**5. Tumor stroma: A permissive substrate for breast cancer development and progression**

The stroma of carcinomas is an intricate ecosystem where heterogeneous cell populations coexist. This structural and functional connective tissue niche is inhabited by immune and inflammatory cells such as macrophages and monocytes, mesenchymal bone marrowderived stem cells, endothelial and pericyte cells, lipocytes, additional smooth muscle cells and activated fibroblastic cells known as myofibroblasts, which are believed to be responsible for producing and maintaining the altered extracellular matrix (ECM) (Beacham & Cukierman, 2005; Li et al., 2007; Xouri & Christian, 2010). It is well accepted that the altered and excessive deposition of ECM, which is part of a process named desmoplasia, is directly associated with rapid progression and bad prognosis in carcinomas such as breast, pancreas, colon and prostate to name a few (Beacham & Cukierman, 2005; Arendt et al., 2010; Franco et al., 2010). In fact, we and others have suggested that stroma progression could be staged (analogously to classic tumor staging) into discrete stromagenic stages (Bissell et al., 2002; Mueller & Fusenig, 2002; Beacham & Cukierman, 2005; Quiros et al., 2008; Castello-Cros et al., 2009). Briefly, under normal (i.e., homeostatic) conditions, the breast stroma maintains the tissue architecture where a specialized ECM rich in collagen IV and laminin-1 known as basement membrane (BM) demarks a barrier between epithelium and the mesenchyme (Gudjonsson et al., 2002). A particular feature of the glandular epithelium in breast tissue is that both alveolar and ductal epithelial cells are not in direct contact with the BM. Instead, they are supported by a monolayer of myoepithelial cells that resides in between. Myoepithelial cells play an important role in supporting epithelial cell differentiation and controlling proliferation and cell polarity. These cells secrete the BM proteins and together with adjacent stromal fibroblasts maintain the integrity of this

harmful to another. However, this problem can be circumvented by chemokine typing every tumor prior to deciding on an appropriate therapy regime. They may be used as an adjunct to increase the efficacy of currently available therapies. Targeting specific chemokines can also modulate tumor infiltrating leukocytes or angiogenesis. High CXCL8 expression levels render tumor cells highly tumorigenic, angiogenic and invasive (Chavey, et al., 2007, Freund, et al., 2003, Freund, et al., 2004). In a murine model of breast cancer treatment with Met-CCL5, an antagonist of CCR1 and CCR5 led to a reduction in the total number of infiltrating inflammatory cells, in particular a decrease in macrophage infiltration and reduced growth of tumors (Liang, et al., 2004, Robinson, et al., 2003). The 7-transmembrane structure of chemokine receptors makes them attractive targets for small molecule inhibitors

In summary, the exploration and manipulation of the chemokine network has just started and is likely to improve efficiency of current tumor therapies. However, since these chemotactic cytokines are also utilized in a plethora of normal interactions, caution is needed especially when extrapolating *in vitro* data into the clinical situation. Differences amongst tumor entities are obvious and the same chemokine/chemokine-receptor system seems to have divergent functions in different tumor entities. A more in-depth analysis of the real players in tumor immunosuppression, for example characterization of the subtypes of infiltrating immune cells and thorough analysis of the cytokine and chemokine milieu of primary tumors, will be necessary to pave the way for more efficient therapeutic

**5. Tumor stroma: A permissive substrate for breast cancer development and** 

The stroma of carcinomas is an intricate ecosystem where heterogeneous cell populations coexist. This structural and functional connective tissue niche is inhabited by immune and inflammatory cells such as macrophages and monocytes, mesenchymal bone marrowderived stem cells, endothelial and pericyte cells, lipocytes, additional smooth muscle cells and activated fibroblastic cells known as myofibroblasts, which are believed to be responsible for producing and maintaining the altered extracellular matrix (ECM) (Beacham & Cukierman, 2005; Li et al., 2007; Xouri & Christian, 2010). It is well accepted that the altered and excessive deposition of ECM, which is part of a process named desmoplasia, is directly associated with rapid progression and bad prognosis in carcinomas such as breast, pancreas, colon and prostate to name a few (Beacham & Cukierman, 2005; Arendt et al., 2010; Franco et al., 2010). In fact, we and others have suggested that stroma progression could be staged (analogously to classic tumor staging) into discrete stromagenic stages (Bissell et al., 2002; Mueller & Fusenig, 2002; Beacham & Cukierman, 2005; Quiros et al., 2008; Castello-Cros et al., 2009). Briefly, under normal (i.e., homeostatic) conditions, the breast stroma maintains the tissue architecture where a specialized ECM rich in collagen IV and laminin-1 known as basement membrane (BM) demarks a barrier between epithelium and the mesenchyme (Gudjonsson et al., 2002). A particular feature of the glandular epithelium in breast tissue is that both alveolar and ductal epithelial cells are not in direct contact with the BM. Instead, they are supported by a monolayer of myoepithelial cells that resides in between. Myoepithelial cells play an important role in supporting epithelial cell differentiation and controlling proliferation and cell polarity. These cells secrete the BM proteins and together with adjacent stromal fibroblasts maintain the integrity of this

(Seaton, et al., 2009).

interventions.

**progression** 

specialized gland (Gudjonsson et al., 2002; Polyak & Kalluri, 2010). Under physiological conditions, a normal stroma preserves and drives regular breast tissue morphogenesis (Kuperwasser et al., 2004) and, at the same time, suppresses the transformation of epithelial cells thus preventing the development of breast carcinoma *in situ* (CIS) and inhibiting progression towards invasive cancer (Hu et al., 2008). Although not much information is available to describe the mechanistic events responsible for normal stroma prevention of carcinoma progression, recent data suggests that the tumor microenvironment lacks the regulatory mechanisms that are necessary to maintain a normal epithelial phenotype (Postovit et al., 2008). As shown by interesting work conducted by Mintz and Illmensee in 1975 where they observed that a normal embryo microenvironment is repressive of teratoma tumorigenesis (Mintz & Illmensee, 1975), more recent work by Postovit *et al* looking at specific human embryonic stem cells-secreted factors also concluded that embryonic microenvironments can control and sustain a normal behaviour of invasive tumor cells (Postovit et al., 2008). In summary, one could state that the normal stroma is a natural barrier or a non-permissive environment for tumor progression.

In an effort to understand premature events that occur during stroma progression (i.e., stromagenesis (Cukierman, 2009)), researchers have used animal models where they have shown stromal cells alterations at early stages of tumorigenesis. For example, prostate smooth muscle cells, known to support homeostasis and epithelium differentiation and considered to be analogous to normal myoepithelial cells in breast, have been shown to undergo alterations during early tumorigenesis (Wong & Tam, 2002). Similar to myoepithelial cells, smooth muscle cells are also lost in advanced stages of tumor progression, but prior to this they lower the expression levels of differentiation markers such as myosin, desmin, and laminin (Wong & Tam, 2002). This fact strongly suggests the advent of a discrete intermediate state between normal and activated stroma. To this end, the up-regulated expression of proteins, such as fibroblast activation protein, has been suggested as potential markers of this intermediate or primed stromal stage (Mathew et al., 1995; Huber et al., 2003; Santos et al., 2009). Another such molecule is tenascin-C, an ECM protein expressed in breast cancer at early stages of the tumorigenesis, which has been shown to have a diagnostic value (Adams et al., 2002; Guttery et al., 2010).

Once the stroma becomes activated, many histological features are evident. This stage is commonly described by pathologists as desmoplasia and is characterized by increased interstitial ECM-deposition. The desmoplastic ECM is believed to be produced by a highly proliferating fibroblastic and alpha-smooth muscle actin (-SMA) expressing myofibroblastic cell population. It is common in many cancers including breast, and it can constitute up to 50% of the tumor mass (Kunz-Schughart & Knuechel, 2002a, b; Desmouliere et al., 2004). The altered architecture of the desmoplastic stroma reaction is characterized by the over expression of ECM proteins such as collagen I and differential spliced fibronectin isoforms such as EDA and EDB (Matsumoto et al., 1999; Desmouliere et al., 2004). The desmoplastic ECM is highly organized in a parallel fiber pattern, which is clearly oriented *in vivo* perpendicular to the tumor border (Provenzano et al., 2006). In fact, this particular feature of the tumor associated-ECM (TA-ECM) has been suggested to facilitate migration of breast cancer cells *in vitro*, in a 1-integrin dependent manner (Castello-Cros et al., 2009). Moreover, there is evidence to suggest that TA-ECM can induce a phenotypic switch upon naïve fibroblasts thus inducing a myofibroblastic (or activated) conformation (Amatangelo et al., 2005). To this end, in a xenograft model of human breast cancer, it was shown that

Novel Insights Into the Role of Inflammation in Promoting Breast Cancer Development 141

myofibroblastic phenotypes is also suggestive of the eliciting of different roles played by these cell populations at the tumor stroma. Interestingly, this heterogeneity could have been originated (i.e., differentiated) by the multiple cell lineages known to produce myofibroblastic CAFs. These are: local fibroblasts (Kalluri & Zeisberg, 2006), bone marrow recruited mesenchymal cells (Ishii et al., 2003; Goldstein et al., 2010), as well as endothelial and tumor (i.e., epithelial) cells (Petersen et al., 2003; Kalluri & Zeisberg, 2006; Zeisberg et al., 2007), among others. In all these cases, TGF-β has been closely associated with tumorinduced myofibroblastic activation or differentiation (Zeisberg et al., 2007; Hinz, 2010; Taylor et al., 2010). The myofibroblastic differentiation is a complex and not yet fully understood process that is believed to play a central role during breast tumorigenesis (Cukierman, 2004; McAllister & Weinberg, 2010). Even though a plethora of molecules has been implicated in regulating fibroblastic activation, the specific desmoplastic response in breast cancer is believed to be driven by four main groups of inducers; i) growth factors, ii) TA-ECM, iii) acute inflammation and iv) microenvironmental stress denoted by nutrient

i. Specific growth factor presence at the tumor microenvironment may constitute the most studied aspect believed to trigger a myofibroblastic switch of the otherwise quiescent homeostatic fibroblasts. Determined mainly *in vitro* by an increment in proliferation rate, induction of -SMA expression, and an up-regulation of ECM components, the growth factors most commonly implicated in this process are TGF-, TGF-, insulinlike growth factors I and II (TGF-I and TGF-II), the platelet-derived growth factor (PDGF), and the basic fibroblast growth factor (bFGF) (Beacham & Cukierman, 2005; Kalluri & Zeisberg, 2006; Rasanen & Vaheri, 2010; Xouri & Christian, 2010). Although many questions remain regarding specific triggers for breast cancer desmoplasia, work from Walker and Dearing implicated TGF-1, TGF-2 and TGF- receptor as vital contributors of breast tumorigenesis associated with a stromal increment of fibronectin and tenascin in the tumor stroma (Walker & Dearing, 1992; Walker et al., 1994). Moreover, TGF- known to induce myofibroblastic differentiation and to increase collagen I deposition during the wound healing process (Desmouliere et al., 2005), has also been implicated as a main factor in inducing breast cancer associated bone marrow-derived myofibroblasts differentiation (Goldstein et al., 2010). Similarly, PDGF has been shown to increase the breast myofibroblastic population by 30% while greatly increasing the amount of interstitial collagen I *in vivo* (Shao et al., 2000). In the context of epithelial to mesenchymal transition (EMT)-derived myofibroblasts, hepatocyte growth factor (HGF) and epidermal growth factor (EGF), in addition to the above-mentioned PDGF and TGF-, have also been implicated (Mimeault & Batra, 2007; Kalluri &

ii. Breast TA-ECMs' features are known to become altered in both their molecular composition (Chen, S.T. et al., 2008; Levental et al., 2009; Ronnov-Jessen & Bissell, 2009) and their architectural characteristics (Provenzano et al., 2006)*.* Together these two altered features can modulate tumorigenic behaviours of cancer cells and promote or delay the evolution of carcinomas in a permissive or restrictive manner (Ronnov-Jessen & Bissell, 2009; Cukierman & Bassi, 2010). In addition, it has been suggested that the physico-chemical characteristics of the ECM also affect the behaviour of mesenchymal cells (Discher et al., 2005). Fibroblasts are influenced by stromal stiffness, which exerts mechanical forces that modulate their cell behaviour. Thus, it has been demonstrated

and oxygen depravation as well as low pH.

Weinberg, 2009).

activated fibroblasts influence the local microenvironment to promote invasion (Orimo et al., 2005; Hu et al., 2008).
