Introduction

Cancer immunotherapy (IT), including treatment with immune checkpoint inhibitors (ICIs), has been approved for treatment of patients with many cancer types. Despite an overwhelming recent success of ICIs, many patients with cancer do not respond to IT.1 The reasons for this are not entirely clear but suggest a failure of the host immune system to be rejuvenated by IT and effectively mediate antitumor immunity. It has been well established that in patients with cancer, especially those with advanced malignancies, antitumor responses are impaired, enabling tumor escape.2 Tumors induce and drive local and systemic immune suppression by a variety of mechanisms that may include (1) inhibitory receptors/ligands (eg, PD-1/PD-L1); (2) soluble factors (eg, inhibitory cytokines, IDO, arginase, galectin 9, adenosine, COX-2, PGE2, oxygen radicals among others); (3) regulatory immune cells (eg, Treg, myeloid-derived suppressor cells (MDSCs); (4) metabolic alterations in the TME (eg, glucose deprivation); and (5) MHC class I downregulation or loss of β2-microglobulin on tumor cells (reviewed in Whiteside3). Tumor cell derived small extracellular vesicles (sEV) or TEX are the latest addition to this broad list of factors that tumors produce and use to silence the immune cells and subvert antitumor immunity.

EVs are heterogeneous in size and biogenesis.4 The EV subset of sEV may contain “ectosomes” formed by budding of the cell membranes and “exosomes” which are formed in the endocytic compartment of parent cells. Exosomes are released into the extracellular space via specific biogenesis pathways and can be recognized by the presence of the endocytic markers on their surface. The sEV are secreted by live cells and are present in all body fluids. They circulate freely and, like other types of EVs, mediate intercellular communication and represent a physiologically indispensable, evolutionarily preserved system present in bacteria, plants, animals, and humans. In patients with cancer, circulating sEV contain a subset of tumor-derived vesicles, referred to as “tumor cell-derived exosomes” or TEX. Vesiculation is a common feature of all tumor cells, and tumors produce and release large numbers of vesicles that vary in size, cellular biogenesis, and molecular cargos (figure 1).

Figure 1

Biogenesis of tumor-derived exosomes. The exosome vesiculation process consists of the initial endocytosis of surface-associated molecules, their uptake by endosomes, reverse vesicle formation in late endosomes and fusion of endosomes into a multivesicular body (MVB). When the MVB fuses with the tumor cell membrane, exosomes are released into the extracellular space. They are heterogeneous in size approximating that of viruses (see the TEM image in upper right) and differ from various larger vesicles in the molecular content as well as the cellular origin. The small EV may also be formed by “budding” of the cell membrane like microvesicles. The latter originate from the cytosol by “budding” of the cell membrane. The vesiculation process is ATP-dependent and occurs in live cells. EV, extracellular vesicle; MVB, multivesicular body.

These vesicles include sEV or TEX (30–150 nm) which are made in late endosomes and multivesicular bodies (MVBs) in tumor cells as well as ectosomes of the same size which derive from cytosol by budding of the cell membrane; larger vesicles (microvesicles or MVs and oncosomes), which pinch off the tumor cell membrane; and the largest, apoptotic bodies, which are products of dying tumor cells. TEX are of special interest: they are packaged in tumor cells by complex sorting mechanisms and carry cargos that are being processed by the tumor cell. They interact with a broad variety of cell types and on delivery to recipient cells, alter cellular metabolism, functions and behavior of these cells.5 TEX differ from EVs produced by non-malignant cells by the molecular cargo, genetic content (RNA, miRNA, and DNA) and by distinct functional changes they induce on interactions with recipient cells. While EVs derived from healthy cells carry cargos that consist of common cellular proteins, TEX are “miniature surrogates of tumor cells’’, and their molecular and genetic content resembles that of parent tumor cells. For example, TEX carry tumor-associated antigens (TAAs),5 and TEX isolated from human melanoma cell lines with mutated BRAF contain DNA with the mutant BRAF alleles in the vesicle lumen.6 In comparison to parent tumor cells, TEX have many advantages that facilitate their pro-tumor and anti-immune cell functions. These advantages include the TEX size approximating that of viruses, vesicular morphology, motility, abilities to cross all tissue barriers and to avoid immune rejection.7 TEX circulate freely, cross all tissue barriers, including the blood–brain barrier,8 and are especially well equipped to carry and deliver messages to immune and non-immune target cells, altering functions of the recipient cells. How tumors imprint the cellular address on TEX they produce is not known. TEX are decorated by and carry in their lumen a variety of immunosuppressive factors tumor cells produce, including those listed above, and suppress activities of immune effector cells in vitro and in vivo.9 Although TEX also carry an abundance of functional costimulatory proteins, TEX largely deliver negative signals in the TME by preferentially engaging those cellular/molecular mechanisms in recipient cells that promote immune suppression.10

Tumors produce large numbers of immunosuppressive exosomes

TEX are a subset of circulating sEV, and their frequency in body fluids of patients with cancer may vary depending on rates of their production and removal by the host reticuloendothelial system (RES). Recently developed methods for separation of TEX from non-malignant sEV in cancer plasma have enabled an estimation of circulating TEX frequencies.11 For example, in patients with metastatic melanoma, numbers of circulating TEX are estimated to be as high as 1012/mL of plasma, and TEX variably accounted for a small (eg, <5%) or large (eg, >50%) fraction of circulating sEV, depending on disease activity at the time of sampling for TEX.10 It has been suggested that tumors produce large numbers of TEX not to discard cellular refuse but rather to reutilize TEX for support of tumor growth via the autocrine pathway. Experiments in mouse models of tumor growth suggest that blocking the autocrine pathway of TEX re-utilization impairs tumor growth on the one hand and promotes antitumor functions of spleen derived T cells on the other. Results of numerous experiments with primary human immune cells coincubated with TEX secreted by various human tumors confirm TEX capabilities to induce dysfunction and death of effector immune cells.12 Specifically, TEX can directly interfere with functions of immune cells by inducing changes in cellular metabolism (eg, mitochondrial dysfunction), in transcription factor signaling (eg, NF-κB signaling) in the cytokine profile (eg, reduction in IL-2 or IFN-γ in T cells), inhibition of migration, proliferation, cytotoxicity and by driving apoptosis.12–14 TEX can also indirectly inhibit immune cell functions by promoting growth and activity of regulatory T cells (Treg) and MDSCs.15 Mechanisms of TEX-mediated immune suppression may involve upregulated expression of death receptor ligands on the vesicle surface, ability to activate immunosuppressive pathways (eg, the adenosine/PGE2 pathways) and/or to enhance activity of inhibitory cytokines (ie, IL-10, TGF-β1) in recipient cells as well as the transfer of oncogenic proteins or genes from the tumor to immune cells.12–14 The simultaneous presence on the TEX surface of numerous immunoinhibitory death receptor ligands (eg, PD-1, CTLA-4, TRAIL, and FasL) suggests that TEX are well equipped to mediate immune suppression in recipient cells expressing complementary receptors.14 Further, immunosuppressive signals that TEX carry are highly amplified, as they are distributed among millions of TEX released by tumor cells to simultaneously engage recipient cells. In aggregate, the emerging data provide a compelling rationale for the key significance of TEX in the tumor promotion and suppression of immune cell functions in tumor-bearing hosts, thereby facilitating tumor immune escape. At the same time, TEX also carry and deliver to recipient cells a variety of costimulatory proteins, including CD40, CD40L, OX40, OX40L, CD80,14 which could promote pro-tumor or antitumor immune responses. Thus, TEX molecular surface profiles confirm their immunoregulatory potential that could be differentially used probably depending on the environmental requirements.

TEX reprogram functions of cells in the TME

An additional and critical consequence of enhanced TEX production by tumors is reprogramming of the TME, including all malignant and non-malignant cells, by imparting to them the capability to produce and release a new crop of sEV which mimic the pro-tumor attributes of TEX. The cells reprogrammed by TEX retain their morphology and phenotypic profiles but make and release sEV that phenotypically and functionally resemble TEX. Thus, TEX-induced reprogramming of the TME results in the amplification of TEX mediated effects, that is, the enhanced production of negatively signaling sEV by immune and non-immune cells in the TME.13 As a result of molecular/genetic changes TEX induce in the TME, healthy cells acquire abilities to promote tumor growth by enhancing tumor resistance to elimination by immune cells, while tumor cells acquire resistance to external or internal antitumor agents.

Research into the role TEX play in reprogramming of T cells has provided some clues about mechanisms that restrain immune cells from harming the tumor while simultaneously boosting tumor resistance to IT.12 Masses of TEX (think millions of vesicles) are constantly produced by a growing tumor. TEX fill extracellular spaces and transmigrate freely into blood vessels, interacting with the surrounding tumor infiltrating and/or circulating immune cells. The unfolding molecular events have been partly defined for CD8+T cells, as described below, but are likely to apply more broadly to TEX interacting with other cell types. On contact with numerous vesicles highly enriched in surface integrins and opsonins, the T cell receives multiple signals that enable TEX to cross the surface cell membrane. TEX entry into the cytosol may involve receptor-ligand interactions, fusion, endocytosis, pinocytosis, phagocytosis, or all of the above, leading to a series of events that culminate in T cell reprogramming. Interestingly, once initiated, the TEX uptake by activated CD8+T cells could not be blocked by inhibitors of the above listed entry mechanisms.14 The simultaneous uptake and signaling by numerous vesicles are stressful for the recipient T cell and result in upregulated expression of stress proteins (PERK, BIP, CHOP, ATF4 or IRE1) and in the unfolded protein response. These changes translate into mitochondrial dysfunction as evidenced by cytochrome C release from mitochondria into the cytosol and initiate intrinsic apoptosis culminating in an irreversible death of the T cell.14

Another scenario involves TEX entry into a recipient T cell followed by TEX disrobing and release of their content comprising nucleic acids, proteins, enzymes, and cytokines, all of which are functional and can be readily engaged into reprogramming mechanisms in recipient cells. Like T cells, other immune cells, B cells, NK cells, monocytes, dendritic cells (DCs) and neutrophils, are susceptible to TEX-mediated reprogramming, by diverse mechanisms. While the B cells uptaking TEX enriched in ectonucleotidases, CD39 and CD73, are induced to upregulate the adenosinergic pathway,16 NK cells downregulate expression of NKG2D receptors as well as cytotoxic activity.1 TEX produced by HNSCC, increased maturation and activation of immature DCs potentiating their ability to phagocytose more vesicles and suppressed functions of mature DCs by downregulating expression of the APM components.17 In NSCLC, TEX polarized macrophages toward a PD-L1high immunosuppressive phenotype through metabolic reprogramming. TEX signaling through TLR2 and NF-KB increases glucose uptake by macrophages and also stimulates INOS expression, which inhibits oxidative phosphorylation in mitochondria leading to elevated conversion of pyruvate to lactate and to the upregulation of PD-L1 on macrophages.18 Neutrophils uptaking TEX derived from HNSCC upregulated adenosinergic signaling and PD-L1 expression and acquired pro-tumor activity that was mediated via the sEV produced by reprogrammed neutrophils.19 The TEX-mediated reprogramming likely involves miRNA and DNA as well as transcription factors and enzymes delivered by TEX to recipient cells and, in effect, leads to transcriptional and translational reorganization of the recipient cell activity. As a result, the former antitumor immune effector cells now produce vesicles that promote tumor growth and induce suppression or apoptosis of circulating antitumor effector immune cells, including those cells entering the TME. Preliminary14 20 and unpublished data from the author’s laboratory support this view of immune cell reprogramming by TEX. However, emerging insights into the various reprogramming mechanisms TEX appear to use begs a question of whether different cancers produce TEX carrying defined reprogramming tools and whether there is a “division of labor” among TEX produced by a given tumor to address different recipient cell types.

Mechanisms of information transfer by TEX may also involve indirect TEX interactions with the recipient immune cells. These are mediated via regulatory T cells (Treg) and/or MDSCs. TEX drive proliferation and differentiation of these regulatory immune cells and enhance their suppressive functions.15 Specifically, TEX were shown to upregulate expression levels of surface ectonucleotidases, CD39 and CD73, and adenosine production by Treg, converting them into highly effective immunosuppressive cells.21 The mechanistic details of TEX-mediated reprogramming of other recipient cells in the TME, for example, endothelial cells or cancer-associated fibroblasts, are still being defined. Nevertheless, emerging data indicate that tumors operate the vesicular system to regulate and promote cancer progression and resistance to therapies. The vesicular system is ubiquitous, and it represents a highly effective barrier for immune therapy. Studies are underway to define cellular mechanisms that TEX harness to induce resistance of tumor cells to cancer therapies, and preliminary results suggest the engagement of autophagy in TEX-mediated resistance of tumor cells to IT.

TEX eliminate adoptively transferred immune cells and induce tumor resistance to IT

In 2017, we reported the results of a phase I clinical trial with Neukoplast (NK-92 cells) in patients with refractory/relapsed AML (IND-BB 84040). In this dose-escalating safety study, cultured NK-92 cells were delivered IV.1 The PBMC obtained from patients prior to and after NK-92 delivery were assessed for NK cell activity. Therapy with NK-92 cells was well tolerated: no DLT, no clinical improvement and no immune recovery as per immune cell phenotypes, functions or cytokine plasma levels were seen. The trial was terminated after 7 patients were treated. Exosomes isolated from patients’ plasma collected before therapy were tested for suppression of NK-92 cell cytotoxicity. Importantly, exosome numbers/mL of patients’ plasma obtained prior to therapy were higher than those seen at AML diagnosis, and these exosomes were highly enriched in immunosuppressive proteins (PD-L1, FasL, TGF-β, CD39, and CD73). They were readily taken up by NK-92 cells, significantly downregulated expression of NKG2D in NK-92 cells and eliminated NK-92-mediated cytotoxicity (figure 2).

Figure 2

Exosomes from plasma of AML patients interfere with functions of NK92 cells. NK-92 cells were coincubated with exosomes isolated from pre-therapy AML plasma for 24 hours. (A) Flow cytometry of NK-92 cells illustrates downregulation of NKG2D protein. (B) Confocal microscopy illustrates the presence of GFP-labeled exosomes in NK-92 cells and downregulation of red-labeled NKG2D protein from the NK-92 cell surface. (C) Quantitation of the NKG2D downregulation in NK-92 cells. (D) A loss of NK-92 cytotoxicity. Reproduced from Hong et al.1

The lack of clinical response in this trial was likely due to the direct interference of circulating TEX with adoptively transferred NK-92 cells. TEX carrying an immunosuppressive cargo were present in patients’ plasma prior to adoptive IT and eliminated infused NK-92 cells thus protecting the tumor and increasing its resistance to IT. The trial serves as an example of how TEX can protect the tumor by interfering with anti-leukemia functions of adoptively delivered NK-92 cells.

More recent in vitro studies show that coincubation of TEX from plasma of patient with various tumors results in rapid dose dependent apoptosis of primary activated CD8+T cells or of CAR-T cells independently of CAR-T cell specificity (TLW, unpublished data). Also, a recent report showed that TEX carrying PD-L1, and targeted antigens preferentially interacted with cognate CAR-T cells inhibiting their proliferation, migration, and function.22 These studies indicate that TEX are armed to directly interfere and block development and functions of antitumor effector cells and thus represent a highly efficient protection mechanism for tumors, contributing to their resistance to therapies.

Another mechanism TEX use to subvert IT involves their capability to bind to and thereby deplete therapeutic antibodies delivered to patients with cancer.23 Large numbers of TEX carrying TAAs circulate freely in patients with cancer and specifically bind therapeutic mAbs. The TEX-mAbs aggregates are removed by the RES, and only a part of the intended therapeutic dose reaches the tumor. This mechanism may be partially responsible for the lack of tumor responsiveness to antibody-based IT.

TEX cargos contain TAA and costimulatory proteins in addition to immunosuppressive molecules.8 14 This provides a rationale for viewing TEX as potential anticancer vaccines. Indeed, numerous attempts have been made to use TEX alone or TEX with various adjuvants to immunize animals against cancer or increase susceptibility of cancers to IT. In most of these studies, including early clinical trials in patients with advanced cancers.24 TEX failed to induce actionable antitumor responses, presumably because of the prevailing immunosuppressive nature of the TME and pre-existing resistance of established tumors to IT. In subjects with the established immunosuppressive TME, where immune cells are targets for TEX induced elimination, little hope and numerous challenges exist for successful immunization without prior silencing of TEX.

Tumors produce TEX and reuse them via an autocrine mechanism that appears to be essential for tumor growth and survival. Blocking of TEX release or TEX reutilization by tumors in experimental in vitro and in vivo systems inhibited tumor progression.25 Preliminary data suggest that massive TEX production by tumors is a critical factor in driving autophagy in tumor cells. The role of TEX in engaging autophagy is under intense scrutiny because this mechanism appears to largely account for TEX ability to promote tumor resistance to IT.

TEX as components of liquid tumor biopsy

Recent data indicate that TEX also emerge as potentially highly specific and sensitive prognostic and predictive biomarkers of cancer progression and survival.26 Two features of TEX account for their efficacy as biomarkers: (1) they carry proteins and transcriptional signatures that associate with the tumor clinicopathologic status, disease progression and outcome26 27 and (2) TEX have the capability to reprogram phenotype and functions of targeted recipient cells thereby revealing the degree of existing cancer-induced dysfunction that can be linked to disease activity, stage, progression and response to therapy.28 TEX profiles, as determined by mass spectrometry and nucleic acid analyses, not only mimic the tumor cell genetic and molecular content (liquid tumor biopsy) but also provide a view of functional competence of immune cells (liquid immune cell biopsy) as recently discussed.12 29 The role of TEX as cancer biomarkers has been enhanced by the discovery that TEX carry genomic DNA, including cancer-specific mutated genes, in their lumen, thus providing a more reliable source of information than cell-free (cf) DNA.6

Conclusions

It is well documented that immune cells in the tumor-bearing hosts are functionally compromised, specifically lacking the ability to arrest cancer progression. TEX carrying an immunosuppressive molecular and genetic cargo are enriched in body fluids of patients with cancer and interact with all cell types in the TME. TEX carry tumor relevant information (proteins and nucleic acids) which on delivery to recipient cells leads to transcriptional and translational processing and results in metabolic alterations and distinct functional phenotypes. TEX convert healthy normal immune and non-immune cells into producers of vesicles that, like TEX, mediate pro-tumor activities. Specifically, CD3(+)CD8(+) effector T cells interacting with TEX are either eliminated by intrinsic apoptosis or reprogrammed to produce CD3(+)sEV that are enriched in immunosuppressive proteins and promote tumor growth. This creates a vicious cycle of immune suppression that leads to increased tumor resistance and may be of great clinical significance in the context of IT. In a broader view, emerging data indicate that TEX-mediated reprogramming of all cells present in the TME and in all body fluids of patients with cancer significantly contributes to functional impairment of antitumor immunity. At the same time, it enhances pro-tumor activity and tumor resistance to cancer therapies by mechanisms involving simultaneous activation of major molecular pathways, for example, NF-κB, STAT/JAK, ATP-to-ADO, TGF-β, MAPK/ERK, PDL-L1/PD-1 and others.5 13 Importantly, the pro-tumor activities induced in T cells by TEX could not be blocked with conventional antibodies or pharmacologic inhibitors targeting death receptor/ligands in preclinical experiments.14 This finding indicates that overcoming TEX mediated pro-tumor effects may be a challenge for cancer IT. Efforts to engineer and produce sEV with antitumor activity that potentially could replace the “danger bearing” TEX are ongoing and could provide novel strategies for eliminating/decreasing tumor-driven immune suppression and resistance of cancers to IT.

Ethics statementsPatient consent for publication

Not applicable.