Alternate Estrogen Receptors Promote Invasion of Inflammatory Breast Cancer Cells via Non-Genomic Signaling

Although Inflammatory Breast Cancer (IBC) is a rare and an aggressive type of locally advanced breast cancer with a generally worst prognosis, little work has been done in identifying the status of non-genomic signaling in the invasiveness of IBC. The present study was performed to explore the status of non-genomic signaling as affected by various estrogenic and anti-estrogenic agents in IBC cell lines SUM149 and SUM190. We have identified the presence of estrogen receptor α (ERα) variant, ERα36 in SUM149 and SUM190 cells. This variant as well as ERβ was present in a substantial concentration in IBC cells. The treatment with estradiol (E2), anti-estrogenic agents 4-hydroxytamoxifen and ICI 182780, ERβ specific ligand DPN and GPR30 agonist G1 led to a rapid activation of p-ERK1/2, suggesting the involvement of ERα36, ERβ and GPR30 in the non-genomic signaling pathway in these cells. We also found a substantial increase in the cell migration and invasiveness of SUM149 cells upon the treatment with these ligands. Both basal and ligand-induced migration and invasiveness of SUM149 cells were drastically reduced in the presence of MEK inhibitor U0126, implicating that the phosphorylation of ERK1/2 by MEK is involved in the observed motility and invasiveness of IBC cells. We also provide evidence for the upregulation of p-ERK1/2 through immunostaining in IBC patient samples. These findings suggest a role of non-genomic signaling through the activation of p-ERK1/2 in the hormonal dependence of IBC by a combination of estrogen receptors. These findings only explain the failure of traditional anti-estrogen therapies in ER-positive IBC which induces the non-genomic signaling, but also opens newer avenues for design of modified therapies targeting these estrogen receptors.


Introduction
Inflammatory Breast Cancer (IBC) is a rare and aggressive form of locally advanced breast cancer affecting approximately 1-6% of breast cancer patients in the United States [1]. It is reported by the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) that while the incidence of normal breast cancer has been steadily decreasing, the incidence of IBC continues to be increasing. SEER further reports that the overall survival rate of IBC patients is significantly lower than non-IBC stage III breast cancer [2]. The aggressive characteristic of IBC enables its malignant cells to invade the dermal lymphatics of the breast, causing the accumulation of fluids within lymphatic vessels and the subsequent edematous red swelling of the breast [3]. By its ability to metastasize rapidly, most IBC tumors are characterized as stage IIIB at the time of detection [4]. These characteristics coupled with the relatively poor prognosis makes IBC one of the most deadly carcinomas.
Since the term IBC was coined by Lee and Tannebaum in 1924, there has been an extensive search for an effective way to treat this disease. Although many groups have reported that the prognosis of the disease can be improved by multimodality treatment, improvement in the overall survival rate is still poor due to the lack of a major molecular signature [5][6][7]. Several studies have found that an alteration in the levels of molecules such as RhoC, WISP-3 and Caveolin contribute significantly to IBC progression, but an IBC-specific therapy still remains elusive [8][9][10][11].
IBC is predominantly negative for estrogen receptor a (ERa) and progesterone receptor (PR) and positive for human epidermal growth factor receptor 2 (HER2) [12]. The absence of ERa makes it difficult to treat IBC with traditional antiestrogens such as tamoxifen. In addition to the wild-type ERa, several groups have provided evidence for the existence of ERa splice variants. Of the splice variants, ERa36 has been relatively well studied. ERa36 is translated from a transcript initiated by a previously unidentified promoter in the first intron of ERa gene and lacks both AF-1 and AF-2 transactivation domains present in ERa [13]. ERa36 was expressed in both ERa positive and negative breast carcinomas and its expression was reduced with the presence of ERa [14]. Furthermore, ERa36 has been shown to be capable of supporting the non-genomic signaling pathways such as ERK and Akt in endometrial carcinomas and HEK293 cells [15,16]. The search for an alternate estrogen receptor led to an identification of a novel G-protein coupled receptor, GPR30 that localized in the plasma membrane, the cytoplasm and the endoplasmic reticulum [17][18][19]. GPR30 is responsible for its rapid activation of non-genomic signaling pathway and is involved in the biological processes including migration, proliferation, adhesion, gene regulation and invasion of the cancer cells [18][19][20][21]. Interestingly, the classical anti-estrogens such as tamoxifen and ICI 182780 are known to act as agonists for ERa36 and GPR30. Therefore, these alternate estrogen receptors are considered to contribute to the tamoxifen resistance in a variety of tumors [22,23].
Stimulation of the estrogenic non-genomic signaling components such as PI3K, Akt and ERK contributes to alterations in the downstream effectors in a cell-type-specific manner and causes rapid physiological responses in target tissues [24]. In spite of the immense importance of non-genomic signaling pathway in other tumors, little research has been done on its role in IBC. In the present study, we discovered the presence of ERa36 and ERb in IBC cells and also found an evidence for the possible contribution of these variants in the stimulation of p-ERK1/2 in response to treatment with agonists and antagonists. Furthermore, we found that this stimulation of p-ERK1/2 leaded to an enhanced migration and invasiveness of IBC cells.

Expression status of estrogen receptors in SUM149 and SUM190 cells
The full-length ERa was negative in the IBC cell lines SUM149 and SUM190 (Fig. 1A), while ERa alternative splicing variant ERa36 was substantially expressed in the both cell lines (Fig. 1A). ERb was detectable in both SUM149 and SUM190 cells. As expected, MDA-MB-231 and SKBR3 cells were positive and negative for ERb expression, respectively. GPR30 were also detected in both SUM149 and SUM190 as well as in GPR30positive cell lines MCF-7 and MDA-MB-468 (Fig.1A). Immunoflourescent labeling followed by visualization through scanning confocal microscopy revealed that while ERa36 was predominantly localized in the cytoplasmic compartment in both SUM149 and SUM190 cells, ERb was detected with low expression levels in the both cytoplasm and nucleus of these cell lines (Fig. 1B). GPR30 was present in both the cytoplasmic and nuclear compartment in these cells.

Rapid activation of p-ERK1/2 by estrogen receptor ligands in SUM149 and SUM190 cells
The localization of the estrogen receptors in the cytoplasmic compartment presented a possibility of the involvement of nongenomic signaling in these cells. To investigate this possibility, SUM149 and SUM190 cells were treated with E2 (10 nM) after maintained in 5% DCC medium for 48 hrs. We noticed a rapid activation of p-ERK1/2 within 5 min of post E2 treatment in SUM149 and SUM190 cells ( Fig. 2A). However, the phosphorylation of the other MAPKs JNK and p38 were hardly activated by E2 treatment in SUM149 cells (supplementary Fig. S1). In addition, Akt phosphorylation did not increased upon E2 treatment. Interestingly, we also found a rapid activation of p-ERK1/2 in these cells upon treatment with 10 nM of ERb specific ligand DPN and GPR30-specific agonist G1 (Fig. 2B, C). However, there was no p-ERK1/2 activation upon the treatment with 10 nM of ERa-specific ligand PPT (Fig. 2B).
The anti-estrogens 4OHT and ICI are known to inhibit the response of E2 mediated by ERa. Next, we tested whether these anti-estrogens could compromise the activation of p-ERK1/2 by E2 treatment. Interestingly, we found that the treatment of IBC cells with OHT or ICI, alone or in combination with E2 potentiated the magnitude of p-ERK1/2 stimulation by E2 (10 nM) or DPN (10 nM) (Fig. 3A). In addition, all of E2, 4OHT or DPN at 10 27 to 10 29 M activated p-ERK1/2 (Fig. 3B). Pretreatment of SUM149 cells with a specific MEK inhibitor U0126 blocked the activation of p-ERK1/2 in response to stimulation with E2, 4OHT, DPN or G1 in SUM149 cells (Fig. 4A). Confocal microscopic analysis showed that E2 treatment induced cytoskeletal reorganization in SUM149 cells as revealed by actin staining (Fig. 4B), indicating that the cell mobility might be positively regulated by E2, which could be effectively blocked by the pretreatment of MEK inhibitor U0126. Activation of p-ERK1/2 promotes motility and invasiveness of SUM149 cells Downstream signaling through the activation of p-ERK1/2 has been implicated in increasing the proliferation and motility in numerous cancer cell models. Therefore, we initially investigated whether E2 could stimulate IBC cell proliferation. The treatment with E2 at different doses (0.1-10 nM) had no effect on the cell proliferation ( Supplementary Fig. S2). We then performed a Boyden chamber migration and invasion assay. The results showed that the treatment with E2 (10 nM) or/and 4OHT (100 nM) drastically increased the ability of the SUM149 cells to migrate and invade into the lower side of the well through the matrigel (Fig. 5A). To confirm if the observed increase in the invasiveness of SUM149 cells is mediated by the ERK pathway, we treated the cells with U0126 (20 mM) alone and in combination with E2 (10 nM), and a substantial reduction in the migration and invasion by inclusion of MEK inhibitor U0126 (Fig. 5B).

Activation of p-ERK1/2 in IBC tissues
To validate the proof-of-principle evidence of p-ERK1/2 in IBC tumors, we examined immunohistochemically p-ERK1/2 expression in the tissues from five IBC patients using anti-p-ERK1/2 antibody (Fig. 6A). Three of five tissues displayed nuclear immunostaining of p-ERK1/2 with strong (#1, #3) and the moderate (#2) intensity, and two cases were negative or weak staining.

Discussion
ERa is known to be absent in many cases of IBC and its absence is associated with poor clinical consequences [25][26][27]. In this study, we did not detect ERa in both IBC cell line SUM149 and SUM190, however, we identified the presence of ERa variant, ERa36. ERa36 was found to be mainly present in plasma membrane and cytoplasm and to be involved in estrogen nongenomic signaling [15,28]. Subsequently, ERa36 has been reported to be potentially involved in the non-genomic signaling in several cancers including endometrial and gastric cancers [16,29]. Recently, E2 treatment has been reported to enhance the cell proliferation of ERa negative cell line, MDA-MB-231 and MDA-MB-468 through EGFR/Src/ERK signaling pathway [30], while the cell proliferation of SUM149 and SUM190 were not affected by E2 treatment, indicating no stimulation of estrogen mitogenic signaling. However, the migration and invasion were promoted by E2 treatment in both SUM149 and SUM190 cells, revealing that the non-genomic estrogen signaling might be associated with the aggressive characteristic of IBC cells to invade into lymph node.
The involvement of GPR30 in the non-genomic signaling is found in both ER-positive and -negative breast cancer cells [18]. The activation of p-ERK1/2 by the treatment of GPR30-specific agonist G1 also suggested the involvement of GPR30 in the nongenomic signaling cascade in IBC cells. The fact that GPR30 localized in the nucleus than in the cytoplasm in the IBC cells might suggest its lower contribution in activating non-genomic signaling by the GPR30 pathway. However, it has been shown that the nuclear localized GPR30 could also participate in the non-genomic signaling [31]. Kang et al. have shown that G1 is also able to stimulate ERa36 [23]. Furthermore, they have shown that the non-genomic signaling pathway through p-ERK1/2 is directly mediated by ERa36 but not by GPR30, because the activities of GPR30 promoted by estrogen were due to its ability to induce ERa36 expression. However, it remains possible that both receptors can play a crucial role in the non-genomic signaling pathway. Although we did not investigate whether G1 treatment upregulated the expression of ERa36 in these cells, there might be another layer of cross-talk between the GPR30 and ERa36 rather than the regulation of ERa36 protein expression by GPR30 as the primary mechanism of the noted non-genomic signaling.
In the IBC cells, the involvement of ERs in the non-genomic signaling seems to be more complicated due to the expression of ERb in addition to the expression of ERa36 and GPR30. The involvement of ERb in the non-genomic signaling pathway has been shown in several cancer cell lines [32,33]. The rapid activation of p-ERK1/2 by treatment with ERb-specific ligand DPN in IBC cells suggested a possible contribution of ERb in the non-genomic signaling. Recent data from the uterine carcinosarcoma suggests a possible combined role of ERb and GPR30 in the non-genomic signaling [34]. In addition, coordinated regulation of GPR30 and ERb in endometrial cells upon E2 treatment also leads to an increased expression of GPR30 mRNA through ERb [35]. Taken together, these data provide us the possibility that all of GPR30, ERa36 and ERb might be involved in the non-genomic signaling through a network of cross-talks among ERs in IBC cells. However, further studies must investigate these possibilities.
We also found that 4OHT and the pure antiestrogen ICI also failed to block the E2-mediated activation of p-ERK1/2 in IBC cells. The antiestrogens including tamoxifen, 4OHT and ICI have been reported to act as agonists for ERa36 and GPR30 in a variety of cancer cells [15,[36][37][38][39]. One of potential mechanisms which causes antiestrogen-resistant in some breast cancers could be the stimulation of anti-estrogen signaling pathways via ERa36 [15]. Since expression of ER/PR is generally lower in IBC cells [25,40,41], such agents are not used for IBC therapeutics. Our findings presented here raise a new possibility that ERa variant ERa36, ERb and GPR30 might be potentially targeted for IBC therapeutic. In this context, a recent study has shown that estriol could act as a GPR30 antagonist by inhibiting the GPR30mediated activation of p-ERK1/2 in ERa-negative breast cancer cells [42]. Search and discovery for the other antiestrogens and antagonists including estriol to block non-genomic signaling components through ERs could provide novel insights into the therapeutic strategy for IBC patients. In addition, we have shown IBC non-genomic signaling function including the migration and the invasiveness of SUM149 cells via estrogenand antiestrogen-mediated p-ERK1/2 activation ( Figure 6B). Our present study suggests that in addition to ERs-targeted agents, the drugs which specifically inhibit ERK1/2 might prove to be effective in IBC, either as a single agent or in combination with targeted therapeutics to evolve an effective anti-IBC therapeutic strategy.

Materials and Methods
Cell culture and reagents SUM149 and SUM190 were obtained from Asterand, plc (Detroit, MI) and maintained in Ham-F-12 (1:1) medium supplemented with 5% Fetal Bovine Serum (FBS), 5 mg/ml insulin, 1 mg/ml hydrocortisone and 10 mM HEPES. SUM149 cell line is obtained from the primary ductal carcinoma of the breast from a patient with a locally advanced disease while SUM190 is obtained from poorly differentiated inflammatory breast carcinoma.

Cell extracts and immunoblotting
To prepare cell extracts, the cells were grown in 5% DCC medium for the period of 48 hrs and then treated with either E2 (10 nM), DPN (10 nM), PPT (10 nM) or G1 (10 nM). When indicated, ICI (1.0 mM) or 4OHT (1.0 mM) was added 1 hr before and MEK inhibitor U0126 (20 mM) was added 15 min before ligand treatment. The cells were washed once with PBS and then lysed in radioimmunoprecipitation assay buffer [50 mmol/l Tris-HCl (pH 7.5), 150 mmol/l NaCl, and 0.5% 16 protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN)] for 10 minutes on ice. The cell lysates were centrifuged at 13,000 rpm for 10 min at 4uC. The cell lysates containing an equal amount of protein were then resolved on a sodium dodecyl sulfate-polyacrylamide gel (8% acrylamide), transferred to a nitrocellulose membrane, probed with the appropriate antibodies, and developed using the ECL detection reagent (Amersham Pharmacia Biosciences, Piscataway, NJ).

Immunoflourescent labeling and confocal microscopy
The cellular localization of proteins was determined by indirect immunofluorescence labeling. SUM149 and SUM190 cells were grown on sterile glass coverslips in 5% DCC media for 48 hrs, fixed in 4% paraformaldehyde, permeabilized in 0.1% Triton X-100, and blocked in 5% normal goat serum-PBS. The cells were incubated with the primary antibodies for 1 hr, washed thrice in PBS, and then incubated with goat anti-mouse or goat anti-rabbit

Matrigel invasion and migration assays
To quantify the cell migration potential of IBC cells, the cells were grown in 5% DCC media for 48 hrs. The cells were then trypsinised for collection, washed in PBS, resuspended in 0% DCC medium containing 0.1% BSA, and loaded into the upper well of an uncoated Boyden chamber (BD Biosciences) at a concentration of approximately 3.3610 4 cells/well. The lower side of the separating filter was filled 0.2% DCC medium containing 0.1% BSA. The cells were fixed after about 16 hrs and the cells which had successfully migrated were stained and counted manually. The same protocol was followed for measuring the invasion potential with principal differences being the use of Matrigelcoated Boyden chamber (BD Biosciences) and loading of 1610 5 cells/well.

Immunohistochemistry
Deparaffinized sections were subjected to antigen retrieval by boiling the sections in 10 mM citric acid buffer (pH 6.0) for 10 min. Sections were then incubated with rabbit p-ERK1/2 antibody (Cell Signaling) at 4uC overnight, followed by incubation with EnVision (Dako, Carpinteria, CA) for 1 hour at room temperature. Immunostained sections were lightly counterstained with hematoxylin, dehydrated in graded ethanol, cleared in xylene, and mounted with the use of the peramount mounting medium. Figure S1 Phosphorylation of Akt, JNK and p38 upon E2 treatment in SUM149 cells. The cells were maintained in 5% DCC for 48 hours and then treated with E2 (10 nM) and then Akt, JNK and p38 phosphorylation were analyzed by western blotting. (TIF) Figure S2 Cell proliferation of SUM149 with E2 treatment. The cells were plated in 6-well plates and treated with E2 the next day. The cell growth was determined by counting cell number using a Coulter Counter. (TIF)