Background
The progesterone receptor (PgR) is a downstream relative of the estrogen receptor (ER), which activates the expression of PgR via the estrogen-responsive element located in the promoter region of the PgR gene. Adequate expression of PgR indicates a functional ER-α and ER-α pathway [
1]. Loss of PgR expression in ER-positive (ER+) breast cancer potentially defines a subgroup with impaired function in the ER pathway, which probably gains limited benefit from endocrine therapy [
2‐
4].
Clinically, it is generally agreed upon that all newly-diagnosed primary breast cancers should be evaluated for ER and PgR protein expression by immunohistochemistry (IHC). Although some researchers have suggested that the ER-negative/PgR-positive (ER−/PgR+) phenotype does not actually exist and may represent technical artifacts [
5‐
7], an increasing body of evidence has shown that ER−/PgR+ tumors exist both biologically and clinically [
8,
9]. Moreover, an ER−/PgR+ breast cancer cell line had been described earlier [
10], indicating a mechanism of PgR expression regulation independent from ER-α.
The recently updated St. Gallen consensus on early-stage breast cancer recommends making clinical treatment decisions based on the surrogates of molecular subtypes (luminal-A, luminal-B, HER2-positive, and basal-like) defined by ER, PgR, HER2, and Ki67 [
11]. The St. Gallen panelists failed to categorize the ER−/PgR+/HER2− phenotype into the four molecular subtypes, while some other guidelines treated the ER−/PgR+/HER2− phenotype as a luminal-B subtype. The ER−/PgR+ group accounts for 1 − 5 % of all breast cancers [
2,
8]. Even after repeated reassessment of ER and PgR in these cases as the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) guidelines recommend [
12], at least 50 % of ER−/PgR+ remained [
2,
5].
Some efforts have been made to reveal the molecular essence of ER−/PgR+ breast cancer. Using gene-expression profile information, Itoh et al. [
13] proposed that, among these patients, 20 % were luminal-like and 65 % were basal-like, indicating for the first time that ER−/PgR+ breast cancer is a mixed group. In the current study, we included four large cohorts of breast cancer cases and systemically studied the clinical features and molecular essence of the ER−/PgR+ phenotype. Furthermore, we established a feasible and reliable IHC-based method to determine the subtype of each ER−/PgR+ case to guide individualized treatment. Because HER2+ breast cancers represent a biologically distinct subgroup [
14], we excluded HER2+ cases from this study.
Discussion
In the present study, we systemically investigated the clinicopathologic features and molecular essence of a clinically rare but biologically occurring ER–/PgR+/HER2– phenotype. We revealed that the clinicopathologic features and survival outcomes of this phenotype fell in between ER+/PgR+ and ER–/PgR– and were more similar to the ER–/PgR– phenotype. For the intrinsic subtype of ER–/PgR+ tumors, about 30 % were luminal-like and 60 % were basal-like. Moreover, we developed a feasible IHC-based method using three markers, TFF1, CK5, and EGFR, to determine the prognosis-relevant subtype of each ER–/PgR+ case, which may assist oncologists in making treatment decisions. ER–/PgR+ cases with basal-like characteristics may eliminate long-term but ineffective endocrine therapy and lead to individualized chemotherapy.
In our series, the majority of ER–/PgR+ tumors occurred in younger women with poorly differentiated tumors, which have been observed in triple-negative cases [
43]. At the molecular level, about 60 % were associated with a basal-like subtype, while only less than 30 % showed luminal features. Currently, routine clinical evaluation of subtype is most valuable in predicting the response to targeted therapy. Clinical guidelines, such as the St. Gallen consensus, recommend tailoring adjuvant systemic treatment according to subtypes [
11]. However, the ER–/PgR+ phenotype is not mentioned in the 2013 St. Gallen consensus, and the recommended treatment is therefore undetermined. Although the ER–/PgR+ phenotype belongs to the “hormone receptor-positive” group and is suggested to use endocrine therapy, its response to endocrine therapy is low [
3,
4,
12]. According to a collaborative meta-analysis of individual patient data from 20 trials (n = 21,457) in early breast cancer of about 5 years of tamoxifen versus no adjuvant tamoxifen by Early Breast Cancer Trialists’ Collaborative Group [
4], the rate ratios were 0.63 (standard error, 0.03) for ER+/PgR+ disease (
P <0.00001) but 0.90 (standard error, 0.10) for ER–/PgR+ disease (
P = 0.35). We herein demonstrated that the majority of ER–/PgR+ cases were actually basal-like, therefore indicating that treatment of ER–/PgR+ cases with long-term endocrine therapy for 5 years or even more is questionable.
Being able to identify the luminal-like subgroup within the ER–/PgR+ phenotype is important. Our study provides, for the first time, an effective and feasible IHC method to distinguish the intrinsic subtype within the ER–/PgR+ phenotype using three markers, TFF1, CK5, and EGFR. TFF1 is an indicator of the functional estrogen-responsive pathway and improves the response to tamoxifen [
44]. KRT5 and EGFR are identified as reliable basal markers [
25]. Moreover, we identified a significant difference in the sensitivity to endocrine therapy between luminal-like ER–/PgR+ and basal-like ER–/PgR+. Basal-like ER–/PgR+ cases obtained limited benefit from endocrine therapy, while luminal-like ER–/PgR+ cases probably benefited from endocrine therapy despite of ER loss. There are some potential explanations for this. First, in these cases, ER-negativity is falsely negative [
5]. Technical failure in ER detection made it difficult to detect positive ER even after re-assessment by IHC. Second, strong evidence exists for the presence of plasma membrane ER (only nuclear staining of ER is recognized as ER-positivity according to the ASCO/CAP guideline [
12]). When estrogen binds cell surface ER, membrane-initiated stimulation is able to induce and potentiate the genomic activation of PgR expression [
45,
46]. In this situation, endocrine therapy by antagonizing or reducing estrogen may also work.
Our study has some limitations. First, we excluded HER2+ cases and thus our findings could not be applicable in the ER–/PgR+/HER2+ phenotype. Second, although it is better to use the same survival endpoint (BCSS, RFS, or DRFS) for analysis, unfortunately, the various cohorts provide different endpoints and it is impossible to use the same endpoint for analysis. Third, because of limited ER–/PgR+/HER2– cases included in analysis of sensitivity to endocrine therapy and rare survival events, it is still too early to conclude the causal association between basal-like ER–/PgR+ tumors and limited benefit from endocrine therapy. Finally, our study is biased by its retrospective nature. However, due to the very low incidence of the ER–/PgR+/HER2– phenotype, it is impractical to conduct a large-scale prospective trial to test our hypothesis; we therefore must rely on data from the present large retrospective study. Our study uses the data from some prospective observational cohorts and provides a piece of state-of-the-art evidence describing the molecular essence of ER–/PgR+ and how to recognize the subtype of a ER–/PgR+ case using an IHC assay.
Acknowledgements
This work was supported by National Natural Science Foundation of China (81001169, 81370075), Training Plan of Excellent Talents in Shanghai Municipality Health System (XYQ2013101), Research and Innovation Project of Shanghai Municipal Education Commission, Shanghai International Science and Technique Cooperation Foundation (12410707700), International S&T Cooperation Program of China (ISTCP No. 09), and the Shanghai Key Laboratory of Breast Cancer (12DZ2260100). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interest
The authors have declared that no competing interests exist.
Authors’ contributions
Conceived and designed: KDY and ZMS. Performed the research: KDY and YZJ. Analyzed the data: KDY, YZJ, SH. Contributed reagents/materials/analysis tools: KDY, YZJ, SH, and ZMS. Wrote the paper: KDY, YZJ, SH, and ZMS. All authors read and approved the final manuscript.