Comparison of the HER2, estrogen and progesterone receptor expression profile of primary tumor, metastases and circulating tumor cells in metastatic breast cancer patients

In primary and metastatic breast cancer (MBC), tumors are usually analyzed for the presence or absence of the estrogen receptor (ER), the progesterone receptor (PR), and for amplification of HER-2, and the results of these analyses direct the types of treatment that patients receive. In MBC, patients may be treated with systemic therapy (chemotherapy, biological therapy, targeted therapy, hormonal therapy), local therapy (surgery, radiation therapy), or a combination of these treatments. The choice of treatment generally depends on the characteristics of the primary tumor because metastatic tissue is often difficult to obtain. Notably, HER-2 as well as ER/PR were shown to be differentially expressed between the primary tumor and corresponding metastases in up to 48 % which might lead to ineffective treatment in the absence of the respective marker [1‐6]. Therefore, reassessment of these markers at the time of disease progression might help to optimize treatment decisions. Although biopsies from most metastatic sites may be obtained by the use of imaging and interventional radiology on a routine basis, these techniques are invasive and may pose some discomfort or may result in complications. Thus, a blood based biomarker would be desirable to bypass these problems.

In this regard, circulating tumor cells (CTCs) would be an ideal ‘surrogate tissue’ to identify prognostic and predictive factors that will help in selecting the optimal therapeutic strategy for each individual patient in case that metastatic tissue is not available.

Our study group has already demonstrated that ER and PR were differentially expressed between primary tumor and CTCs in MBC [7]. It was the purpose of the present study to compare the HER2/ER/PR expression profile of primary tumor and metastases, primary tumor and CTCs as well as metastases and CTCs. To our knowledge, it is the first study comparing histopathological and molecular findings between primary tumor, metastases and CTCs.

In MBC, the choice of therapy generally depends on the size, location, and number of metastatic sites whereas the decision to administer antihormonal- and/or HER2-targeted therapy depends on the expression of these markers on the primary tumor since metastatic tissue is often difficult to obtain. However, several BC studies have indicated that the expression of HER2, ER and PR can change during course of disease [1‐6, 12‐30]. Therefore, reassessment of the predictive markers at the time of disease progression might help to optimize treatment decisions. In this context, characterization of CTCs could be of relevance in the future.

Here we demonstrate that the molecular detection of HER2 overexpression in CTCs using the AdnaTest BreastCancer is able to significantly predict the HER2 status on metastases. However, for ER/PR, a more detailed analysis of expression rates in tissue samples will be necessary to decide whether to use CTCs as a useful tool for treatment decisions. Interestingly, in contrast to some already published studies [1‐6, 12‐30], we could show that primary tumors and their metastases showed a highly significant concordance of the expression of predictive markers. Furthermore, ER- and PR-positive metastases significantly were found to be of visceral origin whereas no trend was seen for HER2 which has to be discussed in more detail.

In a prospective study, Simmons et al. demonstrated a discordant ER/PR and HER2 status in 40 % and 8 % of cases in more than half of the 40 patients analyzed who presented with new lesions suspicious for MBC. Consequently, therapeutic intervention was changed accordingly in 20 % of the patients who agreed to undergo biopsy not only to confirm their metastatic phenotype but also to have reassurance of receiving “targeted therapy” [3]. Another MBC study including 25 patients with liver metastases observed a discordant ER, PR and HER2 receptor status in 14.5 %, 48.6 % and 13.9 % of cases, respectively, which led to change in therapy in 12.1 % of patients [2]. With regard to HER2, subsequently published studies have reported discordant rates from 1 % to 24 % between primary tumor and metastases [1, 2, 12‐29] and a study-level meta-analysis including 26 trials and about 2.500 patients, found a discordance rate for either HER2 loss or gain of 5.5 % [30, 31]. These findings are quite in opposite to our study, showing a significantly high concordance for these markers when comparing primary tumor and metastases.

A discordant expression of these receptors and the primary tumor and corresponding metastases and/or CTCs has already been demonstrated with a discordance between primary BC and HER2 expression on CTCs in the setting of disease recurrence at variable rates, with a gain of HER2 from 9 % to over 60 % in different studies [32‐39]. The fact that ER and PR were differentially expressed between primary tumor and CTCs confirms the results of our previously published study demonstrating a loss of receptor expression on CTCs when compared with the expression on primary tumors [7]. These results can now be extended and confirmed for metastases and CTCs. However, the concordance for biomarker negativity seems to be higher in this study, although the number of these cases is quite small. In fact, the possibility of changes in receptor status during the course of tumor progression in triple-negative BC is very low, up to 8 %, despite changes in receptor positive BC with up to 40 % [40]. One could speculate that escape from antitumor therapy is more effective for CTCs when losing ER and PR on the surface.

In addition, the fact that ER- and PR-positive metastases significantly were found to be of visceral origin with a positive trend also documented in primary disease allows the hypothesis that CTC release as well as their downregulation of hormonal receptors might be recognized as a resistance mechanism to adjuvant endocrine therapy. As a consequence, CTCs released under therapy downregulate the therapeutic target during their phase of epithelial-mesenchymal transition (EMT) but recover ER/PR overexpression during the course of metastasis. This could fairly explain how hormone receptor positive visceral metastasis appear in significant concordance to the metastatic hormonal phenotype but also still seem to be positively correlated with the primary lesion. These results were not found for non-visceral metastases and we can only speculate that probably the different environment might influence the rate of receptor expression.

In our study, applying the AdnaTest BreastCancer, metastases and CTCs displayed a significantly concordant HER2 status in 67 % of cases whereas no significant concordance values could be shown for ER and PR. In contrast, applying the CellSearch® assay, no significant concordance was found when HER2 status on CTCs was compared to primary tumors and metastases. These findings might be explained by the different selection strategies of both assays. CellSearch® as an EpCAM-dependent assay might not detect CTCs that lost the EpCAM epitope and, therefore might result in false negatives with regards to HER2 overexpressing cells [41]. In contrast, the AdnaTest CTC enrichment method consists of an antibody mixture targeting EpCAM and MUC1, which might enable efficient CTC enrichment even in case EpCAM got lost.

However, CTCs are highly heterogeneous and using EpCAM-based capturing methods, it has been shown that this procedure is not able to detect the entire, highly heterogenous population of CTCs in MBC. In this regard, it has been demonstrated that these methods underestimate the most important subpopulations of CTCs involved in cancer dissemination, which often share EMT and stemness features [42‐45]. In the current study, these subpopulations have not been analyzed which might explain discordant findings. Thus, despite the prognostic impact of CTC counts, molecular methods might complement these studies by improving the overall detection rate as well as sensitivity and, thus, permitting the assessment of genomic markers in CTCs of MBC patients as recently published [46].

From the clinical point of view, in a recent review of the literature, Turner and Di Leo concluded that the best management approach for receptor discordance between primary and metastatic disease is currently unknown, and the very limited evidence of alteration in clinical outcomes based on repeated biopsy does not seem to be strong enough to confirm that repeated biopsy is essential in every patient [47].

However, although these discrepancies have been demonstrated, the acquisition of tissue from metastases is not recommended as routine practice in any guideline. Thus, monitoring and phenotypic characterization of CTCs can provide new insights into the clonal selection of tumor cells under palliative therapies which may allow physicians to follow cancer changes over time and tailor treatment accordingly.