It is generally assumed that nearly all cells in the body release vesicles into the surroundings that are termed extracellular vesicles (EV) [1–5]. The term EV is presently used to define various types of vesicles released by normal cells (i. e. erythrocytes, platelets, leukocytes) and cancer cells . EV secretion is particularly active in proliferating cells, such as cancer cells . Small EVs (50–150 nm) are released from the cell surface (microvesicles) or from the endosomal system (exosomes) that can presently not be separated from each other . By the fusion of late endosome with the plasma membrane, exosomes are secreted into the extracellular space [8, 9]. Here we will use the term “exosomes” to refer to small EVs. In contrast, apoptotic vesicles (bodies) are larger (1000–5000 nm) and can be separated by size and density from smaller vesicles [8, 9]. Depending on the cell of origin, exosomes cargo different biologically active molecules such as proteins, mRNA, microRNA (miR) and lipids and act as mediators of intercellular communication [10, 11].
But once released, what is the fate of exosomes? The situation is similar to the constant generation of apoptotic or necrotic cells that result from tissue homeostasis, a necessary process in a multicellular organism . Dead cells display signals essential for phagocytic clearance. The most important "eat me" signal is the exposure of phosphatidyl-serine (PS) on the cell surface membrane . Exosomes also display PS at the surface and may undergo phagocytic clearance similar to dead cells. It can be assumed that their lifetime in the body is rather short.
However, the situation is totally different in cancer. It is known that exosomes play an important role in the immune suppression mediated by cancer cells . Tumor-EVs carry the PD-L1 capable of suppressing T lymphocyte function by binding to its counter receptor PD-1 on T lymphocytes . In addition, exosomes released from melanoma cells are involved in the induction of myeloid-derived suppressor cells (MDSCs) [16, 17] that are of central importance for tumor-mediated immune suppression . Cancer-derived exosomes are believed to travel to target organs to prepare for metastatic niches [19–21]. Thus, to reach distant sites, cancer-derived exosomes need to enter the blood stream where they encounter a hostile environment in lymph nodes and other organs.
Given the apparent similarities between the clearance of apoptotic/necrotic cells and exosomes, we will first give a brief summary of the basic principles of phagocytosis.
Phagocytosis: “find-me”, “eat-me” and “don’t eat-me” signals
Phagocytosis is an important mechanism that allows the innate immune system to clear the body from cellular debris. Professional phagocytes include neutrophils, monocytes, macrophages, DCs, osteoclasts, and eosinophils . Many pathological processes give rise to increased apoptosis and necrosis and are characterized by enhanced phagocytic clearance. In cancer, monocytes and macrophages participate in radio- and chemotherapy by removing apoptotic and necrotic tumor cells and debris with or without inflammation .
Apoptotic or necrotic cells send out “find-me” and “eat-me“ signals for clearance. Important “find-me” signals released from apoptotic cells are among others the nucleotides ATP and UTP , whereas the exposure of PS is a common „eat-me“ signal for the clearance of cellular debris from both apoptotic and necrotic cells . Phagocytic cells can recognize PS using a number of different receptors, such as TIM-1, TIM-3, TIM-4, resulting in clearance of the cellular remainders by phagocytes . In addition to these most important interactions, parallel mechanisms are operating that we spare here for space reasons but are extensively described elsewhere .
In addition to positive “eat-me” signals, there are also “don’t eat me” signals that block phagocytosis. The most prominent one is CD47, a thrombospondin receptor, that is expressed on the surface of all human solid tumor cells and can be released by exosomes [25–27]. CD47 is a ligand for SIRPα, a protein expressed on macrophages and dendritic cells and functions as a "don't eat me" signal for phagocytic cells . Binding of SIRPα with CD47 causes phosphorylation of the immune-receptor tyrosine-based inhibition motif (ITIM) on the cytoplasmic tail of SIRPα . This results in the inhibition of phagocytosis through preventing myosin- IIA accumulation at the phagocytic synapse .
Importantly, blocking mAbs against CD47 have shown remarkable effect in tumor therapy. Thus, CD47 represents a valid target for cancer therapies . CD47 expression on tumor cells does not only inhibit phagocytosis of cancer cells but also affects anti-cancer T-cell responses due to an impaired antigen presentation .
Another molecule recently discovered as “don’t eat-me” signal is CD24. The work of Bakal et al  pointed out that CD24 played an important role at the interphase of the immune system and tumor cells. In search for new phagocytosis inhibitors, they observed that CD24 on ovarian cancer (OvCa) or triple-negative breast cancers (TNBC) acted as an anti-phagocytic surface protein . CD24 on tumor cells interacted with Siglec-10 on macrophages . Genetic ablation or therapeutic blockade of CD24 resulted in macrophage-dependent reduction of tumor growth in vivo and an increased survival .
In the immune system, the CD24-Siglec-10 axis was already known to play a role [32, 33]. Siglec-10 is a member of the family of sialic-acid-binding immunoglobulin-like lectins that regulate the functions of immune cells [34, 35]. Siglec-10 is an inhibitory receptor that can trigger recruitment of SHP1 to its intracellular ITIM motif [34, 35]. Thus, CD47 and CD24 elicit similar downstream intracellular signaling. It is worth mentioning that soluble CD24 as CD24-Fc is already in clinical trials and is exploited as a novel drug for dampening of over-shooting immune reactions for example in auto-immunity [32, 36].
CD24 is predominantly expressed on the surface of B lymphocytes, monocytes, and granulocytes but is found in many cancer entities [37, 38). It has been suggested that, next to the frequency of p53 mutations, expression of CD24 is the most frequently overexpressed antigen in cancer cells . CD24 is a membrane glycoprotein with a small protein core of 31 amino acids, and extensive N- or O-linked glycosylation linked to the membrane by a glycosyl-phosphatidylinositol (GPI)-anchor [37, 39]. In cancer cells, CD24 is known as a regulator of cell migration, invasion and proliferation [37, 40, 41], as bad prognostic marker in pathology [37, 42] and as a marker for cancer stem cells . Due to its GPI-anchor, CD24 is readily recruited into exosomes and can be used as an EV marker. CD24+ exosomes were reported to be present in many body fluids including malignant ascites of OvCa patients , or sera and pleural effusions from breast cancer patients .
Implications for EV phagocytic clearance
Assuming a similar phagocytic clearance of apoptotic cells and exosomes, we can speculate that CD47 and CD24 act as “don’t eat-me” signals for exosomes. Under normal conditions, EVs released from healthy cells displaying PS as „eat-me“ signals at the surface and are removed by phagocytosis (Fig. 1A).
As an alternative mechanism, exosomes my undergo opsonisation by auto-antibodies followed by Fc-receptor mediated phagocytosis or complement-mediated lysis [46, 47] (Fig. 1B). In contrast to normal exosomes, cancer derived exosomes display CD47 and CD24 as phagocytic inhibitory receptors (Fig. 1C). The presence of these molecules will markedly attenuate phagocytic clearance. The protection from phagocytosis may favor a longer survival of tumor-EVs in the tumor microenvironment and the blood.
Is there evidence that “don’t eat me“ signals on EVs such as CD47 and CD24 are functionally operating? Kamerkar et al  showed that therapeutic exosomes collected from mesenchymal stem cells carrying short interfering RNA or short hairpin RNA specific to oncogenic Kras (G12D) were stabilized in the blood circulation when expressing CD47. The authors claimed that the enhanced retention of exosomes, compared to liposomes, in the circulation of mice was due to CD47-mediated protection of exosomes from phagocytosis by immune cells .
Kaur et al  reported that EVs released from cells expressing or lacking CD47 differentially regulate activation of T cells induced by engaging the T cell receptor. Similarly, the uptake of T cell derived EVs by recipient endothelial cells globally altered gene expression in a CD47-dependent manner.
A particular protective role for CD24 on the surface of EVs was not investigated until now. However, Siglec-10 is expressed on myeloid and T and B lymphocytes and its binding could modulate phagocytosis and other responses of the immune system. Thus, in light of the ubiquitous presence of CD24 on cancer cells, it may be important to reconsider its role on tumor-derived exosomes. We suggest, that the presence of CD24 could protect exosomes from phagocytic removal and guarantee their survival.
The peripheral blood of healthy donors contains very few EVs. This could be the result of an efficient clearance system. In contrast, in cancer patients several studies have shown that the amount of EVs in the peripheral blood is much higher than in healthy donors [4, 45]. A trivial explanation could be that the release of EVs is higher under the influence of a growing tumor. We propose here an alternative explanation, that cancer-derived exosomes are protected from removal by phagocytosis by displaying “don’t’eat me” signals. This has advantages for the tumor cells as well as for their released exosomes.
In addition to CD47 and CD24, previous work has identified other “don’t eat me” signals . Such surface molecules comprise programmed cell death ligand (PD-L1), and beta2-microglobulin associated with class I MHC molecules . Antibodies against these molecules that antagonize the interaction of “don’t eat me” signals with their respective receptors on macrophages (termed phagocytic checkpoints) have demonstrated therapeutic potential in several cancer types . Also antibodies to CD24 have shown remarkable therapeutic efficacy in vivo in preclinical animal models . It remains unclear whether these effects are due to the targeting of the tumor cells or of the CD24+ EVs. Future studies will be needed to decipher the mechanisms of such therapeutic effect.