Extracellular vesicles in immunomodulation and tumor progression

Abstract

Extracellular vesicles have emerged as prominent regulators of the immune response during tumor progression. EVs contain a diverse repertoire of molecular cargo that plays a critical role in immunomodulation. Here, we identify the role of EVs as mediators of communication between cancer and immune cells. This expanded role of EVs may shed light on the mechanisms behind tumor progression and provide translational diagnostic and prognostic tools for immunologists.

Cells communicate via contact-dependent and contact-independent interactions through the secretion of chemokines, cytokines, growth factors, and associated cell receptors. Over the past 30 years, the secretion of extracellular vesicles (EVs) has also emerged as a major mechanism for cell-cell and cell-environment interactions. These 30–5000 nm lipid membrane-bound vesicles contain functional proteins, nucleic acids, lipids, and other bioactive molecules that can modulate the behavior of recipient cells¹.

EVs are generally split into three subtypes based on their mechanism of biogenesis – exosomes, shed microvesicles, and apoptotic bodies (Box 1). EV biogenesis and secretion have been well reviewed elsewhere^2,3,4. In brief, exosomes are formed via the inward budding of the cell plasma membrane and the subsequent formation of multi-vesicular bodies. Conversely, shed microvesicles - also referred to as ectosomes - are produced via the outward budding of the plasma membrane. Apoptotic bodies are released during cellular apoptosis. Due to the difficulty of isolating EV subtypes, they are often labeled according to size as small, medium, and large EVs (Box 1). Following the International Society for Extracellular Vesicles (ISEV) nomenclature⁵, we will use the term EV for all lipid bilayer particles secreted by cells, specifying size where appropriate. While some proteins may be shared within an EV subtype, vesicular cargo is highly dependent on the type and state of the donor cell⁶. This diversity in cargo allows EVs to modulate several important cellular processes, including cell differentiation, blood coagulation, and angiogenesis in tissue homeostasis and development⁷. Studies have demonstrated the key role that EVs play in tumor progression and antitumor immune responses⁸. Here, we review the role of EVs as mediators of communication between cancer and immune cells, as well as the potential clinical uses of EVs that arise from these interactions.

Box 1. Subtypes of extracellular vesicles.

Classically defined by size, no specific molecular markers exist to distinguish EV subtypes. While a number of molecules tend to be enriched in EV fractions; many traditional markers (CD9, CD63, CD81, TSG101, Alix, Flotillin-1, HSC70, Actin, MHC I and MHC II) co-isolate with multiple EV subtypes leading to confusion and misinterpretation in the literature. Special consideration should be directed towards extensive characterization of EV isolations prior to publication.

The role of extracellular vesicles in the immune cascade

Immune-derived EVs carry a variety of cargo that function in immune activation and suppression (Fig. 1a). The role of EVs in an immune response was first described in 1996 (Ref.⁹). This study observed that small EVs released by B cells not only carried major histocompatibility complex (MHC) class II molecules, but also elicited immune responses in T cells⁹. Soon after, a second study¹⁰ described the ability of dendritic cell (DC) small EVs to induce tumor suppression in vivo. In 2002, another study elaborated on this work to describe a mechanism for indirect T cell stimulation by DC small EVs¹¹. These discoveries established EVs as a major mechanism of cellular communication and provided a basis for immune EV research over the past two decades.

Figure 1: Heterogeneous cargo of EVs.

Secreted from tumor and immune cells alike, EVs contain molecular cargo that facilitates immune stimulatory and suppressive interactions. Trafficking transmembrane and cytosolic proteins, RNAs, and DNAs, EVs regulate autoimmune and disease responses. A) EVs secreted by immune cells carry cargo that directly controls peripheral immune function by either promoting a regulatory response or conditioning surrounding cells, reducing inflammation. B) Tumor-derived EVs express surface and cytosolic markers from their cellular progenitor, which in turn induce phenotypic changes in recipient cells.

DCs are antigen presenting cells (APCs) that orchestrate the immune response against a specific antigen. DC EVs have become a focus of the field due to their ability to stimulate antitumor immune responses. DC EVs harbor MHC I and MHC II molecules, tetraspanins, adhesion molecules, heat shock proteins, and costimulatory molecules that confer their immunomodulatory capabilities^12–19. It is debated, however, whether EVs can carry out these functions via direct interaction with target cells^18,20,21, or if the presence of DCs is necessary^11,22–24. Interestingly, in vitro modeling of primary mouse DCs might set a new example for the immunomodulatory capabilities of EVs. Shown to transfer tumor antigen loaded vesicles via tight synaptic connections, DCs highlight a potential downstream mechanism for EV antigen presentation²⁵.. Further in vitro modeling involving priming of DCs via antigen laden EVs might reveal internalized EV cargo is also shared between APCs during this synaptic transfer.

Whether their effects are direct or indirect, APC-derived EVs, and specifically DC EVs, play a major role in antigen presentation and immune activation^26–28. EVs have the remarkable ability to transfer pre-formed functional peptide-MHC complexes from APCs to recipient cells^29,30. Amongst the MHC II molecules present on DC EVs, HLA-DQ enables DC EVs to promote T cell proliferation^23,31,32. DC small EVs induce maturation and differentiation in several cell types, including immature DCs and monocytes^33,34. DC EVs also stimulate interferon-gamma (IFN-γ) production by naive CD4⁺ T cells and induce differentiation into T helper 1, T helper 2, and regulatory T (T_reg) cells^23,35–38. Interestingly, the molecular cargo of DC EVs varies based on the presence of surrounding cells. Co-culture assays using primary DCs derived from human monocytes reveal that in the presence of bystander T cells, DCs secrete small EVs enriched in microRNAs (miRNAs) miR-30b, miR-146a, and miR-155³⁹. These miRNAs carried by DC EVs are functionally active in promoting further CD8⁺ T cell activation³⁹.

While T cells generally function to carry out immune responses, their EVs have also emerged as important immunomodulators. Small EVs from activated T cells are enriched in nucleic acids with a variety of functions. Found to contain mitochondrial DNA (mtDNA), small T cell EVs prime DCs to more efficiently respond to future attacks by a familiar antigen⁴⁰. Acting through the cGAS-STING cytosolic DNA-sensing pathway, DNA bound to the surface of small T cell-derived EVs confers an increased resistance to infection in recipient DCs⁴⁰. Activated T cells secrete EVs containing specific tRNA fragments, which would otherwise function to inhibit T cell activation⁴¹. Secretomic analysis of cell culture-derived T cell EVs revealed systemic encapsulation of immunomodulatory cytokines²¹. EV-associated cytokine encapsulation by T cells further stabilize the functional ability of these cytokines in comparison to their free counterparts²¹. This stabilizing effect of EVs has been exploited in drug delivery applications (described in detail below).

While the differences among EV subtypes are well accepted, reports suggest that heterogeneity within EV subtypes remains largely unaccounted for^23,35,42,43. The maturity of donor DCs as well as variation in EV enrichment profiles affect naive T cell development^23,35,42. Small EVs and EVs secreted by mature DCs preferentially induce T helper 1 activation, while larger EVs induce T helper 2 differentiation²³. Similarly, T cell-derived EVs bearing different markers – namely CD47, CD63, and MHC I – are enriched in different sets of RNAs⁴³. These developments have shed light on why EV research often produces sometimes conflicting results and provide more evidence for why it is typically inaccurate to attribute any effect to a single subpopulation of EVs.

Autoimmune suppression via extracellular vesicles

The primary function of T_regs, is to prevent autoimmunity by promoting self-tolerance of the immune system⁴⁴. T_regs employ EVs to suppress the activity of other immune cells⁴⁴. T_reg-derived EVs contain several miRNAs and miRNA precursors^45–50. Uptake of EVs containing miR-142–3p and miR-150–5p by DCs induces a decrease in pro-inflammatory cytokine IL-6 expression and an increase in immunosuppressive cytokine IL-10 expression⁴⁵. These miRNAs interfere with antigen processing and presentation in DCs⁵¹, thus inhibiting immune activation (Fig. 2).

Figure 2: Mechanisms of EV immune regulation.

EVs released by immune cells constantly balance immune regulatory and autoimmune responses of surrounding cells in a highly interwoven network of immune responses. EVs promote cell maturation, polarization, and differentiation in T cells and mononuclear cells. Encapsulation and stabilization of EV-associated cytokines further extend the functional capacity of immune EVs by regulating cytokine release and downstream proliferation.

T_regs also target other T cells as a method of immune regulation. T_reg-derived EVs suppress the proliferation of T cells via EV-mediated transfer of miRNAs and miRNA precursors^46–50. T_reg-derived EVs are enriched in miR-146a-5p, which targets and suppresses signal transducer and activator of transcription 1 (STAT1) and interleukin-1 receptor-associated kinase-like 2 (IRAK2) in recipient CD4⁺ T cells to inhibit proliferation⁴⁹. The miRNA precursor Let-7d is carried by T_reg-derived EVs and transferred to the target cell, where it appears to inhibit Cox-2 and decrease IFN-γ expression⁴⁸. A specialized subset of CD4⁺CD25⁻ T_regs release EVs found to induce a Treg phenotype in naïve CD4⁺ T cells (Fig. 2). These EVs containing miR-9, miR-330, miR-503, and inducible nitric oxide synthase (iNOS) mRNA suppress proliferation and induce an increase in IL-10 secretion in recipient T cells⁴⁷.

In addition to directly influencing immune cell functions, T_reg-derived EVs exert secondary immune suppression effects by transferring proteins to recipient cells. On their surface, T_reg-derived EVs display both the Ebi3 and p35 subunits of IL-35 in association with tetraspanin CD81⁵². Cells that uptake these EVs exogenously express IL-35 on their surface. IL-35 interacts with the cell’s own IL-35 receptors (IL-35R) to induce cell exhaustion via expression of the inhibitory receptors programmed cell death protein 1 (PD-1), T cell immunoglobulin and mucin domain-containing protein 3 (TIM3), and lymphocyte-activation gene 3 (LAG3)⁵³. Other lymphocytes may also interact with IL-35 on the surface of these recipient cells, causing secondary suppression of the neighboring cells. This effect has been observed in CD4⁺ T cells, CD8⁺ T cells, and B cells⁵² (Fig. 2).

Antigen-specific immunosuppression is exhibited by EVs derived from APCs^54,55. APC-derived EVs express MHC II molecules and Fas ligand (FasL), which allow them to induce apoptosis in T cells⁵⁶. Since this mechanism of immunosuppression is antigen-specific, the host must be immunized against the antigen for the effect to occur. Similarly, overactivated T cells undergoing activation-induced cell death secrete EVs expressing FasL, which suppresses the immune response⁵⁷.

The role of cancer-derived extracellular vesicles in immune evasion

A hallmark of cancer is the ability of tumor cells to evade detection by an individual’s immune system⁵⁸. Colloquially termed “immune escape”, successful immune evasion utilizes a variety of mechanisms to stifle both adaptive and innate immune responses (Fig. 3)⁸. Immune cells experience a wide variety of inhibitory interactions from cancer cells, both via direct physical contact and through endogenous soluble factors^58–61. As a significant driver of these interactions, cancer EVs can exhibit both immune evasion^62–65 and immunogenic properties (Fig. 2b)⁶⁶.

Figure 3: Cancer EVs directly regulate tumor progression.

A) EVs released by cancer cells promote tumor progression through the suppression of adaptive and innate immune cells. Impeding effective antigen cross presentation in dendritic cells, EVs also contribute to T cell dysfunction through check point inhibition. Additionally, cancer EVs polarize mononuclear cells towards an immunosuppressive phenotype, while reducing cytotoxicity in NK cells. Less restricted by physical constraints, cancer EVs demonstrate far reaching immunosuppressive effects on circulatory and distal immune cells. B) Cancer EVs function as vessels for tumor recognition. During a functional immune response, EVs deliver foreign antigens to dendritic cells indirectly stimulating cytotoxic T cells. While macrophages actively sequester cancer EVs from circulation promoting cancer cell recognition. Molecules highlighted in red stimulate immune activation; molecules highlighted in black describe immunosuppressive cargo.

Early work on metastatic melanoma revealed that upregulation of the protein tyrosine kinase MET (MET) in small EVs permanently re-educates circulating bone marrow progenitors into pro-tumorigenic mediators of pre-metastatic niche formation in vivo⁶⁷. Similarly, small EVs from pancreatic cancer cells, which contain a high amount of macrophage migration inhibitory factor (MIF), induce transforming growth factor beta (TGF-β) release by Kupffer cells to remodel the extracellular matrix (ECM) in the liver⁶⁸. More recent work has shown that cancer EVs carry an abundance of functional mRNA, non-coding RNA, and proteins thought to be crucial for interacting with both the innate and adaptive immune responses (Fig. 3)^69–71.

Highlighting the diverse role for cancer cell EVs, RNA sequencing of small chronic lymphocytic leukemia EVs revealed the non-coding Y RNA hY4 increased programmed death-ligand 1 (PD-L1) protein expression in circulating monocytes via toll-like receptor (TLR) 7 signaling⁷². PD-1/PD-L1 interactions regulate T cell receptor (TCR) activation to prevent autoimmune response⁷³. Like its tumor cell progenitor, PD-L1 on small EVs isolated from glioblastoma and metastatic melanomas demonstrate the ability to directly activate the PD-1/PD-L1 immune checkpoint^69,74 (Fig. 3). Small EVs isolated from metastatic melanoma feature identical PD-L1 membrane topology⁶⁹ and, as such, functionally bind PD-1^69,74. A syngeneic mouse model of melanoma further revealed that PD-L1⁺ EVs prevent the proliferation of PD-1⁺ CD8⁺ T cells, and consequently the number of tumor infiltrating lymphocytes⁶⁹. Expressed in a concentration-dependent manner, PD-L1⁺ EVs are upregulated by soluble IFN-γ and correlate to overall tumor burden^69,74. The identification of EVs as vessels for PD-L1 has emerged as a prominent subclass of EV-mediated immune suppression. Exosomal PD-L1 in immune evasion has been well reviewed elsewhere⁷⁵.

In addition to functional membrane-bound PD-L1, small EVs can carry miRNAs and enzymatically active Arginase-1 (ARG1) which both directly impact T cell activation, proliferation, and cytokine release via T cell receptor (TCR) downregulation^70,76. Transfer of miR-498 in vitro directly impacts the release of tumor necrosis factor (TNF) release in CD8⁺ T cells in metastatic melanoma⁷⁶. Meanwhile, miR-3187–3p can suppress CD45 membrane expression and consequent TCR activation in melanoma (Fig. 3)⁷⁶. ARG1, found in ovarian carcinoma EVs, has emerged as a metabolic mechanism for T cell dysfunction. A catalyst for the urea cycle, ARG1-mediated depletion of L-arginine suppresses T cell immune response in vitro by down-regulating TCR complex component CD3ζ⁷⁷. and by causing cell cycle arrest in the G1 phase via RICTOR in the mTORC2 complex^78,79. Isolated from patient plasma and ascites, ARG1⁺ ovarian cancer EVs actively suppress T cell proliferation, both directly and via DC antigen cross-presentation⁷⁰. ARG1⁺ cancer EVs have potentially far-reaching effects in T cell inhibition, as ARG1⁺ macrophages have been observed to actively arrest T cell proliferation in the absence of TCR signaling⁷⁸.

There are also indirect routes for EV-mediated suppression of T cells. Since CD8⁺ T cells primarily undergo activation and proliferation via APC upregulation of MHC I, indirect activation via APCs presents an alternative avenue for EV-mediated suppression¹¹. shRNA knockdowns of G-protein Rab27a in an in vitro irradiated prostate cancer model revealed a predominantly immunosuppressive role for EVs during antigen cross presentation⁸⁰. Inducing a non-native CD73⁺ phenotype to DCs, EVs derived from irradiated prostate cancer promote an adenosine mediated suppression of CD8⁺ T cell activation via impaired DC antigen presentation⁸⁰.

While EVs and their contents can drastically affect immune responses, it is important to understand where those signals originate. A significant body of evidence suggests external factors unique to the tumor microenvironment regulate the content and quantity of cancer EVs^81–84. Consequently, environmental stresses are often reflected by EV protein and RNA expression⁸². Hypoxia, an intrinsic property of desmoplastic and late-stage cancers, promotes increased secretion of EVs often rich in immunosuppressive proteins and miRNAs^83,85–87. A syngeneic mouse model of macrophage infiltration revealed that hypoxic small EVs isolated from B16F0 melanoma cells promote an anti-inflammatory M2-like phenotype in infiltrating macrophage cells⁸³. Furthermore, small EV transfer of miR-103a and Let-7a upregulated under hypoxic conditions promote M2-like polarization in lung and melanoma cancers^81,83, the latter working through the downregulation of an insulin-AKT-mTOR signaling pathway⁸³. Meanwhile, microvesicular miR-23a and TGF-β carried by hypoxia-induced EVs has implications for natural killer (NK) cell suppression via downregulation of CD107a and NKG2D in vitro (Fig. 3) ⁸⁷. Although chemical and physical cues may regulate EV content, the mechanism through which EVs systematically disseminate in tissue may also have implications for the downstream immunoregulatory capabilities of EVs.

Strikingly, EVs may acquire innate diffusive capacities within the tumor microenvironment. Physically restricted by a biocompatible nano-porous matrix, labeled EVs from mouse mesenchymal stromal cells were observed to readily diffuse through an otherwise spatially confined environment⁸⁴. Indicative of mechanosensing properties, changes in the matrix stiffness and stress served to enhanced EV transport through the nano-porous matrix. This enhanced diffusion is also accompanied by physical deformation of the EV itself⁸⁴. Mediated by the transport of water through aquaporins, AQP1 depletion increased the Young’s modulus (i.e. stiffness) of EVs and reduced diffusivity within the porous matrix⁸⁴. Interestingly, whether and how the physical properties of the tumor microenvironment can affect EV production, cargo, and function remain unexplored. Further research into the impact of biomechanical cues – including stromal matrix stiffness⁸⁸, topological cues^89,90, fluid shear stresses^91,92 – on EV content and function could also elucidate why many pre-clinical immune therapies based on EVs fail to return clinical benefits.

The role of extracellular vesicles in the antitumor immune response

While cancer EVs can mediate tumor progression, they also play an immunogenic role in the body’s response to cancer^93–103. Cancer EVs carry tumor-associated antigens (TAAs), damage-associated molecular patterns (DAMPs), and other cargo that are taken up by immune cells and leveraged to mount an antitumor response^{93,94,96–101} (Fig. 1). Cancer EVs deliver TAAs and peptide-MHC complexes (pMHC) for antigen presentation and tumor-specific T cell stimulation^{93,95,98,100,103}. More recently, lymphoma and melanoma EVs have been found to affect antigen processing machinery in DCs. Carrying mucin 1 (MUC1), NADPH oxidase 2 (NOX2), and reactive oxygen species (ROS), cancer EVs may alkalinize the DC phagosomal compartment via ROS production, making antigen processing within the DC more efficient^93,103. Cancer EVs may also carry DNA fragments, including GAPDH gene fragments and mtDNA, which induce maturation in DCs via the cGAS-STING pathway^96,100,102. However, the presence of dsDNA in small EVs has been disputed. Shown to co-isolate with small EVs, dsDNA may not be trafficked by small EVs at all, but instead released primarily and independently through a similar endosomal mechanism¹⁰⁴.

Cancer EVs can also exert immune-stimulating effects directly on effector cells¹⁰⁵. Melanoma EVs can increase IFN-γ secretion, promote proliferation, and increase the tumor-cytotoxic activity of NK cells¹⁰⁶. Displaying an activating NKp30 ligand BAG6 on their surfaces, melanoma EVs can trigger an antitumor response in NK cells upon binding¹⁰⁶. Furthermore, pancreatic and colon carcinoma EVs carrying heat shock protein 70 (Hsp70) stimulate NK cells to secrete granzyme B, an apoptosis factor in vitro¹⁰⁷. In the long term, however, this effect is reversed with tumor EVs promoting a decrease in cytotoxic activity¹⁰⁸. The ability of cancer EVs to directly stimulate effector cells is dependent upon the presence of adhesion and costimulatory molecules on their surfaces, in addition to the maturation status of the target cell¹⁰⁹. The mechanisms by which cancer EVs stimulate effector cells require further study.

EVs contribute to the immune response in several other ways. For example, glioma EVs reduce the presence of T_regs at the tumor site, attenuating inherent immune suppression¹¹⁰. EVs from non-small lung cancer activate mast cells in vitro and increase TNF and CCL2 production¹¹¹. Immune cell-derived EVs can also directly carry out antitumor functions^112–115. CD8⁺ T cell EVs are cytotoxic and can directly kill tumor cells^113,114. Furthermore, CD8⁺ T cell EVs deplete mesenchymal tumor cells via the transfer of cytotoxic miR-298 to prevent invasion and metastasis^113,114. DC EVs from hepatocellular carcinoma patients reduce T_regs at the tumor site¹¹⁵. Macrophages have also been found to suppress tumor immune evasion by taking up melanoma EVs to prevent their tumor-promoting interactions with B cells¹¹².

Several studies have noted discrepancies in function between soluble factors and their vesicular counterparts. Tumor antigens seem to be more efficiently taken up by immune cells when associated with EVs rather than in soluble form^93,95,100. Additionally, vesicular Hsp70 can contribute to radiotherapy resistance in tumors¹¹⁶, while immunization with non-vesicular Hsp70-peptide complex is associated with more positive effects^117,118. These findings cast doubt on whether functions are always correctly attributed to EVs instead of soluble factors. This is especially concerning since it is difficult to segregate soluble factors from EV preparations¹⁰⁴. While ideal single EV subtype models and validation techniques remain largely elusive, effects attributed to EV populations should always be extensively validated through multiple isolation and analysis techniques³⁵.

Clinical applications of extracellular vesicles in immunotherapy, drug delivery, and prognostics

The immunotherapeutic potential of EVs was first demonstrated in 1998 (Ref. ¹⁰). The properties (i.e., antigen presentation and costimulatory molecules) that made EVs a functional communication pathway have since been widely explored, leading to the inception of a new class of EV-centered cancer immunotherapies. After more than two decades of probing EV function in immune-cancer interactions, the field has turned its focus to the engineering and targeting of EVs for cancer therapy and diagnosis. As more immunological and oncogenic characteristics of EVs come to light, new therapeutic targets and mechanisms will become available (Fig. 4).

Figure 4: Potential clinical translations for EVs.

EVs have emerged as powerful tools for cancer treatment. Synthetic and re-appropriated natural EVs carry bioactive molecules that enhance the immunological roles of EVs, indicate disease prognosis, or facilitate drug delivery to the tumor site. These capacities are the focus of several ongoing clinical trials.

The involvement of EVs in cancer-immune cell crosstalk presents many opportunities for cancer immunotherapy. Proposed therapies use natural and engineered EVs to enhance the existing immune response to cancer, or to inhibit their immunosuppressive functions^119,120. Currently, there are no approved EV-based immunotherapies. However, a considerable number of clinical trials have started over the past five years.

EVs are ideal candidates for drug development and delivery due to their biocompatibility, stability, targeting capabilities, and scalability. EVs are enriched in adhesion and signaling molecules, which allow them to hone to target cells and stimulate uptake¹²¹. The presence of transmembrane CD47 permits EVs to avoid immune rejection via CD47-SIRPα “don’t eat me’ signaling¹²². This immune escape mechanism contributes to the extended circulation time of EVs in comparison to free drug or cell-based therapies, which are more susceptible to immune clearance. In addition to extended half-life, EVs demonstrate greater cellular targeting and uptake than free drug delivery^123,124.

A major advantage of EVs over cell-based therapies is their amenability to specialized and scaled-up production. Several platforms are being developed for scaled-immunoprecipitation (IP) production and purification of EVs^125,126. Recently described methods even allow for rapid, automated harvest and surface modification of EVs on a microfluidic device¹²⁷. For internal cargo modification, several methods have been established, including sonication, electroporation, and passive diffusion¹²⁸. Furthermore, EVs are better suited to long term storage than cells, experiencing limited loss of function^129,130. Cytokine release syndrome (CRS), however, remains a safety concern for the use of EVs in cancer immunotherapy. CRS is a potentially fatal reaction to immunotherapies that target T cells¹³¹. While it has not been investigated or observed in EV therapies, CRS should be taken into account during the development of EV-based cancer immunotherapies.

The presence of TAAs on cancer EVs makes DCs a major target for cancer EV-based immunotherapy^132–138. Treatment with EVs from colorectal carcinoma, glioblastoma, myeloid leukemia, renal carcinoma, melanoma, tongue carcinoma, and lung cancer has demonstrated increased T cell activation, proliferation, tumor infiltration, and tumor cytotoxicity with simultaneous decrease in immunosuppression and tumor growth^132–139. Proposed treatments collect, modify, and return tumor EVs to improve in vivo targeting of DCs and enhance the immune response^{132,134–138,140}. Melanoma EVs have been engineered to present immunostimulatory CpG-DNA for preferential uptake by DCs^136,140. Upregulation of IL-12¹³⁵ or the addition of miR-155¹³², pH-sensitive fusogenic GALA peptide¹³⁷, or DAMPs¹³⁸ to cancer EVs enhances antigen presentation and immune activation by DCs. Myeloid leukemia EVs have also been used to condition DCs ex vivo prior to DC vaccination to improve efficacy of treatment¹³⁴ (Fig. 4).

Other EV-based therapies inhibiting tumor progression involve EV-mediated immunosuppression strategies that preserve their immunostimulatory functions^{94,133,135,141}. Depletion of suppressive factors such as Siglec-9 ligands¹³³ or TGF-β^94,135,141 in glioblastoma, colon carcinoma, and leukemia EVs resulted in increased uptake by DCs and an enhanced antitumor immune response. Subjecting colon cancer cells to heat stress produces EVs enriched in Hsp70 that stimulates IL-6 production in DCs, increases T helper 17 polarization, and decreases T_reg polarization leading to an enhanced immune response¹⁴². Similarly, irradiated hepatoma cells secrete EVs enriched in TAAs, like CDCP1, and DAMPs, including Hsp70 and Hsp90⁹⁸. Another potential target to inhibit tumor EV-mediated immune suppression is myeloid-derived suppressor cells (MDSCs)¹⁴³. Hsp70 on breast, lung, and ovarian cancer EVs was found to bind TLR2 on MDSCs, stimulating EV-mediated suppression of T cells and NK cells. Addition of the peptide aptamer A8 blocks TLR2 binding and may prevent immune suppression by MDSCs¹⁴³.

Immune cell-derived EVs also have potential for cancer immunotherapy. DC EVs, in particular, have garnered attention for their ability to stimulate tumor-specific immune responses^144–146. DC EVs loaded with IFN-γ enhance this effect, increasing secretion of IFN-γ and TNF by NK cells¹⁴⁷. It was hypothesized that the addition of melanoma antigen recognized by T cells 1 (MART1) to these IFN-γ-enriched EVs might stimulate a tumor-specific response in NK cells, but the effect was small in humans¹⁴⁷. Other DC EV modifications aim to enhance their immune stimulating capabilities. Adding TAAs – such as melanoma-associated antigen 3 (MAGE-A3), MART1, glycoprotein 100 (gp100), or HPV16 E7 – to the surfaces of DC EVs increases uptake by monocytes and promotes T cell proliferation, stimulation, and cytotoxicity via DCs^127,148. Increasing alpha-fetoprotein (AFP) expression in DC EVs also enhances tumor immunity. AFP promotes enrichment of MHC I, MHC II, and costimulatory molecules on DC EVs, which induces a change in the tumor microenvironment from immunoinhibitory to immunostimulatory by increasing T cell tumor infiltration and reducing the presence of T_regs¹⁴⁹. Similarly, adding ovalbumin, lipopolysaccharides (LPS), and IFN-γ to DC EVs promotes the conversion of immunosuppressive M2 macrophages into immunostimulatory M1 macrophages, promotes antigen presentation in DCs, and directly activates T cells¹⁵⁰.

EVs from NK cells and macrophages have also been investigated for immunotherapy. NK cell EVs contain FasL and TNF, making them cytotoxic to cancer cells^151,152. Vesicles generated from NK cell membranes also exhibit these properties¹⁵¹. Modifications to enhance tumor attack include priming the NK cells with IL-15. This treatment increases TNF-related apoptosis-inducing ligand (TRAIL), and expression of the activatory receptors NKp46, and NKp30 on NK cell EVs, leading to more efficient tumor targeting and cytotoxicity¹⁵³. Macrophage EVs have been employed for antigen presentation to DCs. The addition of hyaluronic acid (HA), 3-(diethylamino)propylamine (DEAP), monophosphoryl lipid A (MPLA), and MUC1 trigger uptake by DCs and the release of TAA in the endocytic compartment for improved antigen presentation and T cell activation¹⁵⁴.

Clinical trials for extracellular vesicle-based cancer diagnostics and therapy

Early clinical trials investigating the use of EVs in cancer immunotherapy showed little more than that EVs are safe for human use^144–146. This has not deterred further attempts, however, as our knowledge of EV function continues to expand. Current clinical studies aim to determine the effects of immunotherapy, among other treatments, on vesicular cargo, including PD-L1 and miRNA expression profiles^155–160. Other clinical research is investigating the role of EVs in therapeutic resistance^159,161. EVs are hypothesized to contribute to therapeutic resistance via enrichment of immunotherapeutic targets, like CD20 and PD-L1. A possible solution, currently undergoing Early Feasibility Phase I clinical trials, is to deplete circulating EVs via a proprietary hemopurifier device ¹⁶², neutralizing their immunosuppressive and therapeutic resistance effects. Another Phase I clinical trial is investigating the immune modulation and anticancer effects of curcumin-loaded plant EVs ¹⁶³. An upcoming Phase 2 clinical trial investigating the use of an antisense oligodeoxynucleotide drug (IMV-001) against insulin-like growth factor type I has shown dependence on TAA-bearing tumor EVs to stimulate antitumor immunity. The IMV-001 drug released from an implanted bio-diffusion chamber for the treatment of malignant gliomas, induces apoptotic cell death in surrounding tumor cells. IMV-001 is hypothesized to work together with these TAA-bearing tumor EVs released from apoptotic cells to stimulate antitumor immunity in the surrounding tissue¹⁶⁴. The results of this treatment strategy have been positive thus far and emphasize the downstream potential for TAA loaded EVs to promote immune response^165–167.

Engineered and synthetic vesicles are gaining popularity for application in drug delivery due to their stability and homing capabilities. For example, oncolytic virus is an established cancer treatment, but due to rapid immune clearance, it must be administered locally¹⁶⁸. Encapsulation of the virus in cancer EVs enables systemic delivery and tumor site homing for the targeting of cancer metastases¹⁶⁸. Macrophage EVs have likewise been used to deliver the antitumor drug doxorubicin to the tumor site in vivo¹⁵⁴. DNA vaccinations may even be targeted to EVs to enhance immunogenicity. An expression plasmid encoding antigen-fused CD63 was able to deliver the antigen onto EVs in vivo, eliciting a stronger antitumor response in a syngeneic lymphoma model ¹⁶⁹.

EVs derived from the human HEK293 cell line are another common platform for cancer therapy. HEK293 EVs have been modified with signal regulatory protein alpha (SIRPα) to block CD47 at the tumor site, resulting in increased phagocytosis of tumor cells by macrophages and increased CD8⁺ T cell tumor infiltration¹²⁰. HEK293 EVs can also be loaded with PH20 hyaluronidase to break down high molecular weight HA in the tumor microenvironment. The resulting oligo-HA induce DC maturation via TLR4 activation and elicits a more potent antitumor response¹⁷⁰. Vesicles made from HEK293 cell membrane expressing PD-1 were used to block tumor PD-L1 and deliver indoleamine 2,3-dioxygenase-1 (IDO) inhibitor to the tumor microenvironment, resulting in reduced T_reg presence and improved antitumor response¹⁷¹. Efforts to decrease liver uptake of HEK293 EVs, thus increasing circulation time and tumor site accumulation, use dextran sulfate to block scavenger receptor class A (SR-A) on the EVs, optimizing their performance¹²⁵.

Efforts to engineer synthetic vesicles attempt to recapitulate the tumor-targeting and immune-stimulating properties of EVs, via the repurposing of cell membranes, for instance. Leukocyte membranes have been harvested and extruded into vesicles for the coating of synthetic microcapsules¹⁷². Melanoma cell membranes were similarly used to generate poly(ethylene glycol) (PEG)ylated nanovesicles capable of inducing an antitumor immune response¹¹⁹. Synthetic multivalent antibodies retargeted exosomes (SMART-Exos) are modified with tumor antigen and CD3 to induce antigen-specific immune responses^173,174.

EV isolation remains bound by labor-intensive purification techniques. Techniques such as immunoprecipitation (IP) and size exclusion chromatography (SEC) supplement the ad hoc-standard ultracentrifugation in a laboratory setting. However, extensive sample preparation and incubation steps negate the practical benefits of IP and SEC techniques clinically. While advancements in microfluidic “chip” isolations promise to bridge this laboratory-clinic divide, it reflects a time-untested methodology¹⁷⁵.

The quickest route to clinical practicality of EVs is as prognostic biomarkers. Currently limited by low throughput of isolation techniques, patient-derived EVs isolated from bodily fluids – whole blood, plasma, urine, ascites, etc. – and tissue biopsy have the potential to be a critical prognostic tool for immunologists. In addition to catapulting EVs into mainstream clinical use, high throughput EV isolation methods for liquid and solid biopsies might further reveal systematic vs local EV function. The focus of over 100 completed and ongoing clinical trials, the EV prognostic field is rapidly expanding. EVs from patient samples could possibly indicate diseases progression, as well as response to therapy, allowing for more personalized treatment.

Characterization of patient samples has, however, correlated several immune-related vesicular biomarkers with poor prognosis^176,177. Vesicular annexin II is involved in the activation of proinflammatory signaling in macrophages leading to breast cancer metastasis¹⁷⁷. This role and the correlation of vesicular annexin II with breast cancer progression suggest it may serve as a biomarker for breast cancer prognosis (Fig. 4). Similarly, elevated serum levels of vesicular miR-200b and miR-200c have been correlated with the spread of epithelial ovarian cancer to the lymph nodes (FIGO stage III-IV)¹⁷⁶.

In addition to predicting and measuring disease progression, EVs may indicate patient response to cancer treatment^178,179. In general, patients with long-term clinical remission of acute myeloid leukemia, as well as patients tested after induction of chemotherapy, have lower vesicular expression levels of TGF-β than those tested at diagnosis¹⁷⁸. Another study in anaplastic astrocytoma patients found that decreases in vesicular IL-8 and TGF-β correspond to an increase in immune activity after antitumor vaccination¹⁷⁹. The relationship between vesicular TGF-β and cancer progression is likely due to its role in NK cell suppression¹⁷⁸. TGF-β may, therefore, serve as a prognostic biomarker for patient response to chemotherapy.

EVs may emerge as powerful indicators of immunotherapeutic resistance as well^69,180–182. The monoclonal antibody drug trastuzumab is used to treat HER2⁺ breast cancer, but not all patients respond to the treatment. HER2 and epithelial cell adhesion molecule (EpCAM) present on breast cancer EVs have been found to sequester therapeutic antibodies and interfere with tumor targeting (Fig. 4)^180,181. Meanwhile circulating vesicular PD-L1 has proved to be a predictor of clinical response to the landmark immunotherapy pembrolizumab (anti-PD-1)^69,182. Thus, the presence of vesicular tumor antigens may serve as indicators of immunotherapeutic resistance.

Conclusion

Burdened by low yield and labor-intensive isolation techniques, the purity, quality, and characterization of EVs remain paramount for both pre-clinical and clinical research. Pre-clinical studies have revealed a diverse repertoire of EV functions, particularly with regards to cell maturation, antigen presentation, and immune suppression. Essential to both the antitumor and immune regulatory responses, the emergence of EVs as immune checkpoint vessels continue to be an exciting avenue for clinical translation. Moreover, the contribution of the physical properties of the extracellular matrix (e.g. matrix stiffness and porosity) to EV secretion, cargo, and consequent immune suppression remain largely unexplored. A holistic approach to immune suppression beginning with environment origin, immune cell migration, and consequent adaptive immunity may elucidate why and when EVs are able to suppress immune response.

Currently secondary in clinical applications, significant progress has been made in the past five years towards the use of EVs in cancer immunotherapy. Their safety, targeting capabilities, and clinical practicality make EVs an attractive candidate for drug development and delivery. Tumor and immune cell EVs have shown significant effects on cancer and the antitumor immune response in disease models. The challenge remains, however, to harness these abilities to produce a meaningful benefit in patients. For the time being, EVs are better suited to diagnostic and prognostic use than treatment. Future work should focus on developing EVs as adjuvant therapies to combat the side effects and drawbacks of established treatments.

Table 1:

Ongoing and completed clinical trials involving EVs at www.ClinicalTrials.gov

Clinical Trial	Identifier
Prediction of Immunotherapeutic Effect of Advanced Non-small Cell Lung Cancer	NCT04427475
Clinical Research for the Consistency Analysis of PD-L1 in Cancer Tissue and Plasma Exosome (RadImm01)	NCT02890849
Clinical Research for the Consistency Analysis of PD-L1 in Lung Cancer Tissue and Plasma Exosome Before and After Radiotherapy (RadImm02)	NCT02869685
Anaplastic Thyroid Cancer and Follicular Thyroid Cancer-derived Exosomal Analysis Via Treatment of Lovastatin and Vildagliptin and Pilot Prognostic Study Via Urine Exosomal Biological Markers in Thyroid Cancer Patients	NCT02862470
Study of Molecular Mechanisms Implicated in the Pathogenesis of Melanoma. Role of Exosomes (EXOSOMES)	NCT02310451
Study of Exosomes in Monitoring Patients With Sarcoma (EXOSARC) (EXOSARC)	NCT03800121
Exosomes and Immunotherapy in Non-Hodgkin B-cell Lymphomas (ExoReBLy)	NCT03985696
Hemopurifier Plus Pembrolizumab in Head and Neck Cancer	NCT04453046
Study Investigating the Ability of Plant Exosomes to Deliver Curcumin to Normal and Colon Cancer Tissue	NCT01294072
Pilot Immunotherapy Trial for Recurrent Malignant Gliomas	NCT01550523
Antisense102: Pilot Immunotherapy for Newly Diagnosed Malignant Glioma	NCT02507583