Please notice that this presentation aims at summarizing currently available data only without the claim of representing established knowledge. Thus information presented in this course may be subject to change in the future due to the rapid evolvement of the field of extracellular vesicles. So the title of this course is a EVs in tumor and immune system interaction and it's been put together by these legends of the field. So Mercedes Tkach and Clotilde Thery from the Institut Curie in Paris and France and Suzanne Gabrielsson from the Karolinska Institute in Stockholm, Sweden. I'm Dave Carter from Oxford Brookes University, and I'm just here to narrate the amazing presentation that these fabulous people have put together. This is the outline for today's presentation. So there's a bit of a background of forward and some general introduction. And then we'll talk about the effects of tumor EVs on the immune system, focusing on various examples of various types of immune cells. We'll talk about attempts to modulate EV secretion in vivo and the consequences on the immune system. We'll talk about the effect of the immune EVs on the tumors. So that's the reciprocal interaction, the immune cells interacting with the tumor cells. And then we'll talk about the effects of circulating EVs on anti-tumor treatments. We know that cells can release many different types of EVs from different intracellular origins, both when alive or when dying. So dying cells, for example, release apoptotic cell derived EVs. So the term exosomes is generally used for small EVs formed inside multi-vesicular late endosomes, also known as multi-vesicular bodies or MVBs and secreted upon fusion of these intracellular compartments with the plasma membrane. Other types of EVs, like microvesicles also known as exosomes, which can be of similar size as exosomes or sometimes larger, bud directly from the plasma membrane. Since when EVs are recovered from outside the cells, it's difficult to determine whether they came from in MVBs or from the plasma membrane. So in this presentation, we will use the generic term EVs, extracellular vesicles, throughout the presentation or sometimes small EVs, or SEVs, even if the authors of some of the papers that were describing chose the term exosome in their studies. In the published literature, authors describe functions of EVs either selected for their small size, for example obtained by filtration through 200 nanometer filters or steps of low speed centrifugation followed by high speed centrifugation or generic precipitation, and these are often called exosomes. Or isolated by lower speed centrifugation without the filtration step, and these are often called microvesicles or large orgosomes or microparticles or mixtures of these. However, there have been very few qualitative or quantitative comparisons of the effects of these different types of EV. Therefore, it's not really possible to determine yet if the described functions are specific of the used EV type or if they're also present in other EVs. Functions of tumor or immune cell EVs are often analyzed using EVs isolated and concentrated in vitro from in vitro cultured cells, often cell lines or more rarely from circulating EVs from tumor bearing mice or humans. The risk with using EVs from cell lines are, one, the cell lines have been selected for what grows in vitro may not represent the tumor in vivo. And two, EVs isolated from cell cultures are often, if grown in FPS or SCS, contaminated by serum RNAs or proteins. The use of EVs from clinical samples when possible is important to confirm and validate results obtained with cell derived EVs. The majority of the immune inhibiting effects are described when tumor EVs are used in vitro or in vivo. Although tumor derived EVs can also be a source of tumor antigens, which antigen presenting cells used to generate anti-tumor immune responses. However, we don't know in most cases if this reflects their function when they are secreted in vivo, as it's difficult to estimate the concentration who of the EVs locally within the tissue or even in the circulation. The few attempts to prevent EV secretion directly in tumors growing in vivo and the consequent effects on the immune system will be described later on at the end of this presentation. The next forward before we get into the presentation proper is the need for caution on EV isolation from plasma or serum of patients. So there's a few things to consider here. That one is that most EVs in plasma, also in cancer patients, come from cells other than tumor cells. And the other thing to consider is that the immune system is affected in cancer patients. And so therefore immune system EVs are also modified. The third thing to consider as well as that the coagulation system is affected in cancer patients and that platelet-derived EVs are increased in cancer patients. So when you're doing these kind of studies what you have to ask is do you really want to be studying or do you want to study immune cell derived EVs, platelet-derived EVs, or cancer cells derived EVs? So these are considerations when doing these kinds of studies. Tumor derived EVs have been proposed to favor the growth and spread of tumor cells in several ways, both by affecting non immune cells and immune cells. And the possibility of EVs to act both locally and distantly makes them capable of acting both within the tumor micro environment, but also more distantly to spread metastasis, to cause the tumor to metastasize. So the the EVs released by tumor cells can be involved in promoting cell motility and invasiveness by several different means, but also in establishing the pre-metastatic niche, again through different potential means. There are reciprocal interactions between tumor cells and the immune system. The immune system can attack tumors by natural killer cells, CD8 T-cells antibodies from B cells, which can induce ADCCs antibody dependent cell-mediated cytotoxicity, and by NK cells, macrophages, neutrophils, etc. So there are many ways in which these tumor cells can be attacked by the immune system. But tumor cells have developed means to evade this attack. So down-regulation of surface receptors, including the MHC class 1, and expression of immune checkpoint receptors. Tumors also developed means to skew immune cells towards promoting tumor growth or metastasis, so pro-angiogenic neutrophils, and tumor macrophages, T regulatory cells, etc. EVs have been proposed to play a role in many both pro and antitumoral effects. So here are some examples of how they can have a pro-tumor roll. So for example, they can immunosuppress dendritic cells through the inhibition of their differentiation capacity. On natural killer cells, they can decrease their cytotoxic capacity via NKG2DL and TGF-Beta. They can also inhibit T cell activation and proliferation and killing capacity. Well, they can even promote differentiation and expansion of T regulatory cells. They can also promote myeloid-derived suppressor cells and also induce alternative activation of macrophages Promoting the M2 phenotype over the M1, which promotes tumor growth and the immune escape. But they can also have anti-tumor effects. So so here are some examples of anti-tumor effects of these EVs. So for example, tumor exosomes are sort of antigen for induction of anti-tumor immune responses by dendritic cells. And they can also promote activation of innate immune cells, like NKs and macrophages, especially when they're released by stressed tumor cells. Tumors EVs can inhibit the functions of NK cells in several ways. NKG2D is a killer activating receptor of NK cells and the production of NKG2D ligand bearing exosomes, and probably EVs in general, has been proposed as a mechanism for tumor cell escape from immune recognition. Indeed, it's been demonstrated that in contrast to the NKG2D ligand ULPB2, released ULBP3 is included into EVs. Remarkably, the ULBP3 containing EVs have been shown to be more potent downregulators of the NKG2D receptor then the soluble form of ULBP2 proteins released by their metalloproteinase is Adam10 and 17. Pre-incubation of NK cells with ULBP3 containing EVs induces a dramatic reduction of NKG2D mediated lysis of MICA expressing cells. The transfer of TD bearing membrane anchored TGF-beta MICA and MICAB leads the down-regulation of NKG2D expression at the surface of NK cells and impairs their cytotoxic function. T-cells can aid in tumor killing both directly via cytotoxic cells and indirectly via T-helper cells which promote for example antibody production. But T-cells are affected by tumor EVs in several ways as shown in this slide here. So for example, EVs can express Galactin-9 and LMP1, which can then inhibit T-cell proliferation. Microvesicles from melanoma cells have been shown to express Fas ligand, which could kill T-cells and EVs from several cancer types express CD39 and and CD73, which suppress T-cells via adenosine production. The smart way for the cancer cell to inhibit the immune system is to induce T regulatory cells. So T-reg cells can act both antigen specifically and non-specifically and by inducing T-reg cells the anti-tumor immune response can be dampened in several ways. Wickowski, for example, showed that T-cells incubated with medium-sized EVs increased the proliferation of CD4 positive CD25 high, FoxP3 positive cells. These cells can also be affected by EVs, as shown in this slide. In a work done by Pucci et al, tumor EVs have been tracked in vivo in draining lymph nodes by injecting tumor cells expressing fluorescently labeled EVs or EVs containing luciferase. These vesicles in the lymph nodes are captured by a specialized macrophage population subcapsular sinus macrophages that are CD169 positive, thus blocking the spread of the tumor EVs. During tumor progression, however, the disruption of this macrophage barrier leads to the direct interaction of tumor EVs with B-cells, fostering tumor promoting humoral responses, although the exact mechanism of these pro-tumor responses still needs to be elucidated. Myeloid cells as well can also be affected by tumor EVs. So monocytes have been described to differentiate into macrophages endowed with more or less tumor promoting abilities. In a simplified model, M1 macrophages are prone to produce IL-12 and can help with stimulating for example, anti-tumor cytotoxic T-cells. While M2 macrophages produce IL-10, which will for example stimulate the generation of T-reg cells. Tumor EVs have been shown to induce the M2 macrophage population by various molecular mechanisms. After the M2 macrophages of developed these cells will produce EVs and other molecules, which can then further promote metastasis. Alternatively, small EVs from pancreatic cancer cells themselves migrate to distant organs and promote the formation of a pre-metastatic niche by creating a fibrotic environment enriched in TGF-beta, fibronectin, and a macrophage attracting chemokine. Neutrophils myeloid derived suppressor cells are affected by tumor EVs. Many of the studies on tumor derived extracellular vesicles are based on injecting in vitro derived EVs into mice. An elegant study has demonstrated how EVs produced in vivo a distributed. Ridder et al produced glioma and carcinoma cells which release EVs containing Cre-mRNA. This made them able to detect cells that took up the in vivo produced tumor derived EVs. The study showed that the majority of recombined cells were CD45 positive leukocytes predominantly GR1 positive CD11B positive myeloid derived suppressor cells. These cells have been shown to inhibit functions of various cell types including dendritic cells, T-cells, macrophages, and NK cells. Additionally, they could show that the cells that took up EVs acquired immunosuppressive properties, including upregulation of PDL1, Arginase-1, and TGF-beta. It's been shown by Bobrie et al that secretion by some tumor cells of small EVs that in conjunction with soluble factors secreted by the cells induced differentiation of precursors into neutrophils and induce their recruitment to the tumor, thus promoting local tumor growth. EVs have been shown to participate in the dissemination of cancer cells. Tumorous EVs can alter the cellular physiology of both surrounding and distant non tumor cells to allow dissemination and growth of cancer cells by conditioning pre-metastatic sites in distant organs. In particular melanoma tumorous EVs bearing the tyrosine kinase receptor MET can promote migration of bone marrow progenitor cells to future sites of metastasis. When it comes to the effect of tumor EVs on dendritic cells, there are contradictory published results. It's been shown that the treatment of dendritic cells with tumor derived EVs impaired their LPS induced maturation. On the contrary, tumor derived EVs have also been trying to activate DCs and to induce the release of IL-6 mediated by HSP70 to and HSP105 on the EV surface. Moreover, Il-6 accretion is toll-like receptor 2 and 4 dependent. This increased IL-6 dramatically promotes tumor invasion by increasing stat3 dependent matrix metalloproteinase-9 To demonstrate functions of tumor EVs in vivo, several studies have attempted to modulate EV secretions in vivo. The intracellular small GTPase, RAB27 promotes secretion of EVs including exosomes, but also soluble molecules, which cooperate to attract immune cells to promote tumor progression. RAB27A knockdown in several tumors decreases secretion of EVs, especially exosomes. When RAB27A knockdown 41 tumors are grafted in vivo in syngeneic mice, local growth of this tumor is decreased which is due to impaired accumulation of neutrophils in the tumor. This tumor secretes cytokines favoring neutrophil to differentiation, such as G-CSF which cooperate with EVs secreted in a RAB27A dependent manner to induce neutrophil invasion in tumors. This effect was not observed in another mammary carcinoma, TS/A, suggesting that tumor cell intrinsic differences may lead to different EV types or functions. However, this study has also shown that a few soluble molecules, ie non-EV associated factors such as MMP9 were also secreted in a RAB27A dependent manner. Thus highlighting the need to perform controls such as reconstitution with exogenous EVs or exosomes to conclude on an exosome function after RAB27A knockdown. In another study, knockdown of RAB27A in a melanoma model reduced EV production and circulating UV concentration. Thus preventing the recruitment of the bone marrow derived cells that are necessary for metastatic progression. In this study, as described in that last slide, Rab27a knockdown led to reduced secretion of a few selected soluble factors. Another intracellular small GTPase, RAB35, was also shown to be required for EV secretion. As described a few slides ago, melanoma EVs are captured by the subcapsular sinus macrophages, which thus control tumor growth. Removal of the CD169+ macrophages accelerated the growth of a wild-type tumor through the induction of B cell dependent humoral immune responses. However, tumors knocked down for RAB35, in other words having impaired EV release capacity, grew similarly with or without CD169+ macrophages. This demonstrates that a normal EV production is needed to support the link between EVs and the subcapsular sinus macrophages and thus tumor growth. However, all of these observations in vivo should be considered with caution, since the inhibition of all these trafficking pathways have plenty of non EV related consequences. And finally since 2008, numerous studies have used inhibitors of sphingomyelinases, such as the drug GW4869 or spiroepoxide to inhibit exosome secretion in vitro or in vivo. And conclude on the role of exosome secretion in various tumor related processes involving cell to cell communication. However, such conclusions should be taken with a big grain of salt, because sphingomyelinases are involved in so many different intracellular physiological processes, that targeting them is unlikely to specifically inhibit secretion of exosomes or even EVs in general. And I'll summarize a few examples here. So for example, it's been shown that affecting sphingomyelinases can affect the production of different proteins and protein lipid protein secretion in EVs. And suggesting that sphingomyelinases are involved in the secretion of multiple types of EV not just exosomes. Also it's been shown recently that the drug GW4869 can decrease the secretion of small EVs, but there's a concomitant increase in the secretion of larger plasma membrane derived EVs and associated changes in protein composition. So again suggesting that it's not just about exosome release. Also, it's been shown that sphingomyelinases can cycle between golgi and plasma membrane but also other systems including early/recycling endosomes, suggesting that it's involved in vesicular trafficking in multiple locations. And it has also been shown that sphingomyelinases are involved in numerous physiological and pathological pathways, including the response to TNF alpha and interferon gamma and so on. So it's not necessarily just related to EVs and or exosome secretion. One approach used to demonstrate a function of EV secretion in vivo has been to force secretion of an antigen on or in EV secreted in vivo. The cDNA coding for the antigen was inserted in frame with a phosphatidyl serine binding C1C2 domain of a protein abundantly found on EVs, the milk fat globule, MGF factor 8, also called lactadherin. And cells expressing this modified antigen secreted EVs bearing this antigen on the EVs. Alternatively, the cDNA was fused with the retroviral gag protein, which is known to induce budding of the virus and cells expressing this modified antigen secreted virus-like particles, a type of EV containing the antigen in the EVs. Modified antigen was expressed in tumor cells, previously translated in vitro, which were then grafted in immunocompetent mice in vivo, were directly transferred in vivo, in muscle or skin cells upon direct injection of plasma DNA in these tissues, a process called DNA vaccination. In both cases, in mice expressing the EV associated antigen, immune responses against the antigen were more potent than against a secreted non EV associated antigen. And these immune responses induced more efficient prevention or rejection of tumors. This effect was probably due to a more efficient capture by dendritic cells of the EV associated than a soluble antigen. In this third section, we'll talk about how immune cells involved in induction of anti-tumor immune responses also secrete exosomes and small EVs and what the therapeutic implications of this might be. Dendritic cells or antigen presenting cells which are potent stimulators of naive T Cell responses. These cells also release EVs capable of antigen presentation. Antigen presenting cells take up tumor cells and antigens and release EVs loaded with antigens and or antigen fragments and or MHC peptide complexes, some of which have formed in multi-vesicular bodies, where endogenous MHC2 and exogenous antigens meet and thus are released on bona fide exosomes. These compounds may potentiate the immune response, although the MHC peptide complexes are probably less abundant and functional than the antigens. This property of dendritic cell derived EVs led to To the idea to use dendritic cell EVs in immunotherapy. Whole dendritic cells have been tested in immunotherapy to increase or decrease antigen-specific immune responses. However, the injection of whole cells is complicated and has drawbacks. Dendritic cells may die during freezing, and therefore it's difficult to control the exact number of injected cells. Also, the properties of dendritic cells can change after injection in vivo. So for example, MHC peptide complexes, cytokines, etc, these can change. Dendritic cell exozomes and small EVs are nanovesicles, which are stable from freezing and are stable after injection in vivo, so therefore, more stable than cells. The first evidence that dendritic cell-derived EVs could induce immune responses capable of eradicating cancer cells was published in 1998. Only syngenic exosomes could eradicate the p815 tumors, and not allergenic ones. Supporting a role for MHC peptide on the exozomes, and of CD8-positive cytotoxic T-cells. This may vary depending on the tumor model, the p815 tumor is known to be sensitive to cytotoxic T-cells The first clinical trials were published in 2005. The idea was to use dendritic cells generated from the patient's own monocytes, and load these with tumor-specific peptides. Whereby the EVs would then carry the peptide MHC complexes together with costimulatory molecules. And that these EVs would then stimulate cancer-specific immune responses in the patient. The first trials were done with immature dendritic cells, and later ones used interferon-gamma matured dendritic cells. The first two clinical trials used EVs from immature dendritic cells. Patients underwent leukapheresis, the monocytes were recovered to generate dendritic cells, and tumor antigen peptides were then added to the cells. The exozomes were harvested, isolated, quality-controlled and then re-infused into the patients. The outcome of the study was that the feasibility and the safety of the procedure was demonstrated. In a later clinical trial published in 2016, EVs derived from interferon-gamma mature dendritic cells were used. 22 patients were included, only 1 presented with grade 3 hepatotoxicity, and thus, safety and feasibility were confirmed. Primary endpoint was 50% of patients with progression-free survival of over four months, this wasn't reached. No clear T-cell responses were seen, however, increase of natural killer cell activity in some patients was seen with initial decrease of NKp30 expression, and this correlated with longer progression-free survival. Later mechanistic, preclinical studies suggest improvements of the conditions of generating or using EVs for clinical trial. In a B16 OVA melanoma, the NK T-cell ligand alpha-Galactoside-Ceramide increased antigen-specific T-cell activation and tumor killing. This has been tested in a clinical trial in hepatitis B infection, with some side effects. But these could probably be reduced, as EV-bound AGC was more potent than soluble AGC, as we'll see in the next slide. And here are those data, so while soluble AGC was more potent than EVs to stimulate NK T-cells, as you can see from this graph on the left, it was only the small EV-bound OVA and AGC that could stimulate the OVA specific CD8 positive T-cells. Another word of caution, though, when it comes to immune responses, the effects of EV is can be very different between in vitro and in vivo response. In this case, it was seen in vitro that peptide-loaded exosomes or EVs were more efficient to stimulate T-cells, compared to exosomes or EVs loaded with OVA protein. In vivo, however, the results were completely the opposite way round. Now this could be due to the bioavailability of the exosomes or EVs. IE, EVs and T-cells do not meet in the spleen unless also B-cells were activated, and could shuttle the EVs to other parts of the spleen. Alternatively, antibodies may have a role that's not yet clear. In B-cell knockout mice, the stimulation was diminished, demonstrating the importance of B-cell activation for T-cell activation by EVS. EVs from antigen-presenting cells such as B-cells or dendritic cells contain several factors that may activate or amplify different parts of the immune response. And this could present an opportunity, they could be engineered to further amplify immune responses to destroy cancer cells. And finally, another consideration is the effects of circulating EVs on anti-tumor treatments. So the large number of both tumor derived and non-tumor derived EVs that circulate in the body might interfere with different cancer therapies. So for example, tumor EVs bearing tumor surface receptors as decoys of tumor-targeting antibodies. An example of this is HER2, so HER2 overexpressing EVs could counteract the effect of based therapies. A second example of this is tumor EVs bearing NK ligands as decoys of NK activity. So NKG2D ligands make A and B and ULBP1 and 2. And acting as a decoy, the ligand-bearing EVs downregulate the NKG2D receptor mediated cytotoxicity and impair NK cell function. So in conclusion, then, during cancer, altered EVs are released from cancer cells, immune cells, and other cells. EVs from cancer of different stages have different effects on the immune cells. The non-cancerous cells in the cancer microenvironment may also affect the immune system, nearby or at a distance. Caution really needs to be taken when applying in vitro findings in vivo. EVs as immuno therapy for cancer may be developed. And EVs as decoys should be considered during cancer therapy.
