Introduction
The development and progression of breast cancer are influenced by steroid hormones, particularly estrogen, via their interaction with specific target receptors. Endocrine therapy has become the most important treatment option for women with estrogen receptor (ER)-positive breast cancer. Nevertheless, many breast cancer patients with tumors expressing high levels of ER are unresponsive to endocrine therapy, and all patients with advanced disease eventually develop resistance to the therapy. The potential mechanisms behind either this intrinsic or acquired endocrine resistance involve ER-coregulatory proteins and crosstalk between the ER pathway and other growth factor signaling networks [
1,
2]. An understanding of the molecular mechanisms that modulate the activity of the estrogen signaling network has enabled new ways of overcoming endocrine resistance to be developed.
ER-α is phosphorylated on multiple amino acid residues [
3]. Serines 104, 106, 118, and 167 are all located within the activation function (AF)1 region of ER-α, and their phosphorylation provides the important mechanism that regulates AF1 activity [
4,
5]. In response to estradiol binding, human ER-α is phosphorylated mainly on Ser118 and to a lesser extent on Ser104 and Ser106 [
4]. Although some authors have also reported that Ser167 is a major estradiol-induced phosphorylation site [
5,
6], this response to estradiol has not been universally observed [
4,
7]. Interestingly, in response to the activation of the mitogen-activated protein kinase (MAPK) pathway, phosphorylation occurs on Ser118 and Ser167 [
8,
9]. However, the role of phosphorylation of Ser118 and Ser167 of ER-α in human breast cancer has not been investigated.
ER-β and its splicing isoforms are widely expressed in both normal and malignant breast tissue [
10]. Although several groups have reported results regarding the possible function of ER-β, and its potential as a prognostic or predictive factor in breast cancer, the data remain inconclusive and are often contradictory [
11,
12]. ER-βcx (also called ER-β2), a splice variant of ER-β, is considered to be a dominant repressor of ER-α; it is identical to ER-β1 (wild-type ER-β) except that the last exon, 8, is replaced by 26 amino acid residues [
13]. The role of ER-β and its isoforms, especially with respect to the response of breast cancer to endocrine therapy, has also not been elucidated.
Progesterone receptors (PRs) occur as two isoforms, PRA and PRB, transcribed from two distinct promoters on a single gene. PRA, but not PRB, lacks the 164 amino acid N-terminal residues that contain AF3, and this is the cause of their functional differences [
14]. In the mammary gland, the overexpression of PRA relative to PRB results in extensive epithelial cell hyperplasia, excessive ductal branching, and a disorganized basement membrane, all features associated with neoplasia [
15]. In contrast, the overexpression of PRB leads to premature arrest of ductal growth and inadequate lobuloalveolar differentiation [
16]. However, little is known about the unique roles of the two PR isoforms in breast cancer.
In this study, we focused on the expression and phosphorylation of the hormone receptors themselves and, using immunohistochemistry (IHC), examined the phosphorylation of ER-α Ser118 and Ser167 and the expression of ER-α, ER-β1, ER-βcx/β2, PR, PRA, and PRB in primary breast tumor specimens from 75 patients with metastatic breast cancer who received first-line treatment with endocrine therapy on relapse. Our results show that patients with primary breast tumors in which there is either high phosphorylation of ER-α Ser167 or high expression of ER-α, PR, PRA, or PRB significantly responded to endocrine therapy and had a better survival after relapse.
Discussion
Using IHC techniques, we investigated the phosphorylation of ER-α Ser118 and ER-α Ser167, and expression of ER-α, ER-β1, ER-βcx/β2, PR, PRA, and PRB, in primary breast tumor specimens from 75 patients with metastatic breast cancer who, on relapse, received endocrine therapy as first-line treatment. Our results indicate that patients with primary breast tumors with high phosphorylation of ER-α Ser167, or high expression of ER-α, PR, PRA, or PRB, significantly respond to endocrine therapy and had a better survival after relapse.
ER-α is phosphorylated on multiple amino acid residues [
3]. In general, phosphorylation of serine residues in the AF1 domain of ER-α appears to influence the recruitment of coactivators, resulting in enhanced ER-mediated transcription. In this study, we measured the phosphorylation of ER-α Ser118 and Ser167 as well as the expression of ER-α in breast cancer by IHC using site-specific anti-ER-α-phosphoserine antibodies. Our results showed that phosphorylation of ER-α Ser118, but not of ER-α Ser167, was significantly correlated with expression levels of HER2. It has been reported that ER-α was significantly phosphorylated on Ser118 in response to either estradiol binding or the activation of the mitogen-activated protein kinase (MAPK) pathway, while Ser167 is phosphorylated by AKT, p90 ribosomal S6 kinase (RSK), and casein kinase II as well as MAPK [
5,
7,
9,
24]. Murphy and colleagues recently reported that in 45 human breast tumor biopsies phosphorylation of ER-α Ser118 correlated with active MAPK [
25]. Because MAPK is located downstream of HER2, it is possible that phosphorylation of ER-α Ser118 is in part caused by HER2-MAPK signaling in breast cancer. On the other hand, phosphorylation of ER-α Ser167 seems to be led by different mechanisms.
Our results also showed that while phosphorylation of both ER-α Ser118 and Ser167 was strongly and positively correlated with expression of ER-β 1 and ER-βcx/β2, there was no observed association between expression of ER-α and ER-β proteins. Both antibodies for ER-β1 and ER-βcx/β2, generated in this study, were specific against their respective C-terminal amino acid residues, and positive nuclear staining was observed in normal breast epithelial cells, noninvasive ductal carcinoma, and invasive carcinoma. Saunders and colleagues also found that there was no quantitative relation between IHC scores for ER-α and ER-β [
26]. However, using IHC, it was reported that ER-β expression was positively correlated with ER-α and PR [
27]. Specific detection of ER-β1 from other isoforms also indicated a positive correlation between expression of ER-β1 and ER-α [
28]. However, no studies have been reported concerning the relation between phosphorylation of ER-α and expression of ER-β proteins.
In our analysis, ER-α expression was positively correlated with PRA but not with PRB. In addition, phosphorylation of both ER-α Ser118 and Ser167 was strongly and positively associated with expression of PRA but not with PRB. This suggests that PRA is preferentially induced following the transcription of ER-α after the phosphorylation of Ser118 and/or Ser167. Two previous studies have reported investigations into the expression of PRA and PRB in breast cancer. The first, an analysis of 202 PR-positive breast cancers by immunoblotting, showed that while there was no significant difference in levels of PRA and PRB in most of the PR-positive tumors, nevertheless expression levels of PRA were higher in 59% of tumors and at least four times as high in 25% [
29]. In the second study, of 32 PR-positive breast cancers, it was reported that excess PRB correlated with the absence of HER2, thereby indicating a good prognosis, whereas excess PRA correlated with a poorly differentiated phenotype and higher tumor grade [
30]. The normally equal expression of PRA and PRB is disrupted early in carcinogenesis. PRA is usually the predominant isoform in tumors of the breast, and it appears, therefore, that disrupted progesterone signaling may play a role in the development or progression of these cancers [
14,
29,
31].
The most important results to come out of this study concern the correlation between clinical response and the phosphorylation and expression of the receptors. We identified that patients with primary breast tumors with high phosphorylation of ER-α Ser167, or high expression of PRA or PRB, significantly responded to endocrine therapy, whereas phosphorylation of ER-α Ser118 and expression of ER-β1 and ER-βcx/β2 did not influence response. Phosphorylation of both ER-α Ser118 and ER-α Ser167 occurs in response to either estradiol binding or activation of growth factor signaling pathways. It is well established that peptide growth factor signaling pathways can activate ER-α, in the absence of its ligand, through phosphorylation of ER-α by MAPK [
8,
32]. In addition, the induction of ER-α by MAPK also enhances ER signaling and promotes tumor growth in the presence of estradiol, and such tumors have been shown to be responsive still to antiestrogen therapy [
33]. Furthermore, Clark and colleagues reported that, independently of MAPK, p90 ribosomal S6 kinase 2 (Rsk2) specifically activates ER-α by phosphorylation of Ser167 and by docking to the hormone-binding domain of ER-α, and that the antiestrogen 4-hydroxytamoxifen blocks Rsk2-mediated activation of ER-α [
7]. Since our results showed that phosphorylation of ER-α Ser167, but not ER-α Ser118, was predictive of response to endocrine therapy, they suggest that, in breast cancer, phosphorylation of ER-α Ser118 occurs frequently without estradiol, whereas phosphorylation of ER-α Ser167 may occur frequently in response to estradiol binding.
It has been reported that HER2-induced MAPK and ER-α activation leads to tamoxifen resistance [
34]. Data from these clinical trials demonstrated that the antiproliferative response to endocrine therapy was impeded in ER-α-positive/HER2-positive primary breast cancers [
35]. In contrast, a Southwest Oncology Group study reported that overexpression of HER2 was not associated with tamoxifen unresponsiveness or a more aggressive phenotype of ER-α-positive metastatic breast cancer [
36]. In our analysis, HER2-positive tumors showed high phosphorylation levels of ER-α Ser118 and were resistant to endocrine therapy.
Finally, our results showed that expression of ER-β1 and ER-βcx/β2 does correlate with response to endocrine therapy. No significant differences in the expression of ER-β1, ER-β2, and ER-β5 mRNAs between tamoxifen-sensitive and -resistant groups, has been reported [
37]. Taken together, these data suggest that the expression of ER-β proteins is not predictive of response to endocrine therapy in breast cancer. However, a significant correlation between a PR-negative phenotype and the presence of ER-βcx/β2 in ER-α-rich tumor foci and expression of ER-βcx/β2 with low PR expression has been shown to correlate with a poor response to tamoxifen [
38].
In our analysis, the expression of PRA and PRB as well as PR was strongly predictive of response to endocrine therapy. In contrast, it has been reported, in a study of T47D human breast tumor xenografts, that tamoxifen treatment preferentially inhibited the growth of PRA tumors, whereas PRB tumors were unaffected [
39]. Another study reported that, although estrogen induced PR expression in all breast cancer cell lines studied, the expression ratio of PRA/PRB induced by estrogen was dependent on the cell line, and that these results suggested that the PRA and PRB promoters were differentially regulated by estrogen in different breast cancer cells [
40]. Further studies are obviously needed to resolve these apparent discrepancies and in order to identify the functional importance of altered PR isoform expression and how this might affect the response of breast tumors to endocrine therapy.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
HY conceived of the study and participated in its design, coordination, and manuscript writing. MN carried out immunostaining experiments. SK, YF, and HI participated in its design and coordination and helped to draft the manuscript. YA, HS, MH, and KM provided tissue samples. ZZ assessed the immunostaining. All authors read and approved the final manuscript.