**2.1 What are mesenchymal stem cells?**

Mesenchymal stem cells, also called multipotent mesenchymal stromal cells, were first described as stromal cells residing in the bone marrow (Friedenstein et al., 1966). They have stem cell-like characteristics (Caplan, 1991; Friedenstein & Kuralesova, 1971), a fibroblastlike appearance and features different from cells of the haematopoietic lineages. Those features include the ability to differentiate to osteoblasts, chondrocytes and adipocytes (Friedenstein et al., 1974; Noth et al., 2002; Pittenger et al., 1999). MSCs may also play a role in haematopoiesis, as MSCs have been shown to be involved in forming niches for the haematopoietic stem cells and to regulate the activities of these cells (Ehninger & Trumpp, 2011; Mendez-Ferrer et al., 2010; Omatsu et al., 2010; Sacchetti et al., 2007). MSCs are rare in the bone marrow. Only 1 of 34,000 nucleated cells in this tissue were determined to be MSCs (Wexler et al., 2003). Though much is known about MSCs today, there are still no specific markers available that clearly define a cell as an MSC. In 2006, the International Society for Cellular Therapy published a list of minimal criteria instead (Dominici et al., 2006) that are now commonly used to identify MSCs. Among these criteria are two functional features, the potential to differentiate to osteoblasts, chondrocytes and adipocytes as mentioned above and the ability to adhere to plastic. The latter feature allows the separation of MSCs from the other bone marrow cell populations, as cells of the haematopoietic lineages are nonadherent cells (Beyer Nardi & da Silva Meirelles, 2006). Other critieria used to characterize

Involvement of Mesenchymal Stem Cells in Breast Cancer Progression 249

MSCs (Crisan et al., 2008). It is possible that MSCs from different sources may not be identical and may behave differently (Zhao et al., 2010). In fact, environmental conditions, such as the supply with growth factors or oxygen, have been shown to change the behavior of MSCs (Krinner et al., 2010; Sanchez et al., 2011). Even a population of MSCs derived from a single source may not be homogenous and may have different developmental potentials (Phinney, 2002). This hypothesis was confirmed by Wicha and his co-workers who demonstrated that MSCs from bone marrow contain at least two subpopulations, one that expresses and one that lacks the stem cell marker ALDH-1 (aldehyde dehydrogenase-1) (Liu

et al., 2011). These two subpopulations behaved also functionally different (see 4.3).

In addition to the ability to mature to osteoblasts, chondrocytes and adipocytes, MSCs are also capable of differentiating to fibroblasts (Mishra & Banerjee, 2011). This conversion may be of particular importance in cancer, where MSCs that colonize a cancerous lesion switch to a particular type of fibroblast-like cells, the carcinoma-associated fibroblast (CAF) (Mishra et al., 2008; Spaeth et al., 2009). This may have consequences for tumor progression (see 4.4). The differentiation potential of MSCs goes far beyond the ability to differentiate towards the mesodermal lineage (Uccelli et al., 2008). Differentiation of MSCs to cells of ectodermal and endodermal lineages have been demonstrated as well. E.g., MSCs derived from adipose tissue were shown to be able to differentiate to endothelial cells (Zuk et al., 2002), while pancreatic MSCs could become hepatocytes (Seeberger et al., 2006). In addition, MSCs from umbilical cord blood were shown to have the potential to switch to cells displaying features of neural cells (Li et al., 2005; Park et al., 2007; Tondreau et al., 2004), though, in some cases, the neural phenotype may have caused by fusions of MSCs with neurons (Krabbe et al., 2005; Wislet-Gendebien et al., 2005). Under certain conditions, MSCs can also become epithelial cells, such as lung or renal epithelium-like cells (Kale et al., 2003; Lin et al., 2003;

MSCs are believed to play an important role in wound healing. Chemokines and cytokines as secreted by inflammatory cells seem to chemoattract MSCs to injured tissues (Brooke et al., 2007). E.g., Kidd and his co-workers reported that, when inoculated into wounded mice, labeled MSCs were preferentially detected in wounds, whereas, in non-injured mice, MSCs settled in lung, liver and spleen (Kidd et al., 2009). MSCs are attracted to many types of organs after injury, such as heart after myocardial infarction (Barbash et al., 2003), kidney after glomeruli damage (Ito et al., 2001), injured muscles (Natsu et al., 2004), bleomycindamaged lung (Ortiz et al., 2003) and brain after stroke (Chen et al., 2001; Mahmood et al., 2003). Interestingly, homing to the injured brain could be specifically blocked by an antibody directed to the chemokine MCP-1 (monocyte chemotactic protein-1)/CCL2 (Wang et al., 2002) suggesting that MCP-1/CCL2 is an important chemoattractant for MSCs. In the injured tissue, MSCs were found to help to regenerate this tissue. MSCs accomplish this goal partly by directly converting to those cells specifically needed to restore the function of the tissue. It is therefore tempting to consider the MSC as a general repair cell (Dittmer, 2010). Numerous reports support this hypothesis. E.g., bone marrow-derived MSCs were demonstrated to facilitate healing of injured muscles by differentiating to muscle progenitor

**2.2 Plasticity of MSCs** 

Ortiz et al., 2003; Rojas et al., 2005).

**3. Attracted to wounds and cancer** 

**3.1 Tropism towards injured tissue: MSCs as "repair" cells** 

Fig. 1. Sources of MSCs. The cartoon depicts the different sources from which MSCs can be isolated (left), the cells that can convert to MSCs (right, bottom) and cells that display MSClike features (right, top). Details are described in the text. ALK-2 = activin-like kinase-2.

MSCs are the expression profiles of certain proteins. MSCs express CD105 (endoglin), CD73 (ecto 5'-nucleotidase) and CD90 (Thy-1) and are deficient of CD45 (pan-leukocyte marker), CD34 (marker for primitive haematopoietic progenitors and endothelial cells), CD14 and CD11 (marker for monocytes and macrophages), CD79 and CD19 (marker for B-cells) and HLA-DR (MSCs not stimulated by IFN-). Bone marrow is not the only source of MSCs, other tissues are suitable to isolate MSCs as well (Fig. 1). Among these tissues are human adipose tissue (Zuk et al., 2002), umbilical cord blood (Sun et al., 2010), fetal dermis tissue (Qiao et al., 2008a), pancreatic tissue (Seeberger et al., 2006) and breast milk (Patki et al., 2010). More MSC sources are expected (Ding et al., 2011). Recently, menstrual blood and endometrium have been shown to contain MSCs. It is likely that most MSCs found in other tissues originated from the bone marrow. However, there is also evidence that some tissues, such as the adipose tissue, may produce their own MSCs (Bianco, 2011; Zhao et al., 2010). The MSC pool of a tissue may be expanded by dedifferentiation of differentiated cells (Fig. 1). This has been demonstrated for vascular endothelial cells that, under certain conditions, can undergo endothelial-to-mesenchymal transition to convert to MSCs (Medici et al., 2010). Some tissue-specific MSCs may be known for many years by other names (Fig. 1). Adiposederived stromal cells or preadipocytes are likely to be MSCs residing in adipose tissue (Locke et al., 2011; Manabe et al., 2003; Zuk et al., 2002). Pericytes isolated from skeletal muscles or non-muscle tissues have recently be found to show the typical characteristics of MSCs (Crisan et al., 2008). It is possible that MSCs from different sources may not be identical and may behave differently (Zhao et al., 2010). In fact, environmental conditions, such as the supply with growth factors or oxygen, have been shown to change the behavior of MSCs (Krinner et al., 2010; Sanchez et al., 2011). Even a population of MSCs derived from a single source may not be homogenous and may have different developmental potentials (Phinney, 2002). This hypothesis was confirmed by Wicha and his co-workers who demonstrated that MSCs from bone marrow contain at least two subpopulations, one that expresses and one that lacks the stem cell marker ALDH-1 (aldehyde dehydrogenase-1) (Liu et al., 2011). These two subpopulations behaved also functionally different (see 4.3).

### **2.2 Plasticity of MSCs**

248 Breast Cancer – Focusing Tumor Microenvironment, Stem Cells and Metastasis

Fig. 1. Sources of MSCs. The cartoon depicts the different sources from which MSCs can be isolated (left), the cells that can convert to MSCs (right, bottom) and cells that display MSClike features (right, top). Details are described in the text. ALK-2 = activin-like kinase-2.

MSCs are the expression profiles of certain proteins. MSCs express CD105 (endoglin), CD73 (ecto 5'-nucleotidase) and CD90 (Thy-1) and are deficient of CD45 (pan-leukocyte marker), CD34 (marker for primitive haematopoietic progenitors and endothelial cells), CD14 and CD11 (marker for monocytes and macrophages), CD79 and CD19 (marker for B-cells) and HLA-DR (MSCs not stimulated by IFN-). Bone marrow is not the only source of MSCs, other tissues are suitable to isolate MSCs as well (Fig. 1). Among these tissues are human adipose tissue (Zuk et al., 2002), umbilical cord blood (Sun et al., 2010), fetal dermis tissue (Qiao et al., 2008a), pancreatic tissue (Seeberger et al., 2006) and breast milk (Patki et al., 2010). More MSC sources are expected (Ding et al., 2011). Recently, menstrual blood and endometrium have been shown to contain MSCs. It is likely that most MSCs found in other tissues originated from the bone marrow. However, there is also evidence that some tissues, such as the adipose tissue, may produce their own MSCs (Bianco, 2011; Zhao et al., 2010). The MSC pool of a tissue may be expanded by dedifferentiation of differentiated cells (Fig. 1). This has been demonstrated for vascular endothelial cells that, under certain conditions, can undergo endothelial-to-mesenchymal transition to convert to MSCs (Medici et al., 2010). Some tissue-specific MSCs may be known for many years by other names (Fig. 1). Adiposederived stromal cells or preadipocytes are likely to be MSCs residing in adipose tissue (Locke et al., 2011; Manabe et al., 2003; Zuk et al., 2002). Pericytes isolated from skeletal muscles or non-muscle tissues have recently be found to show the typical characteristics of In addition to the ability to mature to osteoblasts, chondrocytes and adipocytes, MSCs are also capable of differentiating to fibroblasts (Mishra & Banerjee, 2011). This conversion may be of particular importance in cancer, where MSCs that colonize a cancerous lesion switch to a particular type of fibroblast-like cells, the carcinoma-associated fibroblast (CAF) (Mishra et al., 2008; Spaeth et al., 2009). This may have consequences for tumor progression (see 4.4). The differentiation potential of MSCs goes far beyond the ability to differentiate towards the mesodermal lineage (Uccelli et al., 2008). Differentiation of MSCs to cells of ectodermal and endodermal lineages have been demonstrated as well. E.g., MSCs derived from adipose tissue were shown to be able to differentiate to endothelial cells (Zuk et al., 2002), while pancreatic MSCs could become hepatocytes (Seeberger et al., 2006). In addition, MSCs from umbilical cord blood were shown to have the potential to switch to cells displaying features of neural cells (Li et al., 2005; Park et al., 2007; Tondreau et al., 2004), though, in some cases, the neural phenotype may have caused by fusions of MSCs with neurons (Krabbe et al., 2005; Wislet-Gendebien et al., 2005). Under certain conditions, MSCs can also become epithelial cells, such as lung or renal epithelium-like cells (Kale et al., 2003; Lin et al., 2003; Ortiz et al., 2003; Rojas et al., 2005).

### **3. Attracted to wounds and cancer**

### **3.1 Tropism towards injured tissue: MSCs as "repair" cells**

MSCs are believed to play an important role in wound healing. Chemokines and cytokines as secreted by inflammatory cells seem to chemoattract MSCs to injured tissues (Brooke et al., 2007). E.g., Kidd and his co-workers reported that, when inoculated into wounded mice, labeled MSCs were preferentially detected in wounds, whereas, in non-injured mice, MSCs settled in lung, liver and spleen (Kidd et al., 2009). MSCs are attracted to many types of organs after injury, such as heart after myocardial infarction (Barbash et al., 2003), kidney after glomeruli damage (Ito et al., 2001), injured muscles (Natsu et al., 2004), bleomycindamaged lung (Ortiz et al., 2003) and brain after stroke (Chen et al., 2001; Mahmood et al., 2003). Interestingly, homing to the injured brain could be specifically blocked by an antibody directed to the chemokine MCP-1 (monocyte chemotactic protein-1)/CCL2 (Wang et al., 2002) suggesting that MCP-1/CCL2 is an important chemoattractant for MSCs. In the injured tissue, MSCs were found to help to regenerate this tissue. MSCs accomplish this goal partly by directly converting to those cells specifically needed to restore the function of the tissue. It is therefore tempting to consider the MSC as a general repair cell (Dittmer, 2010). Numerous reports support this hypothesis. E.g., bone marrow-derived MSCs were demonstrated to facilitate healing of injured muscles by differentiating to muscle progenitor

Involvement of Mesenchymal Stem Cells in Breast Cancer Progression 251

MCF-7 cells and mesenchymal (basal-B)-type MDA-MB-231 cells. In some investigations, also luminal A-type T47D, basal A-type MDA-MB-468, murine 4T1 breast cancer cells and primary human breast cancer were used. In all cases, breast cancer cells stimulated MSC migration. However, the chemoattractive potency differed among the different breast cancer cell subtypes. E.g., the highly invasive MDA-MB-231 cells were more potent than the weakly invasive MCF-7 cells in stimulating migration of MSCs *in vitro* and *in vivo* (Dittmer et al., 2009; Goldstein et al., 2010; Ritter et al., 2008). Hence, it seems that MSCs have a higher affinity to more aggressive tumors. It is well established that factors secreted by breast cancer cells are responsible for MSC attraction (Fig. 2). IL-6 is one factor that is secreted by breast cancer cells and acts as a chemoattractant for MSCs (Liu et al., 2011; Rattigan et al., 2010). In response to IL-6, MSCs not only enhance their migratory activity, but also secrete chemokines, such as CXCL7 (see 4.3) (Liu et al., 2011). Interestingly, hypoxic conditions as often found in tumors trigger breast cancer cells to produce more IL-6 which further enhances migration of MSCs (Rattigan et al., 2010). Hypoxia also affects MSCs directly in

Fig. 2. Chemoattraction of MSCs to breast cancer cells. Breast cancer-derived cytokines and growth factors stimulate MSCs to migrate towards the tumor. Irradiation or hypoxia increase the CCL2 or IL-6 secretion, respectively, by breast cancer cells. Basal-type breast cancer cells seem to produce more CCL-2 than luminal A-type breast cancer cells. IL-6(R) = interleukin-6 (receptor), FGF(R) = fibroblast growth factor (receptor), VEGF(R) = vascular

endothelial growth factor (receptor), HDGF = hepatoma-derived growth factor.

cells (Natsu et al., 2004). In bleomycin-injured lung, MSCs switched to a phenotype typical for lung epithelial cells (Ortiz et al., 2003; Rojas et al., 2005). In the damaged myocardium, bone marrow-derived MSCs converted to cardiomyocytes (Toma et al., 2002; Wang et al., 2001). In ischemically injured renal tubules, MSCs are able to become tubular epithelial cells (Kale et al., 2003; Lin et al., 2003). In kidneys after anti-Thy1 antibody-induced glomerulonephritis, MSCs have been shown to mature to mesangial cells (Ito et al., 2001). And in diabetic mice, MSCs induced the number of pancreatic islets to increase and enhanced insulin production (Hess et al., 2003; Lee et al., 2006). The affinity of MSCs to injured tissue can be utilized for therapy (Brooke et al., 2007; Tocci & Forte, 2003). MSCs can be used as vectors to deliver drugs to injured tissues. Examples are BDNF (brain-derived neurotrophic factor)- or insulin-secreting MSCs to improve recovery from stroke (Kurozumi et al., 2004) or to treat diabetes (Xu et al., 2007), respectively. MSCs have also been used in clinical trials (Herberts et al., 2011). Most of the clinical trials with MSCs were carried out to treat patients with heart disease (Prockop & Olson, 2007). In many cases, patients' conditions improved suggesting that MSCs have positive effects on tissue repair also in humans.

#### **3.2 Tropism towards cancer: MSCs are attracted to breast cancer lesions**

Given the fact that MSCs are entering wounds to facilitate tissue repair, MSCs are of great value to maintain body functions. However, the affinity of MSCs to wounds may be of disadvantage to people who are suffering from cancer. In support of the view that a tumor is a wound that never heals (Dvorak, 1986), MSCs were also found to be attracted to cancerous lesions (Kidd et al., 2009) where they may promote tumor progression. Importantly, wounds and cancers secrete a similar cocktail of inflammatory cytokines and chemokines (Kidd et al., 2008). Among them are MSC-attracting factors, such as the growth factors PDGF (platelet-derived growth factor) and IGF-1 (insulin-like growth factor-1), the cytokines IL-6 (interleukin-6) and IL-8 as well as the chemokines MCP-1/CCL2, RANTES/CCL5, MDC (macrophage-derived chemokine)/CCL22 and SDF-1 (stromal-derived factor-1)/CXCL12 (Dwyer et al., 2007,Ponte, 2007 #228; Kim et al., 2011; Liu et al., 2011). It was confirmed that MSCs express the corresponding receptors for these ligands, i.e. PDGFR (PDGF receptor), IGFR (insulin growth factor receptor), IL-6R, gp130, CXCR1, CCR2, CCR3, CCR4 and CXCR4 (Dwyer et al., 2007,Ponte, 2007 #228; Kim et al., 2011; Liu et al., 2011). The susceptibility of MSCs to chemoattractants can be enhanced by certain factors. E.g., TNF (tumor necrosis factor ) was shown to increase the response of MSCs to certain chemokines by upregulating the expression of the receptors CCR2, CCR3 and CCR4 (Ponte et al., 2007). Many studies demonstrated that MSCs are attracted by tumors. In one study, the bone marrow of a mouse was replaced by the bone marrow from a transgenic mouse that expressed beta-galactosidase and MSC migration monitored from the bone marrow towards a prostate tumor xenograft (Ishii et al., 2003). It was found that X-gal positive MSCs colonized the tumor and differentiated to fibroblasts and endothelial cells. In a similar experimental setting, Direkze and co-workers could show that MSCs enter pancreatic insulinoma and convert to myofibroblasts (Direkze et al., 2004). Also breast cancer cells have been shown to chemoattract MSCs *in vitro* as well as *in vivo* (Dittmer et al., 2009; Dwyer et al., 2007; Goldstein et al., 2010; Klopp et al., 2007; Lin et al., 2008; Ling et al., 2010; Liu et al., 2011; Mishra et al., 2008; Pulukuri et al., 2010; Rattigan et al., 2010; Ritter et al., 2008; Zielske et al., 2009). Most breast cancer studies with MSCs were performed with luminal A-type

cells (Natsu et al., 2004). In bleomycin-injured lung, MSCs switched to a phenotype typical for lung epithelial cells (Ortiz et al., 2003; Rojas et al., 2005). In the damaged myocardium, bone marrow-derived MSCs converted to cardiomyocytes (Toma et al., 2002; Wang et al., 2001). In ischemically injured renal tubules, MSCs are able to become tubular epithelial cells (Kale et al., 2003; Lin et al., 2003). In kidneys after anti-Thy1 antibody-induced glomerulonephritis, MSCs have been shown to mature to mesangial cells (Ito et al., 2001). And in diabetic mice, MSCs induced the number of pancreatic islets to increase and enhanced insulin production (Hess et al., 2003; Lee et al., 2006). The affinity of MSCs to injured tissue can be utilized for therapy (Brooke et al., 2007; Tocci & Forte, 2003). MSCs can be used as vectors to deliver drugs to injured tissues. Examples are BDNF (brain-derived neurotrophic factor)- or insulin-secreting MSCs to improve recovery from stroke (Kurozumi et al., 2004) or to treat diabetes (Xu et al., 2007), respectively. MSCs have also been used in clinical trials (Herberts et al., 2011). Most of the clinical trials with MSCs were carried out to treat patients with heart disease (Prockop & Olson, 2007). In many cases, patients' conditions improved suggesting that MSCs have positive effects on tissue repair also in

**3.2 Tropism towards cancer: MSCs are attracted to breast cancer lesions** 

Given the fact that MSCs are entering wounds to facilitate tissue repair, MSCs are of great value to maintain body functions. However, the affinity of MSCs to wounds may be of disadvantage to people who are suffering from cancer. In support of the view that a tumor is a wound that never heals (Dvorak, 1986), MSCs were also found to be attracted to cancerous lesions (Kidd et al., 2009) where they may promote tumor progression. Importantly, wounds and cancers secrete a similar cocktail of inflammatory cytokines and chemokines (Kidd et al., 2008). Among them are MSC-attracting factors, such as the growth factors PDGF (platelet-derived growth factor) and IGF-1 (insulin-like growth factor-1), the cytokines IL-6 (interleukin-6) and IL-8 as well as the chemokines MCP-1/CCL2, RANTES/CCL5, MDC (macrophage-derived chemokine)/CCL22 and SDF-1 (stromal-derived factor-1)/CXCL12 (Dwyer et al., 2007,Ponte, 2007 #228; Kim et al., 2011; Liu et al., 2011). It was confirmed that MSCs express the corresponding receptors for these ligands, i.e. PDGFR (PDGF receptor), IGFR (insulin growth factor receptor), IL-6R, gp130, CXCR1, CCR2, CCR3, CCR4 and CXCR4 (Dwyer et al., 2007,Ponte, 2007 #228; Kim et al., 2011; Liu et al., 2011). The susceptibility of MSCs to chemoattractants can be enhanced by certain factors. E.g., TNF (tumor necrosis factor ) was shown to increase the response of MSCs to certain chemokines by upregulating the expression of the receptors CCR2, CCR3 and CCR4 (Ponte et al., 2007). Many studies demonstrated that MSCs are attracted by tumors. In one study, the bone marrow of a mouse was replaced by the bone marrow from a transgenic mouse that expressed beta-galactosidase and MSC migration monitored from the bone marrow towards a prostate tumor xenograft (Ishii et al., 2003). It was found that X-gal positive MSCs colonized the tumor and differentiated to fibroblasts and endothelial cells. In a similar experimental setting, Direkze and co-workers could show that MSCs enter pancreatic insulinoma and convert to myofibroblasts (Direkze et al., 2004). Also breast cancer cells have been shown to chemoattract MSCs *in vitro* as well as *in vivo* (Dittmer et al., 2009; Dwyer et al., 2007; Goldstein et al., 2010; Klopp et al., 2007; Lin et al., 2008; Ling et al., 2010; Liu et al., 2011; Mishra et al., 2008; Pulukuri et al., 2010; Rattigan et al., 2010; Ritter et al., 2008; Zielske et al., 2009). Most breast cancer studies with MSCs were performed with luminal A-type

humans.

MCF-7 cells and mesenchymal (basal-B)-type MDA-MB-231 cells. In some investigations, also luminal A-type T47D, basal A-type MDA-MB-468, murine 4T1 breast cancer cells and primary human breast cancer were used. In all cases, breast cancer cells stimulated MSC migration. However, the chemoattractive potency differed among the different breast cancer cell subtypes. E.g., the highly invasive MDA-MB-231 cells were more potent than the weakly invasive MCF-7 cells in stimulating migration of MSCs *in vitro* and *in vivo* (Dittmer et al., 2009; Goldstein et al., 2010; Ritter et al., 2008). Hence, it seems that MSCs have a higher affinity to more aggressive tumors. It is well established that factors secreted by breast cancer cells are responsible for MSC attraction (Fig. 2). IL-6 is one factor that is secreted by breast cancer cells and acts as a chemoattractant for MSCs (Liu et al., 2011; Rattigan et al., 2010). In response to IL-6, MSCs not only enhance their migratory activity, but also secrete chemokines, such as CXCL7 (see 4.3) (Liu et al., 2011). Interestingly, hypoxic conditions as often found in tumors trigger breast cancer cells to produce more IL-6 which further enhances migration of MSCs (Rattigan et al., 2010). Hypoxia also affects MSCs directly in

Fig. 2. Chemoattraction of MSCs to breast cancer cells. Breast cancer-derived cytokines and growth factors stimulate MSCs to migrate towards the tumor. Irradiation or hypoxia increase the CCL2 or IL-6 secretion, respectively, by breast cancer cells. Basal-type breast cancer cells seem to produce more CCL-2 than luminal A-type breast cancer cells. IL-6(R) = interleukin-6 (receptor), FGF(R) = fibroblast growth factor (receptor), VEGF(R) = vascular endothelial growth factor (receptor), HDGF = hepatoma-derived growth factor.

Involvement of Mesenchymal Stem Cells in Breast Cancer Progression 253

MSCs may be able to transform to sarcoma cells (Burns et al., 2008; Gjerstorff et al., 2009; Li et al., 2009; Mohseny & Hogendoorn, 2011; Riggi et al., 2008; Rosland et al., 2009). Currently, there is a debate about whether the MSC and not a primitive neuroectodermal cell is the cell

**3.3 Immunosuppression by MSCs: Consequences for wound healing and cancer** 

It is well established that MSCs act anti-inflammatory by modulating the activities of cells of the innate and the adaptive immune system (Rasmusson, 2006; Uccelli et al., 2008; Yagi et al., 2010). Among the affected cells are antigen-presenting dendritic cells, tumor cell-targeting natural killer cells, neutrophils and B- as well as T-lymphocytes. MSCs block antigen presentations by dendritic cells (Jiang et al., 2005; Ramasamy et al., 2007), inhibit the proliferation of activated T-lymphocytes (Bartholomew et al., 2002; Di Nicola et al., 2002; Krampera et al., 2003; Rasmusson et al., 2005), activate regulatory T cells (Tregs) that suppress T-effector cells (Aggarwal & Pittenger, 2005; Patel et al., 2010; Selmani et al., 2008), inhibit the activity of cytotoxic T-lymphocytes (Rasmusson et al., 2003) and block the proliferation of natural killer cells (Aggarwal & Pittenger, 2005; Sotiropoulou et al., 2006; Spaggiari et al., 2008). Direct and indirect interactions of MSCs with immune cells are made responsible for the antiinflammatory activity of the MSCs (Uccelli et al., 2008). The indirect effects are mediated by a number of cyto- and chemokines as secreted by MSCs. Among them are TGF1 (transforming growth factor 1) which stimulates the proliferation of inhibitory Tregs (Patel et al., 2010), IL-6 shown to inhibit neutrophil proliferation (Raffaghello et al., 2008) and prostaglandin E2 that inhibits antigen presentation by dendritic cells as well as proliferation of T-effector cells (Aggarwal & Pittenger, 2005; Bartholomew et al., 2002; Di Nicola et al., 2002; Glennie et al., 2005; Jiang et al., 2005; Krampera et al., 2003; Ramasamy et al., 2007; Rasmusson et al., 2005; Selmani et al., 2008). In the mouse model, the anti-inflammatory effects of MSCs were also linked to increased phagocytosis and enhanced elimination of bacteria (Mei et al., 2010). However, due to differences in the anti-sepsis defense in mice and men, it is unclear whether these data allow the prediction of an MSC-induced anti-sepsis effect also in humans (Monneret, 2009). It is likely that, by down-modulating the immune response, MSCs prevent excessive inflammation in injuries. This is thought to be the second way by which MSCs facilitate regeneration of the injured tissue. While for that reason the anti-inflammatory effect of MSCs may be beneficial for a patient with an injury, it may be however detrimental to a cancer patient. By inducing local immunosuppression cancer-residing MSCs may help cancer

of origin of Ewing's sarcoma (Lin et al., 2011).

cells to escape immune surveillance.

**4.1 The cytokine cocktail secreted by MSCs** 

**4. Communication between MSCs and breast cancer cells** 

MSCs secrete a plethora of cytokines and chemokines. In addition to the immuno-regulatory proteins, such as TGF1, IL-6 and prostaglandin E2, MSCs produce many other interleukins, including IL-7, IL-8 and IL-9, CC-type chemokines (CCL1, 2, 5, 8, 11, 15, 16, 20, 22, 26, and 27), CXC-type chemokines (CXCL1, 5, 6, 10, 11, 12, 13, and 16) and other factors, such as TIMP (tissue inhibitor of metalloproteases) -1 and -2, TNF and , PDGF A and B, G-CSF (granulocyte colony-stimulating factor), HGF (hepatocyte growth factor), VEGF and angiopoietin (Parekkadan et al., 2007). The syntheses of these factors can be further stimulated. E.g., IL-6 induces the expression of CXCL7, which further enhances the

**progression** 

that it increases their proliferative activity and their expression of stem cell and differentiation markers (Grayson et al., 2006). Besides IL-6, breast cancer cell-derived FGF-2, VEGF (vascular endothelial growth factor), cyclophilin B and HDGF (hepatoma-derived growth factor) were demonstrated to induce migration of MSCs (Lin et al., 2003; Ritter et al., 2008). Another important tumor-derived chemoattractant was shown to be the chemokine MCP-1/CCL2 (Dwyer et al., 2007) which is recognized by MSCs via the receptor CCR2 (Lu et al., 2006; Wang et al., 2002). Interestingly, mesenchymal (basal B-type) MDA-MB-231 cells produce more MCP-1/CCL2 than luminal A-type T47D cells, which may explain why more aggressive breast cancer cells have a higher potential to stimulate MSC migration. In primary breast cancer, which contains both an epithelial and a stromal compartment, the stromal compartment seems to be the major source of MCP-1/CCL2 (Dwyer et al., 2007). Irradiation of tumors was found to increase the expression of MCP-1/CCL2 and, along with it, the potential to recruit MSCs to tumors (Zielske et al., 2009). This further supports the notion that MCP-1/CCL2 plays an important role in attracting MSCs to tumors. The efficiency of recruitment of MSCs to tumors may also depend on inherent features of MSCs. MSCs overexpressing uPA (urokinase plasminogen activator) have a higher ability to migrate towards breast and prostate cancer cells than their vector-treated counterparts (Pulukuri et al., 2010). Given their similar tropism to injuries and cancer (Kidd et al., 2009), MSCs are a promising tool for therapeutic intervention of cancer (Motaln et al., 2010) as much as they are for treating injuries. MSCs engineered to express anti-cancer drugs can be used as vectors to deliver toxic loads to tumor cells. In many studies with engineered MSCs, MSCs were forced to express TRAIL (tumor necrosis factor-related apoptosis-inducing ligand), a membrane protein that induces apoptosis of tumor cells, but not of normal cells (Walczak et al., 1999). Using mouse xenografts, it could be shown that TRAIL-expressing MSCs are able to eradicate many kinds of tumor cells, including glioma, cervival, pancreatic, colon and breast cancer cells (Grisendi et al., 2010; Loebinger et al., 2009; Menon et al., 2009; Sonabend et al., 2008; Yang et al., 2009). MSC-delivered TRAIL can induce apoptosis by upregulating caspase 8 (Grisendi et al., 2010). TRAIL-expressing MSCs were also able to attack metastatic breast cancer cells and to significantly reduce pulmonary metastatic load in mice (Loebinger et al., 2009). In contrast to recombinant TRAIL, which has a short half life in plasma, TRAIL-expressing MSCs allow prolonged TRAIL exposure (Grisendi et al., 2010). Other approaches use MSCs that were engineered to express IFN- (interferon-) or transduced with CRAds (conditionally replicating adenoviruses) (Dembinski et al., 2009; Ling et al., 2010; Stoff-Khalili et al., 2007). In another setting, MSCs were transfected with enzymes to locally convert a relatively non-toxic substance into a toxin. Examples are MSCs expressing HSV-TK (herpes simplex virus-thymidine kinase) which catalyses the conversion of the prodrug ganciclovir to a toxic compound (Conrad et al., 2011) and MSCs loaded with cytosine deaminase which induces the deamination of 5-fluorocytosine to the chemotherapeutic drug 5-fluorouracil (Kucerova et al., 2008; You et al., 2009). In both cases, the non-toxic prodrug was systemically administered to tumor-bearing mice. MSCs can also be engineered such that they boost immune responses to cancer cells. MSCs engineered to express Her2 (human epidermal receptor2), a receptor tyrosine kinase often overexpressed in breast cancer (Theillet, 2010), can act as antigen-presenting cells to induce an immune reaction against Her2-exposing breast cancer cells (Romieu-Mourez et al., 2010). However, it should be noted that caution should be exercised when using MSCs as therapeutic tools as

that it increases their proliferative activity and their expression of stem cell and differentiation markers (Grayson et al., 2006). Besides IL-6, breast cancer cell-derived FGF-2, VEGF (vascular endothelial growth factor), cyclophilin B and HDGF (hepatoma-derived growth factor) were demonstrated to induce migration of MSCs (Lin et al., 2003; Ritter et al., 2008). Another important tumor-derived chemoattractant was shown to be the chemokine MCP-1/CCL2 (Dwyer et al., 2007) which is recognized by MSCs via the receptor CCR2 (Lu et al., 2006; Wang et al., 2002). Interestingly, mesenchymal (basal B-type) MDA-MB-231 cells produce more MCP-1/CCL2 than luminal A-type T47D cells, which may explain why more aggressive breast cancer cells have a higher potential to stimulate MSC migration. In primary breast cancer, which contains both an epithelial and a stromal compartment, the stromal compartment seems to be the major source of MCP-1/CCL2 (Dwyer et al., 2007). Irradiation of tumors was found to increase the expression of MCP-1/CCL2 and, along with it, the potential to recruit MSCs to tumors (Zielske et al., 2009). This further supports the notion that MCP-1/CCL2 plays an important role in attracting MSCs to tumors. The efficiency of recruitment of MSCs to tumors may also depend on inherent features of MSCs. MSCs overexpressing uPA (urokinase plasminogen activator) have a higher ability to migrate towards breast and prostate cancer cells than their vector-treated counterparts (Pulukuri et al., 2010). Given their similar tropism to injuries and cancer (Kidd et al., 2009), MSCs are a promising tool for therapeutic intervention of cancer (Motaln et al., 2010) as much as they are for treating injuries. MSCs engineered to express anti-cancer drugs can be used as vectors to deliver toxic loads to tumor cells. In many studies with engineered MSCs, MSCs were forced to express TRAIL (tumor necrosis factor-related apoptosis-inducing ligand), a membrane protein that induces apoptosis of tumor cells, but not of normal cells (Walczak et al., 1999). Using mouse xenografts, it could be shown that TRAIL-expressing MSCs are able to eradicate many kinds of tumor cells, including glioma, cervival, pancreatic, colon and breast cancer cells (Grisendi et al., 2010; Loebinger et al., 2009; Menon et al., 2009; Sonabend et al., 2008; Yang et al., 2009). MSC-delivered TRAIL can induce apoptosis by upregulating caspase 8 (Grisendi et al., 2010). TRAIL-expressing MSCs were also able to attack metastatic breast cancer cells and to significantly reduce pulmonary metastatic load in mice (Loebinger et al., 2009). In contrast to recombinant TRAIL, which has a short half life in plasma, TRAIL-expressing MSCs allow prolonged TRAIL exposure (Grisendi et al., 2010). Other approaches use MSCs that were engineered to express IFN- (interferon-) or transduced with CRAds (conditionally replicating adenoviruses) (Dembinski et al., 2009; Ling et al., 2010; Stoff-Khalili et al., 2007). In another setting, MSCs were transfected with enzymes to locally convert a relatively non-toxic substance into a toxin. Examples are MSCs expressing HSV-TK (herpes simplex virus-thymidine kinase) which catalyses the conversion of the prodrug ganciclovir to a toxic compound (Conrad et al., 2011) and MSCs loaded with cytosine deaminase which induces the deamination of 5-fluorocytosine to the chemotherapeutic drug 5-fluorouracil (Kucerova et al., 2008; You et al., 2009). In both cases, the non-toxic prodrug was systemically administered to tumor-bearing mice. MSCs can also be engineered such that they boost immune responses to cancer cells. MSCs engineered to express Her2 (human epidermal receptor2), a receptor tyrosine kinase often overexpressed in breast cancer (Theillet, 2010), can act as antigen-presenting cells to induce an immune reaction against Her2-exposing breast cancer cells (Romieu-Mourez et al., 2010). However, it should be noted that caution should be exercised when using MSCs as therapeutic tools as MSCs may be able to transform to sarcoma cells (Burns et al., 2008; Gjerstorff et al., 2009; Li et al., 2009; Mohseny & Hogendoorn, 2011; Riggi et al., 2008; Rosland et al., 2009). Currently, there is a debate about whether the MSC and not a primitive neuroectodermal cell is the cell of origin of Ewing's sarcoma (Lin et al., 2011).

### **3.3 Immunosuppression by MSCs: Consequences for wound healing and cancer progression**

It is well established that MSCs act anti-inflammatory by modulating the activities of cells of the innate and the adaptive immune system (Rasmusson, 2006; Uccelli et al., 2008; Yagi et al., 2010). Among the affected cells are antigen-presenting dendritic cells, tumor cell-targeting natural killer cells, neutrophils and B- as well as T-lymphocytes. MSCs block antigen presentations by dendritic cells (Jiang et al., 2005; Ramasamy et al., 2007), inhibit the proliferation of activated T-lymphocytes (Bartholomew et al., 2002; Di Nicola et al., 2002; Krampera et al., 2003; Rasmusson et al., 2005), activate regulatory T cells (Tregs) that suppress T-effector cells (Aggarwal & Pittenger, 2005; Patel et al., 2010; Selmani et al., 2008), inhibit the activity of cytotoxic T-lymphocytes (Rasmusson et al., 2003) and block the proliferation of natural killer cells (Aggarwal & Pittenger, 2005; Sotiropoulou et al., 2006; Spaggiari et al., 2008). Direct and indirect interactions of MSCs with immune cells are made responsible for the antiinflammatory activity of the MSCs (Uccelli et al., 2008). The indirect effects are mediated by a number of cyto- and chemokines as secreted by MSCs. Among them are TGF1 (transforming growth factor 1) which stimulates the proliferation of inhibitory Tregs (Patel et al., 2010), IL-6 shown to inhibit neutrophil proliferation (Raffaghello et al., 2008) and prostaglandin E2 that inhibits antigen presentation by dendritic cells as well as proliferation of T-effector cells (Aggarwal & Pittenger, 2005; Bartholomew et al., 2002; Di Nicola et al., 2002; Glennie et al., 2005; Jiang et al., 2005; Krampera et al., 2003; Ramasamy et al., 2007; Rasmusson et al., 2005; Selmani et al., 2008). In the mouse model, the anti-inflammatory effects of MSCs were also linked to increased phagocytosis and enhanced elimination of bacteria (Mei et al., 2010). However, due to differences in the anti-sepsis defense in mice and men, it is unclear whether these data allow the prediction of an MSC-induced anti-sepsis effect also in humans (Monneret, 2009). It is likely that, by down-modulating the immune response, MSCs prevent excessive inflammation in injuries. This is thought to be the second way by which MSCs facilitate regeneration of the injured tissue. While for that reason the anti-inflammatory effect of MSCs may be beneficial for a patient with an injury, it may be however detrimental to a cancer patient. By inducing local immunosuppression cancer-residing MSCs may help cancer cells to escape immune surveillance.
