Tumorigenesis is not a single-step event, but a consequence of long-term alteration of mutations of genes and functional changes in the TME (85). Emerging evidence suggests that sEVs participate in the initiation, formation and remodeling of the TME in HCC (11). Chronic hepatitis B (CHB) remains a main factor responsible for HCC development, and sEVs are implicated in the spread, immune regulation and antiviral response of HBV infections (86). For example, exosomes from macrophages can deliver IFN-α-related microRNAs (miRNAs or miRs) to HBV-infected hepatocytes, and activate the antiviral response to suppress HBV replication and expression (87). The exosomal long non-coding RNA (lncRNA) HOTTIP has been shown to play a role in mediating the antiviral effect of tenofovir alafenamide following HBV infection (88). It has been demonstrated that sEVs from CD4⁺ T-cells can enhance B cell responses and potentiate the efficacy of the hepatitis B surface antigen vaccine (89). These studies indicate that sEVs can mediate immune regulation and antiviral response in HBV infection. A previous study also indicated that sEVs may exert negative immune regulatory effects, and that they are indispensable in the transformation from liver cirrhosis (LC) to liver cancer (90). The interplay between cancer cells and the TME is an essential activity that supports or prevents tumor development and progression. In HCC, tumor cells co-exist with other non-cancerous cells that constitute the TME and enhance tumor growth via various mechanisms. sEVs can exert an effect on the information and remodeling of the TME. For instance, exosomal miR-21 derived from HCC has been found to promote the conversion of hepatic stellate cells into cancer-associated fibroblasts (CAFs), and to facilitate the formation of the TME (91). Exosomal-miR-1247-3p from HCC cells has been shown to reduce the expression of B4GALT3 in CAFs and stabilize β1-integrin, leading to the activation of fibroblasts via the NF-κB signaling pathway (92). To summarize, sEVs play an essential role in the pathogenesis of HBV-related hepatic diseases, transformation from precancerous diseases to HCC and in the formation of the TME to confer tumorigenesis in HCC.

The development of HCC can be partly attributed to the rapid proliferation and uncontrolled expansion of tumor cells, which also accounts for tumor progression and resistance to therapy. sEVs mediate tumor growth and expansion by affecting the cell cycle, proliferation rate and apoptosis of HCC cells (93-95). Cao et al (71) suggested that exosomal miR-21 can influence HCC by altering the expression of the tumor suppressor genes, PTEN and PTEN pseudogene 1. Sun et al (96) indicated that exosome-specific miR-155 targeted PTEN and consequently stimulated the proliferation of HCC cells. On the contrary, certain sEV-encapsulated cargos, such as miR-338-3p can inhibit cell proliferation, induce cell apoptosis and consequently repress the progression of HCC (97). In addition, another study demonstrated that the proliferative and migratory abilities of HCC cell lines were potentiated, while their apoptosis was counteracted via the enforced expression of the exosomal lncRNA H19 (98). Furthermore, sEV constituents may intervene in the cell cycle to regulate the progression of HCC. It has been corroborated that circ_0061395 silencing can trigger cell cycle arrest and apoptosis, and suppress the proliferation of HCC in vitro, as well as inhibit tumor growth (56). Similarly, miR-4454 inhibitor-mediated exosomes can substantially exacerbate cycle arrest, apoptosis and the formation of reactive oxygen species in HCC (95). Of note, the progression of HCC is a result of the accumulation of several time-intersecting steps, including invasion, migration, angiogenesis, immune escape and metastasis, and sEV cargos may also function via several mechanisms. Huang et al (62) suggested that the silencing of circANTXR1 can suppress HCC progression, not only by inhibiting the proliferative ability of HCC, but also by hampering the migration, invasion and metastasis of tumor cells. The roles of sEVs in other hallmarks of tumor progression will be further discussed in the following section.

Angiogenesis refers to the formation of new blood vessels from pre-existing ones. It is a complex, multistep process involving extracellular matrix remodeling, endothelial cell migration and ultimately generation of microvessels. Angiogenesis not only provides sufficient oxygen and nutrition for cancer cells, but is also essential for HCC proliferation, local invasion and distant metastasis. The significance of sEVs in cancer angiogenesis has been widely explored and documented recently (98). In HCC, exosomal SNHG16 can sponge miR-4500 and activate angiogenesis in HUVECs by regulating polypeptide N-acetylgalactosaminyltransferase 1 via the PI3K/Akt/mTOR pathway (99). Lin et al (100) reported that tumor-derived exosomes (TDEs) containing miR-210 could target SMAD4 and STAT6 in endothelial cells, and thereby promote the angiogenesis of HCC. The functions of these sEV cargos are multifaceted, and they alter the gene expression of the recipient cells, which become more aggressive and exhibit malignant characteristics. Apart from regulating angiogenesis, they may also control phenotypic changes, such as the proliferative or migratory abilities of cancer cells. A previous study demonstrated that circCMTM3-bearing sEVs can drive the angiogenesis of HUVECs, as well as their viability, migration and invasion (58). Certain studies have found that serval sEV cargos may play the opposite role and suppress angiogenesis in HCC (34,101). For instance, HCC-derived exosomes containing miR-3682-3p have been shown to attenuate angiogenesis via targeting ANGPT1, which is dependent on the RAS-MEK1/2-ERK1/2 pathway (101). Taken together, these results indicate that angiogenesis is a complex process that is orchestrated by multiple biological factors, and the treatment of angiogenesis may provide a novel prospective therapeutic approach for HCC.

Widespread metastasis in patients with HCC remains a major challenge for treatment, and a main reason for treatment failure, as well as one of the leading-causes of cancer-associated mortality (102). Metastasis is a multistep process involving EMT, invasion into vessels, intravascular transport and organ-specific seeding. The most common mode of metastasis in HCC is intrahepatic metastasis, followed by lymphatic metastasis and distant metastasis to the lungs. sEVs are involved in multiple steps of HCC metastasis, and the importance of sEVs in HCC metastasis has recently been widely reported. Firstly, sEVs contribute to the EMT of HCC cells. Yang et al (103) demonstrated that exosomal miR-92a-3p from high-metastatic HCC cell lines can potentiate EMT and metastasis by inactivating PTEN and activating Akt/Snail signaling. Similarly, Chen et al (104) suggested that TDEs from HCC cells can accelerate EMT, and induce HCC progression and recurrence by activating the MAPK/ERK signaling pathway; however, those studies did not clarify the specific sEVs-carrying cargo that is involved in this process. sEVs exacerbate the migratory and invasive abilities of HCC, which may promote the metastasis of HCC. It has been observed that miR-374a-5p in exosomes potentiates the migration and invasion of HCC by regulating growth arrest and DNA damage inducible alpha (105). In addition, sEVs can orchestrate the organotropic metastasis of HCC by converting the pre-metastatic microenvironment into a tumor cell-friendly site. A previous study demonstrated that Exo-miR-1247-3p derived from HCC can trigger β1-integrin-NF-κB signaling in fibroblasts in the lungs, and is positively associated with several pro-inflammatory cytokines, such as IL-8 and IL-6, which promote the lung metastasis of HCC (92). Recent studies have identified numerous sEV cargos that are involved in the metastasis of HCC, including FAM138B (106), α(M) β(2) integrin (107), hsa_circ_0074854 (108) and lncRNA TUC339 (109). It should be noted that the aforementioned sEV cargos may not only participate in one step of metastasis, but may play multifaceted roles in the whole process of metastasis. For example, Fang et al (42) indicated that, apart from promoting EMT and enhancing tumor motility in vitro, HCC cells can secrete sEV-encapsulated miR-103, which also potentiates vascular permeability and lung metastasis in mouse models.

The immune system plays a paramount role in recognizing and eliminating malignant cells and foreign invaders. In the processes of tumor initiation and progression, aberrant proliferation and gene alteration in cancer cells can generate abnormally expressed antigens, which should be adequately presented, recognized and eliminated by the immune system (110). During their fight against the immune system, cancer cells also evolve, and may acquire the ability to evade immunosurveillance via various mechanisms. Among the immune escape effects, that contribute to cancer progression and drug resistance, engagement of the attenuation or abrogation of immunocytes is worth mentioning, and sEVs play a pivotal role in this process (111). Recent studies have suggested that HCC-derived sEVs can impair the function of natural killer (NK) and T-cells, as well as activate immuno-suppressive cells such as M2 macrophages. For instance, exosomal circUHRF1 from HCC triggers the exhaustion of NK cells and subsequently induces resistance to therapy (52). Exo-lncRNA TUC339 has also been shown to be internalized by macrophages, and to modulate M1/M2 polarization and suppress the antitumor immune response in HCC (109). Apart from diminishing the activity of the innate immune system, previous studies have indicated that TDEs from HCC can also impede the activation and function of specific immunocytes such as T- and B-cells (32,46,52). Wang et al (112) found that exosomal 14-3-3ζ released by HCC cells suppressed the antineoplastic characteristics of tumor-infiltrating T-lymphocytes. Tumor-derived exosomal HMGB1 can enhance the expansion the T-cell Ig and mucin domain (TIM)-1 (+) regulatory B-cells and facilitate HCC immune evasion (113). Collectively, sEVs play multiple roles in the communication of HCC and immune cells, and are critical for the immune escape of HCC cells and tumor progression. Thus, sEVs may serve as ideal therapeutic targets for HCC, although further investigations into this matter are warranted.

HCC is one of the most aggressive cancer types, and hepatic resection remains the gold standard of treatment for HCC if the patients can withstand surgery. For patients who experience HCC recurrence or cannot tolerate surgery, targeted therapies involving the use of sorafenib, a multi-kinase inhibitor compound, and chemotherapy including paclitaxel and 5-fluorouracil (5-FU) are first-line treatments. Resistance to drugs remains a main obstacle to the effective treatment of these patients. The mechanisms underlying drug resistance remain complex and elusive; however, but the roles of sEVs in this process is emerging and have captured the interest of researchers. For instance, a previous study found that transfection with GRP78 small interfering RNA into bone marrow-derived mesenchymal stem cells could yield sEVs containing siGRP78, thus mediating targeted RNA silencing, which increased the sensitivity of drug-resistant cancer cells to sorafenib and improved the drug resistance reversion (114). Furthermore, as previously demonstrated, sEV-encapsulated miR-23a/b derived from adipocytes was transferred to neighboring HCC cells, which enhanced their chemoresistance to 5-FU by targeting the VHL/HIF axis (115). The upregulation of miR-32-5p-bearing sEVs has been shown to induce multidrug resistance by potentiating EMT and angiogenesis via targeting the PI3K/PTEN/Akt signaling pathway (34). Another study demonstrated that sEVs secreted from cancer stem cells induced regorafenib insensitivity by upregulating Nanog expression (116). These findings highlight the significance of sEVs in the drug resistance of HCC, which results from sEVs directly suppressing drug efficacy against tumor cells, or from sEVs regulating the gene expression of recipient cells to facilitate cancer survival. Since sEVs can alter the drug sensitivity of HCC cells, it is tempting to engineer a sEV-derived vehicle to deliver specific agents for HCC treatment (117). Studies on this topic are currently underway, and the preliminary results are promising. Wang et al (118) indicated that the downregulation of miR-744 in HCC tissue and cell lines was implicated in the chemoresistance to sorafenib, and HCC cell lines treated with miR-744-upregulating sEVs were more sensitive to sorafenib, which provides a potential approach to reduce the occurrence of drug resistance.

Despite the advances in the diagnosis of HCC, the number of new cases and cancer mortalities associated with HCC remain high (1). The diagnosis of HCC relies heavily on imaging analyses, such as magnetic resonance imaging or computed tomography; however, the diagnosis of the majority of patients is confirmed at an advanced stage, and thus these patients miss the optimal treatment period (119). Immense efforts have been made as regards the early diagnosis of HCC, although success has been limited. Alpha-fetoprotein (AFP) is a traditional HCC marker with a low specificity, which has a limited value in the differential diagnosis between HCC and other liver diseases (120,121). As regards other biomarkers, such as golgi glycoprotein 73, AFP-L3, phosphatidylinositol proteoglycan 3 and decarboxylated prothrombin, they do not provide any obvious advantage in the early diagnosis of HCC compared with AFP (122,123). Therefore, a non-invasive method with a high diagnostic sensitivity and specificity is urgently required. Recently, liquid biopsies and particularly circulating tumor cells, have attracted extensive interest for the diagnosis and monitoring of HCC (124). Studies on other tumor-derived components, such as circulating tumor DNA, sEVs and serum miRNAs are also increasing (125-127).

sEVs have the following advantages: i) Due to being protected by the sEV membrane, sEV cargos have a high stability and cannot easily degraded by lysosomes; ii) since the secretion of sEVs is a normal physiological event for tumor cells, sEVs can be detected in the majority of fluids, and their extraction is relatively non-invasive; iii) compared with plasma biomarkers, bioactive molecules from sEVs contain less interference of plasma; and iv) markedly, cargos of sEVs have extensive homology with recipient cells, which can confer sEVs superior sensitivity and specificity than traditional methods (128-130). Differential ultracentrifugation is the most common method used to separate sEVs from the cell culture medium. However, each biological fluid presents specific biophysical and chemical characteristics that render it different from culture conditioned medium. Despite current mainstream commercial kits are based on precipitation, which may result in EV populations bound to or mixed with introduced components, such as antibodies, beads or polymers; the majority of studies on the potential clinical applications of sEVs use this method as it is user-friendly, cost-effective and has potential for scale-up production (17). Furthermore, the efficiency and repeatability of sEVs separated using the ExoQuick^™ kit have been demonstrated to be comparable with those of differential ultracentrifugation (131). Therefore, the articles cited in the current section include those that using commercial kits to isolate sEVs.

Numerous studies on the role of sEVs as HCC promising biomarkers have been conducted (132-136). The importance of sEVs as HCC biomarkers is reflected in numerous aspects, including the fact that sEV cargos may serve as biomarkers for the early detection of HCC; among these sEV cargos, miRNAs are the most extensively investigated ones. For instance, the expression of miR-21 and miR-10b in sEVs is markedly increased in patients with HCC compared with that of healthy individuals and patients with CHB, indicating that sEVs-carrying miR-21 and miR-10b may be used as early diagnostic biomarkers for HCC (80). Similarly, by comparison with that of patients with LC, the expression of miR-221, miR-192 and miR-146a in exosomes was increased in patients with HCC, and Fründt et al (137) indicated that sEVs carrying miR-146a could distinguish patients with HCC from patients with LC with an area under the curve value of 0.80±0.14 in a logistic regression model, and miR-96, miR-122, miR-200a had similar effect (138-140). Other sEV cargos, such as proteins and other non-coding RNAs (ncRNAs), including lncRNAs and circRNAs, may also play a role in the preliminary diagnosis of HCC, and it has been observed that LINC00161, circRNA 0006602, LDHC, sphingosines, dilysocardiolipins, lysophosphatidylserines, and (O-acyl)-1-hydroxy fatty acids are early diagnostic biomarker candidates (48,121,141-143).

Apart from their early diagnostic value, sEV cargos may be involved in the prediction of tumor staging and metastasis (144-147). Exo-miR-1307-5p expression in plasma has been found to be positively associated with tumor stage and progression, while sEVs carrying miR-125b have been shown to possess anti-metastatic features and are indicators of early metastasis in HCC (148-150). Other sEV cargos can play a similar role in tumor staging or metastasis prediction, and the function of lncRNA ATB, hnRNPH1 and ASMTL-AS1 in this regard has been reported (151-153). In addition, certain sEV cargos may serve as prognostic indicators and may predict the prognosis of patients with HCC. It has been corroborated that miR-638, miR-150-3p, lncRNA CRNDE and circAKT3 in the sEVs of patients with HCC are implicated in overall survival and disease-free survival and may serve as independent indicators of a poor prognosis (31,154-157). It should be noted that the abnormal expression of certain sEV cargos, such as miR-718, miR-125b and miR-92b is not only an effective tool to evaluate survival, but is also a potential marker to predict the recurrence of HCC (32,149,158). Notably, sEVs-carrying miR-122, hsa-circRNA-G004213 and DANCR are also potential markers to evaluate the efficacy of HCC surgical and interventional treatment (153,159-161). In addition to the above, a panel of tumor specific biochemical indicators has been proposed for a higher sensitivity and specificity compared with single one (162-166). For example, Sorop et al (167) established exosomal miR HCC Score including serum AFP and the level of plasma sEVs-carrying miR-21-5p and miR-92a-3p with a great diagnostic ability of HCC (AUC=0.85). Taken together, previous studies have demonstrated that certain sEV cargos, including ncRNAs, mRNAs, lipids and proteins, may serve as potential HCC diagnostic and prognostic biomarkers. The potential HCC biomarkers are summarized in Table II and the efficacy of these candidates warrants further validation.

Liver cancer is ranked as the third leading cause of cancer-associated mortality worldwide due to resistance to traditional drugs, as well as diagnosis in the late stage (1,168). The identification of novel drugs for targeted therapy is imperative for patients with HCC (169). As aforementioned, the distinctive property of sEVs in delivering functional molecules and altering the biological behavior of recipient cells highlights their potential application as ideal therapeutic vehicles in cancer therapy, both at the theoretical and practical level. The development of engineered sEVs, with a purpose of acting as alternatives to chemotherapeutic and targeted agents, is currently ongoing. sEVs have several advantages compared with previous drug carriers, such as liposomes: First, sEVs achieve highly efficient drug delivery due to their facility of penetrating biological barriers. In addition, the cellular origin of sEVs makes them well tolerated, and they can easily escape immune clearance, which also reduces drug dose and toxicity (111,170,171); Furthermore, the heterogeneity of proteins on the sEVs membrane facilitates the targeting abilities of sEVs. In addition, sEVs are more biocompatible, safe and stable than liposomes (172,173). Taken together, sEVs have a great potential to serve as nano-carriers in the treatment field. The present section mainly focuses on the advancements made in sEV research as regards their application in therapy.

To achieve a better understanding of sEVs as vehicles for therapeutic agents, methods of sEV preparation, the engineering of sEVs and the selection of cargos are under investigation, and the preliminary results are promising. Multiple therapeutic agents, including chemotherapeutic drugs and nucleic acids or their inhibitors, can be loaded. For example, in a previous study, the subcutaneous injection of sEVs containing miR-let-7a into a breast cancer mouse model exhibited an antitumor ability by targeting EGFR (174). In HCC, recent research has validated the importance of sEVs in the delivery of EV-packaged drugs (175). A previous study provided a prospective approach for generating sEV-associated adeno-associated virus containing inducible caspase 9 (iCasp9) suicide gene (Vexo-AAV6-iCasp9). The engineered sEVs possessed a low immunogenicity and toxicity, and were readily absorbed by HCC cells, consequently increasing HCC regression in an in vivo xenograft model (176). Another study encapsulated erastin (a typical ferroptosis inducer) and rose bengal (RB, a well-known photosensitizer) into sEVs and engineered CD47 on the surface of sEVs to protect the designed sEVs from phagocytosis by macrophages. The sEVs induced obvious ferroptosis in HCC, with minimized toxicity in the liver and kidneys (177). Apart from packing antitumor payloads into sEVs, previous studies have developed nanoparticles targeting specific adhesion or receptor proteins on the surface of sEVs membranes for targeted delivery. Tian et al (80) designed a nano-drug based on the PDCM system by targeting sEVs shuttling miR-21 and miR-10b, which markedly decreased HCC growth and the numbers of metastatic lung nodules. TDEs not only assist in exacerbating tumor progression, but also increase the resistance of cancer cells to antitumor treatments (178). Sorafenib and transarterial chemoembolization have been considered optional treatments for terminal-stage HCC for numerous years. Due to acquired drug resistance to commonly used chemotherapeutic agents, the clinical outcome and overall survival of patients with HCC remain unsatisfactory (179). The expression of programmed cell death protein 1 (PD-1) in HCC tissues from patients with HCC who accepted sorafenib treatment was upregulated and induced T-cell apoptosis (180). Therefore, immune checkpoint inhibitors, such as commonly used PD-1 antibodies and PD-L1 anti-bodies have been introduced into medical practice as part of the HCC regimen; however, the efficacy of the combination of sorafenib and immunotherapy has not yet been fully elucidated. Shi et al (180) treated mouse models of HCC with the triple treatment of sorafenib, PD-1 antibodies and DC-derived exosomes (DEXs), which markedly prolonged the survival time mice with HCC by comparison with the mice treated with sorafenib alone, DEXs alone, or the combination of DEXs and sorafenib. Taken together, these preclinical studies offer encouragement for the application of sEVs as vehicles for HCC treatment.

Despite the rapid development of advanced techniques, major limitations remain in the current understanding of sEVs vs. ideal treatment scenarios: i) For sEV technology to play a role as a drug delivery vector, it is necessary to ensure high purity and adequate production. There are still several obstacles which hamper the efficacy of current methods, such as being time consuming, having a high cost and generating polluting by-products. Oncologists increase the total yield of sEVs by intracellular calcium production, external stress, cytoskeletal blocking, drug stimulation and the induction of gene expression factors. Furthermore, sEVs usually represent heterogeneous populations from different cell sources, and no standard separation process has been established to date to achieve product consistency (181); ii) the efficient incorporation of external antitumor agents and molecules is another demanding challenge that needs to be optimized. The high drug-loading content in sEVs must be sufficient to obtain a therapeutic response. Several approaches, including transfection, electroporation and sonication (182,183), can be applied to upload the desired biomolecules into sEVs. However, it is difficult to ensure the integrity and biostability of the plasma membrane and the function of sEVs; iii) the current evidence for the ability of sEVs to deliver specific messages derives from cell culture studies. The biodistribution and tissue or cellular tropism in vivo will determine the application of therapeutic sEVs in clinical practice. At present, there is insufficient evidence for in vivo and clinical applications, which is a critical topic for future research in this area. Due to being subjected to elimination by the mononuclear phagocyte system, the half-life of sEVs in the systemic circulation is relatively short (184). Thus, further studies on the balance between prolonged circulation time and increased risk of toxicity on major organs are warranted; and iv) currently, sEVs need to be stored under-20 and -80°C in phosphate buffer saline (185). Therefore, identifying a suitable storage method is one of the barriers to be overcome.

In summary, while the application of sEVs as a therapeutic drug delivery system remains in its infancy, the deeper understanding of the aforementioned obstacles will provide a new orientation for cancer nanomedicine and immunotherapy.