Various cells can secrete exosomes. Han et al. have found that exosomes containing miR-145 secreted by tumor-associated stroma in pancreatic ductal adenocarcinoma can act on adjacent pancreatic ductal adenocarcinoma cells to inhibit cancer cell growth (46). Cancer-associated fibroblasts (CAFs) are the main components of the TIME. Studies have found that exosomes secreted by CAFs transmit information, such as miR-181D-5P, miR-21, miR-143, and miR-378e, to BC cells, which promote stem cell characteristics, epithelial-mesenchymal transition (EMT) phenotype, and non-anchoring cell growth of BC cells to encourage the progression of cancer (47–49). CAFs can also secrete exosomes carrying SNHG17 to mediate osteosarcoma proliferation and metastasis (50). Adipocytes also secrete exosomes, leading to increased cancer cell migration and invasion (51). Mesenchymal stem cells (MSCs) also exist in the TIME and have immunomodulatory properties. Mao et al. have shown that BM-MSCs from p53-deficient mice (p53 mBM-MSCs) can regulate the Wnt/β-catenin pathway and deliver UBr2 to target cells, thereby promoting the growth and metastasis of gastric cancer (GC) (52). Ma et al. have found that exosomes derived from BM-MSCs transfected with miR-221 significantly promote the proliferation, migration, invasion, and adhesion of GC BGC-823 and SGC-7901 cells (53). Human BM-MSC exosomal miR-375 inhibits glioma cell progression through solute carrier family 31 member 1(SLC31A1)inhibition (54).

Immune cell-derived exosomes in the tumor microenvironment also play an important role in the occurrence and development of cancers. As early as 2012, Luana et al. have proposed an important role of natural killer (NK) cell-derived exosomes in immune surveillance and homeostasis (55). Aladin and colleagues showed that NK cell-derived exosomes eliminated leukemia cells isolated from patients with leukemia (56). Zitvogel’s earlier study showed that DC cell-derived exosomes were as effective as DC cells in inducing tumor immune responses, which also provided theoretical support for cancer vaccine research (44). Romagnoli et al. indicated that breast cancer cells treated with DC cell-derived exosomes were more effective in stimulating T cell responses than breast cancer cells without exosome treatment (57). Cai and colleagues showed that CD8+ T cell-derived Exos can promote tumor cell invasion and lung metastasis through the Fas pathway (58).Xie et al. have demonstrated that Natural CD8+25+ regulatory T cell-secreted exosomes can inhibit CD8+ T cell responses and antitumor immunity (59). Wang et al. have shown that exosomes derived from Tumor-associated macrophages up-regulate miR-192-5p to promote endometrial cancer cells apoptosis and effectively suppress the progression of endometrial cancer (60). Exosomes secreted by polymorphonuclear (PMN) myeloid-derived suppressor cells (MDSCs), which are immunomodulatory cells present in the TIME, significantly inhibit NK cell-mediated anticancer activity, suggesting that the detection of PMN-MDSC content may have prognostic value (61). miR-326 in M1 macrophage exosomes promotes apoptosis of liver cancer cells and inhibits cancer proliferation and migration by downregulating the expression of NF-κB in liver cancer (62).

Cancers interact with various cells in their TIME. In the above review, we briefly discuss the role of non-tumor derived exosomes in tumors, next, we focused on the role of TDEs.

TDEs function in both autocrine and paracrine ways in the TIME. They can act on the cancer cells themselves or on the surrounding cells and tissues to affect cancer development, which has been confirmed in various cancers ( Table 1 ). Furthermore, TDEs can affect other non-cancer cells and participate in the transfer of metastatic capacity. For example, exosomes released by malignant BC cells are taken up by less malignant cancer cells located within the same cancer, thereby promoting the migration and metastatic potential of low-grade malignant cancer cells (100). In addition, TDEs can act on immune cells in the microenvironment to regulate cancer immunity and affect the balance between cancer promotion and cancer suppression in the TIME.

As mentioned above, exosomes contain various small molecular substances that mediate the occurrence and development of cancers. Therefore, targeting the production of exosomes is a strategy to increase or decrease certain small molecular substances to affect the development of cancers. Guan et al. have indicated that high-dose Proton-pump inhibitors (PPIs) suppress cancer size in cancer-bearing BALB/c nude mice, most likely mediated by inhibiting exosome release through suppressing ATP production (124). Huang and colleagues have found that the anticancer effect of lecithin is mediated by the reduction of exosome production and thus the downregulation of miR-1246 carried by them (125). Wang and colleagues have revealed that antigen-presenting cell (APC)-activated lncRNA (lncrNA-APC1) can directly regulate the stability of Rab5b mRNA to inhibit the production of exosomes in CRC cells, thereby suppressing angiogenesis and limiting cancer growth and proliferation (126). Similarly, promoting the secretion of exosomes can also achieve some anticancer effects. Xu et al. have indicated that hyperthermia can promote the secretion of exosomes through Rab7b, upregulate the transfer of FOS and CREB5 anticancer genes into drug-resistant cancer cells, and increase the sensitivity of BC cells to doxorubicin (Dox) (127). All these studies provide potential therapeutic targets. The RAB27A-targeting inhibitor tipifarnib (128), sphingomyelinase inhibitor GW4869 (129), manumycin A (130), and other drugs have been developed, and the clinical safety and efficacy of these drugs need to be further studied.

Exosomes can be loaded with both small and large molecules, supporting their use as therapeutic tools for treating various diseases, including cancer (131). The construction of engineered exosomes loaded with drugs provides a more efficient, less toxic and more targeted drug delivery strategy (132, 133). It has attracted attention due to its good biocompatibility (134, 135). Exosomes can be derived from milk, MSCs, HEK293 cells, DCs, red blood cells, and cancer cells (136). Agrawal et al. have suggested that milk-derived exosomes used for oral delivery of paclitaxel exhibit more marked inhibitory effects in vivo (137). Kanchanapally and colleagues have found that macrophage-derived adriamycin (DOX)-loaded exosomes are more cytotoxic than TDEs (132). This finding provides a basis for selecting raw materials for engineered exosomes. In addition, exosomes carry different peptides, tetraspanins and integrins to target specific cells, which provides strategies for the construction of more specific drug delivery (138–141). Cui et al. used a bone-targeting peptide to modify exosomes to deliver siShn3, which achieved the anti-osteoporosis effect of targeting osteoblasts (142). Compared with liposomes, exosomes have a stronger targeting effect, mainly because CD47 has the ability to evade the phagocytosis of mononuclear macrophages, which was confirmed in the study of Sushrut Kamerkar (143). Therefore, it is not difficult to think that artificial modification of exosomes can better enhance their targeting ability. Zhang et al. integrated hybrid membrane proteins from red blood cells and MCF-7 cancer cells into exosomes and designed an artificial chimeric exosome. Membrane proteins from red blood cells contain high CD47 levels and can inhibit phagocytosis, while membrane proteins from MCF-7 cancer cells contain specific adhesion proteins that can adhere to homologous cancer cells. This exosome obtained significant anti-cancer ability in mouse experiments (144). In addition, Qi et al. shown that superparamagnetic nanoparticle modified exosomes enhanced cancer targeting under external magnetic field and suppressed liver cancer in mice (145). Therefore, exosomes are highly valuable drug tools.

The exosome carrier strategy can also be combined with sonodynamic therapy and nanomedicine to give full play to the advantages of both sides. Wang et al. have developed Exo^Ce6+R848 by coincubating ultrasound-sensitive Chlorin e6 (Ce6) and R848 into exosomal membranes. Under ultrasound, EXO^Ce6+R848 stimulates robust maturation of cancer-infiltrating DCs, promotes M1 macrophage polarization and proinflammatory cytokine production, and decreases anti-inflammatory cytokines. This process further inhibits Tregs in the TIME and activates effector T cells for anticancer purposes (146). Although much attention has been paid to drugs, such as nanomedicines (147, 148), the exosome carrier strategy has attracted much attention due to controversial toxicity, immune rejection, and stability (149). Wu and colleagues have used catalase (CAT) and the sonosensitive agent indocyanine green-loaded silica nanoparticles (CAT@SiO2), which is known as a biodegradable nanoplatform (termed as CSI), to coat macrophage exosomes with CSI to form CSI@Ex-A. In addition, exosomes overcome the blood-brain barrier obstruction of the original CSI, making it more effective in clinical practice (150). Various methods exist to transfer anticancer drugs into exosomes, including coincubation, electroporation, and chemical transfection (151, 152) However, the transfer efficiency of multiple processes is different, and exploring more effective construction strategies requires more in-depth research.

Dual-targeting exosome immunotherapy based on exosome delivery to activate the immune system and inhibit immune checkpoints is a more effective emerging strategy. Fan et al. have suggested that by using a dual-targeting exosome-loaded drug, cGAMP (cGAMP@dual-anti-Exos), whose surface is modified with two antibodies (anti-PD-L1 and anti-CD40), two-step activation of DCs and blocking of cancer cell PD-L1, the survival rate of mice inoculated with the drug is up to 60%. In contrast, all other mice die after 40 days, indicating that the drug significantly improves cancer suppression (153). Wang et al. have used the homing ability of cancer cells and macrophages to develop a chimera that secretes outside the body. The cancer cell nucleus embedded within macrophages causes the secretion of macrophage cancer signals to carry body secretion. It relies on the homing ability and lymph nodes to achieve double targets, activates the body’s immune system, removes cancer immunosuppression, and increases anticancer ability. When combined with the immune checkpoint inhibitor PD-1 treatment, chimeric exosomes significantly prolong the survival of a mouse model of cancer recurrence (154).

PD-1/PD-L1 is a crucial target of immune checkpoint inhibitor therapy. PD-L1 in TDEs can induce PD-L1 expression on immune cells, such as neutrophils, inhibit T-cell activation, and promote cancer growth (78, 155). In immunotherapy, exosomes are involved in resistance to immune checkpoint inhibitor therapy targeting PD-L1. One possible mechanism is that exosomes deplete PD-L1 antibodies and increase resistance to PD-L1 therapy (156). In a study by Mauro and collaborators, anti-PD-L1 antibodies synergize with exosomal PD-L1 blockers to inhibit cancer growth. This finding suggests that exosomal PD-L1 is a novel target for relieving cancer drug resistance (45). Various strategies have been developed based on the above mechanisms to target exosomal PD-L1. Blocking any step in exosome production can reduce exosomal PD-L1 expression (157). Therefore, drugs that block exosome production can reduce exosome PD-L1. Shen et al. have shown that blockade of the LSD1 target specifically reduces exosomal PD-L1 expression without reducing intracellular PD-L1 expression and activates T-cell killing, providing a more precise choice of targets for inhibiting exosomal PD-L1 expression (158). In addition, increasing the elimination of exosomes can be therapeutic, such as in vitro ultrafiltration (159), while this approach also removes exosomes from normal cells, and its safety needs to be studied. Wang Min et al. have demonstrated that lactadherin promotes the clearance of circulating cancer cell-derived EVs, which also provides an exosome clearance mode (87). Direct inhibition of PD-L1 production is also a strategy. Shin et al. have reported that sulfamethoxazole effectively reduces exosomal PD-L1 content, thus allowing the anti-PD-L1 antibody to exert a potent anticancer effect. In addition, because inhibition of exosomal PD-L1 enhances T-cell immunity, it also increases the response rate to anti-PD-1 antibodies (160).

Exosomes have the potential to be used as cancer vaccines. As carriers of information transmission, exosomes carry information about the original cell and express complexes of MHC class I/II epitopes and costimulatory molecules. Therefore, they can also activate CD8 T cells. Moreover, exosomes have good biological stability and bioavailability, providing a basis for their application in cancer vaccines (44, 161–163). As early as 2004, Altieri and coworkers have reported that plasmacytoma cells release exosomes in vitro, and a single dose (5 µg) of exosomal protein combined with an isogenic vaccine can protect 80% of mice against wild-type cancer challenge, making exosome vaccines possible (164). There are two significant obstacles to the success of DC vaccines: cancer-mediated immunosuppression and the functional limitations of commonly used monocyte-derived DCs, which are effectively avoided by using cancer cell exosomes (165). It has been found that vaccination with DCexos (derived from BMDCs) has better anticancer efficacy than vaccination with DCs (44, 166).

Exosomes can be designed as vaccine adjuvants. Yildirim and colleagues have developed BC cell-derived exosomes carrying two immune adjuvants, the TLR9 ligand, K-type CpG ODN and a TLR3 ligand, p(I:C), to achieve more vigorous cellular and humoral anticancer immunity. In addition, it can inhibit cancer growth in mice (167). Experiments have shown that transfer of CCL22 siRNA into immunologically dead TDEs can inhibit Treg expansion and enhance anticancer effects. Better anticancer efficacy has been achieved when combined with chemotherapy drugs (168). Exosomes derived from M1-polarized macrophages can induce the release of a Th1 cytokine repertoire and antigen-specific cytotoxic T-cell responses, acting as immunopotentiating agents for cancer vaccines (169), most likely through downregulation of Wnt signaling (170). Kavitha et al. have demonstrated that exosomes from mouse embryonic stem cells engineered to produce GM-CSF activate a higher CD8+ T effector, Th1-cell response, and significantly block cancer growth (171). Huang et al. have developed an exosome loaded with Hiltonol (TLR3 agonist) and ICD (immunogenic cell death) inducer human neutrophil elastase (ELANE). In addition, the engineered Hiltonol-elane-α-la exosomes (HELA-Exos) are modified with α-lactalbumin on the exosome surface, which induces strong anticancer immunity (172). Phung and colleagues combined immune checkpoint blockade strategies with cancer vaccines, and modified anti-CTLA-4 antibodies on the surface of exosomes by lipid anchoring, which effectively promoted T cell targeting and increased tumor-specific T cell responses, leading to significant tumor growth inhibition (173). Modifying exosomes with MART-1 peptides to induce a stronger immune response (168). These strategies show the potential of exosomes as cancer vaccines to specifically activate anti-tumor immunity.