**4. Targeting chemokines for breast cancer metastasis**

As a consequence of studies focusing almost exclusively on cancer cells, nearly all of the currently used cancer therapeutic agents target the cancer cells that, due to their inherent genomic instability, frequently acquire therapeutic resistance (Rajagopalan, et al., 2003). In part due to frequent therapeutic failures during the course of treatment of advanced stage tumors, increasing emphasis has been placed on targeting various stromal cells, particularly endothelial cells, via therapeutic interventions. Since these cells are thought to be normal and genetically stable, they are less likely to develop acquired resistance to cancer therapy. Thus, isolating, and characterizing each cell type (epithelial, myoephitelial, and various stromal cells) comprising non-malignant and cancerous breast tissue would not only help us to understand the role these cells play in breast tumorigenesis, but would likely give us new molecular targets for cancer intervention and treatment.

There is now an abundant literature documenting the associations of chemokine receptors with various types of cancer (Zlotnik, 2006) and their importance to mediate the establishment or development of metastatic foci. In fact, some anticancer drugs currently in use -like Herceptin- may involve the downregulation of chemokine receptors as part of their mechanism of action (Li, et al., 2004). This would provide the ultimate validation of the hypothesis, and would also point to future opportunities for therapeutic intervention as we discussed below. Current therapies such as surgery, radiotherapy and chemotherapy are primarily concerned with destruction of cancer. Targeting chemokines and chemokine receptors will allow limiting angiogenesis or metastasis and may enable such therapies to act as chemotherapeutic agents alone or in synergism with conventional agents. The upregulation of certain chemokine molecules in tumor as compared with normal cells offers a potential avenue—where cancer cells and their metastases can be specifically targeted. This selective destruction of cells is also pre-requisite of non-toxic treatment regimens.

Manipulation of the tumor microenvironment by treatment with chemokines can be used to recruit either immature dendritic cells for the initiation of anti-tumor responses or effector cells for cytotoxic responses. Intratumoral delivery of CCL21 using pox virus vaccine into established tumors derived from murine colon cancer line, CT26 results in enhanced infiltration of CD4 T cells which correlated with inhibition of tumor growth (Flanagan, et al., 2004). Non-immunogenic murine breast carcinoma is rejected after transducing cells with CCL19. The rejection of tumor was mediated by activated NK and CD4+ cells (Braun, et al., 2000). Adenoviral delivery of the CCL16 is able to inhibit growth of mammary tumors and prevent metastatic growth (Okada, et al., 2004). Importantly, in treatment involving delivery of chemokines to the tumor environment, there is a major problem of heterogeneity of the tumor cells. Chemokines may have dual effects, can be beneficial to one patient might be

Novel Insights Into the Role of Inflammation in Promoting Breast Cancer Development 139

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

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

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

natural barrier or a non-permissive environment for tumor progression.

shown to have a diagnostic value (Adams et al., 2002; Guttery et al., 2010).

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 (Seaton, et al., 2009).

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 interventions.
