Extracellular vesicles in tumor angiogenesis and resistance to anti‐angiogenic therapy

Abstract Tumor angiogenesis plays an important role in the development of cancer as it allows the delivery of oxygen, nutrients, and growth factors as well as tumor dissemination to distant organs. Although anti‐angiogenic therapy (AAT) has been approved for treating various advanced cancers, this potential strategy has limited efficacy due to resistance over time. Therefore, there is a critical need to understand how resistance develops. Extracellular vesicles (EVs) are nano‐sized membrane‐bound phospholipid vesicles produced by cells. A growing body of evidence suggests that tumor cell‐derived EVs (T‐EVs) directly transfer their cargoes to endothelial cells (ECs) to promote tumor angiogenesis. Importantly, recent studies have reported that T‐EVs may play a major role in the development of resistance to AAT. Moreover, studies have demonstrated the role of EVs from non‐tumor cells in angiogenesis, although the mechanisms involved are still not completely understood. In this review, we provide a comprehensive description of the role of EVs derived from various cells, including tumor cells and non‐tumor cells, in tumor angiogenesis. Moreover, from the perspective of EVs, this review summarized the role of EVs in the resistance to AAT and the mechanisms involved. Due to their role in the resistance of AAT, we here proposed potential strategies to further improve the efficacy of AAT by inhibiting T‐EVs.


cell-derived EVs (T-EVs) directly transfer their cargoes to endothelial cells (ECs) to
promote tumor angiogenesis. Importantly, recent studies have reported that T-EVs may play a major role in the development of resistance to AAT. Moreover, studies have demonstrated the role of EVs from non-tumor cells in angiogenesis, although the mechanisms involved are still not completely understood. In this review, we provide a comprehensive description of the role of EVs derived from various cells, including tumor cells and non-tumor cells, in tumor angiogenesis. Moreover, from the perspective of EVs, this review summarized the role of EVs in the resistance to AAT and the mechanisms involved. Due to their role in the resistance of AAT, we here proposed potential strategies to further improve the efficacy of AAT by inhibiting T-EVs.

K E Y W O R D S
anti-angiogenic therapy, extracellular vesicles, non-tumor cells, resistance, tumor angiogenesis, tumor cells the United States Food and Drug Administration (US FDA) for several types of advanced cancers. 1 However, this potential strategy has shown limited efficacy, with survival benefits ranging from only a few weeks to several months. Many patients who initially respond to AAT eventually develop resistance over time. Therefore, it is vital to understand how primary or acquired resistance develops. 2 EVs are nano-sized membrane-bound phospholipid vesicles that are likely to be produced by all types of cells. 3,4 As a critical intercellular communicator, EVs carry different bioactive molecules (lipids, nucleic acids, and proteins) and can deliver those cargoes to receipt cells to participate in many biological processes, particularly in the development of cancers. 4 A growing body of evidence now suggest that tumor cell-derived T-EVs directly transfer bioactive cargoes to ECs to promote tumor angiogenesis. 5 Recent studies reported that T-EVs play a critical role in the development of resistance to AAT, but the underlying mechanisms remain obscure. In this review, we summarized recent studies on tumor angiogenesis mediated by EVs derived from both tumor cells and non-tumor cells. Moreover, we discussed the resistance of AAT from the perspective of EVs and proposed potential strategies to further improve the efficacy of AAT. We hypothesized that, due to the important role of T-EVs in the resistance of AAT, EVs might be served as an effective and potential target to enhance the efficiency of AAT in the future.

| B I OG ENE S IS AND CL A SS IFI C ATI ON OF E VS
EVs are a diverse population of membrane vesicles generated via diverse mechanisms. On the basis of their formation mechanism and size, EVs are generically categorized as exosomes and microvesicles (MVs), and apoptotic body. 7 Exosomes have a complex biogenesis process. As shown in Figure 1, the plasma membrane buds inwards and then endocytosed, thus forming a lipid bilayer structure known as the early endosome. Subsequently, these endosomes continue to invaginate, forming vesicles containing many intraluminal vesicles (ILVs), known as multivesicular bodies (MVBs). Then, MVBs fuse with the plasma membrane of mother cells and then release exosomes into the extracellular environment. 8 Unlike exosomes, microvesicles arise through direct outward budding and fission of the plasma membrane without going through pathways such as ILV and MVB. Apoptotic F I G U R E 1 The biogenesis of extracellular vesicles (EVs). Exosomes originate from the endocytic pathway, and the plasma membrane invaginates to form endosomes, and endosomes continue to invaginate to form multivesicular bodies (MVBs), then MVBs fuse with the plasma membrane and release exosomes into the extracellular environment. Microvesicles are released directly from the plasma membrane to the outside of the cell. Apoptotic bodies are produced by apoptosis. bodies are vesicles formed by the decomposition of cell contents during the process of programmed cell death. After arriving at the receptor cell, EVs can release their cargo into the cytoplasm of recipient cells to regulate their biological activities. 9 In 2018, the International Society for Extracellular Vesicles (ISEV) put forward the point that, unless the source of EVs in this experiment can be clearly identified, the EV should be classified according to size, which are "small EVs" (sEVs, <200 nm) and "large EVs" (lEVs, >200 nm). 10

| Effects of T-EVs on endothelial cells
A previous study has found that ECs take up EVs through endocytosis and promote angiogenesis through the phosphoinositide 3-kinase/Akt signaling pathway. 11 We recently found that following clathrin-mediated endocytosis by ECs, T-sEVs are transported to the perinuclear region in a typical three-stage pattern. Importantly, T-sEVs frequently interact with and finally enter lysosomes, followed by a quick release of their carried miRNAs. 12 The proliferation, migration, and tube formation of ECs are finely controlled by T-EVs via multiple pathways. T-EV-mediated angiogenesis depends on their bioactive cargoes, such as proteins, miRNAs, and lncRNAs (Table 1).
Among these cargoes, miRNA is the most reported molecule in T-EV-mediated angiogenesis. The tumor promoter miRNA-23a is found in T-EVs secreted by some types of cancer cells. T-EV-derived miRNA-23a downregulates prolyl hydroxylase domain 2 (PHD2), a key oxygen sensor that negatively regulates hypoxia-inducible factor (HIF) protein, thus leading to enhance angiogenesis. 25 In addition, miRNA-23a targets and damages ZO-1, the EC junction protein, to increase vascular permeability. 37 Another study confirmed that miRNA-23a in NPC-derived EVs can promote angiogenesis by inhibiting TSGA10 (an anti-angiogenic factor). 38 The high level of miRNA-23a in EVs endows EVs with enhanced pro-angiogenic effects in vivo and in vitro. 26 MiR-143-3p and miR-145-5p are two tumor suppressors that inhibit the proliferation of colorectal cancer cells with mutated insulin-like growth factor 1 receptor (IGF1R). 39 Another study revealed that miR-143-3p and miR-145-5p within EVs derived from lung adenocarcinoma cells could increase tube formation in ECs by reducing the level of CAMK1D in ECs. 24 The production and content of EVs are susceptible to the growth environment of parent cells, including hypoxia. 40 Hypoxic or nor- MLL3 and DACT2 to promote angiogenesis. 53 The level of miR-181a in T-EVs under hypoxic has been shown to be higher than it is under normoxic in papillary thyroid cancer. 54 YAP, the Hippo pathway effector, works as a crucial signal transducer to mediate VEGF-VEGFR2 signaling during angiogenesis. 53 The binding of VEGF and VEGFR activates YAP/TAZ in ECs, leading to cytoskeletal remodeling, cell migration, and protein transmission, which is necessary for angiogenesis. YAP knockout will impair the transfer of VEGFR2 and change the distribution of VEGFR2, which influences angiogenesis. 53 In the hypoxic microenvironment, miR-23a carried by liver cancer cell-derived EVs can inhibit the SIRT1 in the ECs from inducing angiogenesis. 27 In addition, EVs released by NPC cells under hypoxia can carry HIF-1α, which are taken up by other cells, and then propagate hypoxic signals, 55 forming a cycle of hypoxic signal transduction.

| Effects of T-EVs on macrophages
Macrophages in the TME can differentiate into TAMs, and most of them exhibit the M2-like phenotype. The differentiation of macrophages is influenced by induction conditions and the type of cargoes in the T-EVs. 60 For example, EVs derived from head and neck cancer cells, which are enriched with adenosine A 2B receptor, mediate the differentiation of macrophage into an M2-like phenotype. 17 The tumor suppressor phosphatase and tensin homolog (PTEN) 61 and hypoxia are both considered to be critical factors within the TME. Under hypoxic conditions, EVs derived from ovarian cancer cells can induce macrophages to differentiate into an M2-like phenotype 62 and inhibit PTEN to upregulate some angiogenic factors (e.g., VEGF-A and Ang-1) in lung cancer. 46 Anx-II in EVs derived from BC cells was previously shown to activate macrophages to promote angiogenesis in a tPAdependent manner. In addition, macrophages create a pre-metastatic niche by releasing inflammatory factors such as IL-6 and TNFα. 63 The pre-metastatic niche destroys the integrity of the matrix and is conducive to the formation of tumor angiogenesis.

| Effects of TAM-derived EVs on angiogenesis
Clinical research has shown that the proportion of TAM is highly correlated with tumor angiogenesis, invasion, and prognosis in the center of the tumor and the tumor-infiltrating front. 64 Signaling molecules (primarily nucleic acids, such as miRNA-155-5p, miRNA-221-5p, miRNA-942, and miRNA-130b-3p) that are involved in the process of TAM-derived EVs mediated angiogenesis are listed in Table 3.

| Effects of MSC-derived EVs on angiogenesis
The precise effect of MSC-derived EVs on angiogenesis remains controversial currently. A previous study has shown that MSC-EVs suppress angiogenesis in BC. 68 Other studies have revealed that MSC-derived EVs, carrying miR-100, are internalized by BC cells, and that these EVs reduce the expression of VEGF by regulating the mTOR/HIFα axis, thereby inhibiting angiogenesis. 69 In contrast, some studies have shown that, in the TME of BC, MSC-EVs activate YAP and TAZ proteins to promote tumor angiogenesis and tumor progression through the Hippo signaling pathway. 70 Furthermore, external stress, such as hypoxia, 71 may promote MSC-EVs to induce angiogenesis. The disparity may result from differences between the cellular context and variation of the external microenvironment.

| Effects of CAF-derived EVs on angiogenesis
Fibroblast-derived EVs begin to play the pro-angiogenic role after fibroblasts are transdifferentiated into CAFs in the TME. The levels of T-box transcription factor 5 (TBX5) and Thioredoxin-interacting protein (TXNIP) were negatively correlated with angiogenesis, indicating their negative role in angiogenesis. In cervical cancer, miR-10a-5p from CAF-derived EVs promotes angiogenesis through the hedgehog signaling pathway by targeting TBX5 and inhibiting the expression of TBX5. 72 Another study showed that CAF-derived EVs deliver miR-135b-5p to colorectal cancer cells, thereby inhibiting TXNIP and promoting angiogenesis. 73 However, the mechanisms of CAFs-derived EVs promote angiogenesis remain to be explored, perhaps mainly determined by internal cargoes, including membrane proteins, cytosolic, nuclear proteins, and nucleic acids, especially non-coding RNAs. 74

| THE REL ATI ON S HIP B E T WEEN E VS AND A AT FAILURE
Several bioactive molecules (e.g., VEGF, PlGF, FGF, PDGF) are known to be involved in tumor angiogenesis. Among them, since VEGF is the primary regulator of angiogenesis and promotes proliferation, migration, and tube formation in ECs, VEGF/VEGFR has become the principal target for AAT. 75  Another study found that CAF-derived EVs in the tumor stroma promoted angiogenesis through the VEGFR2-AKT-ERK signaling pathway in ECs both in vivo and in vitro. However, the TA B L E 3 TAM-derived EVs in tumor angiogenesis and mechanisms that underlie these events.
VEGF bound to CAFs-derived EVs is not neutralized by bevacizumab, which may also be one of the reasons for the resistance of tumor AAT therapy. VEGF is anchored on the surface of EVs via the link of heparan sulfate proteoglycans (HSPGs). When heparinase was used to release VEGF bound to CAFs-derived EVs, the VEGF bound to bevacizumab increased, indicating that the use of heparanase can restore the sensitivity of tumor cells to bevacizumab to some degree. 78 Interestingly, bevacizumab had little effect on the number of glioblastoma-derived EVs. Annexin A2 is a protein that can promote angiogenesis and tumor progression.
The mechanism is that annexin A2 carried by T-EVs promotes EC migration and invasion in a tPA-dependent manner, which are important steps in angiogenesis. 63 The expression of annexin A2 was increased in glioblastoma-derived EVs after treatment with bevacizumab. Similarly, glioblastoma-derived EVs did not bind to bevacizumab, which appears to be one of the ways in which a glioblastoma protects itself from AAT. In addition to T-EVs, annexin A2 was also found in perivascular cell-derived EVs (TPC-EVs), which also play a role in tumor angiogenesis. 63 More interestingly, the experiment used the combination of GW4869, an inhibitor that reduces EV biogenesis, 79 and bevacizumab. The combination therapy substantially increased anti-tumor efficacy compared with bevacizumab use alone. 80 82 The general principle and situation are depicted in Figure 2. to EVs, which is not sensitive to anti-angiogenic drugs, and restore the sensitivity of anti-angiogenic drugs. As mentioned above, VEGF is not the only signal for T-EVs to mediate tumor angiogenesis, simply blocking the connection between T-EVs and VEGF may insufficiently improve the effect of AAT. Due to the important role of T-EVs in the resistance to AAT, the use of GW4869, an inhibition for the biogenesis and secretion of EVs, may serve as an effective strategy to enhance the efficiency of AAT in the future. The combination of T-EV inhibition and AAT may be a promising strategy to effectively suppress tumor growth in the clinic, yielding new methods to overcome the resistance of AAT. In addition, cargoes of EVs (e.g., Hsp90) may be used as biomarkers to predict susceptibility to AAT and assist in evaluating response.

| CON CLUS I ON AND PROS PEC TS
This review discussed the pro-angiogenic effects of EVs secreted from both tumor and non-tumor cells. The interrelation between EVs and tumor angiogenesis is complex and forms a specific network ( Figure 3). We also discussed the potential role of EVs in the resistance of AAT as well as the underlying mechanisms. Due to the important role of T-EVs in AAT resistance, strategies that target EVs, such as restoring the sensitivity of VEGF on EV surfaces to anti-angiogenic drugs or inhibiting the release of EVs, might be served as effective methods to enhance the efficiency of AAT in the future. In addition to T-EVs, EVs released by non-tumor cells, such as TAMs, can also be considered targets for AAT (e.g., induce M2-type macrophages into M1-type macrophages). Although EVs released by various cells induce tumor angiogenesis in vitro, strong empirical evidence in vivo is lacking. This suggests that there is still a long way from bench to bed, and the anti-tumor angiogenesis effect of targeting EVs is worthy of further study in vitro and in vivo.

AUTH O R CO NTR I B UTI O N S
Zi-Wu Ye and Zi-Li Yu contributed equally to this study. Zi-Wu Ye drafted the manuscript and drew the figures. Zi-Wu Ye and Zi-Li Yu discussed and revised the manuscript. Gang Chen and Jun Jia designed the study. All authors read and approved the final manuscript.

ACK N OWLED G M ENTS
Not applicable.

FU N D I N G I N FO R M ATI O N
National Key R&D Program of China (2019YFA0210500), National Natural Science Foundation of China (81922038, 81801842).

CO N FLI C T O F I NTE R E S T S TATE M E NT
All other authors declare that they have no competing interests.