving trastuzumab develop resistance within 1 year, and 15%–25% of women diagnosed HER2+ in early-stage disease develop tumor relapse within 3 years, despite therapy (Perez, 2011).

These observations produce indirect evidence of the existence of different molecular patterns among HER2/neu-positive tumors, due to the co-existence of additional mutations or chromosomic alterations able to disturb therapeutic efficacy, or due to the acquisition of new mutations during tumor progression. Thus, it can happen that some tumor clones show sensitivity, while others become resistant and the net therapeutic effect is insufficient to reach a positive outcome.

What is responsible for making a group of HER2+ tumors so heterogeneous in response to the drug? To address this issue, HER2+ genomic profiles were characterized by several authors (Chin et al., 2006; Hu et al., 2009; Leary et al., 2008; Marchiò et al., 2008; Sircoulomb et al., 2010; Staaf et al., 2010; 2011; Theillet, 2010) (Table 1) to establish the most common chromosome copy number alterations (CNA) and to identify and characterize the driver genes promoting neoplastic progression through their altered function.

Table 1.

Main HER2+ Associated CNAs

Features	Gain	Loss	Citation
HER2+	1q, 7p, 8q, 16p, 20q	1p, 8p, 13q, 18q.	Chin. et al. 2006
HER2+/ER+	1q, 6p, 8q, 11q, 17q	1p, 7q, 8p, 9p, 11q, 17p, 19p, 22q	Marchiò C. et al. 2008
HER2+/ER−	1q, 59, 8q, 17q	1p, 8p, 16q, 19p	Marchiò C. et al. 2008
HER2+	10p, 11q13, 17q12, 17q21, 17q23	5q	Hu X. et al. 2009
HER2+	1q, 5p, 6p, 8q, 9q, 11q, 12p, 12q, 16p, 17q, 19p, 19q, 20p, 20q, 21q	1p, 3p, 8p, 9p,11q, 16q, 17p, 17q, 18q	Staaf J. et al. 2010
HER2+/ER+	11q13	1p, 11q, 6q	Staaf J. et al. 2010
HER2+/ER−		5q	Staaf J. et al. 2010
HER2+/ER+	8p, 8q, 17q, 8q	1p36, 4p16, 11p15	Sircoulomb F. et al. 2010
HER2+/ER−	11q		Sircoulomb F. et al. 2010
HER2+	6q, 8p, 8q, 10p, 11q, 17q, 19q, 20q	11q, 16q, 17p, 17q, 18q	Staaf J. et al. 2011

A synoptic description of the main chromosome copy number alterations (CNA) observed in HER2/neu positive Breast Cancer by different authors.

Tumor heterogeneity, low tumor cell percentages, and the presence of different tumor clones hinder the identification of these alterations. However, to accurately detect allelic changes in carcinomas, the application of genome wide analysis on flow-sorted tumor fractions, separated according to the DNA-index, can help the isolation of different clones and the characterization of genomic alterations (Corver, 2008).

Staaf and co-workers (2010) recently characterized a large series of 200 HER2+ BC samples, by aCGH (array-comparative genomic hybridization), comparing these profiles with a series of 554 HER2-negative (HER−) breast tumors. Although they failed to observe significant differences between HER2+ and HER2− tumors with regard to the incidence of specific recurrent amplifications other than HER2, they observed high levels of amplification at multiple sites, some of which correspond to known oncogenes (FGFR1, MYC, CCND1, and ZNF217), and the mutually exclusive occurrence of 8p12, 11q13, and 20q13 amplification in HER2+ tumors. However, in this series of cases, the most common amplifications regarded the HER2 neighboring loci on 17q and higher HER2 copy numbers indicated a worse prognosis. Interestingly, they could also distinguish two kinds of HER2 amplification patterns: one is a smaller amplicon of 86 kb surrounding HER2 gene, associated with estrogen receptor (ER) negative tumors, the other is a larger amplicon of 1.1 Mb covering the region from centromere to the TOP2A telomeric gene, found interrupted, associated to ER-positive (ER+) tumors and better survival. According to Leary et al. (2008), a significant fraction of ductal breast tumors bears TOP2A alterations. This observation could be of clinical utility to select those HER2+ patients who could benefit from the administration of topoisomerase targeted therapies in addition to the standard therapeutic transtuzumab regimen.

The International Cancer Genome Consortium (ICGC) recently funded a large study aiming to list somatic mutations comprehensively, including base substitutions, insertions, deletions, copy number changes, translocations, and other chromosomal rearrangements, as molecular signatures typical of BC subgroups (at least 50 tumors for each subtype). Although the project itself is primarily focused on ER+/HER2− tumors, the French partner will focus on HER2+ subtypes.

Hu and co-workers (2009) recently conducted a similar work, characterizing similarities and differences in genomic signatures of BC cases and cell line samples representative of typical breast tumor subtypes. In this analysis, HER2+ tumors were found to have the highest frequency of gains, including the HER2 amplicon on chromosome 17, a feature that remained significant even excluding chr.17 from the statistical analysis. The gain of 10p, 11q13, 17q12, 17q21, 17q23, and the loss of 5q13-14 were recurrently observed in HER2+ tumor samples. Furthermore, the authors generated a bioinformatic tool to select those genes that could play a role as driver genes, on the basis of three features: 1) involvement in chromosome alterations, 2) significant differential gene expression in tumor samples compared to normal tissue, 3) correlation between altered gene expression and poor clinical outcome. By these criteria, HER2 was clearly identified as a driver gene. In addition, GRB7 belonging to the same amplicon, MYST2, PPM1D, CCND1, and RASA1 located on 17q21, 17q23, 11q13, and 5q13-14, respectively, also met the same criteria.

The integration of data from the analysis of this BC series with those from a series of colorectal cancers and previous mutation analyses of genes from the Reference Sequence database (RefSeq) conducted in these same tumor types, allowed Leary and co-workers (2008) to observe that converging gene pathways, including the HER2, EGFR, and PI3K ones, were most frequently affected by copy number changes and/or point alterations in both breast and colorectal cancers, including candidate driver genes (MYC, HER2, CCNE1, CCND1, EGFR, FGFR2, and IRS2).

Concerning the MYC gene (chr. 8), its role as a proliferative driver signal downstream of HER2 activation, is controversial. In HER2+ BCs, c-MYC gene alterations seem to affect the response to targeted therapies (Chen and Olopade, 2008; Perez et al., 2011) and the co-existence of HER2/MYC amplification was correlated with a poor prognosis. On the other hand, preliminary observations from the adjuvant trastuzumab NSABP B31 trial showed that trastuzumab-treated patients with MYC/HER2 co-amplification benefited significantly more than patients with only HER2 amplification (Perez et al., 2011). However, the latest work from this same group did not reveal any significant association between MYC amplification and a gain in terms of prolonged disease free survival in trastuzumab-treated patients. Another essential aspect to be discussed regards the influence of ER status on HER2 amplification. As is known, luminal B type are typically HER2+/ER+ breast carcinomas, which represent approximately 50% of HER2+ cases, whereas HER2+/ER− are classically known as pure HER2 subgroup (Gruver, 2011). Marchiò and co-workers (2008) studied the genomic differences occurring in 26 HER2+ tumors (11 ER+ and 15 ER−) in comparison with a series of 70 HER2− breast carcinomas.

The authors observed the following recurrent chromosomal copy number alterations in ER+ tumors: gain of 1q, 6p, 8q, 11q, 17q; loss of 1p, 7q, 8p, 9p, 11q, 17p, 19p, 22q, and in ER− tumors: gain of 1q, 5p, 8q, 17q; loss of 1p, 8p, 16q, 17p, 19p. As can be observed, HER2+ tumor subtypes share some of their respective genomic alterations, though the arising of 5p15–p12 was significantly associated with ER− cancers, whereas two regions of amplification, 11q13.3 and 17q21.2, encompassing GJC1, IGFBP4, TNS4, and TOP2A genes, were significantly associated with ER+ carcinomas.

Actually, data and studies reported herein suggest indirect evidence that heterogeneity within HER2+ BCs, producing diversity in response to therapy, should be interpreted considering HER2/neu as the major driver gene in tumor progression, while only a few other additional mutations occurring on HER2, EGFR, and PI3K pathways might be responsible for therapy failure. Namely, despite the number of accumulated CNA or the tumor ploidy status, only particular mutations or genomic alterations, impairing crucial genes belonging to HER2 signaling cascade, importantly affect therapeutic outcomes (Esteva et al., 2010).

Since experimental evidences suggested that mutational activation of the PI3K pathway confers resistance to HER2-directed therapies, patients with drug-resistant HER2+ BC participate in clinical trials to evaluate the combined treatment of PI3K pathway inhibitors and trastuzumab. Indeed, HER2+ BC cell lines are highly sensitive to PI3K and mTOR inhibitors before and after acquiring resistance to trastuzumab or lapatinib (Eichhorn et al., 2008), suggesting PI3K pathway dependence even in the presence of HER2 targeted downregulation (Kumandan et al., 2012). In addition, there is also evidence that patients with PIK3CA mutations and/or PTEN decreased levels have a poor outcome following treatment with trastuzumab, or they show a less effective response rate to trastuzumab/chemotherapy combined treatment compared to HER2+ tumors having a normal PI3K pathway (Esteva et al., 2010). Instead, HER2 dual blockade with trastuzumab and lapatinib seem to be more effective against HER2+/PTEN−deficient BCs. On the other hand, the addition of the TORC1 inhibitor everolimus to trastuzumab and chemotherapy seem to improve, by 20%–45%, the response rate to treatment (Jerusalem et al., 2011).

Epigenetics of HER2+ Tumors

Epigenetic modifications are molecular changes directly affecting gene expression by mechanisms comprising histone modifications, DNA methylation, and the annealing of noncoding antisense RNAs (Kouzarides, 2007; Surani et al., 2007). In addition to the occurrence of these changes during life-time, epigenetic alterations can also be heritable. These changes produce indirect functional effects on cell processes able to strongly modify cellular phenotypic behavior. DNA hypomethylation can be associated with gene reactivation and chromosomal instabilities, and might lead to upregulation or overexpression of proto-oncogenes. Instead, DNA hypermethylation is frequently associated with gene repression and genomic instability (e.g., through silencing of DNA repair genes) and can lead to the suppression of tumor-suppressor genes and chromatin compaction. Aberrant DNA methylation is deeply involved in the development and progression of human cancers (Baylin and Ohm, 2006; Esteller, 2007; Jones and Baylin, 2007). The presence of frequent CpG island (CGI) methylation, designated as the CpG island methylator phenotype, is associated with distinct pathological characteristics in several tumor types (Gaudet et al., 2003; Rodriguez et al., 2006).

Aberrant methylation of CpG island promoters, frequently observed in BC, permit the correlation of distinct epigenetic patterns with BC molecular subtypes (basal-like, luminal A, luminal B, HER2+). Indeed, analysis of 807 cancer-related genes in 189 breast tumors, showed different methylation profiles associated to different BC subtypes, especially luminal A, luminal B, and basal-like (Holm et al., 2010). Identification of subtype-specific methylation patterns revealed that, for example, RASSF1 and GSTP1 were specifically methylated in luminal B (ER+/HER2+) tumors and unmethylated in basal-like tumors (Holm et al., 2010). According to Hu and colleagues (2006), seven genes significantly hypermethylated in one subtype (ARHGDIB, GRB7, and SEMA3B in basal-like; MMP7 and PEG10 in luminal A; GSTP1 and CHI3L2 in luminal B) were also less expressed in the corresponding breast subtype (Hu et al., 2006). Also, Bediaga and co-workers (2010) found that basal-like, HER2+, luminal A, and luminal B molecular subtypes displayed distinct methylation profiles. Specifically, HER2 enriched breast tumors (HER2+ and luminal B) were associated with the hypermethylation of several genes related to cancer development, such as, NPY, DBC1, FGF2, RASSF1, HS3ST2, SOX17, and TNFRS10D, while basal-like tumors showed lack of methylation of these markers. These differences in the methylation profiles of basal-like and HER2-overexpressing tumors support the utilization of these biomarkers for prognosis and therapeutic stratification of BC patients (Bediaga et al., 2010).

In addition, Terada and co-workers (2009) demonstrated that frequent CpG island methylation is highly associated with HER2 amplification. They selected 11 genes that are unlikely to confer any growth advantage to cells (LOC346978, 3OST2, GREM1, XT3, PCDH10, FLNc, THBD, COE2, CLDN3, F2R, and AK5) from 20 and 14 methylated genes in human breast and gastric cancers, respectively, and quantified their promoter CGI methylation levels in 63 BCs. The incidence of HER2 amplification was significantly higher in cancers with frequent CpG island methylation and the number of methylated genes correlated with the degree of HER2 amplification. This study showed that elevated methylation levels have a strong association with HER2 amplification in BCs, suggesting clinical implications in the possibility of predicting the development of metastatic HER2+ tumors (Terada et al., 2009).

Fiegl and co-workers (2006) used 143 cases of primary BCs known for their HER2/neu status to correlate 35 DNA methylation markers with the HER2/neu immunohistochemical staining intensity. They identified 5 of 35 genes, showing significant correlation coefficients: CDH13, MYOD1, PGR, and HSD17B4, positively associated with HER2/neu expression, whereas BRCA1 was negatively associated with HER2/neu expression. They also found that most of the genes showing significant positive correlations between their methylation profile and the HER2/neu status are involved in hormonal regulation (PGR and HSD17B4) or are members of the cadherin family (CDH13). PGR codes for the progesteron receptor (PR) and HER2/neu positive cancers often show PR lower levels. Moreover, a strong association between estrogen receptor expression and progesterone receptor gene methylation was described, since PGR methylation revealed the absence of the estrogen receptor. In addition, significant levels of HSD17B4 methylation were present only in HER2/neu positive BCs which consequently had low HSD17B4 mRNA levels (Fiegl et al., 2006). The HSD17B4 gene codes for an enzyme responsible for metabolizing estradiol, converting active estrogen into inactive metabolites. Estrogen deprivation therapy, by aromatase inhibition, seems to be more effective in HER2/neu positive BCs, compared to the estrogen receptor blockade by tamoxifen. Indeed high intra-tumor estrogen concentrations might prevent antiestrogens from blocking ER action, producing a resistant phenotype. This led the group of Fiegl et al. (2006) to assume that HER2/neu-positive cancers create an environment that may prevent the antitumor activities of tamoxifen through either low levels of functional ER (evidence of PGR methylation) or low expression of estradiol metabolizing enzymes (evidence of DNA methylation-mediated HSD17B4 low expression). Furthermore, the same authors observed a significant correlation between the CDH13 methylation pattern and HER2/neu status, where CDH13 has a tumor suppressor effect when suppressed by DNA methylation. Indeed, decreased expression of cadherin molecules in invasive carcinomas results in cell scattering and decreased cell–cell adhesion, which may enhance tumor progression and invasion. It is well known that breast tumors bearing BRCA1 germline mutations are more often negative for HER2/neu. In frozen tissue, Fiegl and co-workers (2006) showed significant levels of BRCA1 methylation only in HER2/neu-negative tumors. Also, they detected DNA methylation changes not only in cancer cells but also in the stroma of HER2/neu-positive cancers. Several data demonstrate that HER2/neu overexpression in the tumor epithelial fraction has a strong effect on the activity of the tumor stroma: in the mouse model, mammary tumorigenesis was triggered in a single step by HER2/neu transgene overexpression in the epithelial compartment of the mammary gland. A myofibroblast-like cell line derived from this tumor not expressing the HER2/neu transgene was highly aggressive and gave rise to sarcomatoid tumors. This suggests that HER2/neu cancer cell signaling affects the surrounding stroma by epigenetic mechanisms (Galiè et al., 2005).

In addition to the described molecular mechanisms, microRNAs (miRNAs) play an important role in regulating gene expression in BC by epigenetic events. miRNAs are a recently discovered class of regulatory RNAs, influencing the stability and translational efficiency of target mRNAs. Indeed, these are very small, single-stranded RNAs, not coding for proteins, but playing a role in regulating the amount of protein expressed from coding mRNAs (Bentwich et al., 2005). They have been implicated in an increasing number of biological processes, including neoplasia. Several studies have shown an involvement of these regulatory molecules in BC (Castilla et al., 2012; Esquela-Kerscher and Slack, 2006; Lowery et al., 2008; Lu and Clark, 2012; Wee et al., 2012). Furthermore, miRNA expression has been examined in a wide range of BC cell lines and in normal and cancer breast tissues. Specific miRNA profiles have been associated with BC subgroups, since miRNA profiling identified distinct miRNA subsets able to distinguish HER2+ from HER2− and ER+ from ER–BCs, independently of other clinically important parameters. Subsets of HER2-specific miRNAs (let-7f, let-7g, miR-107, miR-10b, miR-126, miR-154, and miR-195) and ER/PR specific miRNAs (miR-142-5p, miR-200a, miR-205, and miR-25) were established (Mattie et al., 2006). Furthermore, functional studies have uncovered miRNA roles in BC, both as tumor suppressor genes (e.g., miR-335) and oncogenes (e.g., miR-21). miR-125a and miR-125b, identified as potential tumor suppressors in tissue culture models, were significantly downregulated in HER2 amplified and overexpressing BCs (Mattie et al., 2006). Bioinformatic comparisons of HER2 and HER3 3’-UTR sequences revealed the presence of target sites for these miRNAs in these regions. miR-125a or miR-125b overexpression in the SKBR3 cells (a HER2-overexpressing human BC cell line) suppressed HER2 and HER3 transcript and protein levels, thereby reducing anchorage-dependent growth and enhancing cell motility and invasiveness. However, this effect was only marginal in the HER2-independent BC cell line MCF-10A (Scott et al., 2007).

The implications of these data concern the ability of these miRNAs to affect different cancer subtype behavior and their potential role as biomarkers.

Gene Expression of HER2+ Tumors

Gene expression profiling studies have provided a BC molecular classification into clinically relevant subtypes, new tools to predict disease recurrence, response to different treatments, metastatic progression, and new insights into related oncogenic pathways (Sotiriou and Piccart, 2007).

Microarray-based expression studies have demonstrated that BC is a clinically diverse and molecularly heterogeneous disease comprising subtypes with distinct gene expression patterns, associated with different prognoses and outcomes (Wirapati et al., 2008).

Landmark studies by Perou et al. (2000) and Sorlie et al. (2001) have identified five distinct subtypes, each different from the other in its gene expression profile. These subtypes have been shown to be mainly driven by the expression of ER+ and ER negative (ER−) related genes, proliferation-related genes and, to a lesser extent, by HER2 and genes mapping on the HER2 amplicon on chromosome 17 (Reis-Filho and Pusztai, 2011).

Dvorkin-Gheva and Hassell (2011) proposed a “14-gene HER2 signature” that they used to separate HER2+ from HER2-tumors with an accuracy of 93.18%, sensitivity of 77.78%, and specificity of 94.94%. The HER2 predictive gene signature consists of 14 genes: CRK7 (CDK12), HER2, F2RL1, GRB7, IDI1, ITGB6, PERLD1, PPARBP, SEC63, STARD3, TRIM26, DIRAS2, DUSP24, and UBTF, including HER2, and 5 genes located within the HER2 17q12-q21 amplicon (Dvorkin-Gheva and Hassell, 2011).

HER2 pathway deregulation, resulting in injurious effects in cancer cells, is primarily due to an excess of signal, rather than to mutations on related genes, and this allows tumor cells to grow and spread. However, resistance to trastuzumab is frequent and the mechanisms are still poorly understood.

In studies designed to evaluate markers predictive for anthracycline responsiveness, a strong adverse prognostic impact has been suggested for high levels of HER2/TOP2A mRNA co-expression, rather than for TOP2A gene amplification, probably due to a selective advantage of cells having a more effective DNA replication and remodeling processes.

The TOP2A gene, located at ∼700 kb telomeric with respect to HER2, is co-amplified in 30%–40% of the HER2 amplified tumors, even though deletions are also frequently observed. As TOP2A is a crucial factor in DNA remodeling and replication, as well as being a target for anthracyclines and other chemotherapeutic agents, the assessment of HER2/TOP2A co-expression could be helpful in the setting of anti-HER2 plus adjuvant treatments in profiled patients (Fountzilas et al., 2012).

In addition, several 17q12-q21 genes are variably co-amplified and co-expressed with HER2, influencing the response to trastuzumab and/or constituting accessory targets to obtain synergistic treatments (Slamon and Press, 2009). Therefore, a better knowledge of HER2 amplified BCs may help to design new therapeutic strategies. To better characterize this particular BC subtype, Sircoulomb and colleagues (2010) integrated aCGH and RNA expression profiling of 54 HER2+ BCs, finding 37 genes whose expression levels were deregulated with respect to DNA copy number losses and gains (or amplification). These genes were C8orf68, C8orf53, MAL2, LOC286052, SQLE, KIAA0196, RSF1, INTS4, KCTD21, DDX52, MRPL45, SOCS7, ARHGAP23, SNIP, MLLT6, CISD3, PCGF2, PSMB3, PIP4K2B, CCDC49, LASP1, LOC642808, CACNB1, LOC90110, FBXL20, MED1, CRKRS, STARD3, PERLD1, HER2, C17orf37, GRB7, RAPGEFL1, WIPF2, CDC6, RARA, and CCNE1. Among these genes, GRB7, encoding for an adaptor-type signaling protein, able to bind a variety of tyrosine kinases receptors, including EGFR and HER2, is of particular interest. Indeed, GRB7 may facilitate HER2-mediated signal transduction and tumor formation (Bai and Luoh, 2008) and it has been suggested as a therapeutic target (Sircoulomb et al., 2010).

Furthermore, Sircoulomb and co-workers also described gene expression differences between inflammatory (IBC) and noninflammatory (NIBC) HER2 amplified BCs tumors. Twelve genes were deregulated in HER2 amplified IBCs. Ten of them (SNIP, MLLT6, CISD3, PCGF2, PSMB3, PIP4K2B, FBXL20, STARD3, GRB7, and RARA) are located within the 17q12-q21-amplicon, suggesting that this region has a particular influence on the IBC phenotype in HER2 amplified BCs. Another important gene expression tool to capture the complexity of BC is the substantial amount of data available from public databases. Cheng and co-workers (2012) collected a large collection of breast tumor gene expression data (n=4.010) derived from 23 datasets. They were able to identify a series of highly expressed genes correlated with decreased survival among patients with HER2−/ER+ BC. Particularly, upregulated HSP90 mRNA expression renders HER2− BCs more aggressive, resulting in poor prognosis. However, HSP90 is required for the stabilization of many proteins playing key roles in cancer growth and survival, such as ER, PR, EGFR, essential components of HER2 signaling (HER2, AKT, c-SRC, RAF, and HIF-1a) and HER2 itself, which is among the most sensitive HSP90 substrate. In particular, HSP90 inhibition mediates HER2 as well as PI3K and AKT degradation in HER2 overexpressing cancer cells, consequently, HSP90 inhibitors plus trastuzumab could have significant anticancer activity in HER2+ patients, especially those with metastases and already being treated with trastuzumab (Cheng et al., 2012; Modi et al., 2011).

Concerning the role of tumor-initiating cells (TICs), Liu and colleagues (2012) searched for prognostic signatures based on gene expression of enriched TIC fractions specific for the BC subtype, as this fraction of cells within a whole tumor is responsible of sustaining tumor growth.

HER2/neu drives asymmetrical cell division, increasing the number of TICs with respect to mammary stem cells; furthermore its continuous expression is required to sustain tumorigenesis. A 17-gene signature was obtained from enriched HER2/neu mammary TICs, generated on the basis of differentially expressed genes in enriched TICs samples versus non-TIC fractions, consisting of eight upregulated (AURKB, CCNA2, SCRN1, NPY, ATP7b, CHAF1b, CCNB1, CLDN8) and nine downregulated genes (NRP1, CCR2, C1qb, CD74, VCAM1, CD180, ITGB2, CD72, ST8SIA4). Such signatures have been observed to predict clinical outcomes for human HER2+/ER− BCs (Korkaya et al., 2008; Liu et al., 2012).

Notably, HER2 expression in peripheral blood can be considered an independent prognostic predictor of BC metastasis, indicating the presence of circulating tumor cells (CTCs). In a study conducted by Markou and colleagues (2011), mRNA isolated from immunomagnetically enriched CTCs was subjected to multiplex PCRs to detect CK19 and HER2. Recently, also EPCAM and MUC1 transcripts were detected in HER2+ BC CTCs.

However, to understand how HER2 overexpression may lead to a poor prognosis, it is necessary to better focus on elements participating in the downstream signaling cascade (RAS-MAPK, PI3K-Akt, and NF-κB pathways). The PI3K-Akt cascade is the most frequently mutated pathway in BC. Several drugs targeting multiple molecules belonging to the PI3K network (PI3K, AKT, mTOR) are under development for BC care. Elevated Akt1 kinase activity has been shown to be important for the development and proliferation of human cancers, including BC (Wu et al., 2010). Akt activation results in the downstream regulation of targeted genes, such as GSK-3, caspase-9, Bad, and FOXO family transcription factors. The final outcome may result in cell proliferation or anti-apoptosis (Wu et al., 2010). HER2 downregulation results in pAkt decreased levels and in FOXO1A increased expression, a mechanism producing resistance to therapy (Faltus et al., 2004). Instead, NF-κB is mostly activated in HER2+/ER− tumors. Several studies demonstrated that Akt through NFκB, induce COX-2 expression in mutated PTEN tumor cells, which in turn significantly contributes to tumorigenesis through increased angiogenesis, invasiveness, and inhibition of apoptosis (St-Germain et al., 2004). Finally, although the defined HER2 intracellular signalling pathways are based on HER2 membrane localization, there is increasing evidence of nuclear HER2 translocation, associated with increased transcription of several genes, notably COX-2, significantly mediating a worse prognosis (Boone et al., 2009).

Proteomics of HER2+ Tumors

Proteomic sciences represent one of the latest technological developments, allowing scientists to complement genomic information and increase levels of knowledge about the destiny of codified sequences from DNA to one or more proteins due to different stimuli.

Proteomic comparison of healthy and pathological samples, using experimental approaches normally sorted in “non-mass spectroscopy (MS)-based and MS-based” studies, strongly improves the potential discovery of new diagnostic and prognostic biomarkers, as well as novel therapeutic targets (Goncalves and Bertucci, 2011).

In this section, the most relevant HER2+ BC proteomic studies are reviewed, classifying them on the basis of the experimental method used, in order to highlight the proteomic contribution in understanding specific features and mechanisms of HER2+ BCs. The studies reported herein also demonstrate the proteomic flexibility of investigation, since different methodological approaches can be chosen to perform classification and or functional studies.

First, protein array applications will be discussed. This technique, considered the DNA microarray protein counterpart, as antibodies are arrayed on solid supports (biochip or proteinchip), is a qualitative/quantitative assay, able to determine the presence of specific proteins and/or their amount in biological samples such as liquids and tumor tissues.

A characterization study was recently conducted by Oliveras-Ferraros and co-workers (2010), who applied an antibody-based array technology to investigate the histogenesis of the intrinsically trastuzumab-resistant basal/HER2+ phenotype in breast carcinomas, to search for biomarker signatures that may differentiate trastuzumab-responsive from nonresponsive tumors.

On comparing trastuzumab-refractory JIMT-1 and trastuzumab-sensitive SKBR3 cell culture extracellular milieus, the JIMT-1 cells were found to secrete higher amounts of several growth factors, including amphiregulin, EGF, IGFBP-6, PDGF-AA, neurotrophins, TGFß, and VEGF.

Another interesting study, conducted by the group of Vazquez-Martin (2008), focused on the HER2+ induced “cytokine signature” in an engineered MCF-7 clone, stably expressing HER2/neu cDNA, using a Human Cytokine Array that detected 42 cytokines and growth factors. The group identified a 10-fold increased expression of IL-8 and of the growth-related oncogene (GRO). Importantly, the same authors confirmed significantly higher IL-8 and GRO circulating levels in sera from HER2+ BC patients.

Today, the latest methodological advance is the reverse phase type approach of protein microarray, RPPA (reverse phase protein array). In this case, proteins from one or more samples are immobilized on a solid support that can be screened, in a single step, by a large collection of probes, for the presence of one distinct target protein. An appealing facet of RPPA technology for molecular diagnostic applications is the possibility of searching for specific activated molecules (e.g., phosphorylated proteins at specific sites) associated with crucial pathways. Using the RRPA approach, Boyd et al. (2008) studied the phosphorylation status of 100 proteins representative of crucial signaling pathways necessary for the control of proliferation in a panel of 30 BC cell lines, in order to evaluate signaling modifications associable with different BC subtypes. Through this analysis the authors observed typical phosphorylated protein profiles for cancer subtypes. It was noted that HER2 amplified cell lines are distinct from the others in that they have high levels of pHER3, pFAK, and pEGFR.

Based on a similar approach, Luminex technology allows the simultaneous quantification of protein patterns in a biological sample, making use of different fluorescent and spectrally resolvable beads coated by specific antibodies whose signals are detected by a flow cytometer. The serum profiling of 22 cytokines in BC patients was analyzed by Luminex technology to evaluate the possibility of distinguishing them from healthy controls, node-negative from node-positive patients, and pre- from post-vaccination phases with a HER2/neu E75 peptide vaccine, an immunogenic peptide developed from the HER2/neu protein (Dehqanzada et al., 2007). Significantly elevated levels of MCP-1, Eotaxin, RANTES, and GM-CSF and decreased IL-1α and IL-4 levels were detected in patients compared to controls, whereas Eotaxin and IL-13 were increased in node-positive patients with respect to node-negative ones. In addition, response to vaccination was associated with a significant upregulation of MCP-1, Eotaxin, and IL-13 levels, supporting the serum cytokine profiling role in the monitoring of cancer vaccine trials.

Conversely, MS-based proteomic experiments permit the examination and quantification of a large number of initially unknown proteins, without the necessity of a priori assumptions or previous biological knowledge.

The MS-based approach could be used to identify and quantify biomarkers, using sophisticated 2D gels followed by MS analysis or generating a protein profiling as a multi-protein signature (Bloom et al., 2007).

Of particular interest is the SERPA (serological proteomics) assay, useful for the identification of tumor antigens able to commonly induce humoral immune response. In this case, tumor lysate proteins, separated by 2D gels and then transferred onto membranes, are incubated with the patient's serum. Proteins recognized by serum antibodies are subsequently identified by MS (Hamrita et al., 2008). This technique efficiently distinguishes protein post-translational modifications and permits the isolation of specific antibodies directed against aberrant proteins. Thus, this assay could provide specific cancer autoantibody profiles for each patient, which could be helpful in establishment individualized cancer therapies.

Using this approach, Mojtahedi and co-workers (2011) studied the differences in the anti-cancer humoral immune response occurring in 9 HER2+ and 9 HER2− node positive BCs, comparing each group with healthy individuals. Sera from these groups of subjects were tested on the MCF7 cell line proteome, and reactive spots were identified by MS. LDH-A, GAPDH, ENO1A, 3-PGDH, and PSMD13 resulted preferentially associated with HER2+ cases, whereas 14-3-3ɛ, the PP2A regulatory subunit and RNH1 commonly reacted with HER2− sera. Several glycolytic enzymes were observed to stimulate antibody production among HER2+ cases. Enhanced glycolysis is a mechanism that may help cancer cells proliferate in hypoxic environments, such as conditions within tumors, and to escape apoptosis. Importantly, the antigenicity of GAPDH, LDH-A, and enolase-α has been observed in other cancer types, and was recommended as a diagnostic, prognostic, and/or therapeutic monitoring marker for cancer patients (Chang et al., 2006).

The two-dimensional difference gel electrophoresis (2D-DIGE) technique is an approach created to improve sensitivity and reproducibility of 2D-PAGE, overcoming traditional 2D-PAGE inter-gel variability.

By 2D-DIGE and MS analysis, Schulz and colleagues (2009) compared the protein expression patterns of 15 triple-negative (HER2−/ER−/PR−) versus 19 HER2+/ER−/PR− type breast carcinomas. The authors reported 33 differentially expressed proteins, including cytokeratins (CKs), glycolytic proteins, annexins, lactoferrin, peroxiredoxin, and NME. Among these proteins, the authors reproduced previous observations regarding elevated CK8 and CK9 expression in HER2+ tumors (Brouillard et al., 2008; Zhang et al., 2005). Conversely, CK14 and CK17, known to be highly expressed in basal-like BCs were found upregulated in the triple negative tumors (Rakha et al., 2007). Furthermore, HER2+ tumors have a higher expression of some glycolytic enzymes compared to triple negative samples such as PGK1, GAPDH, TLK1, TPI, and ENO1A. The association of glycolytic markers with HER2+ tumors was also confirmed by the above described study of Mojtahedi and colleagues (2011). Note that PGK1 expression can be reduced by partially switching off HER2 signaling with Herceptin treatment (Zhang et al., 2005).

Finally, MS-based proteomics, like MALDI (matrix assisted laser desorption ionization)–TOF (time of flight) and SELDI (surface-enhanced laser desorption/ionization)–TOF technologies, can also be used to determine protein profiling on biological samples (generally liquid samples). This approach can be chosen to generate a protein peak profile, characteristic of a phenotype or a physio-pathological state (Streckfus et al., 2012).

Shi and colleagues (2006) applied SELDI-TOF technology to characterize HER2+ plasma samples in comparison to healthy subjects. Performing several statistical tests, including curve ROC analysis, they identified seven characteristic biomarkers, of which the most resolving power addressed a fragment of fibrinogen alpha (FGA), encompassing residues 605–629. Indeed, FGA605–629 is decreased in HER2+ BC patients compared to controls, and reverts to normal levels after surgical tumor removal, suggesting it to be a useful diagnostic and treatment-monitoring marker.

In another recent study, Mazouni and colleagues (2010) analyzed pre- and post-chemotherapy serum profiles of 39 HER2+ BC patients who received 6 months of preoperative chemotherapy, using LC-MALDI-TOF/MS technology. Qualitative and quantitative differences were revealed on comparing the group of pre-and post-treated patients who achieved complete response (pCR, n=21) with those with residual disease (n=18). In synthesis, 34 (32 decreased, 2 increased) and 304 peaks (157 decreased, 147 increased) were significantly changed (p<0.01, false discovery rate ≤20%) after treatment in responders and nonresponders, respectively. Of the 11 most significantly altered peaks, A2M, C3, HPX, and APCS were distinctive of the responding group, whereas chains C and A of APOA1, HPX precursor, complement C and amyloid P component were identified in the nonresponding group. These data highlight the serum profiling utility in the treatment monitoring phase, allowing the prediction of clinical outcomes (Ma et al., 2012).

Conclusions

In the current postgenomics era, high-throughput molecular technologies permit investigations into the genome, transcriptome, and proteome of cancer cells. The studies reported herein clearly highlight the HER2/neu+ BC tumor complexity, affecting an unpredictable response to targeted therapies. In order to produce conclusive results from the large quantity of data described, and to verify potential interactions among markers identified by separated studies, we generated a molecular network by introducing all the above cited HER2-associated markers into the http://string-db.org/, a molecular tool able to elaborate physical and functional associations among proteins (Szklarczyk et al., 2011). Despite data heterogeneity, Figure 1 clearly displays the tight relationships existing among molecules, showing that 107 out of 140 are closely connected in a unique network. Moreover, the analysis reveals that the most significant GO biological processes characterizing this network are: “regulation of cell proliferation”, “positive regulation of epithelial cell proliferation”, and “intracellular signal transduction” (Table 2), in agreement with clinical aggressive growth and poor prognosis featuring HER2+ tumors. As the above reported studies suggest, proteogenomic alterations occurring on a few key molecules belonging to HER2, EGFR, and PI3K signaling, significantly affect targeted therapy efficacy. In Figure 1, these markers correspond to nodes having the most protein interactions. Therefore, in order to achieve the more effective cancer patient treatments, much effort is needed to maximize the effectiveness of transtuzumab, with the support of other targeted drugs. According to our assumption, the systematic search of molecular alteration in key nodes of Figure 1 could be used to select the best appropriate supporting drug to overcome HER2+ biological tumor complexity.

FIG. 1.

HER2+ associated protein network. Protein network elaboration by STRING molecular tool (http://string-db.org/) of the above cited HER2+ associated markers. STRING is a database of known and predicted protein interactions. The interactions include direct (physical) and indirect (functional) associations derived from four sources: genomic context, high-throughput experiments, conserved co-expression, and previous bibliographic knowledge. Nonetheless, these markers resulted from separated studies, the figure displays the tight relationships existing among molecules, showing that 107 out of 140 are closely connected in a single network. The complex network generated reflects the biological complexity of HER2+ cancer cell signaling.

Table 2.

HER2+ Associated GO

GO_ID	Term	Number of Genes	p-value	p-value_fdr
GO:0042127	regulation of cell proliferation	32	3.2E-11	3.2E-7
GO:0050679	positive regulation of epithelial cell proliferation	12	6.7E-11	3.2E-7
GO:0035556	intracellular signal transduction	33	8.64E-11	3.2E-7
GO:0051246	regulation of protein metabolic process	34	1.34E-10	3.53E-7
GO:0050730	regulation of peptidyl-tyrosine phosphorylation	13	1.6E-10	3.53E-7
GO:0008284	positive regulation of cell proliferation	23	1.91E-10	3.53E-7
GO:0032101	regulation of response to external stimulus	18	2.22E-10	3.53E-7
GO:0031325	positive regulation of cellular metabolic process	39	2.78E-10	3.68E-7
GO:0022612	gland morphogenesis	11	2.98E-10	3.68E-7
GO:0031401	positive regulation of protein modification process	24	4.06E-10	4.52E-7

Significant GO biological processes characterizing the network represented in Figure 1.

Footnotes

Acknowledgments

The authors wish to thank Dr. Alexandros Xynos for English manuscript editing. Special thanks to Prof. Cecilia Gelfi, Prof. Maria Carla Gilardi, and Prof. Cristina Messa for their helpful discussions.

Author Disclosure Statement

The authors disclose no commercial and competing financial interest. This work was supported by FIRB/MERIT (RBNE089KHH) and “Proteogenomica e Bioimaging in Medicina” project (n. DM45602).

References

Bai

, Luoh

. 2008. GRB-7 facilitates HER-2/Neu-mediated signal transduction and tumor formation. Carcinogenesis, 29:473–479.

Barginear

, John

, Budman

. 2012. Trastuzumab-DM1: A clinical update of the novel antibody-drug conjugate for HER2-overexpressing breast cancer. Mol Med 10.2119/molmed.2012.00302.

Baylin

, Ohm

. 2006. Epigenetic gene silencing in cancer mechanism for early oncogenic pathway addiction? Nat Rev Cancer, 6:107–116.

Bediaga

, Acha-Sagredo

, Guerra

et al. 2010. DNA methylation epigenotypes in breast cancer molecular subtypes. Breast Cancer Res, 12:R77.

Bentwich

, Avniel

, Karov

et al. 2005. Identification of hundreds of conserved and nonconserved human microRNAs. Nat Genet, 37:766–770.

Bloom

, Eschrich

, Zhou

, Coppola

, Yeatman

. 2007. Elucidation of a protein signature discriminating six common types of adenocarcinoma. Int J Cancer, 120:769–775.

Boone

, Bhosle

, Tilby

, Hartley

, Hochhauser

. 2009. Involvement of the HER2 pathway in repair of DNA damage produced by chemotherapeutic agents. Mol. Cancer Ther, 9:3015–3023.

Boyd

, Wu

, O'Brien

et al. 2008. Proteomic analysis of breast cancer molecular subtypes and biomarkers of response to targeted kinase inhibitors using reverse-phase protein microarrays. Mol Cancer Ther, 7:3695–3706.

Brouillard

, Fritsch

, Edelman

, Ollero

. 2008. Contribution of proteomics to the study of the role of cytokeratins in disease and physiopathology. Proteomics Clin Appl, 2:264–285.

10.

Castilla

MÁ

, Díaz-Martín

, Sarrió

et al. 2012. MicroRNA–200 family modulation in distinct breast cancer phenotypes. PLoS One, 7:e47709.

11.

Chang

, Liu

, Hsieh

et al. 2006. Identification of alpha-enolase as an autoantigen in lung cancer: Its overexpression is associated with clinical outcomes. Clin Cancer Res, 12:5746–5754.

12.

Chen

, Olopade

. 2008. MYC in breast tumor progression. Expert Rev Anticancer Ther, 8:1689–1698.

13.

Cheng

, Chang

, Geradts

et al. 2012. Amplification and high-level expression of heat shock protein 90 marks aggressive phenotypes of human epidermal growth factor receptor 2 negative breast cancer. Breast Cancer Res, 14:R62.

14.

Chin

, DeVries

, Fridlyand

et al. 2006. Genomic and transcriptional aberrations linked to breast cancer pathophysiologies. Cancer Cell, 10:529–541.

15.

Corver

, Middeldorp

, ter Haar

et al. 2008. Genome-wide allelic state analysis on flow-sorted tumor fractions provides an accurate measure of chromosomal aberrations. Cancer Res, 68:10333–10340.

16.

Dehqanzada

, Storrer

, Hueman

et al. 2007. Assessing serum cytokine profiles in breast cancer patients receiving a HER2/neu vaccine using Luminex technology. Oncol Rep, 17:687–694.

17.

Dvorkin-Gheva

, Hassell

. 2011. Hormone receptor and ERBB2 status in gene expression profiles of human breast tumor samples. PLoS One, 6:e26023.

18.

Eichhorn

, Gili

, Scaltriti

et al. 2008. Phosphatidylinositol 3-kinase hyperactivation results in lapatinib resistance that is reversed by the mTOR/phosphatidylinositol 3-kinase inhibitor NVP-BEZ235. Cancer Res, 68:9221–9230.

19.

Esquela-Kerscher

, Slack

. 2006. Oncomirs—microRNAs with a role in cancer. Nat Rev Cancer, 6:259–269.

20.

Esteller

. 2007. Cancer epigenomics: DNA methylomes and histone modification maps. Nat Rev Genet, 8:286–298.

21.

Esteva

, Guo

, Zhang

et al. 2010. PTEN, PIK3CA, p-AKT, and p-p70S6K status: Association with trastuzumab response and survival in patients with HER2-positive metastatic breast cancer. Am J Pathol, 177:1647–1656.

22.

Faltus

, Yuan

, Zimmer

et al. 2004. Silencing of the HER2/neu gene by siRNA inhibits proliferation and induces apoptosis in HER2/neu-overexpressing breast cancer cells. Neoplasia, 6:786–795.

23.

Fiegl

, Millinger

, Goebel

et al. 2006. Breast cancer DNA methylation profiles in cancer cells and tumor stroma: Association with HER-2/neu status in primary breast cancer. Cancer Res, 66:29–33.

24.

Fountzilas

, Valavanis

, Kotoula

et al. 2012. HER2 and TOP2A in high-risk early breast cancer patients treated with adjuvant epirubicin-based dose-dense sequential chemotherapy. J Transl Med, 10:10.

25.

Galiè

, Sorrentino

, Montani

et al. 2005. Mammary carcinoma provides highly tumorigenic and invasive reactive stromal cells. Carcinogenesis, 26:1868–1878.

26.

Gaudet

, Hodgson

, Eden

et al. 2003. Induction of tumors in mice by genomic hypomethylation. Science, 300:489–492.

27.

Goddard

, Aiello Bowles

, Spencer Feigelson

et al. 2012. Utilization of HER2 genetic testing in a multi-institutional observational study. Am J Manag Care, 18:704–712.

28.

Goncalves

, Bertucci

. 2011. Clinical application of proteomics in breast cancer: State of the art and perspectives. Med Princ Pract, 20:4–18.

29.

Gown

. 2008. Current issues in ER and HER2 testing by IHC in breast cancer. Mod Pathol, 21:S8–15.

30.

Gruver

. 2011. Molecular pathology of breast cancer: The journey from traditional practice toward embracing the complexity of a molecular classification. Arch Pathol Lab Med, 135:546–557.

31.

Hamrita

, Chahed

, Kabbage

et al. 2008. Identification of tumor antigens that elicit a humoral immune response in breast cancer patients' sera by serological proteome analysis (SERPA) Clin Chim Acta, 393:95–102.

32.

Holm

, Hegardt

, Staaf

et al. 2010. Molecular subtypes of breast cancer are associated with characteristic DNA methylation patterns. Breast Cancer Res, 12:R36.

33.

, Stern

, Ge

et al. 2009. Genetic alterations and oncogenic pathways associated with breast cancer subtypes. Mol Cancer Res, 7:511–522.

34.

, Fan

, Oh

et al. 2006. The molecular portraits of breast tumors are conserved across microarray platforms. BMC Genom, 7:96.

35.

Jerusalem

, Fasolo

, Dieras

et al. 2011. Phase I trial of oral mTOR inhibitor everolimus in combination with trastuzumab and vinorelbine in pretreated patients with HER2-overexpressing metastatic breast cancer. Breast Cancer Res Treat, 125:447–455.

36.

Jones

, Baylin

. 2007. The epigenomics of cancer. Cell, 128:683–692.

37.

Kennecke

, Yerushalmi

, Woods

et al. 2010. Metastatic behavior of breast cancer subtypes. J Clin Oncol, 28:3271–3277.

38.

Korkaya

, Paulson

, Iovino

, Wicha

. 2008. HER2 regulates the mammary stem/progenitor cell population driving tumorigenesis and invasion. Oncogene, 27:6120–6130.

39.

Kouzarides

. 2007. Chromatin modifications and their function. Cell, 128:693–705.

40.

Kumandan

, Mahadevan

, Chiu

, Delaney

, Zanetti

. 2012. Activation of the unfolded protein response bypasses trastuzumab-mediated inhibition of the PI-3K pathway. Cancer Lett, 11:014.

41.

Leary

, Lin

, Cummins

et al. 2008. Integrated analysis of homozygous deletions, focal amplifications, and sequence alterations in breast and colorectal cancers. Proc Natl Acad Sci USA, 105:16224–16229.

42.

Liu

, Voisin

, Bader

et al. 2012. Seventeen-gene signature from enriched Her2/Neu mammary tumor-initiating cells predicts clinical outcome for human HER2+:ERα− breast cancer. Proc Natl Acad Sci USA, 109:5832–5837.

43.

Lowery

, Miller

, RE

, Kerin

. 2008. MicroRNAs as prognostic indicators and therapeutic targets: Potential effect on breast cancer management. Clin Cancer Res, 14:360–365.

44.

, Clark

. 2012. Impact of microRNA regulation on variation in human gene expression. Genome Res., 22:1243–1254.

45.

, Yang

, Han

et al. 2012. Relationship between serum HER2 extracellular domain levels, tissue HER2 expression, and clinico-pathological parameters in early stage breast cancer. Chin Med J, 125:4104–4110.

46.

Marchiò

, Natrajan

, Shiu

et al. 2008. The genomic profile of HER2-amplified breast cancers: The influence of ER status. J Pathol, 216:399–407.

47.

Markou

, Strati

, Malamos

, Georgoulias

, Lianidou

. 2011. Molecular characterization of circulating tumor cells in breast cancer by a liquid bead array hybridization assay. Clin Chem, 57:421–430.

48.

Mattie

, Benz

, Bowers

et al. 2006. Optimized high-throughput microRNA expression profiling provides novel biomarker assessment of clinical prostate and breast cancer biopsies. Mol Cancer, 19:5–24.

49.

Mazouni

, Baggerly

, Hawke

et al. 2010. Evaluation of changes in serum protein profiles during neoadjuvant chemotherapy in HER2-positive breast cancer using an LC-MALDI-TOF/MS procedure. Proteomics, 10:3525–3532.

50.

Minafra

, Norata

, Bravatà

et al. 2012. Unmasking epithelial-mesenchymal transition in a breast cancer primary culture: A study report. BMC Res. Notes, 5:343.

51.

Modi

, Stopeck

, Linden

et al. 2011. HSP90 inhibition is effective in breast cancer: A phase II trial of tanespimycin (17-AAG) plus trastuzumab in patients with HER2-positive metastatic breast cancer progressing on tastuzumab. Clin Cancer Res, 17:5132–5139.

52.

Mojtahedi

, Safaei

, Yousefi

, Ghaderi

. 2011. Immunoproteomics of HER2-positive and HER2-negative breast cancer patients with positive lymph nodes. OMICS, 15:409–418.

53.

Oliveras-Ferraros

, Vazquez-Martin

, Martin-Castilló

et al. 2010. Pathway-focused proteomic signatures in HER2-overexpressing breast cancer with a basal-like phenotype: New insights into de novo resistance to trastuzumab (Herceptin) Int J Oncol, 37:669–678.

54.

Pauletti

, Godolphin

, Press

, Slamon

. 1996. Detection and quantitation of HER-2/neu gene amplification in human breast cancer archival material using fluorescence in situ hybridization. Oncogene, 13:63–72.

55.

Perez

, Jenkins

, Dueck

et al. 2011. C-MYC alterations and association with patient outcome in early-stage HER2-positive breast cancer from the north central cancer treatment group N9831 adjuvant trastuzumab trial. J Clin Oncol, 29:651–659.

56.

Perou

, Sørlie

, Eisen

et al. 2000. Molecular portraits of human breast tumors. Nature, 406:747–752.

57.

Rakha

, El-Sayed

, Green

, Lee

. 2007. Prognostic markers in triple-negative breast cancer. Cancer, 109:25–32.

58.

Reis-Filho

, Pusztai

. 2011. Gene expression profiling in breast cancer: Classification, prognostication, and prediction. Lancet, 378:1812–1823.

59.

Rodriguez

, Frigola

, Vendrell

et al. 2006. Chromosomal instability correlates with genome-wide DNA demethylation in human primary colorectal cancers. Cancer Res, 66:8462–8468.

60.

Schulz

, Böllner

, Thomas

et al. 2009. Identification of differentially expressed proteins in triple-negative breast carcinomas using DIGE and mass spectrometry. Proteome Res, 8:3430–3438.

61.

Scott

, Goga

, Bhaumik

, Berger

, Sullivan

, Benz

. 2007. Coordinate suppression of ERBB2 and ERBB3 by enforced expression of micro-RNA miR-125a or miR-125b. J Biol Chem, 282:1479–1486.

62.

Shi

, Harris

, Lu

et al. 2006. A declining plasma fibrinogen alpha fragment identifies HER2-positive breast cancer patients and reverts to normal levels after surgery. J Proteome Res, 5:2947–2955.

63.

Sircoulomb

, Bekhouche

, Finetti

et al. 2010. Genome profiling of ERBB2-amplified breast cancers. BMC Cancer, 10:539.

64.

Slamon

, Press

. 2009. Alterations in the TOP2A and HER2 genes: Association with adjuvant anthracycline sensitivity in human breast cancers. J Natl Cancer Inst, 101:615–618.

65.

Smith

, Reardon

, Olson

. 2012. Pertuzumab for the treatment of patients with previously untreated HER2-positive metastatic breast cancer. Drugs Today (Barc), 48:713–722.

66.

Sørlie

, Perou

, Tibshirani

et al. 2001. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA, 98:10869–10874.

67.

Sotiriou

, Piccart

. 2007. Taking gene-expression profiling to the clinic: When will molecular signatures become relevant to patient care? Nat Rev Cancer, 7:545–553.

68.

Staaf

, Jönsson

, Ringnér

, Baldetorp

, Borg

. 2011. Landscape of somatic allelic imbalances and copy number alterations in HER2-amplified breast cancer. Breast Cancer Res, 13:R129.

69.

Staaf

, Jönsson

, Ringnér

et al. 2010. High-resolution genomic and expression analyses of copy number alterations in HER2-amplified breast cancer. Breast Cancer Res, 12:R25.

70.

St-Germain

, Gagnon

, Parent

, Asselin

. 2004. Regulation of COX-2 protein expression by Akt in endometrial cancer cells is mediated through NFκB/IκB pathway. Mol Cancer, 3:7.

71.

Streckfus

, Arreola

, Edwards

, Bigler

. 2012. Salivary protein profiles among HER2/neu-receptor-positive and -negative breast cancer patients: Support for using salivary protein profiles for modeling breast cancer progression. J Oncol, 2012:413256.

72.

Surani

, Hayashi

, Hajkova

. 2007. Genetic and epigenetic regulators of pluripotency. Cell, 128:747–762.

73.

Szklarczyk

, Franceschini

, Kuhn

et al. 2011. The STRING database in 2011: Functional interaction networks of proteins, globally integrated and scored. Nucleic Acids Res, 39:D561–568.

74.

Terada

, Okochi-Takada

, Akashi-Tanaka

et al. 2009. Association between frequent CpG island methylation and HER2 amplification in human breast cancers. Carcinogenesis, 30:466–479.

75.

Theillet

. 2010. What do we learn from HER2-positive breast cancer genomic profiles? Breast Cancer Res, 12:107.

76.

Tubbs

, Pettay

, Roche

, Stoler

, Jenkins

, Grogan

. 2001. Discrepancies in clinical laboratory testing of eligibility for trastuzumab therapy: Apparent immunohistochemical false-positives do not get the message. J Clin Oncol, 19:2714–2721.

77.

Vazquez-Martin

, Colomer

, Menendez

. 2008. Her-2/neu-induced “cytokine signature” in breast cancer. Adv Exp Med Biol, 617:311–319.

78.

Voduc

, Cheang

, Tyldesley

, Gelmon

, Nielsen

, Kennecke

. 2010. Breast cancer subtypes and the risk of local and regional relapse. J Clin Oncol, 28:1684–1691.

79.

Wang

, Hossein Saboorian

, Frenkel

et al. 2002. Aneusomy 17 in breast cancer: Its role in HER-2/neu protein expression and implication for clinical assessment of HER-2/neu status. Mod Pathol, 15:137–145.

80.

Wee

EJH

, Peters

, Nair

et al. 2012. Mapping the regulatory sequences controlling 93 breast cancer-associated miRNA genes leads to the identification of two functional promoters of the Hsa-mir-200b cluster, methylation of which is associated with metastasis or hormone receptor status in advanced breast. Oncogene, 20:82–95.

81.

Wirapati

, Sotiriou

, Kunkel

et al. 2008. Meta-analysis of gene expression profiles in breast cancer: Toward a unified understanding of breast cancer subtyping and prognosis signatures. Breast Cancer Res, 10:R65.

82.

Wolff

, Hammond

, Schwartz

et al. 2007. 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, 131:18–43.

83.

, Shang

, Sarkissyan

, Slamon

, Vadgama

. 2010. FOXO1A is a target for HER2-overexpressing breast tumors. Cancer Res, 70:5475–5485.

84.

Zhang

, Tai

, Wong

, Chiu

, Sethi

, Koay

. 2005. Proteomic study reveals that proteins involved in metabolic and detoxification pathways are highly expressed in HER-2/neu-positive breast cancer. Mol Cell Proteomics, 4:1686–1696.