Minireview: Inflammation: An Instigator of More Aggressive Estrogen Receptor (ER) Positive Breast Cancers

Abstract

Approximately 75% of breast tumors express the estrogen receptor (ER), and women with these tumors will receive endocrine therapy. Unfortunately, up to 50% of these patients will fail ER-targeted therapies due to either de novo or acquired resistance. ER-positive tumors can be classified based on gene expression profiles into Luminal A- and Luminal B-intrinsic subtypes, with distinctly different responses to endocrine therapy and overall patient outcome. However, the underlying biology causing this tumor heterogeneity has yet to become clear. This review will explore the role of inflammation as a risk factor in breast cancer as well as a player in the development of more aggressive, therapy-resistant ER-positive breast cancers. First, breast cancer risk factors, such as obesity and mammary gland involution after pregnancy, which can foster an inflammatory microenvironment within the breast, will be described. Second, inflammatory components of the tumor microenvironment, including tumor-associated macrophages and proinflammatory cytokines, which can act on nearby breast cancer cells and modulate tumor phenotype, will be explored. Finally, activation of the nuclear factor κB (NF-κB) pathway and its cross talk with ER in the regulation of key genes in the promotion of more aggressive breast cancers will be reviewed. From these multiple lines of evidence, we propose that inflammation may promote more aggressive ER-positive tumors and that combination therapy targeting both inflammation and estrogen production or actions could benefit a significant portion of women whose ER-positive breast tumors fail to respond to endocrine therapy.

Estrogen receptor (ER)α is expressed in approximately 75% of breast tumors at the time of diagnosis, and ER status serves as a major prognostic marker and determinant of the course of therapy that a patient will receive. Patients with ER-positive tumors generally have a better prognosis than those with tumors that lack ER, and these patients usually undergo treatment with endocrine therapy such as the selective ER modulator (SERM) tamoxifen or aromatase inhibitors. Unfortunately, many patients with ER-positive tumors fail to respond to endocrine therapy, and most tumors that are initially responsive acquire resistance (1). These tumors typically continue to express ER (2) and demonstrate earlier metastatic recurrence (3). Currently, there is still no clear method to distinguish tumors that will or will not respond to endocrine therapy. As a result, one goal of research efforts is to find predictive markers so that women with ER-positive tumors can be treated appropriately, meaning that women with tumors that are most likely to respond to endocrine therapy can avoid detrimental overtreatment, whereas women whose tumors are least likely to respond will receive more aggressive chemotherapies earlier.

In recent years, numerous studies using gene expression profiling have attempted to uncover signatures that predict response to endocrine therapy to steer women into more appropriate therapeutic avenues. For example, profiling of ER-positive tumors has led to a number of signatures that are predictive of clinical outcome for patients taking tamoxifen (4, 5). In broader gene profiling studies of both ER-positive and ER-negative breast tumors, ER-positive tumors were classified into two intrinsic subtypes, Luminal A and Luminal B, that differ significantly in relapse-free and overall survival (6). Tumors of the Luminal A subtype are associated with greater overall patient survival, whereas the Luminal B subtype is associated with significantly worse patient outcome. These findings led to the development of the PAM50 Breast Cancer Intrinsic Classifier assay that uses expression levels of 50 genes to classify the intrinsic tumor subtype, including Luminal A and Luminal B, and can provide additional predictive and prognostic information beyond current pathological tumor characterization (7, 8). Another multigene assay, called Oncotype DX, has been developed to predict recurrence risk within 10 yr of diagnosis for patients with node negative, ER-positive, stage I or II invasive breast cancer. Expression levels of 16 cancer-related genes and five control genes are used to quantify the likelihood of recurrence (9) and predict the benefit that a patient with an ER-positive tumor may gain from adjuvant chemotherapy treatment (10, 11).

Although gene expression studies have the potential to provide greater predictive ability for responsiveness of some ER-positive breast tumors to endocrine therapy, the underlying biology causing tumor heterogeneity has yet to become fully clear. What is known is that ER-positive tumors with poor response to endocrine therapy tend to have lower ER expression (12, 13), low to no progesterone receptor expression (14, 15), and high levels of proliferation-associated genes (4, 16). One accepted mechanism of resistance to endocrine therapy is overexpression of human epidermal growth factor receptor 2 (HER2) but this occurs in only 10% of ER-positive breast cancers (17–19). A number of studies suggest that other growth factors and signaling pathways, such as IGF-I, phosphatidylinositol 3 kinase (PI3K)/protein kinase B (AKT), and protein kinase C alpha (PKCα), contribute to poor outcome for women with ER-positive tumors (13, 20–24). Current thinking is that human epidermal growth factor receptor 2 (HER2) signaling or activation of these other pathways can either override or alter ER activity, such that tamoxifen or aromatase inhibitors are no longer effective. Thus, a number of mechanisms have been proposed to contribute to tumor escape from endocrine therapy, suggesting that the etiology of endocrine resistance may be multifactorial. In this review, we will explore inflammation as another biological factor that may cause increased risk of ER-positive breast cancer and/or poor therapeutic response to endocrine therapy.

Inflammation is now considered a hallmark of cancer (25) and can play a role in virtually all aspects of tumor biology, including initiation, promotion, angiogenesis, and metastasis (26). An inflammatory tumor microenvironment consists of infiltrating immune cells and activated fibroblasts, both of which can secrete cytokines, chemokines, and growth factors, as well as DNA-damaging agents (27). Downstream of cytokines and chemokines, the nuclear factor κB (NF-κB) pathway is known to be a major player in many aspects of tumor biology (28). Tumors themselves are able to both generate and respond to inflammatory microenvironments (26). Oncogenic changes within the tumor, hypoxia, secretion of molecules that attract inflammatory cells, and the extensive tumor cell death and necrosis initiated by cancer therapy are all contributors to an inflammatory microenvironment. Additionally, factors outside of the tumor, such as obesity, can create an inflammatory environment that contributes to tumorigenesis.

The structural and cellular composition of the breast provides a unique microenvironment for both tumor growth and localized inflammation. Hormonally responsive epithelial cells with specialized functions for milk production and ejection, along with many types of stromal and immune cells, are embedded in adipose tissue, which is now clearly recognized as both an endocrine and an inflammatory organ (29). In this review, several lines of clinical, preclinical, and cell-based evidence will be presented that suggest that an inflammatory microenvironment in the breast may be both a mediator of various risk factors associated with breast cancer as well as a player in the development of more aggressive breast tumors that fail to respond to therapies. More specifically, two risk factors, mammary gland involution, after pregnancy, and obesity, and their ability to create an inflammatory tumor microenvironment, will be discussed (Fig. 1). In addition, the ability of locally produced estrogens and proinflammatory cytokines, and subsequent activation of ER and NF-κB transcription factors, to promote a more aggressive ER-positive breast tumor phenotype will be reviewed. We will focus on the role of inflammation in ER-positive cancers, because studies have shown that regular use of nonsteroidal antiinflammatory drugs (NSAIDS), such as aspirin, significantly reduce the risk of ER-positive but not ER-negative breast cancers (30). Furthermore, although there is an inverse correlation between most markers of inflammation and ER, those ER-positive tumors with an inflammatory component tend to be more aggressive and are associated with a poor outcome and therapeutic failure.

Fig. 1.

Inflammation promotes an aggressive phenotype of ER-positive tumors. Pregnancy, particularly mammary gland involution, and obesity are known breast cancer risk factors that can promote an inflammatory microenvironment locally within the breast. TAMs (M), adipocytes (A), and fibroblasts (F) within the tumor microenvironment produce both estrogens and cytokines that can activate ER and the NF-κB pathway, respectively, in breast cancer cells. ER and NF-κB can repress or enhance each other's activity at particular genes to promote an aggressive phenotype of ER-positive breast cancer with endocrine therapy resistance, chemoresistance, and increased cell survival and proliferation.

Inflammation and Risk Factors for Breast Cancer

The risk factors for breast cancer are well established and include age, family history of breast cancer, certain genetic changes, radiation exposure, hormone exposure, reproductive history, obesity (particularly after menopause), physical inactivity, and alcohol use (www.cancer.gov). Interestingly, some risk factors, such as menopause and increased age, are associated with systemic inflammation, as indicated by increased levels of circulating proinflammatory cytokines (31, 32). Other risk factors have the potential to promote or maintain an inflammatory microenvironment locally in the breast and thereby influence breast tumorigenesis. Here, two of these risk factors—pregnancy, specifically involution of the mammary gland after pregnancy, and obesity—will be discussed further.

Pregnancy

Initial epidemiological studies revealed that pregnancy is associated with a reduced lifetime risk of developing breast cancer, and large prospective studies confirmed this notion. However, more recently, it has been realized that the protection conferred by pregnancy is not “immediate or constant,” because breast cancer risk actually increases for approximately 10 yr after parturition (33). Unfortunately, these pregnancy-associated breast cancers, which are breast cancers diagnosed after parturition, are associated with a poorer prognosis and decreased overall survival when compared with women who have not recently given birth (34) or age-matched nulliparous women (35). Although delayed diagnosis and gestational hormones can contribute to pregnancy-associated breast cancers and their poor outcome, emerging data suggest that the microenvironment of an involuting mammary gland is an equally important factor (33).

Involution is the period of controlled apoptosis and dramatic tissue remodeling that occurs after giving birth or after cessation of lactation in women who nurse. Involution is an orderly process that is considered noninflammatory by some (36). However, many inflammatory mediators are present and active during mammary gland involution and are thought to contribute to both the increased risk and poor prognosis of pregnancy-associated breast cancers. For example, parous women showed increased inflammation-associated gene expression in breast tissue for up to 10 yr postpartum compared with age-matched nonparous women (37). Additionally, macrophage influx is a prominent feature of involuting glands (38). Characterization of the involuting mammary gland and its stroma, through microarray studies and morphological examination, has revealed extensive immune cell infiltration and similarities to acute-phase and wound-healing responses (33). Furthermore, one of the major molecular mediators of involution appears to be NF-κB. In rodent studies, expression and activity of NF-κB is abundant during involution but not lactation, and constitutive activation of NF-κB in mammary epithelium results in rapid loss of milk production and an involution phenotype (39, 40). NF-κB appears to work in conjunction with signal transducer and activator of transcription 3 (STAT3) to mediate these inflammatory effects, because activation of both is involved in involution and induction of the acute-phase-response profile (41).

Factors from the involuting mammary gland can promote tumor growth, invasion, and metastasis. Initially, it was shown that extracellular matrix from involuting mammary glands could support enhanced cancer cell invasiveness in vitro and metastasis in vivo (42). More recently, the same group demonstrated that the involuting mammary gland promotes growth of larger tumors with more invasive properties than a noninvoluting gland. These effects are accompanied by increased fibrillar collagen deposition and up-regulation of cyclooxygenase-2 (COX-2) expression in the tumors (43). Inhibition of COX-2 with nonsteroidal antiinflammatory drugs (NSAIDS) reduced both tumor growth and metastasis, suggesting that these antiinflammatory drugs may be beneficial for pregnancy-associated breast cancers.

The ER status of pregnancy-associated breast cancers is not clear, with conflicting data in the literature (44, 45). However, COX-2 expression in breast cancers has been positively correlated with the expression of aromatase, the enzyme responsible for converting androgens to estrogens (46). In addition, prostaglandin E2, a product of the COX-2 prostaglandin synthesis pathway, has been shown to stimulate aromatase gene expression (47) through a mechanism involving cAMP and inhibition of the AMP-activated protein kinase pathway (AMPK), recently reviewed by Brown and Simpson (48). Taken together, these findings imply that up-regulation of COX-2, as observed in tumors in an involuting mammary gland, may lead to increased prostaglandin levels, which could promote expression of aromatase. The resulting increase in estrogen production could, in turn, stimulate the proliferation of ER-positive cancers.

Obesity

Obesity is associated with an increased risk of breast cancer, as well as an increased risk of recurrence and higher death rates, especially in postmenopausal women (49–51). Additionally, postmenopausal weight gain is associated with ER-positive tumors rather than ER-negative tumors (52, 53). A recent study found that obese women with ER-positive breast tumors have a higher rate of recurrence than lean women after treatment with tamoxifen or aromatase inhibitors (54). Overall, the literature suggests that obese women tend to have ER-positive tumors that are resistant to endocrine therapies, more likely to recur, and associated with higher death rates.

Obesity is associated with elevated levels of proinflammatory cytokines in adipose tissue and in the circulation, which generates a low-grade, chronic inflammatory state. One hallmark of obesity-associated inflammation is the recruitment of macrophages into adipose tissue, where they form characteristic “crown-like” structures around apoptotic adipocytes. Macrophages and adipocytes are able to produce inflammatory factors, such as adipokines and cytokines (29), leading to activation of the proinflammatory transcription factor NF-κB in adipose tissue and liver (55). Inflammation and NF-κB activation were also observed in mammary adipose in an animal model of diet-induced obesity (56). In mammary tissue of women undergoing breast surgery, crown-like structures of macrophages and NF-κB activation have been observed (57). Importantly, the extent of inflammation was correlated with both body mass index and adipocyte size, suggesting a strong connection between mammary gland inflammation and obesity in women.

One likely explanation for the tight link between obesity, inflammation, and hormone-dependent breast cancer can be explained by the observation that adipocytes express aromatase and that this enzyme is up-regulated in the adipose tissue of obese women (57). Aromatase expression is regulated not only by prostaglandins but also by proinflammatory cytokines. A study by Irahara et al. (58) found a significant association between aromatase expression and expression of TNFα and IL-6 in mammary adipose tissue and axillary adipose tissue from patients with breast cancer but not in adjacent normal breast tissue. This suggests that TNFα and IL-6 produced by adipocytes could act through an autocrine or paracrine manner to increase expression of aromatase in the tumor microenvironment. In studies of adipose stromal cells, Zhao et al. (59) found that the proinflammatory cytokine TNFα regulates aromatase expression by stimulating the binding of c-Fos and c-Jun transcription factors to an activation protein 1 (AP-1) element located upstream of promoter I.4 of the aromatase gene. An increase in adipose mass, as well as proinflammatory cytokines, in obese women could produce higher levels of both circulating and local estrogens, which, in turn, can stimulate the growth of ER-positive cancers (60). In addition to increasing estrogen levels, obesity-associated cytokines can have a number of direct effects on breast cancers, many of which may be mediated by NF-κB (discussed below), to contribute to the more aggressive, endocrine therapy-resistant ER-positive tumors seen in obese women.

Proinflammatory Factors in Breast Cancer

Tumor-associated macrophages (TAMs)

TAMs generally arise from circulating monocytes that migrate into tissues in response to chemical signals and differentiate into macrophages, and they are distinctly different from tissue histiocytes that naturally reside in specific tissues. In breast cancer, TAMs have been found to comprise up to 50% of the breast tumor mass (61), and a greater extent of macrophage infiltration in breast tumors is positively correlated with increased angiogenesis as well as reduced relapse-free and overall survival (62).

In the polyoma middle T oncogene (PyMT) mouse model of mammary cancer, which closely mimics human breast cancer progression (63), it was found that macrophage recruitment may be required for progression of mammary tumors to a metastatic stage (64). Lin et al. (64) demonstrated that polyoma middle T oncogene (PyMT) mice that are homozygous for colony-stimulating factor-1 (CSF-1) null mutation lacked macrophage infiltration at the period of time marked by transition to carcinoma in heterozygous mice. Although these mice showed no difference in the incidence or growth of the primary tumor, transition to malignancy, as well as pulmonary metastasis, was markedly delayed in the CSF-1 homozygous mutants. This study suggests that CSF-1 is a mediator of macrophage recruitment into mammary tumors and that these TAMs may play a critical role in the processes of tumor progression and metastasis.

In human breast cancer, the association between TAMs and ER status remains unclear. Although one study found an inverse relationship between the macrophage content and ER expression in breast tumors (65), other studies have suggested that macrophages can regulate proliferation (66) and invasiveness (67) of ER-positive breast tumors. A study by Mor et al. (66) demonstrates that macrophages produce estrogen, both in vitro and in human breast tumors, and that conditioned media from cultured macrophages stimulated growth of ER-positive breast cancer cells. Treatment with the aromatase inhibitor, letrozole, prevented the stimulatory effects of macrophage-conditioned media, suggesting that estrogen produced by TAMs can have a direct proliferative effect on breast cancer cells. Additionally, findings by Hagemann et al. (67) reveal that coculture of macrophages and ER-positive breast cancer cells increases cancer cell invasiveness through a mechanism involving TNFα-dependent activation of the NF-κB and c-Jun N-terminal kinase (JNK) pathways. Notably, the Oncotype DX assay, discussed previously, employs the macrophage marker CD68 in its analysis of breast tumors. Higher expression of CD68 is associated with higher recurrence score, suggesting that the presence of macrophages in ER-positive tumors correlates with poor prognosis and response to endocrine therapy (9).

Interleukin-6

Serum levels of IL-6 are generally higher in breast cancer patients than in healthy women, and increased levels of IL-6 are correlated with poorer survival and diminished response to endocrine therapy in patients with metastatic breast cancer (68, 69). The source of higher IL-6 serum levels is not precisely known, because immune cells, stromal cells, tumor cells, or a combination of the three could serve as the source (69). An in vitro study of breast cancer cell lines revealed that they can produce IL-6, but ER-positive cells secrete lower levels of IL-6 than ER-negative cells (70). However, treatment of ER-positive cells with IL-6 has been found to promote proliferation (70) and a more aggressive phenotype. For example, Sullivan et al. (71) found that overexpression of IL-6 in MCF-7 cells induced epithelial-mesenchymal transition as well as increased invasiveness. In addition, IL-6 produced by fibroblasts from breast tissue, as well as from bone and lung, which are common sites of breast cancer metastasis, stimulate growth and invasion of MCF-7 cells (72). Furthermore, adipose stromal cells were found to induce invasion and metastasis of MCF-7 cells, and IL-6 was determined to be a critical mediator of this increased invasive phenotype (73). These studies suggest that although ER-positive breast cancer cells produce low levels of IL-6, they are highly responsive to the tumor-promoting effects of this cytokine.

Interleukin-1β

IL-1β is expressed by both cancer and stromal cells, particularly in ER-negative breast tumors, and increased serum and tumor levels of IL-1β are correlated with invasiveness and poor prognosis (74). In contrast to the reported correlation of IL-1β levels with ER-negative status, a study of 77 breast tumors found that all tumors expressing IL-1β were ER-positive (75). The authors subsequently demonstrated that IL-1β was able to directly stimulate ERα transcriptional activity using luciferase reporter assays. Additionally, Zhu et al. (76) demonstrated that in the presence of IL-1β, the selective ER modulator (SERM) 4-hydroxytamoxifen (4-OHT) activates, rather than represses, ER target gene transcription. 4-OHT requires SMRT (silencing mediator of retinoic acid and thyroid hormone receptor) and N-CoR (nuclear receptor corepressor) to inhibit estrogen-dependent breast cancer proliferation and ER activity (77). IL-1β is able to displace an inhibitory complex that contains NCoR from ER target gene promoters, leading to a derepression of gene transcription in response to 4-OHT (76). These findings suggest that IL-1β may play a role in regulating ER activity and responsiveness to endocrine therapy.

Additional studies suggest that IL-1β promotes migration and metastasis of ER-positive breast cancer cells (78, 79). Franco-Barraza et al. (78) found that IL-1β, acting through its receptor IL-1RI, stimulated morphological changes of MCF-7 cells to a more fibroblast-like appearance and actin cytoskeletal reorganization. This was accompanied by an increase in cell motility, as well as an increase in matrix metalloproteinase-9 and -2 activity, both of which can mediate aspects of breast tumor angiogenesis, invasion, and metastasis (80). Wang et al. (79) confirmed that IL-1β increased migration and further reported that IL-1β promotes invasion and metastasis of ER-positive breast cancer cells.

Tumor Necrosis Factor-α

Initial studies of TNFα in breast cancer demonstrated that increased circulating levels of this cytokine were correlated with an increased tumor stage and lymph node metastasis (81). Also, levels of TNFα were found to be increased in invasive breast tumor samples compared with benign tissue (82), specifically in the stromal compartment of the tumor. Studies of TNFα action in ER-positive breast cancer cell lines have shown contrasting results. Although TNFα promotes apoptosis in MCF-7 cells (83), it stimulates proliferation in T47D cells (84, 85). This proliferation was found to be mediated through increased expression of cyclin D1 and was NF-κB dependent, because treatment with a specific NF-κB inhibitor blocked TNFα-induced proliferation (84, 85). Additional studies in MCF-7 cells have shown that TNFα can promote tumor cell invasion in vitro, as well as up-regulate a number of genes associated with proliferation, invasion, and metastasis (86). TNFα has also been found to promote chemotherapeutic resistance in MCF-7 cells through up-regulation of ABCG2 expression (87, 88), an ATP-binding cassette transporter that effluxes chemotherapeutic drugs out of cancer cells (89).

ER and NF-κB Cross Talk in Breast Cancer

The ER pathway

Estrogen promotes breast cancer cell proliferation and survival by regulating the expression of a wide variety of genes, including genes that promote cell-cycle progression, prevent apoptosis, as well as genes that encode growth factors, transcription factors, hormones, and signaling molecules (90). In addition to protein encoding transcripts, estrogen can also regulate the expression of intergenic transcripts as well as several classes of noncoding RNA (91). Estrogen action is mediated through its two receptors, ERα and ERβ, which can alter gene transcription by binding to DNA at specific sequences, known as estrogen response elements (ERE), or by interacting with other transcription factors, such as proteins of the AP-1 complex, specificity protein-1 (Sp1), or NF-κB, without directly binding to DNA. In fact, genome-wide studies have shown that ER binding sites are enriched not only for ERE but also for a number of other transcription factor-response elements, suggesting that cooperativity between ER and other factors may be a major mechanism of ER action (92, 93).

The NF-κB pathway

A key downstream mediator of proinflammatory cytokines is the NF-κB family of transcription factors, which is known to play a critical role in the development and progression of a variety of tumors (28, 94). The NF-κB pathway can be activated by a number of stimuli, including cytokines, such as TNFα and IL-1β, growth factors, chemotherapeutic drugs, physiological stress, as well as many others. In the canonical pathway, NF-κB family members are bound to inhibitors of NF-κB (IκB) in the cytoplasm and rendered inactive. Members of the NF-κB family include RelA/p65, RelB, and c-Rel, which all contain transcriptional activation domains at their C terminus, as well as the p105/p50 and p100/p52 family members that lack this domain. When a cell is stimulated by an appropriate extracellular signal the IκB kinase (IKK) complex, consisting of the IKKα and IKKβ catalytic subunits, as well as the IKKγ and ELKS [an IκB kinase regulatory unit rich in glutamic acid (E), leucine (L), lysine (K), and serine (S)] regulatory subunits, is activated. This activated complex phosphorylates IκB, which then undergoes proteasomal-mediated degradation and releases NF-κB dimers that travel to the nucleus and bind to DNA at NF-κB-response elements to modulate the expression of a wide variety of genes, including inflammatory cytokines, chemokines, immune receptors, and cell adhesion molecules. These genes are important mediators of inflammation and survival and can contribute to both promotion and progression of tumors (28).

In an alternative pathway of NF-κB activation, certain TNF family members trigger the activation of IKKα, independent of IKKβ or the other regulatory subunits of the IKK complex. Activated IKKα phosphorylates p100, resulting in its proteasomal processing to p52, which can form a transcriptionally active dimer with RelB (95). IKKα has been shown to be an essential regulator of receptor activator of nuclear factor κB (RANK) signaling and NF-κB activation in mammary gland development (96). Additionally, IKKα has nuclear functions independent of NF-κB transcriptional activity (97) through interactions with other transcription factors, such as ER, to regulate a number of genes important for proliferation and cell-cycle progression in breast cancer (98, 99), as discussed below.

NF-κB and ER-Positive breast cancer

Constitutive NF-κB activity has been observed in both human breast cancers as well as breast cancer cell lines (100), and studies have found NF-κB activation to be predominantly associated with ER-negative breast tumors (101, 102). However, there is an increasing amount of evidence that NF-κB activation occurs in a subset of ER-positive tumors and cell lines with poor response to endocrine therapy. The cause of NF-κB activation is largely unknown, but the status of progesterone receptor, which has been shown to have an antiinflammatory role in breast cancer cells (103, 104) and is lost in a subset of endocrine therapy-resistant ER-positive tumors, may be one contributing factor. Regardless of the cause, NF-κB activation in ER-positive tumors is associated with a more aggressive phenotype. Zhou et al. (105) found that tumors with lower ER levels have significantly higher p50 and p65 DNA-binding activity when compared with tumors that expressed higher levels of ER. Also, in this study, p50 binding was observed to be significantly greater in tumors that were destined to relapse. Additionally, activation of the NF-κB pathway can confer resistance to endocrine therapy upon breast cancer cells. In two models of resistance, treatment with parthenolide, a small molecule inhibitor of the NF-κB pathway, was able to restore endocrine sensitivity (106, 107). Furthermore, a subset of genes that is synergistically up-regulated by the combination of estrogen and proinflammatory cytokines was recently identified (108), and many of these genes have been confirmed to be regulated by positive cross talk between ER and NF-κB. Importantly, this gene signature of positive cross talk was found to be significantly overexpressed in ER-positive human breast tumors of the Luminal B-intrinsic subtype, which is associated with poor overall patient outcome, but not the Luminal A subtype. Together, these studies indicate that activation of the NF-κB pathway occurs in a subtype of more aggressive, endocrine-resistant ER-positive tumors and suggests that cross talk between ER and NF-κB could be an important mechanism in the progression of these tumors.

ER and NF-κB: Transrepression

Numerous studies on the cross talk between ER and NF-κB have demonstrated that transrepression, in which ER represses NF-κB activity and NF-κB represses ER activity, is a major mechanism of interaction between these factors at specific target genes. In general, the ability of ER to repress NF-κB action is associated with the antiinflammatory properties of estrogen and occurs primarily on classical inflammatory target genes of NF-κB (109). Inhibition of NF-κB by ER occurs through different mechanisms that can affect either NF-κB pathway activation, NF-κB DNA binding, or NF-κB transcriptional activity (110). For example, ER may inhibit NF-κB activation by maintaining a steady state of IκB levels (111, 112). Alternatively, ER can inhibit NF-κB binding to DNA, which appears to be particularly important for the regulation of IL-6 gene expression (113, 114). In breast cancer cells, ER represses RelB expression by preventing NF-κB and AP-1 interaction with the RelB promoter (115). In addition, ER can repress NF-κB transcriptional activity at target genes through direct interactions between the two transcription factors (116, 117) or indirect interactions between ER, NF-κB, and coactivators proteins, leading to formation of complexes on DNA that are unfavorable for transcriptional activation (118). Several studies describe a role for the histone acetyltransferase coactivator cAMP response element-binding protein-binding protein (CBP) or the related factor p300 in transrepression of NF-κB activity by ER (119–121). These studies suggest that transrepression may occur through a competition between ER and NF-κB for CBP/p300 or through displacement of CBP from the promoters of specific NF-κB target genes (120, 121).

Although study of ER-mediated repression of NF-κB is more extensive, NF-κB family members, p50 and p65, have both been found to repress ER activity at target genes (122, 123). In ovarian granulosa cells, high NF-κB activity was found be responsible for repression of ERβ transcriptional activity (124). In addition, Wang et al. (125) find that RelB expression inhibits ERα synthesis in breast cancer cells through the zinc finger repressor protein B lymphocyte induced maturation protein 1 (Blimp1), which leads to a more migratory phenotype of these cells.

ER and NF-κB: Positive cross talk

Less is known about positive cross talk between ER and NF-κB than transrepression, but there are currently two emerging concepts on how this might occur. First, ER and p65 can cooperate to synergistically regulate expression of specific genes that are important in modulating breast cancer cell survival and chemoresistance. Three examples of this type of cross talk are described below:

• 1) Prostaglandin-E synthase-1 (PTGES) is an enzyme that acts downstream of COX-1/2 to produce prostaglandin E2 (126), a factor that increases breast cancer cell motility and invasiveness, tumor metastasis and angiogenesis, and stromal expression of aromatase (47, 127–129). The combination of estrogen and proinflammatory cytokines, such as IL-1β or TNFα, synergistically up-regulates PTGES gene transcription. The synergistic effect was found to be ER and NF-κB dependent with enhanced binding of both ER and p65 to an ERE upstream of the PTGES gene promoter (130).

• 2) ABCG2 is a drug transporter that can cause efflux of a number of endogenous and exogenous agents and can confer a drug-resistant phenotype to breast cancer cells (89). TNFα potentiates the up-regulation of ABCG2 gene expression by 17β-estradiol, and this potentiation is mediated by cooperative binding of ER and p65 to adjacent response elements (88).

• 3) Baculoviral IAP repeat containing 3 (BIRC3)/cellular inhibitor of apoptosis protein 2 (cIAP-2) is a member of the inhibitor of apoptosis family of survival genes that is important for estrogen-dependent protection of breast cancer cells from TNFα-induced cell death (131). TNFα-dependent expression of BIRC3 is potentiated by 17β-estradiol through adjacent ERE and NF-κB-response elements, and NF-κB is required for histone acetylation and subsequent ER recruitment to the region (132).

A second mechanism by which positive cross talk between ER and the NF-κB pathway may occur is through interactions between ER and IKKα at the promoters of estrogen-regulated genes. Many of these genes are known to drive proliferation and cell-cycle progression in breast cancer cells, and in fact, knockdown of IKKα inhibits estrogen-mediated proliferation (98). IKKα has been shown to phosphorylate ER on serine 118 (98, 133), and after phosphorylation, both proteins are recruited to the promoter of cyclin D1. At this gene, they form a complex with steroid receptor coactivator 3 (SRC-3), leading to the phosphorylation of histone H3 and additional histone modifications that favor gene expression (98). A later study revealed that ER, IKKα, and SRC-3 are also recruited to the promoter of another essential cell-cycle regulator, E2F1 (99). Together, these findings suggest that ER and IKKα interaction could be particularly important for breast cancer cell proliferation.

Summary and Conclusions

Although ER-positive breast tumors generally have a better prognosis than ER-negative tumors, they actually display varying levels of responsiveness to endocrine therapy. Assays such as PAM50 and Oncotype DX may be able to predict a Luminal B subtype or recurrence after endocrine therapy, but they provide little insight into the biology that underlies these ER-positive tumors with more aggressive phenotypes. As reviewed here, an inflammatory microenvironment within the breast may be a contributing factor (Fig. 1). A positive correlation between TAMs and tumor angiogenesis and metastasis is established. Proinflammatory cytokines, which are produced not only by TAMs but also by cancer cells, adipocytes, and fibroblasts, can act to increase tumor proliferation, invasion, metastasis, and resistance to chemotherapy and endocrine therapy. Further exacerbating the problem, cytokines and prostaglandins can stimulate expression of aromatase, which is expressed in multiple cell types within the mammary tumor microenvironment, thereby leading to elevated local estrogen production.

Estrogens and cytokines can activate ER and the NF-κB pathway, respectively, within breast cancer cells to modulate the expression of a wide variety of genes and influence tumor phenotype. Although many genes are regulated by ER or NF-κB individually, subsets of genes can be regulated by either repressive or positive cross talk between the two factors, with both types of cross talk having the potential to influence tumor biology. Hypothetically, the ability of ER to repress NF-κB activity could explain why some ER-positive tumors, particularly of the Luminal A subtype, have a good prognosis. These would be tumors characterized by high ER expression, low NF-κB activity, low proinflammatory cytokine production, and less infiltration of TAMs into the tumor. On the other hand, NF-κB repression of ER could explain why some ER-positive tumors, such as Luminal B, are associated with poor disease-free and overall survival. The combination of low ER expression and high NF-κB activity could lead to increased migration, invasion, and metastatic characteristics of breast cancer cells. Positive cross talk between ER and NF-κB suggests these factors can also work together to promote a more aggressive type of ER-positive tumor. Cooperativity between ER and p65 and interaction between ER and IKKα can increase the expression of particular genes that are involved in proliferation, cell survival, and chemoresistance. However, the extent to which these modes of cross talk between ER and NF-κB play a role in ER-positive breast cancer biology has yet to be established. It is also unclear why ER and NF-κB can both repress and enhance each other's activity in the same cell. Evidence thus far suggests that the type of cross talk that occurs at a particular gene depends on the nature of that gene's regulatory elements. The arrangement, sequence, and accessibility of response elements, as well as the composition of the transcriptional complex formed at a specific regulatory element, could determine which type of cross talk occurs.

Overall, we propose a model in which risk factors, such as obesity and mammary gland involution, promote an inflammatory microenvironment within the breast to incite an aggressive phenotype of ER-positive breast cancer (Fig. 1). We suggest that cross talk between ER and the NF-κB pathway, including both transrepression and positive cross talk, could modulate gene expression within breast cancer cells to promote this aggressive phenotype. The molecular cross talk between ER and NF-κB, in addition to the numerous other effects that inflammatory factors can have on breast cancer cells, raises the possibility that cotargeting ER and inflammation may be an effective therapeutic strategy for some women with more aggressive, endocrine-resistant ER-positive tumors.