**5. Excessive IL-6 in human breast cancer**

Aberrantly elevated IL-6 is associated with a poor prognosis in breast cancer (Bachelot *et al.*, 2003; Salgado *et al.*, 2003; Zhang and Adachi, 1999). Human breast tumors produce more IL-6 when compared to matched healthy breast tissue, and tumor IL-6 levels concurrently increase with tumor grade. In addition, increased serum IL-6 has been demonstrated in

Interleukin-6 in the Breast Tumor Microenvironment 171

Mesenchymal stem cells (MSC) are a bone marrow-derived fibroblast cell population that can be recruited to the breast tumor stroma, acquire a TAF phenotype, and produce high levels of IL-6. MSC enhance the growth of ERα-positive breast cancer cells, which do not express IL-6 or activated STAT3. In contrast, MSC have no effect on IL-6-producing ERαnegative breast cancer cells, which express constitutively activated STAT3. Moreover, ERαpositive breast cancer cells orthotopically co-injected with MSC or MSC conditioned medium and ERα-positive breast cancer cells that ectopically express IL-6 demonstrate enhanced xenograft tumor growth in the absence of exogenous 17β-estradiol (Sasser *et al.*, 2007). Similar differential growth enhancement was demonstrated *in vivo* with ERα-positive and ERα-negative breast cancer cells co-injected with MSC, which also promoted metastasis (Karnoub *et al.*, 2007). Interestingly, IL-6 has been reported to facilitate the recruitment of MSC to hypoxic breast tumor microenvironments (Rattigan *et al.*, 2010). Likewise, IL-6 secreted from breast cancer cells has been shown to contribute to a recently characterized phenomenon termed "self-seeding" in which aggressive circulating tumor cells engraft within their original xenograft tumor (Kim *et al.*, 2009). MSC have also been shown to mediate the self-renewal capacity of breast cancer stem cells, in part, through a reciprocal IL-6 loop (Liu *et al.*, 2010). Taken together, preceding evidence strongly suggests that IL-6 promotes breast cancer cell growth by activating STAT3, which culminates with the upregulation of proliferative oncogenes such as c-Myc and cyclin D1 and and growth factors such as IL-6, hepatocyte growth factor (HGF), vascular endothelial growth factor (VEGF),

**7. IL-6 promotes epithelial-mesenchymal transition in breast cancer cells** 

Normal polarized epithelial cells exhibit 'cobblestone' homophilic morphology and express E-cadherin, which is required for epithelial cell polarization, phenotype, and consequent homeostasis (Jeanes *et al.*, 2008). E-cadherin is a key prognostic molecular biomarker clinically utilized to predict the metastatic propensity of breast cancer. Whereas very few studies have failed to demonstrate E-cadherin as an independent prognostic biomarker in breast cancer patients (Lipponen *et al.*, 1994; Parker *et al.*, 2001), the overwhelming majority of relevant studies have revealed E-cadherin as one of the strongest predictors of patient survival. Specifically, impaired E-cadherin expression in human breast tumors correlates with enhanced invasiveness, metastatic potential (Oka *et al.*, 1993), and decreased breast cancer patient survival (Heimann and Hellman, 2000; Pedersen *et al.*, 2002). While appropriate E-cadherin function is essential to the maintenance of epithelial cell morphology, phenotype, and homeostasis, regulation of E-cadherin expression is of equal importance. *CDH1*, the gene that encodes E-cadherin, is located on human chromosome 16q22.1 (Rakha *et al.*, 2006) and is susceptible to inactivation by promoter hypermethylation, somatic mutation, or aberrant overexpression of repressive transcription factors including Twist, Snail, and Slug among others (Hirohashi, 1998). Likewise, E-cadherin loss of function can arise due to extracellular domain-specific proteolytic cleavage. Although uncommon, germline mutations of *CDH1* predispose individuals to hereditary diffuse gastric cancer (HDGC) syndrome, and a proportion of these patients present with other cancers, including

E-cadherin was initially termed uvomorulin in mice and L-CAM in chicks following its discovery as a 120 kDa calcium-dependent trypsin-labile cell surface glycoprotein required for intercellular adhesion in mouse blastomeres (Hyafil *et al.*, 1981) and chick embryos

and epidermal growth factor (EGF) (Yu *et al.*, 2009a).

breast cancer (Guilford, 1999).

breast cancer patients compared to normal donors and correlates with advanced breast tumor stage (Kozlowski *et al.*, 2003) and increased number of metastatic sites (Salgado *et al.*, 2003). Furthermore, a single nucleotide polymorphism (SNP) exists at position -174 in the IL-6 gene promoter region, noted as IL-6 (-174 G>C), with the following allele frequency in a Caucasion population: 36% G/G, 44% G/C, and 18% C/C. An inflammatory stimulus such as *Salmonella typhii* vaccination induced higher serum IL-6 in those individuals with the G/G allele (Bennermo *et al.*, 2004). Although the IL-6 (-174 G>C) SNP is not associated with increased risk of developing breast cancer (Gonzalez-Zuloeta Ladd *et al.*, 2006; Litovkin *et al.*, 2007; Yu *et al.*, 2009b), it is significantly associated with disease-free and overall survival in breast cancer patients (DeMichele *et al.*, 2003).

ERα is expressed in luminal subtype breast tumors (Perou *et al.*, 2000) and therefore associated with improved patient survival (Buyse *et al.*, 2006; Sorlie *et al.*, 2001). A clear and well-characterized inverse correlation exists between breast cancer ERα status and IL-6. In fact, ERα directly binds to NF-κB, thus preventing transactivation of *IL6* gene expression (Galien and Garcia, 1997), which demonstrates a direct mechanism for such a correlation. Furthermore, ERα-negative human breast tumors produce more IL-6 than tumors that express ERα (Chavey *et al.*, 2007), and IL-6 serum levels are higher in ERα-negative breast cancer patients compared to ERα-positive patients (Jiang *et al.*, 2000). Likewise, ERαnegative breast cancer cell lines produce autocrine IL-6 whereas ERα-positive breast cancer cell lines do not (Sasser *et al.*, 2007). Therefore, this strongly suggests that ERα-negative breast cancer cells would exploit both paracrine (i.e., stromal cell-derived) and autocrine IL-6 signaling, whereas ERα-positive breast cancer cells could only utilize paracrine IL-6 signaling. In addition, ERα-negative breast cancer patients, whose tumors produce more IL-6 than those that express ERα (Chavey *et al.*, 2007), showed no difference in survival between the G/G allele (higher inducible serum IL-6) and any C allele (lower inducible serum IL-6) at the IL-6 (-174 G>C) promoter SNP. In contrast, ERα-positive breast cancer patients with any C allele at the IL-6 (-174 G>C) promoter SNP demonstrated improved disease-free and overall survival compared to those with the G/G allele (DeMichele *et al.*, 2003).

#### **6. IL-6 promotes breast cancer cell growth**

Stromal fibroblasts isolated from multiple types of tumors (i.e., TAF) or cancers (i.e., CAF) are now appreciated as influential players in cancer progression and metastasis (Orimo and Weinberg, 2006). CAF derived from multiple cancer types, including murine mammary cancers, exhibit an activated, proinflammatory phenotype with increased IL-6 production (Erez *et al.*, 2010). Furthermore, work from our laboratory has demonstated that fibroblasts isolated from breast tissue and common sites of breast cancer metastasis such as bone and lung enhance the growth of breast cancer cells in an IL-6-dependent manner, and IL-6 is the major fibroblast-derived soluble factor that induced STAT3 activation in breast cancer cells (Sasser *et al.*, 2007; Studebaker *et al.*, 2008). MDA-MB-231 breast cancer cells are commonly utilized to model triple negative breast cancer and produce autocrine IL-6. MDA-MB-231 cells expressing a dominant negative isoform of gp130 lacked constitutively active STAT3 and exhibited impaired tumorigenicity in an orthotopic xenograft model (Selander *et al.*, 2004), thus suggesting that IL-6 may drive tumor progression in this model. In addition, STAT3 is estimated to be constitutively activated in more than half of primary breast cancers due to IL-6 signaling (Berishaj *et al.*, 2007).

breast cancer patients compared to normal donors and correlates with advanced breast tumor stage (Kozlowski *et al.*, 2003) and increased number of metastatic sites (Salgado *et al.*, 2003). Furthermore, a single nucleotide polymorphism (SNP) exists at position -174 in the IL-6 gene promoter region, noted as IL-6 (-174 G>C), with the following allele frequency in a Caucasion population: 36% G/G, 44% G/C, and 18% C/C. An inflammatory stimulus such as *Salmonella typhii* vaccination induced higher serum IL-6 in those individuals with the G/G allele (Bennermo *et al.*, 2004). Although the IL-6 (-174 G>C) SNP is not associated with increased risk of developing breast cancer (Gonzalez-Zuloeta Ladd *et al.*, 2006; Litovkin *et al.*, 2007; Yu *et al.*, 2009b), it is significantly associated with disease-free and overall survival

ERα is expressed in luminal subtype breast tumors (Perou *et al.*, 2000) and therefore associated with improved patient survival (Buyse *et al.*, 2006; Sorlie *et al.*, 2001). A clear and well-characterized inverse correlation exists between breast cancer ERα status and IL-6. In fact, ERα directly binds to NF-κB, thus preventing transactivation of *IL6* gene expression (Galien and Garcia, 1997), which demonstrates a direct mechanism for such a correlation. Furthermore, ERα-negative human breast tumors produce more IL-6 than tumors that express ERα (Chavey *et al.*, 2007), and IL-6 serum levels are higher in ERα-negative breast cancer patients compared to ERα-positive patients (Jiang *et al.*, 2000). Likewise, ERαnegative breast cancer cell lines produce autocrine IL-6 whereas ERα-positive breast cancer cell lines do not (Sasser *et al.*, 2007). Therefore, this strongly suggests that ERα-negative breast cancer cells would exploit both paracrine (i.e., stromal cell-derived) and autocrine IL-6 signaling, whereas ERα-positive breast cancer cells could only utilize paracrine IL-6 signaling. In addition, ERα-negative breast cancer patients, whose tumors produce more IL-6 than those that express ERα (Chavey *et al.*, 2007), showed no difference in survival between the G/G allele (higher inducible serum IL-6) and any C allele (lower inducible serum IL-6) at the IL-6 (-174 G>C) promoter SNP. In contrast, ERα-positive breast cancer patients with any C allele at the IL-6 (-174 G>C) promoter SNP demonstrated improved disease-free and overall survival compared to those with the G/G allele (DeMichele *et al.*,

Stromal fibroblasts isolated from multiple types of tumors (i.e., TAF) or cancers (i.e., CAF) are now appreciated as influential players in cancer progression and metastasis (Orimo and Weinberg, 2006). CAF derived from multiple cancer types, including murine mammary cancers, exhibit an activated, proinflammatory phenotype with increased IL-6 production (Erez *et al.*, 2010). Furthermore, work from our laboratory has demonstated that fibroblasts isolated from breast tissue and common sites of breast cancer metastasis such as bone and lung enhance the growth of breast cancer cells in an IL-6-dependent manner, and IL-6 is the major fibroblast-derived soluble factor that induced STAT3 activation in breast cancer cells (Sasser *et al.*, 2007; Studebaker *et al.*, 2008). MDA-MB-231 breast cancer cells are commonly utilized to model triple negative breast cancer and produce autocrine IL-6. MDA-MB-231 cells expressing a dominant negative isoform of gp130 lacked constitutively active STAT3 and exhibited impaired tumorigenicity in an orthotopic xenograft model (Selander *et al.*, 2004), thus suggesting that IL-6 may drive tumor progression in this model. In addition, STAT3 is estimated to be constitutively activated in more than half of primary breast cancers

in breast cancer patients (DeMichele *et al.*, 2003).

**6. IL-6 promotes breast cancer cell growth** 

due to IL-6 signaling (Berishaj *et al.*, 2007).

2003).

Mesenchymal stem cells (MSC) are a bone marrow-derived fibroblast cell population that can be recruited to the breast tumor stroma, acquire a TAF phenotype, and produce high levels of IL-6. MSC enhance the growth of ERα-positive breast cancer cells, which do not express IL-6 or activated STAT3. In contrast, MSC have no effect on IL-6-producing ERαnegative breast cancer cells, which express constitutively activated STAT3. Moreover, ERαpositive breast cancer cells orthotopically co-injected with MSC or MSC conditioned medium and ERα-positive breast cancer cells that ectopically express IL-6 demonstrate enhanced xenograft tumor growth in the absence of exogenous 17β-estradiol (Sasser *et al.*, 2007). Similar differential growth enhancement was demonstrated *in vivo* with ERα-positive and ERα-negative breast cancer cells co-injected with MSC, which also promoted metastasis (Karnoub *et al.*, 2007). Interestingly, IL-6 has been reported to facilitate the recruitment of MSC to hypoxic breast tumor microenvironments (Rattigan *et al.*, 2010). Likewise, IL-6 secreted from breast cancer cells has been shown to contribute to a recently characterized phenomenon termed "self-seeding" in which aggressive circulating tumor cells engraft within their original xenograft tumor (Kim *et al.*, 2009). MSC have also been shown to mediate the self-renewal capacity of breast cancer stem cells, in part, through a reciprocal IL-6 loop (Liu *et al.*, 2010). Taken together, preceding evidence strongly suggests that IL-6 promotes breast cancer cell growth by activating STAT3, which culminates with the upregulation of proliferative oncogenes such as c-Myc and cyclin D1 and and growth factors such as IL-6, hepatocyte growth factor (HGF), vascular endothelial growth factor (VEGF), and epidermal growth factor (EGF) (Yu *et al.*, 2009a).
