Colocalization of somatostatin receptors and epidermal growth factor receptors in breast cancer cells

Somatostatin (SST) is an endogenously produced peptide in neuroendocrine and immune cells. It exists as two biologically active forms, SST-14 and SST-28, which are produced by tissue-specific proteolytic processing of a common precursor [1]. SST is a potent inhibitor of hormone and growth factor secretion as well as a modulator of cell proliferation [2, 3]. These actions are mediated by a family of G protein-coupled receptors (GPCR) with five known subtypes (SSTR1-5). SST exerts antiproliferative effects on normal dividing cells, such as intestinal mucosal cells, activated lymphocytes and inflammatory cells as well as on solid tumors and cultured cells derived from both endocrine and epithelial tumors. These effects include cytostatic (growth arrest) and cytotoxic (apoptotic) actions and are mediated (i) directly by SSTRs present on tumor cells, and (ii) indirectly via SSTRs present on non-tumor cell targets. SST inhibits the secretion of hormones and growth factors that promote tumor growth, inhibits growth factor-induced DNA synthesis, inhibits angiogenesis, promotes vasoconstriction and modulates immune cell function [1]. Moreover, immunoreactive SST has been identified, by immunohistochemistry, in 30% of breast cancer samples and in several breast cancer cell lines [4, 5]. Whether SST is synthesized and secreted from these cells and acts as a paracrine/autocrine growth inhibitor remains to be established.

All five SSTRs have been implicated in antiproliferative signaling in a subtype selective manner. When studied as individual isotypes, four of the receptors (SSTR1, 2, 4, 5) induce cell cycle arrest whereas SSTR3 uniquely triggers apoptosis [3, 6]. Previous studies have demonstrated the presence of SSTRs in a large variety of tumors and cancer cell lines [7‐9]. In addition, 15–66% of primary human breast tumors are SSTR-positive by binding analysis [10‐14]. Consistent with previous studies, we have recently shown that SSTRs are expressed in breast cancers in variable amounts and are correlated with various histological markers in a receptor-specific manner [15]. We have also shown the effects of estradiol and tamoxifen on SSTR1 and SSTR2 expression in breast cancer cells [16].

Epidermal growth factor receptors, members of the type I receptor tyrosine kinase (RTK) family commonly known as ErbBs, are also variably distributed in breast tumors and breast cancer cell lines as are SSTRs [17, 18]. ErbBs can be detected in all tumors with variable degrees of expression. There are currently four known ErbB receptors with ErbB1 (also known as EGFR) and ErbB2 (also known as Neu or HER2) being the most likely to be overexpressed in cancers, and, therefore, the most studied [19‐22]. ErbB3 and ErbB4 (also known as HER3 and HER4, respectively) have been investigated the least. ErbBs exist as monomers and, upon ligand activation or when overexpressed, form homo- and heterodimers [23, 24].

Previous studies showed that ErbB1 is expressed in 40–50% of breast cancer cases and is inversely related with estrogen receptor (ER) levels and survival [25‐27]. This is associated with more aggressive proliferation and unresponsiveness to hormone treatment [12, 14, 27]. Similarly, ErbB2 is present in 10–40% of breast cancer cases and is associated with poor survival [19, 21, 25, 26]. ErbB3 is also expressed in breast cancer [28, 29]. Associations with ErbB1 and ER have been shown in some studies but not in others [20]. This discrepancy may be due to the techniques employed, antibodies used, sample size or tumor type. In contrast with ErbB1-3, ErbB4 is generally reported to be associated with favorable prognostic factors [20, 21, 25, 30, 31].

While ErbBs are involved in tumor growth and cell proliferation and are often associated with poor response to endocrine therapy and reduced survival, SSTRs play a major role in the control of tumor growth and tumor cell proliferation [32‐34]. SSTR expression is positively correlated with tumor size and inversely correlated with ErbB levels and tumor differentiation [12, 14]. Several recent reports have shown GPCRs to directly interact with RTKs via scaffolding proteins when both receptors are present together in the large signaling complexes [35‐37]. Alternatively, GPCRs can indirectly transactivate RTKs via G proteins which ultimately lead to increased intracellular calcium levels and activation of PKC [38]. Indirect RTK transactivation has also been reported to occur via membrane-bound metalloproteinases (MMPs) or metalloproteinase-disintegrin proteins (ADAMs) which process ErbB transmembrane ligands [35, 39, 40]. In general, RTK transactivation by GPCRs results in altered mitogen activated protein kinase (MAPK) signaling and, subsequently, in altered cell growth and proliferation [39, 41, 42]. It is not known if SSTRs (GPCR) and ErbBs (RTK) are coexpressed within the same cells. Hence, before defining the mechanisms for functional interactions between ErbBs and SSTRs, it is essential to determine if this occurs. We have therefore determined, in the current study, SSTR1-5 and ErbB1-4 expression at the protein and mRNA levels. In addition, since ER has been shown to be associated with ErbB levels, we investigated their colocalization in ER-positive (ER+) and negative (ER-) breast cancer cells. Our data showed that SSTRs and ErbBs are well expressed in both cell lines and, significantly, exhibited variable colocalization.

The present study represents the first comprehensive description showing SSTR1-5 and ErbB1-4 colocalization in ER+ and ER-breast cancer cells. All five SSTRs were detected in MCF-7 and MDA-MB-231 with a rich expression of subtypes 1 and 4, moderate expression of SSTR2 and relatively weak expression of subtypes 3 and 5. Our data also demonstrate a potential correlation between SSTR and ErbB expression and estrogen dependency. We found higher levels of expression of ErbB1 and lower levels of SSTR1, SSTR4 and ErbB3 in ERα – (MDA-MB-231) cells when compared to ER+ (MCF-7) breast cancer cells. In addition, we showed that there was more colocalization of SSTRs with ErbBs in MCF-7 cells than in MDA-MB-231 cells. We also detected preferential colocalization among ErbBs in both MCF-7 and MDA-MB-231 cells.

Overall expression levels of SSTR subtypes in cultured breast cancer cell lines were comparatively less than in solid tumors. Significantly, SSTR3, which is well expressed in breast tumor tissues, was relatively poorly expressed in these cell lines [15]. These results indicate that the various breast cancer cell lines, although useful for studying SSTR biology, do not necessarily reflect endogenous tumor SSTR expression or function. Possible explanations for the difference are the probable induction of SSTR expression in solid tumors by circulating hormones, or, locally, by growth factors, cytokines, and other mediators produced from peritumoral structures such as the stroma, blood vessels and immune cells [45]. Increasing evidence points to the occurrence of multiple SSTR subtypes in many different types of tumor cells as well as normal cells [46, 47]. All five SSTR isoforms bind the natural ligands SST-14 and SST-28 with nanomolar affinity and share common signaling pathways, such as the inhibition of adenylyl cyclase, making the functional significance of expressing more than one SSTR subtype in the same cell unclear [2]. Whether the different SSTRs subserve different biological roles in the same cell or cooperate through dimerization to create greater signaling diversity remains to be determined. In this regard, we have recently shown that SSTR1 and SSTR5 heterodimerization, in stably transfected HEK and CHO-K1 cells, results in a new receptor with enhanced signaling properties [48, 49]. We further anticipate such a possibility of heterodimerization between SSTR1 and SSTR5 and, additionally, between SSTRs and ErbBs in breast cancer cells.

Whereas SSTRs have been associated with antiproliferative signaling, several previous studies, using a variety of tumors including MCF-7 and MDA-MB-231 cells, have correlated ErbBs with tumor progression and poor prognosis [19, 22, 50, 51]. However, the data have been inconsistent and controversial [52‐54]. These inconsistencies may have arisen due to the techniques employed, the variation between cell stocks studied in different laboratories and, most significantly, the different passages at which the cells were used [45]. In this regard, we have seen significant variation in receptor expression/levels at different passages (data not shown). In keeping with ErbBs roles in tumor progression and poor prognosis, overexpression of ErbBs in breast carcinomas has been correlated with a lack of ER [44, 52]. Furthermore, blocking ER using antisense strategies resulted in increased ErbB1, no change in ErbB2 and a slight decrease in ErbB3 expression in breast cancer cells [22]. Consistent with these observations, we found higher levels of expression of ErbB1 and decreased levels of ErbB3 in ERα – (MDA-MB-231) than in ER+ (MCF-7) cells. In accordance with previous studies, our findings strongly support the concept that the presence of ER could be a determining factor in ErbB expression in both breast cancer cells and tumors.

Previous reports state that specific ErbB heterodimers, i.e., ErbB1/ErbB2 and ErbB2/ErbB3, result in increased tumor growth and cell proliferation. We report that, in MCF-7 and MDA-MB-231 cells, there is preferential colocalization of ErbBs with other ErbBs. We found greater colocalization between ErbB1 and ErbB3 in both MCF-7 and MDA-MB-231 cells. We also detected a high degree of colocalization between ErbB2 and ErbB4 in MCF-7 cells. These data strongly support previous observations whereby heterodimerization between ErbB1 and ErbB2 was correlated with tumor progression [22, 51]. These alternate heterodimer pairs, i.e., ErbB1/ErbB3 and ErbB2/ErbB4, may account for the less aggressive proliferation rates reported for both cell lines. Furthermore, in agreement with previous studies, we detected fewer cells showing ErbB colocalization in ERα – cells (MDA-MB-231) than in ER+ (MCF-7) cells with the exception of those coexpressing ErbB1 and ErbB3. Altogether, the higher degree of colocalization of ErbBs in MCF-7 cells than in MDA-MB-231 cells may be partially associated with slower tumor growth and better response to hormonal therapy. Our data provide direct evidence that ErbB1 and ErbB3 are the prominent subtypes which may interact as heterodimers, in these cells. Nothing is currently known regarding the physiological responses and functional consequences of these observations suggesting that further studies are required in this direction.

In addition to heterodimerization within receptor subfamilies, there have been several reports demonstrating that crosstalk between RTKs and GPCRs modulates downstream signaling pathways [35‐37]. Even so, direct evidence for functional interactions between ErbBs and SSTRs have not yet been demonstrated despite the critical roles they play in tumor progression. We showed here that there was increased colocalization of SSTRs with ErbBs in MCF-7 cells (ER+) compared with MDA-MB-231 (ERα-) cells. This may help elucidate why estrogen-sensitive tumors show less aggressive proliferation than estrogen-insensitive tumors. This pattern of colocalization may also explain the superior response of ER+ patients to SST analog therapy [55]. In MCF-7 cells, the preferentially greater colocalization of SSTRs with ErbB2 may serve to counteract any deleterious effects of ErbB2. Whether this colocalization exists in vivo and is lost during tumor progression needs to be determined. Furthermore, colocalization of SSTR1 and SSTR5 with ErbB4 supports the antiproliferative effects of both SSTRs. SSTR interactions with ErbB4 may also serve to potentiate ErbB4's previously reported role in differentiation and apoptosis [30]. Furthermore, by preventing ErbB4's downregulation, SSTRs may be indirectly circumventing ErbB1-3's growth promoting effects. However, whether such interactions exist in vivo in solid tumors needs to be determined.

Despite SSTR and ErbB colocalization, low abundance of SSTRs alongside high expression of ErbBs within the same cell may account for the failure of SST treatment of breast tumor or other ErbB-expressing tumors. Furthermore, it is anticipated but not yet proven that SSTRs would reverse the effects of ErbBs with respect to MAPK activation and subsequent cell proliferation [56‐58]. In addition, some reports suggest that the ER is involved in MAPK activation [59‐61]. Previous studies have also demonstrated that ER presence is required for cbl-induced ubiquitination of ErbB1 and that ubiquitination of ErbB1 results in its degradation [62]. This could result in different levels of activation of downstream pathways in ER+ (MCF-7) and ERα – (MDA-MB-231) breast cancer cells. In addition, SST-induced internalization and subsequent downregulation of SSTR2-5 on the membrane may release ErbBs from complexes and result in cell proliferation [63‐65]. Altogether, this suggests that not only do we need to activate SSTRs to counteract ErbBs effects on cell proliferation but we also need a mechanism to upregulate, or at least maintain, SSTRs on the membrane in order to reduce or modify ErbB signaling.

In summary, the present results have important functional and therapeutic implications. Predominant SSTR1 expression and weak SSTR5 expression in breast cancer cells may help explain their poor sensitivity to hormonal therapy. These data may also explain the differential effects of the SST analog octreotide in breast cancer therapy. Since there is evidence of crosstalk between GPCRs and RTKs, cells displaying SSTR colocalization with ErbB suggest that, within these cells, both receptor families may functionally interact through hetero-oligomerization. If such a process exists, it may account for the diversification of receptor signaling. Most significantly, developing a new therapeutic agent that could both activate SSTRs and inhibit ErbB overexpression could potentially be a way to block tumor progression.