The overexpressed human epidermal growth factor receptor (HER) 2 receptor tyrosine kinase (RTK) is a useful therapeutic target in breast cancer patients. In pathological diagnostics, the HER2 expression level is routinely evaluated and an elevated receptor expression (scored 3+) is usually based on HER2 gene amplification, which can reliably be identified by in situ hybridisation 1. Without contraindication HER2 protein overexpression (or HER2 gene amplification) represents the worldwide accepted rationale for antibody-targeted therapy using trastuzumab (Herceptin™). Herceptin™ was approved in Europe in 2000 for the treatment of breast cancer patients suffering from metastatic and aggressive tumour growth that is indicative for a poor disease outcome 2. In fall 2006, Herceptin™ treatment was extended by the approval for adjuvant patient treatment 3. Statistically considered, antigen-specific targeting of tumour cells has brought significant clinical benefit to these patients, both in terms of overall survival (OS) and recurrence free survival (RFS) documented in many clinical trials 45. Although HER2 overexpression or HER2 gene amplification is a prerequisite for therapeutic antibody treatment, the extent of sensitivity to Herceptin™ therapy varies enormously and ranges from excellent individual responsiveness over acquired insensitivity to complete resistance from the outset 6789. Consequently, HER2 overexpression does not represent a reliable predictive marker for responsiveness to Herceptin™.

Numerous molecular biomarkers implicated in individual sensitivity to Herceptin™ treatment have been suggested, but have not been precisely and reproducibly defined to date. The panel of suggested molecules comprises proteins involved in intracellular survival signalling (PTEN) 10 and cell proliferation (p21, p27) 1112131415, as well as cell membrane located molecules featuring the capacity to directly interact with HER2 and thereby modulate lateral and transversal HER2 activity 16. Among them the HER2-related RTK family members HER1 (also known as epidermal growth factor receptor (EGFR)), HER3 and HER4 are in the focus of interest 17. Depending on their coexpression profile HER receptors act in concert on ligand binding and trigger transmembraneous signal transduction as a functional unit rather than individually and independently. The recruitment of intracellular signalling pathways is a result of intensively cross-talking receptors and well-defined three-dimensional intracellular activation domains 18. Hence a highly organised spatiotemporal coexpression of HER receptors plays a pivotal role in growth, development and differentiation both on the cellular and organ level in healthy organisms 1920. However, HER1 and/or HER2 overexpression, as well as atypical HER receptor coexpression profiles, causing receptor hyperactivation and enhanced mitogenic signalling, could have been frequently attributed to malignant cell growth 17.

Moreover, HER3 has been found to preferentially interact with HER2 21 and, although kinase deficient, to significantly enhance mitogenic signalling 22. As a consequence even antiproliferative effects mediated by Herceptin™ can be extenuated by HER3 coexpression 23. Remarkably, receptor coexpression does not necessarily result in hyperstimulation and hypermitogenic cellular activity. Signal modulation, fine tuning and control can also be achieved by extensive receptor communication and even antiproliferative and induction of differentiation responses on growth factor binding (i.e. heregulin) to HER4 have been previously described 2425.

Conceivably the tissue-specific coexpression of HER receptors has immediate impact on the cellular response to Herceptin™, which specifically binds to HER2. In a previous study we have experimentally shown that the cellular sensitivity to Herceptin™ treatment is directly contingent on the EGFR coexpression density and can be enhanced by EGFR downregulation 26. In another study undertaken in our laboratory, we have shown that amplified HER3 and HER4 gene copy numbers have additional prognostic impact on the course of breast cancer disease in patients both with and without HER2 gene amplification 27.

Here, in a series of 48 breast cancer patients with HER2 gene amplification we retrospectively investigated numerous additional biomolecular parameters to have potential impact on the course of disease under Herceptin™ treatment (prognostic value) and to be implicated in responsiveness to Herceptin™ treatment (predictive value). Therefore, we quantified the gene copy numbers of all HER receptors via fluorescent in situ hybridisation (FISH) and immunohistochemically investigated the coexpression profile of all HER receptors. In addition, the HER2 phosphorylation state, the ratio of native to truncated HER2, the expression of p27^kip1 and PTEN, and the presence of the proliferation associated antigen Ki-67 were studied. As inhibition of activated HER2 might be primarily responsible for Herceptin™ sensitivity rather than the total receptor targeting 28, we evaluated the pY1248 phosphorylation, which has been linked to the mitogenic ras/MAP kinase signalling pathway 29. Herceptin™ is targeted to the juxtamembrane domain of HER2 30, so the loss of binding site by receptor truncation could be suggestive for antibody resistance 31. Furthermore, the identification of proteolytically shed HER2 (sHER2/p95) in relation to total HER2 expression could be informative in terms of receptor activation and consequently the efficacy of Herceptin™ binding. Moreover PTEN, a negative regulator of PKB/Akt, and p27^kip1 both downstream targets of HER2, have been proposed to be involved in Herceptin™ resistance 1032.

The aim of this study was to cytogenetically and immunohistochemically identify molecular biomarkers within the context of HER receptor signalling, which affects the susceptibility of metastatic breast cancer to Herceptin™ treatment. Of all the markers investigated in this study, the presence of truncated and phosphorylated receptors appeared to have an impact by trend on the course of disease of Herceptin™-treated (HT) patients. All Kaplan-Meier and Cox regression analyses revealed a significant positive impact of HER4 (co-)expression on OS from the beginning of antibody therapy. Finally both HER4 expression and HER4 gene amplification emerged as independent prognostic markers in HT breast cancer patients. HER4 receptor expression is even supposed to provide predictive value because it apparently affects susceptibility to Herceptin™ treatment.

This study was approved by the Institutional Review Board of the University of Regensburg, Germany.

Screening of HER2 aberration (both on a protein or DNA level) as a prerequisite for Herceptin™ therapy does not allow the courses of disease or individual therapy responses to be predicted, so numerous biomarkers with potential impact on Herceptin™ responsiveness have moved into the focus of interest. Utilising immunohistochemistry and FISH, we retrospectively investigated the most likely parameters with immediate impact on HER2 activity and Herceptin™ responsiveness in a cohort of 48 HT breast cancer patients with HER2 amplification. Here, we present first-hand data comprising expression of all HER receptors as well as p27^Kip1 receptor phosphorylation and native vs. truncated HER2 staining.

Our study demonstrates for the first time that HER4 expression improves the outcome in a sub-collective of HER2-positive, HT breast cancer patients. First evidence indicating that, irrespective of therapy, HER4 coexpression has a positive impact on the outcome of breast cancer patients was published in 2003 35, indicating a potential anti-tumourigenic effect mediated by the fourth member of the HER receptor family discovered in 1993 25. From there on a significant reduction of the proliferation indices in HER4-positive invasive breast carcinomas was reported, suggesting HER4 itself has a functional anti-proliferative or even protective capacity 36. An association of HER4 expression and a favourable outcome in breast cancer patients has independently been described by Suo and colleagues 37 who found that HER4 coexpression obviously antagonises the effect of HER2 (over-)expression on the course of breast cancer disease.

Our observation that HER4 coexpression apparently favours the outcome of HER2-positive, HT breast cancer patients indicates that HER4 coexpression favours the HER2 targeting with Herceptin™. This suggests a non-autonomous role of HER4 but rather some degree of HER4/HER2 interplay what is in agreement with data provided by Sartor and colleagues 38. They transfected HER2-positive cancer cells to overexpress HER4, achieving a reduction in proliferation and an increase in apoptosis which is suggestive of HER4 slowing down HER2 signalling activity. Moreover, a differentiation inducing effect by heregulin treatment in HER4 expressing SUM44 and SUM102 breast cancer cells has been experimentally demonstrated. This observation appeared to be independent of HER2 expression, because HER2 elimination could not eradicate the HER4 dependent decrease in cell growth. However, we previously found the coexpressed EGFR to contribute to cell susceptibility to Herceptin™ (and pertuzumab) and thereby provided evidence for EGFR/HER2 interaction playing a pivotal role in anti-HER2 targeting 26 in which HER4 is also most likely to be involved in.

Although HER4 expression has been frequently attributed to low-grade tumours with low proliferation activity 39, the discussion on the impact of HER4 on breast cancer progression and outcome of disease appeared contradictory over the past couple of years. This debate can be attributed to the ambivalent function of HER4 representing either oncogenic or tumour-suppressing activity 40. Four differentially spliced HER4 isoforms have been identified but never been distinguished in descriptive studies addressing HER4 expression in primary cancer 41. The JM-a/CYT-1 and JM-a/CYT-2 isoforms can be proteolytically cleaved by tumour necrosis factor-α converting enzyme 4042 and subsequently by γ-secretase 43, whereas JM-b/CYT-1 and JM-b/CYT-2 variants represent non-cleavable counterparts 44 characterised by ligand-independent activity that promotes cancer cell growth 43. Commonly, only JM-a/CYT-1 and JM-a/CYT-2 variants are expressed in tumour tissues at the same time 45. The intracellular CYT isoforms differ by having (CYT-1) or not having (CYT-2) a cytoplasmic binding site for phosphoinositide 3-kinase. Upon cleavage of JM-a isoforms these intracellular released HER4 domains (4ICD) can be differentially translocated into intracellular compartments and, if cytoplasmically located, subsequently trigger anti-proliferative and pro-apoptotic signals 4647. In contrast, if translocated into the nucleus, they have the capacity to trigger pro-proliferative and survival promoting signalling 48495051 and accordingly a nuclear localisation of 4ICD has been associated to shortened survival of breast cancer patients 45. In contrast, a mitochondrial accumulation of 4ICD, which evidently shares structural and functional homology with BH3-only pro-apoptotic BCL-2 family members, has also been reported 46. This in turn results in a disturbance of mitochondrial membrane integrity. This process typically entails an apoptotic cell death characterised by mitochondrial permeabilisation, cytochrome-c efflux and caspase activation, and can explain the protective role of HER4.

The positive impact of HER4 on a patient's course of disease as shown in this study is not only suggestive of its differentiation promoting and potentially pro-apoptotic function but rather for a Herceptin™ augmenting effect. As HER2 is primarily phosphorylated on Herceptin™ treatment, as we have previously shown 5216, it might be conceivable that HER4 can be crossactivated and a subsequent release of 4ICD might possibly enhance Herceptin™/HER2 triggered signalling. HER4, if coexpressed to HER2, could synergistically extenuate proliferative and survival signalling, for example via downregulation of ERK1/2 and PKB/Akt activity, respectively 53. Numerous observations suggest that irrespective of any therapeutic treatment HER4 expression is associated with reduced tumour breast cancer aggressiveness, which is indicative of its prognostic impact 2735363738395154. Moreover, based on Kaplan-Meier analysis as illustrated in Figure 4, we here provide evidence that the outcome of HER2/HER4 double-positive patients is significantly better both under CT and HT compared with those patients who were HER2-positive but HER4 negative. The latter observation does not necessarily designate a predictive value of HER4 but is indicative of the positive effect of HER4 among patients characterised by a well defined therapeutic setting (either CT or HT treatment). Nevertheless, the molecular signalling background responsible for the higher efficiency of Herceptin™ treatment in HER2-positive/HER4-positive patients versus HER2-positive/HER4-negative patients needs to be experimentally elucidated in detail and the experimental design should allow for investigating the predictive relevance of HER4 in consideration of its oncogenic and tumour suppressing potential, respectively. Deciphering the molecular effect of HER4, which enhances therapy efficiency of Herceptin™ treatment, would affect the decision to undergo Herceptin™-related therapy but the presence of HER4 could be potentially therapeutically addressed in the future as well. It would not be advisable, for instance, to use novel anti-HER2 drugs, which would eradicate the therapy-enhancing effect of HER4. Alternatively one could consider therapeutically eliciting or intensifying the Herceptin™ enhancing effect of HER4.

The induction of the cyclin-dependent kinase inhibitor p27^Kip1 protein is considered as a key mechanism on anti-HER2 targeting using Herceptin™ by which at least six signalling targets and pathways might be modulated 12. Also, in this study HER2 overexpressing patients with upregulated p27^Kip1, benefit significantly from antibody therapy indicating a prognostic value of this protein in HT patients, which is in agreement with other reports 5556.

Patients respond more efficiently to Herceptin™ treatment if HER2-Tyr1248 56 is phosphorylated suggesting that Herceptin™ affects an activated rather than an inactive receptor. However, other sites may influence the response to Herceptin™ 57.

Even if no correlation between HER4 and ER status could be found in our study, the signalling crosstalk of HER receptors with the ERα needs to be functionally explored in more detail. 4ICD has been shown to potentially activate ERα and in turn to stimulate tumour cell proliferation 51. Accordingly, Tovey and colleagues showed a significantly poorer survival in ER-positive breast cancer patients by detecting nuclear HER4 localisation 58, highlighting the importance of 4ICD/ERα interplay.

Current approaches in developing novel anti-HER receptor targeted drugs that block the entire type 1 RTK family via 'pan-HER' blockade 59 may be counterproductive due to the ambivalent function of HER4. HER4 overexpression or amplification dependent on the prevalent isoform confers a reduced risk 2754. Hence development of selective type 1 RTK inhibitors leaving HER4 signalling unaffected may prove more advantageous. On the contrary, antibodies selectively targeting HER4 JM-a, such as the recently introduced monoclonal antibody 1479 60, could have therapeutic potential, assuming that JM-a-specific antibodies are expected not to cause adverse effects in tissues, such as the heart, that exclusively express ErbB4 JM-b.

Further studies addressed to elucidate HER4 receptor function in the context of Herceptin™ treatment are necessary to guide the design of highly efficient therapeutic strategies based on HER receptor targeting. Extending the data presented here by stratifying larger patient cohorts characterised by their HER2/HER4 coexpression profile will disclose the impact of HER4 on anti-HER2 targeted Herceptin™ treatment compared with conventional therapy of HER2-positive breast cancer patients. Potentially the evaluation of HER4 expression, by which valuable additional prognostic or even predictive information might possibly be revealed either from immunohistochemistry or FISH even without differentiating between HER4 splice variants, could contribute to therapy decisions and/or optimisation in the future.

CI: confidence interval; CT: conventionally treated; DAPI: 4',6-diamidino-2-phenylindole; EGFR: epidermal growth factor receptor; ER: oestrogen receptor; FISH: fluorescence in situ hybridisation; H&E: haematoxylin and eosin; HER: human epidermal growth factor receptor; HR: hazard ratio; HT: Herceptin™ treated; IRS: immunoreactive score; OS: overall survival; PR: progesterone receptor; RFS: recurrence free survival; RTK: receptor tyrosine kinase; TMA: tissue microarray.

The authors declare that they have no competing interests.

AS wrote the manuscript, performed the image analysis and FISH evaluation and analysed the statistical data. MS helped to analyse the statistical data and performed FISH evaluation. SD-D helped to write the manuscript. OO performed the histological analysis and interpreted the data. FH provided material to be analysed and performed primary tissue-based diagnostics. SS performed the histological analysis and interpreted the data. GB is the research group leader, evaluated the data, and finally approved the manuscript to be published. All authors read and approved the final manuscript.