Human epidermal growth factor receptor 2 (HER2) is amplified or over-expressed in around 20% of breast cancer cases, and the amplification of HER2 is usually associated with worse prognosis [1–3]. Trastuzumab, a humanized monoclonal antibody, was the first drug developed to target HER2 amplified breast cancer. The addition of trastuzumab to cytotoxic chemotherapy showed significant improved time to progression, overall response rate, response duration and overall survival (OS) in advanced HER2-positive breast cancer (HER2+), resulting in FDA approval of the drug in 1998 [4]. In 2006, FDA approval was extended to use of the drug in combination with chemotherapy in the adjuvant setting in early stage HER2 positive breast cancer [5, 6]. There are several proposed mechanisms of action. Some studies suggest that the drug disrupts ligand-independent transmembrane signaling triggered by the formation of HER2:HER2 homodimers or HER2:HER3 heterodimers, thereby diminishing Akt pathway activation, which ultimately leads to cell apoptosis [7]. Other mechanisms of cytotoxicity, including activation of antibody-dependent cell-mediated cytotoxicity (ADCC) [8] or blocking the cleavage of HER2 extracellular domain [9] have also been described.

Pre-surgical or neoadjuvant chemotherapy is standard therapy for inflammatory and locally advanced breast cancer. The addition of trastuzumab to chemotherapy in the pre-surgical setting in HER2+ patients has been tested in several phase II studies [10–14], with pathological complete response (pathCR) rates ranging from 18% to 47%. A phase II-III randomized pre-surgical trial conducted by the M.D. Anderson Cancer Center shows significant improvement in the pathCR rate with the addition of trastuzumab to chemotherapy [15]. The NOAH (NeO-Adjuvant Herceptin) trial is a phase III trial that evaluated the addition of trastuzumab to anthracycline- and taxane -based chemotherapy for HER2-positive patients with locally advanced or inflammatory breast cancer. Again, the addition of trastuzumab resulted in an increased pathCR rate, which translated into improved event free survival and OS [16]. In the phase III GeparQuattro trial, evaluating the effect of the addition of capecitabine to epribubin/cyclophosphamide/docetaxel regime, 445 HER2+ patients among 1509 patients with operable or locally advanced tumors were also given trastuzumab. The pathCR rate in the HER2+ subset was 31.7% [17]. Taken together, these trials suggest that the addition of trastuzumab to pre-surgical chemotherapy significantly improves outcomes in HER2+ breast cancer patients.

Although the use of trastuzumab as part of the pre-surgical regimen for breast cancer has increased, a uniform clinical benefit of trastuzumab in combination with chemotherapy is not observed. The likelihood of achieving a pathCR in this cohort is higher in trastuzumab treated patients with hormone receptor- negative HER2+ cancers compared to hormone receptor-positive HER2+ cancers [18]. Other groups have used mRNA measurements to predict pathCR. For example Denkert and colleagues found that quantitative assessment of mRNA for ESR1 and HER2 can predict pathCR [19]. Still, there remains no uniformly accepted method to predict which HER2+ patients are more or less likely to achieve a pathCR. Currently the only standard diagnostic tests to support the addition of trastuzumab are IHC and FISH and neither predicts pathological complete response. As treatment options expand for treatment of HER2 driven tumors, increased assay predictability and specificity will be sought. Here, we seek to identify quantitative measurement of HER2 protein level that are able to specifically select the patients who will most benefit from a trastuzumab containing treatment regime, and those who are not likely to achieve the desired pathCR status.

BrUOG BR-211B (NCT00617942) was a pre-surgical trial for stage II-III HER2+ breast cancer patients led by the Brown University Oncology Group (BrUOG) and the Yale Cancer Center. Patients were treated with q3week carboplatin, weekly nab-paclitaxel and trastuzumab for 18 weeks. In this trial, research biopsies were collected before initiation of treatment and again after a brief "run-in" exposure to either trastuzumab or nab-paclitaxel before patients received their first doses of the entire treatment regimen, allowing the testing of two hypotheses: 1) that the level of HER2 or pHER2 is associated with the likelihood of achieving a pathological complete response; and 2) that short exposure "run-in" treatment with trastuzumab alters the expression of HER2 or pHER2. In this paper, we quantitatively measured the level of HER2 and pHER2 (pY1248) to address these two issues.

Few biomarker studies have been able to predict the clinical benefit of the addition of trastuzumab to chemotherapy in the pre-surgical setting, partially because the concept of adding trastuzumab is relatively recent. In addition, patients undergoing pre-surgical therapy typically have limited tissue available for biomarker studies. As a result, studies on pre-surgical therapy tissue are challenging. Furthermore, tissues from the initial core biopsies are small, and often partially or completely exhausted by the diagnostic process. Therefore the availability of tissue for biomarker research in the pre-surgical setting represents a challenge. In order to obtain more tissue, this study was designed to obtain tissue specifically for research by core biopsy at both the initial time point and 10 days after "run in" therapy. This design allowed assessment of biological changes after one dose pre-surgical treatment. While the study is not large, the dedicated tissue collection allowed us to demonstrate that the level of HER2 is higher in patients that respond, and that the levels do not change between the initial biopsy and the post-treatment biopsy.

Biomarker studies have found several biomarkers, such as HER2 level [23] and PI3Kinase Akt pathway activity [24, 25], that are associated with response to trastuzumab in adjuvant and metastatic settings. Consistent with that data, in this study, we were able to demonstrate that the higher level of quantitatively measured HER2 is associated with pathCR. However, we did not find an association between pHER2 (pY1248) level and outcome. Increased phospho-HER2 (pY1248) levels have been associated with worse prognosis [26, 27], but this data has been hard to confirm due to challenges associated with phospho-epitope stability. In human tumors, the association between the level of phospho-HER2 and outcome remains controversial [28, 29]. While our cohort is small, we find no evidence for an association between phospho-HER2 and response to trastuzumab in the pre-surgical setting.

Measurement of phospho-HER2 is challenging since phospho-epitopes are known to be substantially more susceptible to degradation as a function of time to fixation [30, 31]. Our recent work (Neumiester et al., in press) and that of others [32] has shown that rapid fixation, as can be achieved by core needle biopsy, is the most effective way to preserve phospho-epitopes. In this study, dedicated core needle biopsy tissues were fixed immediately to minimize artifacts associated with pre-analytic cold ischemic time. Although some previous studies predicted that phospho-HER2 should be associated with pathway activity and potentially with drug response, we believe the phospho-epitopes measured here were accurately determined and they do not support a relationship with response in the pre-surgical setting.

Overall, this work should be considered as a pilot or proof of concept study. Although limited by small sample size, our data suggests that careful measurement of HER2 levels could provide increased specificity in selection of patients for trastuzumab therapy. As the number of therapies that target the HER2 signaling pathway increases it may be important to select patients less likely to respond to trastuzumab so they can be studied with alternative or combination therapies. Future studies that quantify the levels of HER2 in the pre-surgical setting are needed to confirm this observation.