**4. Proteome of cancer membrane vesicles**

Mass spectrometry-based proteomic tools coupled with advanced purification methods for exosomes, has allowed more in-depth proteome analyses, contributing immensely to our understanding of the molecular composition of exosomes. Proteomic analysis of exosomes from diverse cell types, including cancers has revealed a common set of membrane and cytosolic proteins, suggesting the evolutionary importance of these membrane particles. In addition, exosomes express an array of proteins that reflect the originating host cell. The excessive release of exosomes in tumor cells, as evidenced by their increased levels in body fluids during the late stage of a disease and their overexpression of certain tumor cell biomarkers, suggests an important role of exosomes in diagnosis and biomarker studies (Simpson *et al.*, 2009).

By proteomic analysis we can enrich low abundance membrane proteins from underrepresented conventional cell lysates and unfractionated biological fluids. Identification of a conserved set of common proteins that are essential for vesicle biogenesis, structure and trafficking mechanisms can be explored. We can also detect cell-specific biomarkers. These concepts suggest that analyzing the composition and abundance of such proteins in exosomes may be useful to reveal different cell behaviors.

The Application of Membrane Vesicles for Cancer Therapy 31

Characterization of urinary exosomal composition may be proposed as a potential source of diagnostic markers in bladder cancer. Two different approaches (Smalley *et al.*, 2008; Welton *et al.*, 2010) were taken to characterize these exosomes. In the first, urine exosomes were isolated from a limited number of individuals with bladder cancer and their protein composition compared to that of healthy controls. Eight proteins were found elevated, among which five have been linked to the EGF receptor pathway (Smalley *et al.*, 2008). In the second study, extensive steps were taken to produce high purity and quality-assured exosome preparations prior to beginning proteomics workflows. Working with conditioned media from cultured bladder cancer cell lines, 350 proteins were identified. Eighteen were proven to be present in exosomes isolated from the urine of three bladder cancer patients (Welton *et al.*, 2010). This suggests that conditioned media from cultured cell lines could represent an interesting starting model to detect exosomal proteomic alterations, which must then be confirmed *in vivo*, using biological fluids from a wide cohort of patients, in

Exosomes prepared from urine of prostate cancer patients contain typical markers of such a tumor (PSA and PCA3) (Mitchell *et al.*, 2009; Nilsson *et al.*, 2009). Moreover, -catenin immunoreactivity was identified in vesicles prepared from culture media of PC3 cells and found significantly increased in prostasomes of the urine of prostate cancer patients (Lu *et* 

Studies confirm that urinary exosome protein profiling is an important topic and may be a valuable tool for biomarker discovery in the field of urinary tract pathology. Proteomic approaches to investigate membranous vesicles and exosomes are still immature. However, they show a great potential for future developments in the diagnostic and prognostic

Tumor cells release large quantities of exosomes containing procoagulant, growth regulatory, and oncogenic cargo, which can be transferred throughout the cancer cell population and to transformed stromal cells, endothelial cells and possibly to the inflammatory infiltrates. These events likely impact tumor invasion, angiogenesis, metastasis, drug resistance, and cancer stem cells. Instead of physical contact, the influence of exosomes on the target cell may also involve pericellular discharge/activation of the bioactive cargo (Dolo *et al.*, 2005; Hendrix *et al.*, 2010; Muralidharan-Chari *et al.*, 2010). For instance, this may involve proteolytic remodelling of the extracellular microenvironment, modulation of ligand-receptor interactions, and a variety of other effects that could change the behaviour of target cells and properties of their surroundings (Hendrix *et al.*, 2010) **(Figure 5)**. In some instances, such interactions could be rather complex and multifactorial. The recently described Rab27B-regulated exosomal release of MMPs and HSP90a from metastatic cancer cells is believed to control invasive cellular behaviour by inducing changes in the extracellular matrix (ECM) as well as through modification of growth factor responses (Hendrix *et al.*, 2010). Likewise, procoagulant exosomes may facilitate tumor initiation, invasion, and dissemination by activating the clotting cascade extracellularly and coagulation-dependent signalling intracellularly (Milsom *et al.*, 2007). Exosome-mediated

**5. Targeting interactions of membrane vesicles with neighboring tumor** 

order to supply a non-invasive source for biomarker discovery.

*al.*, 2009).

**4.2 Summary** 

applications.

**microenvironment** 

#### **4.1 Proteomics of cell-type dependent exosomes**

Exosomes have a unique protein composition that varies depending on cellular origin **(Table 2)**. Analysis of exosomes from a wide variety of cells and body fluids have been identified. Exosomes from various cancer cells expose Fas ligand (FasL, CD95L), a ligand of the death receptor Fas (CD95), which induces T-cell apoptosis and diminishes the function of adaptive immune cells (Andreola *et al.*, 2002; Huber *et al.*, 2005). It was shown that a modest correlation exists between lymph node infiltration and tumor burden and the numbers of circulating FasL-exposing exosomes in blood from patients with oral squamous cell cancers (Kim *et al.*, 2005). Exosomes from lymphoblastoma cells exposed latent membrane protein-1 (LMP-1), another immune suppressing transmembrane protein, thereby inhibiting leukocyte proliferation. In addition, low numbers of circulating exosomes, in cancer patients, stained positive for MUC1, a cancer cell antigen, and glycoprotein IIIa (integrin β3), which is mainly present on platelets and platelet-derived exosomes. Exosomes are released after fusion of exosomes from malignant epithelial cells with platelets (Tesselaar *et al.*, 2007). Alternatively, platelet-derived exosomes were shown to transfer integrins to breast and lung cancer cells (Janowska-Wieczorek *et al.*, 2001; Janowska-Wieczorek *et al.*, 2005). Thus, cancer cells can fuse with non-cancer cell-derived exosomes, thereby receiving lipids and membrane specific proteins which may help them escape from immune surveillance.

Degradation of the extracellular matrix (ECM) is essential for tumor growth (Hotary *et al.*, 2003). Exosomes expose and contain proteases, including matrix metalloproteinase 2 (MMP-2) and MMP-9 and its zymogens, and urokinase-type plasminogen activator (uPA). Neovascularization is also responsible for the increased entry of tumor cells into the circulation and metastasis. It is believed that tumor and stromal cells secrete angiogenic factors such as VEGF and basic fibroblast growth factor (bFGF) and that tumor-associated angiogenesis occurs by the action of these factors (Carmeliet, 2005). Furthermore, growing evidence suggests that exosomes derived from tumor cells and platelets also possess angiogenic activities. Vesicular components such as sphingomyelin, CD147, tetraspanin-8, VEGF, and bFGF are likely involved in exosome-mediated neovascularization (Kim *et al.*, 2002; Brill *et al.*, 2004; Gesierich *et al.*, 2006; Millimaggi *et al.*, 2007).

In a colorectal cancer cell line study, several exosomal proteins have been identified that are believed to be involved in tumor-associated angiogenesis: ADAM 10, CD44, NG2, ephrin-B1, macrophage migration inhibitory factor (MIF), RACK1, and tetraspanin-8 (Dong-Sic *et al.*, 2007). Clinical exosome analysis may also prove useful for solid cancers (Mathivanan *et al.*, 2010). Using exosomes from ovarian carcinoma cell lines, malignant ascites and sera from ovarian carcinoma patients, it was found that malignant ascites-derived exosomes cargo tumor progression related proteins (L1CAM, CD24, ADAM10 and EMMPRIN). It was also observed that exosomes move systemically via the blood stream (Keller *et al.*, 2009). Therefore, if some membrane proteins are typically and specifically expressed by a certain tumor, their detection on circulating exosomes, (which could be isolated from only 1mL of blood), may be exploited for diagnostic purposes as an early signal of cancer presence. Proteomic analysis of exosomes was also performed on human mesothelioma cell lines and malignant pleural effusions. Bard and colleagues, described exosomes which contained antigen presenting molecules, cytoskeletal proteins, and signal tranduction-involved proteins were in mesothelioma, lung, breast, and ovarian cancers. In addition, SNx25, BTG1, PEDF, and Thrombospondin were also identified (Bard *et al.*, 2004).

Characterization of urinary exosomal composition may be proposed as a potential source of diagnostic markers in bladder cancer. Two different approaches (Smalley *et al.*, 2008; Welton *et al.*, 2010) were taken to characterize these exosomes. In the first, urine exosomes were isolated from a limited number of individuals with bladder cancer and their protein composition compared to that of healthy controls. Eight proteins were found elevated, among which five have been linked to the EGF receptor pathway (Smalley *et al.*, 2008). In the second study, extensive steps were taken to produce high purity and quality-assured exosome preparations prior to beginning proteomics workflows. Working with conditioned media from cultured bladder cancer cell lines, 350 proteins were identified. Eighteen were proven to be present in exosomes isolated from the urine of three bladder cancer patients (Welton *et al.*, 2010). This suggests that conditioned media from cultured cell lines could represent an interesting starting model to detect exosomal proteomic alterations, which must then be confirmed *in vivo*, using biological fluids from a wide cohort of patients, in order to supply a non-invasive source for biomarker discovery.

Exosomes prepared from urine of prostate cancer patients contain typical markers of such a tumor (PSA and PCA3) (Mitchell *et al.*, 2009; Nilsson *et al.*, 2009). Moreover, -catenin immunoreactivity was identified in vesicles prepared from culture media of PC3 cells and found significantly increased in prostasomes of the urine of prostate cancer patients (Lu *et al.*, 2009).

#### **4.2 Summary**

30 Advances in Cancer Therapy

Exosomes have a unique protein composition that varies depending on cellular origin **(Table 2)**. Analysis of exosomes from a wide variety of cells and body fluids have been identified. Exosomes from various cancer cells expose Fas ligand (FasL, CD95L), a ligand of the death receptor Fas (CD95), which induces T-cell apoptosis and diminishes the function of adaptive immune cells (Andreola *et al.*, 2002; Huber *et al.*, 2005). It was shown that a modest correlation exists between lymph node infiltration and tumor burden and the numbers of circulating FasL-exposing exosomes in blood from patients with oral squamous cell cancers (Kim *et al.*, 2005). Exosomes from lymphoblastoma cells exposed latent membrane protein-1 (LMP-1), another immune suppressing transmembrane protein, thereby inhibiting leukocyte proliferation. In addition, low numbers of circulating exosomes, in cancer patients, stained positive for MUC1, a cancer cell antigen, and glycoprotein IIIa (integrin β3), which is mainly present on platelets and platelet-derived exosomes. Exosomes are released after fusion of exosomes from malignant epithelial cells with platelets (Tesselaar *et al.*, 2007). Alternatively, platelet-derived exosomes were shown to transfer integrins to breast and lung cancer cells (Janowska-Wieczorek *et al.*, 2001; Janowska-Wieczorek *et al.*, 2005). Thus, cancer cells can fuse with non-cancer cell-derived exosomes, thereby receiving lipids and membrane specific proteins which may help them

Degradation of the extracellular matrix (ECM) is essential for tumor growth (Hotary *et al.*, 2003). Exosomes expose and contain proteases, including matrix metalloproteinase 2 (MMP-2) and MMP-9 and its zymogens, and urokinase-type plasminogen activator (uPA). Neovascularization is also responsible for the increased entry of tumor cells into the circulation and metastasis. It is believed that tumor and stromal cells secrete angiogenic factors such as VEGF and basic fibroblast growth factor (bFGF) and that tumor-associated angiogenesis occurs by the action of these factors (Carmeliet, 2005). Furthermore, growing evidence suggests that exosomes derived from tumor cells and platelets also possess angiogenic activities. Vesicular components such as sphingomyelin, CD147, tetraspanin-8, VEGF, and bFGF are likely involved in exosome-mediated neovascularization (Kim *et al.*,

In a colorectal cancer cell line study, several exosomal proteins have been identified that are believed to be involved in tumor-associated angiogenesis: ADAM 10, CD44, NG2, ephrin-B1, macrophage migration inhibitory factor (MIF), RACK1, and tetraspanin-8 (Dong-Sic *et al.*, 2007). Clinical exosome analysis may also prove useful for solid cancers (Mathivanan *et al.*, 2010). Using exosomes from ovarian carcinoma cell lines, malignant ascites and sera from ovarian carcinoma patients, it was found that malignant ascites-derived exosomes cargo tumor progression related proteins (L1CAM, CD24, ADAM10 and EMMPRIN). It was also observed that exosomes move systemically via the blood stream (Keller *et al.*, 2009). Therefore, if some membrane proteins are typically and specifically expressed by a certain tumor, their detection on circulating exosomes, (which could be isolated from only 1mL of blood), may be exploited for diagnostic purposes as an early signal of cancer presence. Proteomic analysis of exosomes was also performed on human mesothelioma cell lines and malignant pleural effusions. Bard and colleagues, described exosomes which contained antigen presenting molecules, cytoskeletal proteins, and signal tranduction-involved proteins were in mesothelioma, lung, breast, and ovarian cancers. In addition, SNx25, BTG1,

2002; Brill *et al.*, 2004; Gesierich *et al.*, 2006; Millimaggi *et al.*, 2007).

PEDF, and Thrombospondin were also identified (Bard *et al.*, 2004).

**4.1 Proteomics of cell-type dependent exosomes** 

escape from immune surveillance.

Studies confirm that urinary exosome protein profiling is an important topic and may be a valuable tool for biomarker discovery in the field of urinary tract pathology. Proteomic approaches to investigate membranous vesicles and exosomes are still immature. However, they show a great potential for future developments in the diagnostic and prognostic applications.

### **5. Targeting interactions of membrane vesicles with neighboring tumor microenvironment**

Tumor cells release large quantities of exosomes containing procoagulant, growth regulatory, and oncogenic cargo, which can be transferred throughout the cancer cell population and to transformed stromal cells, endothelial cells and possibly to the inflammatory infiltrates. These events likely impact tumor invasion, angiogenesis, metastasis, drug resistance, and cancer stem cells. Instead of physical contact, the influence of exosomes on the target cell may also involve pericellular discharge/activation of the bioactive cargo (Dolo *et al.*, 2005; Hendrix *et al.*, 2010; Muralidharan-Chari *et al.*, 2010). For instance, this may involve proteolytic remodelling of the extracellular microenvironment, modulation of ligand-receptor interactions, and a variety of other effects that could change the behaviour of target cells and properties of their surroundings (Hendrix *et al.*, 2010) **(Figure 5)**. In some instances, such interactions could be rather complex and multifactorial. The recently described Rab27B-regulated exosomal release of MMPs and HSP90a from metastatic cancer cells is believed to control invasive cellular behaviour by inducing changes in the extracellular matrix (ECM) as well as through modification of growth factor responses (Hendrix *et al.*, 2010). Likewise, procoagulant exosomes may facilitate tumor initiation, invasion, and dissemination by activating the clotting cascade extracellularly and coagulation-dependent signalling intracellularly (Milsom *et al.*, 2007). Exosome-mediated

The Application of Membrane Vesicles for Cancer Therapy 33

production was detected in association with increased oncogenic activity of protein kinase B (PKB/Akt), or upon stimulation with growth factors (EGF), and depending on the status of the actin regulating protein known as diaphanous related formin 3 (DRF3) (Di Vizio *et al.*, 2009). In this case, inhibition of DRF3 expression through RNA interference enhanced the rate of exosome formation, and membrane blebbing activity, suggesting that DRF3 may be an inhibitor of ectosome release (Di Vizio *et al.*, 2009). Interestingly, DRF3 expression is lost during the progression of prostate cancer to metastatic disease, which suggests an intriguing

link between oncogenesis, vesiculation and metastasis (Di Vizio *et al.*, 2009).

Fig. 5. Tumor microenvironment. The heterotypic interactions within the tumor

within the tumor microenvironment.

**5.3 Effect of TEX on tumor microenvironment** 

microenvironment and their give and take of exosomes and their contents provide many targets for possible therapy. The goal of targetting these interactions will interrupt the heterotypic signaling that would thus deprive the cancer cells of the support they have

Exosome release by colorectal cancer cells is a function of K-ras and p53 status (Yu *et al.*, 2005). It is noteworthy that oncoproteins not only stimulate exosome formation but also become incorporated into their cargo (Al-Nedawi *et al.*, 2008; Al-Nedawi *et al.*, 2009). As a result, oncogene-containing exosomes (sometimes refered to as *oncosomes*) may serve as vehicles that carry oncogenic cargo and mediate its transfer between cells (Al-Nedawi *et al.*, 2009). At least four different modes of such oncogenic transfer have been described: (a) intercellular passage of active oncoproteins (Al-Nedawi *et al.*, 2008), (b) transfer of oncogenic mRNA transcripts

emission of various factors including tetraspanins, chemoattractants, adhesion molecules and proteases from cancer cells, platelets, and other cellular sources contributes to metastatic regulation in several experimental systems (Janowska-Wieczorek *et al.*, 2005; Jung *et al.*, 2009). As mentioned earlier, exosomes may also act as important reservoirs of cytokines and mediators of inflammatory and immune responses (Bianco *et al.*, 2009; Thery *et al.*, 2009).

#### **5.1 Mediators of intercellular communication**

Contact with the cell death ligand (FasL) exposed on certain tumor cell-derived exosomes is lethal for Fas-expressing lymphoid cytotoxic effector cells, a process implicated in the induction of immunotolerance in colorectal cancer and possibly other malignancies (Albanese *et al.*, 1998). These influences may affect recipient cells via a random distribution of exosomes in tissue and body fluids, or more directional exosome homing/uptake mechanisms. For instance, an acidic pH commonly present in hypo-perfused areas of solid tumors may lead to localized disruption of exosomes and consequent discharge of their proangiogenic and pro-inflammatory cargo such as VEGF and other factors (Taraboletti *et al.*, 2006). Exosomes may also be directed to specific sites due to the molecular addresses they carry on their surfaces (Celi *et al.*, 2004; Zhou *et al.*, 2008; Xiao *et al.*, 2009). The nature, directionality, and efficiency of this molecular exchange depends on several factors. For instance, the physical properties of vesicular plasma membranes affect the fusion rate between exosomes and target cells, which may increase their exosome uptake under acidic pH (Parolini *et al.*, 2009). In some instances, exosome transfer could also be directed by specific molecular addresses, for example, a high concentration of phosphatydyl serines (PS) on the surface of certain exosomes (e.g. ectosomes or procoagulant microparticles) may enable their recognition by PS receptors (PSRs) on the surface of specific types of target cells. Many of such PSRs have been described within the context of phagocytosis of apoptotic cells by mononuclear cells; examples of such PSRs include Tim1, Tim4, stabilin 2 and BAI1 (Park *et al.*, 2007; Zhao *et al.*, 2010), at least some of which could be expressed more widely and may be involved in the uptake of exosomes. Indeed, blocking PSRs often obliterates exosome incorporation by endothelial cells, platelets and cancer cells (Del Conde *et al.*, 2005; Al-Nedawi *et al.*, 2008; Al-Nedawi *et al.*, 2009). A corollary to this point would be that phagocytes could be particularly susceptible to molecular influences of PS-positive exosomes, beyond their simple destruction. It has also been proposed that Tim1/4 receptors on two adjacent cells could allow formation of exosome bridges, thereby promoting additional indirect intercellular interactions (Xiao *et al.*, 2009). Similarly, the presence of PSGL-1 (P-selectin ligand) on the surface of procoagulant exosomes directs them to Pselectin-expressing platelets and endothelial cells (Thomas *et al.*, 2009).

#### **5.2 Oncosomes: Oncogene-driven vesiculation**

During malignant transformation, the action of mutant oncogenes, such as K-ras, EGFR, or its constitutively active mutant EGFR (variant III) (EGFRvIII), as well as several others, appear to stimulate the formation and release of exosomes (Yu *et al.*, 2005; Al-Nedawi *et al.*, 2008). Similarly, the activation or loss of specific tumor suppressor proteins appears to impact cellular vesiculation either positively or negatively (Yu *et al.*, 2005; Yu *et al.*, 2006). While the exact nature of the signalling pathways involved in oncogene-driven exosome biogenesis remains largely unknown, recent studies have begun to shed more light on the underlying processes. For instance, in cultures of prostate cancer cells, elevated exosome

emission of various factors including tetraspanins, chemoattractants, adhesion molecules and proteases from cancer cells, platelets, and other cellular sources contributes to metastatic regulation in several experimental systems (Janowska-Wieczorek *et al.*, 2005; Jung *et al.*, 2009). As mentioned earlier, exosomes may also act as important reservoirs of cytokines and mediators of inflammatory and immune responses (Bianco *et al.*, 2009; Thery

Contact with the cell death ligand (FasL) exposed on certain tumor cell-derived exosomes is lethal for Fas-expressing lymphoid cytotoxic effector cells, a process implicated in the induction of immunotolerance in colorectal cancer and possibly other malignancies (Albanese *et al.*, 1998). These influences may affect recipient cells via a random distribution of exosomes in tissue and body fluids, or more directional exosome homing/uptake mechanisms. For instance, an acidic pH commonly present in hypo-perfused areas of solid tumors may lead to localized disruption of exosomes and consequent discharge of their proangiogenic and pro-inflammatory cargo such as VEGF and other factors (Taraboletti *et al.*, 2006). Exosomes may also be directed to specific sites due to the molecular addresses they carry on their surfaces (Celi *et al.*, 2004; Zhou *et al.*, 2008; Xiao *et al.*, 2009). The nature, directionality, and efficiency of this molecular exchange depends on several factors. For instance, the physical properties of vesicular plasma membranes affect the fusion rate between exosomes and target cells, which may increase their exosome uptake under acidic pH (Parolini *et al.*, 2009). In some instances, exosome transfer could also be directed by specific molecular addresses, for example, a high concentration of phosphatydyl serines (PS) on the surface of certain exosomes (e.g. ectosomes or procoagulant microparticles) may enable their recognition by PS receptors (PSRs) on the surface of specific types of target cells. Many of such PSRs have been described within the context of phagocytosis of apoptotic cells by mononuclear cells; examples of such PSRs include Tim1, Tim4, stabilin 2 and BAI1 (Park *et al.*, 2007; Zhao *et al.*, 2010), at least some of which could be expressed more widely and may be involved in the uptake of exosomes. Indeed, blocking PSRs often obliterates exosome incorporation by endothelial cells, platelets and cancer cells (Del Conde *et al.*, 2005; Al-Nedawi *et al.*, 2008; Al-Nedawi *et al.*, 2009). A corollary to this point would be that phagocytes could be particularly susceptible to molecular influences of PS-positive exosomes, beyond their simple destruction. It has also been proposed that Tim1/4 receptors on two adjacent cells could allow formation of exosome bridges, thereby promoting additional indirect intercellular interactions (Xiao *et al.*, 2009). Similarly, the presence of PSGL-1 (P-selectin ligand) on the surface of procoagulant exosomes directs them to P-

selectin-expressing platelets and endothelial cells (Thomas *et al.*, 2009).

During malignant transformation, the action of mutant oncogenes, such as K-ras, EGFR, or its constitutively active mutant EGFR (variant III) (EGFRvIII), as well as several others, appear to stimulate the formation and release of exosomes (Yu *et al.*, 2005; Al-Nedawi *et al.*, 2008). Similarly, the activation or loss of specific tumor suppressor proteins appears to impact cellular vesiculation either positively or negatively (Yu *et al.*, 2005; Yu *et al.*, 2006). While the exact nature of the signalling pathways involved in oncogene-driven exosome biogenesis remains largely unknown, recent studies have begun to shed more light on the underlying processes. For instance, in cultures of prostate cancer cells, elevated exosome

**5.2 Oncosomes: Oncogene-driven vesiculation** 

*et al.*, 2009).

**5.1 Mediators of intercellular communication** 

production was detected in association with increased oncogenic activity of protein kinase B (PKB/Akt), or upon stimulation with growth factors (EGF), and depending on the status of the actin regulating protein known as diaphanous related formin 3 (DRF3) (Di Vizio *et al.*, 2009). In this case, inhibition of DRF3 expression through RNA interference enhanced the rate of exosome formation, and membrane blebbing activity, suggesting that DRF3 may be an inhibitor of ectosome release (Di Vizio *et al.*, 2009). Interestingly, DRF3 expression is lost during the progression of prostate cancer to metastatic disease, which suggests an intriguing link between oncogenesis, vesiculation and metastasis (Di Vizio *et al.*, 2009).

Fig. 5. Tumor microenvironment. The heterotypic interactions within the tumor microenvironment and their give and take of exosomes and their contents provide many targets for possible therapy. The goal of targetting these interactions will interrupt the heterotypic signaling that would thus deprive the cancer cells of the support they have within the tumor microenvironment.

#### **5.3 Effect of TEX on tumor microenvironment**

Exosome release by colorectal cancer cells is a function of K-ras and p53 status (Yu *et al.*, 2005). It is noteworthy that oncoproteins not only stimulate exosome formation but also become incorporated into their cargo (Al-Nedawi *et al.*, 2008; Al-Nedawi *et al.*, 2009). As a result, oncogene-containing exosomes (sometimes refered to as *oncosomes*) may serve as vehicles that carry oncogenic cargo and mediate its transfer between cells (Al-Nedawi *et al.*, 2009). At least four different modes of such oncogenic transfer have been described: (a) intercellular passage of active oncoproteins (Al-Nedawi *et al.*, 2008), (b) transfer of oncogenic mRNA transcripts

The Application of Membrane Vesicles for Cancer Therapy 35

antigens to the immune system and, when derived from certain cell types, are capable of presenting antigens to immune cells directly. Antigen presentation by antigen presenting cells (APCs) requires several important steps to elicit an immune response: 1. the APC internalizes and processes the antigen, 2. the processed antigen is inserted into Major Histocompatibility Complex (MHC) molecules and displayed on the cell surface, 3. MHC molecules interact with T cell receptors to start a signaling cascade to activate the T cell (Kim *et al.*, 2004). Dendritic cells are the primary APC type that stimulates T cells *in vivo*, which requires MHC molecules on the dendritic cell, as well as expression of co-stimulatory molecules that enhance T cell activation, such as CD40, CD80 and CD86 (Kim *et al.*, 2004). Like many other cell types, dendritic cells release membrane vesicles, particularly exosomes, which help modulate the immune response (Chaput *et al.*, 2006; Schorey & Bhatnagar, 2008;

Antigen presenting cells, such as dendritic cells and B cells, release exosomes equipped with MHC class I and class II molecules that allow direct presentation of antigens to cytotoxic and helper T cells, respectively (Kim *et al.*, 2004; Chaput *et al.*, 2006). Antigens processed by APCs are loaded into MHC molecules and the MHC-antigen complex is released into the extracellular space within exosomes. These exosomes then travel throughout the body and induce an immune response by stimulating antigen-specific T cells. In addition to MHC molecules, APC exosomes express surface co-stimulatory molecules CD40, CD80 and CD86 to enhance T cell activation (Raposo *et al.*, 1996). The ability of exosomes to induce a T cell response is dependent on the expression of these molecules, as introduction of antibodies against CD40, CD80 and CD86 inhibit antigen presentation and T cell activation by APC exosomes. Additionally, the tetraspanin molecule CD54 (ICAM1) plays a crucial role in this process by enhancing exosome-T cell contact through its interaction with CD11a (LFA-1) on the T cell (Kim *et al.*, 2004). This interaction allows MHC and co-stimulatory molecules to bind to their receptors on the T cells long enough to provide sufficient signaling. As observed with inhibition of the co-stimulatory molecules, antibody blockade of CD54 on

exosomes dramatically reduces T cell activation by APC exosomes (Kim *et al.*, 2004).

In addition to their ability to present processed antigens, exosomes can also transfer antigens from one cell to another. Exosomes containing cellular antigens are taken up by APCs, where the antigens are processed and inserted into MHC molecules for presentation (Obregon *et al.*, 2006). This is a very important mechanism in disease states, where infected or malignant cells release exosomes into the bloodstream that can be processed and presented by APCs throughout the body to induce an immune response (Bhatnagar & Schorey, 2007). In cancer, TEXs are taken up by APCs and activate tumor-specific T cells (Andre *et al.*, 2002a). *In vitro* studies have shown that when T cells are incubated with TEX in the presence of naïve dendritic cells, both helper and cytotoxic T cells are activated in an antigen-specific manner (Thery *et al.*, 2009). Additionally, *in vivo* vaccination studies comparing TEX and irradiated tumor cells showed a stronger immune response in animals vaccinated with TEXs. These studies have led to two clinical trials, which will be discussed

Thery *et al.*, 2009).

**6.1 Antigen presentation by immune cell exosomes** 

**6.2 Transfer of antigens by exosomes** 

in detail later.

(Skog *et al.*, 2008), (c) exchange of oncogenic miR and/or (d) passage of genomic sequences containing oncogenic DNA (Bergsmedh *et al.*, 2001). In many instances, this horizontal transfer may have marked biological (transforming) consequences. Thus, oncosomes containing EGFRvIII may emanate from malignant tumor cells and be taken up by their indolent counterparts inducing their growth, survival, and clonogenic and angiogenic capacity (Al-Nedawi *et al.*, 2008). These exosomes may also act on endothelial cells and reprogram their responses such that they exhibit an increase in angiogenic activity (Skog *et al.*, 2008) or switch to an autocrine mode of secretory pathway, e.g. by turning on VEGF production (Al-Nedawi *et al.*, 2008). Indeed, blocking exosome uptake using the Annexin V analogue (Diannexin) is associated with a measurable anti-angiogenic effect *in vivo* (Al-Nedawi *et al.*, 2008). In chronic lymphoblastic leukaemia (CLL), exosomes containing AXL kinase conditioned the bone marrow stroma to support disease progression (Ghosh *et al.*, 2010). These and similar effects identify exosomes as possible effectors of oncogenic and proangiogenic *field effects*, long postulated to exist in cancer (Slaughter *et al.*, 1953; Al-Nedawi *et al.*, 2009) and viewed as a mechanism of cell recruitment to the malignant process.

#### **5.4 Intercellular exchange**

Interactions between exosomes and their target cells may depend on the specific ligandreceptor recognition events. For instance, platelets take up procoagulant exosomes in a manner that depends on the expression of P-selectin and its ligand (PSGL-1) on the respective surfaces (Falati *et al.*, 2003). After the uptake, exosome-associated material was shown to penetrate into the cytoplasm of the acceptor cell (Skog *et al.*, 2008) or to remain on the cell surface, potentially in the immediately active form (Del Conde *et al.*, 2005; Al-Nedawi *et al.*, 2008). Interestingly, cloaking PS by exposure of exosomes to Annexin V often obliterates their uptake by target cells. Recent studies suggested that exosomal transfer would encompass multiple effectors at once (Skog *et al.*, 2008). Such exosomal exchange could contribute to tumor angiogenesis by mechanisms dependent on transfer of oncogenes (Rak, 2010) (e.g., EGFRvIII4 or mRNA), but possibly also through bidirectional trafficking of other molecules (e.g., VEGFRs, Tie, Notch), including ligands traditionally assumed to act almost exclusively in a cell-associated manner (Dll4, ephrins) (Kerbel, 2008). This prototype has been already validated in the case of exosomal emission of coagulation factors (e.g., tissue factor—TF21), chemokine receptors (Mack *et al.*, 2000) adhesion molecules (Falati *et al.*, 2003), immunomodulators (Valenti *et al.*, 2007), cell surface antigens (Dolo *et al.*, 2005), intact RNA species (Ratajczak *et al.*, 2006; Valadi *et al.*, 2007), and oncogenic proteins (Koga *et al.*, 2005; Al-Nedawi *et al.*, 2008; Sanderson *et al.*, 2008).

#### **5.5 Summary**

So far in the literature, exosomal cargo associated proteins and their importance in angiogenesis, inflammation, proteolysis regulators, membrane receptors, soluble factors, oncoproteins and tumor suppressors, lipids, nucleic acids are described briefly. Therefore, exosomes represent an integral part of both physiological regulation and disease pathogenesis, which may influence new therapeutic and diagnostic modalities.

#### **6. Membrane vesicles and antigen presentation**

As described earlier in this chapter, various cell types release exosomes, which contain a proteomic sampling from the cell of origin. Proteins within exosomes can be presented as

(Skog *et al.*, 2008), (c) exchange of oncogenic miR and/or (d) passage of genomic sequences containing oncogenic DNA (Bergsmedh *et al.*, 2001). In many instances, this horizontal transfer may have marked biological (transforming) consequences. Thus, oncosomes containing EGFRvIII may emanate from malignant tumor cells and be taken up by their indolent counterparts inducing their growth, survival, and clonogenic and angiogenic capacity (Al-Nedawi *et al.*, 2008). These exosomes may also act on endothelial cells and reprogram their responses such that they exhibit an increase in angiogenic activity (Skog *et al.*, 2008) or switch to an autocrine mode of secretory pathway, e.g. by turning on VEGF production (Al-Nedawi *et al.*, 2008). Indeed, blocking exosome uptake using the Annexin V analogue (Diannexin) is associated with a measurable anti-angiogenic effect *in vivo* (Al-Nedawi *et al.*, 2008). In chronic lymphoblastic leukaemia (CLL), exosomes containing AXL kinase conditioned the bone marrow stroma to support disease progression (Ghosh *et al.*, 2010). These and similar effects identify exosomes as possible effectors of oncogenic and proangiogenic *field effects*, long postulated to exist in cancer (Slaughter *et al.*, 1953; Al-Nedawi *et al.*, 2009) and viewed as a

Interactions between exosomes and their target cells may depend on the specific ligandreceptor recognition events. For instance, platelets take up procoagulant exosomes in a manner that depends on the expression of P-selectin and its ligand (PSGL-1) on the respective surfaces (Falati *et al.*, 2003). After the uptake, exosome-associated material was shown to penetrate into the cytoplasm of the acceptor cell (Skog *et al.*, 2008) or to remain on the cell surface, potentially in the immediately active form (Del Conde *et al.*, 2005; Al-Nedawi *et al.*, 2008). Interestingly, cloaking PS by exposure of exosomes to Annexin V often obliterates their uptake by target cells. Recent studies suggested that exosomal transfer would encompass multiple effectors at once (Skog *et al.*, 2008). Such exosomal exchange could contribute to tumor angiogenesis by mechanisms dependent on transfer of oncogenes (Rak, 2010) (e.g., EGFRvIII4 or mRNA), but possibly also through bidirectional trafficking of other molecules (e.g., VEGFRs, Tie, Notch), including ligands traditionally assumed to act almost exclusively in a cell-associated manner (Dll4, ephrins) (Kerbel, 2008). This prototype has been already validated in the case of exosomal emission of coagulation factors (e.g., tissue factor—TF21), chemokine receptors (Mack *et al.*, 2000) adhesion molecules (Falati *et al.*, 2003), immunomodulators (Valenti *et al.*, 2007), cell surface antigens (Dolo *et al.*, 2005), intact RNA species (Ratajczak *et al.*, 2006; Valadi *et al.*, 2007), and oncogenic proteins (Koga

So far in the literature, exosomal cargo associated proteins and their importance in angiogenesis, inflammation, proteolysis regulators, membrane receptors, soluble factors, oncoproteins and tumor suppressors, lipids, nucleic acids are described briefly. Therefore, exosomes represent an integral part of both physiological regulation and disease

As described earlier in this chapter, various cell types release exosomes, which contain a proteomic sampling from the cell of origin. Proteins within exosomes can be presented as

pathogenesis, which may influence new therapeutic and diagnostic modalities.

mechanism of cell recruitment to the malignant process.

*et al.*, 2005; Al-Nedawi *et al.*, 2008; Sanderson *et al.*, 2008).

**6. Membrane vesicles and antigen presentation** 

**5.4 Intercellular exchange** 

**5.5 Summary** 

antigens to the immune system and, when derived from certain cell types, are capable of presenting antigens to immune cells directly. Antigen presentation by antigen presenting cells (APCs) requires several important steps to elicit an immune response: 1. the APC internalizes and processes the antigen, 2. the processed antigen is inserted into Major Histocompatibility Complex (MHC) molecules and displayed on the cell surface, 3. MHC molecules interact with T cell receptors to start a signaling cascade to activate the T cell (Kim *et al.*, 2004). Dendritic cells are the primary APC type that stimulates T cells *in vivo*, which requires MHC molecules on the dendritic cell, as well as expression of co-stimulatory molecules that enhance T cell activation, such as CD40, CD80 and CD86 (Kim *et al.*, 2004). Like many other cell types, dendritic cells release membrane vesicles, particularly exosomes, which help modulate the immune response (Chaput *et al.*, 2006; Schorey & Bhatnagar, 2008; Thery *et al.*, 2009).

#### **6.1 Antigen presentation by immune cell exosomes**

Antigen presenting cells, such as dendritic cells and B cells, release exosomes equipped with MHC class I and class II molecules that allow direct presentation of antigens to cytotoxic and helper T cells, respectively (Kim *et al.*, 2004; Chaput *et al.*, 2006). Antigens processed by APCs are loaded into MHC molecules and the MHC-antigen complex is released into the extracellular space within exosomes. These exosomes then travel throughout the body and induce an immune response by stimulating antigen-specific T cells. In addition to MHC molecules, APC exosomes express surface co-stimulatory molecules CD40, CD80 and CD86 to enhance T cell activation (Raposo *et al.*, 1996). The ability of exosomes to induce a T cell response is dependent on the expression of these molecules, as introduction of antibodies against CD40, CD80 and CD86 inhibit antigen presentation and T cell activation by APC exosomes. Additionally, the tetraspanin molecule CD54 (ICAM1) plays a crucial role in this process by enhancing exosome-T cell contact through its interaction with CD11a (LFA-1) on the T cell (Kim *et al.*, 2004). This interaction allows MHC and co-stimulatory molecules to bind to their receptors on the T cells long enough to provide sufficient signaling. As observed with inhibition of the co-stimulatory molecules, antibody blockade of CD54 on exosomes dramatically reduces T cell activation by APC exosomes (Kim *et al.*, 2004).

#### **6.2 Transfer of antigens by exosomes**

In addition to their ability to present processed antigens, exosomes can also transfer antigens from one cell to another. Exosomes containing cellular antigens are taken up by APCs, where the antigens are processed and inserted into MHC molecules for presentation (Obregon *et al.*, 2006). This is a very important mechanism in disease states, where infected or malignant cells release exosomes into the bloodstream that can be processed and presented by APCs throughout the body to induce an immune response (Bhatnagar & Schorey, 2007). In cancer, TEXs are taken up by APCs and activate tumor-specific T cells (Andre *et al.*, 2002a). *In vitro* studies have shown that when T cells are incubated with TEX in the presence of naïve dendritic cells, both helper and cytotoxic T cells are activated in an antigen-specific manner (Thery *et al.*, 2009). Additionally, *in vivo* vaccination studies comparing TEX and irradiated tumor cells showed a stronger immune response in animals vaccinated with TEXs. These studies have led to two clinical trials, which will be discussed in detail later.

The Application of Membrane Vesicles for Cancer Therapy 37

cell exosomes or non-heat shocked tumor exosomes (Chen *et al.*, 2006c). Additionally, the dendritic cells treated with heat shocked TEXs exhibited increased production of proinflammatory cytokines tumor necrosis factor alpha (TNFα), IL-1β, and IL-12 (Chen *et al.*, 2006c). This is thought to be mediated by the increased amount of Hsps in the heat-shocked

Fig. 6. Dendritic cell-derived exosomes interact with T cells through MHC class I, II,

**7.2.2 Tumor cell exosomes inhibit natural killer cells via NKG2D** 

immune detection (Thery *et al.*, 2009).

tetraspanins such as CD9, CD63, CD81 and through co-stimulatory molecules such as CD86.

NK cells kill cancer cells by release of granules and perforins (**Figure 7**). NK cell activity is often lost in cancer patients, resulting in a reduced ability of the immune system to eliminate malignant cells (Clayton *et al.*, 2008; Viaud *et al.*, 2009; Ahiru *et al.*, 2010). The NK cell receptor NKG2D is important in regulation of NK cell function, with some ligands stimulating and others inhibiting cytotoxic function (Clayton *et al.*, 2008). As discussed earlier, DEXs expressing activating NKG2D ligands can enhance NK cell function and promote tumor clearance (Viaud *et al.*, 2009). Other NKG2D ligands, such as MHC class Irelated chain A (MICA) can reduce NK cell function. Tumor cells abuse this normal ligand by upregulating MICA on the cell surface as well as on TEXs (Ahiru *et al.*, 2010). Ovarian cancer exosomes expressing high levels of MICA were shown to decrease NK cell function *in vitro* by reducing their NKG2D receptor expression and their responsiveness to activating NKG2D ligands (Ahiru *et al.*, 2010). This reduction in NK cell function is highly detrimental to the anti-tumor response and is one of many mechanisms by which tumors escape

TEXs, though the exact mechanism is still under investigation.

#### **6.3 Summary**

Antigen presentation to T cells is the first and most critical step in the adaptive immune response. The ability of exosomes to supply antigens to dendritic cells for presentation as well as present antigens directly via exosomal MHC molecules is an important mechanism for detection of infection and malignancy. Induction of anti-cancer responses via tumorreleased exosome antigens is likely a key mechanism in immune surveillance to prevent tumor progression. Whether this can be utilized in tumor immunotherapies is under investigation and could prove useful in combination therapies.
